diff options
Diffstat (limited to 'fs')
439 files changed, 23137 insertions, 9280 deletions
diff --git a/fs/9p/v9fs.h b/fs/9p/v9fs.h index 698c43dd5dc8..9defa12208f9 100644 --- a/fs/9p/v9fs.h +++ b/fs/9p/v9fs.h @@ -179,16 +179,13 @@ extern int v9fs_vfs_rename(struct mnt_idmap *idmap, struct inode *old_dir, struct dentry *old_dentry, struct inode *new_dir, struct dentry *new_dentry, unsigned int flags); -extern struct inode *v9fs_inode_from_fid(struct v9fs_session_info *v9ses, - struct p9_fid *fid, - struct super_block *sb, int new); +extern struct inode *v9fs_fid_iget(struct super_block *sb, struct p9_fid *fid); extern const struct inode_operations v9fs_dir_inode_operations_dotl; extern const struct inode_operations v9fs_file_inode_operations_dotl; extern const struct inode_operations v9fs_symlink_inode_operations_dotl; extern const struct netfs_request_ops v9fs_req_ops; -extern struct inode *v9fs_inode_from_fid_dotl(struct v9fs_session_info *v9ses, - struct p9_fid *fid, - struct super_block *sb, int new); +extern struct inode *v9fs_fid_iget_dotl(struct super_block *sb, + struct p9_fid *fid); /* other default globals */ #define V9FS_PORT 564 @@ -230,27 +227,9 @@ v9fs_get_inode_from_fid(struct v9fs_session_info *v9ses, struct p9_fid *fid, struct super_block *sb) { if (v9fs_proto_dotl(v9ses)) - return v9fs_inode_from_fid_dotl(v9ses, fid, sb, 0); + return v9fs_fid_iget_dotl(sb, fid); else - return v9fs_inode_from_fid(v9ses, fid, sb, 0); -} - -/** - * v9fs_get_new_inode_from_fid - Helper routine to populate an inode by - * issuing a attribute request - * @v9ses: session information - * @fid: fid to issue attribute request for - * @sb: superblock on which to create inode - * - */ -static inline struct inode * -v9fs_get_new_inode_from_fid(struct v9fs_session_info *v9ses, struct p9_fid *fid, - struct super_block *sb) -{ - if (v9fs_proto_dotl(v9ses)) - return v9fs_inode_from_fid_dotl(v9ses, fid, sb, 1); - else - return v9fs_inode_from_fid(v9ses, fid, sb, 1); + return v9fs_fid_iget(sb, fid); } #endif diff --git a/fs/9p/v9fs_vfs.h b/fs/9p/v9fs_vfs.h index 0e8418066a48..7923c3c347cb 100644 --- a/fs/9p/v9fs_vfs.h +++ b/fs/9p/v9fs_vfs.h @@ -40,13 +40,16 @@ extern struct kmem_cache *v9fs_inode_cache; struct inode *v9fs_alloc_inode(struct super_block *sb); void v9fs_free_inode(struct inode *inode); -struct inode *v9fs_get_inode(struct super_block *sb, umode_t mode, - dev_t rdev); void v9fs_set_netfs_context(struct inode *inode); int v9fs_init_inode(struct v9fs_session_info *v9ses, - struct inode *inode, umode_t mode, dev_t rdev); + struct inode *inode, struct p9_qid *qid, umode_t mode, dev_t rdev); void v9fs_evict_inode(struct inode *inode); -ino_t v9fs_qid2ino(struct p9_qid *qid); +#if (BITS_PER_LONG == 32) +#define QID2INO(q) ((ino_t) (((q)->path+2) ^ (((q)->path) >> 32))) +#else +#define QID2INO(q) ((ino_t) ((q)->path+2)) +#endif + void v9fs_stat2inode(struct p9_wstat *stat, struct inode *inode, struct super_block *sb, unsigned int flags); void v9fs_stat2inode_dotl(struct p9_stat_dotl *stat, struct inode *inode, diff --git a/fs/9p/vfs_dir.c b/fs/9p/vfs_dir.c index 4102759a5cb5..e0d34e4e9076 100644 --- a/fs/9p/vfs_dir.c +++ b/fs/9p/vfs_dir.c @@ -127,7 +127,7 @@ static int v9fs_dir_readdir(struct file *file, struct dir_context *ctx) } over = !dir_emit(ctx, st.name, strlen(st.name), - v9fs_qid2ino(&st.qid), dt_type(&st)); + QID2INO(&st.qid), dt_type(&st)); p9stat_free(&st); if (over) return 0; @@ -184,7 +184,7 @@ static int v9fs_dir_readdir_dotl(struct file *file, struct dir_context *ctx) if (!dir_emit(ctx, curdirent.d_name, strlen(curdirent.d_name), - v9fs_qid2ino(&curdirent.qid), + QID2INO(&curdirent.qid), curdirent.d_type)) return 0; diff --git a/fs/9p/vfs_inode.c b/fs/9p/vfs_inode.c index 32572982f72e..360a5304ec03 100644 --- a/fs/9p/vfs_inode.c +++ b/fs/9p/vfs_inode.c @@ -253,9 +253,12 @@ void v9fs_set_netfs_context(struct inode *inode) } int v9fs_init_inode(struct v9fs_session_info *v9ses, - struct inode *inode, umode_t mode, dev_t rdev) + struct inode *inode, struct p9_qid *qid, umode_t mode, dev_t rdev) { int err = 0; + struct v9fs_inode *v9inode = V9FS_I(inode); + + memcpy(&v9inode->qid, qid, sizeof(struct p9_qid)); inode_init_owner(&nop_mnt_idmap, inode, NULL, mode); inode->i_blocks = 0; @@ -332,36 +335,6 @@ error: } /** - * v9fs_get_inode - helper function to setup an inode - * @sb: superblock - * @mode: mode to setup inode with - * @rdev: The device numbers to set - */ - -struct inode *v9fs_get_inode(struct super_block *sb, umode_t mode, dev_t rdev) -{ - int err; - struct inode *inode; - struct v9fs_session_info *v9ses = sb->s_fs_info; - - p9_debug(P9_DEBUG_VFS, "super block: %p mode: %ho\n", sb, mode); - - inode = new_inode(sb); - if (!inode) { - pr_warn("%s (%d): Problem allocating inode\n", - __func__, task_pid_nr(current)); - return ERR_PTR(-ENOMEM); - } - err = v9fs_init_inode(v9ses, inode, mode, rdev); - if (err) { - iput(inode); - return ERR_PTR(err); - } - v9fs_set_netfs_context(inode); - return inode; -} - -/** * v9fs_evict_inode - Remove an inode from the inode cache * @inode: inode to release * @@ -384,82 +357,40 @@ void v9fs_evict_inode(struct inode *inode) #endif } -static int v9fs_test_inode(struct inode *inode, void *data) -{ - int umode; - dev_t rdev; - struct v9fs_inode *v9inode = V9FS_I(inode); - struct p9_wstat *st = (struct p9_wstat *)data; - struct v9fs_session_info *v9ses = v9fs_inode2v9ses(inode); - - umode = p9mode2unixmode(v9ses, st, &rdev); - /* don't match inode of different type */ - if (inode_wrong_type(inode, umode)) - return 0; - - /* compare qid details */ - if (memcmp(&v9inode->qid.version, - &st->qid.version, sizeof(v9inode->qid.version))) - return 0; - - if (v9inode->qid.type != st->qid.type) - return 0; - - if (v9inode->qid.path != st->qid.path) - return 0; - return 1; -} - -static int v9fs_test_new_inode(struct inode *inode, void *data) -{ - return 0; -} - -static int v9fs_set_inode(struct inode *inode, void *data) -{ - struct v9fs_inode *v9inode = V9FS_I(inode); - struct p9_wstat *st = (struct p9_wstat *)data; - - memcpy(&v9inode->qid, &st->qid, sizeof(st->qid)); - return 0; -} - -static struct inode *v9fs_qid_iget(struct super_block *sb, - struct p9_qid *qid, - struct p9_wstat *st, - int new) +struct inode *v9fs_fid_iget(struct super_block *sb, struct p9_fid *fid) { dev_t rdev; int retval; umode_t umode; - unsigned long i_ino; struct inode *inode; + struct p9_wstat *st; struct v9fs_session_info *v9ses = sb->s_fs_info; - int (*test)(struct inode *inode, void *data); - if (new) - test = v9fs_test_new_inode; - else - test = v9fs_test_inode; - - i_ino = v9fs_qid2ino(qid); - inode = iget5_locked(sb, i_ino, test, v9fs_set_inode, st); - if (!inode) + inode = iget_locked(sb, QID2INO(&fid->qid)); + if (unlikely(!inode)) return ERR_PTR(-ENOMEM); if (!(inode->i_state & I_NEW)) return inode; + /* * initialize the inode with the stat info * FIXME!! we may need support for stale inodes * later. */ - inode->i_ino = i_ino; + st = p9_client_stat(fid); + if (IS_ERR(st)) { + retval = PTR_ERR(st); + goto error; + } + umode = p9mode2unixmode(v9ses, st, &rdev); - retval = v9fs_init_inode(v9ses, inode, umode, rdev); + retval = v9fs_init_inode(v9ses, inode, &fid->qid, umode, rdev); + v9fs_stat2inode(st, inode, sb, 0); + p9stat_free(st); + kfree(st); if (retval) goto error; - v9fs_stat2inode(st, inode, sb, 0); v9fs_set_netfs_context(inode); v9fs_cache_inode_get_cookie(inode); unlock_new_inode(inode); @@ -470,23 +401,6 @@ error: } -struct inode * -v9fs_inode_from_fid(struct v9fs_session_info *v9ses, struct p9_fid *fid, - struct super_block *sb, int new) -{ - struct p9_wstat *st; - struct inode *inode = NULL; - - st = p9_client_stat(fid); - if (IS_ERR(st)) - return ERR_CAST(st); - - inode = v9fs_qid_iget(sb, &st->qid, st, new); - p9stat_free(st); - kfree(st); - return inode; -} - /** * v9fs_at_to_dotl_flags- convert Linux specific AT flags to * plan 9 AT flag. @@ -633,7 +547,7 @@ v9fs_create(struct v9fs_session_info *v9ses, struct inode *dir, /* * instantiate inode and assign the unopened fid to the dentry */ - inode = v9fs_get_new_inode_from_fid(v9ses, fid, dir->i_sb); + inode = v9fs_get_inode_from_fid(v9ses, fid, dir->i_sb); if (IS_ERR(inode)) { err = PTR_ERR(inode); p9_debug(P9_DEBUG_VFS, @@ -761,10 +675,8 @@ struct dentry *v9fs_vfs_lookup(struct inode *dir, struct dentry *dentry, inode = NULL; else if (IS_ERR(fid)) inode = ERR_CAST(fid); - else if (v9ses->cache & (CACHE_META|CACHE_LOOSE)) - inode = v9fs_get_inode_from_fid(v9ses, fid, dir->i_sb); else - inode = v9fs_get_new_inode_from_fid(v9ses, fid, dir->i_sb); + inode = v9fs_get_inode_from_fid(v9ses, fid, dir->i_sb); /* * If we had a rename on the server and a parallel lookup * for the new name, then make sure we instantiate with @@ -1187,26 +1099,6 @@ v9fs_stat2inode(struct p9_wstat *stat, struct inode *inode, } /** - * v9fs_qid2ino - convert qid into inode number - * @qid: qid to hash - * - * BUG: potential for inode number collisions? - */ - -ino_t v9fs_qid2ino(struct p9_qid *qid) -{ - u64 path = qid->path + 2; - ino_t i = 0; - - if (sizeof(ino_t) == sizeof(path)) - memcpy(&i, &path, sizeof(ino_t)); - else - i = (ino_t) (path ^ (path >> 32)); - - return i; -} - -/** * v9fs_vfs_get_link - follow a symlink path * @dentry: dentry for symlink * @inode: inode for symlink diff --git a/fs/9p/vfs_inode_dotl.c b/fs/9p/vfs_inode_dotl.c index 3505227e1704..ef9db3e03506 100644 --- a/fs/9p/vfs_inode_dotl.c +++ b/fs/9p/vfs_inode_dotl.c @@ -52,78 +52,33 @@ static kgid_t v9fs_get_fsgid_for_create(struct inode *dir_inode) return current_fsgid(); } -static int v9fs_test_inode_dotl(struct inode *inode, void *data) -{ - struct v9fs_inode *v9inode = V9FS_I(inode); - struct p9_stat_dotl *st = (struct p9_stat_dotl *)data; - - /* don't match inode of different type */ - if (inode_wrong_type(inode, st->st_mode)) - return 0; - - if (inode->i_generation != st->st_gen) - return 0; - - /* compare qid details */ - if (memcmp(&v9inode->qid.version, - &st->qid.version, sizeof(v9inode->qid.version))) - return 0; - - if (v9inode->qid.type != st->qid.type) - return 0; - - if (v9inode->qid.path != st->qid.path) - return 0; - return 1; -} - -/* Always get a new inode */ -static int v9fs_test_new_inode_dotl(struct inode *inode, void *data) -{ - return 0; -} - -static int v9fs_set_inode_dotl(struct inode *inode, void *data) -{ - struct v9fs_inode *v9inode = V9FS_I(inode); - struct p9_stat_dotl *st = (struct p9_stat_dotl *)data; - - memcpy(&v9inode->qid, &st->qid, sizeof(st->qid)); - inode->i_generation = st->st_gen; - return 0; -} - -static struct inode *v9fs_qid_iget_dotl(struct super_block *sb, - struct p9_qid *qid, - struct p9_fid *fid, - struct p9_stat_dotl *st, - int new) +struct inode *v9fs_fid_iget_dotl(struct super_block *sb, struct p9_fid *fid) { int retval; - unsigned long i_ino; struct inode *inode; + struct p9_stat_dotl *st; struct v9fs_session_info *v9ses = sb->s_fs_info; - int (*test)(struct inode *inode, void *data); - - if (new) - test = v9fs_test_new_inode_dotl; - else - test = v9fs_test_inode_dotl; - i_ino = v9fs_qid2ino(qid); - inode = iget5_locked(sb, i_ino, test, v9fs_set_inode_dotl, st); - if (!inode) + inode = iget_locked(sb, QID2INO(&fid->qid)); + if (unlikely(!inode)) return ERR_PTR(-ENOMEM); if (!(inode->i_state & I_NEW)) return inode; + /* * initialize the inode with the stat info * FIXME!! we may need support for stale inodes * later. */ - inode->i_ino = i_ino; - retval = v9fs_init_inode(v9ses, inode, + st = p9_client_getattr_dotl(fid, P9_STATS_BASIC | P9_STATS_GEN); + if (IS_ERR(st)) { + retval = PTR_ERR(st); + goto error; + } + + retval = v9fs_init_inode(v9ses, inode, &fid->qid, st->st_mode, new_decode_dev(st->st_rdev)); + kfree(st); if (retval) goto error; @@ -135,6 +90,7 @@ static struct inode *v9fs_qid_iget_dotl(struct super_block *sb, goto error; unlock_new_inode(inode); + return inode; error: iget_failed(inode); @@ -142,22 +98,6 @@ error: } -struct inode * -v9fs_inode_from_fid_dotl(struct v9fs_session_info *v9ses, struct p9_fid *fid, - struct super_block *sb, int new) -{ - struct p9_stat_dotl *st; - struct inode *inode = NULL; - - st = p9_client_getattr_dotl(fid, P9_STATS_BASIC | P9_STATS_GEN); - if (IS_ERR(st)) - return ERR_CAST(st); - - inode = v9fs_qid_iget_dotl(sb, &st->qid, fid, st, new); - kfree(st); - return inode; -} - struct dotl_openflag_map { int open_flag; int dotl_flag; @@ -307,7 +247,7 @@ v9fs_vfs_atomic_open_dotl(struct inode *dir, struct dentry *dentry, p9_debug(P9_DEBUG_VFS, "p9_client_walk failed %d\n", err); goto out; } - inode = v9fs_get_new_inode_from_fid(v9ses, fid, dir->i_sb); + inode = v9fs_fid_iget_dotl(dir->i_sb, fid); if (IS_ERR(inode)) { err = PTR_ERR(inode); p9_debug(P9_DEBUG_VFS, "inode creation failed %d\n", err); @@ -402,32 +342,17 @@ static int v9fs_vfs_mkdir_dotl(struct mnt_idmap *idmap, } /* instantiate inode and assign the unopened fid to the dentry */ - if (v9ses->cache & (CACHE_META|CACHE_LOOSE)) { - inode = v9fs_get_new_inode_from_fid(v9ses, fid, dir->i_sb); - if (IS_ERR(inode)) { - err = PTR_ERR(inode); - p9_debug(P9_DEBUG_VFS, "inode creation failed %d\n", - err); - goto error; - } - v9fs_fid_add(dentry, &fid); - v9fs_set_create_acl(inode, fid, dacl, pacl); - d_instantiate(dentry, inode); - err = 0; - } else { - /* - * Not in cached mode. No need to populate - * inode with stat. We need to get an inode - * so that we can set the acl with dentry - */ - inode = v9fs_get_inode(dir->i_sb, mode, 0); - if (IS_ERR(inode)) { - err = PTR_ERR(inode); - goto error; - } - v9fs_set_create_acl(inode, fid, dacl, pacl); - d_instantiate(dentry, inode); + inode = v9fs_fid_iget_dotl(dir->i_sb, fid); + if (IS_ERR(inode)) { + err = PTR_ERR(inode); + p9_debug(P9_DEBUG_VFS, "inode creation failed %d\n", + err); + goto error; } + v9fs_fid_add(dentry, &fid); + v9fs_set_create_acl(inode, fid, dacl, pacl); + d_instantiate(dentry, inode); + err = 0; inc_nlink(dir); v9fs_invalidate_inode_attr(dir); error: @@ -709,14 +634,11 @@ v9fs_vfs_symlink_dotl(struct mnt_idmap *idmap, struct inode *dir, kgid_t gid; const unsigned char *name; struct p9_qid qid; - struct inode *inode; struct p9_fid *dfid; struct p9_fid *fid = NULL; - struct v9fs_session_info *v9ses; name = dentry->d_name.name; p9_debug(P9_DEBUG_VFS, "%lu,%s,%s\n", dir->i_ino, name, symname); - v9ses = v9fs_inode2v9ses(dir); dfid = v9fs_parent_fid(dentry); if (IS_ERR(dfid)) { @@ -736,36 +658,6 @@ v9fs_vfs_symlink_dotl(struct mnt_idmap *idmap, struct inode *dir, } v9fs_invalidate_inode_attr(dir); - if (v9ses->cache & (CACHE_META|CACHE_LOOSE)) { - /* Now walk from the parent so we can get an unopened fid. */ - fid = p9_client_walk(dfid, 1, &name, 1); - if (IS_ERR(fid)) { - err = PTR_ERR(fid); - p9_debug(P9_DEBUG_VFS, "p9_client_walk failed %d\n", - err); - goto error; - } - - /* instantiate inode and assign the unopened fid to dentry */ - inode = v9fs_get_new_inode_from_fid(v9ses, fid, dir->i_sb); - if (IS_ERR(inode)) { - err = PTR_ERR(inode); - p9_debug(P9_DEBUG_VFS, "inode creation failed %d\n", - err); - goto error; - } - v9fs_fid_add(dentry, &fid); - d_instantiate(dentry, inode); - err = 0; - } else { - /* Not in cached mode. No need to populate inode with stat */ - inode = v9fs_get_inode(dir->i_sb, S_IFLNK, 0); - if (IS_ERR(inode)) { - err = PTR_ERR(inode); - goto error; - } - d_instantiate(dentry, inode); - } error: p9_fid_put(fid); @@ -888,33 +780,17 @@ v9fs_vfs_mknod_dotl(struct mnt_idmap *idmap, struct inode *dir, err); goto error; } - - /* instantiate inode and assign the unopened fid to the dentry */ - if (v9ses->cache & (CACHE_META|CACHE_LOOSE)) { - inode = v9fs_get_new_inode_from_fid(v9ses, fid, dir->i_sb); - if (IS_ERR(inode)) { - err = PTR_ERR(inode); - p9_debug(P9_DEBUG_VFS, "inode creation failed %d\n", - err); - goto error; - } - v9fs_set_create_acl(inode, fid, dacl, pacl); - v9fs_fid_add(dentry, &fid); - d_instantiate(dentry, inode); - err = 0; - } else { - /* - * Not in cached mode. No need to populate inode with stat. - * socket syscall returns a fd, so we need instantiate - */ - inode = v9fs_get_inode(dir->i_sb, mode, rdev); - if (IS_ERR(inode)) { - err = PTR_ERR(inode); - goto error; - } - v9fs_set_create_acl(inode, fid, dacl, pacl); - d_instantiate(dentry, inode); + inode = v9fs_fid_iget_dotl(dir->i_sb, fid); + if (IS_ERR(inode)) { + err = PTR_ERR(inode); + p9_debug(P9_DEBUG_VFS, "inode creation failed %d\n", + err); + goto error; } + v9fs_set_create_acl(inode, fid, dacl, pacl); + v9fs_fid_add(dentry, &fid); + d_instantiate(dentry, inode); + err = 0; error: p9_fid_put(fid); v9fs_put_acl(dacl, pacl); diff --git a/fs/9p/vfs_super.c b/fs/9p/vfs_super.c index 941f7d0e0bfa..4236058c7bbd 100644 --- a/fs/9p/vfs_super.c +++ b/fs/9p/vfs_super.c @@ -110,7 +110,6 @@ static struct dentry *v9fs_mount(struct file_system_type *fs_type, int flags, struct inode *inode = NULL; struct dentry *root = NULL; struct v9fs_session_info *v9ses = NULL; - umode_t mode = 0777 | S_ISVTX; struct p9_fid *fid; int retval = 0; @@ -140,7 +139,7 @@ static struct dentry *v9fs_mount(struct file_system_type *fs_type, int flags, else sb->s_d_op = &v9fs_dentry_operations; - inode = v9fs_get_inode(sb, S_IFDIR | mode, 0); + inode = v9fs_get_inode_from_fid(v9ses, fid, sb); if (IS_ERR(inode)) { retval = PTR_ERR(inode); goto release_sb; @@ -152,32 +151,6 @@ static struct dentry *v9fs_mount(struct file_system_type *fs_type, int flags, goto release_sb; } sb->s_root = root; - if (v9fs_proto_dotl(v9ses)) { - struct p9_stat_dotl *st = NULL; - - st = p9_client_getattr_dotl(fid, P9_STATS_BASIC); - if (IS_ERR(st)) { - retval = PTR_ERR(st); - goto release_sb; - } - d_inode(root)->i_ino = v9fs_qid2ino(&st->qid); - v9fs_stat2inode_dotl(st, d_inode(root), 0); - kfree(st); - } else { - struct p9_wstat *st = NULL; - - st = p9_client_stat(fid); - if (IS_ERR(st)) { - retval = PTR_ERR(st); - goto release_sb; - } - - d_inode(root)->i_ino = v9fs_qid2ino(&st->qid); - v9fs_stat2inode(st, d_inode(root), sb, 0); - - p9stat_free(st); - kfree(st); - } retval = v9fs_get_acl(inode, fid); if (retval) goto release_sb; @@ -271,21 +244,6 @@ done: return res; } -static int v9fs_drop_inode(struct inode *inode) -{ - struct v9fs_session_info *v9ses; - - v9ses = v9fs_inode2v9ses(inode); - if (v9ses->cache & (CACHE_META|CACHE_LOOSE)) - return generic_drop_inode(inode); - /* - * in case of non cached mode always drop the - * inode because we want the inode attribute - * to always match that on the server. - */ - return 1; -} - static int v9fs_write_inode(struct inode *inode, struct writeback_control *wbc) { @@ -320,7 +278,6 @@ static const struct super_operations v9fs_super_ops_dotl = { .alloc_inode = v9fs_alloc_inode, .free_inode = v9fs_free_inode, .statfs = v9fs_statfs, - .drop_inode = v9fs_drop_inode, .evict_inode = v9fs_evict_inode, .show_options = v9fs_show_options, .umount_begin = v9fs_umount_begin, diff --git a/fs/Kconfig b/fs/Kconfig index ea2f77446080..a46b0cbc4d8f 100644 --- a/fs/Kconfig +++ b/fs/Kconfig @@ -60,7 +60,6 @@ endif # BLOCK config FS_DAX bool "File system based Direct Access (DAX) support" depends on MMU - depends on !(ARM || MIPS || SPARC) depends on ZONE_DEVICE || FS_DAX_LIMITED select FS_IOMAP select DAX @@ -261,6 +260,7 @@ menuconfig HUGETLBFS depends on X86 || SPARC64 || ARCH_SUPPORTS_HUGETLBFS || BROKEN depends on (SYSFS || SYSCTL) select MEMFD_CREATE + select PADATA if SMP help hugetlbfs is a filesystem backing for HugeTLB pages, based on ramfs. For architectures that support it, say Y here and read diff --git a/fs/afs/dir.c b/fs/afs/dir.c index 8a67fc427e74..67afe68972d5 100644 --- a/fs/afs/dir.c +++ b/fs/afs/dir.c @@ -474,16 +474,6 @@ static int afs_dir_iterate_block(struct afs_vnode *dvnode, continue; } - /* Don't expose silly rename entries to userspace. */ - if (nlen > 6 && - dire->u.name[0] == '.' && - ctx->actor != afs_lookup_filldir && - ctx->actor != afs_lookup_one_filldir && - memcmp(dire->u.name, ".__afs", 6) == 0) { - ctx->pos = blkoff + next * sizeof(union afs_xdr_dirent); - continue; - } - /* found the next entry */ if (!dir_emit(ctx, dire->u.name, nlen, ntohl(dire->u.vnode), diff --git a/fs/afs/rotate.c b/fs/afs/rotate.c index 700a27bc8c25..ed04bd1eeae8 100644 --- a/fs/afs/rotate.c +++ b/fs/afs/rotate.c @@ -602,6 +602,8 @@ iterate_address: goto wait_for_more_probe_results; alist = op->estate->addresses; + best_prio = -1; + addr_index = 0; for (i = 0; i < alist->nr_addrs; i++) { if (alist->addrs[i].prio > best_prio) { addr_index = i; @@ -609,9 +611,7 @@ iterate_address: } } - addr_index = READ_ONCE(alist->preferred); - if (!test_bit(addr_index, &set)) - addr_index = __ffs(set); + alist->preferred = addr_index; op->addr_index = addr_index; set_bit(addr_index, &op->addr_tried); @@ -656,12 +656,6 @@ wait_for_more_probe_results: next_server: trace_afs_rotate(op, afs_rotate_trace_next_server, 0); _debug("next"); - ASSERT(op->estate); - alist = op->estate->addresses; - if (op->call_responded && - op->addr_index != READ_ONCE(alist->preferred) && - test_bit(alist->preferred, &op->addr_tried)) - WRITE_ONCE(alist->preferred, op->addr_index); op->estate = NULL; goto pick_server; @@ -690,14 +684,7 @@ no_more_servers: failed: trace_afs_rotate(op, afs_rotate_trace_failed, 0); op->flags |= AFS_OPERATION_STOP; - if (op->estate) { - alist = op->estate->addresses; - if (op->call_responded && - op->addr_index != READ_ONCE(alist->preferred) && - test_bit(alist->preferred, &op->addr_tried)) - WRITE_ONCE(alist->preferred, op->addr_index); - op->estate = NULL; - } + op->estate = NULL; _leave(" = f [failed %d]", afs_op_error(op)); return false; } diff --git a/fs/afs/validation.c b/fs/afs/validation.c index 46b37f2cce7d..32a53fc8dfb2 100644 --- a/fs/afs/validation.c +++ b/fs/afs/validation.c @@ -122,6 +122,9 @@ bool afs_check_validity(const struct afs_vnode *vnode) const struct afs_volume *volume = vnode->volume; time64_t deadline = ktime_get_real_seconds() + 10; + if (test_bit(AFS_VNODE_DELETED, &vnode->flags)) + return true; + if (atomic_read(&volume->cb_v_check) != atomic_read(&volume->cb_v_break) || atomic64_read(&vnode->cb_expires_at) <= deadline || volume->cb_expires_at <= deadline || @@ -389,12 +392,17 @@ int afs_validate(struct afs_vnode *vnode, struct key *key) key_serial(key)); if (afs_check_validity(vnode)) - return 0; + return test_bit(AFS_VNODE_DELETED, &vnode->flags) ? -ESTALE : 0; ret = down_write_killable(&vnode->validate_lock); if (ret < 0) goto error; + if (test_bit(AFS_VNODE_DELETED, &vnode->flags)) { + ret = -ESTALE; + goto error_unlock; + } + /* Validate a volume after the v_break has changed or the volume * callback expired. We only want to do this once per volume per * v_break change. The actual work will be done when parsing the @@ -448,12 +456,6 @@ int afs_validate(struct afs_vnode *vnode, struct key *key) vnode->cb_ro_snapshot = cb_ro_snapshot; vnode->cb_scrub = cb_scrub; - if (test_bit(AFS_VNODE_DELETED, &vnode->flags)) { - _debug("file already deleted"); - ret = -ESTALE; - goto error_unlock; - } - /* if the vnode's data version number changed then its contents are * different */ zap |= test_and_clear_bit(AFS_VNODE_ZAP_DATA, &vnode->flags); diff --git a/fs/bcachefs/Makefile b/fs/bcachefs/Makefile index 1a05cecda7cc..b02796c8a595 100644 --- a/fs/bcachefs/Makefile +++ b/fs/bcachefs/Makefile @@ -82,6 +82,7 @@ bcachefs-y := \ super-io.o \ sysfs.o \ tests.o \ + time_stats.o \ thread_with_file.o \ trace.o \ two_state_shared_lock.o \ @@ -90,3 +91,6 @@ bcachefs-y := \ xattr.o obj-$(CONFIG_MEAN_AND_VARIANCE_UNIT_TEST) += mean_and_variance_test.o + +# Silence "note: xyz changed in GCC X.X" messages +subdir-ccflags-y += $(call cc-disable-warning, psabi) diff --git a/fs/bcachefs/alloc_background.c b/fs/bcachefs/alloc_background.c index fd3e175d8342..893e38f9db80 100644 --- a/fs/bcachefs/alloc_background.c +++ b/fs/bcachefs/alloc_background.c @@ -29,6 +29,8 @@ #include <linux/sched/task.h> #include <linux/sort.h> +static void bch2_discard_one_bucket_fast(struct bch_fs *c, struct bpos bucket); + /* Persistent alloc info: */ static const unsigned BCH_ALLOC_V1_FIELD_BYTES[] = { @@ -530,13 +532,13 @@ int bch2_bucket_gens_init(struct bch_fs *c) u8 gen = bch2_alloc_to_v4(k, &a)->gen; unsigned offset; struct bpos pos = alloc_gens_pos(iter.pos, &offset); + int ret2 = 0; if (have_bucket_gens_key && bkey_cmp(iter.pos, pos)) { - ret = commit_do(trans, NULL, NULL, - BCH_TRANS_COMMIT_no_enospc, - bch2_btree_insert_trans(trans, BTREE_ID_bucket_gens, &g.k_i, 0)); - if (ret) - break; + ret2 = bch2_btree_insert_trans(trans, BTREE_ID_bucket_gens, &g.k_i, 0) ?: + bch2_trans_commit(trans, NULL, NULL, BCH_TRANS_COMMIT_no_enospc); + if (ret2) + goto iter_err; have_bucket_gens_key = false; } @@ -547,7 +549,8 @@ int bch2_bucket_gens_init(struct bch_fs *c) } g.v.gens[offset] = gen; - 0; +iter_err: + ret2; })); if (have_bucket_gens_key && !ret) @@ -850,7 +853,7 @@ int bch2_trigger_alloc(struct btree_trans *trans, bucket_journal_seq); if (ret) { bch2_fs_fatal_error(c, - "error setting bucket_needs_journal_commit: %i", ret); + "setting bucket_needs_journal_commit: %s", bch2_err_str(ret)); return ret; } } @@ -860,23 +863,28 @@ int bch2_trigger_alloc(struct btree_trans *trans, *bucket_gen(ca, new.k->p.offset) = new_a->gen; bch2_dev_usage_update(c, ca, old_a, new_a, journal_seq, false); + percpu_up_read(&c->mark_lock); + +#define eval_state(_a, expr) ({ const struct bch_alloc_v4 *a = _a; expr; }) +#define statechange(expr) !eval_state(old_a, expr) && eval_state(new_a, expr) +#define bucket_flushed(a) (!a->journal_seq || a->journal_seq <= c->journal.flushed_seq_ondisk) - if (new_a->data_type == BCH_DATA_free && - (!new_a->journal_seq || new_a->journal_seq < c->journal.flushed_seq_ondisk)) + if (statechange(a->data_type == BCH_DATA_free) && + bucket_flushed(new_a)) closure_wake_up(&c->freelist_wait); - if (new_a->data_type == BCH_DATA_need_discard && - (!bucket_journal_seq || bucket_journal_seq < c->journal.flushed_seq_ondisk)) - bch2_do_discards(c); + if (statechange(a->data_type == BCH_DATA_need_discard) && + !bch2_bucket_is_open(c, new.k->p.inode, new.k->p.offset) && + bucket_flushed(new_a)) + bch2_discard_one_bucket_fast(c, new.k->p); - if (old_a->data_type != BCH_DATA_cached && - new_a->data_type == BCH_DATA_cached && + if (statechange(a->data_type == BCH_DATA_cached) && + !bch2_bucket_is_open(c, new.k->p.inode, new.k->p.offset) && should_invalidate_buckets(ca, bch2_dev_usage_read(ca))) bch2_do_invalidates(c); - if (new_a->data_type == BCH_DATA_need_gc_gens) + if (statechange(a->data_type == BCH_DATA_need_gc_gens)) bch2_do_gc_gens(c); - percpu_up_read(&c->mark_lock); } if ((flags & BTREE_TRIGGER_GC) && @@ -1045,14 +1053,13 @@ int bch2_check_alloc_key(struct btree_trans *trans, if (ret) goto err; - if (k.k->type != discard_key_type && - (c->opts.reconstruct_alloc || - fsck_err(c, need_discard_key_wrong, - "incorrect key in need_discard btree (got %s should be %s)\n" - " %s", - bch2_bkey_types[k.k->type], - bch2_bkey_types[discard_key_type], - (bch2_bkey_val_to_text(&buf, c, alloc_k), buf.buf)))) { + if (fsck_err_on(k.k->type != discard_key_type, + c, need_discard_key_wrong, + "incorrect key in need_discard btree (got %s should be %s)\n" + " %s", + bch2_bkey_types[k.k->type], + bch2_bkey_types[discard_key_type], + (bch2_bkey_val_to_text(&buf, c, alloc_k), buf.buf))) { struct bkey_i *update = bch2_trans_kmalloc(trans, sizeof(*update)); @@ -1076,15 +1083,14 @@ int bch2_check_alloc_key(struct btree_trans *trans, if (ret) goto err; - if (k.k->type != freespace_key_type && - (c->opts.reconstruct_alloc || - fsck_err(c, freespace_key_wrong, - "incorrect key in freespace btree (got %s should be %s)\n" - " %s", - bch2_bkey_types[k.k->type], - bch2_bkey_types[freespace_key_type], - (printbuf_reset(&buf), - bch2_bkey_val_to_text(&buf, c, alloc_k), buf.buf)))) { + if (fsck_err_on(k.k->type != freespace_key_type, + c, freespace_key_wrong, + "incorrect key in freespace btree (got %s should be %s)\n" + " %s", + bch2_bkey_types[k.k->type], + bch2_bkey_types[freespace_key_type], + (printbuf_reset(&buf), + bch2_bkey_val_to_text(&buf, c, alloc_k), buf.buf))) { struct bkey_i *update = bch2_trans_kmalloc(trans, sizeof(*update)); @@ -1108,14 +1114,13 @@ int bch2_check_alloc_key(struct btree_trans *trans, if (ret) goto err; - if (a->gen != alloc_gen(k, gens_offset) && - (c->opts.reconstruct_alloc || - fsck_err(c, bucket_gens_key_wrong, - "incorrect gen in bucket_gens btree (got %u should be %u)\n" - " %s", - alloc_gen(k, gens_offset), a->gen, - (printbuf_reset(&buf), - bch2_bkey_val_to_text(&buf, c, alloc_k), buf.buf)))) { + if (fsck_err_on(a->gen != alloc_gen(k, gens_offset), + c, bucket_gens_key_wrong, + "incorrect gen in bucket_gens btree (got %u should be %u)\n" + " %s", + alloc_gen(k, gens_offset), a->gen, + (printbuf_reset(&buf), + bch2_bkey_val_to_text(&buf, c, alloc_k), buf.buf))) { struct bkey_i_bucket_gens *g = bch2_trans_kmalloc(trans, sizeof(*g)); @@ -1167,14 +1172,13 @@ int bch2_check_alloc_hole_freespace(struct btree_trans *trans, *end = bkey_min(k.k->p, *end); - if (k.k->type != KEY_TYPE_set && - (c->opts.reconstruct_alloc || - fsck_err(c, freespace_hole_missing, - "hole in alloc btree missing in freespace btree\n" - " device %llu buckets %llu-%llu", - freespace_iter->pos.inode, - freespace_iter->pos.offset, - end->offset))) { + if (fsck_err_on(k.k->type != KEY_TYPE_set, + c, freespace_hole_missing, + "hole in alloc btree missing in freespace btree\n" + " device %llu buckets %llu-%llu", + freespace_iter->pos.inode, + freespace_iter->pos.offset, + end->offset)) { struct bkey_i *update = bch2_trans_kmalloc(trans, sizeof(*update)); @@ -1604,6 +1608,36 @@ int bch2_check_alloc_to_lru_refs(struct bch_fs *c) return ret; } +static int discard_in_flight_add(struct bch_fs *c, struct bpos bucket) +{ + int ret; + + mutex_lock(&c->discard_buckets_in_flight_lock); + darray_for_each(c->discard_buckets_in_flight, i) + if (bkey_eq(*i, bucket)) { + ret = -EEXIST; + goto out; + } + + ret = darray_push(&c->discard_buckets_in_flight, bucket); +out: + mutex_unlock(&c->discard_buckets_in_flight_lock); + return ret; +} + +static void discard_in_flight_remove(struct bch_fs *c, struct bpos bucket) +{ + mutex_lock(&c->discard_buckets_in_flight_lock); + darray_for_each(c->discard_buckets_in_flight, i) + if (bkey_eq(*i, bucket)) { + darray_remove_item(&c->discard_buckets_in_flight, i); + goto found; + } + BUG(); +found: + mutex_unlock(&c->discard_buckets_in_flight_lock); +} + struct discard_buckets_state { u64 seen; u64 open; @@ -1642,6 +1676,7 @@ static int bch2_discard_one_bucket(struct btree_trans *trans, struct bch_dev *ca; struct bkey_i_alloc_v4 *a; struct printbuf buf = PRINTBUF; + bool discard_locked = false; int ret = 0; ca = bch_dev_bkey_exists(c, pos.inode); @@ -1709,6 +1744,11 @@ static int bch2_discard_one_bucket(struct btree_trans *trans, goto out; } + if (discard_in_flight_add(c, SPOS(iter.pos.inode, iter.pos.offset, true))) + goto out; + + discard_locked = true; + if (!bkey_eq(*discard_pos_done, iter.pos) && ca->mi.discard && !c->opts.nochanges) { /* @@ -1740,6 +1780,8 @@ write: count_event(c, bucket_discard); s->discarded++; out: + if (discard_locked) + discard_in_flight_remove(c, iter.pos); s->seen++; bch2_trans_iter_exit(trans, &iter); percpu_ref_put(&ca->io_ref); @@ -1779,6 +1821,93 @@ void bch2_do_discards(struct bch_fs *c) bch2_write_ref_put(c, BCH_WRITE_REF_discard); } +static int bch2_clear_bucket_needs_discard(struct btree_trans *trans, struct bpos bucket) +{ + struct btree_iter iter; + bch2_trans_iter_init(trans, &iter, BTREE_ID_alloc, bucket, BTREE_ITER_INTENT); + struct bkey_s_c k = bch2_btree_iter_peek_slot(&iter); + int ret = bkey_err(k); + if (ret) + goto err; + + struct bkey_i_alloc_v4 *a = bch2_alloc_to_v4_mut(trans, k); + ret = PTR_ERR_OR_ZERO(a); + if (ret) + goto err; + + SET_BCH_ALLOC_V4_NEED_DISCARD(&a->v, false); + a->v.data_type = alloc_data_type(a->v, a->v.data_type); + + ret = bch2_trans_update(trans, &iter, &a->k_i, 0); +err: + bch2_trans_iter_exit(trans, &iter); + return ret; +} + +static void bch2_do_discards_fast_work(struct work_struct *work) +{ + struct bch_fs *c = container_of(work, struct bch_fs, discard_fast_work); + + while (1) { + bool got_bucket = false; + struct bpos bucket; + struct bch_dev *ca; + + mutex_lock(&c->discard_buckets_in_flight_lock); + darray_for_each(c->discard_buckets_in_flight, i) { + if (i->snapshot) + continue; + + ca = bch_dev_bkey_exists(c, i->inode); + + if (!percpu_ref_tryget(&ca->io_ref)) { + darray_remove_item(&c->discard_buckets_in_flight, i); + continue; + } + + got_bucket = true; + bucket = *i; + i->snapshot = true; + break; + } + mutex_unlock(&c->discard_buckets_in_flight_lock); + + if (!got_bucket) + break; + + if (ca->mi.discard && !c->opts.nochanges) + blkdev_issue_discard(ca->disk_sb.bdev, + bucket.offset * ca->mi.bucket_size, + ca->mi.bucket_size, + GFP_KERNEL); + + int ret = bch2_trans_do(c, NULL, NULL, + BCH_WATERMARK_btree| + BCH_TRANS_COMMIT_no_enospc, + bch2_clear_bucket_needs_discard(trans, bucket)); + bch_err_fn(c, ret); + + percpu_ref_put(&ca->io_ref); + discard_in_flight_remove(c, bucket); + + if (ret) + break; + } + + bch2_write_ref_put(c, BCH_WRITE_REF_discard_fast); +} + +static void bch2_discard_one_bucket_fast(struct bch_fs *c, struct bpos bucket) +{ + struct bch_dev *ca = bch_dev_bkey_exists(c, bucket.inode); + + if (!percpu_ref_is_dying(&ca->io_ref) && + !discard_in_flight_add(c, bucket) && + bch2_write_ref_tryget(c, BCH_WRITE_REF_discard_fast) && + !queue_work(c->write_ref_wq, &c->discard_fast_work)) + bch2_write_ref_put(c, BCH_WRITE_REF_discard_fast); +} + static int invalidate_one_bucket(struct btree_trans *trans, struct btree_iter *lru_iter, struct bkey_s_c lru_k, @@ -2210,9 +2339,16 @@ void bch2_dev_allocator_add(struct bch_fs *c, struct bch_dev *ca) set_bit(ca->dev_idx, c->rw_devs[i].d); } +void bch2_fs_allocator_background_exit(struct bch_fs *c) +{ + darray_exit(&c->discard_buckets_in_flight); +} + void bch2_fs_allocator_background_init(struct bch_fs *c) { spin_lock_init(&c->freelist_lock); + mutex_init(&c->discard_buckets_in_flight_lock); INIT_WORK(&c->discard_work, bch2_do_discards_work); + INIT_WORK(&c->discard_fast_work, bch2_do_discards_fast_work); INIT_WORK(&c->invalidate_work, bch2_do_invalidates_work); } diff --git a/fs/bcachefs/alloc_background.h b/fs/bcachefs/alloc_background.h index e7f7e842ee1b..052b2fac25d6 100644 --- a/fs/bcachefs/alloc_background.h +++ b/fs/bcachefs/alloc_background.h @@ -269,6 +269,7 @@ u64 bch2_min_rw_member_capacity(struct bch_fs *); void bch2_dev_allocator_remove(struct bch_fs *, struct bch_dev *); void bch2_dev_allocator_add(struct bch_fs *, struct bch_dev *); +void bch2_fs_allocator_background_exit(struct bch_fs *); void bch2_fs_allocator_background_init(struct bch_fs *); #endif /* _BCACHEFS_ALLOC_BACKGROUND_H */ diff --git a/fs/bcachefs/alloc_foreground.c b/fs/bcachefs/alloc_foreground.c index 633d3223b353..214b15c84d1f 100644 --- a/fs/bcachefs/alloc_foreground.c +++ b/fs/bcachefs/alloc_foreground.c @@ -236,8 +236,7 @@ static struct open_bucket *__try_alloc_bucket(struct bch_fs *c, struct bch_dev * if (cl) closure_wait(&c->open_buckets_wait, cl); - track_event_change(&c->times[BCH_TIME_blocked_allocate_open_bucket], - &c->blocked_allocate_open_bucket, true); + track_event_change(&c->times[BCH_TIME_blocked_allocate_open_bucket], true); spin_unlock(&c->freelist_lock); return ERR_PTR(-BCH_ERR_open_buckets_empty); } @@ -263,11 +262,8 @@ static struct open_bucket *__try_alloc_bucket(struct bch_fs *c, struct bch_dev * ca->nr_open_buckets++; bch2_open_bucket_hash_add(c, ob); - track_event_change(&c->times[BCH_TIME_blocked_allocate_open_bucket], - &c->blocked_allocate_open_bucket, false); - - track_event_change(&c->times[BCH_TIME_blocked_allocate], - &c->blocked_allocate, false); + track_event_change(&c->times[BCH_TIME_blocked_allocate_open_bucket], false); + track_event_change(&c->times[BCH_TIME_blocked_allocate], false); spin_unlock(&c->freelist_lock); return ob; @@ -555,8 +551,7 @@ again: goto again; } - track_event_change(&c->times[BCH_TIME_blocked_allocate], - &c->blocked_allocate, true); + track_event_change(&c->times[BCH_TIME_blocked_allocate], true); ob = ERR_PTR(-BCH_ERR_freelist_empty); goto err; @@ -1361,15 +1356,17 @@ retry: /* Don't retry from all devices if we're out of open buckets: */ if (bch2_err_matches(ret, BCH_ERR_open_buckets_empty)) { - int ret = open_bucket_add_buckets(trans, &ptrs, wp, devs_have, + int ret2 = open_bucket_add_buckets(trans, &ptrs, wp, devs_have, target, erasure_code, nr_replicas, &nr_effective, &have_cache, watermark, flags, cl); - if (!ret || - bch2_err_matches(ret, BCH_ERR_transaction_restart) || - bch2_err_matches(ret, BCH_ERR_open_buckets_empty)) + if (!ret2 || + bch2_err_matches(ret2, BCH_ERR_transaction_restart) || + bch2_err_matches(ret2, BCH_ERR_open_buckets_empty)) { + ret = ret2; goto alloc_done; + } } /* diff --git a/fs/bcachefs/backpointers.c b/fs/bcachefs/backpointers.c index 569b97904da4..8cb35ea572cb 100644 --- a/fs/bcachefs/backpointers.c +++ b/fs/bcachefs/backpointers.c @@ -131,8 +131,7 @@ static noinline int backpointer_mod_err(struct btree_trans *trans, printbuf_exit(&buf); if (c->curr_recovery_pass > BCH_RECOVERY_PASS_check_extents_to_backpointers) { - bch2_inconsistent_error(c); - return -EIO; + return bch2_inconsistent_error(c) ? BCH_ERR_erofs_unfixed_errors : 0; } else { return 0; } @@ -478,8 +477,7 @@ missing: prt_printf(&buf, "\nbp pos "); bch2_bpos_to_text(&buf, bp_iter.pos); - if (c->opts.reconstruct_alloc || - fsck_err(c, ptr_to_missing_backpointer, "%s", buf.buf)) + if (fsck_err(c, ptr_to_missing_backpointer, "%s", buf.buf)) ret = bch2_bucket_backpointer_mod(trans, bucket, bp, orig_k, true); goto out; @@ -555,60 +553,61 @@ static inline struct bbpos bp_to_bbpos(struct bch_backpointer bp) }; } -static size_t btree_nodes_fit_in_ram(struct bch_fs *c) +static u64 mem_may_pin_bytes(struct bch_fs *c) { struct sysinfo i; - u64 mem_bytes; - si_meminfo(&i); - mem_bytes = i.totalram * i.mem_unit; - return div_u64(mem_bytes >> 1, c->opts.btree_node_size); + + u64 mem_bytes = i.totalram * i.mem_unit; + return div_u64(mem_bytes * c->opts.fsck_memory_usage_percent, 100); +} + +static size_t btree_nodes_fit_in_ram(struct bch_fs *c) +{ + return div_u64(mem_may_pin_bytes(c), c->opts.btree_node_size); } static int bch2_get_btree_in_memory_pos(struct btree_trans *trans, - unsigned btree_leaf_mask, - unsigned btree_interior_mask, + u64 btree_leaf_mask, + u64 btree_interior_mask, struct bbpos start, struct bbpos *end) { - struct btree_iter iter; - struct bkey_s_c k; - size_t btree_nodes = btree_nodes_fit_in_ram(trans->c); - enum btree_id btree; + struct bch_fs *c = trans->c; + s64 mem_may_pin = mem_may_pin_bytes(c); int ret = 0; - for (btree = start.btree; btree < BTREE_ID_NR && !ret; btree++) { - unsigned depth = ((1U << btree) & btree_leaf_mask) ? 1 : 2; + btree_interior_mask |= btree_leaf_mask; + + c->btree_cache.pinned_nodes_leaf_mask = btree_leaf_mask; + c->btree_cache.pinned_nodes_interior_mask = btree_interior_mask; + c->btree_cache.pinned_nodes_start = start; + c->btree_cache.pinned_nodes_end = *end = BBPOS_MAX; + + for (enum btree_id btree = start.btree; + btree < BTREE_ID_NR && !ret; + btree++) { + unsigned depth = ((1U << btree) & btree_leaf_mask) ? 0 : 1; + struct btree_iter iter; + struct btree *b; if (!((1U << btree) & btree_leaf_mask) && !((1U << btree) & btree_interior_mask)) continue; - bch2_trans_node_iter_init(trans, &iter, btree, - btree == start.btree ? start.pos : POS_MIN, - 0, depth, 0); - /* - * for_each_btree_key_contineu() doesn't check the return value - * from bch2_btree_iter_advance(), which is needed when - * iterating over interior nodes where we'll see keys at - * SPOS_MAX: - */ - do { - k = __bch2_btree_iter_peek_and_restart(trans, &iter, 0); - ret = bkey_err(k); - if (!k.k || ret) - break; - - --btree_nodes; - if (!btree_nodes) { - *end = BBPOS(btree, k.k->p); + __for_each_btree_node(trans, iter, btree, + btree == start.btree ? start.pos : POS_MIN, + 0, depth, BTREE_ITER_PREFETCH, b, ret) { + mem_may_pin -= btree_buf_bytes(b); + if (mem_may_pin <= 0) { + c->btree_cache.pinned_nodes_end = *end = + BBPOS(btree, b->key.k.p); bch2_trans_iter_exit(trans, &iter); return 0; } - } while (bch2_btree_iter_advance(&iter)); + } bch2_trans_iter_exit(trans, &iter); } - *end = BBPOS_MAX; return ret; } @@ -666,62 +665,6 @@ static int bch2_check_extents_to_backpointers_pass(struct btree_trans *trans, return 0; } -static struct bpos bucket_pos_to_bp_safe(const struct bch_fs *c, - struct bpos bucket) -{ - return bch2_dev_exists2(c, bucket.inode) - ? bucket_pos_to_bp(c, bucket, 0) - : bucket; -} - -static int bch2_get_alloc_in_memory_pos(struct btree_trans *trans, - struct bpos start, struct bpos *end) -{ - struct btree_iter alloc_iter; - struct btree_iter bp_iter; - struct bkey_s_c alloc_k, bp_k; - size_t btree_nodes = btree_nodes_fit_in_ram(trans->c); - bool alloc_end = false, bp_end = false; - int ret = 0; - - bch2_trans_node_iter_init(trans, &alloc_iter, BTREE_ID_alloc, - start, 0, 1, 0); - bch2_trans_node_iter_init(trans, &bp_iter, BTREE_ID_backpointers, - bucket_pos_to_bp_safe(trans->c, start), 0, 1, 0); - while (1) { - alloc_k = !alloc_end - ? __bch2_btree_iter_peek_and_restart(trans, &alloc_iter, 0) - : bkey_s_c_null; - bp_k = !bp_end - ? __bch2_btree_iter_peek_and_restart(trans, &bp_iter, 0) - : bkey_s_c_null; - - ret = bkey_err(alloc_k) ?: bkey_err(bp_k); - if ((!alloc_k.k && !bp_k.k) || ret) { - *end = SPOS_MAX; - break; - } - - --btree_nodes; - if (!btree_nodes) { - *end = alloc_k.k ? alloc_k.k->p : SPOS_MAX; - break; - } - - if (bpos_lt(alloc_iter.pos, SPOS_MAX) && - bpos_lt(bucket_pos_to_bp_safe(trans->c, alloc_iter.pos), bp_iter.pos)) { - if (!bch2_btree_iter_advance(&alloc_iter)) - alloc_end = true; - } else { - if (!bch2_btree_iter_advance(&bp_iter)) - bp_end = true; - } - } - bch2_trans_iter_exit(trans, &bp_iter); - bch2_trans_iter_exit(trans, &alloc_iter); - return ret; -} - int bch2_check_extents_to_backpointers(struct bch_fs *c) { struct btree_trans *trans = bch2_trans_get(c); @@ -732,10 +675,16 @@ int bch2_check_extents_to_backpointers(struct bch_fs *c) bkey_init(&s.last_flushed.k->k); while (1) { - ret = bch2_get_alloc_in_memory_pos(trans, s.bucket_start, &s.bucket_end); + struct bbpos end; + ret = bch2_get_btree_in_memory_pos(trans, + BIT_ULL(BTREE_ID_backpointers), + BIT_ULL(BTREE_ID_backpointers), + BBPOS(BTREE_ID_backpointers, s.bucket_start), &end); if (ret) break; + s.bucket_end = end.pos; + if ( bpos_eq(s.bucket_start, POS_MIN) && !bpos_eq(s.bucket_end, SPOS_MAX)) bch_verbose(c, "%s(): alloc info does not fit in ram, running in multiple passes with %zu nodes per pass", @@ -763,6 +712,9 @@ int bch2_check_extents_to_backpointers(struct bch_fs *c) bch2_trans_put(trans); bch2_bkey_buf_exit(&s.last_flushed, c); + c->btree_cache.pinned_nodes_leaf_mask = 0; + c->btree_cache.pinned_nodes_interior_mask = 0; + bch_err_fn(c, ret); return ret; } @@ -868,6 +820,9 @@ int bch2_check_backpointers_to_extents(struct bch_fs *c) } bch2_trans_put(trans); + c->btree_cache.pinned_nodes_leaf_mask = 0; + c->btree_cache.pinned_nodes_interior_mask = 0; + bch_err_fn(c, ret); return ret; } diff --git a/fs/bcachefs/bbpos_types.h b/fs/bcachefs/bbpos_types.h index 5198e94cf3b8..f63893344f80 100644 --- a/fs/bcachefs/bbpos_types.h +++ b/fs/bcachefs/bbpos_types.h @@ -13,6 +13,6 @@ static inline struct bbpos BBPOS(enum btree_id btree, struct bpos pos) } #define BBPOS_MIN BBPOS(0, POS_MIN) -#define BBPOS_MAX BBPOS(BTREE_ID_NR - 1, POS_MAX) +#define BBPOS_MAX BBPOS(BTREE_ID_NR - 1, SPOS_MAX) #endif /* _BCACHEFS_BBPOS_TYPES_H */ diff --git a/fs/bcachefs/bcachefs.h b/fs/bcachefs/bcachefs.h index 69d0d60d50e3..799aa32b6b4d 100644 --- a/fs/bcachefs/bcachefs.h +++ b/fs/bcachefs/bcachefs.h @@ -212,6 +212,7 @@ #include "recovery_types.h" #include "sb-errors_types.h" #include "seqmutex.h" +#include "time_stats.h" #include "util.h" #ifdef CONFIG_BCACHEFS_DEBUG @@ -266,6 +267,9 @@ do { \ #define bch2_fmt(_c, fmt) bch2_log_msg(_c, fmt "\n") __printf(2, 3) +void bch2_print_opts(struct bch_opts *, const char *, ...); + +__printf(2, 3) void __bch2_print(struct bch_fs *c, const char *fmt, ...); #define maybe_dev_to_fs(_c) _Generic((_c), \ @@ -504,6 +508,7 @@ enum gc_phase { GC_PHASE_BTREE_deleted_inodes, GC_PHASE_BTREE_logged_ops, GC_PHASE_BTREE_rebalance_work, + GC_PHASE_BTREE_subvolume_children, GC_PHASE_PENDING_DELETE, }; @@ -593,7 +598,7 @@ struct bch_dev { /* The rest of this all shows up in sysfs */ atomic64_t cur_latency[2]; - struct bch2_time_stats io_latency[2]; + struct bch2_time_stats_quantiles io_latency[2]; #define CONGESTED_MAX 1024 atomic_t congested; @@ -663,6 +668,8 @@ struct journal_seq_blacklist_table { }; struct journal_keys { + /* must match layout in darray_types.h */ + size_t nr, size; struct journal_key { u64 journal_seq; u32 journal_offset; @@ -671,15 +678,13 @@ struct journal_keys { bool allocated; bool overwritten; struct bkey_i *k; - } *d; + } *data; /* * Gap buffer: instead of all the empty space in the array being at the * end of the buffer - from @nr to @size - the empty space is at @gap. * This means that sequential insertions are O(n) instead of O(n^2). */ size_t gap; - size_t nr; - size_t size; atomic_t ref; bool initial_ref_held; }; @@ -703,6 +708,7 @@ struct btree_trans_buf { x(reflink) \ x(fallocate) \ x(discard) \ + x(discard_fast) \ x(invalidate) \ x(delete_dead_snapshots) \ x(snapshot_delete_pagecache) \ @@ -843,6 +849,8 @@ struct bch_fs { struct workqueue_struct *btree_interior_update_worker; struct work_struct btree_interior_update_work; + struct workqueue_struct *btree_node_rewrite_worker; + struct list_head pending_node_rewrites; struct mutex pending_node_rewrites_lock; @@ -919,8 +927,6 @@ struct bch_fs { /* ALLOCATOR */ spinlock_t freelist_lock; struct closure_waitlist freelist_wait; - u64 blocked_allocate; - u64 blocked_allocate_open_bucket; open_bucket_idx_t open_buckets_freelist; open_bucket_idx_t open_buckets_nr_free; @@ -940,8 +946,11 @@ struct bch_fs { unsigned write_points_nr; struct buckets_waiting_for_journal buckets_waiting_for_journal; - struct work_struct discard_work; struct work_struct invalidate_work; + struct work_struct discard_work; + struct mutex discard_buckets_in_flight_lock; + DARRAY(struct bpos) discard_buckets_in_flight; + struct work_struct discard_fast_work; /* GARBAGE COLLECTION */ struct task_struct *gc_thread; diff --git a/fs/bcachefs/bcachefs_format.h b/fs/bcachefs/bcachefs_format.h index 0668b682a21c..bff8750ac0d7 100644 --- a/fs/bcachefs/bcachefs_format.h +++ b/fs/bcachefs/bcachefs_format.h @@ -189,7 +189,11 @@ struct bversion { __u32 hi; __u64 lo; #endif -} __packed __aligned(4); +} __packed +#if __BYTE_ORDER__ == __ORDER_LITTLE_ENDIAN__ +__aligned(4) +#endif +; struct bkey { /* Size of combined key and value, in u64s */ @@ -222,7 +226,36 @@ struct bkey { __u8 pad[1]; #endif -} __packed __aligned(8); +} __packed +#if __BYTE_ORDER__ == __ORDER_LITTLE_ENDIAN__ +/* + * The big-endian version of bkey can't be compiled by rustc with the "aligned" + * attr since it doesn't allow types to have both "packed" and "aligned" attrs. + * So for Rust compatibility, don't include this. It can be included in the LE + * version because the "packed" attr is redundant in that case. + * + * History: (quoting Kent) + * + * Specifically, when i was designing bkey, I wanted the header to be no + * bigger than necessary so that bkey_packed could use the rest. That means that + * decently offten extent keys will fit into only 8 bytes, instead of spilling over + * to 16. + * + * But packed_bkey treats the part after the header - the packed section - + * as a single multi word, variable length integer. And bkey, the unpacked + * version, is just a special case version of a bkey_packed; all the packed + * bkey code will work on keys in any packed format, the in-memory + * representation of an unpacked key also is just one type of packed key... + * + * So that constrains the key part of a bkig endian bkey to start right + * after the header. + * + * If we ever do a bkey_v2 and need to expand the hedaer by another byte for + * some reason - that will clean up this wart. + */ +__aligned(8) +#endif +; struct bkey_packed { __u64 _data[0]; @@ -840,7 +873,9 @@ struct bch_sb_field_downgrade { x(snapshot_skiplists, BCH_VERSION(1, 1)) \ x(deleted_inodes, BCH_VERSION(1, 2)) \ x(rebalance_work, BCH_VERSION(1, 3)) \ - x(member_seq, BCH_VERSION(1, 4)) + x(member_seq, BCH_VERSION(1, 4)) \ + x(subvolume_fs_parent, BCH_VERSION(1, 5)) \ + x(btree_subvolume_children, BCH_VERSION(1, 6)) enum bcachefs_metadata_version { bcachefs_metadata_version_min = 9, @@ -1275,7 +1310,8 @@ static inline __u64 __bset_magic(struct bch_sb *sb) x(dev_usage, 8) \ x(log, 9) \ x(overwrite, 10) \ - x(write_buffer_keys, 11) + x(write_buffer_keys, 11) \ + x(datetime, 12) enum { #define x(f, nr) BCH_JSET_ENTRY_##f = nr, @@ -1376,6 +1412,11 @@ struct jset_entry_log { u8 d[]; } __packed __aligned(8); +struct jset_entry_datetime { + struct jset_entry entry; + __le64 seconds; +} __packed __aligned(8); + /* * On disk format for a journal entry: * seq is monotonically increasing; every journal entry has its own unique @@ -1482,7 +1523,9 @@ enum btree_id_flags { BIT_ULL(KEY_TYPE_logged_op_truncate)| \ BIT_ULL(KEY_TYPE_logged_op_finsert)) \ x(rebalance_work, 18, BTREE_ID_SNAPSHOT_FIELD, \ - BIT_ULL(KEY_TYPE_set)|BIT_ULL(KEY_TYPE_cookie)) + BIT_ULL(KEY_TYPE_set)|BIT_ULL(KEY_TYPE_cookie)) \ + x(subvolume_children, 19, 0, \ + BIT_ULL(KEY_TYPE_set)) enum btree_id { #define x(name, nr, ...) BTREE_ID_##name = nr, diff --git a/fs/bcachefs/bkey.h b/fs/bcachefs/bkey.h index 831be01809f2..cf23ff47bed8 100644 --- a/fs/bcachefs/bkey.h +++ b/fs/bcachefs/bkey.h @@ -4,7 +4,7 @@ #include <linux/bug.h> #include "bcachefs_format.h" - +#include "bkey_types.h" #include "btree_types.h" #include "util.h" #include "vstructs.h" @@ -31,57 +31,6 @@ void bch2_bkey_packed_to_binary_text(struct printbuf *, const struct bkey_format *, const struct bkey_packed *); -/* bkey with split value, const */ -struct bkey_s_c { - const struct bkey *k; - const struct bch_val *v; -}; - -/* bkey with split value */ -struct bkey_s { - union { - struct { - struct bkey *k; - struct bch_val *v; - }; - struct bkey_s_c s_c; - }; -}; - -#define bkey_p_next(_k) vstruct_next(_k) - -static inline struct bkey_i *bkey_next(struct bkey_i *k) -{ - return (struct bkey_i *) ((u64 *) k->_data + k->k.u64s); -} - -#define bkey_val_u64s(_k) ((_k)->u64s - BKEY_U64s) - -static inline size_t bkey_val_bytes(const struct bkey *k) -{ - return bkey_val_u64s(k) * sizeof(u64); -} - -static inline void set_bkey_val_u64s(struct bkey *k, unsigned val_u64s) -{ - unsigned u64s = BKEY_U64s + val_u64s; - - BUG_ON(u64s > U8_MAX); - k->u64s = u64s; -} - -static inline void set_bkey_val_bytes(struct bkey *k, unsigned bytes) -{ - set_bkey_val_u64s(k, DIV_ROUND_UP(bytes, sizeof(u64))); -} - -#define bkey_val_end(_k) ((void *) (((u64 *) (_k).v) + bkey_val_u64s((_k).k))) - -#define bkey_deleted(_k) ((_k)->type == KEY_TYPE_deleted) - -#define bkey_whiteout(_k) \ - ((_k)->type == KEY_TYPE_deleted || (_k)->type == KEY_TYPE_whiteout) - enum bkey_lr_packed { BKEY_PACKED_BOTH, BKEY_PACKED_RIGHT, @@ -362,10 +311,7 @@ static inline struct bpos bkey_start_pos(const struct bkey *k) static inline unsigned bkeyp_key_u64s(const struct bkey_format *format, const struct bkey_packed *k) { - unsigned ret = bkey_packed(k) ? format->key_u64s : BKEY_U64s; - - EBUG_ON(k->u64s < ret); - return ret; + return bkey_packed(k) ? format->key_u64s : BKEY_U64s; } static inline unsigned bkeyp_key_bytes(const struct bkey_format *format, @@ -553,155 +499,6 @@ static inline void bkey_reassemble(struct bkey_i *dst, memcpy_u64s_small(&dst->v, src.v, bkey_val_u64s(src.k)); } -#define bkey_s_null ((struct bkey_s) { .k = NULL }) -#define bkey_s_c_null ((struct bkey_s_c) { .k = NULL }) - -#define bkey_s_err(err) ((struct bkey_s) { .k = ERR_PTR(err) }) -#define bkey_s_c_err(err) ((struct bkey_s_c) { .k = ERR_PTR(err) }) - -static inline struct bkey_s bkey_to_s(struct bkey *k) -{ - return (struct bkey_s) { .k = k, .v = NULL }; -} - -static inline struct bkey_s_c bkey_to_s_c(const struct bkey *k) -{ - return (struct bkey_s_c) { .k = k, .v = NULL }; -} - -static inline struct bkey_s bkey_i_to_s(struct bkey_i *k) -{ - return (struct bkey_s) { .k = &k->k, .v = &k->v }; -} - -static inline struct bkey_s_c bkey_i_to_s_c(const struct bkey_i *k) -{ - return (struct bkey_s_c) { .k = &k->k, .v = &k->v }; -} - -/* - * For a given type of value (e.g. struct bch_extent), generates the types for - * bkey + bch_extent - inline, split, split const - and also all the conversion - * functions, which also check that the value is of the correct type. - * - * We use anonymous unions for upcasting - e.g. converting from e.g. a - * bkey_i_extent to a bkey_i - since that's always safe, instead of conversion - * functions. - */ -#define x(name, ...) \ -struct bkey_i_##name { \ - union { \ - struct bkey k; \ - struct bkey_i k_i; \ - }; \ - struct bch_##name v; \ -}; \ - \ -struct bkey_s_c_##name { \ - union { \ - struct { \ - const struct bkey *k; \ - const struct bch_##name *v; \ - }; \ - struct bkey_s_c s_c; \ - }; \ -}; \ - \ -struct bkey_s_##name { \ - union { \ - struct { \ - struct bkey *k; \ - struct bch_##name *v; \ - }; \ - struct bkey_s_c_##name c; \ - struct bkey_s s; \ - struct bkey_s_c s_c; \ - }; \ -}; \ - \ -static inline struct bkey_i_##name *bkey_i_to_##name(struct bkey_i *k) \ -{ \ - EBUG_ON(!IS_ERR_OR_NULL(k) && k->k.type != KEY_TYPE_##name); \ - return container_of(&k->k, struct bkey_i_##name, k); \ -} \ - \ -static inline const struct bkey_i_##name * \ -bkey_i_to_##name##_c(const struct bkey_i *k) \ -{ \ - EBUG_ON(!IS_ERR_OR_NULL(k) && k->k.type != KEY_TYPE_##name); \ - return container_of(&k->k, struct bkey_i_##name, k); \ -} \ - \ -static inline struct bkey_s_##name bkey_s_to_##name(struct bkey_s k) \ -{ \ - EBUG_ON(!IS_ERR_OR_NULL(k.k) && k.k->type != KEY_TYPE_##name); \ - return (struct bkey_s_##name) { \ - .k = k.k, \ - .v = container_of(k.v, struct bch_##name, v), \ - }; \ -} \ - \ -static inline struct bkey_s_c_##name bkey_s_c_to_##name(struct bkey_s_c k)\ -{ \ - EBUG_ON(!IS_ERR_OR_NULL(k.k) && k.k->type != KEY_TYPE_##name); \ - return (struct bkey_s_c_##name) { \ - .k = k.k, \ - .v = container_of(k.v, struct bch_##name, v), \ - }; \ -} \ - \ -static inline struct bkey_s_##name name##_i_to_s(struct bkey_i_##name *k)\ -{ \ - return (struct bkey_s_##name) { \ - .k = &k->k, \ - .v = &k->v, \ - }; \ -} \ - \ -static inline struct bkey_s_c_##name \ -name##_i_to_s_c(const struct bkey_i_##name *k) \ -{ \ - return (struct bkey_s_c_##name) { \ - .k = &k->k, \ - .v = &k->v, \ - }; \ -} \ - \ -static inline struct bkey_s_##name bkey_i_to_s_##name(struct bkey_i *k) \ -{ \ - EBUG_ON(!IS_ERR_OR_NULL(k) && k->k.type != KEY_TYPE_##name); \ - return (struct bkey_s_##name) { \ - .k = &k->k, \ - .v = container_of(&k->v, struct bch_##name, v), \ - }; \ -} \ - \ -static inline struct bkey_s_c_##name \ -bkey_i_to_s_c_##name(const struct bkey_i *k) \ -{ \ - EBUG_ON(!IS_ERR_OR_NULL(k) && k->k.type != KEY_TYPE_##name); \ - return (struct bkey_s_c_##name) { \ - .k = &k->k, \ - .v = container_of(&k->v, struct bch_##name, v), \ - }; \ -} \ - \ -static inline struct bkey_i_##name *bkey_##name##_init(struct bkey_i *_k)\ -{ \ - struct bkey_i_##name *k = \ - container_of(&_k->k, struct bkey_i_##name, k); \ - \ - bkey_init(&k->k); \ - memset(&k->v, 0, sizeof(k->v)); \ - k->k.type = KEY_TYPE_##name; \ - set_bkey_val_bytes(&k->k, sizeof(k->v)); \ - \ - return k; \ -} - -BCH_BKEY_TYPES(); -#undef x - /* byte order helpers */ #if __BYTE_ORDER__ == __ORDER_LITTLE_ENDIAN__ diff --git a/fs/bcachefs/bkey_types.h b/fs/bcachefs/bkey_types.h new file mode 100644 index 000000000000..c9ae9e42b385 --- /dev/null +++ b/fs/bcachefs/bkey_types.h @@ -0,0 +1,213 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +#ifndef _BCACHEFS_BKEY_TYPES_H +#define _BCACHEFS_BKEY_TYPES_H + +#include "bcachefs_format.h" + +/* + * bkey_i - bkey with inline value + * bkey_s - bkey with split value + * bkey_s_c - bkey with split value, const + */ + +#define bkey_p_next(_k) vstruct_next(_k) + +static inline struct bkey_i *bkey_next(struct bkey_i *k) +{ + return (struct bkey_i *) ((u64 *) k->_data + k->k.u64s); +} + +#define bkey_val_u64s(_k) ((_k)->u64s - BKEY_U64s) + +static inline size_t bkey_val_bytes(const struct bkey *k) +{ + return bkey_val_u64s(k) * sizeof(u64); +} + +static inline void set_bkey_val_u64s(struct bkey *k, unsigned val_u64s) +{ + unsigned u64s = BKEY_U64s + val_u64s; + + BUG_ON(u64s > U8_MAX); + k->u64s = u64s; +} + +static inline void set_bkey_val_bytes(struct bkey *k, unsigned bytes) +{ + set_bkey_val_u64s(k, DIV_ROUND_UP(bytes, sizeof(u64))); +} + +#define bkey_val_end(_k) ((void *) (((u64 *) (_k).v) + bkey_val_u64s((_k).k))) + +#define bkey_deleted(_k) ((_k)->type == KEY_TYPE_deleted) + +#define bkey_whiteout(_k) \ + ((_k)->type == KEY_TYPE_deleted || (_k)->type == KEY_TYPE_whiteout) + +/* bkey with split value, const */ +struct bkey_s_c { + const struct bkey *k; + const struct bch_val *v; +}; + +/* bkey with split value */ +struct bkey_s { + union { + struct { + struct bkey *k; + struct bch_val *v; + }; + struct bkey_s_c s_c; + }; +}; + +#define bkey_s_null ((struct bkey_s) { .k = NULL }) +#define bkey_s_c_null ((struct bkey_s_c) { .k = NULL }) + +#define bkey_s_err(err) ((struct bkey_s) { .k = ERR_PTR(err) }) +#define bkey_s_c_err(err) ((struct bkey_s_c) { .k = ERR_PTR(err) }) + +static inline struct bkey_s bkey_to_s(struct bkey *k) +{ + return (struct bkey_s) { .k = k, .v = NULL }; +} + +static inline struct bkey_s_c bkey_to_s_c(const struct bkey *k) +{ + return (struct bkey_s_c) { .k = k, .v = NULL }; +} + +static inline struct bkey_s bkey_i_to_s(struct bkey_i *k) +{ + return (struct bkey_s) { .k = &k->k, .v = &k->v }; +} + +static inline struct bkey_s_c bkey_i_to_s_c(const struct bkey_i *k) +{ + return (struct bkey_s_c) { .k = &k->k, .v = &k->v }; +} + +/* + * For a given type of value (e.g. struct bch_extent), generates the types for + * bkey + bch_extent - inline, split, split const - and also all the conversion + * functions, which also check that the value is of the correct type. + * + * We use anonymous unions for upcasting - e.g. converting from e.g. a + * bkey_i_extent to a bkey_i - since that's always safe, instead of conversion + * functions. + */ +#define x(name, ...) \ +struct bkey_i_##name { \ + union { \ + struct bkey k; \ + struct bkey_i k_i; \ + }; \ + struct bch_##name v; \ +}; \ + \ +struct bkey_s_c_##name { \ + union { \ + struct { \ + const struct bkey *k; \ + const struct bch_##name *v; \ + }; \ + struct bkey_s_c s_c; \ + }; \ +}; \ + \ +struct bkey_s_##name { \ + union { \ + struct { \ + struct bkey *k; \ + struct bch_##name *v; \ + }; \ + struct bkey_s_c_##name c; \ + struct bkey_s s; \ + struct bkey_s_c s_c; \ + }; \ +}; \ + \ +static inline struct bkey_i_##name *bkey_i_to_##name(struct bkey_i *k) \ +{ \ + EBUG_ON(!IS_ERR_OR_NULL(k) && k->k.type != KEY_TYPE_##name); \ + return container_of(&k->k, struct bkey_i_##name, k); \ +} \ + \ +static inline const struct bkey_i_##name * \ +bkey_i_to_##name##_c(const struct bkey_i *k) \ +{ \ + EBUG_ON(!IS_ERR_OR_NULL(k) && k->k.type != KEY_TYPE_##name); \ + return container_of(&k->k, struct bkey_i_##name, k); \ +} \ + \ +static inline struct bkey_s_##name bkey_s_to_##name(struct bkey_s k) \ +{ \ + EBUG_ON(!IS_ERR_OR_NULL(k.k) && k.k->type != KEY_TYPE_##name); \ + return (struct bkey_s_##name) { \ + .k = k.k, \ + .v = container_of(k.v, struct bch_##name, v), \ + }; \ +} \ + \ +static inline struct bkey_s_c_##name bkey_s_c_to_##name(struct bkey_s_c k)\ +{ \ + EBUG_ON(!IS_ERR_OR_NULL(k.k) && k.k->type != KEY_TYPE_##name); \ + return (struct bkey_s_c_##name) { \ + .k = k.k, \ + .v = container_of(k.v, struct bch_##name, v), \ + }; \ +} \ + \ +static inline struct bkey_s_##name name##_i_to_s(struct bkey_i_##name *k)\ +{ \ + return (struct bkey_s_##name) { \ + .k = &k->k, \ + .v = &k->v, \ + }; \ +} \ + \ +static inline struct bkey_s_c_##name \ +name##_i_to_s_c(const struct bkey_i_##name *k) \ +{ \ + return (struct bkey_s_c_##name) { \ + .k = &k->k, \ + .v = &k->v, \ + }; \ +} \ + \ +static inline struct bkey_s_##name bkey_i_to_s_##name(struct bkey_i *k) \ +{ \ + EBUG_ON(!IS_ERR_OR_NULL(k) && k->k.type != KEY_TYPE_##name); \ + return (struct bkey_s_##name) { \ + .k = &k->k, \ + .v = container_of(&k->v, struct bch_##name, v), \ + }; \ +} \ + \ +static inline struct bkey_s_c_##name \ +bkey_i_to_s_c_##name(const struct bkey_i *k) \ +{ \ + EBUG_ON(!IS_ERR_OR_NULL(k) && k->k.type != KEY_TYPE_##name); \ + return (struct bkey_s_c_##name) { \ + .k = &k->k, \ + .v = container_of(&k->v, struct bch_##name, v), \ + }; \ +} \ + \ +static inline struct bkey_i_##name *bkey_##name##_init(struct bkey_i *_k)\ +{ \ + struct bkey_i_##name *k = \ + container_of(&_k->k, struct bkey_i_##name, k); \ + \ + bkey_init(&k->k); \ + memset(&k->v, 0, sizeof(k->v)); \ + k->k.type = KEY_TYPE_##name; \ + set_bkey_val_bytes(&k->k, sizeof(k->v)); \ + \ + return k; \ +} + +BCH_BKEY_TYPES(); +#undef x + +#endif /* _BCACHEFS_BKEY_TYPES_H */ diff --git a/fs/bcachefs/btree_cache.c b/fs/bcachefs/btree_cache.c index d7c81beac14a..562561a9a510 100644 --- a/fs/bcachefs/btree_cache.c +++ b/fs/bcachefs/btree_cache.c @@ -1,6 +1,7 @@ // SPDX-License-Identifier: GPL-2.0 #include "bcachefs.h" +#include "bbpos.h" #include "bkey_buf.h" #include "btree_cache.h" #include "btree_io.h" @@ -60,7 +61,7 @@ static void btree_node_data_free(struct bch_fs *c, struct btree *b) clear_btree_node_just_written(b); - kvpfree(b->data, btree_buf_bytes(b)); + kvfree(b->data); b->data = NULL; #ifdef __KERNEL__ kvfree(b->aux_data); @@ -94,7 +95,7 @@ static int btree_node_data_alloc(struct bch_fs *c, struct btree *b, gfp_t gfp) { BUG_ON(b->data || b->aux_data); - b->data = kvpmalloc(btree_buf_bytes(b), gfp); + b->data = kvmalloc(btree_buf_bytes(b), gfp); if (!b->data) return -BCH_ERR_ENOMEM_btree_node_mem_alloc; #ifdef __KERNEL__ @@ -107,7 +108,7 @@ static int btree_node_data_alloc(struct bch_fs *c, struct btree *b, gfp_t gfp) b->aux_data = NULL; #endif if (!b->aux_data) { - kvpfree(b->data, btree_buf_bytes(b)); + kvfree(b->data); b->data = NULL; return -BCH_ERR_ENOMEM_btree_node_mem_alloc; } @@ -208,6 +209,18 @@ static int __btree_node_reclaim(struct bch_fs *c, struct btree *b, bool flush) int ret = 0; lockdep_assert_held(&bc->lock); + + struct bbpos pos = BBPOS(b->c.btree_id, b->key.k.p); + + u64 mask = b->c.level + ? bc->pinned_nodes_interior_mask + : bc->pinned_nodes_leaf_mask; + + if ((mask & BIT_ULL(b->c.btree_id)) && + bbpos_cmp(bc->pinned_nodes_start, pos) < 0 && + bbpos_cmp(bc->pinned_nodes_end, pos) >= 0) + return -BCH_ERR_ENOMEM_btree_node_reclaim; + wait_on_io: if (b->flags & ((1U << BTREE_NODE_dirty)| (1U << BTREE_NODE_read_in_flight)| @@ -408,7 +421,7 @@ void bch2_fs_btree_cache_exit(struct bch_fs *c) if (c->verify_data) list_move(&c->verify_data->list, &bc->live); - kvpfree(c->verify_ondisk, c->opts.btree_node_size); + kvfree(c->verify_ondisk); for (i = 0; i < btree_id_nr_alive(c); i++) { struct btree_root *r = bch2_btree_id_root(c, i); @@ -711,6 +724,9 @@ static noinline struct btree *bch2_btree_node_fill(struct btree_trans *trans, b = bch2_btree_node_mem_alloc(trans, level != 0); if (bch2_err_matches(PTR_ERR_OR_ZERO(b), ENOMEM)) { + if (!path) + return b; + trans->memory_allocation_failure = true; trace_and_count(c, trans_restart_memory_allocation_failure, trans, _THIS_IP_, path); return ERR_PTR(btree_trans_restart(trans, BCH_ERR_transaction_restart_fill_mem_alloc_fail)); @@ -760,8 +776,9 @@ static noinline struct btree *bch2_btree_node_fill(struct btree_trans *trans, } if (!six_relock_type(&b->c.lock, lock_type, seq)) { - if (path) - trace_and_count(c, trans_restart_relock_after_fill, trans, _THIS_IP_, path); + BUG_ON(!path); + + trace_and_count(c, trans_restart_relock_after_fill, trans, _THIS_IP_, path); return ERR_PTR(btree_trans_restart(trans, BCH_ERR_transaction_restart_relock_after_fill)); } @@ -901,7 +918,7 @@ retry: if (unlikely(btree_node_read_error(b))) { six_unlock_type(&b->c.lock, lock_type); - return ERR_PTR(-EIO); + return ERR_PTR(-BCH_ERR_btree_node_read_error); } EBUG_ON(b->c.btree_id != path->btree_id); @@ -992,7 +1009,7 @@ struct btree *bch2_btree_node_get(struct btree_trans *trans, struct btree_path * if (unlikely(btree_node_read_error(b))) { six_unlock_type(&b->c.lock, lock_type); - return ERR_PTR(-EIO); + return ERR_PTR(-BCH_ERR_btree_node_read_error); } EBUG_ON(b->c.btree_id != path->btree_id); @@ -1075,7 +1092,7 @@ lock_node: if (unlikely(btree_node_read_error(b))) { six_unlock_read(&b->c.lock); - b = ERR_PTR(-EIO); + b = ERR_PTR(-BCH_ERR_btree_node_read_error); goto out; } @@ -1096,7 +1113,7 @@ int bch2_btree_node_prefetch(struct btree_trans *trans, struct btree_cache *bc = &c->btree_cache; struct btree *b; - BUG_ON(trans && !btree_node_locked(path, level + 1)); + BUG_ON(path && !btree_node_locked(path, level + 1)); BUG_ON(level >= BTREE_MAX_DEPTH); b = btree_cache_find(bc, k); diff --git a/fs/bcachefs/btree_gc.c b/fs/bcachefs/btree_gc.c index 1102995643b1..bdaed29f084a 100644 --- a/fs/bcachefs/btree_gc.c +++ b/fs/bcachefs/btree_gc.c @@ -389,7 +389,8 @@ again: have_child = dropped_children = false; bch2_bkey_buf_init(&prev_k); bch2_bkey_buf_init(&cur_k); - bch2_btree_and_journal_iter_init_node_iter(&iter, c, b); + bch2_btree_and_journal_iter_init_node_iter(trans, &iter, b); + iter.prefetch = true; while ((k = bch2_btree_and_journal_iter_peek(&iter)).k) { BUG_ON(bpos_lt(k.k->p, b->data->min_key)); @@ -406,7 +407,7 @@ again: printbuf_reset(&buf); bch2_bkey_val_to_text(&buf, c, bkey_i_to_s_c(cur_k.k)); - if (mustfix_fsck_err_on(ret == -EIO, c, + if (mustfix_fsck_err_on(bch2_err_matches(ret, EIO), c, btree_node_unreadable, "Topology repair: unreadable btree node at btree %s level %u:\n" " %s", @@ -478,7 +479,8 @@ again: goto err; bch2_btree_and_journal_iter_exit(&iter); - bch2_btree_and_journal_iter_init_node_iter(&iter, c, b); + bch2_btree_and_journal_iter_init_node_iter(trans, &iter, b); + iter.prefetch = true; while ((k = bch2_btree_and_journal_iter_peek(&iter)).k) { bch2_bkey_buf_reassemble(&cur_k, c, k); @@ -591,16 +593,15 @@ static int bch2_check_fix_ptrs(struct btree_trans *trans, enum btree_id btree_id struct bucket *g = PTR_GC_BUCKET(ca, &p.ptr); enum bch_data_type data_type = bch2_bkey_ptr_data_type(*k, &entry_c->ptr); - if (!g->gen_valid && - (c->opts.reconstruct_alloc || - fsck_err(c, ptr_to_missing_alloc_key, - "bucket %u:%zu data type %s ptr gen %u missing in alloc btree\n" - "while marking %s", - p.ptr.dev, PTR_BUCKET_NR(ca, &p.ptr), - bch2_data_type_str(ptr_data_type(k->k, &p.ptr)), - p.ptr.gen, - (printbuf_reset(&buf), - bch2_bkey_val_to_text(&buf, c, *k), buf.buf)))) { + if (fsck_err_on(!g->gen_valid, + c, ptr_to_missing_alloc_key, + "bucket %u:%zu data type %s ptr gen %u missing in alloc btree\n" + "while marking %s", + p.ptr.dev, PTR_BUCKET_NR(ca, &p.ptr), + bch2_data_type_str(ptr_data_type(k->k, &p.ptr)), + p.ptr.gen, + (printbuf_reset(&buf), + bch2_bkey_val_to_text(&buf, c, *k), buf.buf))) { if (!p.ptr.cached) { g->gen_valid = true; g->gen = p.ptr.gen; @@ -609,16 +610,15 @@ static int bch2_check_fix_ptrs(struct btree_trans *trans, enum btree_id btree_id } } - if (gen_cmp(p.ptr.gen, g->gen) > 0 && - (c->opts.reconstruct_alloc || - fsck_err(c, ptr_gen_newer_than_bucket_gen, - "bucket %u:%zu data type %s ptr gen in the future: %u > %u\n" - "while marking %s", - p.ptr.dev, PTR_BUCKET_NR(ca, &p.ptr), - bch2_data_type_str(ptr_data_type(k->k, &p.ptr)), - p.ptr.gen, g->gen, - (printbuf_reset(&buf), - bch2_bkey_val_to_text(&buf, c, *k), buf.buf)))) { + if (fsck_err_on(gen_cmp(p.ptr.gen, g->gen) > 0, + c, ptr_gen_newer_than_bucket_gen, + "bucket %u:%zu data type %s ptr gen in the future: %u > %u\n" + "while marking %s", + p.ptr.dev, PTR_BUCKET_NR(ca, &p.ptr), + bch2_data_type_str(ptr_data_type(k->k, &p.ptr)), + p.ptr.gen, g->gen, + (printbuf_reset(&buf), + bch2_bkey_val_to_text(&buf, c, *k), buf.buf))) { if (!p.ptr.cached) { g->gen_valid = true; g->gen = p.ptr.gen; @@ -631,28 +631,26 @@ static int bch2_check_fix_ptrs(struct btree_trans *trans, enum btree_id btree_id } } - if (gen_cmp(g->gen, p.ptr.gen) > BUCKET_GC_GEN_MAX && - (c->opts.reconstruct_alloc || - fsck_err(c, ptr_gen_newer_than_bucket_gen, - "bucket %u:%zu gen %u data type %s: ptr gen %u too stale\n" - "while marking %s", - p.ptr.dev, PTR_BUCKET_NR(ca, &p.ptr), g->gen, - bch2_data_type_str(ptr_data_type(k->k, &p.ptr)), - p.ptr.gen, - (printbuf_reset(&buf), - bch2_bkey_val_to_text(&buf, c, *k), buf.buf)))) + if (fsck_err_on(gen_cmp(g->gen, p.ptr.gen) > BUCKET_GC_GEN_MAX, + c, ptr_gen_newer_than_bucket_gen, + "bucket %u:%zu gen %u data type %s: ptr gen %u too stale\n" + "while marking %s", + p.ptr.dev, PTR_BUCKET_NR(ca, &p.ptr), g->gen, + bch2_data_type_str(ptr_data_type(k->k, &p.ptr)), + p.ptr.gen, + (printbuf_reset(&buf), + bch2_bkey_val_to_text(&buf, c, *k), buf.buf))) do_update = true; - if (!p.ptr.cached && gen_cmp(p.ptr.gen, g->gen) < 0 && - (c->opts.reconstruct_alloc || - fsck_err(c, stale_dirty_ptr, - "bucket %u:%zu data type %s stale dirty ptr: %u < %u\n" - "while marking %s", - p.ptr.dev, PTR_BUCKET_NR(ca, &p.ptr), - bch2_data_type_str(ptr_data_type(k->k, &p.ptr)), - p.ptr.gen, g->gen, - (printbuf_reset(&buf), - bch2_bkey_val_to_text(&buf, c, *k), buf.buf)))) + if (fsck_err_on(!p.ptr.cached && gen_cmp(p.ptr.gen, g->gen) < 0, + c, stale_dirty_ptr, + "bucket %u:%zu data type %s stale dirty ptr: %u < %u\n" + "while marking %s", + p.ptr.dev, PTR_BUCKET_NR(ca, &p.ptr), + bch2_data_type_str(ptr_data_type(k->k, &p.ptr)), + p.ptr.gen, g->gen, + (printbuf_reset(&buf), + bch2_bkey_val_to_text(&buf, c, *k), buf.buf))) do_update = true; if (data_type != BCH_DATA_btree && p.ptr.gen != g->gen) @@ -931,7 +929,7 @@ static int bch2_gc_btree_init_recurse(struct btree_trans *trans, struct btree *b struct printbuf buf = PRINTBUF; int ret = 0; - bch2_btree_and_journal_iter_init_node_iter(&iter, c, b); + bch2_btree_and_journal_iter_init_node_iter(trans, &iter, b); bch2_bkey_buf_init(&prev); bch2_bkey_buf_init(&cur); bkey_init(&prev.k->k); @@ -963,7 +961,8 @@ static int bch2_gc_btree_init_recurse(struct btree_trans *trans, struct btree *b if (b->c.level > target_depth) { bch2_btree_and_journal_iter_exit(&iter); - bch2_btree_and_journal_iter_init_node_iter(&iter, c, b); + bch2_btree_and_journal_iter_init_node_iter(trans, &iter, b); + iter.prefetch = true; while ((k = bch2_btree_and_journal_iter_peek(&iter)).k) { struct btree *child; @@ -976,7 +975,7 @@ static int bch2_gc_btree_init_recurse(struct btree_trans *trans, struct btree *b false); ret = PTR_ERR_OR_ZERO(child); - if (ret == -EIO) { + if (bch2_err_matches(ret, EIO)) { bch2_topology_error(c); if (__fsck_err(c, @@ -1190,9 +1189,7 @@ static void bch2_gc_free(struct bch_fs *c) genradix_free(&c->gc_stripes); for_each_member_device(c, ca) { - kvpfree(rcu_dereference_protected(ca->buckets_gc, 1), - sizeof(struct bucket_array) + - ca->mi.nbuckets * sizeof(struct bucket)); + kvfree(rcu_dereference_protected(ca->buckets_gc, 1)); ca->buckets_gc = NULL; free_percpu(ca->usage_gc); @@ -1365,11 +1362,10 @@ static int bch2_alloc_write_key(struct btree_trans *trans, { struct bch_fs *c = trans->c; struct bch_dev *ca = bch_dev_bkey_exists(c, iter->pos.inode); - struct bucket gc, *b; + struct bucket old_gc, gc, *b; struct bkey_i_alloc_v4 *a; struct bch_alloc_v4 old_convert, new; const struct bch_alloc_v4 *old; - enum bch_data_type type; int ret; old = bch2_alloc_to_v4(k, &old_convert); @@ -1377,28 +1373,29 @@ static int bch2_alloc_write_key(struct btree_trans *trans, percpu_down_read(&c->mark_lock); b = gc_bucket(ca, iter->pos.offset); + old_gc = *b; + + if ((old->data_type == BCH_DATA_sb || + old->data_type == BCH_DATA_journal) && + !bch2_dev_is_online(ca)) { + b->data_type = old->data_type; + b->dirty_sectors = old->dirty_sectors; + } /* * b->data_type doesn't yet include need_discard & need_gc_gen states - * fix that here: */ - type = __alloc_data_type(b->dirty_sectors, - b->cached_sectors, - b->stripe, - *old, - b->data_type); - if (b->data_type != type) { - struct bch_dev_usage *u; - - preempt_disable(); - u = this_cpu_ptr(ca->usage_gc); - u->d[b->data_type].buckets--; - b->data_type = type; - u->d[b->data_type].buckets++; - preempt_enable(); - } - + b->data_type = __alloc_data_type(b->dirty_sectors, + b->cached_sectors, + b->stripe, + *old, + b->data_type); gc = *b; + + if (gc.data_type != old_gc.data_type || + gc.dirty_sectors != old_gc.dirty_sectors) + bch2_dev_usage_update_m(c, ca, &old_gc, &gc); percpu_up_read(&c->mark_lock); if (metadata_only && @@ -1410,8 +1407,7 @@ static int bch2_alloc_write_key(struct btree_trans *trans, if (gen_after(old->gen, gc.gen)) return 0; - if (c->opts.reconstruct_alloc || - fsck_err_on(new.data_type != gc.data_type, c, + if (fsck_err_on(new.data_type != gc.data_type, c, alloc_key_data_type_wrong, "bucket %llu:%llu gen %u has wrong data_type" ": got %s, should be %s", @@ -1422,8 +1418,7 @@ static int bch2_alloc_write_key(struct btree_trans *trans, new.data_type = gc.data_type; #define copy_bucket_field(_errtype, _f) \ - if (c->opts.reconstruct_alloc || \ - fsck_err_on(new._f != gc._f, c, _errtype, \ + if (fsck_err_on(new._f != gc._f, c, _errtype, \ "bucket %llu:%llu gen %u data type %s has wrong " #_f \ ": got %u, should be %u", \ iter->pos.inode, iter->pos.offset, \ @@ -1491,7 +1486,7 @@ static int bch2_gc_alloc_done(struct bch_fs *c, bool metadata_only) static int bch2_gc_alloc_start(struct bch_fs *c, bool metadata_only) { for_each_member_device(c, ca) { - struct bucket_array *buckets = kvpmalloc(sizeof(struct bucket_array) + + struct bucket_array *buckets = kvmalloc(sizeof(struct bucket_array) + ca->mi.nbuckets * sizeof(struct bucket), GFP_KERNEL|__GFP_ZERO); if (!buckets) { @@ -1585,8 +1580,7 @@ static int bch2_gc_write_reflink_key(struct btree_trans *trans, " should be %u", (bch2_bkey_val_to_text(&buf, c, k), buf.buf), r->refcount)) { - struct bkey_i *new = bch2_bkey_make_mut(trans, iter, &k, 0); - + struct bkey_i *new = bch2_bkey_make_mut_noupdate(trans, k); ret = PTR_ERR_OR_ZERO(new); if (ret) return ret; @@ -1595,6 +1589,7 @@ static int bch2_gc_write_reflink_key(struct btree_trans *trans, new->k.type = KEY_TYPE_deleted; else *bkey_refcount(bkey_i_to_s(new)) = cpu_to_le64(r->refcount); + ret = bch2_trans_update(trans, iter, new, 0); } fsck_err: printbuf_exit(&buf); @@ -1817,10 +1812,10 @@ out: if (!ret) { bch2_journal_block(&c->journal); - ret = bch2_gc_stripes_done(c, metadata_only) ?: - bch2_gc_reflink_done(c, metadata_only) ?: - bch2_gc_alloc_done(c, metadata_only) ?: - bch2_gc_done(c, initial, metadata_only); + ret = bch2_gc_alloc_done(c, metadata_only) ?: + bch2_gc_done(c, initial, metadata_only) ?: + bch2_gc_stripes_done(c, metadata_only) ?: + bch2_gc_reflink_done(c, metadata_only); bch2_journal_unblock(&c->journal); } diff --git a/fs/bcachefs/btree_io.c b/fs/bcachefs/btree_io.c index aa9b6cbe3226..34df8ccc5fec 100644 --- a/fs/bcachefs/btree_io.c +++ b/fs/bcachefs/btree_io.c @@ -103,7 +103,7 @@ static void btree_bounce_free(struct bch_fs *c, size_t size, if (used_mempool) mempool_free(p, &c->btree_bounce_pool); else - vpfree(p, size); + kvfree(p); } static void *btree_bounce_alloc(struct bch_fs *c, size_t size, @@ -115,7 +115,7 @@ static void *btree_bounce_alloc(struct bch_fs *c, size_t size, BUG_ON(size > c->opts.btree_node_size); *used_mempool = false; - p = vpmalloc(size, __GFP_NOWARN|GFP_NOWAIT); + p = kvmalloc(size, __GFP_NOWARN|GFP_NOWAIT); if (!p) { *used_mempool = true; p = mempool_alloc(&c->btree_bounce_pool, GFP_NOFS); @@ -581,8 +581,7 @@ static int __btree_err(int ret, break; case -BCH_ERR_btree_node_read_err_bad_node: bch2_print_string_as_lines(KERN_ERR, out.buf); - bch2_topology_error(c); - ret = bch2_run_explicit_recovery_pass(c, BCH_RECOVERY_PASS_check_topology) ?: -EIO; + ret = bch2_topology_error(c); break; case -BCH_ERR_btree_node_read_err_incompatible: bch2_print_string_as_lines(KERN_ERR, out.buf); @@ -840,6 +839,9 @@ static bool __bkey_valid(struct bch_fs *c, struct btree *b, if (k->format > KEY_FORMAT_CURRENT) return false; + if (k->u64s < bkeyp_key_u64s(&b->format, k)) + return false; + struct printbuf buf = PRINTBUF; struct bkey tmp; struct bkey_s u = __bkey_disassemble(b, k, &tmp); @@ -881,7 +883,13 @@ static int validate_bset_keys(struct bch_fs *c, struct btree *b, "invalid bkey format %u", k->format)) goto drop_this_key; - /* XXX: validate k->u64s */ + if (btree_err_on(k->u64s < bkeyp_key_u64s(&b->format, k), + -BCH_ERR_btree_node_read_err_fixable, + c, NULL, b, i, + btree_node_bkey_bad_u64s, + "k->u64s too small (%u < %u)", k->u64s, bkeyp_key_u64s(&b->format, k))) + goto drop_this_key; + if (!write) bch2_bkey_compat(b->c.level, b->c.btree_id, version, BSET_BIG_ENDIAN(i), write, @@ -1058,7 +1066,7 @@ int bch2_btree_node_read_done(struct bch_fs *c, struct bch_dev *ca, ret = bset_encrypt(c, i, b->written << 9); if (bch2_fs_fatal_err_on(ret, c, - "error decrypting btree node: %i", ret)) + "decrypting btree node: %s", bch2_err_str(ret))) goto fsck_err; btree_err_on(btree_node_type_is_extents(btree_node_type(b)) && @@ -1099,7 +1107,7 @@ int bch2_btree_node_read_done(struct bch_fs *c, struct bch_dev *ca, ret = bset_encrypt(c, i, b->written << 9); if (bch2_fs_fatal_err_on(ret, c, - "error decrypting btree node: %i\n", ret)) + "decrypting btree node: %s", bch2_err_str(ret))) goto fsck_err; sectors = vstruct_sectors(bne, c->block_bits); @@ -1330,7 +1338,7 @@ start: if (saw_error && !btree_node_read_error(b)) { printbuf_reset(&buf); bch2_bpos_to_text(&buf, b->key.k.p); - bch_info(c, "%s: rewriting btree node at btree=%s level=%u %s due to error", + bch_err_ratelimited(c, "%s: rewriting btree node at btree=%s level=%u %s due to error", __func__, bch2_btree_id_str(b->c.btree_id), b->c.level, buf.buf); bch2_btree_node_rewrite_async(c, b); @@ -1737,7 +1745,7 @@ static int __bch2_btree_root_read(struct btree_trans *trans, enum btree_id id, list_move(&b->list, &c->btree_cache.freeable); mutex_unlock(&c->btree_cache.lock); - ret = -EIO; + ret = -BCH_ERR_btree_node_read_error; goto err; } @@ -1841,7 +1849,7 @@ static void btree_node_write_work(struct work_struct *work) bch2_dev_list_has_dev(wbio->wbio.failed, ptr->dev)); if (!bch2_bkey_nr_ptrs(bkey_i_to_s_c(&wbio->key))) { - ret = -BCH_ERR_btree_write_all_failed; + ret = -BCH_ERR_btree_node_write_all_failed; goto err; } @@ -1866,8 +1874,8 @@ out: return; err: set_btree_node_noevict(b); - if (!bch2_err_matches(ret, EROFS)) - bch2_fs_fatal_error(c, "fatal error writing btree node: %s", bch2_err_str(ret)); + bch2_fs_fatal_err_on(!bch2_err_matches(ret, EROFS), c, + "writing btree node: %s", bch2_err_str(ret)); goto out; } @@ -2123,7 +2131,7 @@ do_write: ret = bset_encrypt(c, i, b->written << 9); if (bch2_fs_fatal_err_on(ret, c, - "error encrypting btree node: %i\n", ret)) + "encrypting btree node: %s", bch2_err_str(ret))) goto err; nonce = btree_nonce(i, b->written << 9); diff --git a/fs/bcachefs/btree_iter.c b/fs/bcachefs/btree_iter.c index 3ef338df82f5..51bcdc6c6d1c 100644 --- a/fs/bcachefs/btree_iter.c +++ b/fs/bcachefs/btree_iter.c @@ -891,7 +891,7 @@ static noinline int btree_node_iter_and_journal_peek(struct btree_trans *trans, struct bkey_s_c k; int ret = 0; - __bch2_btree_and_journal_iter_init_node_iter(&jiter, c, l->b, l->iter, path->pos); + __bch2_btree_and_journal_iter_init_node_iter(trans, &jiter, l->b, l->iter, path->pos); k = bch2_btree_and_journal_iter_peek(&jiter); @@ -1146,7 +1146,7 @@ int bch2_btree_path_traverse_one(struct btree_trans *trans, path = &trans->paths[path_idx]; if (unlikely(path->level >= BTREE_MAX_DEPTH)) - goto out; + goto out_uptodate; path->level = btree_path_up_until_good_node(trans, path, 0); @@ -1179,7 +1179,7 @@ int bch2_btree_path_traverse_one(struct btree_trans *trans, goto out; } } - +out_uptodate: path->uptodate = BTREE_ITER_UPTODATE; out: if (bch2_err_matches(ret, BCH_ERR_transaction_restart) != !!trans->restarted) @@ -1520,7 +1520,7 @@ static noinline void btree_paths_realloc(struct btree_trans *trans) { unsigned nr = trans->nr_paths * 2; - void *p = kzalloc(BITS_TO_LONGS(nr) * sizeof(unsigned long) + + void *p = kvzalloc(BITS_TO_LONGS(nr) * sizeof(unsigned long) + sizeof(struct btree_trans_paths) + nr * sizeof(struct btree_path) + nr * sizeof(btree_path_idx_t) + 8 + @@ -1729,7 +1729,9 @@ bch2_btree_iter_traverse(struct btree_iter *iter) if (ret) return ret; - btree_path_set_should_be_locked(trans->paths + iter->path); + struct btree_path *path = btree_iter_path(trans, iter); + if (btree_path_node(path, path->level)) + btree_path_set_should_be_locked(path); return 0; } @@ -2305,7 +2307,7 @@ struct bkey_s_c bch2_btree_iter_peek_prev(struct btree_iter *iter) btree_iter_path(trans, iter)->level); if (iter->flags & BTREE_ITER_WITH_JOURNAL) - return bkey_s_c_err(-EIO); + return bkey_s_c_err(-BCH_ERR_btree_iter_with_journal_not_supported); bch2_btree_iter_verify(iter); bch2_btree_iter_verify_entry_exit(iter); @@ -2503,6 +2505,7 @@ struct bkey_s_c bch2_btree_iter_peek_slot(struct btree_iter *iter) k = bch2_btree_iter_peek_upto(&iter2, end); if (k.k && !bkey_err(k)) { + swap(iter->key_cache_path, iter2.key_cache_path); iter->k = iter2.k; k.k = &iter->k; } @@ -2762,6 +2765,9 @@ void bch2_trans_copy_iter(struct btree_iter *dst, struct btree_iter *src) struct btree_trans *trans = src->trans; *dst = *src; +#ifdef TRACK_PATH_ALLOCATED + dst->ip_allocated = _RET_IP_; +#endif if (src->path) __btree_path_get(trans->paths + src->path, src->flags & BTREE_ITER_INTENT); if (src->update_path) @@ -3085,7 +3091,7 @@ void bch2_trans_put(struct btree_trans *trans) trans->paths = NULL; if (paths_allocated != trans->_paths_allocated) - kfree_rcu_mightsleep(paths_allocated); + kvfree_rcu_mightsleep(paths_allocated); if (trans->mem_bytes == BTREE_TRANS_MEM_MAX) mempool_free(trans->mem, &c->btree_trans_mem_pool); diff --git a/fs/bcachefs/btree_journal_iter.c b/fs/bcachefs/btree_journal_iter.c index 719a94a84950..50e04356d72c 100644 --- a/fs/bcachefs/btree_journal_iter.c +++ b/fs/bcachefs/btree_journal_iter.c @@ -1,7 +1,9 @@ // SPDX-License-Identifier: GPL-2.0 #include "bcachefs.h" +#include "bkey_buf.h" #include "bset.h" +#include "btree_cache.h" #include "btree_journal_iter.h" #include "journal_io.h" @@ -40,7 +42,7 @@ static inline size_t idx_to_pos(struct journal_keys *keys, size_t idx) static inline struct journal_key *idx_to_key(struct journal_keys *keys, size_t idx) { - return keys->d + idx_to_pos(keys, idx); + return keys->data + idx_to_pos(keys, idx); } static size_t __bch2_journal_key_search(struct journal_keys *keys, @@ -180,10 +182,10 @@ int bch2_journal_key_insert_take(struct bch_fs *c, enum btree_id id, BUG_ON(test_bit(BCH_FS_rw, &c->flags)); if (idx < keys->size && - journal_key_cmp(&n, &keys->d[idx]) == 0) { - if (keys->d[idx].allocated) - kfree(keys->d[idx].k); - keys->d[idx] = n; + journal_key_cmp(&n, &keys->data[idx]) == 0) { + if (keys->data[idx].allocated) + kfree(keys->data[idx].k); + keys->data[idx] = n; return 0; } @@ -196,17 +198,17 @@ int bch2_journal_key_insert_take(struct bch_fs *c, enum btree_id id, .size = max_t(size_t, keys->size, 8) * 2, }; - new_keys.d = kvmalloc_array(new_keys.size, sizeof(new_keys.d[0]), GFP_KERNEL); - if (!new_keys.d) { + new_keys.data = kvmalloc_array(new_keys.size, sizeof(new_keys.data[0]), GFP_KERNEL); + if (!new_keys.data) { bch_err(c, "%s: error allocating new key array (size %zu)", __func__, new_keys.size); return -BCH_ERR_ENOMEM_journal_key_insert; } /* Since @keys was full, there was no gap: */ - memcpy(new_keys.d, keys->d, sizeof(keys->d[0]) * keys->nr); - kvfree(keys->d); - keys->d = new_keys.d; + memcpy(new_keys.data, keys->data, sizeof(keys->data[0]) * keys->nr); + kvfree(keys->data); + keys->data = new_keys.data; keys->nr = new_keys.nr; keys->size = new_keys.size; @@ -216,11 +218,10 @@ int bch2_journal_key_insert_take(struct bch_fs *c, enum btree_id id, journal_iters_move_gap(c, keys->gap, idx); - move_gap(keys->d, keys->nr, keys->size, keys->gap, idx); - keys->gap = idx; + move_gap(keys, idx); keys->nr++; - keys->d[keys->gap++] = n; + keys->data[keys->gap++] = n; journal_iters_fix(c); @@ -267,10 +268,10 @@ void bch2_journal_key_overwritten(struct bch_fs *c, enum btree_id btree, size_t idx = bch2_journal_key_search(keys, btree, level, pos); if (idx < keys->size && - keys->d[idx].btree_id == btree && - keys->d[idx].level == level && - bpos_eq(keys->d[idx].k->k.p, pos)) - keys->d[idx].overwritten = true; + keys->data[idx].btree_id == btree && + keys->data[idx].level == level && + bpos_eq(keys->data[idx].k->k.p, pos)) + keys->data[idx].overwritten = true; } static void bch2_journal_iter_advance(struct journal_iter *iter) @@ -284,16 +285,16 @@ static void bch2_journal_iter_advance(struct journal_iter *iter) static struct bkey_s_c bch2_journal_iter_peek(struct journal_iter *iter) { - struct journal_key *k = iter->keys->d + iter->idx; + struct journal_key *k = iter->keys->data + iter->idx; - while (k < iter->keys->d + iter->keys->size && + while (k < iter->keys->data + iter->keys->size && k->btree_id == iter->btree_id && k->level == iter->level) { if (!k->overwritten) return bkey_i_to_s_c(k->k); bch2_journal_iter_advance(iter); - k = iter->keys->d + iter->idx; + k = iter->keys->data + iter->idx; } return bkey_s_c_null; @@ -334,9 +335,38 @@ void bch2_btree_and_journal_iter_advance(struct btree_and_journal_iter *iter) iter->pos = bpos_successor(iter->pos); } +static void btree_and_journal_iter_prefetch(struct btree_and_journal_iter *_iter) +{ + struct btree_and_journal_iter iter = *_iter; + struct bch_fs *c = iter.trans->c; + unsigned level = iter.journal.level; + struct bkey_buf tmp; + unsigned nr = test_bit(BCH_FS_started, &c->flags) + ? (level > 1 ? 0 : 2) + : (level > 1 ? 1 : 16); + + iter.prefetch = false; + bch2_bkey_buf_init(&tmp); + + while (nr--) { + bch2_btree_and_journal_iter_advance(&iter); + struct bkey_s_c k = bch2_btree_and_journal_iter_peek(&iter); + if (!k.k) + break; + + bch2_bkey_buf_reassemble(&tmp, c, k); + bch2_btree_node_prefetch(iter.trans, NULL, tmp.k, iter.journal.btree_id, level - 1); + } + + bch2_bkey_buf_exit(&tmp, c); +} + struct bkey_s_c bch2_btree_and_journal_iter_peek(struct btree_and_journal_iter *iter) { struct bkey_s_c btree_k, journal_k, ret; + + if (iter->prefetch && iter->journal.level) + btree_and_journal_iter_prefetch(iter); again: if (iter->at_end) return bkey_s_c_null; @@ -376,17 +406,18 @@ void bch2_btree_and_journal_iter_exit(struct btree_and_journal_iter *iter) bch2_journal_iter_exit(&iter->journal); } -void __bch2_btree_and_journal_iter_init_node_iter(struct btree_and_journal_iter *iter, - struct bch_fs *c, +void __bch2_btree_and_journal_iter_init_node_iter(struct btree_trans *trans, + struct btree_and_journal_iter *iter, struct btree *b, struct btree_node_iter node_iter, struct bpos pos) { memset(iter, 0, sizeof(*iter)); + iter->trans = trans; iter->b = b; iter->node_iter = node_iter; - bch2_journal_iter_init(c, &iter->journal, b->c.btree_id, b->c.level, pos); + bch2_journal_iter_init(trans->c, &iter->journal, b->c.btree_id, b->c.level, pos); INIT_LIST_HEAD(&iter->journal.list); iter->pos = b->data->min_key; iter->at_end = false; @@ -396,15 +427,15 @@ void __bch2_btree_and_journal_iter_init_node_iter(struct btree_and_journal_iter * this version is used by btree_gc before filesystem has gone RW and * multithreaded, so uses the journal_iters list: */ -void bch2_btree_and_journal_iter_init_node_iter(struct btree_and_journal_iter *iter, - struct bch_fs *c, +void bch2_btree_and_journal_iter_init_node_iter(struct btree_trans *trans, + struct btree_and_journal_iter *iter, struct btree *b) { struct btree_node_iter node_iter; bch2_btree_node_iter_init_from_start(&node_iter, b); - __bch2_btree_and_journal_iter_init_node_iter(iter, c, b, node_iter, b->data->min_key); - list_add(&iter->journal.list, &c->journal_iters); + __bch2_btree_and_journal_iter_init_node_iter(trans, iter, b, node_iter, b->data->min_key); + list_add(&iter->journal.list, &trans->c->journal_iters); } /* sort and dedup all keys in the journal: */ @@ -415,9 +446,7 @@ void bch2_journal_entries_free(struct bch_fs *c) struct genradix_iter iter; genradix_for_each(&c->journal_entries, iter, i) - if (*i) - kvpfree(*i, offsetof(struct journal_replay, j) + - vstruct_bytes(&(*i)->j)); + kvfree(*i); genradix_free(&c->journal_entries); } @@ -437,22 +466,20 @@ static int journal_sort_key_cmp(const void *_l, const void *_r) void bch2_journal_keys_put(struct bch_fs *c) { struct journal_keys *keys = &c->journal_keys; - struct journal_key *i; BUG_ON(atomic_read(&keys->ref) <= 0); if (!atomic_dec_and_test(&keys->ref)) return; - move_gap(keys->d, keys->nr, keys->size, keys->gap, keys->nr); - keys->gap = keys->nr; + move_gap(keys, keys->nr); - for (i = keys->d; i < keys->d + keys->nr; i++) + darray_for_each(*keys, i) if (i->allocated) kfree(i->k); - kvfree(keys->d); - keys->d = NULL; + kvfree(keys->data); + keys->data = NULL; keys->nr = keys->gap = keys->size = 0; bch2_journal_entries_free(c); @@ -460,83 +487,38 @@ void bch2_journal_keys_put(struct bch_fs *c) static void __journal_keys_sort(struct journal_keys *keys) { - struct journal_key *src, *dst; + sort(keys->data, keys->nr, sizeof(keys->data[0]), journal_sort_key_cmp, NULL); - sort(keys->d, keys->nr, sizeof(keys->d[0]), journal_sort_key_cmp, NULL); + struct journal_key *dst = keys->data; - src = dst = keys->d; - while (src < keys->d + keys->nr) { - while (src + 1 < keys->d + keys->nr && - !journal_key_cmp(src, src + 1)) - src++; + darray_for_each(*keys, src) { + if (src + 1 < &darray_top(*keys) && + !journal_key_cmp(src, src + 1)) + continue; - *dst++ = *src++; + *dst++ = *src; } - keys->nr = dst - keys->d; + keys->nr = dst - keys->data; } int bch2_journal_keys_sort(struct bch_fs *c) { struct genradix_iter iter; struct journal_replay *i, **_i; - struct jset_entry *entry; - struct bkey_i *k; struct journal_keys *keys = &c->journal_keys; - size_t nr_keys = 0, nr_read = 0; - - genradix_for_each(&c->journal_entries, iter, _i) { - i = *_i; - - if (!i || i->ignore) - continue; - - for_each_jset_key(k, entry, &i->j) - nr_keys++; - } - - if (!nr_keys) - return 0; - - keys->size = roundup_pow_of_two(nr_keys); - - keys->d = kvmalloc_array(keys->size, sizeof(keys->d[0]), GFP_KERNEL); - if (!keys->d) { - bch_err(c, "Failed to allocate buffer for sorted journal keys (%zu keys); trying slowpath", - nr_keys); - - do { - keys->size >>= 1; - keys->d = kvmalloc_array(keys->size, sizeof(keys->d[0]), GFP_KERNEL); - } while (!keys->d && keys->size > nr_keys / 8); - - if (!keys->d) { - bch_err(c, "Failed to allocate %zu size buffer for sorted journal keys; exiting", - keys->size); - return -BCH_ERR_ENOMEM_journal_keys_sort; - } - } + size_t nr_read = 0; genradix_for_each(&c->journal_entries, iter, _i) { i = *_i; - if (!i || i->ignore) + if (journal_replay_ignore(i)) continue; cond_resched(); for_each_jset_key(k, entry, &i->j) { - if (keys->nr == keys->size) { - __journal_keys_sort(keys); - - if (keys->nr > keys->size * 7 / 8) { - bch_err(c, "Too many journal keys for slowpath; have %zu compacted, buf size %zu, processed %zu/%zu", - keys->nr, keys->size, nr_read, nr_keys); - return -BCH_ERR_ENOMEM_journal_keys_sort; - } - } - - keys->d[keys->nr++] = (struct journal_key) { + struct journal_key n = (struct journal_key) { .btree_id = entry->btree_id, .level = entry->level, .k = k, @@ -544,6 +526,18 @@ int bch2_journal_keys_sort(struct bch_fs *c) .journal_offset = k->_data - i->j._data, }; + if (darray_push(keys, n)) { + __journal_keys_sort(keys); + + if (keys->nr * 8 > keys->size * 7) { + bch_err(c, "Too many journal keys for slowpath; have %zu compacted, buf size %zu, processed %zu keys at seq %llu", + keys->nr, keys->size, nr_read, le64_to_cpu(i->j.seq)); + return -BCH_ERR_ENOMEM_journal_keys_sort; + } + + BUG_ON(darray_push(keys, n)); + } + nr_read++; } } @@ -551,6 +545,6 @@ int bch2_journal_keys_sort(struct bch_fs *c) __journal_keys_sort(keys); keys->gap = keys->nr; - bch_verbose(c, "Journal keys: %zu read, %zu after sorting and compacting", nr_keys, keys->nr); + bch_verbose(c, "Journal keys: %zu read, %zu after sorting and compacting", nr_read, keys->nr); return 0; } diff --git a/fs/bcachefs/btree_journal_iter.h b/fs/bcachefs/btree_journal_iter.h index 8ca4c100b2e3..c9d19da3ea04 100644 --- a/fs/bcachefs/btree_journal_iter.h +++ b/fs/bcachefs/btree_journal_iter.h @@ -15,6 +15,7 @@ struct journal_iter { */ struct btree_and_journal_iter { + struct btree_trans *trans; struct btree *b; struct btree_node_iter node_iter; struct bkey unpacked; @@ -22,6 +23,7 @@ struct btree_and_journal_iter { struct journal_iter journal; struct bpos pos; bool at_end; + bool prefetch; }; struct bkey_i *bch2_journal_keys_peek_upto(struct bch_fs *, enum btree_id, @@ -29,6 +31,9 @@ struct bkey_i *bch2_journal_keys_peek_upto(struct bch_fs *, enum btree_id, struct bkey_i *bch2_journal_keys_peek_slot(struct bch_fs *, enum btree_id, unsigned, struct bpos); +int bch2_btree_and_journal_iter_prefetch(struct btree_trans *, struct btree_path *, + struct btree_and_journal_iter *); + int bch2_journal_key_insert_take(struct bch_fs *, enum btree_id, unsigned, struct bkey_i *); int bch2_journal_key_insert(struct bch_fs *, enum btree_id, @@ -42,12 +47,11 @@ void bch2_btree_and_journal_iter_advance(struct btree_and_journal_iter *); struct bkey_s_c bch2_btree_and_journal_iter_peek(struct btree_and_journal_iter *); void bch2_btree_and_journal_iter_exit(struct btree_and_journal_iter *); -void __bch2_btree_and_journal_iter_init_node_iter(struct btree_and_journal_iter *, - struct bch_fs *, struct btree *, +void __bch2_btree_and_journal_iter_init_node_iter(struct btree_trans *, + struct btree_and_journal_iter *, struct btree *, struct btree_node_iter, struct bpos); -void bch2_btree_and_journal_iter_init_node_iter(struct btree_and_journal_iter *, - struct bch_fs *, - struct btree *); +void bch2_btree_and_journal_iter_init_node_iter(struct btree_trans *, + struct btree_and_journal_iter *, struct btree *); void bch2_journal_keys_put(struct bch_fs *); diff --git a/fs/bcachefs/btree_key_cache.c b/fs/bcachefs/btree_key_cache.c index 74e52fd28abe..581edcb0911b 100644 --- a/fs/bcachefs/btree_key_cache.c +++ b/fs/bcachefs/btree_key_cache.c @@ -380,9 +380,11 @@ static int btree_key_cache_fill(struct btree_trans *trans, struct bkey_i *new_k = NULL; int ret; - k = bch2_bkey_get_iter(trans, &iter, ck->key.btree_id, ck->key.pos, - BTREE_ITER_KEY_CACHE_FILL| - BTREE_ITER_CACHED_NOFILL); + bch2_trans_iter_init(trans, &iter, ck->key.btree_id, ck->key.pos, + BTREE_ITER_KEY_CACHE_FILL| + BTREE_ITER_CACHED_NOFILL); + iter.flags &= ~BTREE_ITER_WITH_JOURNAL; + k = bch2_btree_iter_peek_slot(&iter); ret = bkey_err(k); if (ret) goto err; @@ -674,7 +676,7 @@ static int btree_key_cache_flush_pos(struct btree_trans *trans, !bch2_err_matches(ret, BCH_ERR_transaction_restart) && !bch2_err_matches(ret, BCH_ERR_journal_reclaim_would_deadlock) && !bch2_journal_error(j), c, - "error flushing key cache: %s", bch2_err_str(ret)); + "flushing key cache: %s", bch2_err_str(ret)); if (ret) goto out; diff --git a/fs/bcachefs/btree_locking.c b/fs/bcachefs/btree_locking.c index 684397442338..b9b151e693ed 100644 --- a/fs/bcachefs/btree_locking.c +++ b/fs/bcachefs/btree_locking.c @@ -747,7 +747,8 @@ void bch2_trans_downgrade(struct btree_trans *trans) return; trans_for_each_path(trans, path, i) - bch2_btree_path_downgrade(trans, path); + if (path->ref) + bch2_btree_path_downgrade(trans, path); } int bch2_trans_relock(struct btree_trans *trans) diff --git a/fs/bcachefs/btree_types.h b/fs/bcachefs/btree_types.h index 4a5a64499eb7..9404d96c38f3 100644 --- a/fs/bcachefs/btree_types.h +++ b/fs/bcachefs/btree_types.h @@ -5,6 +5,7 @@ #include <linux/list.h> #include <linux/rhashtable.h> +#include "bbpos_types.h" #include "btree_key_cache_types.h" #include "buckets_types.h" #include "darray.h" @@ -173,6 +174,11 @@ struct btree_cache { */ struct task_struct *alloc_lock; struct closure_waitlist alloc_wait; + + struct bbpos pinned_nodes_start; + struct bbpos pinned_nodes_end; + u64 pinned_nodes_leaf_mask; + u64 pinned_nodes_interior_mask; }; struct btree_node_iter { @@ -654,6 +660,7 @@ const char *bch2_btree_node_type_str(enum btree_node_type); BIT_ULL(BKEY_TYPE_inodes)| \ BIT_ULL(BKEY_TYPE_stripes)| \ BIT_ULL(BKEY_TYPE_reflink)| \ + BIT_ULL(BKEY_TYPE_subvolumes)| \ BIT_ULL(BKEY_TYPE_btree)) #define BTREE_NODE_TYPE_HAS_ATOMIC_TRIGGERS \ @@ -727,7 +734,7 @@ struct btree_root { __BKEY_PADDED(key, BKEY_BTREE_PTR_VAL_U64s_MAX); u8 level; u8 alive; - s8 error; + s16 error; }; enum btree_gc_coalesce_fail_reason { diff --git a/fs/bcachefs/btree_update.c b/fs/bcachefs/btree_update.c index c3ff365acce9..a4b40c1656a5 100644 --- a/fs/bcachefs/btree_update.c +++ b/fs/bcachefs/btree_update.c @@ -452,7 +452,7 @@ bch2_trans_update_by_path(struct btree_trans *trans, btree_path_idx_t path_idx, * the key cache - but the key has to exist in the btree for that to * work: */ - if (path->cached && bkey_deleted(&i->old_k)) + if (path->cached && !i->old_btree_u64s) return flush_new_cached_update(trans, i, flags, ip); return 0; @@ -788,6 +788,27 @@ int bch2_btree_delete_range(struct bch_fs *c, enum btree_id id, int bch2_btree_bit_mod(struct btree_trans *trans, enum btree_id btree, struct bpos pos, bool set) { + struct bkey_i *k = bch2_trans_kmalloc(trans, sizeof(*k)); + int ret = PTR_ERR_OR_ZERO(k); + if (ret) + return ret; + + bkey_init(&k->k); + k->k.type = set ? KEY_TYPE_set : KEY_TYPE_deleted; + k->k.p = pos; + + struct btree_iter iter; + bch2_trans_iter_init(trans, &iter, btree, pos, BTREE_ITER_INTENT); + + ret = bch2_btree_iter_traverse(&iter) ?: + bch2_trans_update(trans, &iter, k, 0); + bch2_trans_iter_exit(trans, &iter); + return ret; +} + +int bch2_btree_bit_mod_buffered(struct btree_trans *trans, enum btree_id btree, + struct bpos pos, bool set) +{ struct bkey_i k; bkey_init(&k.k); diff --git a/fs/bcachefs/btree_update.h b/fs/bcachefs/btree_update.h index b9382b7b288b..cc7c53e83f89 100644 --- a/fs/bcachefs/btree_update.h +++ b/fs/bcachefs/btree_update.h @@ -63,11 +63,12 @@ int bch2_btree_delete_range(struct bch_fs *, enum btree_id, struct bpos, struct bpos, unsigned, u64 *); int bch2_btree_bit_mod(struct btree_trans *, enum btree_id, struct bpos, bool); +int bch2_btree_bit_mod_buffered(struct btree_trans *, enum btree_id, struct bpos, bool); static inline int bch2_btree_delete_at_buffered(struct btree_trans *trans, enum btree_id btree, struct bpos pos) { - return bch2_btree_bit_mod(trans, btree, pos, false); + return bch2_btree_bit_mod_buffered(trans, btree, pos, false); } int __bch2_insert_snapshot_whiteouts(struct btree_trans *, enum btree_id, diff --git a/fs/bcachefs/btree_update_interior.c b/fs/bcachefs/btree_update_interior.c index 4530b14ff2c3..b2f5f2e50f7e 100644 --- a/fs/bcachefs/btree_update_interior.c +++ b/fs/bcachefs/btree_update_interior.c @@ -25,8 +25,7 @@ #include <linux/random.h> static int bch2_btree_insert_node(struct btree_update *, struct btree_trans *, - btree_path_idx_t, struct btree *, - struct keylist *, unsigned); + btree_path_idx_t, struct btree *, struct keylist *); static void bch2_btree_update_add_new_node(struct btree_update *, struct btree *); static btree_path_idx_t get_unlocked_mut_path(struct btree_trans *trans, @@ -647,7 +646,7 @@ static void btree_update_nodes_written(struct btree_update *as) bch2_trans_unlock(trans); bch2_fs_fatal_err_on(ret && !bch2_journal_error(&c->journal), c, - "%s(): error %s", __func__, bch2_err_str(ret)); + "%s", bch2_err_str(ret)); err: if (as->b) { @@ -1068,13 +1067,18 @@ bch2_btree_update_start(struct btree_trans *trans, struct btree_path *path, flags &= ~BCH_WATERMARK_MASK; flags |= watermark; - if (!(flags & BCH_TRANS_COMMIT_journal_reclaim) && - watermark < c->journal.watermark) { + if (watermark < c->journal.watermark) { struct journal_res res = { 0 }; + unsigned journal_flags = watermark|JOURNAL_RES_GET_CHECK; + + if ((flags & BCH_TRANS_COMMIT_journal_reclaim) && + watermark != BCH_WATERMARK_reclaim) + journal_flags |= JOURNAL_RES_GET_NONBLOCK; ret = drop_locks_do(trans, - bch2_journal_res_get(&c->journal, &res, 1, - watermark|JOURNAL_RES_GET_CHECK)); + bch2_journal_res_get(&c->journal, &res, 1, journal_flags)); + if (bch2_err_matches(ret, BCH_ERR_operation_blocked)) + ret = -BCH_ERR_journal_reclaim_would_deadlock; if (ret) return ERR_PTR(ret); } @@ -1118,6 +1122,7 @@ bch2_btree_update_start(struct btree_trans *trans, struct btree_path *path, closure_init(&as->cl, NULL); as->c = c; as->start_time = start_time; + as->ip_started = _RET_IP_; as->mode = BTREE_INTERIOR_NO_UPDATE; as->took_gc_lock = true; as->btree_id = path->btree_id; @@ -1193,7 +1198,8 @@ bch2_btree_update_start(struct btree_trans *trans, struct btree_path *path, err: bch2_btree_update_free(as, trans); if (!bch2_err_matches(ret, ENOSPC) && - !bch2_err_matches(ret, EROFS)) + !bch2_err_matches(ret, EROFS) && + ret != -BCH_ERR_journal_reclaim_would_deadlock) bch_err_fn_ratelimited(c, ret); return ERR_PTR(ret); } @@ -1208,10 +1214,6 @@ static void bch2_btree_set_root_inmem(struct bch_fs *c, struct btree *b) mutex_unlock(&c->btree_cache.lock); mutex_lock(&c->btree_root_lock); - BUG_ON(btree_node_root(c, b) && - (b->c.level < btree_node_root(c, b)->c.level || - !btree_node_dying(btree_node_root(c, b)))); - bch2_btree_id_root(c, b->c.btree_id)->b = b; mutex_unlock(&c->btree_root_lock); @@ -1477,7 +1479,7 @@ static void btree_split_insert_keys(struct btree_update *as, static int btree_split(struct btree_update *as, struct btree_trans *trans, btree_path_idx_t path, struct btree *b, - struct keylist *keys, unsigned flags) + struct keylist *keys) { struct bch_fs *c = as->c; struct btree *parent = btree_node_parent(trans->paths + path, b); @@ -1578,7 +1580,7 @@ static int btree_split(struct btree_update *as, struct btree_trans *trans, if (parent) { /* Split a non root node */ - ret = bch2_btree_insert_node(as, trans, path, parent, &as->parent_keys, flags); + ret = bch2_btree_insert_node(as, trans, path, parent, &as->parent_keys); if (ret) goto err; } else if (n3) { @@ -1673,7 +1675,6 @@ bch2_btree_insert_keys_interior(struct btree_update *as, * @path_idx: path that points to current node * @b: node to insert keys into * @keys: list of keys to insert - * @flags: transaction commit flags * * Returns: 0 on success, typically transaction restart error on failure * @@ -1683,7 +1684,7 @@ bch2_btree_insert_keys_interior(struct btree_update *as, */ static int bch2_btree_insert_node(struct btree_update *as, struct btree_trans *trans, btree_path_idx_t path_idx, struct btree *b, - struct keylist *keys, unsigned flags) + struct keylist *keys) { struct bch_fs *c = as->c; struct btree_path *path = trans->paths + path_idx; @@ -1739,7 +1740,7 @@ split: return btree_trans_restart(trans, BCH_ERR_transaction_restart_split_race); } - return btree_split(as, trans, path_idx, b, keys, flags); + return btree_split(as, trans, path_idx, b, keys); } int bch2_btree_split_leaf(struct btree_trans *trans, @@ -1747,7 +1748,6 @@ int bch2_btree_split_leaf(struct btree_trans *trans, unsigned flags) { /* btree_split & merge may both cause paths array to be reallocated */ - struct btree *b = path_l(trans->paths + path)->b; struct btree_update *as; unsigned l; @@ -1759,7 +1759,7 @@ int bch2_btree_split_leaf(struct btree_trans *trans, if (IS_ERR(as)) return PTR_ERR(as); - ret = btree_split(as, trans, path, b, NULL, flags); + ret = btree_split(as, trans, path, b, NULL); if (ret) { bch2_btree_update_free(as, trans); return ret; @@ -1775,6 +1775,60 @@ int bch2_btree_split_leaf(struct btree_trans *trans, return ret; } +static void __btree_increase_depth(struct btree_update *as, struct btree_trans *trans, + btree_path_idx_t path_idx) +{ + struct bch_fs *c = as->c; + struct btree_path *path = trans->paths + path_idx; + struct btree *n, *b = bch2_btree_id_root(c, path->btree_id)->b; + + BUG_ON(!btree_node_locked(path, b->c.level)); + + n = __btree_root_alloc(as, trans, b->c.level + 1); + + bch2_btree_update_add_new_node(as, n); + six_unlock_write(&n->c.lock); + + path->locks_want++; + BUG_ON(btree_node_locked(path, n->c.level)); + six_lock_increment(&n->c.lock, SIX_LOCK_intent); + mark_btree_node_locked(trans, path, n->c.level, BTREE_NODE_INTENT_LOCKED); + bch2_btree_path_level_init(trans, path, n); + + n->sib_u64s[0] = U16_MAX; + n->sib_u64s[1] = U16_MAX; + + bch2_keylist_add(&as->parent_keys, &b->key); + btree_split_insert_keys(as, trans, path_idx, n, &as->parent_keys); + + bch2_btree_set_root(as, trans, path, n); + bch2_btree_update_get_open_buckets(as, n); + bch2_btree_node_write(c, n, SIX_LOCK_intent, 0); + bch2_trans_node_add(trans, path, n); + six_unlock_intent(&n->c.lock); + + mutex_lock(&c->btree_cache.lock); + list_add_tail(&b->list, &c->btree_cache.live); + mutex_unlock(&c->btree_cache.lock); + + bch2_trans_verify_locks(trans); +} + +int bch2_btree_increase_depth(struct btree_trans *trans, btree_path_idx_t path, unsigned flags) +{ + struct bch_fs *c = trans->c; + struct btree *b = bch2_btree_id_root(c, trans->paths[path].btree_id)->b; + struct btree_update *as = + bch2_btree_update_start(trans, trans->paths + path, + b->c.level, true, flags); + if (IS_ERR(as)) + return PTR_ERR(as); + + __btree_increase_depth(as, trans, path); + bch2_btree_update_done(as, trans); + return 0; +} + int __bch2_foreground_maybe_merge(struct btree_trans *trans, btree_path_idx_t path, unsigned level, @@ -1845,8 +1899,7 @@ int __bch2_foreground_maybe_merge(struct btree_trans *trans, __func__, buf1.buf, buf2.buf); printbuf_exit(&buf1); printbuf_exit(&buf2); - bch2_topology_error(c); - ret = -EIO; + ret = bch2_topology_error(c); goto err; } @@ -1916,7 +1969,7 @@ int __bch2_foreground_maybe_merge(struct btree_trans *trans, bch2_trans_verify_paths(trans); - ret = bch2_btree_insert_node(as, trans, path, parent, &as->parent_keys, flags); + ret = bch2_btree_insert_node(as, trans, path, parent, &as->parent_keys); if (ret) goto err_free_update; @@ -1987,8 +2040,7 @@ int bch2_btree_node_rewrite(struct btree_trans *trans, if (parent) { bch2_keylist_add(&as->parent_keys, &n->key); - ret = bch2_btree_insert_node(as, trans, iter->path, - parent, &as->parent_keys, flags); + ret = bch2_btree_insert_node(as, trans, iter->path, parent, &as->parent_keys); if (ret) goto err; } else { @@ -2069,7 +2121,7 @@ static void async_btree_node_rewrite_work(struct work_struct *work) ret = bch2_trans_do(c, NULL, NULL, 0, async_btree_node_rewrite_trans(trans, a)); - bch_err_fn(c, ret); + bch_err_fn_ratelimited(c, ret); bch2_write_ref_put(c, BCH_WRITE_REF_node_rewrite); kfree(a); } @@ -2116,7 +2168,7 @@ void bch2_btree_node_rewrite_async(struct bch_fs *c, struct btree *b) bch2_write_ref_get(c, BCH_WRITE_REF_node_rewrite); } - queue_work(c->btree_interior_update_worker, &a->work); + queue_work(c->btree_node_rewrite_worker, &a->work); } void bch2_do_pending_node_rewrites(struct bch_fs *c) @@ -2128,7 +2180,7 @@ void bch2_do_pending_node_rewrites(struct bch_fs *c) list_del(&a->list); bch2_write_ref_get(c, BCH_WRITE_REF_node_rewrite); - queue_work(c->btree_interior_update_worker, &a->work); + queue_work(c->btree_node_rewrite_worker, &a->work); } mutex_unlock(&c->pending_node_rewrites_lock); } @@ -2396,12 +2448,12 @@ void bch2_btree_updates_to_text(struct printbuf *out, struct bch_fs *c) mutex_lock(&c->btree_interior_update_lock); list_for_each_entry(as, &c->btree_interior_update_list, list) - prt_printf(out, "%p m %u w %u r %u j %llu\n", - as, - as->mode, - as->nodes_written, - closure_nr_remaining(&as->cl), - as->journal.seq); + prt_printf(out, "%ps: mode=%u nodes_written=%u cl.remaining=%u journal_seq=%llu\n", + (void *) as->ip_started, + as->mode, + as->nodes_written, + closure_nr_remaining(&as->cl), + as->journal.seq); mutex_unlock(&c->btree_interior_update_lock); } @@ -2465,6 +2517,8 @@ bch2_btree_roots_to_journal_entries(struct bch_fs *c, void bch2_fs_btree_interior_update_exit(struct bch_fs *c) { + if (c->btree_node_rewrite_worker) + destroy_workqueue(c->btree_node_rewrite_worker); if (c->btree_interior_update_worker) destroy_workqueue(c->btree_interior_update_worker); mempool_exit(&c->btree_interior_update_pool); @@ -2485,10 +2539,15 @@ void bch2_fs_btree_interior_update_init_early(struct bch_fs *c) int bch2_fs_btree_interior_update_init(struct bch_fs *c) { c->btree_interior_update_worker = - alloc_workqueue("btree_update", WQ_UNBOUND|WQ_MEM_RECLAIM, 1); + alloc_workqueue("btree_update", WQ_UNBOUND|WQ_MEM_RECLAIM, 8); if (!c->btree_interior_update_worker) return -BCH_ERR_ENOMEM_btree_interior_update_worker_init; + c->btree_node_rewrite_worker = + alloc_ordered_workqueue("btree_node_rewrite", WQ_UNBOUND); + if (!c->btree_node_rewrite_worker) + return -BCH_ERR_ENOMEM_btree_interior_update_worker_init; + if (mempool_init_kmalloc_pool(&c->btree_interior_update_pool, 1, sizeof(struct btree_update))) return -BCH_ERR_ENOMEM_btree_interior_update_pool_init; diff --git a/fs/bcachefs/btree_update_interior.h b/fs/bcachefs/btree_update_interior.h index c593c925d1e3..f651dd48aaa0 100644 --- a/fs/bcachefs/btree_update_interior.h +++ b/fs/bcachefs/btree_update_interior.h @@ -32,6 +32,7 @@ struct btree_update { struct closure cl; struct bch_fs *c; u64 start_time; + unsigned long ip_started; struct list_head list; struct list_head unwritten_list; @@ -119,6 +120,8 @@ struct btree *__bch2_btree_node_alloc_replacement(struct btree_update *, int bch2_btree_split_leaf(struct btree_trans *, btree_path_idx_t, unsigned); +int bch2_btree_increase_depth(struct btree_trans *, btree_path_idx_t, unsigned); + int __bch2_foreground_maybe_merge(struct btree_trans *, btree_path_idx_t, unsigned, unsigned, enum btree_node_sibling); diff --git a/fs/bcachefs/btree_write_buffer.c b/fs/bcachefs/btree_write_buffer.c index ac7844861966..5cbad8445782 100644 --- a/fs/bcachefs/btree_write_buffer.c +++ b/fs/bcachefs/btree_write_buffer.c @@ -378,7 +378,7 @@ static int bch2_btree_write_buffer_flush_locked(struct btree_trans *trans) } } err: - bch2_fs_fatal_err_on(ret, c, "%s: insert error %s", __func__, bch2_err_str(ret)); + bch2_fs_fatal_err_on(ret, c, "%s", bch2_err_str(ret)); trace_write_buffer_flush(trans, wb->flushing.keys.nr, skipped, fast, 0); bch2_journal_pin_drop(j, &wb->flushing.pin); wb->flushing.keys.nr = 0; @@ -574,8 +574,6 @@ void bch2_journal_keys_to_write_buffer_end(struct bch_fs *c, struct journal_keys static int bch2_journal_keys_to_write_buffer(struct bch_fs *c, struct journal_buf *buf) { struct journal_keys_to_wb dst; - struct jset_entry *entry; - struct bkey_i *k; int ret = 0; bch2_journal_keys_to_write_buffer_start(c, &dst, le64_to_cpu(buf->data->seq)); @@ -590,7 +588,9 @@ static int bch2_journal_keys_to_write_buffer(struct bch_fs *c, struct journal_bu entry->type = BCH_JSET_ENTRY_btree_keys; } + spin_lock(&c->journal.lock); buf->need_flush_to_write_buffer = false; + spin_unlock(&c->journal.lock); out: bch2_journal_keys_to_write_buffer_end(c, &dst); return ret; diff --git a/fs/bcachefs/buckets.c b/fs/bcachefs/buckets.c index 54f7826ac498..96edf2c34d43 100644 --- a/fs/bcachefs/buckets.c +++ b/fs/bcachefs/buckets.c @@ -990,8 +990,8 @@ static int __trigger_extent(struct btree_trans *trans, ret = !gc ? bch2_update_cached_sectors_list(trans, p.ptr.dev, disk_sectors) : update_cached_sectors(c, k, p.ptr.dev, disk_sectors, 0, true); - bch2_fs_fatal_err_on(ret && gc, c, "%s(): no replicas entry while updating cached sectors", - __func__); + bch2_fs_fatal_err_on(ret && gc, c, "%s: no replicas entry while updating cached sectors", + bch2_err_str(ret)); if (ret) return ret; } @@ -1020,7 +1020,7 @@ static int __trigger_extent(struct btree_trans *trans, struct printbuf buf = PRINTBUF; bch2_bkey_val_to_text(&buf, c, k); - bch2_fs_fatal_error(c, "%s(): no replicas entry for %s", __func__, buf.buf); + bch2_fs_fatal_error(c, ": no replicas entry for %s", buf.buf); printbuf_exit(&buf); } if (ret) @@ -1053,7 +1053,8 @@ int bch2_trigger_extent(struct btree_trans *trans, (int) bch2_bkey_needs_rebalance(c, old); if (mod) { - int ret = bch2_btree_bit_mod(trans, BTREE_ID_rebalance_work, new.k->p, mod > 0); + int ret = bch2_btree_bit_mod_buffered(trans, BTREE_ID_rebalance_work, + new.k->p, mod > 0); if (ret) return ret; } @@ -1335,7 +1336,7 @@ static void bucket_gens_free_rcu(struct rcu_head *rcu) struct bucket_gens *buckets = container_of(rcu, struct bucket_gens, rcu); - kvpfree(buckets, sizeof(*buckets) + buckets->nbuckets); + kvfree(buckets); } int bch2_dev_buckets_resize(struct bch_fs *c, struct bch_dev *ca, u64 nbuckets) @@ -1345,16 +1346,16 @@ int bch2_dev_buckets_resize(struct bch_fs *c, struct bch_dev *ca, u64 nbuckets) bool resize = ca->bucket_gens != NULL; int ret; - if (!(bucket_gens = kvpmalloc(sizeof(struct bucket_gens) + nbuckets, - GFP_KERNEL|__GFP_ZERO))) { + if (!(bucket_gens = kvmalloc(sizeof(struct bucket_gens) + nbuckets, + GFP_KERNEL|__GFP_ZERO))) { ret = -BCH_ERR_ENOMEM_bucket_gens; goto err; } if ((c->opts.buckets_nouse && - !(buckets_nouse = kvpmalloc(BITS_TO_LONGS(nbuckets) * - sizeof(unsigned long), - GFP_KERNEL|__GFP_ZERO)))) { + !(buckets_nouse = kvmalloc(BITS_TO_LONGS(nbuckets) * + sizeof(unsigned long), + GFP_KERNEL|__GFP_ZERO)))) { ret = -BCH_ERR_ENOMEM_buckets_nouse; goto err; } @@ -1397,8 +1398,7 @@ int bch2_dev_buckets_resize(struct bch_fs *c, struct bch_dev *ca, u64 nbuckets) ret = 0; err: - kvpfree(buckets_nouse, - BITS_TO_LONGS(nbuckets) * sizeof(unsigned long)); + kvfree(buckets_nouse); if (bucket_gens) call_rcu(&bucket_gens->rcu, bucket_gens_free_rcu); @@ -1407,27 +1407,21 @@ err: void bch2_dev_buckets_free(struct bch_dev *ca) { - unsigned i; - - kvpfree(ca->buckets_nouse, - BITS_TO_LONGS(ca->mi.nbuckets) * sizeof(unsigned long)); - kvpfree(rcu_dereference_protected(ca->bucket_gens, 1), - sizeof(struct bucket_gens) + ca->mi.nbuckets); + kvfree(ca->buckets_nouse); + kvfree(rcu_dereference_protected(ca->bucket_gens, 1)); - for (i = 0; i < ARRAY_SIZE(ca->usage); i++) + for (unsigned i = 0; i < ARRAY_SIZE(ca->usage); i++) free_percpu(ca->usage[i]); kfree(ca->usage_base); } int bch2_dev_buckets_alloc(struct bch_fs *c, struct bch_dev *ca) { - unsigned i; - ca->usage_base = kzalloc(sizeof(struct bch_dev_usage), GFP_KERNEL); if (!ca->usage_base) return -BCH_ERR_ENOMEM_usage_init; - for (i = 0; i < ARRAY_SIZE(ca->usage); i++) { + for (unsigned i = 0; i < ARRAY_SIZE(ca->usage); i++) { ca->usage[i] = alloc_percpu(struct bch_dev_usage); if (!ca->usage[i]) return -BCH_ERR_ENOMEM_usage_init; diff --git a/fs/bcachefs/chardev.c b/fs/bcachefs/chardev.c index 226b39c17667..38defa19d52d 100644 --- a/fs/bcachefs/chardev.c +++ b/fs/bcachefs/chardev.c @@ -22,12 +22,6 @@ #include <linux/slab.h> #include <linux/uaccess.h> -__must_check -static int copy_to_user_errcode(void __user *to, const void *from, unsigned long n) -{ - return copy_to_user(to, from, n) ? -EFAULT : 0; -} - /* returns with ref on ca->ref */ static struct bch_dev *bch2_device_lookup(struct bch_fs *c, u64 dev, unsigned flags) @@ -155,19 +149,35 @@ static void bch2_fsck_thread_exit(struct thread_with_stdio *_thr) kfree(thr); } -static int bch2_fsck_offline_thread_fn(void *arg) +static int bch2_fsck_offline_thread_fn(struct thread_with_stdio *stdio) { - struct fsck_thread *thr = container_of(arg, struct fsck_thread, thr); + struct fsck_thread *thr = container_of(stdio, struct fsck_thread, thr); struct bch_fs *c = bch2_fs_open(thr->devs, thr->nr_devs, thr->opts); - thr->thr.thr.ret = PTR_ERR_OR_ZERO(c); - if (!thr->thr.thr.ret) - bch2_fs_stop(c); + if (IS_ERR(c)) + return PTR_ERR(c); - thread_with_stdio_done(&thr->thr); - return 0; + int ret = 0; + if (test_bit(BCH_FS_errors_fixed, &c->flags)) + ret |= 1; + if (test_bit(BCH_FS_error, &c->flags)) + ret |= 4; + + bch2_fs_stop(c); + + if (ret & 1) + bch2_stdio_redirect_printf(&stdio->stdio, false, "%s: errors fixed\n", c->name); + if (ret & 4) + bch2_stdio_redirect_printf(&stdio->stdio, false, "%s: still has errors\n", c->name); + + return ret; } +static const struct thread_with_stdio_ops bch2_offline_fsck_ops = { + .exit = bch2_fsck_thread_exit, + .fn = bch2_fsck_offline_thread_fn, +}; + static long bch2_ioctl_fsck_offline(struct bch_ioctl_fsck_offline __user *user_arg) { struct bch_ioctl_fsck_offline arg; @@ -220,9 +230,7 @@ static long bch2_ioctl_fsck_offline(struct bch_ioctl_fsck_offline __user *user_a opt_set(thr->opts, stdio, (u64)(unsigned long)&thr->thr.stdio); - ret = bch2_run_thread_with_stdio(&thr->thr, - bch2_fsck_thread_exit, - bch2_fsck_offline_thread_fn); + ret = bch2_run_thread_with_stdio(&thr->thr, &bch2_offline_fsck_ops); err: if (ret < 0) { if (thr) @@ -763,9 +771,9 @@ static long bch2_ioctl_disk_resize_journal(struct bch_fs *c, return ret; } -static int bch2_fsck_online_thread_fn(void *arg) +static int bch2_fsck_online_thread_fn(struct thread_with_stdio *stdio) { - struct fsck_thread *thr = container_of(arg, struct fsck_thread, thr); + struct fsck_thread *thr = container_of(stdio, struct fsck_thread, thr); struct bch_fs *c = thr->c; c->stdio_filter = current; @@ -793,13 +801,16 @@ static int bch2_fsck_online_thread_fn(void *arg) c->stdio_filter = NULL; c->opts.fix_errors = old_fix_errors; - thread_with_stdio_done(&thr->thr); - up(&c->online_fsck_mutex); bch2_ro_ref_put(c); - return 0; + return ret; } +static const struct thread_with_stdio_ops bch2_online_fsck_ops = { + .exit = bch2_fsck_thread_exit, + .fn = bch2_fsck_online_thread_fn, +}; + static long bch2_ioctl_fsck_online(struct bch_fs *c, struct bch_ioctl_fsck_online arg) { @@ -840,9 +851,7 @@ static long bch2_ioctl_fsck_online(struct bch_fs *c, goto err; } - ret = bch2_run_thread_with_stdio(&thr->thr, - bch2_fsck_thread_exit, - bch2_fsck_online_thread_fn); + ret = bch2_run_thread_with_stdio(&thr->thr, &bch2_online_fsck_ops); err: if (ret < 0) { bch_err_fn(c, ret); diff --git a/fs/bcachefs/checksum.c b/fs/bcachefs/checksum.c index 3c761ad6b1c8..4701457f6381 100644 --- a/fs/bcachefs/checksum.c +++ b/fs/bcachefs/checksum.c @@ -558,7 +558,7 @@ got_key: return 0; } -#include "../crypto.h" +#include "crypto.h" #endif int bch2_request_key(struct bch_sb *sb, struct bch_key *key) diff --git a/fs/bcachefs/compress.c b/fs/bcachefs/compress.c index 33df8cf86bd8..1410365a8891 100644 --- a/fs/bcachefs/compress.c +++ b/fs/bcachefs/compress.c @@ -601,13 +601,13 @@ static int __bch2_fs_compress_init(struct bch_fs *c, u64 features) return 0; if (!mempool_initialized(&c->compression_bounce[READ]) && - mempool_init_kvpmalloc_pool(&c->compression_bounce[READ], - 1, c->opts.encoded_extent_max)) + mempool_init_kvmalloc_pool(&c->compression_bounce[READ], + 1, c->opts.encoded_extent_max)) return -BCH_ERR_ENOMEM_compression_bounce_read_init; if (!mempool_initialized(&c->compression_bounce[WRITE]) && - mempool_init_kvpmalloc_pool(&c->compression_bounce[WRITE], - 1, c->opts.encoded_extent_max)) + mempool_init_kvmalloc_pool(&c->compression_bounce[WRITE], + 1, c->opts.encoded_extent_max)) return -BCH_ERR_ENOMEM_compression_bounce_write_init; for (i = compression_types; @@ -622,15 +622,15 @@ static int __bch2_fs_compress_init(struct bch_fs *c, u64 features) if (mempool_initialized(&c->compress_workspace[i->type])) continue; - if (mempool_init_kvpmalloc_pool( + if (mempool_init_kvmalloc_pool( &c->compress_workspace[i->type], 1, i->compress_workspace)) return -BCH_ERR_ENOMEM_compression_workspace_init; } if (!mempool_initialized(&c->decompress_workspace) && - mempool_init_kvpmalloc_pool(&c->decompress_workspace, - 1, decompress_workspace_size)) + mempool_init_kvmalloc_pool(&c->decompress_workspace, + 1, decompress_workspace_size)) return -BCH_ERR_ENOMEM_decompression_workspace_init; return 0; diff --git a/fs/bcachefs/debug.c b/fs/bcachefs/debug.c index 7bdba8507fc9..208ce6f0fc43 100644 --- a/fs/bcachefs/debug.c +++ b/fs/bcachefs/debug.c @@ -137,7 +137,7 @@ void __bch2_btree_verify(struct bch_fs *c, struct btree *b) mutex_lock(&c->verify_lock); if (!c->verify_ondisk) { - c->verify_ondisk = kvpmalloc(btree_buf_bytes(b), GFP_KERNEL); + c->verify_ondisk = kvmalloc(btree_buf_bytes(b), GFP_KERNEL); if (!c->verify_ondisk) goto out; } @@ -170,7 +170,7 @@ void __bch2_btree_verify(struct bch_fs *c, struct btree *b) struct printbuf buf = PRINTBUF; bch2_bkey_val_to_text(&buf, c, bkey_i_to_s_c(&b->key)); - bch2_fs_fatal_error(c, "btree node verify failed for : %s\n", buf.buf); + bch2_fs_fatal_error(c, ": btree node verify failed for: %s\n", buf.buf); printbuf_exit(&buf); } out: @@ -199,7 +199,7 @@ void bch2_btree_node_ondisk_to_text(struct printbuf *out, struct bch_fs *c, return; } - n_ondisk = kvpmalloc(btree_buf_bytes(b), GFP_KERNEL); + n_ondisk = kvmalloc(btree_buf_bytes(b), GFP_KERNEL); if (!n_ondisk) { prt_printf(out, "memory allocation failure\n"); goto out; @@ -293,7 +293,7 @@ void bch2_btree_node_ondisk_to_text(struct printbuf *out, struct bch_fs *c, out: if (bio) bio_put(bio); - kvpfree(n_ondisk, btree_buf_bytes(b)); + kvfree(n_ondisk); percpu_ref_put(&ca->io_ref); } diff --git a/fs/bcachefs/dirent.c b/fs/bcachefs/dirent.c index 4ae1e9f002a0..d37bd07afbfe 100644 --- a/fs/bcachefs/dirent.c +++ b/fs/bcachefs/dirent.c @@ -144,19 +144,21 @@ fsck_err: return ret; } -void bch2_dirent_to_text(struct printbuf *out, struct bch_fs *c, - struct bkey_s_c k) +void bch2_dirent_to_text(struct printbuf *out, struct bch_fs *c, struct bkey_s_c k) { struct bkey_s_c_dirent d = bkey_s_c_to_dirent(k); struct qstr d_name = bch2_dirent_get_name(d); - prt_printf(out, "%.*s -> %llu type %s", - d_name.len, - d_name.name, - d.v->d_type != DT_SUBVOL - ? le64_to_cpu(d.v->d_inum) - : le32_to_cpu(d.v->d_child_subvol), - bch2_d_type_str(d.v->d_type)); + prt_printf(out, "%.*s -> ", d_name.len, d_name.name); + + if (d.v->d_type != DT_SUBVOL) + prt_printf(out, "%llu", le64_to_cpu(d.v->d_inum)); + else + prt_printf(out, "%u -> %u", + le32_to_cpu(d.v->d_parent_subvol), + le32_to_cpu(d.v->d_child_subvol)); + + prt_printf(out, " type %s", bch2_d_type_str(d.v->d_type)); } static struct bkey_i_dirent *dirent_create_key(struct btree_trans *trans, @@ -199,17 +201,17 @@ static struct bkey_i_dirent *dirent_create_key(struct btree_trans *trans, } int bch2_dirent_create_snapshot(struct btree_trans *trans, - u64 dir, u32 snapshot, + u32 dir_subvol, u64 dir, u32 snapshot, const struct bch_hash_info *hash_info, u8 type, const struct qstr *name, u64 dst_inum, u64 *dir_offset, bch_str_hash_flags_t str_hash_flags) { - subvol_inum zero_inum = { 0 }; + subvol_inum dir_inum = { .subvol = dir_subvol, .inum = dir }; struct bkey_i_dirent *dirent; int ret; - dirent = dirent_create_key(trans, zero_inum, type, name, dst_inum); + dirent = dirent_create_key(trans, dir_inum, type, name, dst_inum); ret = PTR_ERR_OR_ZERO(dirent); if (ret) return ret; @@ -217,10 +219,10 @@ int bch2_dirent_create_snapshot(struct btree_trans *trans, dirent->k.p.inode = dir; dirent->k.p.snapshot = snapshot; - ret = bch2_hash_set_snapshot(trans, bch2_dirent_hash_desc, hash_info, - zero_inum, snapshot, - &dirent->k_i, str_hash_flags, - BTREE_UPDATE_INTERNAL_SNAPSHOT_NODE); + ret = bch2_hash_set_in_snapshot(trans, bch2_dirent_hash_desc, hash_info, + dir_inum, snapshot, + &dirent->k_i, str_hash_flags, + BTREE_UPDATE_INTERNAL_SNAPSHOT_NODE); *dir_offset = dirent->k.p.offset; return ret; @@ -291,12 +293,10 @@ int bch2_dirent_rename(struct btree_trans *trans, struct bkey_i_dirent *new_src = NULL, *new_dst = NULL; struct bpos dst_pos = POS(dst_dir.inum, bch2_dirent_hash(dst_hash, dst_name)); - unsigned src_type = 0, dst_type = 0, src_update_flags = 0; + unsigned src_update_flags = 0; + bool delete_src, delete_dst; int ret = 0; - if (src_dir.subvol != dst_dir.subvol) - return -EXDEV; - memset(src_inum, 0, sizeof(*src_inum)); memset(dst_inum, 0, sizeof(*dst_inum)); @@ -317,12 +317,6 @@ int bch2_dirent_rename(struct btree_trans *trans, if (ret) goto out; - src_type = bkey_s_c_to_dirent(old_src).v->d_type; - - if (src_type == DT_SUBVOL && mode == BCH_RENAME_EXCHANGE) - return -EOPNOTSUPP; - - /* Lookup dst: */ if (mode == BCH_RENAME) { /* @@ -350,11 +344,6 @@ int bch2_dirent_rename(struct btree_trans *trans, bkey_s_c_to_dirent(old_dst), dst_inum); if (ret) goto out; - - dst_type = bkey_s_c_to_dirent(old_dst).v->d_type; - - if (dst_type == DT_SUBVOL) - return -EOPNOTSUPP; } if (mode != BCH_RENAME_EXCHANGE) @@ -424,28 +413,55 @@ int bch2_dirent_rename(struct btree_trans *trans, } } + if (new_dst->v.d_type == DT_SUBVOL) + new_dst->v.d_parent_subvol = cpu_to_le32(dst_dir.subvol); + + if ((mode == BCH_RENAME_EXCHANGE) && + new_src->v.d_type == DT_SUBVOL) + new_src->v.d_parent_subvol = cpu_to_le32(src_dir.subvol); + ret = bch2_trans_update(trans, &dst_iter, &new_dst->k_i, 0); if (ret) goto out; out_set_src: - /* - * If we're deleting a subvolume, we need to really delete the dirent, - * not just emit a whiteout in the current snapshot: + * If we're deleting a subvolume we need to really delete the dirent, + * not just emit a whiteout in the current snapshot - there can only be + * single dirent that points to a given subvolume. + * + * IOW, we don't maintain multiple versions in different snapshots of + * dirents that point to subvolumes - dirents that point to subvolumes + * are only visible in one particular subvolume so it's not necessary, + * and it would be particularly confusing for fsck to have to deal with. */ - if (src_type == DT_SUBVOL) { - bch2_btree_iter_set_snapshot(&src_iter, old_src.k->p.snapshot); - ret = bch2_btree_iter_traverse(&src_iter); + delete_src = bkey_s_c_to_dirent(old_src).v->d_type == DT_SUBVOL && + new_src->k.p.snapshot != old_src.k->p.snapshot; + + delete_dst = old_dst.k && + bkey_s_c_to_dirent(old_dst).v->d_type == DT_SUBVOL && + new_dst->k.p.snapshot != old_dst.k->p.snapshot; + + if (!delete_src || !bkey_deleted(&new_src->k)) { + ret = bch2_trans_update(trans, &src_iter, &new_src->k_i, src_update_flags); if (ret) goto out; + } - new_src->k.p = src_iter.pos; - src_update_flags |= BTREE_UPDATE_INTERNAL_SNAPSHOT_NODE; + if (delete_src) { + bch2_btree_iter_set_snapshot(&src_iter, old_src.k->p.snapshot); + ret = bch2_btree_iter_traverse(&src_iter) ?: + bch2_btree_delete_at(trans, &src_iter, BTREE_UPDATE_INTERNAL_SNAPSHOT_NODE); + if (ret) + goto out; } - ret = bch2_trans_update(trans, &src_iter, &new_src->k_i, src_update_flags); - if (ret) - goto out; + if (delete_dst) { + bch2_btree_iter_set_snapshot(&dst_iter, old_dst.k->p.snapshot); + ret = bch2_btree_iter_traverse(&dst_iter) ?: + bch2_btree_delete_at(trans, &dst_iter, BTREE_UPDATE_INTERNAL_SNAPSHOT_NODE); + if (ret) + goto out; + } if (mode == BCH_RENAME_EXCHANGE) *src_offset = new_src->k.p.offset; @@ -456,41 +472,29 @@ out: return ret; } -int __bch2_dirent_lookup_trans(struct btree_trans *trans, - struct btree_iter *iter, - subvol_inum dir, - const struct bch_hash_info *hash_info, - const struct qstr *name, subvol_inum *inum, - unsigned flags) +int bch2_dirent_lookup_trans(struct btree_trans *trans, + struct btree_iter *iter, + subvol_inum dir, + const struct bch_hash_info *hash_info, + const struct qstr *name, subvol_inum *inum, + unsigned flags) { - struct bkey_s_c k; - struct bkey_s_c_dirent d; - u32 snapshot; - int ret; - - ret = bch2_subvolume_get_snapshot(trans, dir.subvol, &snapshot); - if (ret) - return ret; - - ret = bch2_hash_lookup(trans, iter, bch2_dirent_hash_desc, - hash_info, dir, name, flags); + int ret = bch2_hash_lookup(trans, iter, bch2_dirent_hash_desc, + hash_info, dir, name, flags); if (ret) return ret; - k = bch2_btree_iter_peek_slot(iter); + struct bkey_s_c k = bch2_btree_iter_peek_slot(iter); ret = bkey_err(k); if (ret) goto err; - d = bkey_s_c_to_dirent(k); - - ret = bch2_dirent_read_target(trans, dir, d, inum); + ret = bch2_dirent_read_target(trans, dir, bkey_s_c_to_dirent(k), inum); if (ret > 0) ret = -ENOENT; err: if (ret) bch2_trans_iter_exit(trans, iter); - return ret; } @@ -502,13 +506,13 @@ u64 bch2_dirent_lookup(struct bch_fs *c, subvol_inum dir, struct btree_iter iter = { NULL }; int ret = lockrestart_do(trans, - __bch2_dirent_lookup_trans(trans, &iter, dir, hash_info, name, inum, 0)); + bch2_dirent_lookup_trans(trans, &iter, dir, hash_info, name, inum, 0)); bch2_trans_iter_exit(trans, &iter); bch2_trans_put(trans); return ret; } -int bch2_empty_dir_snapshot(struct btree_trans *trans, u64 dir, u32 snapshot) +int bch2_empty_dir_snapshot(struct btree_trans *trans, u64 dir, u32 subvol, u32 snapshot) { struct btree_iter iter; struct bkey_s_c k; @@ -518,7 +522,10 @@ int bch2_empty_dir_snapshot(struct btree_trans *trans, u64 dir, u32 snapshot) SPOS(dir, 0, snapshot), POS(dir, U64_MAX), 0, k, ret) if (k.k->type == KEY_TYPE_dirent) { - ret = -ENOTEMPTY; + struct bkey_s_c_dirent d = bkey_s_c_to_dirent(k); + if (d.v->d_type == DT_SUBVOL && le32_to_cpu(d.v->d_parent_subvol) != subvol) + continue; + ret = -BCH_ERR_ENOTEMPTY_dir_not_empty; break; } bch2_trans_iter_exit(trans, &iter); @@ -531,7 +538,7 @@ int bch2_empty_dir_trans(struct btree_trans *trans, subvol_inum dir) u32 snapshot; return bch2_subvolume_get_snapshot(trans, dir.subvol, &snapshot) ?: - bch2_empty_dir_snapshot(trans, dir.inum, snapshot); + bch2_empty_dir_snapshot(trans, dir.inum, dir.subvol, snapshot); } int bch2_readdir(struct bch_fs *c, subvol_inum inum, struct dir_context *ctx) diff --git a/fs/bcachefs/dirent.h b/fs/bcachefs/dirent.h index 21ffeb78f02e..bee55cca2aa0 100644 --- a/fs/bcachefs/dirent.h +++ b/fs/bcachefs/dirent.h @@ -35,7 +35,7 @@ static inline unsigned dirent_val_u64s(unsigned len) int bch2_dirent_read_target(struct btree_trans *, subvol_inum, struct bkey_s_c_dirent, subvol_inum *); -int bch2_dirent_create_snapshot(struct btree_trans *, u64, u32, +int bch2_dirent_create_snapshot(struct btree_trans *, u32, u64, u32, const struct bch_hash_info *, u8, const struct qstr *, u64, u64 *, bch_str_hash_flags_t); @@ -62,14 +62,14 @@ int bch2_dirent_rename(struct btree_trans *, const struct qstr *, subvol_inum *, u64 *, enum bch_rename_mode); -int __bch2_dirent_lookup_trans(struct btree_trans *, struct btree_iter *, +int bch2_dirent_lookup_trans(struct btree_trans *, struct btree_iter *, subvol_inum, const struct bch_hash_info *, const struct qstr *, subvol_inum *, unsigned); u64 bch2_dirent_lookup(struct bch_fs *, subvol_inum, const struct bch_hash_info *, const struct qstr *, subvol_inum *); -int bch2_empty_dir_snapshot(struct btree_trans *, u64, u32); +int bch2_empty_dir_snapshot(struct btree_trans *, u64, u32, u32); int bch2_empty_dir_trans(struct btree_trans *, subvol_inum); int bch2_readdir(struct bch_fs *, subvol_inum, struct dir_context *); diff --git a/fs/bcachefs/ec.c b/fs/bcachefs/ec.c index d503af270024..082075244e16 100644 --- a/fs/bcachefs/ec.c +++ b/fs/bcachefs/ec.c @@ -448,7 +448,7 @@ int bch2_trigger_stripe(struct btree_trans *trans, struct printbuf buf = PRINTBUF; bch2_bkey_val_to_text(&buf, c, new); - bch2_fs_fatal_error(c, "no replicas entry for %s", buf.buf); + bch2_fs_fatal_error(c, ": no replicas entry for %s", buf.buf); printbuf_exit(&buf); return ret; } @@ -504,7 +504,7 @@ static void ec_stripe_buf_exit(struct ec_stripe_buf *buf) unsigned i; for (i = 0; i < s->v.nr_blocks; i++) { - kvpfree(buf->data[i], buf->size << 9); + kvfree(buf->data[i]); buf->data[i] = NULL; } } @@ -531,7 +531,7 @@ static int ec_stripe_buf_init(struct ec_stripe_buf *buf, memset(buf->valid, 0xFF, sizeof(buf->valid)); for (i = 0; i < v->nr_blocks; i++) { - buf->data[i] = kvpmalloc(buf->size << 9, GFP_KERNEL); + buf->data[i] = kvmalloc(buf->size << 9, GFP_KERNEL); if (!buf->data[i]) goto err; } @@ -1868,10 +1868,10 @@ static int __bch2_ec_stripe_head_reuse(struct btree_trans *trans, struct ec_stri return -BCH_ERR_stripe_alloc_blocked; ret = get_stripe_key_trans(trans, idx, &h->s->existing_stripe); + bch2_fs_fatal_err_on(ret && !bch2_err_matches(ret, BCH_ERR_transaction_restart), c, + "reading stripe key: %s", bch2_err_str(ret)); if (ret) { bch2_stripe_close(c, h->s); - if (!bch2_err_matches(ret, BCH_ERR_transaction_restart)) - bch2_fs_fatal_error(c, "error reading stripe key: %s", bch2_err_str(ret)); return ret; } diff --git a/fs/bcachefs/errcode.c b/fs/bcachefs/errcode.c index d260ff9bbfeb..43557bebd0f8 100644 --- a/fs/bcachefs/errcode.c +++ b/fs/bcachefs/errcode.c @@ -2,6 +2,7 @@ #include "bcachefs.h" #include "errcode.h" +#include "trace.h" #include <linux/errname.h> @@ -49,15 +50,17 @@ bool __bch2_err_matches(int err, int class) return err == class; } -int __bch2_err_class(int err) +int __bch2_err_class(int bch_err) { - err = -err; - BUG_ON((unsigned) err >= BCH_ERR_MAX); + int std_err = -bch_err; + BUG_ON((unsigned) std_err >= BCH_ERR_MAX); - while (err >= BCH_ERR_START && bch2_errcode_parents[err - BCH_ERR_START]) - err = bch2_errcode_parents[err - BCH_ERR_START]; + while (std_err >= BCH_ERR_START && bch2_errcode_parents[std_err - BCH_ERR_START]) + std_err = bch2_errcode_parents[std_err - BCH_ERR_START]; + + trace_error_downcast(bch_err, std_err, _RET_IP_); - return -err; + return -std_err; } const char *bch2_blk_status_to_str(blk_status_t status) diff --git a/fs/bcachefs/errcode.h b/fs/bcachefs/errcode.h index 8c40c2067a04..af25d8ec60f2 100644 --- a/fs/bcachefs/errcode.h +++ b/fs/bcachefs/errcode.h @@ -5,6 +5,10 @@ #define BCH_ERRCODES() \ x(ERANGE, ERANGE_option_too_small) \ x(ERANGE, ERANGE_option_too_big) \ + x(EINVAL, mount_option) \ + x(BCH_ERR_mount_option, option_name) \ + x(BCH_ERR_mount_option, option_value) \ + x(BCH_ERR_mount_option, option_not_bool) \ x(ENOMEM, ENOMEM_stripe_buf) \ x(ENOMEM, ENOMEM_replicas_table) \ x(ENOMEM, ENOMEM_cpu_replicas) \ @@ -78,6 +82,7 @@ x(ENOMEM, ENOMEM_fs_name_alloc) \ x(ENOMEM, ENOMEM_fs_other_alloc) \ x(ENOMEM, ENOMEM_dev_alloc) \ + x(ENOMEM, ENOMEM_disk_accounting) \ x(ENOSPC, ENOSPC_disk_reservation) \ x(ENOSPC, ENOSPC_bucket_alloc) \ x(ENOSPC, ENOSPC_disk_label_add) \ @@ -109,6 +114,8 @@ x(ENOENT, ENOENT_dirent_doesnt_match_inode) \ x(ENOENT, ENOENT_dev_not_found) \ x(ENOENT, ENOENT_dev_idx_not_found) \ + x(ENOTEMPTY, ENOTEMPTY_dir_not_empty) \ + x(ENOTEMPTY, ENOTEMPTY_subvol_not_empty) \ x(0, open_buckets_empty) \ x(0, freelist_empty) \ x(BCH_ERR_freelist_empty, no_buckets_found) \ @@ -176,6 +183,9 @@ x(EINVAL, invalid) \ x(EINVAL, internal_fsck_err) \ x(EINVAL, opt_parse_error) \ + x(EINVAL, remove_with_metadata_missing_unimplemented)\ + x(EINVAL, remove_would_lose_data) \ + x(EINVAL, btree_iter_with_journal_not_supported) \ x(EROFS, erofs_trans_commit) \ x(EROFS, erofs_no_writes) \ x(EROFS, erofs_journal_err) \ @@ -225,7 +235,10 @@ x(BCH_ERR_operation_blocked, nocow_lock_blocked) \ x(EIO, btree_node_read_err) \ x(EIO, sb_not_downgraded) \ - x(EIO, btree_write_all_failed) \ + x(EIO, btree_node_write_all_failed) \ + x(EIO, btree_node_read_error) \ + x(EIO, btree_node_read_validate_error) \ + x(EIO, btree_need_topology_repair) \ x(BCH_ERR_btree_node_read_err, btree_node_read_err_fixable) \ x(BCH_ERR_btree_node_read_err, btree_node_read_err_want_retry) \ x(BCH_ERR_btree_node_read_err, btree_node_read_err_must_retry) \ @@ -238,7 +251,8 @@ x(BCH_ERR_nopromote, nopromote_congested) \ x(BCH_ERR_nopromote, nopromote_in_flight) \ x(BCH_ERR_nopromote, nopromote_no_writes) \ - x(BCH_ERR_nopromote, nopromote_enomem) + x(BCH_ERR_nopromote, nopromote_enomem) \ + x(0, need_inode_lock) enum bch_errcode { BCH_ERR_START = 2048, diff --git a/fs/bcachefs/error.c b/fs/bcachefs/error.c index d32c8bebe46c..043431206799 100644 --- a/fs/bcachefs/error.c +++ b/fs/bcachefs/error.c @@ -1,6 +1,7 @@ // SPDX-License-Identifier: GPL-2.0 #include "bcachefs.h" #include "error.h" +#include "recovery.h" #include "super.h" #include "thread_with_file.h" @@ -25,11 +26,16 @@ bool bch2_inconsistent_error(struct bch_fs *c) } } -void bch2_topology_error(struct bch_fs *c) +int bch2_topology_error(struct bch_fs *c) { set_bit(BCH_FS_topology_error, &c->flags); - if (!test_bit(BCH_FS_fsck_running, &c->flags)) + if (!test_bit(BCH_FS_fsck_running, &c->flags)) { bch2_inconsistent_error(c); + return -BCH_ERR_btree_need_topology_repair; + } else { + return bch2_run_explicit_recovery_pass(c, BCH_RECOVERY_PASS_check_topology) ?: + -BCH_ERR_btree_node_read_validate_error; + } } void bch2_fatal_error(struct bch_fs *c) diff --git a/fs/bcachefs/error.h b/fs/bcachefs/error.h index fec17d1353d1..ae1d6674c512 100644 --- a/fs/bcachefs/error.h +++ b/fs/bcachefs/error.h @@ -30,7 +30,7 @@ struct work_struct; bool bch2_inconsistent_error(struct bch_fs *); -void bch2_topology_error(struct bch_fs *); +int bch2_topology_error(struct bch_fs *); #define bch2_fs_inconsistent(c, ...) \ ({ \ @@ -191,9 +191,9 @@ do { \ void bch2_fatal_error(struct bch_fs *); -#define bch2_fs_fatal_error(c, ...) \ +#define bch2_fs_fatal_error(c, _msg, ...) \ do { \ - bch_err(c, __VA_ARGS__); \ + bch_err(c, "%s(): fatal error " _msg, __func__, ##__VA_ARGS__); \ bch2_fatal_error(c); \ } while (0) diff --git a/fs/bcachefs/extents.h b/fs/bcachefs/extents.h index 6bf839d69e84..fd2669cdd76f 100644 --- a/fs/bcachefs/extents.h +++ b/fs/bcachefs/extents.h @@ -43,6 +43,11 @@ enum bkey_invalid_flags; #define extent_entry_next(_entry) \ ((typeof(_entry)) ((void *) (_entry) + extent_entry_bytes(_entry))) +#define extent_entry_next_safe(_entry, _end) \ + (likely(__extent_entry_type(_entry) < BCH_EXTENT_ENTRY_MAX) \ + ? extent_entry_next(_entry) \ + : _end) + static inline unsigned __extent_entry_type(const union bch_extent_entry *e) { @@ -103,17 +108,17 @@ static inline void extent_entry_drop(struct bkey_s k, union bch_extent_entry *en static inline bool extent_entry_is_ptr(const union bch_extent_entry *e) { - return extent_entry_type(e) == BCH_EXTENT_ENTRY_ptr; + return __extent_entry_type(e) == BCH_EXTENT_ENTRY_ptr; } static inline bool extent_entry_is_stripe_ptr(const union bch_extent_entry *e) { - return extent_entry_type(e) == BCH_EXTENT_ENTRY_stripe_ptr; + return __extent_entry_type(e) == BCH_EXTENT_ENTRY_stripe_ptr; } static inline bool extent_entry_is_crc(const union bch_extent_entry *e) { - switch (extent_entry_type(e)) { + switch (__extent_entry_type(e)) { case BCH_EXTENT_ENTRY_crc32: case BCH_EXTENT_ENTRY_crc64: case BCH_EXTENT_ENTRY_crc128: @@ -280,7 +285,7 @@ static inline struct bkey_ptrs bch2_bkey_ptrs(struct bkey_s k) #define __bkey_extent_entry_for_each_from(_start, _end, _entry) \ for ((_entry) = (_start); \ (_entry) < (_end); \ - (_entry) = extent_entry_next(_entry)) + (_entry) = extent_entry_next_safe(_entry, _end)) #define __bkey_ptr_next(_ptr, _end) \ ({ \ @@ -318,7 +323,7 @@ static inline struct bkey_ptrs bch2_bkey_ptrs(struct bkey_s k) (_ptr).has_ec = false; \ \ __bkey_extent_entry_for_each_from(_entry, _end, _entry) \ - switch (extent_entry_type(_entry)) { \ + switch (__extent_entry_type(_entry)) { \ case BCH_EXTENT_ENTRY_ptr: \ (_ptr).ptr = _entry->ptr; \ goto out; \ @@ -344,7 +349,7 @@ out: \ for ((_ptr).crc = bch2_extent_crc_unpack(_k, NULL), \ (_entry) = _start; \ __bkey_ptr_next_decode(_k, _end, _ptr, _entry); \ - (_entry) = extent_entry_next(_entry)) + (_entry) = extent_entry_next_safe(_entry, _end)) #define bkey_for_each_ptr_decode(_k, _p, _ptr, _entry) \ __bkey_for_each_ptr_decode(_k, (_p).start, (_p).end, \ diff --git a/fs/bcachefs/fifo.h b/fs/bcachefs/fifo.h index 66b945be10c2..d8153fe27037 100644 --- a/fs/bcachefs/fifo.h +++ b/fs/bcachefs/fifo.h @@ -24,12 +24,12 @@ struct { \ (fifo)->mask = (fifo)->size \ ? roundup_pow_of_two((fifo)->size) - 1 \ : 0; \ - (fifo)->data = kvpmalloc(fifo_buf_size(fifo), (_gfp)); \ + (fifo)->data = kvmalloc(fifo_buf_size(fifo), (_gfp)); \ }) #define free_fifo(fifo) \ do { \ - kvpfree((fifo)->data, fifo_buf_size(fifo)); \ + kvfree((fifo)->data); \ (fifo)->data = NULL; \ } while (0) diff --git a/fs/bcachefs/fs-common.c b/fs/bcachefs/fs-common.c index 1c1ea0f0c692..624e6f963240 100644 --- a/fs/bcachefs/fs-common.c +++ b/fs/bcachefs/fs-common.c @@ -107,6 +107,7 @@ int bch2_create_trans(struct btree_trans *trans, u32 new_subvol, dir_snapshot; ret = bch2_subvolume_create(trans, new_inode->bi_inum, + dir.subvol, snapshot_src.subvol, &new_subvol, &snapshot, (flags & BCH_CREATE_SNAPSHOT_RO) != 0); @@ -242,7 +243,7 @@ int bch2_unlink_trans(struct btree_trans *trans, struct bch_inode_unpacked *dir_u, struct bch_inode_unpacked *inode_u, const struct qstr *name, - bool deleting_snapshot) + bool deleting_subvol) { struct bch_fs *c = trans->c; struct btree_iter dir_iter = { NULL }; @@ -260,8 +261,8 @@ int bch2_unlink_trans(struct btree_trans *trans, dir_hash = bch2_hash_info_init(c, dir_u); - ret = __bch2_dirent_lookup_trans(trans, &dirent_iter, dir, &dir_hash, - name, &inum, BTREE_ITER_INTENT); + ret = bch2_dirent_lookup_trans(trans, &dirent_iter, dir, &dir_hash, + name, &inum, BTREE_ITER_INTENT); if (ret) goto err; @@ -270,18 +271,25 @@ int bch2_unlink_trans(struct btree_trans *trans, if (ret) goto err; - if (!deleting_snapshot && S_ISDIR(inode_u->bi_mode)) { + if (!deleting_subvol && S_ISDIR(inode_u->bi_mode)) { ret = bch2_empty_dir_trans(trans, inum); if (ret) goto err; } - if (deleting_snapshot && !inode_u->bi_subvol) { + if (deleting_subvol && !inode_u->bi_subvol) { ret = -BCH_ERR_ENOENT_not_subvol; goto err; } - if (deleting_snapshot || inode_u->bi_subvol) { + if (inode_u->bi_subvol) { + /* Recursive subvolume destroy not allowed (yet?) */ + ret = bch2_subvol_has_children(trans, inode_u->bi_subvol); + if (ret) + goto err; + } + + if (deleting_subvol || inode_u->bi_subvol) { ret = bch2_subvolume_unlink(trans, inode_u->bi_subvol); if (ret) goto err; @@ -349,6 +357,22 @@ bool bch2_reinherit_attrs(struct bch_inode_unpacked *dst_u, return ret; } +static int subvol_update_parent(struct btree_trans *trans, u32 subvol, u32 new_parent) +{ + struct btree_iter iter; + struct bkey_i_subvolume *s = + bch2_bkey_get_mut_typed(trans, &iter, + BTREE_ID_subvolumes, POS(0, subvol), + BTREE_ITER_CACHED, subvolume); + int ret = PTR_ERR_OR_ZERO(s); + if (ret) + return ret; + + s->v.fs_path_parent = cpu_to_le32(new_parent); + bch2_trans_iter_exit(trans, &iter); + return 0; +} + int bch2_rename_trans(struct btree_trans *trans, subvol_inum src_dir, struct bch_inode_unpacked *src_dir_u, subvol_inum dst_dir, struct bch_inode_unpacked *dst_dir_u, @@ -410,6 +434,36 @@ int bch2_rename_trans(struct btree_trans *trans, goto err; } + if (src_inode_u->bi_subvol && + dst_dir.subvol != src_inode_u->bi_parent_subvol) { + ret = subvol_update_parent(trans, src_inode_u->bi_subvol, dst_dir.subvol); + if (ret) + goto err; + } + + if (mode == BCH_RENAME_EXCHANGE && + dst_inode_u->bi_subvol && + src_dir.subvol != dst_inode_u->bi_parent_subvol) { + ret = subvol_update_parent(trans, dst_inode_u->bi_subvol, src_dir.subvol); + if (ret) + goto err; + } + + /* Can't move across subvolumes, unless it's a subvolume root: */ + if (src_dir.subvol != dst_dir.subvol && + (!src_inode_u->bi_subvol || + (dst_inum.inum && !dst_inode_u->bi_subvol))) { + ret = -EXDEV; + goto err; + } + + if (src_inode_u->bi_parent_subvol) + src_inode_u->bi_parent_subvol = dst_dir.subvol; + + if ((mode == BCH_RENAME_EXCHANGE) && + dst_inode_u->bi_parent_subvol) + dst_inode_u->bi_parent_subvol = src_dir.subvol; + src_inode_u->bi_dir = dst_dir_u->bi_inum; src_inode_u->bi_dir_offset = dst_offset; @@ -432,10 +486,10 @@ int bch2_rename_trans(struct btree_trans *trans, goto err; } - if (S_ISDIR(dst_inode_u->bi_mode) && - bch2_empty_dir_trans(trans, dst_inum)) { - ret = -ENOTEMPTY; - goto err; + if (S_ISDIR(dst_inode_u->bi_mode)) { + ret = bch2_empty_dir_trans(trans, dst_inum); + if (ret) + goto err; } } diff --git a/fs/bcachefs/fs-io-buffered.c b/fs/bcachefs/fs-io-buffered.c index 27710cdd5710..39292e7ef342 100644 --- a/fs/bcachefs/fs-io-buffered.c +++ b/fs/bcachefs/fs-io-buffered.c @@ -810,7 +810,8 @@ static noinline void folios_trunc(folios *fs, struct folio **fi) static int __bch2_buffered_write(struct bch_inode_info *inode, struct address_space *mapping, struct iov_iter *iter, - loff_t pos, unsigned len) + loff_t pos, unsigned len, + bool inode_locked) { struct bch_fs *c = inode->v.i_sb->s_fs_info; struct bch2_folio_reservation res; @@ -835,6 +836,15 @@ static int __bch2_buffered_write(struct bch_inode_info *inode, BUG_ON(!fs.nr); + /* + * If we're not using the inode lock, we need to lock all the folios for + * atomiticity of writes vs. other writes: + */ + if (!inode_locked && folio_end_pos(darray_last(fs)) < end) { + ret = -BCH_ERR_need_inode_lock; + goto out; + } + f = darray_first(fs); if (pos != folio_pos(f) && !folio_test_uptodate(f)) { ret = bch2_read_single_folio(f, mapping); @@ -929,8 +939,10 @@ static int __bch2_buffered_write(struct bch_inode_info *inode, end = pos + copied; spin_lock(&inode->v.i_lock); - if (end > inode->v.i_size) + if (end > inode->v.i_size) { + BUG_ON(!inode_locked); i_size_write(&inode->v, end); + } spin_unlock(&inode->v.i_lock); f_pos = pos; @@ -974,12 +986,68 @@ static ssize_t bch2_buffered_write(struct kiocb *iocb, struct iov_iter *iter) struct file *file = iocb->ki_filp; struct address_space *mapping = file->f_mapping; struct bch_inode_info *inode = file_bch_inode(file); - loff_t pos = iocb->ki_pos; - ssize_t written = 0; - int ret = 0; + loff_t pos; + bool inode_locked = false; + ssize_t written = 0, written2 = 0, ret = 0; + + /* + * We don't take the inode lock unless i_size will be changing. Folio + * locks provide exclusion with other writes, and the pagecache add lock + * provides exclusion with truncate and hole punching. + * + * There is one nasty corner case where atomicity would be broken + * without great care: when copying data from userspace to the page + * cache, we do that with faults disable - a page fault would recurse + * back into the filesystem, taking filesystem locks again, and + * deadlock; so it's done with faults disabled, and we fault in the user + * buffer when we aren't holding locks. + * + * If we do part of the write, but we then race and in the userspace + * buffer have been evicted and are no longer resident, then we have to + * drop our folio locks to re-fault them in, breaking write atomicity. + * + * To fix this, we restart the write from the start, if we weren't + * holding the inode lock. + * + * There is another wrinkle after that; if we restart the write from the + * start, and then get an unrecoverable error, we _cannot_ claim to + * userspace that we did not write data we actually did - so we must + * track (written2) the most we ever wrote. + */ + + if ((iocb->ki_flags & IOCB_APPEND) || + (iocb->ki_pos + iov_iter_count(iter) > i_size_read(&inode->v))) { + inode_lock(&inode->v); + inode_locked = true; + } + + ret = generic_write_checks(iocb, iter); + if (ret <= 0) + goto unlock; + + ret = file_remove_privs_flags(file, !inode_locked ? IOCB_NOWAIT : 0); + if (ret) { + if (!inode_locked) { + inode_lock(&inode->v); + inode_locked = true; + ret = file_remove_privs_flags(file, 0); + } + if (ret) + goto unlock; + } + + ret = file_update_time(file); + if (ret) + goto unlock; + + pos = iocb->ki_pos; bch2_pagecache_add_get(inode); + if (!inode_locked && + (iocb->ki_pos + iov_iter_count(iter) > i_size_read(&inode->v))) + goto get_inode_lock; + do { unsigned offset = pos & (PAGE_SIZE - 1); unsigned bytes = iov_iter_count(iter); @@ -1004,12 +1072,17 @@ again: } } + if (unlikely(bytes != iov_iter_count(iter) && !inode_locked)) + goto get_inode_lock; + if (unlikely(fatal_signal_pending(current))) { ret = -EINTR; break; } - ret = __bch2_buffered_write(inode, mapping, iter, pos, bytes); + ret = __bch2_buffered_write(inode, mapping, iter, pos, bytes, inode_locked); + if (ret == -BCH_ERR_need_inode_lock) + goto get_inode_lock; if (unlikely(ret < 0)) break; @@ -1030,50 +1103,46 @@ again: } pos += ret; written += ret; + written2 = max(written, written2); + + if (ret != bytes && !inode_locked) + goto get_inode_lock; ret = 0; balance_dirty_pages_ratelimited(mapping); - } while (iov_iter_count(iter)); + if (0) { +get_inode_lock: + bch2_pagecache_add_put(inode); + inode_lock(&inode->v); + inode_locked = true; + bch2_pagecache_add_get(inode); + + iov_iter_revert(iter, written); + pos -= written; + written = 0; + ret = 0; + } + } while (iov_iter_count(iter)); bch2_pagecache_add_put(inode); +unlock: + if (inode_locked) + inode_unlock(&inode->v); + + iocb->ki_pos += written; - return written ? written : ret; + ret = max(written, written2) ?: ret; + if (ret > 0) + ret = generic_write_sync(iocb, ret); + return ret; } -ssize_t bch2_write_iter(struct kiocb *iocb, struct iov_iter *from) +ssize_t bch2_write_iter(struct kiocb *iocb, struct iov_iter *iter) { - struct file *file = iocb->ki_filp; - struct bch_inode_info *inode = file_bch_inode(file); - ssize_t ret; - - if (iocb->ki_flags & IOCB_DIRECT) { - ret = bch2_direct_write(iocb, from); - goto out; - } - - inode_lock(&inode->v); - - ret = generic_write_checks(iocb, from); - if (ret <= 0) - goto unlock; - - ret = file_remove_privs(file); - if (ret) - goto unlock; - - ret = file_update_time(file); - if (ret) - goto unlock; - - ret = bch2_buffered_write(iocb, from); - if (likely(ret > 0)) - iocb->ki_pos += ret; -unlock: - inode_unlock(&inode->v); + ssize_t ret = iocb->ki_flags & IOCB_DIRECT + ? bch2_direct_write(iocb, iter) + : bch2_buffered_write(iocb, iter); - if (ret > 0) - ret = generic_write_sync(iocb, ret); -out: return bch2_err_class(ret); } diff --git a/fs/bcachefs/fs-io-pagecache.h b/fs/bcachefs/fs-io-pagecache.h index 8cbaba6565b4..828c3d7c8f19 100644 --- a/fs/bcachefs/fs-io-pagecache.h +++ b/fs/bcachefs/fs-io-pagecache.h @@ -51,13 +51,10 @@ enum bch_folio_sector_state { struct bch_folio_sector { /* Uncompressed, fully allocated replicas (or on disk reservation): */ - unsigned nr_replicas:4; - + u8 nr_replicas:4, /* Owns PAGE_SECTORS * replicas_reserved sized in memory reservation: */ - unsigned replicas_reserved:4; - - /* i_sectors: */ - enum bch_folio_sector_state state:8; + replicas_reserved:4; + u8 state; }; struct bch_folio { diff --git a/fs/bcachefs/fs.c b/fs/bcachefs/fs.c index 77ae65542db9..0ccee05f6887 100644 --- a/fs/bcachefs/fs.c +++ b/fs/bcachefs/fs.c @@ -108,7 +108,8 @@ retry: goto retry; bch2_fs_fatal_err_on(bch2_err_matches(ret, ENOENT), c, - "inode %u:%llu not found when updating", + "%s: inode %u:%llu not found when updating", + bch2_err_str(ret), inode_inum(inode).subvol, inode_inum(inode).inum); @@ -176,45 +177,88 @@ static unsigned bch2_inode_hash(subvol_inum inum) return jhash_3words(inum.subvol, inum.inum >> 32, inum.inum, JHASH_INITVAL); } -struct inode *bch2_vfs_inode_get(struct bch_fs *c, subvol_inum inum) +static struct bch_inode_info *bch2_inode_insert(struct bch_fs *c, struct bch_inode_info *inode) { - struct bch_inode_unpacked inode_u; - struct bch_inode_info *inode; - struct btree_trans *trans; - struct bch_subvolume subvol; - int ret; + subvol_inum inum = inode_inum(inode); + struct bch_inode_info *old = to_bch_ei(inode_insert5(&inode->v, + bch2_inode_hash(inum), + bch2_iget5_test, + bch2_iget5_set, + &inum)); + BUG_ON(!old); - inode = to_bch_ei(iget5_locked(c->vfs_sb, - bch2_inode_hash(inum), - bch2_iget5_test, - bch2_iget5_set, - &inum)); - if (unlikely(!inode)) - return ERR_PTR(-ENOMEM); - if (!(inode->v.i_state & I_NEW)) - return &inode->v; + if (unlikely(old != inode)) { + discard_new_inode(&inode->v); + inode = old; + } else { + mutex_lock(&c->vfs_inodes_lock); + list_add(&inode->ei_vfs_inode_list, &c->vfs_inodes_list); + mutex_unlock(&c->vfs_inodes_lock); + /* + * we really don't want insert_inode_locked2() to be setting + * I_NEW... + */ + unlock_new_inode(&inode->v); + } - trans = bch2_trans_get(c); - ret = lockrestart_do(trans, - bch2_subvolume_get(trans, inum.subvol, true, 0, &subvol) ?: - bch2_inode_find_by_inum_trans(trans, inum, &inode_u)); + return inode; +} - if (!ret) - bch2_vfs_inode_init(trans, inum, inode, &inode_u, &subvol); - bch2_trans_put(trans); +#define memalloc_flags_do(_flags, _do) \ +({ \ + unsigned _saved_flags = memalloc_flags_save(_flags); \ + typeof(_do) _ret = _do; \ + memalloc_noreclaim_restore(_saved_flags); \ + _ret; \ +}) - if (ret) { - iget_failed(&inode->v); - return ERR_PTR(bch2_err_class(ret)); +/* + * Allocate a new inode, dropping/retaking btree locks if necessary: + */ +static struct bch_inode_info *bch2_new_inode(struct btree_trans *trans) +{ + struct bch_fs *c = trans->c; + + struct bch_inode_info *inode = + memalloc_flags_do(PF_MEMALLOC_NORECLAIM|PF_MEMALLOC_NOWARN, + to_bch_ei(new_inode(c->vfs_sb))); + + if (unlikely(!inode)) { + int ret = drop_locks_do(trans, (inode = to_bch_ei(new_inode(c->vfs_sb))) ? 0 : -ENOMEM); + if (ret && inode) + discard_new_inode(&inode->v); + if (ret) + return ERR_PTR(ret); } - mutex_lock(&c->vfs_inodes_lock); - list_add(&inode->ei_vfs_inode_list, &c->vfs_inodes_list); - mutex_unlock(&c->vfs_inodes_lock); + return inode; +} - unlock_new_inode(&inode->v); +struct inode *bch2_vfs_inode_get(struct bch_fs *c, subvol_inum inum) +{ + struct bch_inode_info *inode = + to_bch_ei(ilookup5_nowait(c->vfs_sb, + bch2_inode_hash(inum), + bch2_iget5_test, + &inum)); + if (inode) + return &inode->v; - return &inode->v; + struct btree_trans *trans = bch2_trans_get(c); + + struct bch_inode_unpacked inode_u; + struct bch_subvolume subvol; + int ret = lockrestart_do(trans, + bch2_subvolume_get(trans, inum.subvol, true, 0, &subvol) ?: + bch2_inode_find_by_inum_trans(trans, inum, &inode_u)) ?: + PTR_ERR_OR_ZERO(inode = bch2_new_inode(trans)); + if (!ret) { + bch2_vfs_inode_init(trans, inum, inode, &inode_u, &subvol); + inode = bch2_inode_insert(c, inode); + } + bch2_trans_put(trans); + + return ret ? ERR_PTR(ret) : &inode->v; } struct bch_inode_info * @@ -226,7 +270,7 @@ __bch2_create(struct mnt_idmap *idmap, struct bch_fs *c = dir->v.i_sb->s_fs_info; struct btree_trans *trans; struct bch_inode_unpacked dir_u; - struct bch_inode_info *inode, *old; + struct bch_inode_info *inode; struct bch_inode_unpacked inode_u; struct posix_acl *default_acl = NULL, *acl = NULL; subvol_inum inum; @@ -293,7 +337,6 @@ err_before_quota: mutex_unlock(&dir->ei_update_lock); } - bch2_iget5_set(&inode->v, &inum); bch2_vfs_inode_init(trans, inum, inode, &inode_u, &subvol); set_cached_acl(&inode->v, ACL_TYPE_ACCESS, acl); @@ -304,36 +347,7 @@ err_before_quota: * bch2_trans_exit() and dropping locks, else we could race with another * thread pulling the inode in and modifying it: */ - - inode->v.i_state |= I_CREATING; - - old = to_bch_ei(inode_insert5(&inode->v, - bch2_inode_hash(inum), - bch2_iget5_test, - bch2_iget5_set, - &inum)); - BUG_ON(!old); - - if (unlikely(old != inode)) { - /* - * We raced, another process pulled the new inode into cache - * before us: - */ - make_bad_inode(&inode->v); - iput(&inode->v); - - inode = old; - } else { - mutex_lock(&c->vfs_inodes_lock); - list_add(&inode->ei_vfs_inode_list, &c->vfs_inodes_list); - mutex_unlock(&c->vfs_inodes_lock); - /* - * we really don't want insert_inode_locked2() to be setting - * I_NEW... - */ - unlock_new_inode(&inode->v); - } - + inode = bch2_inode_insert(c, inode); bch2_trans_put(trans); err: posix_acl_release(default_acl); @@ -352,23 +366,78 @@ err_trans: /* methods */ +static struct bch_inode_info *bch2_lookup_trans(struct btree_trans *trans, + subvol_inum dir, struct bch_hash_info *dir_hash_info, + const struct qstr *name) +{ + struct bch_fs *c = trans->c; + struct btree_iter dirent_iter = {}; + subvol_inum inum = {}; + + int ret = bch2_hash_lookup(trans, &dirent_iter, bch2_dirent_hash_desc, + dir_hash_info, dir, name, 0); + if (ret) + return ERR_PTR(ret); + + struct bkey_s_c k = bch2_btree_iter_peek_slot(&dirent_iter); + ret = bkey_err(k); + if (ret) + goto err; + + ret = bch2_dirent_read_target(trans, dir, bkey_s_c_to_dirent(k), &inum); + if (ret > 0) + ret = -ENOENT; + if (ret) + goto err; + + struct bch_inode_info *inode = + to_bch_ei(ilookup5_nowait(c->vfs_sb, + bch2_inode_hash(inum), + bch2_iget5_test, + &inum)); + if (inode) + goto out; + + struct bch_subvolume subvol; + struct bch_inode_unpacked inode_u; + ret = bch2_subvolume_get(trans, inum.subvol, true, 0, &subvol) ?: + bch2_inode_find_by_inum_nowarn_trans(trans, inum, &inode_u) ?: + PTR_ERR_OR_ZERO(inode = bch2_new_inode(trans)); + if (bch2_err_matches(ret, ENOENT)) { + struct printbuf buf = PRINTBUF; + + bch2_bkey_val_to_text(&buf, c, k); + bch_err(c, "%s points to missing inode", buf.buf); + printbuf_exit(&buf); + } + if (ret) + goto err; + + bch2_vfs_inode_init(trans, inum, inode, &inode_u, &subvol); + inode = bch2_inode_insert(c, inode); +out: + bch2_trans_iter_exit(trans, &dirent_iter); + return inode; +err: + inode = ERR_PTR(ret); + goto out; +} + static struct dentry *bch2_lookup(struct inode *vdir, struct dentry *dentry, unsigned int flags) { struct bch_fs *c = vdir->i_sb->s_fs_info; struct bch_inode_info *dir = to_bch_ei(vdir); struct bch_hash_info hash = bch2_hash_info_init(c, &dir->ei_inode); - struct inode *vinode = NULL; - subvol_inum inum = { .subvol = 1 }; - int ret; - ret = bch2_dirent_lookup(c, inode_inum(dir), &hash, - &dentry->d_name, &inum); - - if (!ret) - vinode = bch2_vfs_inode_get(c, inum); + struct bch_inode_info *inode; + bch2_trans_do(c, NULL, NULL, 0, + PTR_ERR_OR_ZERO(inode = bch2_lookup_trans(trans, inode_inum(dir), + &hash, &dentry->d_name))); + if (IS_ERR(inode)) + inode = NULL; - return d_splice_alias(vinode, dentry); + return d_splice_alias(&inode->v, dentry); } static int bch2_mknod(struct mnt_idmap *idmap, @@ -1372,6 +1441,7 @@ static void bch2_vfs_inode_init(struct btree_trans *trans, subvol_inum inum, struct bch_inode_unpacked *bi, struct bch_subvolume *subvol) { + bch2_iget5_set(&inode->v, &inum); bch2_inode_update_after_write(trans, inode, bi, ~0); if (BCH_SUBVOLUME_SNAP(subvol)) @@ -1572,7 +1642,6 @@ static int bch2_statfs(struct dentry *dentry, struct kstatfs *buf) * number: */ u64 avail_inodes = ((usage.capacity - usage.used) << 3); - u64 fsid; buf->f_type = BCACHEFS_STATFS_MAGIC; buf->f_bsize = sb->s_blocksize; @@ -1583,10 +1652,7 @@ static int bch2_statfs(struct dentry *dentry, struct kstatfs *buf) buf->f_files = usage.nr_inodes + avail_inodes; buf->f_ffree = avail_inodes; - fsid = le64_to_cpup((void *) c->sb.user_uuid.b) ^ - le64_to_cpup((void *) c->sb.user_uuid.b + sizeof(u64)); - buf->f_fsid.val[0] = fsid & 0xFFFFFFFFUL; - buf->f_fsid.val[1] = (fsid >> 32) & 0xFFFFFFFFUL; + buf->f_fsid = uuid_to_fsid(c->sb.user_uuid.b); buf->f_namelen = BCH_NAME_MAX; return 0; @@ -1805,8 +1871,10 @@ static struct dentry *bch2_mount(struct file_system_type *fs_type, opt_set(opts, read_only, (flags & SB_RDONLY) != 0); ret = bch2_parse_mount_opts(NULL, &opts, data); - if (ret) + if (ret) { + ret = bch2_err_class(ret); return ERR_PTR(ret); + } if (!dev_name || strlen(dev_name) == 0) return ERR_PTR(-EINVAL); @@ -1882,6 +1950,7 @@ got_sb: sb->s_time_gran = c->sb.nsec_per_time_unit; sb->s_time_min = div_s64(S64_MIN, c->sb.time_units_per_sec) + 1; sb->s_time_max = div_s64(S64_MAX, c->sb.time_units_per_sec); + sb->s_uuid = c->sb.user_uuid; c->vfs_sb = sb; strscpy(sb->s_id, c->name, sizeof(sb->s_id)); diff --git a/fs/bcachefs/fsck.c b/fs/bcachefs/fsck.c index 6a760777bafb..47d4eefaba7b 100644 --- a/fs/bcachefs/fsck.c +++ b/fs/bcachefs/fsck.c @@ -100,8 +100,8 @@ err: } static int lookup_inode(struct btree_trans *trans, u64 inode_nr, - struct bch_inode_unpacked *inode, - u32 *snapshot) + struct bch_inode_unpacked *inode, + u32 *snapshot) { struct btree_iter iter; struct bkey_s_c k; @@ -142,34 +142,6 @@ static int lookup_dirent_in_snapshot(struct btree_trans *trans, return 0; } -static int __write_inode(struct btree_trans *trans, - struct bch_inode_unpacked *inode, - u32 snapshot) -{ - struct bkey_inode_buf *inode_p = - bch2_trans_kmalloc(trans, sizeof(*inode_p)); - - if (IS_ERR(inode_p)) - return PTR_ERR(inode_p); - - bch2_inode_pack(inode_p, inode); - inode_p->inode.k.p.snapshot = snapshot; - - return bch2_btree_insert_nonextent(trans, BTREE_ID_inodes, - &inode_p->inode.k_i, - BTREE_UPDATE_INTERNAL_SNAPSHOT_NODE); -} - -static int fsck_write_inode(struct btree_trans *trans, - struct bch_inode_unpacked *inode, - u32 snapshot) -{ - int ret = commit_do(trans, NULL, NULL, BCH_TRANS_COMMIT_no_enospc, - __write_inode(trans, inode, snapshot)); - bch_err_fn(trans->c, ret); - return ret; -} - static int __remove_dirent(struct btree_trans *trans, struct bpos pos) { struct bch_fs *c = trans->c; @@ -280,7 +252,7 @@ create_lostfound: goto err; ret = bch2_dirent_create_snapshot(trans, - root_inode.bi_inum, snapshot, &root_hash_info, + 0, root_inode.bi_inum, snapshot, &root_hash_info, mode_to_type(lostfound->bi_mode), &lostfound_str, lostfound->bi_inum, @@ -303,30 +275,47 @@ static int reattach_inode(struct btree_trans *trans, char name_buf[20]; struct qstr name; u64 dir_offset = 0; + u32 dirent_snapshot = inode_snapshot; int ret; - ret = lookup_lostfound(trans, inode_snapshot, &lostfound); + if (inode->bi_subvol) { + inode->bi_parent_subvol = BCACHEFS_ROOT_SUBVOL; + + u64 root_inum; + ret = subvol_lookup(trans, inode->bi_parent_subvol, + &dirent_snapshot, &root_inum); + if (ret) + return ret; + + snprintf(name_buf, sizeof(name_buf), "subvol-%u", inode->bi_subvol); + } else { + snprintf(name_buf, sizeof(name_buf), "%llu", inode->bi_inum); + } + + ret = lookup_lostfound(trans, dirent_snapshot, &lostfound); if (ret) return ret; if (S_ISDIR(inode->bi_mode)) { lostfound.bi_nlink++; - ret = __write_inode(trans, &lostfound, U32_MAX); + ret = __bch2_fsck_write_inode(trans, &lostfound, U32_MAX); if (ret) return ret; } dir_hash = bch2_hash_info_init(trans->c, &lostfound); - snprintf(name_buf, sizeof(name_buf), "%llu", inode->bi_inum); name = (struct qstr) QSTR(name_buf); ret = bch2_dirent_create_snapshot(trans, - lostfound.bi_inum, inode_snapshot, + inode->bi_parent_subvol, lostfound.bi_inum, + dirent_snapshot, &dir_hash, inode_d_type(inode), - &name, inode->bi_inum, &dir_offset, + &name, + inode->bi_subvol ?: inode->bi_inum, + &dir_offset, BCH_HASH_SET_MUST_CREATE); if (ret) return ret; @@ -334,7 +323,7 @@ static int reattach_inode(struct btree_trans *trans, inode->bi_dir = lostfound.bi_inum; inode->bi_dir_offset = dir_offset; - return __write_inode(trans, inode, inode_snapshot); + return __bch2_fsck_write_inode(trans, inode, inode_snapshot); } static int remove_backpointer(struct btree_trans *trans, @@ -353,6 +342,27 @@ static int remove_backpointer(struct btree_trans *trans, return ret; } +static int reattach_subvol(struct btree_trans *trans, struct bkey_s_c_subvolume s) +{ + struct bch_fs *c = trans->c; + + struct bch_inode_unpacked inode; + int ret = bch2_inode_find_by_inum_trans(trans, + (subvol_inum) { s.k->p.offset, le64_to_cpu(s.v->inode) }, + &inode); + if (ret) + return ret; + + ret = remove_backpointer(trans, &inode); + bch_err_msg(c, ret, "removing dirent"); + if (ret) + return ret; + + ret = reattach_inode(trans, &inode, le32_to_cpu(s.v->snapshot)); + bch_err_msg(c, ret, "reattaching inode %llu", inode.bi_inum); + return ret; +} + struct snapshots_seen_entry { u32 id; u32 equiv; @@ -592,13 +602,12 @@ static int get_inodes_all_snapshots(struct btree_trans *trans, } static struct inode_walker_entry * -lookup_inode_for_snapshot(struct bch_fs *c, struct inode_walker *w, - u32 snapshot, bool is_whiteout) +lookup_inode_for_snapshot(struct bch_fs *c, struct inode_walker *w, struct bkey_s_c k) { - struct inode_walker_entry *i; - - snapshot = bch2_snapshot_equiv(c, snapshot); + bool is_whiteout = k.k->type == KEY_TYPE_whiteout; + u32 snapshot = bch2_snapshot_equiv(c, k.k->p.snapshot); + struct inode_walker_entry *i; __darray_for_each(w->inodes, i) if (bch2_snapshot_is_ancestor(c, snapshot, i->snapshot)) goto found; @@ -609,20 +618,24 @@ found: if (snapshot != i->snapshot && !is_whiteout) { struct inode_walker_entry new = *i; - size_t pos; - int ret; new.snapshot = snapshot; new.count = 0; - bch_info(c, "have key for inode %llu:%u but have inode in ancestor snapshot %u", - w->last_pos.inode, snapshot, i->snapshot); + struct printbuf buf = PRINTBUF; + bch2_bkey_val_to_text(&buf, c, k); + + bch_info(c, "have key for inode %llu:%u but have inode in ancestor snapshot %u\n" + "unexpected because we should always update the inode when we update a key in that inode\n" + "%s", + w->last_pos.inode, snapshot, i->snapshot, buf.buf); + printbuf_exit(&buf); while (i > w->inodes.data && i[-1].snapshot > snapshot) --i; - pos = i - w->inodes.data; - ret = darray_insert_item(&w->inodes, pos, new); + size_t pos = i - w->inodes.data; + int ret = darray_insert_item(&w->inodes, pos, new); if (ret) return ERR_PTR(ret); @@ -633,21 +646,21 @@ found: } static struct inode_walker_entry *walk_inode(struct btree_trans *trans, - struct inode_walker *w, struct bpos pos, - bool is_whiteout) + struct inode_walker *w, + struct bkey_s_c k) { - if (w->last_pos.inode != pos.inode) { - int ret = get_inodes_all_snapshots(trans, w, pos.inode); + if (w->last_pos.inode != k.k->p.inode) { + int ret = get_inodes_all_snapshots(trans, w, k.k->p.inode); if (ret) return ERR_PTR(ret); - } else if (bkey_cmp(w->last_pos, pos)) { + } else if (bkey_cmp(w->last_pos, k.k->p)) { darray_for_each(w->inodes, i) i->seen_this_pos = false; } - w->last_pos = pos; + w->last_pos = k.k->p; - return lookup_inode_for_snapshot(trans->c, w, pos.snapshot, is_whiteout); + return lookup_inode_for_snapshot(trans->c, w, k); } static int __get_visible_inodes(struct btree_trans *trans, @@ -722,7 +735,7 @@ static int hash_redo_key(struct btree_trans *trans, delete->k.p = k_iter->pos; return bch2_btree_iter_traverse(k_iter) ?: bch2_trans_update(trans, k_iter, delete, 0) ?: - bch2_hash_set_snapshot(trans, desc, hash_info, + bch2_hash_set_in_snapshot(trans, desc, hash_info, (subvol_inum) { 0, k.k->p.inode }, k.k->p.snapshot, tmp, BCH_HASH_SET_MUST_CREATE, @@ -795,16 +808,93 @@ fsck_err: goto out; } +static struct bkey_s_c_dirent dirent_get_by_pos(struct btree_trans *trans, + struct btree_iter *iter, + struct bpos pos) +{ + return bch2_bkey_get_iter_typed(trans, iter, BTREE_ID_dirents, pos, 0, dirent); +} + +static struct bkey_s_c_dirent inode_get_dirent(struct btree_trans *trans, + struct btree_iter *iter, + struct bch_inode_unpacked *inode, + u32 *snapshot) +{ + if (inode->bi_subvol) { + u64 inum; + int ret = subvol_lookup(trans, inode->bi_parent_subvol, snapshot, &inum); + if (ret) + return ((struct bkey_s_c_dirent) { .k = ERR_PTR(ret) }); + } + + return dirent_get_by_pos(trans, iter, SPOS(inode->bi_dir, inode->bi_dir_offset, *snapshot)); +} + +static bool inode_points_to_dirent(struct bch_inode_unpacked *inode, + struct bkey_s_c_dirent d) +{ + return inode->bi_dir == d.k->p.inode && + inode->bi_dir_offset == d.k->p.offset; +} + +static bool dirent_points_to_inode(struct bkey_s_c_dirent d, + struct bch_inode_unpacked *inode) +{ + return d.v->d_type == DT_SUBVOL + ? le32_to_cpu(d.v->d_child_subvol) == inode->bi_subvol + : le64_to_cpu(d.v->d_inum) == inode->bi_inum; +} + static int check_inode_deleted_list(struct btree_trans *trans, struct bpos p) { struct btree_iter iter; struct bkey_s_c k = bch2_bkey_get_iter(trans, &iter, BTREE_ID_deleted_inodes, p, 0); - int ret = bkey_err(k); - if (ret) + int ret = bkey_err(k) ?: k.k->type == KEY_TYPE_set; + bch2_trans_iter_exit(trans, &iter); + return ret; +} + +static int check_inode_dirent_inode(struct btree_trans *trans, struct bkey_s_c inode_k, + struct bch_inode_unpacked *inode, + u32 inode_snapshot, bool *write_inode) +{ + struct bch_fs *c = trans->c; + struct printbuf buf = PRINTBUF; + + struct btree_iter dirent_iter = {}; + struct bkey_s_c_dirent d = inode_get_dirent(trans, &dirent_iter, inode, &inode_snapshot); + int ret = bkey_err(d); + if (ret && !bch2_err_matches(ret, ENOENT)) return ret; - bch2_trans_iter_exit(trans, &iter); - return k.k->type == KEY_TYPE_set; + if (fsck_err_on(ret, + c, inode_points_to_missing_dirent, + "inode points to missing dirent\n%s", + (bch2_bkey_val_to_text(&buf, c, inode_k), buf.buf)) || + fsck_err_on(!ret && !dirent_points_to_inode(d, inode), + c, inode_points_to_wrong_dirent, + "inode points to dirent that does not point back:\n%s", + (bch2_bkey_val_to_text(&buf, c, inode_k), + prt_newline(&buf), + bch2_bkey_val_to_text(&buf, c, d.s_c), buf.buf))) { + /* + * We just clear the backpointer fields for now. If we find a + * dirent that points to this inode in check_dirents(), we'll + * update it then; then when we get to check_path() if the + * backpointer is still 0 we'll reattach it. + */ + inode->bi_dir = 0; + inode->bi_dir_offset = 0; + inode->bi_flags &= ~BCH_INODE_backptr_untrusted; + *write_inode = true; + } + + ret = 0; +fsck_err: + bch2_trans_iter_exit(trans, &dirent_iter); + printbuf_exit(&buf); + bch_err_fn(c, ret); + return ret; } static int check_inode(struct btree_trans *trans, @@ -861,7 +951,8 @@ static int check_inode(struct btree_trans *trans, u.bi_flags &= ~BCH_INODE_i_size_dirty|BCH_INODE_unlinked; - ret = __write_inode(trans, &u, iter->pos.snapshot); + ret = __bch2_fsck_write_inode(trans, &u, iter->pos.snapshot); + bch_err_msg(c, ret, "in fsck updating inode"); if (ret) return ret; @@ -876,7 +967,7 @@ static int check_inode(struct btree_trans *trans, if (ret < 0) return ret; - fsck_err_on(ret, c, unlinked_inode_not_on_deleted_list, + fsck_err_on(!ret, c, unlinked_inode_not_on_deleted_list, "inode %llu:%u unlinked, but not on deleted list", u.bi_inum, k.k->p.snapshot); ret = 0; @@ -950,8 +1041,49 @@ static int check_inode(struct btree_trans *trans, do_update = true; } + if (u.bi_dir || u.bi_dir_offset) { + ret = check_inode_dirent_inode(trans, k, &u, k.k->p.snapshot, &do_update); + if (ret) + goto err; + } + + if (fsck_err_on(u.bi_parent_subvol && + (u.bi_subvol == 0 || + u.bi_subvol == BCACHEFS_ROOT_SUBVOL), + c, inode_bi_parent_nonzero, + "inode %llu:%u has subvol %u but nonzero parent subvol %u", + u.bi_inum, k.k->p.snapshot, u.bi_subvol, u.bi_parent_subvol)) { + u.bi_parent_subvol = 0; + do_update = true; + } + + if (u.bi_subvol) { + struct bch_subvolume s; + + ret = bch2_subvolume_get(trans, u.bi_subvol, false, 0, &s); + if (ret && !bch2_err_matches(ret, ENOENT)) + goto err; + + if (fsck_err_on(ret, + c, inode_bi_subvol_missing, + "inode %llu:%u bi_subvol points to missing subvolume %u", + u.bi_inum, k.k->p.snapshot, u.bi_subvol) || + fsck_err_on(le64_to_cpu(s.inode) != u.bi_inum || + !bch2_snapshot_is_ancestor(c, le32_to_cpu(s.snapshot), + k.k->p.snapshot), + c, inode_bi_subvol_wrong, + "inode %llu:%u points to subvol %u, but subvol points to %llu:%u", + u.bi_inum, k.k->p.snapshot, u.bi_subvol, + le64_to_cpu(s.inode), + le32_to_cpu(s.snapshot))) { + u.bi_subvol = 0; + u.bi_parent_subvol = 0; + do_update = true; + } + } + if (do_update) { - ret = __write_inode(trans, &u, iter->pos.snapshot); + ret = __bch2_fsck_write_inode(trans, &u, iter->pos.snapshot); bch_err_msg(c, ret, "in fsck updating inode"); if (ret) return ret; @@ -982,32 +1114,9 @@ int bch2_check_inodes(struct bch_fs *c) return ret; } -static struct bkey_s_c_dirent dirent_get_by_pos(struct btree_trans *trans, - struct btree_iter *iter, - struct bpos pos) -{ - return bch2_bkey_get_iter_typed(trans, iter, BTREE_ID_dirents, pos, 0, dirent); -} - -static bool inode_points_to_dirent(struct bch_inode_unpacked *inode, - struct bkey_s_c_dirent d) -{ - return inode->bi_dir == d.k->p.inode && - inode->bi_dir_offset == d.k->p.offset; -} - -static bool dirent_points_to_inode(struct bkey_s_c_dirent d, - struct bch_inode_unpacked *inode) -{ - return d.v->d_type == DT_SUBVOL - ? le32_to_cpu(d.v->d_child_subvol) == inode->bi_subvol - : le64_to_cpu(d.v->d_inum) == inode->bi_inum; -} - -static int check_i_sectors(struct btree_trans *trans, struct inode_walker *w) +static int check_i_sectors_notnested(struct btree_trans *trans, struct inode_walker *w) { struct bch_fs *c = trans->c; - u32 restart_count = trans->restart_count; int ret = 0; s64 count2; @@ -1032,14 +1141,21 @@ static int check_i_sectors(struct btree_trans *trans, struct inode_walker *w) w->last_pos.inode, i->snapshot, i->inode.bi_sectors, i->count)) { i->inode.bi_sectors = i->count; - ret = fsck_write_inode(trans, &i->inode, i->snapshot); + ret = bch2_fsck_write_inode(trans, &i->inode, i->snapshot); if (ret) break; } } fsck_err: bch_err_fn(c, ret); - return ret ?: trans_was_restarted(trans, restart_count); + return ret; +} + +static int check_i_sectors(struct btree_trans *trans, struct inode_walker *w) +{ + u32 restart_count = trans->restart_count; + return check_i_sectors_notnested(trans, w) ?: + trans_was_restarted(trans, restart_count); } struct extent_end { @@ -1312,7 +1428,7 @@ static int check_extent(struct btree_trans *trans, struct btree_iter *iter, goto err; } - i = walk_inode(trans, inode, equiv, k.k->type == KEY_TYPE_whiteout); + i = walk_inode(trans, inode, k); ret = PTR_ERR_OR_ZERO(i); if (ret) goto err; @@ -1423,7 +1539,7 @@ int bch2_check_extents(struct bch_fs *c) check_extent(trans, &iter, k, &w, &s, &extent_ends) ?: check_extent_overbig(trans, &iter, k); })) ?: - check_i_sectors(trans, &w)); + check_i_sectors_notnested(trans, &w)); bch2_disk_reservation_put(c, &res); extent_ends_exit(&extent_ends); @@ -1453,10 +1569,9 @@ int bch2_check_indirect_extents(struct bch_fs *c) return ret; } -static int check_subdir_count(struct btree_trans *trans, struct inode_walker *w) +static int check_subdir_count_notnested(struct btree_trans *trans, struct inode_walker *w) { struct bch_fs *c = trans->c; - u32 restart_count = trans->restart_count; int ret = 0; s64 count2; @@ -1481,96 +1596,123 @@ static int check_subdir_count(struct btree_trans *trans, struct inode_walker *w) "directory %llu:%u with wrong i_nlink: got %u, should be %llu", w->last_pos.inode, i->snapshot, i->inode.bi_nlink, i->count)) { i->inode.bi_nlink = i->count; - ret = fsck_write_inode(trans, &i->inode, i->snapshot); + ret = bch2_fsck_write_inode(trans, &i->inode, i->snapshot); if (ret) break; } } fsck_err: bch_err_fn(c, ret); - return ret ?: trans_was_restarted(trans, restart_count); + return ret; } -static int check_dirent_target(struct btree_trans *trans, - struct btree_iter *iter, - struct bkey_s_c_dirent d, - struct bch_inode_unpacked *target, - u32 target_snapshot) +static int check_subdir_count(struct btree_trans *trans, struct inode_walker *w) +{ + u32 restart_count = trans->restart_count; + return check_subdir_count_notnested(trans, w) ?: + trans_was_restarted(trans, restart_count); +} + +static int check_dirent_inode_dirent(struct btree_trans *trans, + struct btree_iter *iter, + struct bkey_s_c_dirent d, + struct bch_inode_unpacked *target, + u32 target_snapshot) { struct bch_fs *c = trans->c; - struct bkey_i_dirent *n; struct printbuf buf = PRINTBUF; - struct btree_iter bp_iter = { NULL }; int ret = 0; + if (inode_points_to_dirent(target, d)) + return 0; + if (!target->bi_dir && !target->bi_dir_offset) { target->bi_dir = d.k->p.inode; target->bi_dir_offset = d.k->p.offset; - - ret = __write_inode(trans, target, target_snapshot); - if (ret) - goto err; + return __bch2_fsck_write_inode(trans, target, target_snapshot); } - if (!inode_points_to_dirent(target, d)) { - struct bkey_s_c_dirent bp_dirent = dirent_get_by_pos(trans, &bp_iter, - SPOS(target->bi_dir, target->bi_dir_offset, target_snapshot)); - ret = bkey_err(bp_dirent); - if (ret && !bch2_err_matches(ret, ENOENT)) - goto err; + struct btree_iter bp_iter = { NULL }; + struct bkey_s_c_dirent bp_dirent = dirent_get_by_pos(trans, &bp_iter, + SPOS(target->bi_dir, target->bi_dir_offset, target_snapshot)); + ret = bkey_err(bp_dirent); + if (ret && !bch2_err_matches(ret, ENOENT)) + goto err; - bool backpointer_exists = !ret; - ret = 0; + bool backpointer_exists = !ret; + ret = 0; + + if (fsck_err_on(!backpointer_exists, + c, inode_wrong_backpointer, + "inode %llu:%u has wrong backpointer:\n" + "got %llu:%llu\n" + "should be %llu:%llu", + target->bi_inum, target_snapshot, + target->bi_dir, + target->bi_dir_offset, + d.k->p.inode, + d.k->p.offset)) { + target->bi_dir = d.k->p.inode; + target->bi_dir_offset = d.k->p.offset; + ret = __bch2_fsck_write_inode(trans, target, target_snapshot); + goto out; + } - bch2_bkey_val_to_text(&buf, c, d.s_c); - prt_newline(&buf); - if (backpointer_exists) - bch2_bkey_val_to_text(&buf, c, bp_dirent.s_c); + bch2_bkey_val_to_text(&buf, c, d.s_c); + prt_newline(&buf); + if (backpointer_exists) + bch2_bkey_val_to_text(&buf, c, bp_dirent.s_c); + + if (fsck_err_on(backpointer_exists && + (S_ISDIR(target->bi_mode) || + target->bi_subvol), + c, inode_dir_multiple_links, + "%s %llu:%u with multiple links\n%s", + S_ISDIR(target->bi_mode) ? "directory" : "subvolume", + target->bi_inum, target_snapshot, buf.buf)) { + ret = __remove_dirent(trans, d.k->p); + goto out; + } - if (fsck_err_on(S_ISDIR(target->bi_mode) && backpointer_exists, - c, inode_dir_multiple_links, - "directory %llu:%u with multiple links\n%s", - target->bi_inum, target_snapshot, buf.buf)) { - ret = __remove_dirent(trans, d.k->p); - goto out; - } + /* + * hardlinked file with nlink 0: + * We're just adjusting nlink here so check_nlinks() will pick + * it up, it ignores inodes with nlink 0 + */ + if (fsck_err_on(backpointer_exists && !target->bi_nlink, + c, inode_multiple_links_but_nlink_0, + "inode %llu:%u type %s has multiple links but i_nlink 0\n%s", + target->bi_inum, target_snapshot, bch2_d_types[d.v->d_type], buf.buf)) { + target->bi_nlink++; + target->bi_flags &= ~BCH_INODE_unlinked; + ret = __bch2_fsck_write_inode(trans, target, target_snapshot); + if (ret) + goto err; + } +out: +err: +fsck_err: + bch2_trans_iter_exit(trans, &bp_iter); + printbuf_exit(&buf); + bch_err_fn(c, ret); + return ret; +} - /* - * hardlinked file with nlink 0: - * We're just adjusting nlink here so check_nlinks() will pick - * it up, it ignores inodes with nlink 0 - */ - if (fsck_err_on(backpointer_exists && !target->bi_nlink, - c, inode_multiple_links_but_nlink_0, - "inode %llu:%u type %s has multiple links but i_nlink 0\n%s", - target->bi_inum, target_snapshot, bch2_d_types[d.v->d_type], buf.buf)) { - target->bi_nlink++; - target->bi_flags &= ~BCH_INODE_unlinked; - - ret = __write_inode(trans, target, target_snapshot); - if (ret) - goto err; - } +static int check_dirent_target(struct btree_trans *trans, + struct btree_iter *iter, + struct bkey_s_c_dirent d, + struct bch_inode_unpacked *target, + u32 target_snapshot) +{ + struct bch_fs *c = trans->c; + struct bkey_i_dirent *n; + struct printbuf buf = PRINTBUF; + int ret = 0; - if (fsck_err_on(!backpointer_exists, - c, inode_wrong_backpointer, - "inode %llu:%u has wrong backpointer:\n" - "got %llu:%llu\n" - "should be %llu:%llu", - target->bi_inum, target_snapshot, - target->bi_dir, - target->bi_dir_offset, - d.k->p.inode, - d.k->p.offset)) { - target->bi_dir = d.k->p.inode; - target->bi_dir_offset = d.k->p.offset; - - ret = __write_inode(trans, target, target_snapshot); - if (ret) - goto err; - } - } + ret = check_dirent_inode_dirent(trans, iter, d, target, target_snapshot); + if (ret) + goto err; if (fsck_err_on(d.v->d_type != inode_d_type(target), c, dirent_d_type_wrong, @@ -1586,6 +1728,12 @@ static int check_dirent_target(struct btree_trans *trans, bkey_reassemble(&n->k_i, d.s_c); n->v.d_type = inode_d_type(target); + if (n->v.d_type == DT_SUBVOL) { + n->v.d_parent_subvol = cpu_to_le32(target->bi_parent_subvol); + n->v.d_child_subvol = cpu_to_le32(target->bi_subvol); + } else { + n->v.d_inum = cpu_to_le64(target->bi_inum); + } ret = bch2_trans_update(trans, iter, &n->k_i, 0); if (ret) @@ -1593,33 +1741,134 @@ static int check_dirent_target(struct btree_trans *trans, d = dirent_i_to_s_c(n); } +err: +fsck_err: + printbuf_exit(&buf); + bch_err_fn(c, ret); + return ret; +} - if (fsck_err_on(d.v->d_type == DT_SUBVOL && - target->bi_parent_subvol != le32_to_cpu(d.v->d_parent_subvol), - c, dirent_d_parent_subvol_wrong, - "dirent has wrong d_parent_subvol field: got %u, should be %u", - le32_to_cpu(d.v->d_parent_subvol), - target->bi_parent_subvol)) { - n = bch2_trans_kmalloc(trans, bkey_bytes(d.k)); - ret = PTR_ERR_OR_ZERO(n); +/* find a subvolume that's a descendent of @snapshot: */ +static int find_snapshot_subvol(struct btree_trans *trans, u32 snapshot, u32 *subvolid) +{ + struct btree_iter iter; + struct bkey_s_c k; + int ret; + + for_each_btree_key_norestart(trans, iter, BTREE_ID_subvolumes, POS_MIN, 0, k, ret) { + if (k.k->type != KEY_TYPE_subvolume) + continue; + + struct bkey_s_c_subvolume s = bkey_s_c_to_subvolume(k); + if (bch2_snapshot_is_ancestor(trans->c, le32_to_cpu(s.v->snapshot), snapshot)) { + bch2_trans_iter_exit(trans, &iter); + *subvolid = k.k->p.offset; + goto found; + } + } + if (!ret) + ret = -ENOENT; +found: + bch2_trans_iter_exit(trans, &iter); + return ret; +} + +static int check_dirent_to_subvol(struct btree_trans *trans, struct btree_iter *iter, + struct bkey_s_c_dirent d) +{ + struct bch_fs *c = trans->c; + struct btree_iter subvol_iter = {}; + struct bch_inode_unpacked subvol_root; + u32 parent_subvol = le32_to_cpu(d.v->d_parent_subvol); + u32 target_subvol = le32_to_cpu(d.v->d_child_subvol); + u32 parent_snapshot; + u64 parent_inum; + struct printbuf buf = PRINTBUF; + int ret = 0; + + ret = subvol_lookup(trans, parent_subvol, &parent_snapshot, &parent_inum); + if (ret && !bch2_err_matches(ret, ENOENT)) + return ret; + + if (fsck_err_on(ret, c, dirent_to_missing_parent_subvol, + "dirent parent_subvol points to missing subvolume\n%s", + (bch2_bkey_val_to_text(&buf, c, d.s_c), buf.buf)) || + fsck_err_on(!ret && !bch2_snapshot_is_ancestor(c, parent_snapshot, d.k->p.snapshot), + c, dirent_not_visible_in_parent_subvol, + "dirent not visible in parent_subvol (not an ancestor of subvol snap %u)\n%s", + parent_snapshot, + (bch2_bkey_val_to_text(&buf, c, d.s_c), buf.buf))) { + u32 new_parent_subvol; + ret = find_snapshot_subvol(trans, d.k->p.snapshot, &new_parent_subvol); if (ret) goto err; - bkey_reassemble(&n->k_i, d.s_c); - n->v.d_parent_subvol = cpu_to_le32(target->bi_parent_subvol); + struct bkey_i_dirent *new_dirent = bch2_bkey_make_mut_typed(trans, iter, &d.s_c, 0, dirent); + ret = PTR_ERR_OR_ZERO(new_dirent); + if (ret) + goto err; - ret = bch2_trans_update(trans, iter, &n->k_i, 0); + new_dirent->v.d_parent_subvol = cpu_to_le32(new_parent_subvol); + } + + struct bkey_s_c_subvolume s = + bch2_bkey_get_iter_typed(trans, &subvol_iter, + BTREE_ID_subvolumes, POS(0, target_subvol), + 0, subvolume); + ret = bkey_err(s.s_c); + if (ret && !bch2_err_matches(ret, ENOENT)) + return ret; + + if (ret) { + if (fsck_err(c, dirent_to_missing_subvol, + "dirent points to missing subvolume\n%s", + (bch2_bkey_val_to_text(&buf, c, d.s_c), buf.buf))) + return __remove_dirent(trans, d.k->p); + ret = 0; + goto out; + } + + if (fsck_err_on(le32_to_cpu(s.v->fs_path_parent) != parent_subvol, + c, subvol_fs_path_parent_wrong, + "subvol with wrong fs_path_parent, should be be %u\n%s", + parent_subvol, + (bch2_bkey_val_to_text(&buf, c, s.s_c), buf.buf))) { + struct bkey_i_subvolume *n = + bch2_bkey_make_mut_typed(trans, &subvol_iter, &s.s_c, 0, subvolume); + ret = PTR_ERR_OR_ZERO(n); if (ret) goto err; - d = dirent_i_to_s_c(n); + n->v.fs_path_parent = cpu_to_le32(parent_subvol); + } + + u64 target_inum = le64_to_cpu(s.v->inode); + u32 target_snapshot = le32_to_cpu(s.v->snapshot); + + ret = lookup_inode(trans, target_inum, &subvol_root, &target_snapshot); + if (ret && !bch2_err_matches(ret, ENOENT)) + return ret; + + if (fsck_err_on(parent_subvol != subvol_root.bi_parent_subvol, + c, inode_bi_parent_wrong, + "subvol root %llu has wrong bi_parent_subvol: got %u, should be %u", + target_inum, + subvol_root.bi_parent_subvol, parent_subvol)) { + subvol_root.bi_parent_subvol = parent_subvol; + ret = __bch2_fsck_write_inode(trans, &subvol_root, target_snapshot); + if (ret) + return ret; } + + ret = check_dirent_target(trans, iter, d, &subvol_root, + target_snapshot); + if (ret) + return ret; out: err: fsck_err: - bch2_trans_iter_exit(trans, &bp_iter); + bch2_trans_iter_exit(trans, &subvol_iter); printbuf_exit(&buf); - bch_err_fn(c, ret); return ret; } @@ -1661,7 +1910,7 @@ static int check_dirent(struct btree_trans *trans, struct btree_iter *iter, BUG_ON(!btree_iter_path(trans, iter)->should_be_locked); - i = walk_inode(trans, dir, equiv, k.k->type == KEY_TYPE_whiteout); + i = walk_inode(trans, dir, k); ret = PTR_ERR_OR_ZERO(i); if (ret < 0) goto err; @@ -1707,50 +1956,7 @@ static int check_dirent(struct btree_trans *trans, struct btree_iter *iter, d = bkey_s_c_to_dirent(k); if (d.v->d_type == DT_SUBVOL) { - struct bch_inode_unpacked subvol_root; - u32 target_subvol = le32_to_cpu(d.v->d_child_subvol); - u32 target_snapshot; - u64 target_inum; - - ret = subvol_lookup(trans, target_subvol, - &target_snapshot, &target_inum); - if (ret && !bch2_err_matches(ret, ENOENT)) - goto err; - - if (fsck_err_on(ret, c, dirent_to_missing_subvol, - "dirent points to missing subvolume %u", - le32_to_cpu(d.v->d_child_subvol))) { - ret = __remove_dirent(trans, d.k->p); - goto err; - } - - ret = lookup_inode(trans, target_inum, - &subvol_root, &target_snapshot); - if (ret && !bch2_err_matches(ret, ENOENT)) - goto err; - - if (fsck_err_on(ret, c, subvol_to_missing_root, - "subvolume %u points to missing subvolume root %llu", - target_subvol, - target_inum)) { - bch_err(c, "repair not implemented yet"); - ret = -EINVAL; - goto err; - } - - if (fsck_err_on(subvol_root.bi_subvol != target_subvol, - c, subvol_root_wrong_bi_subvol, - "subvol root %llu has wrong bi_subvol field: got %u, should be %u", - target_inum, - subvol_root.bi_subvol, target_subvol)) { - subvol_root.bi_subvol = target_subvol; - ret = __write_inode(trans, &subvol_root, target_snapshot); - if (ret) - goto err; - } - - ret = check_dirent_target(trans, iter, d, &subvol_root, - target_snapshot); + ret = check_dirent_to_subvol(trans, iter, d); if (ret) goto err; } else { @@ -1776,12 +1982,11 @@ static int check_dirent(struct btree_trans *trans, struct btree_iter *iter, if (ret) goto err; } - } - - if (d.v->d_type == DT_DIR) - for_each_visible_inode(c, s, dir, equiv.snapshot, i) - i->count++; + if (d.v->d_type == DT_DIR) + for_each_visible_inode(c, s, dir, equiv.snapshot, i) + i->count++; + } out: err: fsck_err: @@ -1810,7 +2015,8 @@ int bch2_check_dirents(struct bch_fs *c) k, NULL, NULL, BCH_TRANS_COMMIT_no_enospc, - check_dirent(trans, &iter, k, &hash_info, &dir, &target, &s))); + check_dirent(trans, &iter, k, &hash_info, &dir, &target, &s)) ?: + check_subdir_count_notnested(trans, &dir)); snapshots_seen_exit(&s); inode_walker_exit(&dir); @@ -1829,10 +2035,12 @@ static int check_xattr(struct btree_trans *trans, struct btree_iter *iter, int ret; ret = check_key_has_snapshot(trans, iter, k); - if (ret) + if (ret < 0) return ret; + if (ret) + return 0; - i = walk_inode(trans, inode, k.k->p, k.k->type == KEY_TYPE_whiteout); + i = walk_inode(trans, inode, k); ret = PTR_ERR_OR_ZERO(i); if (ret) return ret; @@ -1919,7 +2127,7 @@ static int check_root_trans(struct btree_trans *trans) 0, NULL); root_inode.bi_inum = inum; - ret = __write_inode(trans, &root_inode, snapshot); + ret = __bch2_fsck_write_inode(trans, &root_inode, snapshot); bch_err_msg(c, ret, "writing root inode"); } err: @@ -1936,6 +2144,107 @@ int bch2_check_root(struct bch_fs *c) return ret; } +typedef DARRAY(u32) darray_u32; + +static bool darray_u32_has(darray_u32 *d, u32 v) +{ + darray_for_each(*d, i) + if (*i == v) + return true; + return false; +} + +/* + * We've checked that inode backpointers point to valid dirents; here, it's + * sufficient to check that the subvolume root has a dirent: + */ +static int subvol_has_dirent(struct btree_trans *trans, struct bkey_s_c_subvolume s) +{ + struct bch_inode_unpacked inode; + int ret = bch2_inode_find_by_inum_trans(trans, + (subvol_inum) { s.k->p.offset, le64_to_cpu(s.v->inode) }, + &inode); + if (ret) + return ret; + + return inode.bi_dir != 0; +} + +static int check_subvol_path(struct btree_trans *trans, struct btree_iter *iter, struct bkey_s_c k) +{ + struct bch_fs *c = trans->c; + struct btree_iter parent_iter = {}; + darray_u32 subvol_path = {}; + struct printbuf buf = PRINTBUF; + int ret = 0; + + if (k.k->type != KEY_TYPE_subvolume) + return 0; + + while (k.k->p.offset != BCACHEFS_ROOT_SUBVOL) { + ret = darray_push(&subvol_path, k.k->p.offset); + if (ret) + goto err; + + struct bkey_s_c_subvolume s = bkey_s_c_to_subvolume(k); + + ret = subvol_has_dirent(trans, s); + if (ret < 0) + break; + + if (fsck_err_on(!ret, + c, subvol_unreachable, + "unreachable subvolume %s", + (bch2_bkey_val_to_text(&buf, c, s.s_c), + buf.buf))) { + ret = reattach_subvol(trans, s); + break; + } + + u32 parent = le32_to_cpu(s.v->fs_path_parent); + + if (darray_u32_has(&subvol_path, parent)) { + if (fsck_err(c, subvol_loop, "subvolume loop")) + ret = reattach_subvol(trans, s); + break; + } + + bch2_trans_iter_exit(trans, &parent_iter); + bch2_trans_iter_init(trans, &parent_iter, + BTREE_ID_subvolumes, POS(0, parent), 0); + k = bch2_btree_iter_peek_slot(&parent_iter); + ret = bkey_err(k); + if (ret) + goto err; + + if (fsck_err_on(k.k->type != KEY_TYPE_subvolume, + c, subvol_unreachable, + "unreachable subvolume %s", + (bch2_bkey_val_to_text(&buf, c, s.s_c), + buf.buf))) { + ret = reattach_subvol(trans, s); + break; + } + } +fsck_err: +err: + printbuf_exit(&buf); + darray_exit(&subvol_path); + bch2_trans_iter_exit(trans, &parent_iter); + return ret; +} + +int bch2_check_subvolume_structure(struct bch_fs *c) +{ + int ret = bch2_trans_run(c, + for_each_btree_key_commit(trans, iter, + BTREE_ID_subvolumes, POS_MIN, BTREE_ITER_PREFETCH, k, + NULL, NULL, BCH_TRANS_COMMIT_no_enospc, + check_subvol_path(trans, &iter, k))); + bch_err_fn(c, ret); + return ret; +} + struct pathbuf_entry { u64 inum; u32 snapshot; @@ -1952,89 +2261,71 @@ static bool path_is_dup(pathbuf *p, u64 inum, u32 snapshot) return false; } -static int path_down(struct bch_fs *c, pathbuf *p, - u64 inum, u32 snapshot) -{ - int ret = darray_push(p, ((struct pathbuf_entry) { - .inum = inum, - .snapshot = snapshot, - })); - - if (ret) - bch_err(c, "fsck: error allocating memory for pathbuf, size %zu", - p->size); - return ret; -} - /* - * Check that a given inode is reachable from the root: + * Check that a given inode is reachable from its subvolume root - we already + * verified subvolume connectivity: * * XXX: we should also be verifying that inodes are in the right subvolumes */ -static int check_path(struct btree_trans *trans, - pathbuf *p, - struct bch_inode_unpacked *inode, - u32 snapshot) +static int check_path(struct btree_trans *trans, pathbuf *p, struct bkey_s_c inode_k) { struct bch_fs *c = trans->c; + struct btree_iter inode_iter = {}; + struct bch_inode_unpacked inode; + struct printbuf buf = PRINTBUF; + u32 snapshot = bch2_snapshot_equiv(c, inode_k.k->p.snapshot); int ret = 0; - snapshot = bch2_snapshot_equiv(c, snapshot); p->nr = 0; - while (!(inode->bi_inum == BCACHEFS_ROOT_INO && - inode->bi_subvol == BCACHEFS_ROOT_SUBVOL)) { + BUG_ON(bch2_inode_unpack(inode_k, &inode)); + + while (!inode.bi_subvol) { struct btree_iter dirent_iter; struct bkey_s_c_dirent d; u32 parent_snapshot = snapshot; - if (inode->bi_subvol) { - u64 inum; - - ret = subvol_lookup(trans, inode->bi_parent_subvol, - &parent_snapshot, &inum); - if (ret) - break; - } - - d = dirent_get_by_pos(trans, &dirent_iter, - SPOS(inode->bi_dir, inode->bi_dir_offset, - parent_snapshot)); + d = inode_get_dirent(trans, &dirent_iter, &inode, &parent_snapshot); ret = bkey_err(d.s_c); if (ret && !bch2_err_matches(ret, ENOENT)) break; - if (!ret && !dirent_points_to_inode(d, inode)) { + if (!ret && !dirent_points_to_inode(d, &inode)) { bch2_trans_iter_exit(trans, &dirent_iter); ret = -BCH_ERR_ENOENT_dirent_doesnt_match_inode; } if (bch2_err_matches(ret, ENOENT)) { - if (fsck_err(c, inode_unreachable, - "unreachable inode %llu:%u, type %s nlink %u backptr %llu:%llu", - inode->bi_inum, snapshot, - bch2_d_type_str(inode_d_type(inode)), - inode->bi_nlink, - inode->bi_dir, - inode->bi_dir_offset)) - ret = reattach_inode(trans, inode, snapshot); - break; + ret = 0; + if (fsck_err(c, inode_unreachable, + "unreachable inode\n%s", + (printbuf_reset(&buf), + bch2_bkey_val_to_text(&buf, c, inode_k), + buf.buf))) + ret = reattach_inode(trans, &inode, snapshot); + goto out; } bch2_trans_iter_exit(trans, &dirent_iter); - if (!S_ISDIR(inode->bi_mode)) + if (!S_ISDIR(inode.bi_mode)) break; - ret = path_down(c, p, inode->bi_inum, snapshot); - if (ret) { - bch_err(c, "memory allocation failure"); + ret = darray_push(p, ((struct pathbuf_entry) { + .inum = inode.bi_inum, + .snapshot = snapshot, + })); + if (ret) return ret; - } snapshot = parent_snapshot; - ret = lookup_inode(trans, inode->bi_dir, inode, &snapshot); + bch2_trans_iter_exit(trans, &inode_iter); + inode_k = bch2_bkey_get_iter(trans, &inode_iter, BTREE_ID_inodes, + SPOS(0, inode.bi_dir, snapshot), 0); + ret = bkey_err(inode_k) ?: + !bkey_is_inode(inode_k.k) ? -BCH_ERR_ENOENT_inode + : bch2_inode_unpack(inode_k, &inode); if (ret) { /* Should have been caught in dirents pass */ if (!bch2_err_matches(ret, BCH_ERR_transaction_restart)) @@ -2042,30 +2333,32 @@ static int check_path(struct btree_trans *trans, break; } - if (path_is_dup(p, inode->bi_inum, snapshot)) { + snapshot = inode_k.k->p.snapshot; + + if (path_is_dup(p, inode.bi_inum, snapshot)) { /* XXX print path */ bch_err(c, "directory structure loop"); darray_for_each(*p, i) pr_err("%llu:%u", i->inum, i->snapshot); - pr_err("%llu:%u", inode->bi_inum, snapshot); - - if (!fsck_err(c, dir_loop, "directory structure loop")) - return 0; + pr_err("%llu:%u", inode.bi_inum, snapshot); - ret = remove_backpointer(trans, inode); - if (ret && !bch2_err_matches(ret, BCH_ERR_transaction_restart)) + if (fsck_err(c, dir_loop, "directory structure loop")) { + ret = remove_backpointer(trans, &inode); bch_err_msg(c, ret, "removing dirent"); - if (ret) - break; + if (ret) + break; - ret = reattach_inode(trans, inode, snapshot); - if (ret && !bch2_err_matches(ret, BCH_ERR_transaction_restart)) - bch_err_msg(c, ret, "reattaching inode %llu", inode->bi_inum); + ret = reattach_inode(trans, &inode, snapshot); + bch_err_msg(c, ret, "reattaching inode %llu", inode.bi_inum); + } break; } } +out: fsck_err: + bch2_trans_iter_exit(trans, &inode_iter); + printbuf_exit(&buf); bch_err_fn(c, ret); return ret; } @@ -2077,7 +2370,6 @@ fsck_err: */ int bch2_check_directory_structure(struct bch_fs *c) { - struct bch_inode_unpacked u; pathbuf path = { 0, }; int ret; @@ -2090,12 +2382,10 @@ int bch2_check_directory_structure(struct bch_fs *c) if (!bkey_is_inode(k.k)) continue; - BUG_ON(bch2_inode_unpack(k, &u)); - - if (u.bi_flags & BCH_INODE_unlinked) + if (bch2_inode_flags(k) & BCH_INODE_unlinked) continue; - check_path(trans, &path, &u, iter.pos.snapshot); + check_path(trans, &path, k); }))); darray_exit(&path); @@ -2291,7 +2581,7 @@ static int check_nlinks_update_inode(struct btree_trans *trans, struct btree_ite u.bi_inum, bch2_d_types[mode_to_type(u.bi_mode)], bch2_inode_nlink_get(&u), link->count)) { bch2_inode_nlink_set(&u, link->count); - ret = __write_inode(trans, &u, k.k->p.snapshot); + ret = __bch2_fsck_write_inode(trans, &u, k.k->p.snapshot); } fsck_err: return ret; diff --git a/fs/bcachefs/fsck.h b/fs/bcachefs/fsck.h index da991e8cf27e..a4ef94271784 100644 --- a/fs/bcachefs/fsck.h +++ b/fs/bcachefs/fsck.h @@ -8,6 +8,7 @@ int bch2_check_indirect_extents(struct bch_fs *); int bch2_check_dirents(struct bch_fs *); int bch2_check_xattrs(struct bch_fs *); int bch2_check_root(struct bch_fs *); +int bch2_check_subvolume_structure(struct bch_fs *); int bch2_check_directory_structure(struct bch_fs *); int bch2_check_nlinks(struct bch_fs *); int bch2_fix_reflink_p(struct bch_fs *); diff --git a/fs/bcachefs/inode.c b/fs/bcachefs/inode.c index 086f0090b03a..2b5e06770ab3 100644 --- a/fs/bcachefs/inode.c +++ b/fs/bcachefs/inode.c @@ -324,7 +324,7 @@ int bch2_inode_unpack(struct bkey_s_c k, return bch2_inode_unpack_slowpath(k, unpacked); } -static int bch2_inode_peek_nowarn(struct btree_trans *trans, +int bch2_inode_peek_nowarn(struct btree_trans *trans, struct btree_iter *iter, struct bch_inode_unpacked *inode, subvol_inum inum, unsigned flags) @@ -384,6 +384,34 @@ int bch2_inode_write_flags(struct btree_trans *trans, return bch2_trans_update(trans, iter, &inode_p->inode.k_i, flags); } +int __bch2_fsck_write_inode(struct btree_trans *trans, + struct bch_inode_unpacked *inode, + u32 snapshot) +{ + struct bkey_inode_buf *inode_p = + bch2_trans_kmalloc(trans, sizeof(*inode_p)); + + if (IS_ERR(inode_p)) + return PTR_ERR(inode_p); + + bch2_inode_pack(inode_p, inode); + inode_p->inode.k.p.snapshot = snapshot; + + return bch2_btree_insert_nonextent(trans, BTREE_ID_inodes, + &inode_p->inode.k_i, + BTREE_UPDATE_INTERNAL_SNAPSHOT_NODE); +} + +int bch2_fsck_write_inode(struct btree_trans *trans, + struct bch_inode_unpacked *inode, + u32 snapshot) +{ + int ret = commit_do(trans, NULL, NULL, BCH_TRANS_COMMIT_no_enospc, + __bch2_fsck_write_inode(trans, inode, snapshot)); + bch_err_fn(trans->c, ret); + return ret; +} + struct bkey_i *bch2_inode_to_v3(struct btree_trans *trans, struct bkey_i *k) { struct bch_inode_unpacked u; @@ -592,7 +620,8 @@ int bch2_trigger_inode(struct btree_trans *trans, bool old_deleted = bkey_is_deleted_inode(old); bool new_deleted = bkey_is_deleted_inode(new.s_c); if (old_deleted != new_deleted) { - int ret = bch2_btree_bit_mod(trans, BTREE_ID_deleted_inodes, new.k->p, new_deleted); + int ret = bch2_btree_bit_mod_buffered(trans, BTREE_ID_deleted_inodes, + new.k->p, new_deleted); if (ret) return ret; } @@ -1088,8 +1117,9 @@ static int may_delete_deleted_inode(struct btree_trans *trans, goto out; if (S_ISDIR(inode.bi_mode)) { - ret = bch2_empty_dir_snapshot(trans, pos.offset, pos.snapshot); - if (fsck_err_on(ret == -ENOTEMPTY, c, deleted_inode_is_dir, + ret = bch2_empty_dir_snapshot(trans, pos.offset, 0, pos.snapshot); + if (fsck_err_on(bch2_err_matches(ret, ENOTEMPTY), + c, deleted_inode_is_dir, "non empty directory %llu:%u in deleted_inodes btree", pos.offset, pos.snapshot)) goto delete; @@ -1141,7 +1171,7 @@ fsck_err: bch2_trans_iter_exit(trans, &inode_iter); return ret; delete: - ret = bch2_btree_bit_mod(trans, BTREE_ID_deleted_inodes, pos, false); + ret = bch2_btree_bit_mod_buffered(trans, BTREE_ID_deleted_inodes, pos, false); goto out; } @@ -1151,6 +1181,15 @@ int bch2_delete_dead_inodes(struct bch_fs *c) bool need_another_pass; int ret; again: + /* + * if we ran check_inodes() unlinked inodes will have already been + * cleaned up but the write buffer will be out of sync; therefore we + * alway need a write buffer flush + */ + ret = bch2_btree_write_buffer_flush_sync(trans); + if (ret) + goto err; + need_another_pass = false; /* @@ -1183,12 +1222,8 @@ again: ret; })); - if (!ret && need_another_pass) { - ret = bch2_btree_write_buffer_flush_sync(trans); - if (ret) - goto err; + if (!ret && need_another_pass) goto again; - } err: bch2_trans_put(trans); return ret; diff --git a/fs/bcachefs/inode.h b/fs/bcachefs/inode.h index b63f312581cf..056298050550 100644 --- a/fs/bcachefs/inode.h +++ b/fs/bcachefs/inode.h @@ -95,6 +95,8 @@ struct bkey_i *bch2_inode_to_v3(struct btree_trans *, struct bkey_i *); void bch2_inode_unpacked_to_text(struct printbuf *, struct bch_inode_unpacked *); +int bch2_inode_peek_nowarn(struct btree_trans *, struct btree_iter *, + struct bch_inode_unpacked *, subvol_inum, unsigned); int bch2_inode_peek(struct btree_trans *, struct btree_iter *, struct bch_inode_unpacked *, subvol_inum, unsigned); @@ -108,6 +110,9 @@ static inline int bch2_inode_write(struct btree_trans *trans, return bch2_inode_write_flags(trans, iter, inode, 0); } +int __bch2_fsck_write_inode(struct btree_trans *, struct bch_inode_unpacked *, u32); +int bch2_fsck_write_inode(struct btree_trans *, struct bch_inode_unpacked *, u32); + void bch2_inode_init_early(struct bch_fs *, struct bch_inode_unpacked *); void bch2_inode_init_late(struct bch_inode_unpacked *, u64, @@ -172,6 +177,20 @@ static inline u8 inode_d_type(struct bch_inode_unpacked *inode) return inode->bi_subvol ? DT_SUBVOL : mode_to_type(inode->bi_mode); } +static inline u32 bch2_inode_flags(struct bkey_s_c k) +{ + switch (k.k->type) { + case KEY_TYPE_inode: + return le32_to_cpu(bkey_s_c_to_inode(k).v->bi_flags); + case KEY_TYPE_inode_v2: + return le64_to_cpu(bkey_s_c_to_inode_v2(k).v->bi_flags); + case KEY_TYPE_inode_v3: + return le64_to_cpu(bkey_s_c_to_inode_v3(k).v->bi_flags); + default: + return 0; + } +} + /* i_nlink: */ static inline unsigned nlink_bias(umode_t mode) diff --git a/fs/bcachefs/io_read.c b/fs/bcachefs/io_read.c index 3c574d8873a1..8a556e6d1ab6 100644 --- a/fs/bcachefs/io_read.c +++ b/fs/bcachefs/io_read.c @@ -174,7 +174,7 @@ static struct promote_op *__promote_alloc(struct btree_trans *trans, if (!bch2_write_ref_tryget(c, BCH_WRITE_REF_promote)) return ERR_PTR(-BCH_ERR_nopromote_no_writes); - op = kzalloc(sizeof(*op) + sizeof(struct bio_vec) * pages, GFP_KERNEL); + op = kzalloc(struct_size(op, bi_inline_vecs, pages), GFP_KERNEL); if (!op) { ret = -BCH_ERR_nopromote_enomem; goto err; diff --git a/fs/bcachefs/io_write.c b/fs/bcachefs/io_write.c index 2c098ac017b3..f137252bccc5 100644 --- a/fs/bcachefs/io_write.c +++ b/fs/bcachefs/io_write.c @@ -88,7 +88,7 @@ void bch2_latency_acct(struct bch_dev *ca, u64 submit_time, int rw) bch2_congested_acct(ca, io_latency, now, rw); - __bch2_time_stats_update(&ca->io_latency[rw], submit_time, now); + __bch2_time_stats_update(&ca->io_latency[rw].stats, submit_time, now); } #endif @@ -530,7 +530,8 @@ static void __bch2_write_index(struct bch_write_op *op) bch_err_inum_offset_ratelimited(c, insert->k.p.inode, insert->k.p.offset << 9, - "write error while doing btree update: %s", + "%s write error while doing btree update: %s", + op->flags & BCH_WRITE_MOVE ? "move" : "user", bch2_err_str(ret)); } @@ -1067,7 +1068,8 @@ do_write: *_dst = dst; return more; csum_err: - bch_err(c, "error verifying existing checksum while rewriting existing data (memory corruption?)"); + bch_err(c, "%s writ error: error verifying existing checksum while rewriting existing data (memory corruption?)", + op->flags & BCH_WRITE_MOVE ? "move" : "user"); ret = -EIO; err: if (to_wbio(dst)->bounce) @@ -1169,7 +1171,8 @@ static void bch2_nocow_write_convert_unwritten(struct bch_write_op *op) bch_err_inum_offset_ratelimited(c, insert->k.p.inode, insert->k.p.offset << 9, - "write error while doing btree update: %s", + "%s write error while doing btree update: %s", + op->flags & BCH_WRITE_MOVE ? "move" : "user", bch2_err_str(ret)); } @@ -1449,7 +1452,9 @@ err: bch_err_inum_offset_ratelimited(c, op->pos.inode, op->pos.offset << 9, - "%s(): error: %s", __func__, bch2_err_str(ret)); + "%s(): %s error: %s", __func__, + op->flags & BCH_WRITE_MOVE ? "move" : "user", + bch2_err_str(ret)); op->error = ret; break; } @@ -1573,7 +1578,8 @@ CLOSURE_CALLBACK(bch2_write) bch_err_inum_offset_ratelimited(c, op->pos.inode, op->pos.offset << 9, - "misaligned write"); + "%s write error: misaligned write", + op->flags & BCH_WRITE_MOVE ? "move" : "user"); op->error = -EIO; goto err; } diff --git a/fs/bcachefs/journal.c b/fs/bcachefs/journal.c index bc890776eb57..9c9a25dbd613 100644 --- a/fs/bcachefs/journal.c +++ b/fs/bcachefs/journal.c @@ -27,33 +27,71 @@ static const char * const bch2_journal_errors[] = { NULL }; +static inline bool journal_seq_unwritten(struct journal *j, u64 seq) +{ + return seq > j->seq_ondisk; +} + +static bool __journal_entry_is_open(union journal_res_state state) +{ + return state.cur_entry_offset < JOURNAL_ENTRY_CLOSED_VAL; +} + +static inline unsigned nr_unwritten_journal_entries(struct journal *j) +{ + return atomic64_read(&j->seq) - j->seq_ondisk; +} + +static bool journal_entry_is_open(struct journal *j) +{ + return __journal_entry_is_open(j->reservations); +} + static void bch2_journal_buf_to_text(struct printbuf *out, struct journal *j, u64 seq) { union journal_res_state s = READ_ONCE(j->reservations); unsigned i = seq & JOURNAL_BUF_MASK; struct journal_buf *buf = j->buf + i; - prt_printf(out, "seq:"); + prt_str(out, "seq:"); prt_tab(out); prt_printf(out, "%llu", seq); prt_newline(out); printbuf_indent_add(out, 2); - prt_printf(out, "refcount:"); + prt_str(out, "refcount:"); prt_tab(out); prt_printf(out, "%u", journal_state_count(s, i)); prt_newline(out); - prt_printf(out, "size:"); + prt_str(out, "size:"); prt_tab(out); prt_human_readable_u64(out, vstruct_bytes(buf->data)); prt_newline(out); - prt_printf(out, "expires"); + prt_str(out, "expires:"); prt_tab(out); prt_printf(out, "%li jiffies", buf->expires - jiffies); prt_newline(out); + prt_str(out, "flags:"); + prt_tab(out); + if (buf->noflush) + prt_str(out, "noflush "); + if (buf->must_flush) + prt_str(out, "must_flush "); + if (buf->separate_flush) + prt_str(out, "separate_flush "); + if (buf->need_flush_to_write_buffer) + prt_str(out, "need_flush_to_write_buffer "); + if (buf->write_started) + prt_str(out, "write_started "); + if (buf->write_allocated) + prt_str(out, "write allocated "); + if (buf->write_done) + prt_str(out, "write done"); + prt_newline(out); + printbuf_indent_sub(out, 2); } @@ -66,26 +104,7 @@ static void bch2_journal_bufs_to_text(struct printbuf *out, struct journal *j) seq <= journal_cur_seq(j); seq++) bch2_journal_buf_to_text(out, j, seq); -} - -static inline bool journal_seq_unwritten(struct journal *j, u64 seq) -{ - return seq > j->seq_ondisk; -} - -static bool __journal_entry_is_open(union journal_res_state state) -{ - return state.cur_entry_offset < JOURNAL_ENTRY_CLOSED_VAL; -} - -static inline unsigned nr_unwritten_journal_entries(struct journal *j) -{ - return atomic64_read(&j->seq) - j->seq_ondisk; -} - -static bool journal_entry_is_open(struct journal *j) -{ - return __journal_entry_is_open(j->reservations); + prt_printf(out, "last buf %s\n", journal_entry_is_open(j) ? "open" : "closed"); } static inline struct journal_buf * @@ -174,21 +193,40 @@ journal_error_check_stuck(struct journal *j, int error, unsigned flags) return stuck; } +void bch2_journal_do_writes(struct journal *j) +{ + for (u64 seq = journal_last_unwritten_seq(j); + seq <= journal_cur_seq(j); + seq++) { + unsigned idx = seq & JOURNAL_BUF_MASK; + struct journal_buf *w = j->buf + idx; + + if (w->write_started && !w->write_allocated) + break; + if (w->write_started) + continue; + + if (!journal_state_count(j->reservations, idx)) { + w->write_started = true; + closure_call(&w->io, bch2_journal_write, j->wq, NULL); + } + + break; + } +} + /* * Final processing when the last reference of a journal buffer has been * dropped. Drop the pin list reference acquired at journal entry open and write * the buffer, if requested. */ -void bch2_journal_buf_put_final(struct journal *j, u64 seq, bool write) +void bch2_journal_buf_put_final(struct journal *j, u64 seq) { - struct bch_fs *c = container_of(j, struct bch_fs, journal); - lockdep_assert_held(&j->lock); if (__bch2_journal_pin_put(j, seq)) bch2_journal_reclaim_fast(j); - if (write) - closure_call(&j->io, bch2_journal_write, c->io_complete_wq, NULL); + bch2_journal_do_writes(j); } /* @@ -380,11 +418,14 @@ static int journal_entry_open(struct journal *j) BUG_ON(j->buf + (journal_cur_seq(j) & JOURNAL_BUF_MASK) != buf); bkey_extent_init(&buf->key); - buf->noflush = false; - buf->must_flush = false; - buf->separate_flush = false; - buf->flush_time = 0; + buf->noflush = false; + buf->must_flush = false; + buf->separate_flush = false; + buf->flush_time = 0; buf->need_flush_to_write_buffer = true; + buf->write_started = false; + buf->write_allocated = false; + buf->write_done = false; memset(buf->data, 0, sizeof(*buf->data)); buf->data->seq = cpu_to_le64(journal_cur_seq(j)); @@ -418,9 +459,10 @@ static int journal_entry_open(struct journal *j) } while ((v = atomic64_cmpxchg(&j->reservations.counter, old.v, new.v)) != old.v); - mod_delayed_work(c->io_complete_wq, - &j->write_work, - msecs_to_jiffies(c->opts.journal_flush_delay)); + if (nr_unwritten_journal_entries(j) == 1) + mod_delayed_work(j->wq, + &j->write_work, + msecs_to_jiffies(c->opts.journal_flush_delay)); journal_wake(j); if (j->early_journal_entries.nr) @@ -445,20 +487,16 @@ static void journal_quiesce(struct journal *j) static void journal_write_work(struct work_struct *work) { struct journal *j = container_of(work, struct journal, write_work.work); - struct bch_fs *c = container_of(j, struct bch_fs, journal); - long delta; spin_lock(&j->lock); - if (!__journal_entry_is_open(j->reservations)) - goto unlock; + if (__journal_entry_is_open(j->reservations)) { + long delta = journal_cur_buf(j)->expires - jiffies; - delta = journal_cur_buf(j)->expires - jiffies; - - if (delta > 0) - mod_delayed_work(c->io_complete_wq, &j->write_work, delta); - else - __journal_entry_close(j, JOURNAL_ENTRY_CLOSED_VAL, true); -unlock: + if (delta > 0) + mod_delayed_work(j->wq, &j->write_work, delta); + else + __journal_entry_close(j, JOURNAL_ENTRY_CLOSED_VAL, true); + } spin_unlock(&j->lock); } @@ -476,30 +514,29 @@ retry: if (bch2_journal_error(j)) return -BCH_ERR_erofs_journal_err; - spin_lock(&j->lock); + if (j->blocked) + return -BCH_ERR_journal_res_get_blocked; - /* check once more in case somebody else shut things down... */ - if (bch2_journal_error(j)) { - spin_unlock(&j->lock); - return -BCH_ERR_erofs_journal_err; + if ((flags & BCH_WATERMARK_MASK) < j->watermark) { + ret = JOURNAL_ERR_journal_full; + can_discard = j->can_discard; + goto out; + } + + if (nr_unwritten_journal_entries(j) == ARRAY_SIZE(j->buf) && !journal_entry_is_open(j)) { + ret = JOURNAL_ERR_max_in_flight; + goto out; } + spin_lock(&j->lock); + /* * Recheck after taking the lock, so we don't race with another thread * that just did journal_entry_open() and call bch2_journal_entry_close() * unnecessarily */ if (journal_res_get_fast(j, res, flags)) { - spin_unlock(&j->lock); - return 0; - } - - if ((flags & BCH_WATERMARK_MASK) < j->watermark) { - /* - * Don't want to close current journal entry, just need to - * invoke reclaim: - */ - ret = JOURNAL_ERR_journal_full; + ret = 0; goto unlock; } @@ -515,30 +552,30 @@ retry: j->buf_size_want = max(j->buf_size_want, buf->buf_size << 1); __journal_entry_close(j, JOURNAL_ENTRY_CLOSED_VAL, false); - ret = journal_entry_open(j); - - if (ret == JOURNAL_ERR_max_in_flight) { - track_event_change(&c->times[BCH_TIME_blocked_journal_max_in_flight], - &j->max_in_flight_start, true); - if (trace_journal_entry_full_enabled()) { - struct printbuf buf = PRINTBUF; - buf.atomic++; - - bch2_journal_bufs_to_text(&buf, j); - trace_journal_entry_full(c, buf.buf); - printbuf_exit(&buf); - } - count_event(c, journal_entry_full); - } + ret = journal_entry_open(j) ?: JOURNAL_ERR_retry; unlock: can_discard = j->can_discard; spin_unlock(&j->lock); - - if (!ret) +out: + if (ret == JOURNAL_ERR_retry) goto retry; + if (!ret) + return 0; + if (journal_error_check_stuck(j, ret, flags)) ret = -BCH_ERR_journal_res_get_blocked; + if (ret == JOURNAL_ERR_max_in_flight && + track_event_change(&c->times[BCH_TIME_blocked_journal_max_in_flight], true)) { + + struct printbuf buf = PRINTBUF; + prt_printf(&buf, "seq %llu\n", journal_cur_seq(j)); + bch2_journal_bufs_to_text(&buf, j); + trace_journal_entry_full(c, buf.buf); + printbuf_exit(&buf); + count_event(c, journal_entry_full); + } + /* * Journal is full - can't rely on reclaim from work item due to * freezing: @@ -674,7 +711,7 @@ recheck_need_open: return ret; seq = res.seq; - buf = j->buf + (seq & JOURNAL_BUF_MASK); + buf = journal_seq_to_buf(j, seq); buf->must_flush = true; if (!buf->flush_time) { @@ -692,8 +729,8 @@ recheck_need_open: } /* - * if write was kicked off without a flush, flush the next sequence - * number instead + * if write was kicked off without a flush, or if we promised it + * wouldn't be a flush, flush the next sequence number instead */ buf = journal_seq_to_buf(j, seq); if (buf->noflush) { @@ -771,8 +808,8 @@ bool bch2_journal_noflush_seq(struct journal *j, u64 seq) unwritten_seq++) { struct journal_buf *buf = journal_seq_to_buf(j, unwritten_seq); - /* journal write is already in flight, and was a flush write: */ - if (unwritten_seq == journal_last_unwritten_seq(j) && !buf->noflush) + /* journal flush already in flight, or flush requseted */ + if (buf->must_flush) goto out; buf->noflush = true; @@ -1157,13 +1194,12 @@ int bch2_fs_journal_start(struct journal *j, u64 cur_seq) struct journal_replay *i, **_i; struct genradix_iter iter; bool had_entries = false; - unsigned ptr; u64 last_seq = cur_seq, nr, seq; genradix_for_each_reverse(&c->journal_entries, iter, _i) { i = *_i; - if (!i || i->ignore) + if (journal_replay_ignore(i)) continue; last_seq = le64_to_cpu(i->j.last_seq); @@ -1196,7 +1232,7 @@ int bch2_fs_journal_start(struct journal *j, u64 cur_seq) genradix_for_each(&c->journal_entries, iter, _i) { i = *_i; - if (!i || i->ignore) + if (journal_replay_ignore(i)) continue; seq = le64_to_cpu(i->j.seq); @@ -1211,8 +1247,8 @@ int bch2_fs_journal_start(struct journal *j, u64 cur_seq) p = journal_seq_pin(j, seq); p->devs.nr = 0; - for (ptr = 0; ptr < i->nr_ptrs; ptr++) - bch2_dev_list_add_dev(&p->devs, i->ptrs[ptr].dev); + darray_for_each(i->ptrs, ptr) + bch2_dev_list_add_dev(&p->devs, ptr->dev); had_entries = true; } @@ -1240,13 +1276,17 @@ int bch2_fs_journal_start(struct journal *j, u64 cur_seq) void bch2_dev_journal_exit(struct bch_dev *ca) { - kfree(ca->journal.bio); - kfree(ca->journal.buckets); - kfree(ca->journal.bucket_seq); + struct journal_device *ja = &ca->journal; + + for (unsigned i = 0; i < ARRAY_SIZE(ja->bio); i++) { + kfree(ja->bio[i]); + ja->bio[i] = NULL; + } - ca->journal.bio = NULL; - ca->journal.buckets = NULL; - ca->journal.bucket_seq = NULL; + kfree(ja->buckets); + kfree(ja->bucket_seq); + ja->buckets = NULL; + ja->bucket_seq = NULL; } int bch2_dev_journal_init(struct bch_dev *ca, struct bch_sb *sb) @@ -1256,14 +1296,13 @@ int bch2_dev_journal_init(struct bch_dev *ca, struct bch_sb *sb) bch2_sb_field_get(sb, journal); struct bch_sb_field_journal_v2 *journal_buckets_v2 = bch2_sb_field_get(sb, journal_v2); - unsigned i, nr_bvecs; ja->nr = 0; if (journal_buckets_v2) { unsigned nr = bch2_sb_field_journal_v2_nr_entries(journal_buckets_v2); - for (i = 0; i < nr; i++) + for (unsigned i = 0; i < nr; i++) ja->nr += le64_to_cpu(journal_buckets_v2->d[i].nr); } else if (journal_buckets) { ja->nr = bch2_nr_journal_buckets(journal_buckets); @@ -1273,13 +1312,18 @@ int bch2_dev_journal_init(struct bch_dev *ca, struct bch_sb *sb) if (!ja->bucket_seq) return -BCH_ERR_ENOMEM_dev_journal_init; - nr_bvecs = DIV_ROUND_UP(JOURNAL_ENTRY_SIZE_MAX, PAGE_SIZE); + unsigned nr_bvecs = DIV_ROUND_UP(JOURNAL_ENTRY_SIZE_MAX, PAGE_SIZE); - ca->journal.bio = bio_kmalloc(nr_bvecs, GFP_KERNEL); - if (!ca->journal.bio) - return -BCH_ERR_ENOMEM_dev_journal_init; + for (unsigned i = 0; i < ARRAY_SIZE(ja->bio); i++) { + ja->bio[i] = kmalloc(struct_size(ja->bio[i], bio.bi_inline_vecs, + nr_bvecs), GFP_KERNEL); + if (!ja->bio[i]) + return -BCH_ERR_ENOMEM_dev_journal_init; - bio_init(ca->journal.bio, NULL, ca->journal.bio->bi_inline_vecs, nr_bvecs, 0); + ja->bio[i]->ca = ca; + ja->bio[i]->buf_idx = i; + bio_init(&ja->bio[i]->bio, NULL, ja->bio[i]->bio.bi_inline_vecs, nr_bvecs, 0); + } ja->buckets = kcalloc(ja->nr, sizeof(u64), GFP_KERNEL); if (!ja->buckets) @@ -1287,14 +1331,14 @@ int bch2_dev_journal_init(struct bch_dev *ca, struct bch_sb *sb) if (journal_buckets_v2) { unsigned nr = bch2_sb_field_journal_v2_nr_entries(journal_buckets_v2); - unsigned j, dst = 0; + unsigned dst = 0; - for (i = 0; i < nr; i++) - for (j = 0; j < le64_to_cpu(journal_buckets_v2->d[i].nr); j++) + for (unsigned i = 0; i < nr; i++) + for (unsigned j = 0; j < le64_to_cpu(journal_buckets_v2->d[i].nr); j++) ja->buckets[dst++] = le64_to_cpu(journal_buckets_v2->d[i].start) + j; } else if (journal_buckets) { - for (i = 0; i < ja->nr; i++) + for (unsigned i = 0; i < ja->nr; i++) ja->buckets[i] = le64_to_cpu(journal_buckets->buckets[i]); } @@ -1303,19 +1347,19 @@ int bch2_dev_journal_init(struct bch_dev *ca, struct bch_sb *sb) void bch2_fs_journal_exit(struct journal *j) { - unsigned i; + if (j->wq) + destroy_workqueue(j->wq); darray_exit(&j->early_journal_entries); - for (i = 0; i < ARRAY_SIZE(j->buf); i++) - kvpfree(j->buf[i].data, j->buf[i].buf_size); + for (unsigned i = 0; i < ARRAY_SIZE(j->buf); i++) + kvfree(j->buf[i].data); free_fifo(&j->pin); } int bch2_fs_journal_init(struct journal *j) { static struct lock_class_key res_key; - unsigned i; mutex_init(&j->buf_lock); spin_lock_init(&j->lock); @@ -1336,14 +1380,20 @@ int bch2_fs_journal_init(struct journal *j) if (!(init_fifo(&j->pin, JOURNAL_PIN, GFP_KERNEL))) return -BCH_ERR_ENOMEM_journal_pin_fifo; - for (i = 0; i < ARRAY_SIZE(j->buf); i++) { + for (unsigned i = 0; i < ARRAY_SIZE(j->buf); i++) { j->buf[i].buf_size = JOURNAL_ENTRY_SIZE_MIN; - j->buf[i].data = kvpmalloc(j->buf[i].buf_size, GFP_KERNEL); + j->buf[i].data = kvmalloc(j->buf[i].buf_size, GFP_KERNEL); if (!j->buf[i].data) return -BCH_ERR_ENOMEM_journal_buf; + j->buf[i].idx = i; } j->pin.front = j->pin.back = 1; + + j->wq = alloc_workqueue("bcachefs_journal", + WQ_HIGHPRI|WQ_FREEZABLE|WQ_UNBOUND|WQ_MEM_RECLAIM, 512); + if (!j->wq) + return -BCH_ERR_ENOMEM_fs_other_alloc; return 0; } @@ -1381,6 +1431,7 @@ void __bch2_journal_debug_to_text(struct printbuf *out, struct journal *j) prt_printf(out, "reclaim kicked:\t\t%u\n", j->reclaim_kicked); prt_printf(out, "reclaim runs in:\t%u ms\n", time_after(j->next_reclaim, now) ? jiffies_to_msecs(j->next_reclaim - jiffies) : 0); + prt_printf(out, "blocked:\t\t%u\n", j->blocked); prt_printf(out, "current entry sectors:\t%u\n", j->cur_entry_sectors); prt_printf(out, "current entry error:\t%s\n", bch2_journal_errors[j->cur_entry_error]); prt_printf(out, "current entry:\t\t"); @@ -1455,7 +1506,6 @@ bool bch2_journal_seq_pins_to_text(struct printbuf *out, struct journal *j, u64 { struct journal_entry_pin_list *pin_list; struct journal_entry_pin *pin; - unsigned i; spin_lock(&j->lock); *seq = max(*seq, j->pin.front); @@ -1473,7 +1523,7 @@ bool bch2_journal_seq_pins_to_text(struct printbuf *out, struct journal *j, u64 prt_newline(out); printbuf_indent_add(out, 2); - for (i = 0; i < ARRAY_SIZE(pin_list->list); i++) + for (unsigned i = 0; i < ARRAY_SIZE(pin_list->list); i++) list_for_each_entry(pin, &pin_list->list[i], list) { prt_printf(out, "\t%px %ps", pin, pin->flush); prt_newline(out); diff --git a/fs/bcachefs/journal.h b/fs/bcachefs/journal.h index 4544ce24bb8a..7c7528f839c5 100644 --- a/fs/bcachefs/journal.h +++ b/fs/bcachefs/journal.h @@ -264,7 +264,8 @@ static inline union journal_res_state journal_state_buf_put(struct journal *j, u } bool bch2_journal_entry_close(struct journal *); -void bch2_journal_buf_put_final(struct journal *, u64, bool); +void bch2_journal_do_writes(struct journal *); +void bch2_journal_buf_put_final(struct journal *, u64); static inline void __bch2_journal_buf_put(struct journal *j, unsigned idx, u64 seq) { @@ -272,7 +273,7 @@ static inline void __bch2_journal_buf_put(struct journal *j, unsigned idx, u64 s s = journal_state_buf_put(j, idx); if (!journal_state_count(s, idx)) - bch2_journal_buf_put_final(j, seq, idx == s.unwritten_idx); + bch2_journal_buf_put_final(j, seq); } static inline void bch2_journal_buf_put(struct journal *j, unsigned idx, u64 seq) @@ -282,7 +283,7 @@ static inline void bch2_journal_buf_put(struct journal *j, unsigned idx, u64 seq s = journal_state_buf_put(j, idx); if (!journal_state_count(s, idx)) { spin_lock(&j->lock); - bch2_journal_buf_put_final(j, seq, idx == s.unwritten_idx); + bch2_journal_buf_put_final(j, seq); spin_unlock(&j->lock); } } diff --git a/fs/bcachefs/journal_io.c b/fs/bcachefs/journal_io.c index 47805193f18c..725fcf46f631 100644 --- a/fs/bcachefs/journal_io.c +++ b/fs/bcachefs/journal_io.c @@ -17,6 +17,37 @@ #include "sb-clean.h" #include "trace.h" +void bch2_journal_ptrs_to_text(struct printbuf *out, struct bch_fs *c, + struct journal_replay *j) +{ + darray_for_each(j->ptrs, i) { + struct bch_dev *ca = bch_dev_bkey_exists(c, i->dev); + u64 offset; + + div64_u64_rem(i->sector, ca->mi.bucket_size, &offset); + + if (i != j->ptrs.data) + prt_printf(out, " "); + prt_printf(out, "%u:%u:%u (sector %llu)", + i->dev, i->bucket, i->bucket_offset, i->sector); + } +} + +static void bch2_journal_replay_to_text(struct printbuf *out, struct bch_fs *c, + struct journal_replay *j) +{ + prt_printf(out, "seq %llu ", le64_to_cpu(j->j.seq)); + + bch2_journal_ptrs_to_text(out, c, j); + + for_each_jset_entry_type(entry, &j->j, BCH_JSET_ENTRY_datetime) { + struct jset_entry_datetime *datetime = + container_of(entry, struct jset_entry_datetime, entry); + bch2_prt_datetime(out, le64_to_cpu(datetime->seconds)); + break; + } +} + static struct nonce journal_nonce(const struct jset *jset) { return (struct nonce) {{ @@ -52,13 +83,15 @@ static void __journal_replay_free(struct bch_fs *c, BUG_ON(*p != i); *p = NULL; - kvpfree(i, offsetof(struct journal_replay, j) + - vstruct_bytes(&i->j)); + kvfree(i); } -static void journal_replay_free(struct bch_fs *c, struct journal_replay *i) +static void journal_replay_free(struct bch_fs *c, struct journal_replay *i, bool blacklisted) { - i->ignore = true; + if (blacklisted) + i->ignore_blacklisted = true; + else + i->ignore_not_dirty = true; if (!c->opts.read_entire_journal) __journal_replay_free(c, i); @@ -84,9 +117,9 @@ static int journal_entry_add(struct bch_fs *c, struct bch_dev *ca, { struct genradix_iter iter; struct journal_replay **_i, *i, *dup; - struct journal_ptr *ptr; size_t bytes = vstruct_bytes(j); u64 last_seq = !JSET_NO_FLUSH(j) ? le64_to_cpu(j->last_seq) : 0; + struct printbuf buf = PRINTBUF; int ret = JOURNAL_ENTRY_ADD_OK; /* Is this entry older than the range we need? */ @@ -108,12 +141,13 @@ static int journal_entry_add(struct bch_fs *c, struct bch_dev *ca, journal_entry_radix_idx(c, jlist->last_seq)) { i = *_i; - if (!i || i->ignore) + if (journal_replay_ignore(i)) continue; if (le64_to_cpu(i->j.seq) >= last_seq) break; - journal_replay_free(c, i); + + journal_replay_free(c, i, false); } } @@ -131,72 +165,62 @@ static int journal_entry_add(struct bch_fs *c, struct bch_dev *ca, */ dup = *_i; if (dup) { - if (bytes == vstruct_bytes(&dup->j) && - !memcmp(j, &dup->j, bytes)) { - i = dup; - goto found; - } + bool identical = bytes == vstruct_bytes(&dup->j) && + !memcmp(j, &dup->j, bytes); + bool not_identical = !identical && + entry_ptr.csum_good && + dup->csum_good; + + bool same_device = false; + darray_for_each(dup->ptrs, ptr) + if (ptr->dev == ca->dev_idx) + same_device = true; + + ret = darray_push(&dup->ptrs, entry_ptr); + if (ret) + goto out; - if (!entry_ptr.csum_good) { - i = dup; - goto found; - } + bch2_journal_replay_to_text(&buf, c, dup); - if (!dup->csum_good) + fsck_err_on(same_device, + c, journal_entry_dup_same_device, + "duplicate journal entry on same device\n %s", + buf.buf); + + fsck_err_on(not_identical, + c, journal_entry_replicas_data_mismatch, + "found duplicate but non identical journal entries\n %s", + buf.buf); + + if (entry_ptr.csum_good && !identical) goto replace; - fsck_err(c, journal_entry_replicas_data_mismatch, - "found duplicate but non identical journal entries (seq %llu)", - le64_to_cpu(j->seq)); - i = dup; - goto found; + goto out; } replace: - i = kvpmalloc(offsetof(struct journal_replay, j) + bytes, GFP_KERNEL); + i = kvmalloc(offsetof(struct journal_replay, j) + bytes, GFP_KERNEL); if (!i) return -BCH_ERR_ENOMEM_journal_entry_add; - i->nr_ptrs = 0; - i->csum_good = entry_ptr.csum_good; - i->ignore = false; + darray_init(&i->ptrs); + i->csum_good = entry_ptr.csum_good; + i->ignore_blacklisted = false; + i->ignore_not_dirty = false; unsafe_memcpy(&i->j, j, bytes, "embedded variable length struct"); - i->ptrs[i->nr_ptrs++] = entry_ptr; if (dup) { - if (dup->nr_ptrs >= ARRAY_SIZE(dup->ptrs)) { - bch_err(c, "found too many copies of journal entry %llu", - le64_to_cpu(i->j.seq)); - dup->nr_ptrs = ARRAY_SIZE(dup->ptrs) - 1; - } - /* The first ptr should represent the jset we kept: */ - memcpy(i->ptrs + i->nr_ptrs, - dup->ptrs, - sizeof(dup->ptrs[0]) * dup->nr_ptrs); - i->nr_ptrs += dup->nr_ptrs; + darray_for_each(dup->ptrs, ptr) + darray_push(&i->ptrs, *ptr); __journal_replay_free(c, dup); + } else { + darray_push(&i->ptrs, entry_ptr); } *_i = i; - return 0; -found: - for (ptr = i->ptrs; ptr < i->ptrs + i->nr_ptrs; ptr++) { - if (ptr->dev == ca->dev_idx) { - bch_err(c, "duplicate journal entry %llu on same device", - le64_to_cpu(i->j.seq)); - goto out; - } - } - - if (i->nr_ptrs >= ARRAY_SIZE(i->ptrs)) { - bch_err(c, "found too many copies of journal entry %llu", - le64_to_cpu(i->j.seq)); - goto out; - } - - i->ptrs[i->nr_ptrs++] = entry_ptr; out: fsck_err: + printbuf_exit(&buf); return ret; } @@ -374,7 +398,6 @@ static int journal_entry_btree_keys_validate(struct bch_fs *c, static void journal_entry_btree_keys_to_text(struct printbuf *out, struct bch_fs *c, struct jset_entry *entry) { - struct bkey_i *k; bool first = true; jset_entry_for_each_key(entry, k) { @@ -741,6 +764,37 @@ static void journal_entry_write_buffer_keys_to_text(struct printbuf *out, struct journal_entry_btree_keys_to_text(out, c, entry); } +static int journal_entry_datetime_validate(struct bch_fs *c, + struct jset *jset, + struct jset_entry *entry, + unsigned version, int big_endian, + enum bkey_invalid_flags flags) +{ + unsigned bytes = vstruct_bytes(entry); + unsigned expected = 16; + int ret = 0; + + if (journal_entry_err_on(vstruct_bytes(entry) < expected, + c, version, jset, entry, + journal_entry_dev_usage_bad_size, + "bad size (%u < %u)", + bytes, expected)) { + journal_entry_null_range(entry, vstruct_next(entry)); + return ret; + } +fsck_err: + return ret; +} + +static void journal_entry_datetime_to_text(struct printbuf *out, struct bch_fs *c, + struct jset_entry *entry) +{ + struct jset_entry_datetime *datetime = + container_of(entry, struct jset_entry_datetime, entry); + + bch2_prt_datetime(out, le64_to_cpu(datetime->seconds)); +} + struct jset_entry_ops { int (*validate)(struct bch_fs *, struct jset *, struct jset_entry *, unsigned, int, @@ -913,11 +967,11 @@ static int journal_read_buf_realloc(struct journal_read_buf *b, return -BCH_ERR_ENOMEM_journal_read_buf_realloc; new_size = roundup_pow_of_two(new_size); - n = kvpmalloc(new_size, GFP_KERNEL); + n = kvmalloc(new_size, GFP_KERNEL); if (!n) return -BCH_ERR_ENOMEM_journal_read_buf_realloc; - kvpfree(b->data, b->size); + kvfree(b->data); b->data = n; b->size = new_size; return 0; @@ -1028,9 +1082,7 @@ reread: ret = bch2_encrypt(c, JSET_CSUM_TYPE(j), journal_nonce(j), j->encrypted_start, vstruct_end(j) - (void *) j->encrypted_start); - bch2_fs_fatal_err_on(ret, c, - "error decrypting journal entry: %s", - bch2_err_str(ret)); + bch2_fs_fatal_err_on(ret, c, "decrypting journal entry: %s", bch2_err_str(ret)); mutex_lock(&jlist->lock); ret = journal_entry_add(c, ca, (struct journal_ptr) { @@ -1102,16 +1154,15 @@ static CLOSURE_CALLBACK(bch2_journal_read_device) if (!r) continue; - for (i = 0; i < r->nr_ptrs; i++) { - if (r->ptrs[i].dev == ca->dev_idx) { - unsigned wrote = bucket_remainder(ca, r->ptrs[i].sector) + + darray_for_each(r->ptrs, i) + if (i->dev == ca->dev_idx) { + unsigned wrote = bucket_remainder(ca, i->sector) + vstruct_sectors(&r->j, c->block_bits); - ja->cur_idx = r->ptrs[i].bucket; + ja->cur_idx = i->bucket; ja->sectors_free = ca->mi.bucket_size - wrote; goto found; } - } } found: mutex_unlock(&jlist->lock); @@ -1144,7 +1195,7 @@ found: ja->dirty_idx = (ja->cur_idx + 1) % ja->nr; out: bch_verbose(c, "journal read done on device %s, ret %i", ca->name, ret); - kvpfree(buf.data, buf.size); + kvfree(buf.data); percpu_ref_put(&ca->io_ref); closure_return(cl); return; @@ -1155,27 +1206,6 @@ err: goto out; } -void bch2_journal_ptrs_to_text(struct printbuf *out, struct bch_fs *c, - struct journal_replay *j) -{ - unsigned i; - - for (i = 0; i < j->nr_ptrs; i++) { - struct bch_dev *ca = bch_dev_bkey_exists(c, j->ptrs[i].dev); - u64 offset; - - div64_u64_rem(j->ptrs[i].sector, ca->mi.bucket_size, &offset); - - if (i) - prt_printf(out, " "); - prt_printf(out, "%u:%u:%u (sector %llu)", - j->ptrs[i].dev, - j->ptrs[i].bucket, - j->ptrs[i].bucket_offset, - j->ptrs[i].sector); - } -} - int bch2_journal_read(struct bch_fs *c, u64 *last_seq, u64 *blacklist_seq, @@ -1228,20 +1258,20 @@ int bch2_journal_read(struct bch_fs *c, i = *_i; - if (!i || i->ignore) + if (journal_replay_ignore(i)) continue; if (!*start_seq) *blacklist_seq = *start_seq = le64_to_cpu(i->j.seq) + 1; if (JSET_NO_FLUSH(&i->j)) { - i->ignore = true; + i->ignore_blacklisted = true; continue; } if (!last_write_torn && !i->csum_good) { last_write_torn = true; - i->ignore = true; + i->ignore_blacklisted = true; continue; } @@ -1280,12 +1310,12 @@ int bch2_journal_read(struct bch_fs *c, genradix_for_each(&c->journal_entries, radix_iter, _i) { i = *_i; - if (!i || i->ignore) + if (journal_replay_ignore(i)) continue; seq = le64_to_cpu(i->j.seq); if (seq < *last_seq) { - journal_replay_free(c, i); + journal_replay_free(c, i, false); continue; } @@ -1293,7 +1323,7 @@ int bch2_journal_read(struct bch_fs *c, fsck_err_on(!JSET_NO_FLUSH(&i->j), c, jset_seq_blacklisted, "found blacklisted journal entry %llu", seq); - i->ignore = true; + i->ignore_blacklisted = true; } } @@ -1302,7 +1332,7 @@ int bch2_journal_read(struct bch_fs *c, genradix_for_each(&c->journal_entries, radix_iter, _i) { i = *_i; - if (!i || i->ignore) + if (journal_replay_ignore(i)) continue; BUG_ON(seq > le64_to_cpu(i->j.seq)); @@ -1353,32 +1383,31 @@ int bch2_journal_read(struct bch_fs *c, .e.data_type = BCH_DATA_journal, .e.nr_required = 1, }; - unsigned ptr; i = *_i; - if (!i || i->ignore) + if (journal_replay_ignore(i)) continue; - for (ptr = 0; ptr < i->nr_ptrs; ptr++) { - struct bch_dev *ca = bch_dev_bkey_exists(c, i->ptrs[ptr].dev); + darray_for_each(i->ptrs, ptr) { + struct bch_dev *ca = bch_dev_bkey_exists(c, ptr->dev); - if (!i->ptrs[ptr].csum_good) - bch_err_dev_offset(ca, i->ptrs[ptr].sector, + if (!ptr->csum_good) + bch_err_dev_offset(ca, ptr->sector, "invalid journal checksum, seq %llu%s", le64_to_cpu(i->j.seq), i->csum_good ? " (had good copy on another device)" : ""); } ret = jset_validate(c, - bch_dev_bkey_exists(c, i->ptrs[0].dev), + bch_dev_bkey_exists(c, i->ptrs.data[0].dev), &i->j, - i->ptrs[0].sector, + i->ptrs.data[0].sector, READ); if (ret) goto err; - for (ptr = 0; ptr < i->nr_ptrs; ptr++) - replicas.e.devs[replicas.e.nr_devs++] = i->ptrs[ptr].dev; + darray_for_each(i->ptrs, ptr) + replicas.e.devs[replicas.e.nr_devs++] = ptr->dev; bch2_replicas_entry_sort(&replicas.e); @@ -1547,7 +1576,7 @@ static void journal_buf_realloc(struct journal *j, struct journal_buf *buf) if (bch2_btree_write_buffer_resize(c, btree_write_buffer_size)) return; - new_buf = kvpmalloc(new_size, GFP_NOFS|__GFP_NOWARN); + new_buf = kvmalloc(new_size, GFP_NOFS|__GFP_NOWARN); if (!new_buf) return; @@ -1558,7 +1587,7 @@ static void journal_buf_realloc(struct journal *j, struct journal_buf *buf) swap(buf->buf_size, new_size); spin_unlock(&j->lock); - kvpfree(new_buf, new_size); + kvfree(new_buf); } static inline struct journal_buf *journal_last_unwritten_buf(struct journal *j) @@ -1568,12 +1597,12 @@ static inline struct journal_buf *journal_last_unwritten_buf(struct journal *j) static CLOSURE_CALLBACK(journal_write_done) { - closure_type(j, struct journal, io); + closure_type(w, struct journal_buf, io); + struct journal *j = container_of(w, struct journal, buf[w->idx]); struct bch_fs *c = container_of(j, struct bch_fs, journal); - struct journal_buf *w = journal_last_unwritten_buf(j); struct bch_replicas_padded replicas; union journal_res_state old, new; - u64 v, seq; + u64 v, seq = le64_to_cpu(w->data->seq); int err = 0; bch2_time_stats_update(!JSET_NO_FLUSH(w->data) @@ -1593,63 +1622,68 @@ static CLOSURE_CALLBACK(journal_write_done) if (err) bch2_fatal_error(c); - spin_lock(&j->lock); - seq = le64_to_cpu(w->data->seq); + closure_debug_destroy(cl); + spin_lock(&j->lock); if (seq >= j->pin.front) journal_seq_pin(j, seq)->devs = w->devs_written; + if (err && (!j->err_seq || seq < j->err_seq)) + j->err_seq = seq; + w->write_done = true; + + bool completed = false; - if (!err) { - if (!JSET_NO_FLUSH(w->data)) { + for (seq = journal_last_unwritten_seq(j); + seq <= journal_cur_seq(j); + seq++) { + w = j->buf + (seq & JOURNAL_BUF_MASK); + if (!w->write_done) + break; + + if (!j->err_seq && !JSET_NO_FLUSH(w->data)) { j->flushed_seq_ondisk = seq; j->last_seq_ondisk = w->last_seq; bch2_do_discards(c); closure_wake_up(&c->freelist_wait); - bch2_reset_alloc_cursors(c); } - } else if (!j->err_seq || seq < j->err_seq) - j->err_seq = seq; - j->seq_ondisk = seq; + j->seq_ondisk = seq; - /* - * Updating last_seq_ondisk may let bch2_journal_reclaim_work() discard - * more buckets: - * - * Must come before signaling write completion, for - * bch2_fs_journal_stop(): - */ - if (j->watermark != BCH_WATERMARK_stripe) - journal_reclaim_kick(&c->journal); + /* + * Updating last_seq_ondisk may let bch2_journal_reclaim_work() discard + * more buckets: + * + * Must come before signaling write completion, for + * bch2_fs_journal_stop(): + */ + if (j->watermark != BCH_WATERMARK_stripe) + journal_reclaim_kick(&c->journal); - /* also must come before signalling write completion: */ - closure_debug_destroy(cl); + v = atomic64_read(&j->reservations.counter); + do { + old.v = new.v = v; + BUG_ON(journal_state_count(new, new.unwritten_idx)); + BUG_ON(new.unwritten_idx != (seq & JOURNAL_BUF_MASK)); - v = atomic64_read(&j->reservations.counter); - do { - old.v = new.v = v; - BUG_ON(journal_state_count(new, new.unwritten_idx)); + new.unwritten_idx++; + } while ((v = atomic64_cmpxchg(&j->reservations.counter, old.v, new.v)) != old.v); - new.unwritten_idx++; - } while ((v = atomic64_cmpxchg(&j->reservations.counter, - old.v, new.v)) != old.v); + closure_wake_up(&w->wait); + completed = true; + } - bch2_journal_reclaim_fast(j); - bch2_journal_space_available(j); + if (completed) { + bch2_journal_reclaim_fast(j); + bch2_journal_space_available(j); - track_event_change(&c->times[BCH_TIME_blocked_journal_max_in_flight], - &j->max_in_flight_start, false); + track_event_change(&c->times[BCH_TIME_blocked_journal_max_in_flight], false); - closure_wake_up(&w->wait); - journal_wake(j); + journal_wake(j); + } - if (!journal_state_count(new, new.unwritten_idx) && - journal_last_unwritten_seq(j) <= journal_cur_seq(j)) { - spin_unlock(&j->lock); - closure_call(&j->io, bch2_journal_write, c->io_complete_wq, NULL); - } else if (journal_last_unwritten_seq(j) == journal_cur_seq(j) && + if (journal_last_unwritten_seq(j) == journal_cur_seq(j) && new.cur_entry_offset < JOURNAL_ENTRY_CLOSED_VAL) { struct journal_buf *buf = journal_cur_buf(j); long delta = buf->expires - jiffies; @@ -1659,46 +1693,46 @@ static CLOSURE_CALLBACK(journal_write_done) * previous entries still in flight - the current journal entry * might want to be written now: */ - - spin_unlock(&j->lock); - mod_delayed_work(c->io_complete_wq, &j->write_work, max(0L, delta)); - } else { - spin_unlock(&j->lock); + mod_delayed_work(j->wq, &j->write_work, max(0L, delta)); } + + spin_unlock(&j->lock); } static void journal_write_endio(struct bio *bio) { - struct bch_dev *ca = bio->bi_private; + struct journal_bio *jbio = container_of(bio, struct journal_bio, bio); + struct bch_dev *ca = jbio->ca; struct journal *j = &ca->fs->journal; - struct journal_buf *w = journal_last_unwritten_buf(j); - unsigned long flags; + struct journal_buf *w = j->buf + jbio->buf_idx; if (bch2_dev_io_err_on(bio->bi_status, ca, BCH_MEMBER_ERROR_write, "error writing journal entry %llu: %s", le64_to_cpu(w->data->seq), bch2_blk_status_to_str(bio->bi_status)) || bch2_meta_write_fault("journal")) { + unsigned long flags; + spin_lock_irqsave(&j->err_lock, flags); bch2_dev_list_drop_dev(&w->devs_written, ca->dev_idx); spin_unlock_irqrestore(&j->err_lock, flags); } - closure_put(&j->io); + closure_put(&w->io); percpu_ref_put(&ca->io_ref); } static CLOSURE_CALLBACK(do_journal_write) { - closure_type(j, struct journal, io); + closure_type(w, struct journal_buf, io); + struct journal *j = container_of(w, struct journal, buf[w->idx]); struct bch_fs *c = container_of(j, struct bch_fs, journal); - struct bch_dev *ca; - struct journal_buf *w = journal_last_unwritten_buf(j); - struct bio *bio; unsigned sectors = vstruct_sectors(w->data, c->block_bits); extent_for_each_ptr(bkey_i_to_s_extent(&w->key), ptr) { - ca = bch_dev_bkey_exists(c, ptr->dev); + struct bch_dev *ca = bch_dev_bkey_exists(c, ptr->dev); + struct journal_device *ja = &ca->journal; + if (!percpu_ref_tryget(&ca->io_ref)) { /* XXX: fix this */ bch_err(c, "missing device for journal write\n"); @@ -1708,7 +1742,7 @@ static CLOSURE_CALLBACK(do_journal_write) this_cpu_add(ca->io_done->sectors[WRITE][BCH_DATA_journal], sectors); - bio = ca->journal.bio; + struct bio *bio = &ja->bio[w->idx]->bio; bio_reset(bio, ca->disk_sb.bdev, REQ_OP_WRITE|REQ_SYNC|REQ_META); bio->bi_iter.bi_sector = ptr->offset; bio->bi_end_io = journal_write_endio; @@ -1727,11 +1761,10 @@ static CLOSURE_CALLBACK(do_journal_write) trace_and_count(c, journal_write, bio); closure_bio_submit(bio, cl); - ca->journal.bucket_seq[ca->journal.cur_idx] = - le64_to_cpu(w->data->seq); + ja->bucket_seq[ja->cur_idx] = le64_to_cpu(w->data->seq); } - continue_at(cl, journal_write_done, c->io_complete_wq); + continue_at(cl, journal_write_done, j->wq); } static int bch2_journal_write_prep(struct journal *j, struct journal_buf *w) @@ -1782,11 +1815,11 @@ static int bch2_journal_write_prep(struct journal *j, struct journal_buf *w) if (!wb.wb) bch2_journal_keys_to_write_buffer_start(c, &wb, seq); - struct bkey_i *k; jset_entry_for_each_key(i, k) { ret = bch2_journal_key_to_wb(c, &wb, i->btree_id, k); if (ret) { - bch2_fs_fatal_error(c, "-ENOMEM flushing journal keys to btree write buffer"); + bch2_fs_fatal_error(c, "flushing journal keys to btree write buffer: %s", + bch2_err_str(ret)); bch2_journal_keys_to_write_buffer_end(c, &wb); return ret; } @@ -1798,15 +1831,24 @@ static int bch2_journal_write_prep(struct journal *j, struct journal_buf *w) if (wb.wb) bch2_journal_keys_to_write_buffer_end(c, &wb); + + spin_lock(&c->journal.lock); w->need_flush_to_write_buffer = false; + spin_unlock(&c->journal.lock); start = end = vstruct_last(jset); end = bch2_btree_roots_to_journal_entries(c, end, btree_roots_have); + struct jset_entry_datetime *d = + container_of(jset_entry_init(&end, sizeof(*d)), struct jset_entry_datetime, entry); + d->entry.type = BCH_JSET_ENTRY_datetime; + d->seconds = cpu_to_le64(ktime_get_real_seconds()); + bch2_journal_super_entries_add_common(c, &end, seq); u64s = (u64 *) end - (u64 *) start; - BUG_ON(u64s > j->entry_u64s_reserved); + + WARN_ON(u64s > j->entry_u64s_reserved); le32_add_cpu(&jset->u64s, u64s); @@ -1814,7 +1856,7 @@ static int bch2_journal_write_prep(struct journal *j, struct journal_buf *w) bytes = vstruct_bytes(jset); if (sectors > w->sectors) { - bch2_fs_fatal_error(c, "aieeee! journal write overran available space, %zu > %u (extra %u reserved %u/%u)", + bch2_fs_fatal_error(c, ": journal write overran available space, %zu > %u (extra %u reserved %u/%u)", vstruct_bytes(jset), w->sectors << 9, u64s, w->u64s_reserved, j->entry_u64s_reserved); return -EINVAL; @@ -1842,8 +1884,7 @@ static int bch2_journal_write_prep(struct journal *j, struct journal_buf *w) ret = bch2_encrypt(c, JSET_CSUM_TYPE(jset), journal_nonce(jset), jset->encrypted_start, vstruct_end(jset) - (void *) jset->encrypted_start); - if (bch2_fs_fatal_err_on(ret, c, - "error decrypting journal entry: %i", ret)) + if (bch2_fs_fatal_err_on(ret, c, "decrypting journal entry: %s", bch2_err_str(ret))) return ret; jset->csum = csum_vstruct(c, JSET_CSUM_TYPE(jset), @@ -1893,6 +1934,7 @@ static int bch2_journal_write_pick_flush(struct journal *j, struct journal_buf * j->nr_noflush_writes++; } else { + w->must_flush = true; j->last_flush_write = jiffies; j->nr_flush_writes++; clear_bit(JOURNAL_NEED_FLUSH_WRITE, &j->flags); @@ -1903,20 +1945,28 @@ static int bch2_journal_write_pick_flush(struct journal *j, struct journal_buf * CLOSURE_CALLBACK(bch2_journal_write) { - closure_type(j, struct journal, io); + closure_type(w, struct journal_buf, io); + struct journal *j = container_of(w, struct journal, buf[w->idx]); struct bch_fs *c = container_of(j, struct bch_fs, journal); - struct journal_buf *w = journal_last_unwritten_buf(j); struct bch_replicas_padded replicas; - struct bio *bio; struct printbuf journal_debug_buf = PRINTBUF; unsigned nr_rw_members = 0; int ret; + for_each_rw_member(c, ca) + nr_rw_members++; + BUG_ON(BCH_SB_CLEAN(c->disk_sb.sb)); + BUG_ON(!w->write_started); + BUG_ON(w->write_allocated); + BUG_ON(w->write_done); j->write_start_time = local_clock(); spin_lock(&j->lock); + if (nr_rw_members > 1) + w->separate_flush = true; + ret = bch2_journal_write_pick_flush(j, w); spin_unlock(&j->lock); if (ret) @@ -1956,12 +2006,14 @@ CLOSURE_CALLBACK(bch2_journal_write) * bch2_journal_space_available(): */ w->sectors = 0; + w->write_allocated = true; /* * journal entry has been compacted and allocated, recalculate space * available: */ bch2_journal_space_available(j); + bch2_journal_do_writes(j); spin_unlock(&j->lock); w->devs_written = bch2_bkey_devs(bkey_i_to_s_c(&w->key)); @@ -1969,12 +2021,6 @@ CLOSURE_CALLBACK(bch2_journal_write) if (c->opts.nochanges) goto no_io; - for_each_rw_member(c, ca) - nr_rw_members++; - - if (nr_rw_members > 1) - w->separate_flush = true; - /* * Mark journal replicas before we submit the write to guarantee * recovery will find the journal entries after a crash. @@ -1985,25 +2031,29 @@ CLOSURE_CALLBACK(bch2_journal_write) if (ret) goto err; + if (!JSET_NO_FLUSH(w->data)) + closure_wait_event(&j->async_wait, j->seq_ondisk + 1 == le64_to_cpu(w->data->seq)); + if (!JSET_NO_FLUSH(w->data) && w->separate_flush) { for_each_rw_member(c, ca) { percpu_ref_get(&ca->io_ref); - bio = ca->journal.bio; + struct journal_device *ja = &ca->journal; + struct bio *bio = &ja->bio[w->idx]->bio; bio_reset(bio, ca->disk_sb.bdev, - REQ_OP_WRITE|REQ_PREFLUSH); + REQ_OP_WRITE|REQ_SYNC|REQ_META|REQ_PREFLUSH); bio->bi_end_io = journal_write_endio; bio->bi_private = ca; closure_bio_submit(bio, cl); } } - continue_at(cl, do_journal_write, c->io_complete_wq); + continue_at(cl, do_journal_write, j->wq); return; no_io: - continue_at(cl, journal_write_done, c->io_complete_wq); + continue_at(cl, journal_write_done, j->wq); return; err: bch2_fatal_error(c); - continue_at(cl, journal_write_done, c->io_complete_wq); + continue_at(cl, journal_write_done, j->wq); } diff --git a/fs/bcachefs/journal_io.h b/fs/bcachefs/journal_io.h index c035e7c108e1..4f1e763ab506 100644 --- a/fs/bcachefs/journal_io.h +++ b/fs/bcachefs/journal_io.h @@ -2,26 +2,35 @@ #ifndef _BCACHEFS_JOURNAL_IO_H #define _BCACHEFS_JOURNAL_IO_H +#include "darray.h" + +struct journal_ptr { + bool csum_good; + u8 dev; + u32 bucket; + u32 bucket_offset; + u64 sector; +}; + /* * Only used for holding the journal entries we read in btree_journal_read() * during cache_registration */ struct journal_replay { - struct journal_ptr { - bool csum_good; - u8 dev; - u32 bucket; - u32 bucket_offset; - u64 sector; - } ptrs[BCH_REPLICAS_MAX]; - unsigned nr_ptrs; + DARRAY_PREALLOCATED(struct journal_ptr, 8) ptrs; bool csum_good; - bool ignore; + bool ignore_blacklisted; + bool ignore_not_dirty; /* must be last: */ struct jset j; }; +static inline bool journal_replay_ignore(struct journal_replay *i) +{ + return !i || i->ignore_blacklisted || i->ignore_not_dirty; +} + static inline struct jset_entry *__jset_entry_type_next(struct jset *jset, struct jset_entry *entry, unsigned type) { @@ -36,12 +45,12 @@ static inline struct jset_entry *__jset_entry_type_next(struct jset *jset, } #define for_each_jset_entry_type(entry, jset, type) \ - for (entry = (jset)->start; \ + for (struct jset_entry *entry = (jset)->start; \ (entry = __jset_entry_type_next(jset, entry, type)); \ entry = vstruct_next(entry)) #define jset_entry_for_each_key(_e, _k) \ - for (_k = (_e)->start; \ + for (struct bkey_i *_k = (_e)->start; \ _k < vstruct_last(_e); \ _k = bkey_next(_k)) @@ -62,4 +71,20 @@ int bch2_journal_read(struct bch_fs *, u64 *, u64 *, u64 *); CLOSURE_CALLBACK(bch2_journal_write); +static inline struct jset_entry *jset_entry_init(struct jset_entry **end, size_t size) +{ + struct jset_entry *entry = *end; + unsigned u64s = DIV_ROUND_UP(size, sizeof(u64)); + + memset(entry, 0, u64s * sizeof(u64)); + /* + * The u64s field counts from the start of data, ignoring the shared + * fields. + */ + entry->u64s = cpu_to_le16(u64s - 1); + + *end = vstruct_next(*end); + return entry; +} + #endif /* _BCACHEFS_JOURNAL_IO_H */ diff --git a/fs/bcachefs/journal_reclaim.c b/fs/bcachefs/journal_reclaim.c index c33dca641575..ab811c0dad26 100644 --- a/fs/bcachefs/journal_reclaim.c +++ b/fs/bcachefs/journal_reclaim.c @@ -62,12 +62,9 @@ void bch2_journal_set_watermark(struct journal *j) ? BCH_WATERMARK_reclaim : BCH_WATERMARK_stripe; - if (track_event_change(&c->times[BCH_TIME_blocked_journal_low_on_space], - &j->low_on_space_start, low_on_space) || - track_event_change(&c->times[BCH_TIME_blocked_journal_low_on_pin], - &j->low_on_pin_start, low_on_pin) || - track_event_change(&c->times[BCH_TIME_blocked_write_buffer_full], - &j->write_buffer_full_start, low_on_wb)) + if (track_event_change(&c->times[BCH_TIME_blocked_journal_low_on_space], low_on_space) || + track_event_change(&c->times[BCH_TIME_blocked_journal_low_on_pin], low_on_pin) || + track_event_change(&c->times[BCH_TIME_blocked_write_buffer_full], low_on_wb)) trace_and_count(c, journal_full, c); swap(watermark, j->watermark); @@ -394,8 +391,6 @@ void bch2_journal_pin_copy(struct journal *j, struct journal_entry_pin *src, journal_pin_flush_fn flush_fn) { - bool reclaim; - spin_lock(&j->lock); u64 seq = READ_ONCE(src->seq); @@ -411,44 +406,44 @@ void bch2_journal_pin_copy(struct journal *j, return; } - reclaim = __journal_pin_drop(j, dst); + bool reclaim = __journal_pin_drop(j, dst); bch2_journal_pin_set_locked(j, seq, dst, flush_fn, journal_pin_type(flush_fn)); if (reclaim) bch2_journal_reclaim_fast(j); - spin_unlock(&j->lock); /* * If the journal is currently full, we might want to call flush_fn * immediately: */ - journal_wake(j); + if (seq == journal_last_seq(j)) + journal_wake(j); + spin_unlock(&j->lock); } void bch2_journal_pin_set(struct journal *j, u64 seq, struct journal_entry_pin *pin, journal_pin_flush_fn flush_fn) { - bool reclaim; - spin_lock(&j->lock); BUG_ON(seq < journal_last_seq(j)); - reclaim = __journal_pin_drop(j, pin); + bool reclaim = __journal_pin_drop(j, pin); bch2_journal_pin_set_locked(j, seq, pin, flush_fn, journal_pin_type(flush_fn)); if (reclaim) bch2_journal_reclaim_fast(j); - spin_unlock(&j->lock); - /* * If the journal is currently full, we might want to call flush_fn * immediately: */ - journal_wake(j); + if (seq == journal_last_seq(j)) + journal_wake(j); + + spin_unlock(&j->lock); } /** diff --git a/fs/bcachefs/journal_seq_blacklist.c b/fs/bcachefs/journal_seq_blacklist.c index 0200e299cfbb..b5303874fc35 100644 --- a/fs/bcachefs/journal_seq_blacklist.c +++ b/fs/bcachefs/journal_seq_blacklist.c @@ -43,61 +43,36 @@ static unsigned sb_blacklist_u64s(unsigned nr) return (sizeof(*bl) + sizeof(bl->start[0]) * nr) / sizeof(u64); } -static struct bch_sb_field_journal_seq_blacklist * -blacklist_entry_try_merge(struct bch_fs *c, - struct bch_sb_field_journal_seq_blacklist *bl, - unsigned i) -{ - unsigned nr = blacklist_nr_entries(bl); - - if (le64_to_cpu(bl->start[i].end) >= - le64_to_cpu(bl->start[i + 1].start)) { - bl->start[i].end = bl->start[i + 1].end; - --nr; - memmove(&bl->start[i], - &bl->start[i + 1], - sizeof(bl->start[0]) * (nr - i)); - - bl = bch2_sb_field_resize(&c->disk_sb, journal_seq_blacklist, - sb_blacklist_u64s(nr)); - BUG_ON(!bl); - } - - return bl; -} - -static bool bl_entry_contig_or_overlaps(struct journal_seq_blacklist_entry *e, - u64 start, u64 end) -{ - return !(end < le64_to_cpu(e->start) || le64_to_cpu(e->end) < start); -} - int bch2_journal_seq_blacklist_add(struct bch_fs *c, u64 start, u64 end) { struct bch_sb_field_journal_seq_blacklist *bl; - unsigned i, nr; + unsigned i = 0, nr; int ret = 0; mutex_lock(&c->sb_lock); bl = bch2_sb_field_get(c->disk_sb.sb, journal_seq_blacklist); nr = blacklist_nr_entries(bl); - for (i = 0; i < nr; i++) { + while (i < nr) { struct journal_seq_blacklist_entry *e = bl->start + i; - if (bl_entry_contig_or_overlaps(e, start, end)) { - e->start = cpu_to_le64(min(start, le64_to_cpu(e->start))); - e->end = cpu_to_le64(max(end, le64_to_cpu(e->end))); - - if (i + 1 < nr) - bl = blacklist_entry_try_merge(c, - bl, i); - if (i) - bl = blacklist_entry_try_merge(c, - bl, i - 1); - goto out_write_sb; + if (end < le64_to_cpu(e->start)) + break; + + if (start > le64_to_cpu(e->end)) { + i++; + continue; } + + /* + * Entry is contiguous or overlapping with new entry: merge it + * with new entry, and delete: + */ + + start = min(start, le64_to_cpu(e->start)); + end = max(end, le64_to_cpu(e->end)); + array_remove_item(bl->start, nr, i); } bl = bch2_sb_field_resize(&c->disk_sb, journal_seq_blacklist, @@ -107,9 +82,10 @@ int bch2_journal_seq_blacklist_add(struct bch_fs *c, u64 start, u64 end) goto out; } - bl->start[nr].start = cpu_to_le64(start); - bl->start[nr].end = cpu_to_le64(end); -out_write_sb: + array_insert_item(bl->start, nr, i, ((struct journal_seq_blacklist_entry) { + .start = cpu_to_le64(start), + .end = cpu_to_le64(end), + })); c->disk_sb.sb->features[0] |= cpu_to_le64(1ULL << BCH_FEATURE_journal_seq_blacklist_v3); ret = bch2_write_super(c); @@ -165,8 +141,7 @@ int bch2_blacklist_table_initialize(struct bch_fs *c) if (!bl) return 0; - t = kzalloc(sizeof(*t) + sizeof(t->entries[0]) * nr, - GFP_KERNEL); + t = kzalloc(struct_size(t, entries, nr), GFP_KERNEL); if (!t) return -BCH_ERR_ENOMEM_blacklist_table_init; diff --git a/fs/bcachefs/journal_types.h b/fs/bcachefs/journal_types.h index 38817c7a0851..8c053cb64ca5 100644 --- a/fs/bcachefs/journal_types.h +++ b/fs/bcachefs/journal_types.h @@ -18,6 +18,7 @@ * the journal that are being staged or in flight. */ struct journal_buf { + struct closure io; struct jset *data; __BKEY_PADDED(key, BCH_REPLICAS_MAX); @@ -33,10 +34,14 @@ struct journal_buf { unsigned disk_sectors; /* maximum size entry could have been, if buf_size was bigger */ unsigned u64s_reserved; - bool noflush; /* write has already been kicked off, and was noflush */ - bool must_flush; /* something wants a flush */ - bool separate_flush; - bool need_flush_to_write_buffer; + bool noflush:1; /* write has already been kicked off, and was noflush */ + bool must_flush:1; /* something wants a flush */ + bool separate_flush:1; + bool need_flush_to_write_buffer:1; + bool write_started:1; + bool write_allocated:1; + bool write_done:1; + u8 idx; }; /* @@ -134,6 +139,7 @@ enum journal_flags { /* Reasons we may fail to get a journal reservation: */ #define JOURNAL_ERRORS() \ x(ok) \ + x(retry) \ x(blocked) \ x(max_in_flight) \ x(journal_full) \ @@ -149,6 +155,13 @@ enum journal_errors { typedef DARRAY(u64) darray_u64; +struct journal_bio { + struct bch_dev *ca; + unsigned buf_idx; + + struct bio bio; +}; + /* Embedded in struct bch_fs */ struct journal { /* Fastpath stuff up front: */ @@ -203,8 +216,8 @@ struct journal { wait_queue_head_t wait; struct closure_waitlist async_wait; - struct closure io; struct delayed_work write_work; + struct workqueue_struct *wq; /* Sequence number of most recent journal entry (last entry in @pin) */ atomic64_t seq; @@ -274,11 +287,6 @@ struct journal { u64 nr_noflush_writes; u64 entry_bytes_written; - u64 low_on_space_start; - u64 low_on_pin_start; - u64 max_in_flight_start; - u64 write_buffer_full_start; - struct bch2_time_stats *flush_write_time; struct bch2_time_stats *noflush_write_time; struct bch2_time_stats *flush_seq_time; @@ -313,7 +321,7 @@ struct journal_device { u64 *buckets; /* Bio for journal reads/writes to this device */ - struct bio *bio; + struct journal_bio *bio[JOURNAL_BUF_NR]; /* for bch_journal_read_device */ struct closure read; diff --git a/fs/bcachefs/logged_ops.c b/fs/bcachefs/logged_ops.c index ad598105c587..9fac838d123e 100644 --- a/fs/bcachefs/logged_ops.c +++ b/fs/bcachefs/logged_ops.c @@ -101,8 +101,8 @@ void bch2_logged_op_finish(struct btree_trans *trans, struct bkey_i *k) struct printbuf buf = PRINTBUF; bch2_bkey_val_to_text(&buf, c, bkey_i_to_s_c(k)); - bch2_fs_fatal_error(c, "%s: error deleting logged operation %s: %s", - __func__, buf.buf, bch2_err_str(ret)); + bch2_fs_fatal_error(c, "deleting logged operation %s: %s", + buf.buf, bch2_err_str(ret)); printbuf_exit(&buf); } } diff --git a/fs/bcachefs/lru.c b/fs/bcachefs/lru.c index 7a4ca5a28b3e..26569043e368 100644 --- a/fs/bcachefs/lru.c +++ b/fs/bcachefs/lru.c @@ -44,8 +44,8 @@ static int __bch2_lru_set(struct btree_trans *trans, u16 lru_id, u64 dev_bucket, u64 time, bool set) { return time - ? bch2_btree_bit_mod(trans, BTREE_ID_lru, - lru_pos(lru_id, dev_bucket, time), set) + ? bch2_btree_bit_mod_buffered(trans, BTREE_ID_lru, + lru_pos(lru_id, dev_bucket, time), set) : 0; } @@ -125,8 +125,7 @@ static int bch2_check_lru_key(struct btree_trans *trans, goto out; } - if (c->opts.reconstruct_alloc || - fsck_err(c, lru_entry_bad, + if (fsck_err(c, lru_entry_bad, "incorrect lru entry: lru %s time %llu\n" " %s\n" " for %s", diff --git a/fs/bcachefs/mean_and_variance.c b/fs/bcachefs/mean_and_variance.c index bf0ef668fd38..0ea9f30803a2 100644 --- a/fs/bcachefs/mean_and_variance.c +++ b/fs/bcachefs/mean_and_variance.c @@ -103,14 +103,17 @@ EXPORT_SYMBOL_GPL(mean_and_variance_get_stddev); * mean_and_variance_weighted_update() - exponentially weighted variant of mean_and_variance_update() * @s: mean and variance number of samples and their sums * @x: new value to include in the &mean_and_variance_weighted + * @initted: caller must track whether this is the first use or not + * @weight: ewma weight * * see linked pdf: function derived from equations 140-143 where alpha = 2^w. * values are stored bitshifted for performance and added precision. */ -void mean_and_variance_weighted_update(struct mean_and_variance_weighted *s, s64 x) +void mean_and_variance_weighted_update(struct mean_and_variance_weighted *s, + s64 x, bool initted, u8 weight) { // previous weighted variance. - u8 w = s->weight; + u8 w = weight; u64 var_w0 = s->variance; // new value weighted. s64 x_w = x << w; @@ -119,45 +122,50 @@ void mean_and_variance_weighted_update(struct mean_and_variance_weighted *s, s64 // new mean weighted. s64 u_w1 = s->mean + diff; - if (!s->init) { + if (!initted) { s->mean = x_w; s->variance = 0; } else { s->mean = u_w1; s->variance = ((var_w0 << w) - var_w0 + ((diff_w * (x_w - u_w1)) >> w)) >> w; } - s->init = true; } EXPORT_SYMBOL_GPL(mean_and_variance_weighted_update); /** * mean_and_variance_weighted_get_mean() - get mean from @s * @s: mean and variance number of samples and their sums + * @weight: ewma weight */ -s64 mean_and_variance_weighted_get_mean(struct mean_and_variance_weighted s) +s64 mean_and_variance_weighted_get_mean(struct mean_and_variance_weighted s, + u8 weight) { - return fast_divpow2(s.mean, s.weight); + return fast_divpow2(s.mean, weight); } EXPORT_SYMBOL_GPL(mean_and_variance_weighted_get_mean); /** * mean_and_variance_weighted_get_variance() -- get variance from @s * @s: mean and variance number of samples and their sums + * @weight: ewma weight */ -u64 mean_and_variance_weighted_get_variance(struct mean_and_variance_weighted s) +u64 mean_and_variance_weighted_get_variance(struct mean_and_variance_weighted s, + u8 weight) { // always positive don't need fast divpow2 - return s.variance >> s.weight; + return s.variance >> weight; } EXPORT_SYMBOL_GPL(mean_and_variance_weighted_get_variance); /** * mean_and_variance_weighted_get_stddev() - get standard deviation from @s * @s: mean and variance number of samples and their sums + * @weight: ewma weight */ -u32 mean_and_variance_weighted_get_stddev(struct mean_and_variance_weighted s) +u32 mean_and_variance_weighted_get_stddev(struct mean_and_variance_weighted s, + u8 weight) { - return int_sqrt64(mean_and_variance_weighted_get_variance(s)); + return int_sqrt64(mean_and_variance_weighted_get_variance(s, weight)); } EXPORT_SYMBOL_GPL(mean_and_variance_weighted_get_stddev); diff --git a/fs/bcachefs/mean_and_variance.h b/fs/bcachefs/mean_and_variance.h index 64df11ab422b..4fcf062dd22c 100644 --- a/fs/bcachefs/mean_and_variance.h +++ b/fs/bcachefs/mean_and_variance.h @@ -154,8 +154,6 @@ struct mean_and_variance { /* expontentially weighted variant */ struct mean_and_variance_weighted { - bool init; - u8 weight; /* base 2 logarithim */ s64 mean; u64 variance; }; @@ -192,10 +190,14 @@ s64 mean_and_variance_get_mean(struct mean_and_variance s); u64 mean_and_variance_get_variance(struct mean_and_variance s1); u32 mean_and_variance_get_stddev(struct mean_and_variance s); -void mean_and_variance_weighted_update(struct mean_and_variance_weighted *s, s64 v); +void mean_and_variance_weighted_update(struct mean_and_variance_weighted *s, + s64 v, bool initted, u8 weight); -s64 mean_and_variance_weighted_get_mean(struct mean_and_variance_weighted s); -u64 mean_and_variance_weighted_get_variance(struct mean_and_variance_weighted s); -u32 mean_and_variance_weighted_get_stddev(struct mean_and_variance_weighted s); +s64 mean_and_variance_weighted_get_mean(struct mean_and_variance_weighted s, + u8 weight); +u64 mean_and_variance_weighted_get_variance(struct mean_and_variance_weighted s, + u8 weight); +u32 mean_and_variance_weighted_get_stddev(struct mean_and_variance_weighted s, + u8 weight); #endif // MEAN_AND_VAIRANCE_H_ diff --git a/fs/bcachefs/mean_and_variance_test.c b/fs/bcachefs/mean_and_variance_test.c index 019583c3ca0e..db63b3f3b338 100644 --- a/fs/bcachefs/mean_and_variance_test.c +++ b/fs/bcachefs/mean_and_variance_test.c @@ -31,53 +31,59 @@ static void mean_and_variance_basic_test(struct kunit *test) static void mean_and_variance_weighted_test(struct kunit *test) { - struct mean_and_variance_weighted s = { .weight = 2 }; + struct mean_and_variance_weighted s = { }; - mean_and_variance_weighted_update(&s, 10); - KUNIT_EXPECT_EQ(test, mean_and_variance_weighted_get_mean(s), 10); - KUNIT_EXPECT_EQ(test, mean_and_variance_weighted_get_variance(s), 0); + mean_and_variance_weighted_update(&s, 10, false, 2); + KUNIT_EXPECT_EQ(test, mean_and_variance_weighted_get_mean(s, 2), 10); + KUNIT_EXPECT_EQ(test, mean_and_variance_weighted_get_variance(s, 2), 0); - mean_and_variance_weighted_update(&s, 20); - KUNIT_EXPECT_EQ(test, mean_and_variance_weighted_get_mean(s), 12); - KUNIT_EXPECT_EQ(test, mean_and_variance_weighted_get_variance(s), 18); + mean_and_variance_weighted_update(&s, 20, true, 2); + KUNIT_EXPECT_EQ(test, mean_and_variance_weighted_get_mean(s, 2), 12); + KUNIT_EXPECT_EQ(test, mean_and_variance_weighted_get_variance(s, 2), 18); - mean_and_variance_weighted_update(&s, 30); - KUNIT_EXPECT_EQ(test, mean_and_variance_weighted_get_mean(s), 16); - KUNIT_EXPECT_EQ(test, mean_and_variance_weighted_get_variance(s), 72); + mean_and_variance_weighted_update(&s, 30, true, 2); + KUNIT_EXPECT_EQ(test, mean_and_variance_weighted_get_mean(s, 2), 16); + KUNIT_EXPECT_EQ(test, mean_and_variance_weighted_get_variance(s, 2), 72); - s = (struct mean_and_variance_weighted) { .weight = 2 }; + s = (struct mean_and_variance_weighted) { }; - mean_and_variance_weighted_update(&s, -10); - KUNIT_EXPECT_EQ(test, mean_and_variance_weighted_get_mean(s), -10); - KUNIT_EXPECT_EQ(test, mean_and_variance_weighted_get_variance(s), 0); + mean_and_variance_weighted_update(&s, -10, false, 2); + KUNIT_EXPECT_EQ(test, mean_and_variance_weighted_get_mean(s, 2), -10); + KUNIT_EXPECT_EQ(test, mean_and_variance_weighted_get_variance(s, 2), 0); - mean_and_variance_weighted_update(&s, -20); - KUNIT_EXPECT_EQ(test, mean_and_variance_weighted_get_mean(s), -12); - KUNIT_EXPECT_EQ(test, mean_and_variance_weighted_get_variance(s), 18); + mean_and_variance_weighted_update(&s, -20, true, 2); + KUNIT_EXPECT_EQ(test, mean_and_variance_weighted_get_mean(s, 2), -12); + KUNIT_EXPECT_EQ(test, mean_and_variance_weighted_get_variance(s, 2), 18); - mean_and_variance_weighted_update(&s, -30); - KUNIT_EXPECT_EQ(test, mean_and_variance_weighted_get_mean(s), -16); - KUNIT_EXPECT_EQ(test, mean_and_variance_weighted_get_variance(s), 72); + mean_and_variance_weighted_update(&s, -30, true, 2); + KUNIT_EXPECT_EQ(test, mean_and_variance_weighted_get_mean(s, 2), -16); + KUNIT_EXPECT_EQ(test, mean_and_variance_weighted_get_variance(s, 2), 72); } static void mean_and_variance_weighted_advanced_test(struct kunit *test) { - struct mean_and_variance_weighted s = { .weight = 8 }; + struct mean_and_variance_weighted s = { }; + bool initted = false; s64 i; - for (i = 10; i <= 100; i += 10) - mean_and_variance_weighted_update(&s, i); + for (i = 10; i <= 100; i += 10) { + mean_and_variance_weighted_update(&s, i, initted, 8); + initted = true; + } - KUNIT_EXPECT_EQ(test, mean_and_variance_weighted_get_mean(s), 11); - KUNIT_EXPECT_EQ(test, mean_and_variance_weighted_get_variance(s), 107); + KUNIT_EXPECT_EQ(test, mean_and_variance_weighted_get_mean(s, 8), 11); + KUNIT_EXPECT_EQ(test, mean_and_variance_weighted_get_variance(s, 8), 107); - s = (struct mean_and_variance_weighted) { .weight = 8 }; + s = (struct mean_and_variance_weighted) { }; + initted = false; - for (i = -10; i >= -100; i -= 10) - mean_and_variance_weighted_update(&s, i); + for (i = -10; i >= -100; i -= 10) { + mean_and_variance_weighted_update(&s, i, initted, 8); + initted = true; + } - KUNIT_EXPECT_EQ(test, mean_and_variance_weighted_get_mean(s), -11); - KUNIT_EXPECT_EQ(test, mean_and_variance_weighted_get_variance(s), 107); + KUNIT_EXPECT_EQ(test, mean_and_variance_weighted_get_mean(s, 8), -11); + KUNIT_EXPECT_EQ(test, mean_and_variance_weighted_get_variance(s, 8), 107); } static void do_mean_and_variance_test(struct kunit *test, @@ -92,26 +98,26 @@ static void do_mean_and_variance_test(struct kunit *test, s64 *weighted_stddev) { struct mean_and_variance mv = {}; - struct mean_and_variance_weighted vw = { .weight = weight }; + struct mean_and_variance_weighted vw = { }; for (unsigned i = 0; i < initial_n; i++) { mean_and_variance_update(&mv, initial_value); - mean_and_variance_weighted_update(&vw, initial_value); + mean_and_variance_weighted_update(&vw, initial_value, false, weight); KUNIT_EXPECT_EQ(test, mean_and_variance_get_mean(mv), initial_value); KUNIT_EXPECT_EQ(test, mean_and_variance_get_stddev(mv), 0); - KUNIT_EXPECT_EQ(test, mean_and_variance_weighted_get_mean(vw), initial_value); - KUNIT_EXPECT_EQ(test, mean_and_variance_weighted_get_stddev(vw),0); + KUNIT_EXPECT_EQ(test, mean_and_variance_weighted_get_mean(vw, weight), initial_value); + KUNIT_EXPECT_EQ(test, mean_and_variance_weighted_get_stddev(vw, weight),0); } for (unsigned i = 0; i < n; i++) { mean_and_variance_update(&mv, data[i]); - mean_and_variance_weighted_update(&vw, data[i]); + mean_and_variance_weighted_update(&vw, data[i], true, weight); KUNIT_EXPECT_EQ(test, mean_and_variance_get_mean(mv), mean[i]); KUNIT_EXPECT_EQ(test, mean_and_variance_get_stddev(mv), stddev[i]); - KUNIT_EXPECT_EQ(test, mean_and_variance_weighted_get_mean(vw), weighted_mean[i]); - KUNIT_EXPECT_EQ(test, mean_and_variance_weighted_get_stddev(vw),weighted_stddev[i]); + KUNIT_EXPECT_EQ(test, mean_and_variance_weighted_get_mean(vw, weight), weighted_mean[i]); + KUNIT_EXPECT_EQ(test, mean_and_variance_weighted_get_stddev(vw, weight),weighted_stddev[i]); } KUNIT_EXPECT_EQ(test, mv.n, initial_n + n); diff --git a/fs/bcachefs/migrate.c b/fs/bcachefs/migrate.c index 5623cee3ef86..69098eeb5d48 100644 --- a/fs/bcachefs/migrate.c +++ b/fs/bcachefs/migrate.c @@ -31,7 +31,7 @@ static int drop_dev_ptrs(struct bch_fs *c, struct bkey_s k, nr_good = bch2_bkey_durability(c, k.s_c); if ((!nr_good && !(flags & lost)) || (nr_good < replicas && !(flags & degraded))) - return -EINVAL; + return -BCH_ERR_remove_would_lose_data; return 0; } @@ -111,7 +111,7 @@ static int bch2_dev_metadata_drop(struct bch_fs *c, unsigned dev_idx, int flags) /* don't handle this yet: */ if (flags & BCH_FORCE_IF_METADATA_LOST) - return -EINVAL; + return -BCH_ERR_remove_with_metadata_missing_unimplemented; trans = bch2_trans_get(c); bch2_bkey_buf_init(&k); @@ -132,10 +132,8 @@ retry: ret = drop_dev_ptrs(c, bkey_i_to_s(k.k), dev_idx, flags, true); - if (ret) { - bch_err(c, "Cannot drop device without losing data"); + if (ret) break; - } ret = bch2_btree_node_update_key(trans, &iter, b, k.k, 0, false); if (bch2_err_matches(ret, BCH_ERR_transaction_restart)) { diff --git a/fs/bcachefs/movinggc.c b/fs/bcachefs/movinggc.c index 69e06a84dad4..0d2b82d8d11f 100644 --- a/fs/bcachefs/movinggc.c +++ b/fs/bcachefs/movinggc.c @@ -155,8 +155,7 @@ static int bch2_copygc_get_buckets(struct moving_context *ctxt, if (bch2_err_matches(ret, EROFS)) return ret; - if (bch2_fs_fatal_err_on(ret, c, "%s: error %s from bch2_btree_write_buffer_tryflush()", - __func__, bch2_err_str(ret))) + if (bch2_fs_fatal_err_on(ret, c, "%s: from bch2_btree_write_buffer_tryflush()", bch2_err_str(ret))) return ret; ret = for_each_btree_key_upto(trans, iter, BTREE_ID_lru, diff --git a/fs/bcachefs/opts.c b/fs/bcachefs/opts.c index b1ed0b9a20d3..08ea0cfc4aef 100644 --- a/fs/bcachefs/opts.c +++ b/fs/bcachefs/opts.c @@ -314,7 +314,7 @@ int bch2_opt_parse(struct bch_fs *c, if (ret < 0 || (*res != 0 && *res != 1)) { if (err) prt_printf(err, "%s: must be bool", opt->attr.name); - return ret; + return ret < 0 ? ret : -BCH_ERR_option_not_bool; } break; case BCH_OPT_UINT: @@ -456,7 +456,7 @@ int bch2_parse_mount_opts(struct bch_fs *c, struct bch_opts *opts, copied_opts = kstrdup(options, GFP_KERNEL); if (!copied_opts) - return -1; + return -ENOMEM; copied_opts_start = copied_opts; while ((opt = strsep(&copied_opts, ",")) != NULL) { @@ -501,11 +501,11 @@ int bch2_parse_mount_opts(struct bch_fs *c, struct bch_opts *opts, bad_opt: pr_err("Bad mount option %s", name); - ret = -1; + ret = -BCH_ERR_option_name; goto out; bad_val: pr_err("Invalid mount option %s", err.buf); - ret = -1; + ret = -BCH_ERR_option_value; goto out; out: kfree(copied_opts_start); diff --git a/fs/bcachefs/opts.h b/fs/bcachefs/opts.h index 9a4b7faa3765..136083c11f3a 100644 --- a/fs/bcachefs/opts.h +++ b/fs/bcachefs/opts.h @@ -290,6 +290,11 @@ enum fsck_err_opts { OPT_BOOL(), \ BCH2_NO_SB_OPT, false, \ NULL, "Allow mounting in when data will be missing") \ + x(no_splitbrain_check, u8, \ + OPT_FS|OPT_MOUNT, \ + OPT_BOOL(), \ + BCH2_NO_SB_OPT, false, \ + NULL, "Don't kick drives out when splitbrain detected")\ x(discard, u8, \ OPT_FS|OPT_MOUNT|OPT_DEVICE, \ OPT_BOOL(), \ @@ -332,6 +337,11 @@ enum fsck_err_opts { OPT_BOOL(), \ BCH2_NO_SB_OPT, false, \ NULL, "Run fsck on mount") \ + x(fsck_memory_usage_percent, u8, \ + OPT_FS|OPT_MOUNT, \ + OPT_UINT(20, 70), \ + BCH2_NO_SB_OPT, 50, \ + NULL, "Maximum percentage of system ram fsck is allowed to pin")\ x(fix_errors, u8, \ OPT_FS|OPT_MOUNT, \ OPT_FN(bch2_opt_fix_errors), \ diff --git a/fs/bcachefs/rebalance.c b/fs/bcachefs/rebalance.c index 22d1017aa49b..56336f3dd1d0 100644 --- a/fs/bcachefs/rebalance.c +++ b/fs/bcachefs/rebalance.c @@ -412,11 +412,11 @@ void bch2_rebalance_status_to_text(struct printbuf *out, struct bch_fs *c) u64 now = atomic64_read(&c->io_clock[WRITE].now); prt_str(out, "io wait duration: "); - bch2_prt_human_readable_s64(out, r->wait_iotime_end - r->wait_iotime_start); + bch2_prt_human_readable_s64(out, (r->wait_iotime_end - r->wait_iotime_start) << 9); prt_newline(out); prt_str(out, "io wait remaining: "); - bch2_prt_human_readable_s64(out, r->wait_iotime_end - now); + bch2_prt_human_readable_s64(out, (r->wait_iotime_end - now) << 9); prt_newline(out); prt_str(out, "duration waited: "); diff --git a/fs/bcachefs/recovery.c b/fs/bcachefs/recovery.c index 21e13bb4335b..03f9d6afe467 100644 --- a/fs/bcachefs/recovery.c +++ b/fs/bcachefs/recovery.c @@ -52,15 +52,50 @@ static bool btree_id_is_alloc(enum btree_id id) } /* for -o reconstruct_alloc: */ -static void drop_alloc_keys(struct journal_keys *keys) +static void do_reconstruct_alloc(struct bch_fs *c) { + bch2_journal_log_msg(c, "dropping alloc info"); + bch_info(c, "dropping and reconstructing all alloc info"); + + mutex_lock(&c->sb_lock); + struct bch_sb_field_ext *ext = bch2_sb_field_get(c->disk_sb.sb, ext); + + __set_bit_le64(BCH_RECOVERY_PASS_STABLE_check_allocations, ext->recovery_passes_required); + __set_bit_le64(BCH_RECOVERY_PASS_STABLE_check_alloc_info, ext->recovery_passes_required); + __set_bit_le64(BCH_RECOVERY_PASS_STABLE_check_lrus, ext->recovery_passes_required); + __set_bit_le64(BCH_RECOVERY_PASS_STABLE_check_extents_to_backpointers, ext->recovery_passes_required); + __set_bit_le64(BCH_RECOVERY_PASS_STABLE_check_alloc_to_lru_refs, ext->recovery_passes_required); + + __set_bit_le64(BCH_FSCK_ERR_ptr_to_missing_alloc_key, ext->errors_silent); + __set_bit_le64(BCH_FSCK_ERR_ptr_gen_newer_than_bucket_gen, ext->errors_silent); + __set_bit_le64(BCH_FSCK_ERR_stale_dirty_ptr, ext->errors_silent); + __set_bit_le64(BCH_FSCK_ERR_alloc_key_data_type_wrong, ext->errors_silent); + __set_bit_le64(BCH_FSCK_ERR_alloc_key_gen_wrong, ext->errors_silent); + __set_bit_le64(BCH_FSCK_ERR_alloc_key_dirty_sectors_wrong, ext->errors_silent); + __set_bit_le64(BCH_FSCK_ERR_alloc_key_stripe_wrong, ext->errors_silent); + __set_bit_le64(BCH_FSCK_ERR_alloc_key_stripe_redundancy_wrong, ext->errors_silent); + __set_bit_le64(BCH_FSCK_ERR_need_discard_key_wrong, ext->errors_silent); + __set_bit_le64(BCH_FSCK_ERR_freespace_key_wrong, ext->errors_silent); + __set_bit_le64(BCH_FSCK_ERR_bucket_gens_key_wrong, ext->errors_silent); + __set_bit_le64(BCH_FSCK_ERR_freespace_hole_missing, ext->errors_silent); + __set_bit_le64(BCH_FSCK_ERR_ptr_to_missing_backpointer, ext->errors_silent); + __set_bit_le64(BCH_FSCK_ERR_lru_entry_bad, ext->errors_silent); + c->sb.compat &= ~(1ULL << BCH_COMPAT_alloc_info); + + bch2_write_super(c); + mutex_unlock(&c->sb_lock); + + c->recovery_passes_explicit |= bch2_recovery_passes_from_stable(le64_to_cpu(ext->recovery_passes_required[0])); + + struct journal_keys *keys = &c->journal_keys; size_t src, dst; - for (src = 0, dst = 0; src < keys->nr; src++) - if (!btree_id_is_alloc(keys->d[src].btree_id)) - keys->d[dst++] = keys->d[src]; + move_gap(keys, keys->nr); - keys->nr = dst; + for (src = 0, dst = 0; src < keys->nr; src++) + if (!btree_id_is_alloc(keys->data[src].btree_id)) + keys->data[dst++] = keys->data[src]; + keys->nr = keys->gap = dst; } /* @@ -70,9 +105,7 @@ static void drop_alloc_keys(struct journal_keys *keys) */ static void zero_out_btree_mem_ptr(struct journal_keys *keys) { - struct journal_key *i; - - for (i = keys->d; i < keys->d + keys->nr; i++) + darray_for_each(*keys, i) if (i->k->k.type == KEY_TYPE_btree_ptr_v2) bkey_i_to_btree_ptr_v2(i->k)->v.mem_ptr = 0; } @@ -124,6 +157,17 @@ static int bch2_journal_replay_key(struct btree_trans *trans, if (ret) goto out; + struct btree_path *path = btree_iter_path(trans, &iter); + if (unlikely(!btree_path_node(path, k->level))) { + bch2_trans_iter_exit(trans, &iter); + bch2_trans_node_iter_init(trans, &iter, k->btree_id, k->k->k.p, + BTREE_MAX_DEPTH, 0, iter_flags); + ret = bch2_btree_iter_traverse(&iter) ?: + bch2_btree_increase_depth(trans, iter.path, 0) ?: + -BCH_ERR_transaction_restart_nested; + goto out; + } + /* Must be checked with btree locked: */ if (k->overwritten) goto out; @@ -161,16 +205,16 @@ static int bch2_journal_replay(struct bch_fs *c) BUG_ON(!atomic_read(&keys->ref)); + move_gap(keys, keys->nr); + /* * First, attempt to replay keys in sorted order. This is more * efficient - better locality of btree access - but some might fail if * that would cause a journal deadlock. */ - for (size_t i = 0; i < keys->nr; i++) { + darray_for_each(*keys, k) { cond_resched(); - struct journal_key *k = keys->d + i; - /* Skip fastpath if we're low on space in the journal */ ret = c->journal.watermark ? -1 : commit_do(trans, NULL, NULL, @@ -264,7 +308,7 @@ static int journal_replay_entry_early(struct bch_fs *c, bkey_copy(&r->key, (struct bkey_i *) entry->start); r->error = 0; } else { - r->error = -EIO; + r->error = -BCH_ERR_btree_node_read_error; } r->alive = true; break; @@ -359,7 +403,7 @@ static int journal_replay_early(struct bch_fs *c, genradix_for_each(&c->journal_entries, iter, _i) { i = *_i; - if (!i || i->ignore) + if (journal_replay_ignore(i)) continue; vstruct_for_each(&i->j, entry) { @@ -388,11 +432,8 @@ static int read_btree_roots(struct bch_fs *c) if (!r->alive) continue; - if (btree_id_is_alloc(i) && - c->opts.reconstruct_alloc) { - c->sb.compat &= ~(1ULL << BCH_COMPAT_alloc_info); + if (btree_id_is_alloc(i) && c->opts.reconstruct_alloc) continue; - } if (r->error) { __fsck_err(c, @@ -524,8 +565,7 @@ static int bch2_set_may_go_rw(struct bch_fs *c) * setting journal_key->overwritten: it will be accessed by multiple * threads */ - move_gap(keys->d, keys->nr, keys->size, keys->gap, keys->nr); - keys->gap = keys->nr; + move_gap(keys, keys->nr); set_bit(BCH_FS_may_go_rw, &c->flags); @@ -862,7 +902,7 @@ int bch2_fs_recovery(struct bch_fs *c) goto out; genradix_for_each_reverse(&c->journal_entries, iter, i) - if (*i && !(*i)->ignore) { + if (!journal_replay_ignore(*i)) { last_journal_entry = &(*i)->j; break; } @@ -887,7 +927,8 @@ int bch2_fs_recovery(struct bch_fs *c) genradix_for_each_reverse(&c->journal_entries, iter, i) if (*i) { last_journal_entry = &(*i)->j; - (*i)->ignore = false; + (*i)->ignore_blacklisted = false; + (*i)->ignore_not_dirty= false; /* * This was probably a NO_FLUSH entry, * so last_seq was garbage - but we know @@ -923,10 +964,8 @@ use_clean: c->journal_replay_seq_start = last_seq; c->journal_replay_seq_end = blacklist_seq - 1; - if (c->opts.reconstruct_alloc) { - c->sb.compat &= ~(1ULL << BCH_COMPAT_alloc_info); - drop_alloc_keys(&c->journal_keys); - } + if (c->opts.reconstruct_alloc) + do_reconstruct_alloc(c); zero_out_btree_mem_ptr(&c->journal_keys); @@ -950,7 +989,7 @@ use_clean: bch2_journal_seq_blacklist_add(c, blacklist_seq, journal_seq); if (ret) { - bch_err(c, "error creating new journal seq blacklist entry"); + bch_err_msg(c, ret, "error creating new journal seq blacklist entry"); goto err; } } @@ -961,9 +1000,6 @@ use_clean: if (ret) goto err; - if (c->opts.reconstruct_alloc) - bch2_journal_log_msg(c, "dropping alloc info"); - /* * Skip past versions that might have possibly been used (as nonces), * but hadn't had their pointers written: diff --git a/fs/bcachefs/recovery_types.h b/fs/bcachefs/recovery_types.h index fa0c8efd2a1b..4959e95e7c74 100644 --- a/fs/bcachefs/recovery_types.h +++ b/fs/bcachefs/recovery_types.h @@ -13,11 +13,11 @@ * must never change: */ #define BCH_RECOVERY_PASSES() \ + x(check_topology, 4, 0) \ x(alloc_read, 0, PASS_ALWAYS) \ x(stripes_read, 1, PASS_ALWAYS) \ x(initialize_subvolumes, 2, 0) \ x(snapshots_read, 3, PASS_ALWAYS) \ - x(check_topology, 4, 0) \ x(check_allocations, 5, PASS_FSCK) \ x(trans_mark_dev_sbs, 6, PASS_ALWAYS|PASS_SILENT) \ x(fs_journal_alloc, 7, PASS_ALWAYS|PASS_SILENT) \ @@ -34,6 +34,7 @@ x(check_snapshot_trees, 18, PASS_ONLINE|PASS_FSCK) \ x(check_snapshots, 19, PASS_ONLINE|PASS_FSCK) \ x(check_subvols, 20, PASS_ONLINE|PASS_FSCK) \ + x(check_subvol_children, 35, PASS_ONLINE|PASS_FSCK) \ x(delete_dead_snapshots, 21, PASS_ONLINE|PASS_FSCK) \ x(fs_upgrade_for_subvolumes, 22, 0) \ x(resume_logged_ops, 23, PASS_ALWAYS) \ @@ -43,6 +44,7 @@ x(check_dirents, 27, PASS_FSCK) \ x(check_xattrs, 28, PASS_FSCK) \ x(check_root, 29, PASS_ONLINE|PASS_FSCK) \ + x(check_subvolume_structure, 36, PASS_ONLINE|PASS_FSCK) \ x(check_directory_structure, 30, PASS_ONLINE|PASS_FSCK) \ x(check_nlinks, 31, PASS_FSCK) \ x(delete_dead_inodes, 32, PASS_FSCK|PASS_UNCLEAN) \ diff --git a/fs/bcachefs/sb-clean.c b/fs/bcachefs/sb-clean.c index b6bf0ebe7e84..5980ba2563fe 100644 --- a/fs/bcachefs/sb-clean.c +++ b/fs/bcachefs/sb-clean.c @@ -171,22 +171,6 @@ fsck_err: return ERR_PTR(ret); } -static struct jset_entry *jset_entry_init(struct jset_entry **end, size_t size) -{ - struct jset_entry *entry = *end; - unsigned u64s = DIV_ROUND_UP(size, sizeof(u64)); - - memset(entry, 0, u64s * sizeof(u64)); - /* - * The u64s field counts from the start of data, ignoring the shared - * fields. - */ - entry->u64s = cpu_to_le16(u64s - 1); - - *end = vstruct_next(*end); - return entry; -} - void bch2_journal_super_entries_add_common(struct bch_fs *c, struct jset_entry **end, u64 journal_seq) diff --git a/fs/bcachefs/sb-downgrade.c b/fs/bcachefs/sb-downgrade.c index 441dcb1bf160..e4396cb0bacb 100644 --- a/fs/bcachefs/sb-downgrade.c +++ b/fs/bcachefs/sb-downgrade.c @@ -45,7 +45,13 @@ BIT_ULL(BCH_RECOVERY_PASS_check_inodes), \ BCH_FSCK_ERR_unlinked_inode_not_on_deleted_list) \ x(rebalance_work, \ - BIT_ULL(BCH_RECOVERY_PASS_set_fs_needs_rebalance)) + BIT_ULL(BCH_RECOVERY_PASS_set_fs_needs_rebalance)) \ + x(subvolume_fs_parent, \ + BIT_ULL(BCH_RECOVERY_PASS_check_dirents), \ + BCH_FSCK_ERR_subvol_fs_path_parent_wrong) \ + x(btree_subvolume_children, \ + BIT_ULL(BCH_RECOVERY_PASS_check_subvols), \ + BCH_FSCK_ERR_subvol_children_not_set) #define DOWNGRADE_TABLE() @@ -253,7 +259,7 @@ void bch2_sb_set_downgrade(struct bch_fs *c, unsigned new_minor, unsigned old_mi if (e < BCH_SB_ERR_MAX) __set_bit(e, c->sb.errors_silent); if (e < sizeof(ext->errors_silent) * 8) - ext->errors_silent[e / 64] |= cpu_to_le64(BIT_ULL(e % 64)); + __set_bit_le64(e, ext->errors_silent); } } } diff --git a/fs/bcachefs/sb-errors_types.h b/fs/bcachefs/sb-errors_types.h index c08aacdfd073..5178bf579f7c 100644 --- a/fs/bcachefs/sb-errors_types.h +++ b/fs/bcachefs/sb-errors_types.h @@ -231,7 +231,7 @@ x(dirent_name_dot_or_dotdot, 223) \ x(dirent_name_has_slash, 224) \ x(dirent_d_type_wrong, 225) \ - x(dirent_d_parent_subvol_wrong, 226) \ + x(inode_bi_parent_wrong, 226) \ x(dirent_in_missing_dir_inode, 227) \ x(dirent_in_non_dir_inode, 228) \ x(dirent_to_missing_inode, 229) \ @@ -250,7 +250,22 @@ x(hash_table_key_duplicate, 242) \ x(hash_table_key_wrong_offset, 243) \ x(unlinked_inode_not_on_deleted_list, 244) \ - x(reflink_p_front_pad_bad, 245) + x(reflink_p_front_pad_bad, 245) \ + x(journal_entry_dup_same_device, 246) \ + x(inode_bi_subvol_missing, 247) \ + x(inode_bi_subvol_wrong, 248) \ + x(inode_points_to_missing_dirent, 249) \ + x(inode_points_to_wrong_dirent, 250) \ + x(inode_bi_parent_nonzero, 251) \ + x(dirent_to_missing_parent_subvol, 252) \ + x(dirent_not_visible_in_parent_subvol, 253) \ + x(subvol_fs_path_parent_wrong, 254) \ + x(subvol_root_fs_path_parent_nonzero, 255) \ + x(subvol_children_not_set, 256) \ + x(subvol_children_bad, 257) \ + x(subvol_loop, 258) \ + x(subvol_unreachable, 259) \ + x(btree_node_bkey_bad_u64s, 260) enum bch_sb_error_id { #define x(t, n) BCH_FSCK_ERR_##t = n, diff --git a/fs/bcachefs/snapshot.c b/fs/bcachefs/snapshot.c index ac6ba04d5521..39debe814bf3 100644 --- a/fs/bcachefs/snapshot.c +++ b/fs/bcachefs/snapshot.c @@ -91,18 +91,20 @@ static int bch2_snapshot_tree_create(struct btree_trans *trans, /* Snapshot nodes: */ -static bool bch2_snapshot_is_ancestor_early(struct bch_fs *c, u32 id, u32 ancestor) +static bool __bch2_snapshot_is_ancestor_early(struct snapshot_table *t, u32 id, u32 ancestor) { - struct snapshot_table *t; - - rcu_read_lock(); - t = rcu_dereference(c->snapshots); - while (id && id < ancestor) id = __snapshot_t(t, id)->parent; + return id == ancestor; +} + +static bool bch2_snapshot_is_ancestor_early(struct bch_fs *c, u32 id, u32 ancestor) +{ + rcu_read_lock(); + bool ret = __bch2_snapshot_is_ancestor_early(rcu_dereference(c->snapshots), id, ancestor); rcu_read_unlock(); - return id == ancestor; + return ret; } static inline u32 get_ancestor_below(struct snapshot_table *t, u32 id, u32 ancestor) @@ -120,13 +122,15 @@ static inline u32 get_ancestor_below(struct snapshot_table *t, u32 id, u32 ances bool __bch2_snapshot_is_ancestor(struct bch_fs *c, u32 id, u32 ancestor) { - struct snapshot_table *t; bool ret; - EBUG_ON(c->recovery_pass_done <= BCH_RECOVERY_PASS_check_snapshots); - rcu_read_lock(); - t = rcu_dereference(c->snapshots); + struct snapshot_table *t = rcu_dereference(c->snapshots); + + if (unlikely(c->recovery_pass_done <= BCH_RECOVERY_PASS_check_snapshots)) { + ret = __bch2_snapshot_is_ancestor_early(t, id, ancestor); + goto out; + } while (id && id < ancestor - IS_ANCESTOR_BITMAP) id = get_ancestor_below(t, id, ancestor); @@ -134,11 +138,11 @@ bool __bch2_snapshot_is_ancestor(struct bch_fs *c, u32 id, u32 ancestor) if (id && id < ancestor) { ret = test_bit(ancestor - id - 1, __snapshot_t(t, id)->is_ancestor); - EBUG_ON(ret != bch2_snapshot_is_ancestor_early(c, id, ancestor)); + EBUG_ON(ret != __bch2_snapshot_is_ancestor_early(t, id, ancestor)); } else { ret = id == ancestor; } - +out: rcu_read_unlock(); return ret; @@ -547,7 +551,7 @@ static int check_snapshot_tree(struct btree_trans *trans, "snapshot tree points to missing subvolume:\n %s", (printbuf_reset(&buf), bch2_bkey_val_to_text(&buf, c, st.s_c), buf.buf)) || - fsck_err_on(!bch2_snapshot_is_ancestor_early(c, + fsck_err_on(!bch2_snapshot_is_ancestor(c, le32_to_cpu(subvol.snapshot), root_id), c, snapshot_tree_to_wrong_subvol, diff --git a/fs/bcachefs/str_hash.h b/fs/bcachefs/str_hash.h index fcaa5a888744..3976f80721bf 100644 --- a/fs/bcachefs/str_hash.h +++ b/fs/bcachefs/str_hash.h @@ -259,7 +259,7 @@ int bch2_hash_needs_whiteout(struct btree_trans *trans, } static __always_inline -int bch2_hash_set_snapshot(struct btree_trans *trans, +int bch2_hash_set_in_snapshot(struct btree_trans *trans, const struct bch_hash_desc desc, const struct bch_hash_info *info, subvol_inum inum, u32 snapshot, @@ -328,17 +328,12 @@ int bch2_hash_set(struct btree_trans *trans, struct bkey_i *insert, bch_str_hash_flags_t str_hash_flags) { - u32 snapshot; - int ret; - - ret = bch2_subvolume_get_snapshot(trans, inum.subvol, &snapshot); - if (ret) - return ret; - insert->k.p.inode = inum.inum; - return bch2_hash_set_snapshot(trans, desc, info, inum, - snapshot, insert, str_hash_flags, 0); + u32 snapshot; + return bch2_subvolume_get_snapshot(trans, inum.subvol, &snapshot) ?: + bch2_hash_set_in_snapshot(trans, desc, info, inum, + snapshot, insert, str_hash_flags, 0); } static __always_inline diff --git a/fs/bcachefs/subvolume.c b/fs/bcachefs/subvolume.c index 7c67c28d3ef8..ce7aed121942 100644 --- a/fs/bcachefs/subvolume.c +++ b/fs/bcachefs/subvolume.c @@ -13,13 +13,26 @@ static int bch2_subvolume_delete(struct btree_trans *, u32); +static struct bpos subvolume_children_pos(struct bkey_s_c k) +{ + if (k.k->type != KEY_TYPE_subvolume) + return POS_MIN; + + struct bkey_s_c_subvolume s = bkey_s_c_to_subvolume(k); + if (!s.v->fs_path_parent) + return POS_MIN; + return POS(le32_to_cpu(s.v->fs_path_parent), s.k->p.offset); +} + static int check_subvol(struct btree_trans *trans, struct btree_iter *iter, struct bkey_s_c k) { struct bch_fs *c = trans->c; struct bkey_s_c_subvolume subvol; + struct btree_iter subvol_children_iter = {}; struct bch_snapshot snapshot; + struct printbuf buf = PRINTBUF; unsigned snapid; int ret = 0; @@ -42,6 +55,72 @@ static int check_subvol(struct btree_trans *trans, return ret ?: -BCH_ERR_transaction_restart_nested; } + if (fsck_err_on(subvol.k->p.offset == BCACHEFS_ROOT_SUBVOL && + subvol.v->fs_path_parent, + c, subvol_root_fs_path_parent_nonzero, + "root subvolume has nonzero fs_path_parent\n%s", + (bch2_bkey_val_to_text(&buf, c, k), buf.buf))) { + struct bkey_i_subvolume *n = + bch2_bkey_make_mut_typed(trans, iter, &subvol.s_c, 0, subvolume); + ret = PTR_ERR_OR_ZERO(n); + if (ret) + goto err; + + n->v.fs_path_parent = 0; + } + + if (subvol.v->fs_path_parent) { + struct bpos pos = subvolume_children_pos(k); + + struct bkey_s_c subvol_children_k = + bch2_bkey_get_iter(trans, &subvol_children_iter, + BTREE_ID_subvolume_children, pos, 0); + ret = bkey_err(subvol_children_k); + if (ret) + goto err; + + if (fsck_err_on(subvol_children_k.k->type != KEY_TYPE_set, + c, subvol_children_not_set, + "subvolume not set in subvolume_children btree at %llu:%llu\n%s", + pos.inode, pos.offset, + (printbuf_reset(&buf), + bch2_bkey_val_to_text(&buf, c, k), buf.buf))) { + ret = bch2_btree_bit_mod(trans, BTREE_ID_subvolume_children, pos, true); + if (ret) + goto err; + } + } + + struct bch_inode_unpacked inode; + struct btree_iter inode_iter = {}; + ret = bch2_inode_peek_nowarn(trans, &inode_iter, &inode, + (subvol_inum) { k.k->p.offset, le64_to_cpu(subvol.v->inode) }, + 0); + bch2_trans_iter_exit(trans, &inode_iter); + + if (ret && !bch2_err_matches(ret, ENOENT)) + return ret; + + if (fsck_err_on(ret, c, subvol_to_missing_root, + "subvolume %llu points to missing subvolume root %llu:%u", + k.k->p.offset, le64_to_cpu(subvol.v->inode), + le32_to_cpu(subvol.v->snapshot))) { + ret = bch2_subvolume_delete(trans, iter->pos.offset); + bch_err_msg(c, ret, "deleting subvolume %llu", iter->pos.offset); + return ret ?: -BCH_ERR_transaction_restart_nested; + } + + if (fsck_err_on(inode.bi_subvol != subvol.k->p.offset, + c, subvol_root_wrong_bi_subvol, + "subvol root %llu:%u has wrong bi_subvol field: got %u, should be %llu", + inode.bi_inum, inode_iter.k.p.snapshot, + inode.bi_subvol, subvol.k->p.offset)) { + inode.bi_subvol = subvol.k->p.offset; + ret = __bch2_fsck_write_inode(trans, &inode, le32_to_cpu(subvol.v->snapshot)); + if (ret) + goto err; + } + if (!BCH_SUBVOLUME_SNAP(subvol.v)) { u32 snapshot_root = bch2_snapshot_root(c, le32_to_cpu(subvol.v->snapshot)); u32 snapshot_tree; @@ -72,8 +151,10 @@ static int check_subvol(struct btree_trans *trans, SET_BCH_SUBVOLUME_SNAP(&s->v, true); } } - +err: fsck_err: + bch2_trans_iter_exit(trans, &subvol_children_iter); + printbuf_exit(&buf); return ret; } @@ -88,6 +169,42 @@ int bch2_check_subvols(struct bch_fs *c) return ret; } +static int check_subvol_child(struct btree_trans *trans, + struct btree_iter *child_iter, + struct bkey_s_c child_k) +{ + struct bch_fs *c = trans->c; + struct bch_subvolume s; + int ret = bch2_bkey_get_val_typed(trans, BTREE_ID_subvolumes, POS(0, child_k.k->p.offset), + 0, subvolume, &s); + if (ret && !bch2_err_matches(ret, ENOENT)) + return ret; + + if (fsck_err_on(ret || + le32_to_cpu(s.fs_path_parent) != child_k.k->p.inode, + c, subvol_children_bad, + "incorrect entry in subvolume_children btree %llu:%llu", + child_k.k->p.inode, child_k.k->p.offset)) { + ret = bch2_btree_delete_at(trans, child_iter, 0); + if (ret) + goto err; + } +err: +fsck_err: + return ret; +} + +int bch2_check_subvol_children(struct bch_fs *c) +{ + int ret = bch2_trans_run(c, + for_each_btree_key_commit(trans, iter, + BTREE_ID_subvolume_children, POS_MIN, BTREE_ITER_PREFETCH, k, + NULL, NULL, BCH_TRANS_COMMIT_no_enospc, + check_subvol_child(trans, &iter, k))); + bch_err_fn(c, ret); + return 0; +} + /* Subvolumes: */ int bch2_subvolume_invalid(struct bch_fs *c, struct bkey_s_c k, @@ -112,8 +229,50 @@ void bch2_subvolume_to_text(struct printbuf *out, struct bch_fs *c, le64_to_cpu(s.v->inode), le32_to_cpu(s.v->snapshot)); - if (bkey_val_bytes(s.k) > offsetof(struct bch_subvolume, parent)) - prt_printf(out, " parent %u", le32_to_cpu(s.v->parent)); + if (bkey_val_bytes(s.k) > offsetof(struct bch_subvolume, creation_parent)) { + prt_printf(out, " creation_parent %u", le32_to_cpu(s.v->creation_parent)); + prt_printf(out, " fs_parent %u", le32_to_cpu(s.v->fs_path_parent)); + } +} + +static int subvolume_children_mod(struct btree_trans *trans, struct bpos pos, bool set) +{ + return !bpos_eq(pos, POS_MIN) + ? bch2_btree_bit_mod(trans, BTREE_ID_subvolume_children, pos, set) + : 0; +} + +int bch2_subvolume_trigger(struct btree_trans *trans, + enum btree_id btree_id, unsigned level, + struct bkey_s_c old, struct bkey_s new, + unsigned flags) +{ + if (flags & BTREE_TRIGGER_TRANSACTIONAL) { + struct bpos children_pos_old = subvolume_children_pos(old); + struct bpos children_pos_new = subvolume_children_pos(new.s_c); + + if (!bpos_eq(children_pos_old, children_pos_new)) { + int ret = subvolume_children_mod(trans, children_pos_old, false) ?: + subvolume_children_mod(trans, children_pos_new, true); + if (ret) + return ret; + } + } + + return 0; +} + +int bch2_subvol_has_children(struct btree_trans *trans, u32 subvol) +{ + struct btree_iter iter; + + bch2_trans_iter_init(trans, &iter, BTREE_ID_subvolume_children, POS(subvol, 0), 0); + struct bkey_s_c k = bch2_btree_iter_peek(&iter); + bch2_trans_iter_exit(trans, &iter); + + return bkey_err(k) ?: k.k && k.k->p.inode == subvol + ? -BCH_ERR_ENOTEMPTY_subvol_not_empty + : 0; } static __always_inline int @@ -197,8 +356,8 @@ static int bch2_subvolume_reparent(struct btree_trans *trans, if (k.k->type != KEY_TYPE_subvolume) return 0; - if (bkey_val_bytes(k.k) > offsetof(struct bch_subvolume, parent) && - le32_to_cpu(bkey_s_c_to_subvolume(k).v->parent) != old_parent) + if (bkey_val_bytes(k.k) > offsetof(struct bch_subvolume, creation_parent) && + le32_to_cpu(bkey_s_c_to_subvolume(k).v->creation_parent) != old_parent) return 0; s = bch2_bkey_make_mut_typed(trans, iter, &k, 0, subvolume); @@ -206,7 +365,7 @@ static int bch2_subvolume_reparent(struct btree_trans *trans, if (ret) return ret; - s->v.parent = cpu_to_le32(new_parent); + s->v.creation_parent = cpu_to_le32(new_parent); return 0; } @@ -229,7 +388,7 @@ static int bch2_subvolumes_reparent(struct btree_trans *trans, u32 subvolid_to_d BTREE_ID_subvolumes, POS_MIN, BTREE_ITER_PREFETCH, k, NULL, NULL, BCH_TRANS_COMMIT_no_enospc, bch2_subvolume_reparent(trans, &iter, k, - subvolid_to_delete, le32_to_cpu(s.parent))); + subvolid_to_delete, le32_to_cpu(s.creation_parent))); } /* @@ -360,6 +519,7 @@ int bch2_subvolume_unlink(struct btree_trans *trans, u32 subvolid) } int bch2_subvolume_create(struct btree_trans *trans, u64 inode, + u32 parent_subvolid, u32 src_subvolid, u32 *new_subvolid, u32 *new_snapshotid, @@ -416,12 +576,13 @@ int bch2_subvolume_create(struct btree_trans *trans, u64 inode, if (ret) goto err; - new_subvol->v.flags = 0; - new_subvol->v.snapshot = cpu_to_le32(new_nodes[0]); - new_subvol->v.inode = cpu_to_le64(inode); - new_subvol->v.parent = cpu_to_le32(src_subvolid); - new_subvol->v.otime.lo = cpu_to_le64(bch2_current_time(c)); - new_subvol->v.otime.hi = 0; + new_subvol->v.flags = 0; + new_subvol->v.snapshot = cpu_to_le32(new_nodes[0]); + new_subvol->v.inode = cpu_to_le64(inode); + new_subvol->v.creation_parent = cpu_to_le32(src_subvolid); + new_subvol->v.fs_path_parent = cpu_to_le32(parent_subvolid); + new_subvol->v.otime.lo = cpu_to_le64(bch2_current_time(c)); + new_subvol->v.otime.hi = 0; SET_BCH_SUBVOLUME_RO(&new_subvol->v, ro); SET_BCH_SUBVOLUME_SNAP(&new_subvol->v, src_subvolid != 0); diff --git a/fs/bcachefs/subvolume.h b/fs/bcachefs/subvolume.h index a6f56f66e27c..903c05162c06 100644 --- a/fs/bcachefs/subvolume.h +++ b/fs/bcachefs/subvolume.h @@ -8,17 +8,22 @@ enum bkey_invalid_flags; int bch2_check_subvols(struct bch_fs *); +int bch2_check_subvol_children(struct bch_fs *); int bch2_subvolume_invalid(struct bch_fs *, struct bkey_s_c, enum bkey_invalid_flags, struct printbuf *); void bch2_subvolume_to_text(struct printbuf *, struct bch_fs *, struct bkey_s_c); +int bch2_subvolume_trigger(struct btree_trans *, enum btree_id, unsigned, + struct bkey_s_c, struct bkey_s, unsigned); #define bch2_bkey_ops_subvolume ((struct bkey_ops) { \ .key_invalid = bch2_subvolume_invalid, \ .val_to_text = bch2_subvolume_to_text, \ + .trigger = bch2_subvolume_trigger, \ .min_val_size = 16, \ }) +int bch2_subvol_has_children(struct btree_trans *, u32); int bch2_subvolume_get(struct btree_trans *, unsigned, bool, int, struct bch_subvolume *); int bch2_subvolume_get_snapshot(struct btree_trans *, u32, u32 *); @@ -30,8 +35,7 @@ int bch2_delete_dead_snapshots(struct bch_fs *); void bch2_delete_dead_snapshots_async(struct bch_fs *); int bch2_subvolume_unlink(struct btree_trans *, u32); -int bch2_subvolume_create(struct btree_trans *, u64, u32, - u32 *, u32 *, bool); +int bch2_subvolume_create(struct btree_trans *, u64, u32, u32, u32 *, u32 *, bool); int bch2_fs_subvolumes_init(struct bch_fs *); diff --git a/fs/bcachefs/subvolume_format.h b/fs/bcachefs/subvolume_format.h index af79134b07d6..e029df7ba89f 100644 --- a/fs/bcachefs/subvolume_format.h +++ b/fs/bcachefs/subvolume_format.h @@ -19,8 +19,8 @@ struct bch_subvolume { * This is _not_ necessarily the subvolume of the directory containing * this subvolume: */ - __le32 parent; - __le32 pad; + __le32 creation_parent; + __le32 fs_path_parent; bch_le128 otime; }; diff --git a/fs/bcachefs/super-io.c b/fs/bcachefs/super-io.c index bd64eb68e84a..ad28e370b640 100644 --- a/fs/bcachefs/super-io.c +++ b/fs/bcachefs/super-io.c @@ -470,6 +470,14 @@ static int bch2_sb_validate(struct bch_sb_handle *disk_sb, struct printbuf *out, return ret; } + if (rw == WRITE && + bch2_sb_member_get(sb, sb->dev_idx).seq != sb->seq) { + prt_printf(out, "Invalid superblock: member seq %llu != sb seq %llu", + le64_to_cpu(bch2_sb_member_get(sb, sb->dev_idx).seq), + le64_to_cpu(sb->seq)); + return -BCH_ERR_invalid_sb_members_missing; + } + return 0; } @@ -717,6 +725,7 @@ retry: if (IS_ERR(sb->s_bdev_file)) { ret = PTR_ERR(sb->s_bdev_file); + prt_printf(&err, "error opening %s: %s", path, bch2_err_str(ret)); goto err; } sb->bdev = file_bdev(sb->s_bdev_file); @@ -743,9 +752,9 @@ retry: prt_printf(&err2, "bcachefs (%s): error reading default superblock: %s\n", path, err.buf); if (ret == -BCH_ERR_invalid_sb_magic && ignore_notbchfs_msg) - printk(KERN_INFO "%s", err2.buf); + bch2_print_opts(opts, KERN_INFO "%s", err2.buf); else - printk(KERN_ERR "%s", err2.buf); + bch2_print_opts(opts, KERN_ERR "%s", err2.buf); printbuf_exit(&err2); printbuf_reset(&err); @@ -803,21 +812,20 @@ got_super: goto err; } - ret = 0; sb->have_layout = true; ret = bch2_sb_validate(sb, &err, READ); if (ret) { - printk(KERN_ERR "bcachefs (%s): error validating superblock: %s\n", - path, err.buf); + bch2_print_opts(opts, KERN_ERR "bcachefs (%s): error validating superblock: %s\n", + path, err.buf); goto err_no_print; } out: printbuf_exit(&err); return ret; err: - printk(KERN_ERR "bcachefs (%s): error reading superblock: %s\n", - path, err.buf); + bch2_print_opts(opts, KERN_ERR "bcachefs (%s): error reading superblock: %s\n", + path, err.buf); err_no_print: bch2_free_super(sb); goto out; @@ -977,7 +985,7 @@ int bch2_write_super(struct bch_fs *c) prt_str(&buf, " > "); bch2_version_to_text(&buf, bcachefs_metadata_version_current); prt_str(&buf, ")"); - bch2_fs_fatal_error(c, "%s", buf.buf); + bch2_fs_fatal_error(c, ": %s", buf.buf); printbuf_exit(&buf); return -BCH_ERR_sb_not_downgraded; } @@ -997,7 +1005,7 @@ int bch2_write_super(struct bch_fs *c) if (le64_to_cpu(ca->sb_read_scratch->seq) < ca->disk_sb.seq) { bch2_fs_fatal_error(c, - "Superblock write was silently dropped! (seq %llu expected %llu)", + ": Superblock write was silently dropped! (seq %llu expected %llu)", le64_to_cpu(ca->sb_read_scratch->seq), ca->disk_sb.seq); percpu_ref_put(&ca->io_ref); @@ -1007,7 +1015,7 @@ int bch2_write_super(struct bch_fs *c) if (le64_to_cpu(ca->sb_read_scratch->seq) > ca->disk_sb.seq) { bch2_fs_fatal_error(c, - "Superblock modified by another process (seq %llu expected %llu)", + ": Superblock modified by another process (seq %llu expected %llu)", le64_to_cpu(ca->sb_read_scratch->seq), ca->disk_sb.seq); percpu_ref_put(&ca->io_ref); @@ -1058,7 +1066,7 @@ int bch2_write_super(struct bch_fs *c) !can_mount_with_written || (can_mount_without_written && !can_mount_with_written), c, - "Unable to write superblock to sufficient devices (from %ps)", + ": Unable to write superblock to sufficient devices (from %ps)", (void *) _RET_IP_)) ret = -1; out: diff --git a/fs/bcachefs/super.c b/fs/bcachefs/super.c index 6b23e11825e6..1ad6e5cd9476 100644 --- a/fs/bcachefs/super.c +++ b/fs/bcachefs/super.c @@ -56,6 +56,7 @@ #include "super.h" #include "super-io.h" #include "sysfs.h" +#include "thread_with_file.h" #include "trace.h" #include <linux/backing-dev.h> @@ -86,26 +87,38 @@ const char * const bch2_fs_flag_strs[] = { NULL }; -void __bch2_print(struct bch_fs *c, const char *fmt, ...) +__printf(2, 0) +static void bch2_print_maybe_redirect(struct stdio_redirect *stdio, const char *fmt, va_list args) { - struct stdio_redirect *stdio = bch2_fs_stdio_redirect(c); +#ifdef __KERNEL__ + if (unlikely(stdio)) { + if (fmt[0] == KERN_SOH[0]) + fmt += 2; + + bch2_stdio_redirect_vprintf(stdio, true, fmt, args); + return; + } +#endif + vprintk(fmt, args); +} + +void bch2_print_opts(struct bch_opts *opts, const char *fmt, ...) +{ + struct stdio_redirect *stdio = (void *)(unsigned long)opts->stdio; va_list args; va_start(args, fmt); - if (likely(!stdio)) { - vprintk(fmt, args); - } else { - unsigned long flags; - - if (fmt[0] == KERN_SOH[0]) - fmt += 2; + bch2_print_maybe_redirect(stdio, fmt, args); + va_end(args); +} - spin_lock_irqsave(&stdio->output_lock, flags); - prt_vprintf(&stdio->output_buf, fmt, args); - spin_unlock_irqrestore(&stdio->output_lock, flags); +void __bch2_print(struct bch_fs *c, const char *fmt, ...) +{ + struct stdio_redirect *stdio = bch2_fs_stdio_redirect(c); - wake_up(&stdio->output_wait); - } + va_list args; + va_start(args, fmt); + bch2_print_maybe_redirect(stdio, fmt, args); va_end(args); } @@ -576,7 +589,7 @@ static void __bch2_fs_free(struct bch_fs *c) destroy_workqueue(c->btree_update_wq); bch2_free_super(&c->disk_sb); - kvpfree(c, sizeof(*c)); + kvfree(c); module_put(THIS_MODULE); } @@ -715,7 +728,7 @@ static struct bch_fs *bch2_fs_alloc(struct bch_sb *sb, struct bch_opts opts) unsigned i, iter_size; int ret = 0; - c = kvpmalloc(sizeof(struct bch_fs), GFP_KERNEL|__GFP_ZERO); + c = kvmalloc(sizeof(struct bch_fs), GFP_KERNEL|__GFP_ZERO); if (!c) { c = ERR_PTR(-BCH_ERR_ENOMEM_fs_alloc); goto out; @@ -818,13 +831,13 @@ static struct bch_fs *bch2_fs_alloc(struct bch_sb *sb, struct bch_opts opts) goto err; pr_uuid(&name, c->sb.user_uuid.b); - strscpy(c->name, name.buf, sizeof(c->name)); - printbuf_exit(&name); - ret = name.allocation_failure ? -BCH_ERR_ENOMEM_fs_name_alloc : 0; if (ret) goto err; + strscpy(c->name, name.buf, sizeof(c->name)); + printbuf_exit(&name); + /* Compat: */ if (le16_to_cpu(sb->version) <= bcachefs_metadata_version_inode_v2 && !BCH_SB_JOURNAL_FLUSH_DELAY(sb)) @@ -862,13 +875,13 @@ static struct bch_fs *bch2_fs_alloc(struct bch_sb *sb, struct bch_opts opts) c->inode_shard_bits = ilog2(roundup_pow_of_two(num_possible_cpus())); if (!(c->btree_update_wq = alloc_workqueue("bcachefs", - WQ_FREEZABLE|WQ_UNBOUND|WQ_MEM_RECLAIM, 512)) || + WQ_HIGHPRI|WQ_FREEZABLE|WQ_MEM_RECLAIM|WQ_UNBOUND, 512)) || !(c->btree_io_complete_wq = alloc_workqueue("bcachefs_btree_io", - WQ_FREEZABLE|WQ_MEM_RECLAIM, 1)) || + WQ_HIGHPRI|WQ_FREEZABLE|WQ_MEM_RECLAIM, 1)) || !(c->copygc_wq = alloc_workqueue("bcachefs_copygc", - WQ_FREEZABLE|WQ_MEM_RECLAIM|WQ_CPU_INTENSIVE, 1)) || + WQ_HIGHPRI|WQ_FREEZABLE|WQ_MEM_RECLAIM|WQ_CPU_INTENSIVE, 1)) || !(c->io_complete_wq = alloc_workqueue("bcachefs_io", - WQ_FREEZABLE|WQ_HIGHPRI|WQ_MEM_RECLAIM, 512)) || + WQ_HIGHPRI|WQ_FREEZABLE|WQ_MEM_RECLAIM, 512)) || !(c->write_ref_wq = alloc_workqueue("bcachefs_write_ref", WQ_FREEZABLE, 0)) || #ifndef BCH_WRITE_REF_DEBUG @@ -882,8 +895,8 @@ static struct bch_fs *bch2_fs_alloc(struct bch_sb *sb, struct bch_opts opts) BIOSET_NEED_BVECS) || !(c->pcpu = alloc_percpu(struct bch_fs_pcpu)) || !(c->online_reserved = alloc_percpu(u64)) || - mempool_init_kvpmalloc_pool(&c->btree_bounce_pool, 1, - c->opts.btree_node_size) || + mempool_init_kvmalloc_pool(&c->btree_bounce_pool, 1, + c->opts.btree_node_size) || mempool_init_kmalloc_pool(&c->large_bkey_pool, 1, 2048) || !(c->unused_inode_hints = kcalloc(1U << c->inode_shard_bits, sizeof(u64), GFP_KERNEL))) { @@ -1061,7 +1074,8 @@ static int bch2_dev_may_add(struct bch_sb *sb, struct bch_fs *c) } static int bch2_dev_in_fs(struct bch_sb_handle *fs, - struct bch_sb_handle *sb) + struct bch_sb_handle *sb, + struct bch_opts *opts) { if (fs == sb) return 0; @@ -1102,11 +1116,14 @@ static int bch2_dev_in_fs(struct bch_sb_handle *fs, bch2_prt_datetime(&buf, le64_to_cpu(sb->sb->write_time));; prt_newline(&buf); - prt_printf(&buf, "Not using older sb"); + if (!opts->no_splitbrain_check) + prt_printf(&buf, "Not using older sb"); pr_err("%s", buf.buf); printbuf_exit(&buf); - return -BCH_ERR_device_splitbrain; + + if (!opts->no_splitbrain_check) + return -BCH_ERR_device_splitbrain; } struct bch_member m = bch2_sb_member_get(fs->sb, sb->sb->dev_idx); @@ -1124,17 +1141,22 @@ static int bch2_dev_in_fs(struct bch_sb_handle *fs, prt_newline(&buf); prt_bdevname(&buf, fs->bdev); - prt_str(&buf, "believes seq of "); + prt_str(&buf, " believes seq of "); prt_bdevname(&buf, sb->bdev); prt_printf(&buf, " to be %llu, but ", seq_from_fs); prt_bdevname(&buf, sb->bdev); prt_printf(&buf, " has %llu\n", seq_from_member); - prt_str(&buf, "Not using "); - prt_bdevname(&buf, sb->bdev); + + if (!opts->no_splitbrain_check) { + prt_str(&buf, "Not using "); + prt_bdevname(&buf, sb->bdev); + } pr_err("%s", buf.buf); printbuf_exit(&buf); - return -BCH_ERR_device_splitbrain; + + if (!opts->no_splitbrain_check) + return -BCH_ERR_device_splitbrain; } return 0; @@ -1168,8 +1190,8 @@ static void bch2_dev_free(struct bch_dev *ca) bch2_dev_buckets_free(ca); free_page((unsigned long) ca->sb_read_scratch); - bch2_time_stats_exit(&ca->io_latency[WRITE]); - bch2_time_stats_exit(&ca->io_latency[READ]); + bch2_time_stats_quantiles_exit(&ca->io_latency[WRITE]); + bch2_time_stats_quantiles_exit(&ca->io_latency[READ]); percpu_ref_exit(&ca->io_ref); percpu_ref_exit(&ca->ref); @@ -1260,8 +1282,8 @@ static struct bch_dev *__bch2_dev_alloc(struct bch_fs *c, INIT_WORK(&ca->io_error_work, bch2_io_error_work); - bch2_time_stats_init(&ca->io_latency[READ]); - bch2_time_stats_init(&ca->io_latency[WRITE]); + bch2_time_stats_quantiles_init(&ca->io_latency[READ]); + bch2_time_stats_quantiles_init(&ca->io_latency[WRITE]); ca->mi = bch2_mi_to_cpu(member); @@ -1597,27 +1619,27 @@ int bch2_dev_remove(struct bch_fs *c, struct bch_dev *ca, int flags) __bch2_dev_read_only(c, ca); ret = bch2_dev_data_drop(c, ca->dev_idx, flags); - bch_err_msg(ca, ret, "dropping data"); + bch_err_msg(ca, ret, "bch2_dev_data_drop()"); if (ret) goto err; ret = bch2_dev_remove_alloc(c, ca); - bch_err_msg(ca, ret, "deleting alloc info"); + bch_err_msg(ca, ret, "bch2_dev_remove_alloc()"); if (ret) goto err; ret = bch2_journal_flush_device_pins(&c->journal, ca->dev_idx); - bch_err_msg(ca, ret, "flushing journal"); + bch_err_msg(ca, ret, "bch2_journal_flush_device_pins()"); if (ret) goto err; ret = bch2_journal_flush(&c->journal); - bch_err(ca, "journal error"); + bch_err_msg(ca, ret, "bch2_journal_flush()"); if (ret) goto err; ret = bch2_replicas_gc2(c); - bch_err_msg(ca, ret, "in replicas_gc2()"); + bch_err_msg(ca, ret, "bch2_replicas_gc2()"); if (ret) goto err; @@ -1835,7 +1857,7 @@ int bch2_dev_online(struct bch_fs *c, const char *path) dev_idx = sb.sb->dev_idx; - ret = bch2_dev_in_fs(&c->disk_sb, &sb); + ret = bch2_dev_in_fs(&c->disk_sb, &sb, &c->opts); bch_err_msg(c, ret, "bringing %s online", path); if (ret) goto err; @@ -2023,7 +2045,7 @@ struct bch_fs *bch2_fs_open(char * const *devices, unsigned nr_devices, best = sb; darray_for_each_reverse(sbs, sb) { - ret = bch2_dev_in_fs(best, sb); + ret = bch2_dev_in_fs(best, sb, &opts); if (ret == -BCH_ERR_device_has_been_removed || ret == -BCH_ERR_device_splitbrain) { diff --git a/fs/bcachefs/sysfs.c b/fs/bcachefs/sysfs.c index cee80c47feea..c86a93a8d8fc 100644 --- a/fs/bcachefs/sysfs.c +++ b/fs/bcachefs/sysfs.c @@ -930,10 +930,10 @@ SHOW(bch2_dev) sysfs_print(io_latency_write, atomic64_read(&ca->cur_latency[WRITE])); if (attr == &sysfs_io_latency_stats_read) - bch2_time_stats_to_text(out, &ca->io_latency[READ]); + bch2_time_stats_to_text(out, &ca->io_latency[READ].stats); if (attr == &sysfs_io_latency_stats_write) - bch2_time_stats_to_text(out, &ca->io_latency[WRITE]); + bch2_time_stats_to_text(out, &ca->io_latency[WRITE].stats); sysfs_printf(congested, "%u%%", clamp(atomic_read(&ca->congested), 0, CONGESTED_MAX) diff --git a/fs/bcachefs/thread_with_file.c b/fs/bcachefs/thread_with_file.c index 9220d7de10db..940db15d6a93 100644 --- a/fs/bcachefs/thread_with_file.c +++ b/fs/bcachefs/thread_with_file.c @@ -2,7 +2,6 @@ #ifndef NO_BCACHEFS_FS #include "bcachefs.h" -#include "printbuf.h" #include "thread_with_file.h" #include <linux/anon_inodes.h> @@ -10,6 +9,7 @@ #include <linux/kthread.h> #include <linux/pagemap.h> #include <linux/poll.h> +#include <linux/sched/sysctl.h> void bch2_thread_with_file_exit(struct thread_with_file *thr) { @@ -65,68 +65,82 @@ err: return ret; } -static inline bool thread_with_stdio_has_output(struct thread_with_stdio *thr) +/* stdio_redirect */ + +static bool stdio_redirect_has_input(struct stdio_redirect *stdio) { - return thr->stdio.output_buf.pos || - thr->output2.nr || - thr->thr.done; + return stdio->input.buf.nr || stdio->done; } -static ssize_t thread_with_stdio_read(struct file *file, char __user *buf, - size_t len, loff_t *ppos) +static bool stdio_redirect_has_output(struct stdio_redirect *stdio) { - struct thread_with_stdio *thr = - container_of(file->private_data, struct thread_with_stdio, thr); - size_t copied = 0, b; - int ret = 0; + return stdio->output.buf.nr || stdio->done; +} - if ((file->f_flags & O_NONBLOCK) && - !thread_with_stdio_has_output(thr)) - return -EAGAIN; +#define STDIO_REDIRECT_BUFSIZE 4096 - ret = wait_event_interruptible(thr->stdio.output_wait, - thread_with_stdio_has_output(thr)); - if (ret) - return ret; +static bool stdio_redirect_has_input_space(struct stdio_redirect *stdio) +{ + return stdio->input.buf.nr < STDIO_REDIRECT_BUFSIZE || stdio->done; +} - if (thr->thr.done) - return 0; +static bool stdio_redirect_has_output_space(struct stdio_redirect *stdio) +{ + return stdio->output.buf.nr < STDIO_REDIRECT_BUFSIZE || stdio->done; +} - while (len) { - ret = darray_make_room(&thr->output2, thr->stdio.output_buf.pos); - if (ret) - break; +static void stdio_buf_init(struct stdio_buf *buf) +{ + spin_lock_init(&buf->lock); + init_waitqueue_head(&buf->wait); + darray_init(&buf->buf); +} - spin_lock_irq(&thr->stdio.output_lock); - b = min_t(size_t, darray_room(thr->output2), thr->stdio.output_buf.pos); +/* thread_with_stdio */ - memcpy(&darray_top(thr->output2), thr->stdio.output_buf.buf, b); - memmove(thr->stdio.output_buf.buf, - thr->stdio.output_buf.buf + b, - thr->stdio.output_buf.pos - b); +static void thread_with_stdio_done(struct thread_with_stdio *thr) +{ + thr->thr.done = true; + thr->stdio.done = true; + wake_up(&thr->stdio.input.wait); + wake_up(&thr->stdio.output.wait); +} - thr->output2.nr += b; - thr->stdio.output_buf.pos -= b; - spin_unlock_irq(&thr->stdio.output_lock); +static ssize_t thread_with_stdio_read(struct file *file, char __user *ubuf, + size_t len, loff_t *ppos) +{ + struct thread_with_stdio *thr = + container_of(file->private_data, struct thread_with_stdio, thr); + struct stdio_buf *buf = &thr->stdio.output; + size_t copied = 0, b; + int ret = 0; - b = min(len, thr->output2.nr); - if (!b) - break; + if (!(file->f_flags & O_NONBLOCK)) { + ret = wait_event_interruptible(buf->wait, stdio_redirect_has_output(&thr->stdio)); + if (ret) + return ret; + } else if (!stdio_redirect_has_output(&thr->stdio)) + return -EAGAIN; - b -= copy_to_user(buf, thr->output2.data, b); - if (!b) { + while (len && buf->buf.nr) { + if (fault_in_writeable(ubuf, len) == len) { ret = -EFAULT; break; } - copied += b; - buf += b; - len -= b; - - memmove(thr->output2.data, - thr->output2.data + b, - thr->output2.nr - b); - thr->output2.nr -= b; + spin_lock_irq(&buf->lock); + b = min_t(size_t, len, buf->buf.nr); + + if (b && !copy_to_user_nofault(ubuf, buf->buf.data, b)) { + ubuf += b; + len -= b; + copied += b; + buf->buf.nr -= b; + memmove(buf->buf.data, + buf->buf.data + b, + buf->buf.nr); + } + spin_unlock_irq(&buf->lock); } return copied ?: ret; @@ -137,27 +151,20 @@ static int thread_with_stdio_release(struct inode *inode, struct file *file) struct thread_with_stdio *thr = container_of(file->private_data, struct thread_with_stdio, thr); + thread_with_stdio_done(thr); bch2_thread_with_file_exit(&thr->thr); - printbuf_exit(&thr->stdio.input_buf); - printbuf_exit(&thr->stdio.output_buf); - darray_exit(&thr->output2); - thr->exit(thr); + darray_exit(&thr->stdio.input.buf); + darray_exit(&thr->stdio.output.buf); + thr->ops->exit(thr); return 0; } -#define WRITE_BUFFER 4096 - -static inline bool thread_with_stdio_has_input_space(struct thread_with_stdio *thr) -{ - return thr->stdio.input_buf.pos < WRITE_BUFFER || thr->thr.done; -} - static ssize_t thread_with_stdio_write(struct file *file, const char __user *ubuf, size_t len, loff_t *ppos) { struct thread_with_stdio *thr = container_of(file->private_data, struct thread_with_stdio, thr); - struct printbuf *buf = &thr->stdio.input_buf; + struct stdio_buf *buf = &thr->stdio.input; size_t copied = 0; ssize_t ret = 0; @@ -173,29 +180,30 @@ static ssize_t thread_with_stdio_write(struct file *file, const char __user *ubu break; } - spin_lock(&thr->stdio.input_lock); - if (buf->pos < WRITE_BUFFER) - bch2_printbuf_make_room(buf, min(b, WRITE_BUFFER - buf->pos)); - b = min(len, printbuf_remaining_size(buf)); - - if (b && !copy_from_user_nofault(&buf->buf[buf->pos], ubuf, b)) { - ubuf += b; - len -= b; - copied += b; - buf->pos += b; + spin_lock(&buf->lock); + if (buf->buf.nr < STDIO_REDIRECT_BUFSIZE) + darray_make_room_gfp(&buf->buf, + min(b, STDIO_REDIRECT_BUFSIZE - buf->buf.nr), GFP_NOWAIT); + b = min(len, darray_room(buf->buf)); + + if (b && !copy_from_user_nofault(&darray_top(buf->buf), ubuf, b)) { + buf->buf.nr += b; + ubuf += b; + len -= b; + copied += b; } - spin_unlock(&thr->stdio.input_lock); + spin_unlock(&buf->lock); if (b) { - wake_up(&thr->stdio.input_wait); + wake_up(&buf->wait); } else { if ((file->f_flags & O_NONBLOCK)) { ret = -EAGAIN; break; } - ret = wait_event_interruptible(thr->stdio.input_wait, - thread_with_stdio_has_input_space(thr)); + ret = wait_event_interruptible(buf->wait, + stdio_redirect_has_input_space(&thr->stdio)); if (ret) break; } @@ -209,90 +217,233 @@ static __poll_t thread_with_stdio_poll(struct file *file, struct poll_table_stru struct thread_with_stdio *thr = container_of(file->private_data, struct thread_with_stdio, thr); - poll_wait(file, &thr->stdio.output_wait, wait); - poll_wait(file, &thr->stdio.input_wait, wait); + poll_wait(file, &thr->stdio.output.wait, wait); + poll_wait(file, &thr->stdio.input.wait, wait); __poll_t mask = 0; - if (thread_with_stdio_has_output(thr)) + if (stdio_redirect_has_output(&thr->stdio)) mask |= EPOLLIN; - if (thread_with_stdio_has_input_space(thr)) + if (stdio_redirect_has_input_space(&thr->stdio)) mask |= EPOLLOUT; if (thr->thr.done) mask |= EPOLLHUP|EPOLLERR; return mask; } +static __poll_t thread_with_stdout_poll(struct file *file, struct poll_table_struct *wait) +{ + struct thread_with_stdio *thr = + container_of(file->private_data, struct thread_with_stdio, thr); + + poll_wait(file, &thr->stdio.output.wait, wait); + + __poll_t mask = 0; + + if (stdio_redirect_has_output(&thr->stdio)) + mask |= EPOLLIN; + if (thr->thr.done) + mask |= EPOLLHUP|EPOLLERR; + return mask; +} + +static int thread_with_stdio_flush(struct file *file, fl_owner_t id) +{ + struct thread_with_stdio *thr = + container_of(file->private_data, struct thread_with_stdio, thr); + + return thr->thr.ret; +} + +static long thread_with_stdio_ioctl(struct file *file, unsigned int cmd, unsigned long p) +{ + struct thread_with_stdio *thr = + container_of(file->private_data, struct thread_with_stdio, thr); + + if (thr->ops->unlocked_ioctl) + return thr->ops->unlocked_ioctl(thr, cmd, p); + return -ENOTTY; +} + static const struct file_operations thread_with_stdio_fops = { - .release = thread_with_stdio_release, + .llseek = no_llseek, .read = thread_with_stdio_read, .write = thread_with_stdio_write, .poll = thread_with_stdio_poll, + .flush = thread_with_stdio_flush, + .release = thread_with_stdio_release, + .unlocked_ioctl = thread_with_stdio_ioctl, +}; + +static const struct file_operations thread_with_stdout_fops = { .llseek = no_llseek, + .read = thread_with_stdio_read, + .poll = thread_with_stdout_poll, + .flush = thread_with_stdio_flush, + .release = thread_with_stdio_release, + .unlocked_ioctl = thread_with_stdio_ioctl, }; +static int thread_with_stdio_fn(void *arg) +{ + struct thread_with_stdio *thr = arg; + + thr->thr.ret = thr->ops->fn(thr); + + thread_with_stdio_done(thr); + return 0; +} + int bch2_run_thread_with_stdio(struct thread_with_stdio *thr, - void (*exit)(struct thread_with_stdio *), - int (*fn)(void *)) + const struct thread_with_stdio_ops *ops) { - thr->stdio.input_buf = PRINTBUF; - thr->stdio.input_buf.atomic++; - spin_lock_init(&thr->stdio.input_lock); - init_waitqueue_head(&thr->stdio.input_wait); + stdio_buf_init(&thr->stdio.input); + stdio_buf_init(&thr->stdio.output); + thr->ops = ops; - thr->stdio.output_buf = PRINTBUF; - thr->stdio.output_buf.atomic++; - spin_lock_init(&thr->stdio.output_lock); - init_waitqueue_head(&thr->stdio.output_wait); + return bch2_run_thread_with_file(&thr->thr, &thread_with_stdio_fops, thread_with_stdio_fn); +} - darray_init(&thr->output2); - thr->exit = exit; +int bch2_run_thread_with_stdout(struct thread_with_stdio *thr, + const struct thread_with_stdio_ops *ops) +{ + stdio_buf_init(&thr->stdio.input); + stdio_buf_init(&thr->stdio.output); + thr->ops = ops; - return bch2_run_thread_with_file(&thr->thr, &thread_with_stdio_fops, fn); + return bch2_run_thread_with_file(&thr->thr, &thread_with_stdout_fops, thread_with_stdio_fn); } +EXPORT_SYMBOL_GPL(bch2_run_thread_with_stdout); -int bch2_stdio_redirect_read(struct stdio_redirect *stdio, char *buf, size_t len) +int bch2_stdio_redirect_read(struct stdio_redirect *stdio, char *ubuf, size_t len) { - wait_event(stdio->input_wait, - stdio->input_buf.pos || stdio->done); + struct stdio_buf *buf = &stdio->input; + + /* + * we're waiting on user input (or for the file descriptor to be + * closed), don't want a hung task warning: + */ + do { + wait_event_timeout(buf->wait, stdio_redirect_has_input(stdio), + sysctl_hung_task_timeout_secs * HZ / 2); + } while (!stdio_redirect_has_input(stdio)); if (stdio->done) return -1; - spin_lock(&stdio->input_lock); - int ret = min(len, stdio->input_buf.pos); - stdio->input_buf.pos -= ret; - memcpy(buf, stdio->input_buf.buf, ret); - memmove(stdio->input_buf.buf, - stdio->input_buf.buf + ret, - stdio->input_buf.pos); - spin_unlock(&stdio->input_lock); + spin_lock(&buf->lock); + int ret = min(len, buf->buf.nr); + buf->buf.nr -= ret; + memcpy(ubuf, buf->buf.data, ret); + memmove(buf->buf.data, + buf->buf.data + ret, + buf->buf.nr); + spin_unlock(&buf->lock); - wake_up(&stdio->input_wait); + wake_up(&buf->wait); return ret; } -int bch2_stdio_redirect_readline(struct stdio_redirect *stdio, char *buf, size_t len) +int bch2_stdio_redirect_readline(struct stdio_redirect *stdio, char *ubuf, size_t len) { - wait_event(stdio->input_wait, - stdio->input_buf.pos || stdio->done); - - if (stdio->done) - return -1; + struct stdio_buf *buf = &stdio->input; + size_t copied = 0; + ssize_t ret = 0; +again: + do { + wait_event_timeout(buf->wait, stdio_redirect_has_input(stdio), + sysctl_hung_task_timeout_secs * HZ / 2); + } while (!stdio_redirect_has_input(stdio)); + + if (stdio->done) { + ret = -1; + goto out; + } - spin_lock(&stdio->input_lock); - int ret = min(len, stdio->input_buf.pos); - char *n = memchr(stdio->input_buf.buf, '\n', ret); + spin_lock(&buf->lock); + size_t b = min(len, buf->buf.nr); + char *n = memchr(buf->buf.data, '\n', b); if (n) - ret = min(ret, n + 1 - stdio->input_buf.buf); - stdio->input_buf.pos -= ret; - memcpy(buf, stdio->input_buf.buf, ret); - memmove(stdio->input_buf.buf, - stdio->input_buf.buf + ret, - stdio->input_buf.pos); - spin_unlock(&stdio->input_lock); - - wake_up(&stdio->input_wait); + b = min_t(size_t, b, n + 1 - buf->buf.data); + buf->buf.nr -= b; + memcpy(ubuf, buf->buf.data, b); + memmove(buf->buf.data, + buf->buf.data + b, + buf->buf.nr); + ubuf += b; + len -= b; + copied += b; + spin_unlock(&buf->lock); + + wake_up(&buf->wait); + + if (!n && len) + goto again; +out: + return copied ?: ret; +} + +__printf(3, 0) +static ssize_t bch2_darray_vprintf(darray_char *out, gfp_t gfp, const char *fmt, va_list args) +{ + ssize_t ret; + + do { + va_list args2; + size_t len; + + va_copy(args2, args); + len = vsnprintf(out->data + out->nr, darray_room(*out), fmt, args2); + va_end(args2); + + if (len + 1 <= darray_room(*out)) { + out->nr += len; + return len; + } + + ret = darray_make_room_gfp(out, len + 1, gfp); + } while (ret == 0); + + return ret; +} + +ssize_t bch2_stdio_redirect_vprintf(struct stdio_redirect *stdio, bool nonblocking, + const char *fmt, va_list args) +{ + struct stdio_buf *buf = &stdio->output; + unsigned long flags; + ssize_t ret; + +again: + spin_lock_irqsave(&buf->lock, flags); + ret = bch2_darray_vprintf(&buf->buf, GFP_NOWAIT, fmt, args); + spin_unlock_irqrestore(&buf->lock, flags); + + if (ret < 0) { + if (nonblocking) + return -EAGAIN; + + ret = wait_event_interruptible(buf->wait, + stdio_redirect_has_output_space(stdio)); + if (ret) + return ret; + goto again; + } + + wake_up(&buf->wait); + return ret; +} + +ssize_t bch2_stdio_redirect_printf(struct stdio_redirect *stdio, bool nonblocking, + const char *fmt, ...) +{ + va_list args; + ssize_t ret; + + va_start(args, fmt); + ret = bch2_stdio_redirect_vprintf(stdio, nonblocking, fmt, args); + va_end(args); + return ret; } diff --git a/fs/bcachefs/thread_with_file.h b/fs/bcachefs/thread_with_file.h index 05879c5048c8..af54ea8f5b0f 100644 --- a/fs/bcachefs/thread_with_file.h +++ b/fs/bcachefs/thread_with_file.h @@ -4,6 +4,38 @@ #include "thread_with_file_types.h" +/* + * Thread with file: Run a kthread and connect it to a file descriptor, so that + * it can be interacted with via fd read/write methods and closing the file + * descriptor stops the kthread. + * + * We have two different APIs: + * + * thread_with_file, the low level version. + * You get to define the full file_operations, including your release function, + * which means that you must call bch2_thread_with_file_exit() from your + * .release method + * + * thread_with_stdio, the higher level version + * This implements full piping of input and output, including .poll. + * + * Notes on behaviour: + * - kthread shutdown behaves like writing or reading from a pipe that has been + * closed + * - Input and output buffers are 4096 bytes, although buffers may in some + * situations slightly exceed that limit so as to avoid chopping off a + * message in the middle in nonblocking mode. + * - Input/output buffers are lazily allocated, with GFP_NOWAIT allocations - + * should be fine but might change in future revisions. + * - Output buffer may grow past 4096 bytes to deal with messages that are + * bigger than 4096 bytes + * - Writing may be done blocking or nonblocking; in nonblocking mode, we only + * drop entire messages. + * + * To write, use stdio_redirect_printf() + * To read, use stdio_redirect_read() or stdio_redirect_readline() + */ + struct task_struct; struct thread_with_file { @@ -17,25 +49,28 @@ int bch2_run_thread_with_file(struct thread_with_file *, const struct file_operations *, int (*fn)(void *)); +struct thread_with_stdio; + +struct thread_with_stdio_ops { + void (*exit)(struct thread_with_stdio *); + int (*fn)(struct thread_with_stdio *); + long (*unlocked_ioctl)(struct thread_with_stdio *, unsigned int, unsigned long); +}; + struct thread_with_stdio { struct thread_with_file thr; struct stdio_redirect stdio; - DARRAY(char) output2; - void (*exit)(struct thread_with_stdio *); + const struct thread_with_stdio_ops *ops; }; -static inline void thread_with_stdio_done(struct thread_with_stdio *thr) -{ - thr->thr.done = true; - thr->stdio.done = true; - wake_up(&thr->stdio.input_wait); - wake_up(&thr->stdio.output_wait); -} - int bch2_run_thread_with_stdio(struct thread_with_stdio *, - void (*exit)(struct thread_with_stdio *), - int (*fn)(void *)); + const struct thread_with_stdio_ops *); +int bch2_run_thread_with_stdout(struct thread_with_stdio *, + const struct thread_with_stdio_ops *); int bch2_stdio_redirect_read(struct stdio_redirect *, char *, size_t); int bch2_stdio_redirect_readline(struct stdio_redirect *, char *, size_t); +__printf(3, 0) ssize_t bch2_stdio_redirect_vprintf(struct stdio_redirect *, bool, const char *, va_list); +__printf(3, 4) ssize_t bch2_stdio_redirect_printf(struct stdio_redirect *, bool, const char *, ...); + #endif /* _BCACHEFS_THREAD_WITH_FILE_H */ diff --git a/fs/bcachefs/thread_with_file_types.h b/fs/bcachefs/thread_with_file_types.h index 90b5e645e98c..e0daf4eec341 100644 --- a/fs/bcachefs/thread_with_file_types.h +++ b/fs/bcachefs/thread_with_file_types.h @@ -2,14 +2,21 @@ #ifndef _BCACHEFS_THREAD_WITH_FILE_TYPES_H #define _BCACHEFS_THREAD_WITH_FILE_TYPES_H +#include "darray.h" + +struct stdio_buf { + spinlock_t lock; + wait_queue_head_t wait; + darray_char buf; +}; + struct stdio_redirect { - spinlock_t output_lock; - wait_queue_head_t output_wait; - struct printbuf output_buf; + struct stdio_buf input; + struct stdio_buf output; spinlock_t input_lock; wait_queue_head_t input_wait; - struct printbuf input_buf; + darray_char input_buf; bool done; }; diff --git a/fs/bcachefs/time_stats.c b/fs/bcachefs/time_stats.c new file mode 100644 index 000000000000..4508e9dcbee2 --- /dev/null +++ b/fs/bcachefs/time_stats.c @@ -0,0 +1,165 @@ +// SPDX-License-Identifier: GPL-2.0 + +#include <linux/jiffies.h> +#include <linux/module.h> +#include <linux/percpu.h> +#include <linux/preempt.h> +#include <linux/time.h> +#include <linux/spinlock.h> + +#include "eytzinger.h" +#include "time_stats.h" + +static const struct time_unit time_units[] = { + { "ns", 1 }, + { "us", NSEC_PER_USEC }, + { "ms", NSEC_PER_MSEC }, + { "s", NSEC_PER_SEC }, + { "m", (u64) NSEC_PER_SEC * 60}, + { "h", (u64) NSEC_PER_SEC * 3600}, + { "d", (u64) NSEC_PER_SEC * 3600 * 24}, + { "w", (u64) NSEC_PER_SEC * 3600 * 24 * 7}, + { "y", (u64) NSEC_PER_SEC * ((3600 * 24 * 7 * 365) + (3600 * (24 / 4) * 7))}, /* 365.25d */ + { "eon", U64_MAX }, +}; + +const struct time_unit *bch2_pick_time_units(u64 ns) +{ + const struct time_unit *u; + + for (u = time_units; + u + 1 < time_units + ARRAY_SIZE(time_units) && + ns >= u[1].nsecs << 1; + u++) + ; + + return u; +} + +static void quantiles_update(struct quantiles *q, u64 v) +{ + unsigned i = 0; + + while (i < ARRAY_SIZE(q->entries)) { + struct quantile_entry *e = q->entries + i; + + if (unlikely(!e->step)) { + e->m = v; + e->step = max_t(unsigned, v / 2, 1024); + } else if (e->m > v) { + e->m = e->m >= e->step + ? e->m - e->step + : 0; + } else if (e->m < v) { + e->m = e->m + e->step > e->m + ? e->m + e->step + : U32_MAX; + } + + if ((e->m > v ? e->m - v : v - e->m) < e->step) + e->step = max_t(unsigned, e->step / 2, 1); + + if (v >= e->m) + break; + + i = eytzinger0_child(i, v > e->m); + } +} + +static inline void time_stats_update_one(struct bch2_time_stats *stats, + u64 start, u64 end) +{ + u64 duration, freq; + bool initted = stats->last_event != 0; + + if (time_after64(end, start)) { + struct quantiles *quantiles = time_stats_to_quantiles(stats); + + duration = end - start; + mean_and_variance_update(&stats->duration_stats, duration); + mean_and_variance_weighted_update(&stats->duration_stats_weighted, + duration, initted, TIME_STATS_MV_WEIGHT); + stats->max_duration = max(stats->max_duration, duration); + stats->min_duration = min(stats->min_duration, duration); + stats->total_duration += duration; + + if (quantiles) + quantiles_update(quantiles, duration); + } + + if (stats->last_event && time_after64(end, stats->last_event)) { + freq = end - stats->last_event; + mean_and_variance_update(&stats->freq_stats, freq); + mean_and_variance_weighted_update(&stats->freq_stats_weighted, + freq, initted, TIME_STATS_MV_WEIGHT); + stats->max_freq = max(stats->max_freq, freq); + stats->min_freq = min(stats->min_freq, freq); + } + + stats->last_event = end; +} + +void __bch2_time_stats_clear_buffer(struct bch2_time_stats *stats, + struct time_stat_buffer *b) +{ + for (struct time_stat_buffer_entry *i = b->entries; + i < b->entries + ARRAY_SIZE(b->entries); + i++) + time_stats_update_one(stats, i->start, i->end); + b->nr = 0; +} + +static noinline void time_stats_clear_buffer(struct bch2_time_stats *stats, + struct time_stat_buffer *b) +{ + unsigned long flags; + + spin_lock_irqsave(&stats->lock, flags); + __bch2_time_stats_clear_buffer(stats, b); + spin_unlock_irqrestore(&stats->lock, flags); +} + +void __bch2_time_stats_update(struct bch2_time_stats *stats, u64 start, u64 end) +{ + unsigned long flags; + + if (!stats->buffer) { + spin_lock_irqsave(&stats->lock, flags); + time_stats_update_one(stats, start, end); + + if (mean_and_variance_weighted_get_mean(stats->freq_stats_weighted, TIME_STATS_MV_WEIGHT) < 32 && + stats->duration_stats.n > 1024) + stats->buffer = + alloc_percpu_gfp(struct time_stat_buffer, + GFP_ATOMIC); + spin_unlock_irqrestore(&stats->lock, flags); + } else { + struct time_stat_buffer *b; + + preempt_disable(); + b = this_cpu_ptr(stats->buffer); + + BUG_ON(b->nr >= ARRAY_SIZE(b->entries)); + b->entries[b->nr++] = (struct time_stat_buffer_entry) { + .start = start, + .end = end + }; + + if (unlikely(b->nr == ARRAY_SIZE(b->entries))) + time_stats_clear_buffer(stats, b); + preempt_enable(); + } +} + +void bch2_time_stats_exit(struct bch2_time_stats *stats) +{ + free_percpu(stats->buffer); +} + +void bch2_time_stats_init(struct bch2_time_stats *stats) +{ + memset(stats, 0, sizeof(*stats)); + stats->min_duration = U64_MAX; + stats->min_freq = U64_MAX; + spin_lock_init(&stats->lock); +} diff --git a/fs/bcachefs/time_stats.h b/fs/bcachefs/time_stats.h new file mode 100644 index 000000000000..5df61403744b --- /dev/null +++ b/fs/bcachefs/time_stats.h @@ -0,0 +1,159 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * bch2_time_stats - collect statistics on events that have a duration, with nicely + * formatted textual output on demand + * + * - percpu buffering of event collection: cheap enough to shotgun + * everywhere without worrying about overhead + * + * tracks: + * - number of events + * - maximum event duration ever seen + * - sum of all event durations + * - average event duration, standard and weighted + * - standard deviation of event durations, standard and weighted + * and analagous statistics for the frequency of events + * + * We provide both mean and weighted mean (exponentially weighted), and standard + * deviation and weighted standard deviation, to give an efficient-to-compute + * view of current behaviour versus. average behaviour - "did this event source + * just become wonky, or is this typical?". + * + * Particularly useful for tracking down latency issues. + */ +#ifndef _BCACHEFS_TIME_STATS_H +#define _BCACHEFS_TIME_STATS_H + +#include <linux/sched/clock.h> +#include <linux/spinlock_types.h> +#include <linux/string.h> + +#include "mean_and_variance.h" + +struct time_unit { + const char *name; + u64 nsecs; +}; + +/* + * given a nanosecond value, pick the preferred time units for printing: + */ +const struct time_unit *bch2_pick_time_units(u64 ns); + +/* + * quantiles - do not use: + * + * Only enabled if bch2_time_stats->quantiles_enabled has been manually set - don't + * use in new code. + */ + +#define NR_QUANTILES 15 +#define QUANTILE_IDX(i) inorder_to_eytzinger0(i, NR_QUANTILES) +#define QUANTILE_FIRST eytzinger0_first(NR_QUANTILES) +#define QUANTILE_LAST eytzinger0_last(NR_QUANTILES) + +struct quantiles { + struct quantile_entry { + u64 m; + u64 step; + } entries[NR_QUANTILES]; +}; + +struct time_stat_buffer { + unsigned nr; + struct time_stat_buffer_entry { + u64 start; + u64 end; + } entries[31]; +}; + +struct bch2_time_stats { + spinlock_t lock; + bool have_quantiles; + /* all fields are in nanoseconds */ + u64 min_duration; + u64 max_duration; + u64 total_duration; + u64 max_freq; + u64 min_freq; + u64 last_event; + u64 last_event_start; + + struct mean_and_variance duration_stats; + struct mean_and_variance freq_stats; + +/* default weight for weighted mean and variance calculations */ +#define TIME_STATS_MV_WEIGHT 8 + + struct mean_and_variance_weighted duration_stats_weighted; + struct mean_and_variance_weighted freq_stats_weighted; + struct time_stat_buffer __percpu *buffer; +}; + +struct bch2_time_stats_quantiles { + struct bch2_time_stats stats; + struct quantiles quantiles; +}; + +static inline struct quantiles *time_stats_to_quantiles(struct bch2_time_stats *stats) +{ + return stats->have_quantiles + ? &container_of(stats, struct bch2_time_stats_quantiles, stats)->quantiles + : NULL; +} + +void __bch2_time_stats_clear_buffer(struct bch2_time_stats *, struct time_stat_buffer *); +void __bch2_time_stats_update(struct bch2_time_stats *stats, u64, u64); + +/** + * time_stats_update - collect a new event being tracked + * + * @stats - bch2_time_stats to update + * @start - start time of event, recorded with local_clock() + * + * The end duration of the event will be the current time + */ +static inline void bch2_time_stats_update(struct bch2_time_stats *stats, u64 start) +{ + __bch2_time_stats_update(stats, start, local_clock()); +} + +/** + * track_event_change - track state change events + * + * @stats - bch2_time_stats to update + * @v - new state, true or false + * + * Use this when tracking time stats for state changes, i.e. resource X becoming + * blocked/unblocked. + */ +static inline bool track_event_change(struct bch2_time_stats *stats, bool v) +{ + if (v != !!stats->last_event_start) { + if (!v) { + bch2_time_stats_update(stats, stats->last_event_start); + stats->last_event_start = 0; + } else { + stats->last_event_start = local_clock() ?: 1; + return true; + } + } + + return false; +} + +void bch2_time_stats_exit(struct bch2_time_stats *); +void bch2_time_stats_init(struct bch2_time_stats *); + +static inline void bch2_time_stats_quantiles_exit(struct bch2_time_stats_quantiles *statq) +{ + bch2_time_stats_exit(&statq->stats); +} +static inline void bch2_time_stats_quantiles_init(struct bch2_time_stats_quantiles *statq) +{ + bch2_time_stats_init(&statq->stats); + statq->stats.have_quantiles = true; + memset(&statq->quantiles, 0, sizeof(statq->quantiles)); +} + +#endif /* _BCACHEFS_TIME_STATS_H */ diff --git a/fs/bcachefs/trace.h b/fs/bcachefs/trace.h index 293b90d704fb..6aa81d1e6d36 100644 --- a/fs/bcachefs/trace.h +++ b/fs/bcachefs/trace.h @@ -1431,6 +1431,25 @@ DEFINE_EVENT(fs_str, data_update, TP_ARGS(c, str) ); +TRACE_EVENT(error_downcast, + TP_PROTO(int bch_err, int std_err, unsigned long ip), + TP_ARGS(bch_err, std_err, ip), + + TP_STRUCT__entry( + __array(char, bch_err, 32 ) + __array(char, std_err, 32 ) + __array(char, ip, 32 ) + ), + + TP_fast_assign( + strscpy(__entry->bch_err, bch2_err_str(bch_err), sizeof(__entry->bch_err)); + strscpy(__entry->std_err, bch2_err_str(std_err), sizeof(__entry->std_err)); + snprintf(__entry->ip, sizeof(__entry->ip), "%ps", (void *) ip); + ), + + TP_printk("%s -> %s %s", __entry->bch_err, __entry->std_err, __entry->ip) +); + #endif /* _TRACE_BCACHEFS_H */ /* This part must be outside protection */ diff --git a/fs/bcachefs/util.c b/fs/bcachefs/util.c index 3a32faa86b5c..216fadf16928 100644 --- a/fs/bcachefs/util.c +++ b/fs/bcachefs/util.c @@ -337,157 +337,16 @@ void bch2_prt_datetime(struct printbuf *out, time64_t sec) } #endif -static const struct time_unit { - const char *name; - u64 nsecs; -} time_units[] = { - { "ns", 1 }, - { "us", NSEC_PER_USEC }, - { "ms", NSEC_PER_MSEC }, - { "s", NSEC_PER_SEC }, - { "m", (u64) NSEC_PER_SEC * 60}, - { "h", (u64) NSEC_PER_SEC * 3600}, - { "eon", U64_MAX }, -}; - -static const struct time_unit *pick_time_units(u64 ns) -{ - const struct time_unit *u; - - for (u = time_units; - u + 1 < time_units + ARRAY_SIZE(time_units) && - ns >= u[1].nsecs << 1; - u++) - ; - - return u; -} - void bch2_pr_time_units(struct printbuf *out, u64 ns) { - const struct time_unit *u = pick_time_units(ns); + const struct time_unit *u = bch2_pick_time_units(ns); prt_printf(out, "%llu %s", div_u64(ns, u->nsecs), u->name); } -/* time stats: */ - -#ifndef CONFIG_BCACHEFS_NO_LATENCY_ACCT -static void bch2_quantiles_update(struct bch2_quantiles *q, u64 v) -{ - unsigned i = 0; - - while (i < ARRAY_SIZE(q->entries)) { - struct bch2_quantile_entry *e = q->entries + i; - - if (unlikely(!e->step)) { - e->m = v; - e->step = max_t(unsigned, v / 2, 1024); - } else if (e->m > v) { - e->m = e->m >= e->step - ? e->m - e->step - : 0; - } else if (e->m < v) { - e->m = e->m + e->step > e->m - ? e->m + e->step - : U32_MAX; - } - - if ((e->m > v ? e->m - v : v - e->m) < e->step) - e->step = max_t(unsigned, e->step / 2, 1); - - if (v >= e->m) - break; - - i = eytzinger0_child(i, v > e->m); - } -} - -static inline void bch2_time_stats_update_one(struct bch2_time_stats *stats, - u64 start, u64 end) -{ - u64 duration, freq; - - if (time_after64(end, start)) { - duration = end - start; - mean_and_variance_update(&stats->duration_stats, duration); - mean_and_variance_weighted_update(&stats->duration_stats_weighted, duration); - stats->max_duration = max(stats->max_duration, duration); - stats->min_duration = min(stats->min_duration, duration); - stats->total_duration += duration; - bch2_quantiles_update(&stats->quantiles, duration); - } - - if (stats->last_event && time_after64(end, stats->last_event)) { - freq = end - stats->last_event; - mean_and_variance_update(&stats->freq_stats, freq); - mean_and_variance_weighted_update(&stats->freq_stats_weighted, freq); - stats->max_freq = max(stats->max_freq, freq); - stats->min_freq = min(stats->min_freq, freq); - } - - stats->last_event = end; -} - -static void __bch2_time_stats_clear_buffer(struct bch2_time_stats *stats, - struct bch2_time_stat_buffer *b) -{ - for (struct bch2_time_stat_buffer_entry *i = b->entries; - i < b->entries + ARRAY_SIZE(b->entries); - i++) - bch2_time_stats_update_one(stats, i->start, i->end); - b->nr = 0; -} - -static noinline void bch2_time_stats_clear_buffer(struct bch2_time_stats *stats, - struct bch2_time_stat_buffer *b) -{ - unsigned long flags; - - spin_lock_irqsave(&stats->lock, flags); - __bch2_time_stats_clear_buffer(stats, b); - spin_unlock_irqrestore(&stats->lock, flags); -} - -void __bch2_time_stats_update(struct bch2_time_stats *stats, u64 start, u64 end) -{ - unsigned long flags; - - WARN_ONCE(!stats->duration_stats_weighted.weight || - !stats->freq_stats_weighted.weight, - "uninitialized time_stats"); - - if (!stats->buffer) { - spin_lock_irqsave(&stats->lock, flags); - bch2_time_stats_update_one(stats, start, end); - - if (mean_and_variance_weighted_get_mean(stats->freq_stats_weighted) < 32 && - stats->duration_stats.n > 1024) - stats->buffer = - alloc_percpu_gfp(struct bch2_time_stat_buffer, - GFP_ATOMIC); - spin_unlock_irqrestore(&stats->lock, flags); - } else { - struct bch2_time_stat_buffer *b; - - preempt_disable(); - b = this_cpu_ptr(stats->buffer); - - BUG_ON(b->nr >= ARRAY_SIZE(b->entries)); - b->entries[b->nr++] = (struct bch2_time_stat_buffer_entry) { - .start = start, - .end = end - }; - - if (unlikely(b->nr == ARRAY_SIZE(b->entries))) - bch2_time_stats_clear_buffer(stats, b); - preempt_enable(); - } -} - static void bch2_pr_time_units_aligned(struct printbuf *out, u64 ns) { - const struct time_unit *u = pick_time_units(ns); + const struct time_unit *u = bch2_pick_time_units(ns); prt_printf(out, "%llu ", div64_u64(ns, u->nsecs)); prt_tab_rjust(out); @@ -506,10 +365,9 @@ static inline void pr_name_and_units(struct printbuf *out, const char *name, u64 void bch2_time_stats_to_text(struct printbuf *out, struct bch2_time_stats *stats) { - const struct time_unit *u; + struct quantiles *quantiles = time_stats_to_quantiles(stats); s64 f_mean = 0, d_mean = 0; - u64 q, last_q = 0, f_stddev = 0, d_stddev = 0; - int i; + u64 f_stddev = 0, d_stddev = 0; if (stats->buffer) { int cpu; @@ -571,14 +429,14 @@ void bch2_time_stats_to_text(struct printbuf *out, struct bch2_time_stats *stats prt_tab(out); bch2_pr_time_units_aligned(out, d_mean); prt_tab(out); - bch2_pr_time_units_aligned(out, mean_and_variance_weighted_get_mean(stats->duration_stats_weighted)); + bch2_pr_time_units_aligned(out, mean_and_variance_weighted_get_mean(stats->duration_stats_weighted, TIME_STATS_MV_WEIGHT)); prt_newline(out); prt_printf(out, "stddev:"); prt_tab(out); bch2_pr_time_units_aligned(out, d_stddev); prt_tab(out); - bch2_pr_time_units_aligned(out, mean_and_variance_weighted_get_stddev(stats->duration_stats_weighted)); + bch2_pr_time_units_aligned(out, mean_and_variance_weighted_get_stddev(stats->duration_stats_weighted, TIME_STATS_MV_WEIGHT)); printbuf_indent_sub(out, 2); prt_newline(out); @@ -594,53 +452,38 @@ void bch2_time_stats_to_text(struct printbuf *out, struct bch2_time_stats *stats prt_tab(out); bch2_pr_time_units_aligned(out, f_mean); prt_tab(out); - bch2_pr_time_units_aligned(out, mean_and_variance_weighted_get_mean(stats->freq_stats_weighted)); + bch2_pr_time_units_aligned(out, mean_and_variance_weighted_get_mean(stats->freq_stats_weighted, TIME_STATS_MV_WEIGHT)); prt_newline(out); prt_printf(out, "stddev:"); prt_tab(out); bch2_pr_time_units_aligned(out, f_stddev); prt_tab(out); - bch2_pr_time_units_aligned(out, mean_and_variance_weighted_get_stddev(stats->freq_stats_weighted)); + bch2_pr_time_units_aligned(out, mean_and_variance_weighted_get_stddev(stats->freq_stats_weighted, TIME_STATS_MV_WEIGHT)); printbuf_indent_sub(out, 2); prt_newline(out); printbuf_tabstops_reset(out); - i = eytzinger0_first(NR_QUANTILES); - u = pick_time_units(stats->quantiles.entries[i].m); - - prt_printf(out, "quantiles (%s):\t", u->name); - eytzinger0_for_each(i, NR_QUANTILES) { - bool is_last = eytzinger0_next(i, NR_QUANTILES) == -1; - - q = max(stats->quantiles.entries[i].m, last_q); - prt_printf(out, "%llu ", - div_u64(q, u->nsecs)); - if (is_last) - prt_newline(out); - last_q = q; + if (quantiles) { + int i = eytzinger0_first(NR_QUANTILES); + const struct time_unit *u = + bch2_pick_time_units(quantiles->entries[i].m); + u64 last_q = 0; + + prt_printf(out, "quantiles (%s):\t", u->name); + eytzinger0_for_each(i, NR_QUANTILES) { + bool is_last = eytzinger0_next(i, NR_QUANTILES) == -1; + + u64 q = max(quantiles->entries[i].m, last_q); + prt_printf(out, "%llu ", div_u64(q, u->nsecs)); + if (is_last) + prt_newline(out); + last_q = q; + } } } -#else -void bch2_time_stats_to_text(struct printbuf *out, struct bch2_time_stats *stats) {} -#endif - -void bch2_time_stats_exit(struct bch2_time_stats *stats) -{ - free_percpu(stats->buffer); -} - -void bch2_time_stats_init(struct bch2_time_stats *stats) -{ - memset(stats, 0, sizeof(*stats)); - stats->duration_stats_weighted.weight = 8; - stats->freq_stats_weighted.weight = 8; - stats->min_duration = U64_MAX; - stats->min_freq = U64_MAX; - spin_lock_init(&stats->lock); -} /* ratelimit: */ @@ -1007,28 +850,6 @@ void sort_cmp_size(void *base, size_t num, size_t size, } } -static void mempool_free_vp(void *element, void *pool_data) -{ - size_t size = (size_t) pool_data; - - vpfree(element, size); -} - -static void *mempool_alloc_vp(gfp_t gfp_mask, void *pool_data) -{ - size_t size = (size_t) pool_data; - - return vpmalloc(size, gfp_mask); -} - -int mempool_init_kvpmalloc_pool(mempool_t *pool, int min_nr, size_t size) -{ - return size < PAGE_SIZE - ? mempool_init_kmalloc_pool(pool, min_nr, size) - : mempool_init(pool, min_nr, mempool_alloc_vp, - mempool_free_vp, (void *) size); -} - #if 0 void eytzinger1_test(void) { diff --git a/fs/bcachefs/util.h b/fs/bcachefs/util.h index b414736d59a5..175aee3074c7 100644 --- a/fs/bcachefs/util.h +++ b/fs/bcachefs/util.h @@ -21,6 +21,7 @@ #include "mean_and_variance.h" #include "darray.h" +#include "time_stats.h" struct closure; @@ -53,38 +54,6 @@ static inline size_t buf_pages(void *p, size_t len) PAGE_SIZE); } -static inline void vpfree(void *p, size_t size) -{ - if (is_vmalloc_addr(p)) - vfree(p); - else - free_pages((unsigned long) p, get_order(size)); -} - -static inline void *vpmalloc(size_t size, gfp_t gfp_mask) -{ - return (void *) __get_free_pages(gfp_mask|__GFP_NOWARN, - get_order(size)) ?: - __vmalloc(size, gfp_mask); -} - -static inline void kvpfree(void *p, size_t size) -{ - if (size < PAGE_SIZE) - kfree(p); - else - vpfree(p, size); -} - -static inline void *kvpmalloc(size_t size, gfp_t gfp_mask) -{ - return size < PAGE_SIZE - ? kmalloc(size, gfp_mask) - : vpmalloc(size, gfp_mask); -} - -int mempool_init_kvpmalloc_pool(mempool_t *, int, size_t); - #define HEAP(type) \ struct { \ size_t size, used; \ @@ -97,13 +66,13 @@ struct { \ ({ \ (heap)->used = 0; \ (heap)->size = (_size); \ - (heap)->data = kvpmalloc((heap)->size * sizeof((heap)->data[0]),\ + (heap)->data = kvmalloc((heap)->size * sizeof((heap)->data[0]),\ (gfp)); \ }) #define free_heap(heap) \ do { \ - kvpfree((heap)->data, (heap)->size * sizeof((heap)->data[0])); \ + kvfree((heap)->data); \ (heap)->data = NULL; \ } while (0) @@ -361,84 +330,8 @@ static inline void prt_bdevname(struct printbuf *out, struct block_device *bdev) #endif } -#define NR_QUANTILES 15 -#define QUANTILE_IDX(i) inorder_to_eytzinger0(i, NR_QUANTILES) -#define QUANTILE_FIRST eytzinger0_first(NR_QUANTILES) -#define QUANTILE_LAST eytzinger0_last(NR_QUANTILES) - -struct bch2_quantiles { - struct bch2_quantile_entry { - u64 m; - u64 step; - } entries[NR_QUANTILES]; -}; - -struct bch2_time_stat_buffer { - unsigned nr; - struct bch2_time_stat_buffer_entry { - u64 start; - u64 end; - } entries[32]; -}; - -struct bch2_time_stats { - spinlock_t lock; - /* all fields are in nanoseconds */ - u64 min_duration; - u64 max_duration; - u64 total_duration; - u64 max_freq; - u64 min_freq; - u64 last_event; - struct bch2_quantiles quantiles; - - struct mean_and_variance duration_stats; - struct mean_and_variance_weighted duration_stats_weighted; - struct mean_and_variance freq_stats; - struct mean_and_variance_weighted freq_stats_weighted; - struct bch2_time_stat_buffer __percpu *buffer; -}; - -#ifndef CONFIG_BCACHEFS_NO_LATENCY_ACCT -void __bch2_time_stats_update(struct bch2_time_stats *stats, u64, u64); - -static inline void bch2_time_stats_update(struct bch2_time_stats *stats, u64 start) -{ - __bch2_time_stats_update(stats, start, local_clock()); -} - -static inline bool track_event_change(struct bch2_time_stats *stats, - u64 *start, bool v) -{ - if (v != !!*start) { - if (!v) { - bch2_time_stats_update(stats, *start); - *start = 0; - } else { - *start = local_clock() ?: 1; - return true; - } - } - - return false; -} -#else -static inline void __bch2_time_stats_update(struct bch2_time_stats *stats, u64 start, u64 end) {} -static inline void bch2_time_stats_update(struct bch2_time_stats *stats, u64 start) {} -static inline bool track_event_change(struct bch2_time_stats *stats, - u64 *start, bool v) -{ - bool ret = v && !*start; - *start = v; - return ret; -} -#endif - void bch2_time_stats_to_text(struct printbuf *, struct bch2_time_stats *); -void bch2_time_stats_exit(struct bch2_time_stats *); -void bch2_time_stats_init(struct bch2_time_stats *); - #define ewma_add(ewma, val, weight) \ ({ \ typeof(ewma) _ewma = (ewma); \ @@ -788,8 +681,15 @@ static inline void __move_gap(void *array, size_t element_size, } /* Move the gap in a gap buffer: */ -#define move_gap(_array, _nr, _size, _old_gap, _new_gap) \ - __move_gap(_array, sizeof(_array[0]), _nr, _size, _old_gap, _new_gap) +#define move_gap(_d, _new_gap) \ +do { \ + BUG_ON(_new_gap > (_d)->nr); \ + BUG_ON((_d)->gap > (_d)->nr); \ + \ + __move_gap((_d)->data, sizeof((_d)->data[0]), \ + (_d)->nr, (_d)->size, (_d)->gap, _new_gap); \ + (_d)->gap = _new_gap; \ +} while (0) #define bubble_sort(_base, _nr, _cmp) \ do { \ @@ -876,4 +776,25 @@ static inline bool qstr_eq(const struct qstr l, const struct qstr r) void bch2_darray_str_exit(darray_str *); int bch2_split_devs(const char *, darray_str *); +#ifdef __KERNEL__ + +__must_check +static inline int copy_to_user_errcode(void __user *to, const void *from, unsigned long n) +{ + return copy_to_user(to, from, n) ? -EFAULT : 0; +} + +__must_check +static inline int copy_from_user_errcode(void *to, const void __user *from, unsigned long n) +{ + return copy_from_user(to, from, n) ? -EFAULT : 0; +} + +#endif + +static inline void __set_bit_le64(size_t bit, __le64 *addr) +{ + addr[bit / 64] |= cpu_to_le64(BIT_ULL(bit % 64)); +} + #endif /* _BCACHEFS_UTIL_H */ diff --git a/fs/bcachefs/xattr.c b/fs/bcachefs/xattr.c index 9c0d2316031b..754f17bba68e 100644 --- a/fs/bcachefs/xattr.c +++ b/fs/bcachefs/xattr.c @@ -544,11 +544,11 @@ static int bch2_xattr_bcachefs_set(const struct xattr_handler *handler, kfree(buf); if (ret < 0) - return ret; + goto err_class_exit; ret = bch2_opt_check_may_set(c, opt_id, v); if (ret < 0) - return ret; + goto err_class_exit; s.v = v + 1; s.defined = true; @@ -595,6 +595,7 @@ err: (opt_id == Opt_compression && !inode_opt_get(c, &inode->ei_inode, background_compression)))) bch2_set_rebalance_needs_scan(c, inode->ei_inode.bi_inum); +err_class_exit: return bch2_err_class(ret); } diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c index a2d07fa3cfdf..1dc1f1946ae0 100644 --- a/fs/btrfs/volumes.c +++ b/fs/btrfs/volumes.c @@ -1303,6 +1303,47 @@ int btrfs_forget_devices(dev_t devt) return ret; } +static bool btrfs_skip_registration(struct btrfs_super_block *disk_super, + const char *path, dev_t devt, + bool mount_arg_dev) +{ + struct btrfs_fs_devices *fs_devices; + + /* + * Do not skip device registration for mounted devices with matching + * maj:min but different paths. Booting without initrd relies on + * /dev/root initially, later replaced with the actual root device. + * A successful scan ensures grub2-probe selects the correct device. + */ + list_for_each_entry(fs_devices, &fs_uuids, fs_list) { + struct btrfs_device *device; + + mutex_lock(&fs_devices->device_list_mutex); + + if (!fs_devices->opened) { + mutex_unlock(&fs_devices->device_list_mutex); + continue; + } + + list_for_each_entry(device, &fs_devices->devices, dev_list) { + if (device->bdev && (device->bdev->bd_dev == devt) && + strcmp(device->name->str, path) != 0) { + mutex_unlock(&fs_devices->device_list_mutex); + + /* Do not skip registration. */ + return false; + } + } + mutex_unlock(&fs_devices->device_list_mutex); + } + + if (!mount_arg_dev && btrfs_super_num_devices(disk_super) == 1 && + !(btrfs_super_flags(disk_super) & BTRFS_SUPER_FLAG_SEEDING)) + return true; + + return false; +} + /* * Look for a btrfs signature on a device. This may be called out of the mount path * and we are not allowed to call set_blocksize during the scan. The superblock @@ -1320,6 +1361,7 @@ struct btrfs_device *btrfs_scan_one_device(const char *path, blk_mode_t flags, struct btrfs_device *device = NULL; struct file *bdev_file; u64 bytenr, bytenr_orig; + dev_t devt; int ret; lockdep_assert_held(&uuid_mutex); @@ -1359,19 +1401,13 @@ struct btrfs_device *btrfs_scan_one_device(const char *path, blk_mode_t flags, goto error_bdev_put; } - if (!mount_arg_dev && btrfs_super_num_devices(disk_super) == 1 && - !(btrfs_super_flags(disk_super) & BTRFS_SUPER_FLAG_SEEDING)) { - dev_t devt; + devt = file_bdev(bdev_file)->bd_dev; + if (btrfs_skip_registration(disk_super, path, devt, mount_arg_dev)) { + pr_debug("BTRFS: skip registering single non-seed device %s (%d:%d)\n", + path, MAJOR(devt), MINOR(devt)); - ret = lookup_bdev(path, &devt); - if (ret) - btrfs_warn(NULL, "lookup bdev failed for path %s: %d", - path, ret); - else - btrfs_free_stale_devices(devt, NULL); + btrfs_free_stale_devices(devt, NULL); - pr_debug("BTRFS: skip registering single non-seed device %s (%d:%d)\n", - path, MAJOR(devt), MINOR(devt)); device = NULL; goto free_disk_super; } diff --git a/fs/debugfs/inode.c b/fs/debugfs/inode.c index 034a617cb1a5..a40da0065433 100644 --- a/fs/debugfs/inode.c +++ b/fs/debugfs/inode.c @@ -751,13 +751,28 @@ static void __debugfs_file_removed(struct dentry *dentry) if ((unsigned long)fsd & DEBUGFS_FSDATA_IS_REAL_FOPS_BIT) return; - /* if we hit zero, just wait for all to finish */ - if (!refcount_dec_and_test(&fsd->active_users)) { - wait_for_completion(&fsd->active_users_drained); + /* if this was the last reference, we're done */ + if (refcount_dec_and_test(&fsd->active_users)) return; - } - /* if we didn't hit zero, try to cancel any we can */ + /* + * If there's still a reference, the code that obtained it can + * be in different states: + * - The common case of not using cancellations, or already + * after debugfs_leave_cancellation(), where we just need + * to wait for debugfs_file_put() which signals the completion; + * - inside a cancellation section, i.e. between + * debugfs_enter_cancellation() and debugfs_leave_cancellation(), + * in which case we need to trigger the ->cancel() function, + * and then wait for debugfs_file_put() just like in the + * previous case; + * - before debugfs_enter_cancellation() (but obviously after + * debugfs_file_get()), in which case we may not see the + * cancellation in the list on the first round of the loop, + * but debugfs_enter_cancellation() signals the completion + * after adding it, so this code gets woken up to call the + * ->cancel() function. + */ while (refcount_read(&fsd->active_users)) { struct debugfs_cancellation *c; diff --git a/fs/dlm/dlm_internal.h b/fs/dlm/dlm_internal.h index dfc444dad329..3b4dbce849f0 100644 --- a/fs/dlm/dlm_internal.h +++ b/fs/dlm/dlm_internal.h @@ -246,7 +246,7 @@ struct dlm_lkb { int8_t lkb_highbast; /* highest mode bast sent for */ int8_t lkb_wait_type; /* type of reply waiting for */ - atomic_t lkb_wait_count; + int8_t lkb_wait_count; int lkb_wait_nodeid; /* for debugging */ struct list_head lkb_statequeue; /* rsb g/c/w list */ diff --git a/fs/dlm/lock.c b/fs/dlm/lock.c index 652c51fbbf76..fd752dd03896 100644 --- a/fs/dlm/lock.c +++ b/fs/dlm/lock.c @@ -1407,7 +1407,6 @@ static int add_to_waiters(struct dlm_lkb *lkb, int mstype, int to_nodeid) { struct dlm_ls *ls = lkb->lkb_resource->res_ls; int error = 0; - int wc; mutex_lock(&ls->ls_waiters_mutex); @@ -1429,17 +1428,20 @@ static int add_to_waiters(struct dlm_lkb *lkb, int mstype, int to_nodeid) error = -EBUSY; goto out; } - wc = atomic_inc_return(&lkb->lkb_wait_count); + lkb->lkb_wait_count++; hold_lkb(lkb); log_debug(ls, "addwait %x cur %d overlap %d count %d f %x", - lkb->lkb_id, lkb->lkb_wait_type, mstype, wc, - dlm_iflags_val(lkb)); + lkb->lkb_id, lkb->lkb_wait_type, mstype, + lkb->lkb_wait_count, dlm_iflags_val(lkb)); goto out; } - wc = atomic_fetch_inc(&lkb->lkb_wait_count); - DLM_ASSERT(!wc, dlm_print_lkb(lkb); printk("wait_count %d\n", wc);); + DLM_ASSERT(!lkb->lkb_wait_count, + dlm_print_lkb(lkb); + printk("wait_count %d\n", lkb->lkb_wait_count);); + + lkb->lkb_wait_count++; lkb->lkb_wait_type = mstype; lkb->lkb_wait_nodeid = to_nodeid; /* for debugging */ hold_lkb(lkb); @@ -1502,7 +1504,7 @@ static int _remove_from_waiters(struct dlm_lkb *lkb, int mstype, log_debug(ls, "remwait %x convert_reply zap overlap_cancel", lkb->lkb_id); lkb->lkb_wait_type = 0; - atomic_dec(&lkb->lkb_wait_count); + lkb->lkb_wait_count--; unhold_lkb(lkb); goto out_del; } @@ -1529,15 +1531,16 @@ static int _remove_from_waiters(struct dlm_lkb *lkb, int mstype, if (overlap_done && lkb->lkb_wait_type) { log_error(ls, "remwait error %x reply %d wait_type %d overlap", lkb->lkb_id, mstype, lkb->lkb_wait_type); - atomic_dec(&lkb->lkb_wait_count); + lkb->lkb_wait_count--; unhold_lkb(lkb); lkb->lkb_wait_type = 0; } - DLM_ASSERT(atomic_read(&lkb->lkb_wait_count), dlm_print_lkb(lkb);); + DLM_ASSERT(lkb->lkb_wait_count, dlm_print_lkb(lkb);); clear_bit(DLM_IFL_RESEND_BIT, &lkb->lkb_iflags); - if (atomic_dec_and_test(&lkb->lkb_wait_count)) + lkb->lkb_wait_count--; + if (!lkb->lkb_wait_count) list_del_init(&lkb->lkb_wait_reply); unhold_lkb(lkb); return 0; @@ -2666,7 +2669,7 @@ static int validate_lock_args(struct dlm_ls *ls, struct dlm_lkb *lkb, goto out; /* lock not allowed if there's any op in progress */ - if (lkb->lkb_wait_type || atomic_read(&lkb->lkb_wait_count)) + if (lkb->lkb_wait_type || lkb->lkb_wait_count) goto out; if (is_overlap(lkb)) @@ -2728,7 +2731,7 @@ static int validate_unlock_args(struct dlm_lkb *lkb, struct dlm_args *args) /* normal unlock not allowed if there's any op in progress */ if (!(args->flags & (DLM_LKF_CANCEL | DLM_LKF_FORCEUNLOCK)) && - (lkb->lkb_wait_type || atomic_read(&lkb->lkb_wait_count))) + (lkb->lkb_wait_type || lkb->lkb_wait_count)) goto out; /* an lkb may be waiting for an rsb lookup to complete where the @@ -5011,21 +5014,32 @@ static struct dlm_lkb *find_resend_waiter(struct dlm_ls *ls) return lkb; } -/* Deal with lookups and lkb's marked RESEND from _pre. We may now be the - master or dir-node for r. Processing the lkb may result in it being placed - back on waiters. */ - -/* We do this after normal locking has been enabled and any saved messages - (in requestqueue) have been processed. We should be confident that at - this point we won't get or process a reply to any of these waiting - operations. But, new ops may be coming in on the rsbs/locks here from - userspace or remotely. */ - -/* there may have been an overlap unlock/cancel prior to recovery or after - recovery. if before, the lkb may still have a pos wait_count; if after, the - overlap flag would just have been set and nothing new sent. we can be - confident here than any replies to either the initial op or overlap ops - prior to recovery have been received. */ +/* + * Forced state reset for locks that were in the middle of remote operations + * when recovery happened (i.e. lkbs that were on the waiters list, waiting + * for a reply from a remote operation.) The lkbs remaining on the waiters + * list need to be reevaluated; some may need resending to a different node + * than previously, and some may now need local handling rather than remote. + * + * First, the lkb state for the voided remote operation is forcibly reset, + * equivalent to what remove_from_waiters() would normally do: + * . lkb removed from ls_waiters list + * . lkb wait_type cleared + * . lkb waiters_count cleared + * . lkb ref count decremented for each waiters_count (almost always 1, + * but possibly 2 in case of cancel/unlock overlapping, which means + * two remote replies were being expected for the lkb.) + * + * Second, the lkb is reprocessed like an original operation would be, + * by passing it to _request_lock or _convert_lock, which will either + * process the lkb operation locally, or send it to a remote node again + * and put the lkb back onto the waiters list. + * + * When reprocessing the lkb, we may find that it's flagged for an overlapping + * force-unlock or cancel, either from before recovery began, or after recovery + * finished. If this is the case, the unlock/cancel is done directly, and the + * original operation is not initiated again (no _request_lock/_convert_lock.) + */ int dlm_recover_waiters_post(struct dlm_ls *ls) { @@ -5040,6 +5054,11 @@ int dlm_recover_waiters_post(struct dlm_ls *ls) break; } + /* + * Find an lkb from the waiters list that's been affected by + * recovery node changes, and needs to be reprocessed. Does + * hold_lkb(), adding a refcount. + */ lkb = find_resend_waiter(ls); if (!lkb) break; @@ -5048,6 +5067,11 @@ int dlm_recover_waiters_post(struct dlm_ls *ls) hold_rsb(r); lock_rsb(r); + /* + * If the lkb has been flagged for a force unlock or cancel, + * then the reprocessing below will be replaced by just doing + * the unlock/cancel directly. + */ mstype = lkb->lkb_wait_type; oc = test_and_clear_bit(DLM_IFL_OVERLAP_CANCEL_BIT, &lkb->lkb_iflags); @@ -5061,22 +5085,40 @@ int dlm_recover_waiters_post(struct dlm_ls *ls) r->res_nodeid, lkb->lkb_nodeid, lkb->lkb_wait_nodeid, dlm_dir_nodeid(r), oc, ou); - /* At this point we assume that we won't get a reply to any - previous op or overlap op on this lock. First, do a big - remove_from_waiters() for all previous ops. */ + /* + * No reply to the pre-recovery operation will now be received, + * so a forced equivalent of remove_from_waiters() is needed to + * reset the waiters state that was in place before recovery. + */ clear_bit(DLM_IFL_RESEND_BIT, &lkb->lkb_iflags); + + /* Forcibly clear wait_type */ lkb->lkb_wait_type = 0; - /* drop all wait_count references we still - * hold a reference for this iteration. + + /* + * Forcibly reset wait_count and associated refcount. The + * wait_count will almost always be 1, but in case of an + * overlapping unlock/cancel it could be 2: see where + * add_to_waiters() finds the lkb is already on the waiters + * list and does lkb_wait_count++; hold_lkb(). */ - while (!atomic_dec_and_test(&lkb->lkb_wait_count)) + while (lkb->lkb_wait_count) { + lkb->lkb_wait_count--; unhold_lkb(lkb); + } + /* Forcibly remove from waiters list */ mutex_lock(&ls->ls_waiters_mutex); list_del_init(&lkb->lkb_wait_reply); mutex_unlock(&ls->ls_waiters_mutex); + /* + * The lkb is now clear of all prior waiters state and can be + * processed locally, or sent to remote node again, or directly + * cancelled/unlocked. + */ + if (oc || ou) { /* do an unlock or cancel instead of resending */ switch (mstype) { diff --git a/fs/dlm/user.c b/fs/dlm/user.c index 695e691b38b3..9f9b68448830 100644 --- a/fs/dlm/user.c +++ b/fs/dlm/user.c @@ -806,7 +806,7 @@ static ssize_t device_read(struct file *file, char __user *buf, size_t count, struct dlm_lkb *lkb; DECLARE_WAITQUEUE(wait, current); struct dlm_callback *cb; - int rv, copy_lvb = 0; + int rv, ret, copy_lvb = 0; int old_mode, new_mode; if (count == sizeof(struct dlm_device_version)) { @@ -906,9 +906,9 @@ static ssize_t device_read(struct file *file, char __user *buf, size_t count, trace_dlm_ast(lkb->lkb_resource->res_ls, lkb); } - rv = copy_result_to_user(lkb->lkb_ua, - test_bit(DLM_PROC_FLAGS_COMPAT, &proc->flags), - cb->flags, cb->mode, copy_lvb, buf, count); + ret = copy_result_to_user(lkb->lkb_ua, + test_bit(DLM_PROC_FLAGS_COMPAT, &proc->flags), + cb->flags, cb->mode, copy_lvb, buf, count); kref_put(&cb->ref, dlm_release_callback); @@ -916,7 +916,7 @@ static ssize_t device_read(struct file *file, char __user *buf, size_t count, if (rv == DLM_DEQUEUE_CALLBACK_LAST) dlm_put_lkb(lkb); - return rv; + return ret; } static __poll_t device_poll(struct file *file, poll_table *wait) diff --git a/fs/exfat/dir.c b/fs/exfat/dir.c index 9f9295847a4e..077944d3c2c0 100644 --- a/fs/exfat/dir.c +++ b/fs/exfat/dir.c @@ -448,88 +448,34 @@ static void exfat_init_name_entry(struct exfat_dentry *ep, } } -int exfat_init_dir_entry(struct inode *inode, struct exfat_chain *p_dir, - int entry, unsigned int type, unsigned int start_clu, - unsigned long long size) +void exfat_init_dir_entry(struct exfat_entry_set_cache *es, + unsigned int type, unsigned int start_clu, + unsigned long long size, struct timespec64 *ts) { - struct super_block *sb = inode->i_sb; + struct super_block *sb = es->sb; struct exfat_sb_info *sbi = EXFAT_SB(sb); - struct timespec64 ts = current_time(inode); struct exfat_dentry *ep; - struct buffer_head *bh; - - /* - * We cannot use exfat_get_dentry_set here because file ep is not - * initialized yet. - */ - ep = exfat_get_dentry(sb, p_dir, entry, &bh); - if (!ep) - return -EIO; + ep = exfat_get_dentry_cached(es, ES_IDX_FILE); exfat_set_entry_type(ep, type); - exfat_set_entry_time(sbi, &ts, + exfat_set_entry_time(sbi, ts, &ep->dentry.file.create_tz, &ep->dentry.file.create_time, &ep->dentry.file.create_date, &ep->dentry.file.create_time_cs); - exfat_set_entry_time(sbi, &ts, + exfat_set_entry_time(sbi, ts, &ep->dentry.file.modify_tz, &ep->dentry.file.modify_time, &ep->dentry.file.modify_date, &ep->dentry.file.modify_time_cs); - exfat_set_entry_time(sbi, &ts, + exfat_set_entry_time(sbi, ts, &ep->dentry.file.access_tz, &ep->dentry.file.access_time, &ep->dentry.file.access_date, NULL); - exfat_update_bh(bh, IS_DIRSYNC(inode)); - brelse(bh); - - ep = exfat_get_dentry(sb, p_dir, entry + 1, &bh); - if (!ep) - return -EIO; - + ep = exfat_get_dentry_cached(es, ES_IDX_STREAM); exfat_init_stream_entry(ep, start_clu, size); - exfat_update_bh(bh, IS_DIRSYNC(inode)); - brelse(bh); - - return 0; -} - -int exfat_update_dir_chksum(struct inode *inode, struct exfat_chain *p_dir, - int entry) -{ - struct super_block *sb = inode->i_sb; - int ret = 0; - int i, num_entries; - u16 chksum; - struct exfat_dentry *ep, *fep; - struct buffer_head *fbh, *bh; - - fep = exfat_get_dentry(sb, p_dir, entry, &fbh); - if (!fep) - return -EIO; - - num_entries = fep->dentry.file.num_ext + 1; - chksum = exfat_calc_chksum16(fep, DENTRY_SIZE, 0, CS_DIR_ENTRY); - - for (i = 1; i < num_entries; i++) { - ep = exfat_get_dentry(sb, p_dir, entry + i, &bh); - if (!ep) { - ret = -EIO; - goto release_fbh; - } - chksum = exfat_calc_chksum16(ep, DENTRY_SIZE, chksum, - CS_DEFAULT); - brelse(bh); - } - - fep->dentry.file.checksum = cpu_to_le16(chksum); - exfat_update_bh(fbh, IS_DIRSYNC(inode)); -release_fbh: - brelse(fbh); - return ret; } static void exfat_free_benign_secondary_clusters(struct inode *inode, @@ -551,76 +497,49 @@ static void exfat_free_benign_secondary_clusters(struct inode *inode, exfat_free_cluster(inode, &dir); } -int exfat_init_ext_entry(struct inode *inode, struct exfat_chain *p_dir, - int entry, int num_entries, struct exfat_uni_name *p_uniname) +void exfat_init_ext_entry(struct exfat_entry_set_cache *es, int num_entries, + struct exfat_uni_name *p_uniname) { - struct super_block *sb = inode->i_sb; int i; unsigned short *uniname = p_uniname->name; struct exfat_dentry *ep; - struct buffer_head *bh; - int sync = IS_DIRSYNC(inode); - - ep = exfat_get_dentry(sb, p_dir, entry, &bh); - if (!ep) - return -EIO; + ep = exfat_get_dentry_cached(es, ES_IDX_FILE); ep->dentry.file.num_ext = (unsigned char)(num_entries - 1); - exfat_update_bh(bh, sync); - brelse(bh); - - ep = exfat_get_dentry(sb, p_dir, entry + 1, &bh); - if (!ep) - return -EIO; + ep = exfat_get_dentry_cached(es, ES_IDX_STREAM); ep->dentry.stream.name_len = p_uniname->name_len; ep->dentry.stream.name_hash = cpu_to_le16(p_uniname->name_hash); - exfat_update_bh(bh, sync); - brelse(bh); - - for (i = EXFAT_FIRST_CLUSTER; i < num_entries; i++) { - ep = exfat_get_dentry(sb, p_dir, entry + i, &bh); - if (!ep) - return -EIO; - - if (exfat_get_entry_type(ep) & TYPE_BENIGN_SEC) - exfat_free_benign_secondary_clusters(inode, ep); + for (i = ES_IDX_FIRST_FILENAME; i < num_entries; i++) { + ep = exfat_get_dentry_cached(es, i); exfat_init_name_entry(ep, uniname); - exfat_update_bh(bh, sync); - brelse(bh); uniname += EXFAT_FILE_NAME_LEN; } - exfat_update_dir_chksum(inode, p_dir, entry); - return 0; + exfat_update_dir_chksum(es); } -int exfat_remove_entries(struct inode *inode, struct exfat_chain *p_dir, - int entry, int order, int num_entries) +void exfat_remove_entries(struct inode *inode, struct exfat_entry_set_cache *es, + int order) { - struct super_block *sb = inode->i_sb; int i; struct exfat_dentry *ep; - struct buffer_head *bh; - for (i = order; i < num_entries; i++) { - ep = exfat_get_dentry(sb, p_dir, entry + i, &bh); - if (!ep) - return -EIO; + for (i = order; i < es->num_entries; i++) { + ep = exfat_get_dentry_cached(es, i); if (exfat_get_entry_type(ep) & TYPE_BENIGN_SEC) exfat_free_benign_secondary_clusters(inode, ep); exfat_set_entry_type(ep, TYPE_DELETED); - exfat_update_bh(bh, IS_DIRSYNC(inode)); - brelse(bh); } - return 0; + if (order < es->num_entries) + es->modified = true; } -void exfat_update_dir_chksum_with_entry_set(struct exfat_entry_set_cache *es) +void exfat_update_dir_chksum(struct exfat_entry_set_cache *es) { int chksum_type = CS_DIR_ENTRY, i; unsigned short chksum = 0; @@ -775,7 +694,6 @@ struct exfat_dentry *exfat_get_dentry(struct super_block *sb, } enum exfat_validate_dentry_mode { - ES_MODE_STARTED, ES_MODE_GET_FILE_ENTRY, ES_MODE_GET_STRM_ENTRY, ES_MODE_GET_NAME_ENTRY, @@ -790,11 +708,6 @@ static bool exfat_validate_entry(unsigned int type, return false; switch (*mode) { - case ES_MODE_STARTED: - if (type != TYPE_FILE && type != TYPE_DIR) - return false; - *mode = ES_MODE_GET_FILE_ENTRY; - break; case ES_MODE_GET_FILE_ENTRY: if (type != TYPE_STREAM) return false; @@ -834,7 +747,7 @@ struct exfat_dentry *exfat_get_dentry_cached( } /* - * Returns a set of dentries for a file or dir. + * Returns a set of dentries. * * Note It provides a direct pointer to bh->data via exfat_get_dentry_cached(). * User should call exfat_get_dentry_set() after setting 'modified' to apply @@ -842,22 +755,24 @@ struct exfat_dentry *exfat_get_dentry_cached( * * in: * sb+p_dir+entry: indicates a file/dir - * type: specifies how many dentries should be included. + * num_entries: specifies how many dentries should be included. + * It will be set to es->num_entries if it is not 0. + * If num_entries is 0, es->num_entries will be obtained + * from the first dentry. + * out: + * es: pointer of entry set on success. * return: - * pointer of entry set on success, - * NULL on failure. + * 0 on success + * -error code on failure */ -int exfat_get_dentry_set(struct exfat_entry_set_cache *es, +static int __exfat_get_dentry_set(struct exfat_entry_set_cache *es, struct super_block *sb, struct exfat_chain *p_dir, int entry, - unsigned int type) + unsigned int num_entries) { int ret, i, num_bh; unsigned int off; sector_t sec; struct exfat_sb_info *sbi = EXFAT_SB(sb); - struct exfat_dentry *ep; - int num_entries; - enum exfat_validate_dentry_mode mode = ES_MODE_STARTED; struct buffer_head *bh; if (p_dir->dir == DIR_DELETED) { @@ -880,12 +795,18 @@ int exfat_get_dentry_set(struct exfat_entry_set_cache *es, return -EIO; es->bh[es->num_bh++] = bh; - ep = exfat_get_dentry_cached(es, ES_IDX_FILE); - if (!exfat_validate_entry(exfat_get_entry_type(ep), &mode)) - goto put_es; + if (num_entries == ES_ALL_ENTRIES) { + struct exfat_dentry *ep; + + ep = exfat_get_dentry_cached(es, ES_IDX_FILE); + if (ep->type != EXFAT_FILE) { + brelse(bh); + return -EIO; + } + + num_entries = ep->dentry.file.num_ext + 1; + } - num_entries = type == ES_ALL_ENTRIES ? - ep->dentry.file.num_ext + 1 : type; es->num_entries = num_entries; num_bh = EXFAT_B_TO_BLK_ROUND_UP(off + num_entries * DENTRY_SIZE, sb); @@ -918,8 +839,27 @@ int exfat_get_dentry_set(struct exfat_entry_set_cache *es, es->bh[es->num_bh++] = bh; } + return 0; + +put_es: + exfat_put_dentry_set(es, false); + return -EIO; +} + +int exfat_get_dentry_set(struct exfat_entry_set_cache *es, + struct super_block *sb, struct exfat_chain *p_dir, + int entry, unsigned int num_entries) +{ + int ret, i; + struct exfat_dentry *ep; + enum exfat_validate_dentry_mode mode = ES_MODE_GET_FILE_ENTRY; + + ret = __exfat_get_dentry_set(es, sb, p_dir, entry, num_entries); + if (ret < 0) + return ret; + /* validate cached dentries */ - for (i = ES_IDX_STREAM; i < num_entries; i++) { + for (i = ES_IDX_STREAM; i < es->num_entries; i++) { ep = exfat_get_dentry_cached(es, i); if (!exfat_validate_entry(exfat_get_entry_type(ep), &mode)) goto put_es; @@ -931,6 +871,85 @@ put_es: return -EIO; } +static int exfat_validate_empty_dentry_set(struct exfat_entry_set_cache *es) +{ + struct exfat_dentry *ep; + struct buffer_head *bh; + int i, off; + bool unused_hit = false; + + /* + * ONLY UNUSED OR DELETED DENTRIES ARE ALLOWED: + * Although it violates the specification for a deleted entry to + * follow an unused entry, some exFAT implementations could work + * like this. Therefore, to improve compatibility, let's allow it. + */ + for (i = 0; i < es->num_entries; i++) { + ep = exfat_get_dentry_cached(es, i); + if (ep->type == EXFAT_UNUSED) { + unused_hit = true; + } else if (!IS_EXFAT_DELETED(ep->type)) { + if (unused_hit) + goto err_used_follow_unused; + i++; + goto count_skip_entries; + } + } + + return 0; + +err_used_follow_unused: + off = es->start_off + (i << DENTRY_SIZE_BITS); + bh = es->bh[EXFAT_B_TO_BLK(off, es->sb)]; + + exfat_fs_error(es->sb, + "in sector %lld, dentry %d should be unused, but 0x%x", + bh->b_blocknr, off >> DENTRY_SIZE_BITS, ep->type); + + return -EIO; + +count_skip_entries: + es->num_entries = EXFAT_B_TO_DEN(EXFAT_BLK_TO_B(es->num_bh, es->sb) - es->start_off); + for (; i < es->num_entries; i++) { + ep = exfat_get_dentry_cached(es, i); + if (IS_EXFAT_DELETED(ep->type)) + break; + } + + return i; +} + +/* + * Get an empty dentry set. + * + * in: + * sb+p_dir+entry: indicates the empty dentry location + * num_entries: specifies how many empty dentries should be included. + * out: + * es: pointer of empty dentry set on success. + * return: + * 0 : on success + * >0 : the dentries are not empty, the return value is the number of + * dentries to be skipped for the next lookup. + * <0 : on failure + */ +int exfat_get_empty_dentry_set(struct exfat_entry_set_cache *es, + struct super_block *sb, struct exfat_chain *p_dir, + int entry, unsigned int num_entries) +{ + int ret; + + ret = __exfat_get_dentry_set(es, sb, p_dir, entry, num_entries); + if (ret < 0) + return ret; + + ret = exfat_validate_empty_dentry_set(es); + if (ret) + exfat_put_dentry_set(es, false); + + return ret; +} + static inline void exfat_reset_empty_hint(struct exfat_hint_femp *hint_femp) { hint_femp->eidx = EXFAT_HINT_NONE; @@ -1187,27 +1206,6 @@ found: return dentry - num_ext; } -int exfat_count_ext_entries(struct super_block *sb, struct exfat_chain *p_dir, - int entry, struct exfat_dentry *ep) -{ - int i, count = 0; - unsigned int type; - struct exfat_dentry *ext_ep; - struct buffer_head *bh; - - for (i = 0, entry++; i < ep->dentry.file.num_ext; i++, entry++) { - ext_ep = exfat_get_dentry(sb, p_dir, entry, &bh); - if (!ext_ep) - return -EIO; - - type = exfat_get_entry_type(ext_ep); - brelse(bh); - if (type & TYPE_CRITICAL_SEC || type & TYPE_BENIGN_SEC) - count++; - } - return count; -} - int exfat_count_dir_entries(struct super_block *sb, struct exfat_chain *p_dir) { int i, count = 0; diff --git a/fs/exfat/exfat_fs.h b/fs/exfat/exfat_fs.h index 361595433480..ecc5db952deb 100644 --- a/fs/exfat/exfat_fs.h +++ b/fs/exfat/exfat_fs.h @@ -431,8 +431,6 @@ int exfat_ent_get(struct super_block *sb, unsigned int loc, unsigned int *content); int exfat_ent_set(struct super_block *sb, unsigned int loc, unsigned int content); -int exfat_count_ext_entries(struct super_block *sb, struct exfat_chain *p_dir, - int entry, struct exfat_dentry *p_entry); int exfat_chain_cont_cluster(struct super_block *sb, unsigned int chain, unsigned int len); int exfat_zeroed_cluster(struct inode *dir, unsigned int clu); @@ -480,16 +478,14 @@ int exfat_get_cluster(struct inode *inode, unsigned int cluster, extern const struct inode_operations exfat_dir_inode_operations; extern const struct file_operations exfat_dir_operations; unsigned int exfat_get_entry_type(struct exfat_dentry *p_entry); -int exfat_init_dir_entry(struct inode *inode, struct exfat_chain *p_dir, - int entry, unsigned int type, unsigned int start_clu, - unsigned long long size); -int exfat_init_ext_entry(struct inode *inode, struct exfat_chain *p_dir, - int entry, int num_entries, struct exfat_uni_name *p_uniname); -int exfat_remove_entries(struct inode *inode, struct exfat_chain *p_dir, - int entry, int order, int num_entries); -int exfat_update_dir_chksum(struct inode *inode, struct exfat_chain *p_dir, - int entry); -void exfat_update_dir_chksum_with_entry_set(struct exfat_entry_set_cache *es); +void exfat_init_dir_entry(struct exfat_entry_set_cache *es, + unsigned int type, unsigned int start_clu, + unsigned long long size, struct timespec64 *ts); +void exfat_init_ext_entry(struct exfat_entry_set_cache *es, int num_entries, + struct exfat_uni_name *p_uniname); +void exfat_remove_entries(struct inode *inode, struct exfat_entry_set_cache *es, + int order); +void exfat_update_dir_chksum(struct exfat_entry_set_cache *es); int exfat_calc_num_entries(struct exfat_uni_name *p_uniname); int exfat_find_dir_entry(struct super_block *sb, struct exfat_inode_info *ei, struct exfat_chain *p_dir, struct exfat_uni_name *p_uniname, @@ -501,7 +497,10 @@ struct exfat_dentry *exfat_get_dentry_cached(struct exfat_entry_set_cache *es, int num); int exfat_get_dentry_set(struct exfat_entry_set_cache *es, struct super_block *sb, struct exfat_chain *p_dir, int entry, - unsigned int type); + unsigned int num_entries); +int exfat_get_empty_dentry_set(struct exfat_entry_set_cache *es, + struct super_block *sb, struct exfat_chain *p_dir, int entry, + unsigned int num_entries); int exfat_put_dentry_set(struct exfat_entry_set_cache *es, int sync); int exfat_count_dir_entries(struct super_block *sb, struct exfat_chain *p_dir); diff --git a/fs/exfat/inode.c b/fs/exfat/inode.c index 0687f952956c..dd894e558c91 100644 --- a/fs/exfat/inode.c +++ b/fs/exfat/inode.c @@ -94,7 +94,7 @@ int __exfat_write_inode(struct inode *inode, int sync) ep2->dentry.stream.start_clu = EXFAT_FREE_CLUSTER; } - exfat_update_dir_chksum_with_entry_set(&es); + exfat_update_dir_chksum(&es); return exfat_put_dentry_set(&es, sync); } diff --git a/fs/exfat/namei.c b/fs/exfat/namei.c index 9c549fd11fc8..631ad9e8e32a 100644 --- a/fs/exfat/namei.c +++ b/fs/exfat/namei.c @@ -204,21 +204,16 @@ const struct dentry_operations exfat_utf8_dentry_ops = { .d_compare = exfat_utf8_d_cmp, }; -/* used only in search empty_slot() */ -#define CNT_UNUSED_NOHIT (-1) -#define CNT_UNUSED_HIT (-2) /* search EMPTY CONTINUOUS "num_entries" entries */ static int exfat_search_empty_slot(struct super_block *sb, struct exfat_hint_femp *hint_femp, struct exfat_chain *p_dir, - int num_entries) + int num_entries, struct exfat_entry_set_cache *es) { - int i, dentry, num_empty = 0; + int i, dentry, ret; int dentries_per_clu; - unsigned int type; struct exfat_chain clu; - struct exfat_dentry *ep; struct exfat_sb_info *sbi = EXFAT_SB(sb); - struct buffer_head *bh; + int total_entries = EXFAT_CLU_TO_DEN(p_dir->size, sbi); dentries_per_clu = sbi->dentries_per_clu; @@ -231,7 +226,7 @@ static int exfat_search_empty_slot(struct super_block *sb, * Otherwise, and if "dentry + hint_famp->count" is also equal * to "p_dir->size * dentries_per_clu", it means ENOSPC. */ - if (dentry + hint_femp->count == p_dir->size * dentries_per_clu && + if (dentry + hint_femp->count == total_entries && num_entries > hint_femp->count) return -ENOSPC; @@ -242,69 +237,41 @@ static int exfat_search_empty_slot(struct super_block *sb, dentry = 0; } - while (clu.dir != EXFAT_EOF_CLUSTER) { + while (dentry + num_entries < total_entries && + clu.dir != EXFAT_EOF_CLUSTER) { i = dentry & (dentries_per_clu - 1); - for (; i < dentries_per_clu; i++, dentry++) { - ep = exfat_get_dentry(sb, &clu, i, &bh); - if (!ep) - return -EIO; - type = exfat_get_entry_type(ep); - brelse(bh); - - if (type == TYPE_UNUSED || type == TYPE_DELETED) { - num_empty++; - if (hint_femp->eidx == EXFAT_HINT_NONE) { - hint_femp->eidx = dentry; - hint_femp->count = CNT_UNUSED_NOHIT; - exfat_chain_set(&hint_femp->cur, - clu.dir, clu.size, clu.flags); - } - - if (type == TYPE_UNUSED && - hint_femp->count != CNT_UNUSED_HIT) - hint_femp->count = CNT_UNUSED_HIT; + ret = exfat_get_empty_dentry_set(es, sb, &clu, i, num_entries); + if (ret < 0) + return ret; + else if (ret == 0) + return dentry; + + dentry += ret; + i += ret; + + while (i >= dentries_per_clu) { + if (clu.flags == ALLOC_NO_FAT_CHAIN) { + if (--clu.size > 0) + clu.dir++; + else + clu.dir = EXFAT_EOF_CLUSTER; } else { - if (hint_femp->eidx != EXFAT_HINT_NONE && - hint_femp->count == CNT_UNUSED_HIT) { - /* unused empty group means - * an empty group which includes - * unused dentry - */ - exfat_fs_error(sb, - "found bogus dentry(%d) beyond unused empty group(%d) (start_clu : %u, cur_clu : %u)", - dentry, hint_femp->eidx, - p_dir->dir, clu.dir); + if (exfat_get_next_cluster(sb, &clu.dir)) return -EIO; - } - - num_empty = 0; - hint_femp->eidx = EXFAT_HINT_NONE; } - if (num_empty >= num_entries) { - /* found and invalidate hint_femp */ - hint_femp->eidx = EXFAT_HINT_NONE; - return (dentry - (num_entries - 1)); - } - } - - if (clu.flags == ALLOC_NO_FAT_CHAIN) { - if (--clu.size > 0) - clu.dir++; - else - clu.dir = EXFAT_EOF_CLUSTER; - } else { - if (exfat_get_next_cluster(sb, &clu.dir)) - return -EIO; + i -= dentries_per_clu; } } - hint_femp->eidx = p_dir->size * dentries_per_clu - num_empty; - hint_femp->count = num_empty; - if (num_empty == 0) + hint_femp->eidx = dentry; + hint_femp->count = 0; + if (dentry == total_entries || clu.dir == EXFAT_EOF_CLUSTER) exfat_chain_set(&hint_femp->cur, EXFAT_EOF_CLUSTER, 0, clu.flags); + else + hint_femp->cur = clu; return -ENOSPC; } @@ -325,7 +292,8 @@ static int exfat_check_max_dentries(struct inode *inode) * if there isn't any empty slot, expand cluster chain. */ static int exfat_find_empty_entry(struct inode *inode, - struct exfat_chain *p_dir, int num_entries) + struct exfat_chain *p_dir, int num_entries, + struct exfat_entry_set_cache *es) { int dentry; unsigned int ret, last_clu; @@ -344,7 +312,7 @@ static int exfat_find_empty_entry(struct inode *inode, } while ((dentry = exfat_search_empty_slot(sb, &hint_femp, p_dir, - num_entries)) < 0) { + num_entries, es)) < 0) { if (dentry == -EIO) break; @@ -499,6 +467,8 @@ static int exfat_add_entry(struct inode *inode, const char *path, struct exfat_sb_info *sbi = EXFAT_SB(sb); struct exfat_uni_name uniname; struct exfat_chain clu; + struct timespec64 ts = current_time(inode); + struct exfat_entry_set_cache es; int clu_size = 0; unsigned int start_clu = EXFAT_FREE_CLUSTER; @@ -513,7 +483,7 @@ static int exfat_add_entry(struct inode *inode, const char *path, } /* exfat_find_empty_entry must be called before alloc_cluster() */ - dentry = exfat_find_empty_entry(inode, p_dir, num_entries); + dentry = exfat_find_empty_entry(inode, p_dir, num_entries, &es); if (dentry < 0) { ret = dentry; /* -EIO or -ENOSPC */ goto out; @@ -521,8 +491,10 @@ static int exfat_add_entry(struct inode *inode, const char *path, if (type == TYPE_DIR && !sbi->options.zero_size_dir) { ret = exfat_alloc_new_dir(inode, &clu); - if (ret) + if (ret) { + exfat_put_dentry_set(&es, false); goto out; + } start_clu = clu.dir; clu_size = sbi->cluster_size; } @@ -531,12 +503,10 @@ static int exfat_add_entry(struct inode *inode, const char *path, /* fill the dos name directory entry information of the created file. * the first cluster is not determined yet. (0) */ - ret = exfat_init_dir_entry(inode, p_dir, dentry, type, - start_clu, clu_size); - if (ret) - goto out; + exfat_init_dir_entry(&es, type, start_clu, clu_size, &ts); + exfat_init_ext_entry(&es, num_entries, &uniname); - ret = exfat_init_ext_entry(inode, p_dir, dentry, num_entries, &uniname); + ret = exfat_put_dentry_set(&es, IS_DIRSYNC(inode)); if (ret) goto out; @@ -577,6 +547,7 @@ static int exfat_create(struct mnt_idmap *idmap, struct inode *dir, struct exfat_dir_entry info; loff_t i_pos; int err; + loff_t size = i_size_read(dir); mutex_lock(&EXFAT_SB(sb)->s_lock); exfat_set_volume_dirty(sb); @@ -587,7 +558,7 @@ static int exfat_create(struct mnt_idmap *idmap, struct inode *dir, inode_inc_iversion(dir); inode_set_mtime_to_ts(dir, inode_set_ctime_current(dir)); - if (IS_DIRSYNC(dir)) + if (IS_DIRSYNC(dir) && size != i_size_read(dir)) exfat_sync_inode(dir); else mark_inode_dirty(dir); @@ -795,12 +766,11 @@ unlock: static int exfat_unlink(struct inode *dir, struct dentry *dentry) { struct exfat_chain cdir; - struct exfat_dentry *ep; struct super_block *sb = dir->i_sb; struct inode *inode = dentry->d_inode; struct exfat_inode_info *ei = EXFAT_I(inode); - struct buffer_head *bh; - int num_entries, entry, err = 0; + struct exfat_entry_set_cache es; + int entry, err = 0; mutex_lock(&EXFAT_SB(sb)->s_lock); exfat_chain_dup(&cdir, &ei->dir); @@ -811,26 +781,20 @@ static int exfat_unlink(struct inode *dir, struct dentry *dentry) goto unlock; } - ep = exfat_get_dentry(sb, &cdir, entry, &bh); - if (!ep) { - err = -EIO; - goto unlock; - } - num_entries = exfat_count_ext_entries(sb, &cdir, entry, ep); - if (num_entries < 0) { + err = exfat_get_dentry_set(&es, sb, &cdir, entry, ES_ALL_ENTRIES); + if (err) { err = -EIO; - brelse(bh); goto unlock; } - num_entries++; - brelse(bh); exfat_set_volume_dirty(sb); + /* update the directory entry */ - if (exfat_remove_entries(dir, &cdir, entry, 0, num_entries)) { - err = -EIO; + exfat_remove_entries(inode, &es, ES_IDX_FILE); + + err = exfat_put_dentry_set(&es, IS_DIRSYNC(inode)); + if (err) goto unlock; - } /* This doesn't modify ei */ ei->dir.dir = DIR_DELETED; @@ -838,10 +802,7 @@ static int exfat_unlink(struct inode *dir, struct dentry *dentry) inode_inc_iversion(dir); simple_inode_init_ts(dir); exfat_truncate_inode_atime(dir); - if (IS_DIRSYNC(dir)) - exfat_sync_inode(dir); - else - mark_inode_dirty(dir); + mark_inode_dirty(dir); clear_nlink(inode); simple_inode_init_ts(inode); @@ -862,6 +823,7 @@ static int exfat_mkdir(struct mnt_idmap *idmap, struct inode *dir, struct exfat_chain cdir; loff_t i_pos; int err; + loff_t size = i_size_read(dir); mutex_lock(&EXFAT_SB(sb)->s_lock); exfat_set_volume_dirty(sb); @@ -872,7 +834,7 @@ static int exfat_mkdir(struct mnt_idmap *idmap, struct inode *dir, inode_inc_iversion(dir); inode_set_mtime_to_ts(dir, inode_set_ctime_current(dir)); - if (IS_DIRSYNC(dir)) + if (IS_DIRSYNC(dir) && size != i_size_read(dir)) exfat_sync_inode(dir); else mark_inode_dirty(dir); @@ -946,13 +908,12 @@ static int exfat_check_dir_empty(struct super_block *sb, static int exfat_rmdir(struct inode *dir, struct dentry *dentry) { struct inode *inode = dentry->d_inode; - struct exfat_dentry *ep; struct exfat_chain cdir, clu_to_free; struct super_block *sb = inode->i_sb; struct exfat_sb_info *sbi = EXFAT_SB(sb); struct exfat_inode_info *ei = EXFAT_I(inode); - struct buffer_head *bh; - int num_entries, entry, err; + struct exfat_entry_set_cache es; + int entry, err; mutex_lock(&EXFAT_SB(inode->i_sb)->s_lock); @@ -976,27 +937,20 @@ static int exfat_rmdir(struct inode *dir, struct dentry *dentry) goto unlock; } - ep = exfat_get_dentry(sb, &cdir, entry, &bh); - if (!ep) { - err = -EIO; - goto unlock; - } - - num_entries = exfat_count_ext_entries(sb, &cdir, entry, ep); - if (num_entries < 0) { + err = exfat_get_dentry_set(&es, sb, &cdir, entry, ES_ALL_ENTRIES); + if (err) { err = -EIO; - brelse(bh); goto unlock; } - num_entries++; - brelse(bh); exfat_set_volume_dirty(sb); - err = exfat_remove_entries(dir, &cdir, entry, 0, num_entries); - if (err) { - exfat_err(sb, "failed to exfat_remove_entries : err(%d)", err); + + exfat_remove_entries(inode, &es, ES_IDX_FILE); + + err = exfat_put_dentry_set(&es, IS_DIRSYNC(dir)); + if (err) goto unlock; - } + ei->dir.dir = DIR_DELETED; inode_inc_iversion(dir); @@ -1022,67 +976,52 @@ static int exfat_rename_file(struct inode *inode, struct exfat_chain *p_dir, int oldentry, struct exfat_uni_name *p_uniname, struct exfat_inode_info *ei) { - int ret, num_old_entries, num_new_entries; + int ret, num_new_entries; struct exfat_dentry *epold, *epnew; struct super_block *sb = inode->i_sb; - struct buffer_head *new_bh, *old_bh; + struct exfat_entry_set_cache old_es, new_es; int sync = IS_DIRSYNC(inode); - epold = exfat_get_dentry(sb, p_dir, oldentry, &old_bh); - if (!epold) - return -EIO; - - num_old_entries = exfat_count_ext_entries(sb, p_dir, oldentry, epold); - if (num_old_entries < 0) - return -EIO; - num_old_entries++; - num_new_entries = exfat_calc_num_entries(p_uniname); if (num_new_entries < 0) return num_new_entries; - if (num_old_entries < num_new_entries) { - int newentry; + ret = exfat_get_dentry_set(&old_es, sb, p_dir, oldentry, ES_ALL_ENTRIES); + if (ret) { + ret = -EIO; + return ret; + } - newentry = - exfat_find_empty_entry(inode, p_dir, num_new_entries); - if (newentry < 0) - return newentry; /* -EIO or -ENOSPC */ + epold = exfat_get_dentry_cached(&old_es, ES_IDX_FILE); - epnew = exfat_get_dentry(sb, p_dir, newentry, &new_bh); - if (!epnew) - return -EIO; + if (old_es.num_entries < num_new_entries) { + int newentry; + newentry = exfat_find_empty_entry(inode, p_dir, num_new_entries, + &new_es); + if (newentry < 0) { + ret = newentry; /* -EIO or -ENOSPC */ + goto put_old_es; + } + + epnew = exfat_get_dentry_cached(&new_es, ES_IDX_FILE); *epnew = *epold; if (exfat_get_entry_type(epnew) == TYPE_FILE) { epnew->dentry.file.attr |= cpu_to_le16(EXFAT_ATTR_ARCHIVE); ei->attr |= EXFAT_ATTR_ARCHIVE; } - exfat_update_bh(new_bh, sync); - brelse(old_bh); - brelse(new_bh); - - epold = exfat_get_dentry(sb, p_dir, oldentry + 1, &old_bh); - if (!epold) - return -EIO; - epnew = exfat_get_dentry(sb, p_dir, newentry + 1, &new_bh); - if (!epnew) { - brelse(old_bh); - return -EIO; - } + epold = exfat_get_dentry_cached(&old_es, ES_IDX_STREAM); + epnew = exfat_get_dentry_cached(&new_es, ES_IDX_STREAM); *epnew = *epold; - exfat_update_bh(new_bh, sync); - brelse(old_bh); - brelse(new_bh); - ret = exfat_init_ext_entry(inode, p_dir, newentry, - num_new_entries, p_uniname); + exfat_init_ext_entry(&new_es, num_new_entries, p_uniname); + + ret = exfat_put_dentry_set(&new_es, sync); if (ret) - return ret; + goto put_old_es; - exfat_remove_entries(inode, p_dir, oldentry, 0, - num_old_entries); + exfat_remove_entries(inode, &old_es, ES_IDX_FILE); ei->dir = *p_dir; ei->entry = newentry; } else { @@ -1090,85 +1029,72 @@ static int exfat_rename_file(struct inode *inode, struct exfat_chain *p_dir, epold->dentry.file.attr |= cpu_to_le16(EXFAT_ATTR_ARCHIVE); ei->attr |= EXFAT_ATTR_ARCHIVE; } - exfat_update_bh(old_bh, sync); - brelse(old_bh); - ret = exfat_init_ext_entry(inode, p_dir, oldentry, - num_new_entries, p_uniname); - if (ret) - return ret; - exfat_remove_entries(inode, p_dir, oldentry, num_new_entries, - num_old_entries); + exfat_remove_entries(inode, &old_es, ES_IDX_FIRST_FILENAME + 1); + exfat_init_ext_entry(&old_es, num_new_entries, p_uniname); } - return 0; + return exfat_put_dentry_set(&old_es, sync); + +put_old_es: + exfat_put_dentry_set(&old_es, false); + return ret; } static int exfat_move_file(struct inode *inode, struct exfat_chain *p_olddir, int oldentry, struct exfat_chain *p_newdir, struct exfat_uni_name *p_uniname, struct exfat_inode_info *ei) { - int ret, newentry, num_new_entries, num_old_entries; + int ret, newentry, num_new_entries; struct exfat_dentry *epmov, *epnew; struct super_block *sb = inode->i_sb; - struct buffer_head *mov_bh, *new_bh; - - epmov = exfat_get_dentry(sb, p_olddir, oldentry, &mov_bh); - if (!epmov) - return -EIO; - - num_old_entries = exfat_count_ext_entries(sb, p_olddir, oldentry, - epmov); - if (num_old_entries < 0) - return -EIO; - num_old_entries++; + struct exfat_entry_set_cache mov_es, new_es; num_new_entries = exfat_calc_num_entries(p_uniname); if (num_new_entries < 0) return num_new_entries; - newentry = exfat_find_empty_entry(inode, p_newdir, num_new_entries); - if (newentry < 0) - return newentry; /* -EIO or -ENOSPC */ - - epnew = exfat_get_dentry(sb, p_newdir, newentry, &new_bh); - if (!epnew) + ret = exfat_get_dentry_set(&mov_es, sb, p_olddir, oldentry, + ES_ALL_ENTRIES); + if (ret) return -EIO; + newentry = exfat_find_empty_entry(inode, p_newdir, num_new_entries, + &new_es); + if (newentry < 0) { + ret = newentry; /* -EIO or -ENOSPC */ + goto put_mov_es; + } + + epmov = exfat_get_dentry_cached(&mov_es, ES_IDX_FILE); + epnew = exfat_get_dentry_cached(&new_es, ES_IDX_FILE); *epnew = *epmov; if (exfat_get_entry_type(epnew) == TYPE_FILE) { epnew->dentry.file.attr |= cpu_to_le16(EXFAT_ATTR_ARCHIVE); ei->attr |= EXFAT_ATTR_ARCHIVE; } - exfat_update_bh(new_bh, IS_DIRSYNC(inode)); - brelse(mov_bh); - brelse(new_bh); - - epmov = exfat_get_dentry(sb, p_olddir, oldentry + 1, &mov_bh); - if (!epmov) - return -EIO; - epnew = exfat_get_dentry(sb, p_newdir, newentry + 1, &new_bh); - if (!epnew) { - brelse(mov_bh); - return -EIO; - } + epmov = exfat_get_dentry_cached(&mov_es, ES_IDX_STREAM); + epnew = exfat_get_dentry_cached(&new_es, ES_IDX_STREAM); *epnew = *epmov; - exfat_update_bh(new_bh, IS_DIRSYNC(inode)); - brelse(mov_bh); - brelse(new_bh); - - ret = exfat_init_ext_entry(inode, p_newdir, newentry, num_new_entries, - p_uniname); - if (ret) - return ret; - exfat_remove_entries(inode, p_olddir, oldentry, 0, num_old_entries); + exfat_init_ext_entry(&new_es, num_new_entries, p_uniname); + exfat_remove_entries(inode, &mov_es, ES_IDX_FILE); exfat_chain_set(&ei->dir, p_newdir->dir, p_newdir->size, p_newdir->flags); ei->entry = newentry; - return 0; + + ret = exfat_put_dentry_set(&new_es, IS_DIRSYNC(inode)); + if (ret) + goto put_mov_es; + + return exfat_put_dentry_set(&mov_es, IS_DIRSYNC(inode)); + +put_mov_es: + exfat_put_dentry_set(&mov_es, false); + + return ret; } /* rename or move a old file into a new file */ @@ -1186,7 +1112,6 @@ static int __exfat_rename(struct inode *old_parent_inode, struct exfat_sb_info *sbi = EXFAT_SB(sb); const unsigned char *new_path = new_dentry->d_name.name; struct inode *new_inode = new_dentry->d_inode; - int num_entries; struct exfat_inode_info *new_ei = NULL; unsigned int new_entry_type = TYPE_UNUSED; int new_entry = 0; @@ -1257,25 +1182,21 @@ static int __exfat_rename(struct inode *old_parent_inode, &newdir, &uni_name, ei); if (!ret && new_inode) { + struct exfat_entry_set_cache es; + /* delete entries of new_dir */ - ep = exfat_get_dentry(sb, p_dir, new_entry, &new_bh); - if (!ep) { + ret = exfat_get_dentry_set(&es, sb, p_dir, new_entry, + ES_ALL_ENTRIES); + if (ret) { ret = -EIO; goto del_out; } - num_entries = exfat_count_ext_entries(sb, p_dir, new_entry, ep); - if (num_entries < 0) { - ret = -EIO; - goto del_out; - } - brelse(new_bh); + exfat_remove_entries(new_inode, &es, ES_IDX_FILE); - if (exfat_remove_entries(new_inode, p_dir, new_entry, 0, - num_entries + 1)) { - ret = -EIO; + ret = exfat_put_dentry_set(&es, IS_DIRSYNC(new_inode)); + if (ret) goto del_out; - } /* Free the clusters if new_inode is a dir(as if exfat_rmdir) */ if (new_entry_type == TYPE_DIR && @@ -1317,6 +1238,7 @@ static int exfat_rename(struct mnt_idmap *idmap, struct super_block *sb = old_dir->i_sb; loff_t i_pos; int err; + loff_t size = i_size_read(new_dir); /* * The VFS already checks for existence, so for local filesystems @@ -1338,7 +1260,7 @@ static int exfat_rename(struct mnt_idmap *idmap, simple_rename_timestamp(old_dir, old_dentry, new_dir, new_dentry); EXFAT_I(new_dir)->i_crtime = current_time(new_dir); exfat_truncate_inode_atime(new_dir); - if (IS_DIRSYNC(new_dir)) + if (IS_DIRSYNC(new_dir) && size != i_size_read(new_dir)) exfat_sync_inode(new_dir); else mark_inode_dirty(new_dir); @@ -1359,9 +1281,7 @@ static int exfat_rename(struct mnt_idmap *idmap, } inode_inc_iversion(old_dir); - if (IS_DIRSYNC(old_dir)) - exfat_sync_inode(old_dir); - else + if (new_dir != old_dir) mark_inode_dirty(old_dir); if (new_inode) { diff --git a/fs/ext2/Kconfig b/fs/ext2/Kconfig index 74d98965902e..d6cfb1849580 100644 --- a/fs/ext2/Kconfig +++ b/fs/ext2/Kconfig @@ -1,16 +1,23 @@ # SPDX-License-Identifier: GPL-2.0-only config EXT2_FS - tristate "Second extended fs support" + tristate "Second extended fs support (DEPRECATED)" select BUFFER_HEAD select FS_IOMAP select LEGACY_DIRECT_IO help Ext2 is a standard Linux file system for hard disks. - To compile this file system support as a module, choose M here: the - module will be called ext2. + This filesystem driver is deprecated because it does not properly + support inode time stamps beyond 03:14:07 UTC on 19 January 2038. - If unsure, say Y. + Ext2 users are advised to use ext4 driver to access their filesystem. + The driver is fully compatible, supports filesystems without journal + or extents, and also supports larger time stamps if the filesystem + is created with at least 256 byte inodes. + + This code is kept as a simple reference for filesystem developers. + + If unsure, say N. config EXT2_FS_XATTR bool "Ext2 extended attributes" diff --git a/fs/ext2/balloc.c b/fs/ext2/balloc.c index e124f3d709b2..1bfd6ab11038 100644 --- a/fs/ext2/balloc.c +++ b/fs/ext2/balloc.c @@ -412,7 +412,7 @@ void ext2_init_block_alloc_info(struct inode *inode) struct ext2_block_alloc_info *block_i; struct super_block *sb = inode->i_sb; - block_i = kmalloc(sizeof(*block_i), GFP_NOFS); + block_i = kmalloc(sizeof(*block_i), GFP_KERNEL); if (block_i) { struct ext2_reserve_window_node *rsv = &block_i->rsv_window_node; diff --git a/fs/ext2/ext2.h b/fs/ext2/ext2.h index 677a9ad45dcb..f38bdd46e4f7 100644 --- a/fs/ext2/ext2.h +++ b/fs/ext2/ext2.h @@ -674,7 +674,7 @@ struct ext2_inode_info { struct inode vfs_inode; struct list_head i_orphan; /* unlinked but open inodes */ #ifdef CONFIG_QUOTA - struct dquot *i_dquot[MAXQUOTAS]; + struct dquot __rcu *i_dquot[MAXQUOTAS]; #endif }; diff --git a/fs/ext2/inode.c b/fs/ext2/inode.c index 5a4272b2c6b0..f3d570a9302b 100644 --- a/fs/ext2/inode.c +++ b/fs/ext2/inode.c @@ -754,7 +754,7 @@ static int ext2_get_blocks(struct inode *inode, */ err = sb_issue_zeroout(inode->i_sb, le32_to_cpu(chain[depth-1].key), count, - GFP_NOFS); + GFP_KERNEL); if (err) { mutex_unlock(&ei->truncate_mutex); goto cleanup; diff --git a/fs/ext2/super.c b/fs/ext2/super.c index cabea887314d..37f7ce56adce 100644 --- a/fs/ext2/super.c +++ b/fs/ext2/super.c @@ -319,7 +319,7 @@ static ssize_t ext2_quota_read(struct super_block *sb, int type, char *data, siz static ssize_t ext2_quota_write(struct super_block *sb, int type, const char *data, size_t len, loff_t off); static int ext2_quota_on(struct super_block *sb, int type, int format_id, const struct path *path); -static struct dquot **ext2_get_dquots(struct inode *inode) +static struct dquot __rcu **ext2_get_dquots(struct inode *inode) { return EXT2_I(inode)->i_dquot; } diff --git a/fs/ext2/xattr.c b/fs/ext2/xattr.c index e849241ebb8f..c885dcc3bd0d 100644 --- a/fs/ext2/xattr.c +++ b/fs/ext2/xattr.c @@ -874,7 +874,7 @@ ext2_xattr_cache_insert(struct mb_cache *cache, struct buffer_head *bh) __u32 hash = le32_to_cpu(HDR(bh)->h_hash); int error; - error = mb_cache_entry_create(cache, GFP_NOFS, hash, bh->b_blocknr, + error = mb_cache_entry_create(cache, GFP_KERNEL, hash, bh->b_blocknr, true); if (error) { if (error == -EBUSY) { diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h index 3c0d7d143036..8d126654019e 100644 --- a/fs/ext4/ext4.h +++ b/fs/ext4/ext4.h @@ -1158,7 +1158,7 @@ struct ext4_inode_info { tid_t i_datasync_tid; #ifdef CONFIG_QUOTA - struct dquot *i_dquot[MAXQUOTAS]; + struct dquot __rcu *i_dquot[MAXQUOTAS]; #endif /* Precomputed uuid+inum+igen checksum for seeding inode checksums */ diff --git a/fs/ext4/extents.c b/fs/ext4/extents.c index 7669d154c05e..e57054bdc5fd 100644 --- a/fs/ext4/extents.c +++ b/fs/ext4/extents.c @@ -4111,10 +4111,10 @@ insert_hole: * * Need to be called with * down_read(&EXT4_I(inode)->i_data_sem) if not allocating file system block - * (ie, create is zero). Otherwise down_write(&EXT4_I(inode)->i_data_sem) + * (ie, flags is zero). Otherwise down_write(&EXT4_I(inode)->i_data_sem) * * return > 0, number of blocks already mapped/allocated - * if create == 0 and these are pre-allocated blocks + * if flags doesn't contain EXT4_GET_BLOCKS_CREATE and these are pre-allocated blocks * buffer head is unmapped * otherwise blocks are mapped * @@ -4218,7 +4218,7 @@ int ext4_ext_map_blocks(handle_t *handle, struct inode *inode, /* * requested block isn't allocated yet; - * we couldn't try to create block if create flag is zero + * we couldn't try to create block if flags doesn't contain EXT4_GET_BLOCKS_CREATE */ if ((flags & EXT4_GET_BLOCKS_CREATE) == 0) { ext4_lblk_t len; diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c index 2ccf3b5e3a7c..537803250ca9 100644 --- a/fs/ext4/inode.c +++ b/fs/ext4/inode.c @@ -465,9 +465,10 @@ static void ext4_map_blocks_es_recheck(handle_t *handle, * Otherwise, call with ext4_ind_map_blocks() to handle indirect mapping * based files * - * On success, it returns the number of blocks being mapped or allocated. if - * create==0 and the blocks are pre-allocated and unwritten, the resulting @map - * is marked as unwritten. If the create == 1, it will mark @map as mapped. + * On success, it returns the number of blocks being mapped or allocated. + * If flags doesn't contain EXT4_GET_BLOCKS_CREATE the blocks are + * pre-allocated and unwritten, the resulting @map is marked as unwritten. + * If the flags contain EXT4_GET_BLOCKS_CREATE, it will mark @map as mapped. * * It returns 0 if plain look up failed (blocks have not been allocated), in * that case, @map is returned as unmapped but we still do fill map->m_len to @@ -589,8 +590,7 @@ found: * Returns if the blocks have already allocated * * Note that if blocks have been preallocated - * ext4_ext_get_block() returns the create = 0 - * with buffer head unmapped. + * ext4_ext_map_blocks() returns with buffer head unmapped */ if (retval > 0 && map->m_flags & EXT4_MAP_MAPPED) /* diff --git a/fs/ext4/mballoc-test.c b/fs/ext4/mballoc-test.c index f94901fd3835..044ca5238f41 100644 --- a/fs/ext4/mballoc-test.c +++ b/fs/ext4/mballoc-test.c @@ -5,6 +5,7 @@ #include <kunit/test.h> #include <kunit/static_stub.h> +#include <linux/random.h> #include "ext4.h" @@ -20,41 +21,135 @@ struct mbt_ctx { }; struct mbt_ext4_super_block { - struct super_block sb; + struct ext4_super_block es; + struct ext4_sb_info sbi; struct mbt_ctx mbt_ctx; }; -#define MBT_CTX(_sb) (&(container_of((_sb), struct mbt_ext4_super_block, sb)->mbt_ctx)) +#define MBT_SB(_sb) (container_of((_sb)->s_fs_info, struct mbt_ext4_super_block, sbi)) +#define MBT_CTX(_sb) (&MBT_SB(_sb)->mbt_ctx) #define MBT_GRP_CTX(_sb, _group) (&MBT_CTX(_sb)->grp_ctx[_group]) +static const struct super_operations mbt_sops = { +}; + +static void mbt_kill_sb(struct super_block *sb) +{ + generic_shutdown_super(sb); +} + +static struct file_system_type mbt_fs_type = { + .name = "mballoc test", + .kill_sb = mbt_kill_sb, +}; + +static int mbt_mb_init(struct super_block *sb) +{ + ext4_fsblk_t block; + int ret; + + /* needed by ext4_mb_init->bdev_nonrot(sb->s_bdev) */ + sb->s_bdev = kzalloc(sizeof(*sb->s_bdev), GFP_KERNEL); + if (sb->s_bdev == NULL) + return -ENOMEM; + + sb->s_bdev->bd_queue = kzalloc(sizeof(struct request_queue), GFP_KERNEL); + if (sb->s_bdev->bd_queue == NULL) { + kfree(sb->s_bdev); + return -ENOMEM; + } + + /* + * needed by ext4_mb_init->ext4_mb_init_backend-> sbi->s_buddy_cache = + * new_inode(sb); + */ + INIT_LIST_HEAD(&sb->s_inodes); + sb->s_op = &mbt_sops; + + ret = ext4_mb_init(sb); + if (ret != 0) + goto err_out; + + block = ext4_count_free_clusters(sb); + ret = percpu_counter_init(&EXT4_SB(sb)->s_freeclusters_counter, block, + GFP_KERNEL); + if (ret != 0) + goto err_mb_release; + + ret = percpu_counter_init(&EXT4_SB(sb)->s_dirtyclusters_counter, 0, + GFP_KERNEL); + if (ret != 0) + goto err_freeclusters; + + return 0; + +err_freeclusters: + percpu_counter_destroy(&EXT4_SB(sb)->s_freeclusters_counter); +err_mb_release: + ext4_mb_release(sb); +err_out: + kfree(sb->s_bdev->bd_queue); + kfree(sb->s_bdev); + return ret; +} + +static void mbt_mb_release(struct super_block *sb) +{ + percpu_counter_destroy(&EXT4_SB(sb)->s_dirtyclusters_counter); + percpu_counter_destroy(&EXT4_SB(sb)->s_freeclusters_counter); + ext4_mb_release(sb); + kfree(sb->s_bdev->bd_queue); + kfree(sb->s_bdev); +} + +static int mbt_set(struct super_block *sb, void *data) +{ + return 0; +} + static struct super_block *mbt_ext4_alloc_super_block(void) { - struct ext4_super_block *es = kzalloc(sizeof(*es), GFP_KERNEL); - struct ext4_sb_info *sbi = kzalloc(sizeof(*sbi), GFP_KERNEL); - struct mbt_ext4_super_block *fsb = kzalloc(sizeof(*fsb), GFP_KERNEL); + struct mbt_ext4_super_block *fsb; + struct super_block *sb; + struct ext4_sb_info *sbi; + + fsb = kzalloc(sizeof(*fsb), GFP_KERNEL); + if (fsb == NULL) + return NULL; - if (fsb == NULL || sbi == NULL || es == NULL) + sb = sget(&mbt_fs_type, NULL, mbt_set, 0, NULL); + if (IS_ERR(sb)) goto out; - sbi->s_es = es; - fsb->sb.s_fs_info = sbi; - return &fsb->sb; + sbi = &fsb->sbi; + + sbi->s_blockgroup_lock = + kzalloc(sizeof(struct blockgroup_lock), GFP_KERNEL); + if (!sbi->s_blockgroup_lock) + goto out_deactivate; + + bgl_lock_init(sbi->s_blockgroup_lock); + + sbi->s_es = &fsb->es; + sb->s_fs_info = sbi; + + up_write(&sb->s_umount); + return sb; +out_deactivate: + deactivate_locked_super(sb); out: kfree(fsb); - kfree(sbi); - kfree(es); return NULL; } static void mbt_ext4_free_super_block(struct super_block *sb) { - struct mbt_ext4_super_block *fsb = - container_of(sb, struct mbt_ext4_super_block, sb); + struct mbt_ext4_super_block *fsb = MBT_SB(sb); struct ext4_sb_info *sbi = EXT4_SB(sb); - kfree(sbi->s_es); - kfree(sbi); + kfree(sbi->s_blockgroup_lock); + deactivate_super(sb); kfree(fsb); } @@ -82,6 +177,9 @@ static void mbt_init_sb_layout(struct super_block *sb, sbi->s_clusters_per_group = layout->blocks_per_group >> layout->cluster_bits; sbi->s_desc_size = layout->desc_size; + sbi->s_desc_per_block_bits = + sb->s_blocksize_bits - (fls(layout->desc_size) - 1); + sbi->s_desc_per_block = 1 << sbi->s_desc_per_block_bits; es->s_first_data_block = cpu_to_le32(0); es->s_blocks_count_lo = cpu_to_le32(layout->blocks_per_group * @@ -91,9 +189,13 @@ static void mbt_init_sb_layout(struct super_block *sb, static int mbt_grp_ctx_init(struct super_block *sb, struct mbt_grp_ctx *grp_ctx) { + ext4_grpblk_t max = EXT4_CLUSTERS_PER_GROUP(sb); + grp_ctx->bitmap_bh.b_data = kzalloc(EXT4_BLOCK_SIZE(sb), GFP_KERNEL); if (grp_ctx->bitmap_bh.b_data == NULL) return -ENOMEM; + mb_set_bits(grp_ctx->bitmap_bh.b_data, max, sb->s_blocksize * 8 - max); + ext4_free_group_clusters_set(sb, &grp_ctx->desc, max); return 0; } @@ -112,6 +214,13 @@ static void mbt_ctx_mark_used(struct super_block *sb, ext4_group_t group, mb_set_bits(grp_ctx->bitmap_bh.b_data, start, len); } +static void *mbt_ctx_bitmap(struct super_block *sb, ext4_group_t group) +{ + struct mbt_grp_ctx *grp_ctx = MBT_GRP_CTX(sb, group); + + return grp_ctx->bitmap_bh.b_data; +} + /* called after mbt_init_sb_layout */ static int mbt_ctx_init(struct super_block *sb) { @@ -133,6 +242,8 @@ static int mbt_ctx_init(struct super_block *sb) * block which will fail ext4_sb_block_valid check. */ mb_set_bits(ctx->grp_ctx[0].bitmap_bh.b_data, 0, 1); + ext4_free_group_clusters_set(sb, &ctx->grp_ctx[0].desc, + EXT4_CLUSTERS_PER_GROUP(sb) - 1); return 0; out: @@ -167,6 +278,13 @@ static int ext4_wait_block_bitmap_stub(struct super_block *sb, ext4_group_t block_group, struct buffer_head *bh) { + /* + * real ext4_wait_block_bitmap will set these flags and + * functions like ext4_mb_init_cache will verify the flags. + */ + set_buffer_uptodate(bh); + set_bitmap_uptodate(bh); + set_buffer_verified(bh); return 0; } @@ -232,6 +350,14 @@ static int mbt_kunit_init(struct kunit *test) kunit_activate_static_stub(test, ext4_mb_mark_context, ext4_mb_mark_context_stub); + + /* stub function will be called in mbt_mb_init->ext4_mb_init */ + if (mbt_mb_init(sb) != 0) { + mbt_ctx_release(sb); + mbt_ext4_free_super_block(sb); + return -ENOMEM; + } + return 0; } @@ -239,6 +365,7 @@ static void mbt_kunit_exit(struct kunit *test) { struct super_block *sb = (struct super_block *)test->priv; + mbt_mb_release(sb); mbt_ctx_release(sb); mbt_ext4_free_super_block(sb); } @@ -246,14 +373,19 @@ static void mbt_kunit_exit(struct kunit *test) static void test_new_blocks_simple(struct kunit *test) { struct super_block *sb = (struct super_block *)test->priv; - struct inode inode = { .i_sb = sb, }; + struct inode *inode; struct ext4_allocation_request ar; ext4_group_t i, goal_group = TEST_GOAL_GROUP; int err = 0; ext4_fsblk_t found; struct ext4_sb_info *sbi = EXT4_SB(sb); - ar.inode = &inode; + inode = kunit_kzalloc(test, sizeof(*inode), GFP_KERNEL); + if (!inode) + return; + + inode->i_sb = sb; + ar.inode = inode; /* get block at goal */ ar.goal = ext4_group_first_block_no(sb, goal_group); @@ -297,6 +429,436 @@ static void test_new_blocks_simple(struct kunit *test) "unexpectedly get block when no block is available"); } +#define TEST_RANGE_COUNT 8 + +struct test_range { + ext4_grpblk_t start; + ext4_grpblk_t len; +}; + +static void +mbt_generate_test_ranges(struct super_block *sb, struct test_range *ranges, + int count) +{ + ext4_grpblk_t start, len, max; + int i; + + max = EXT4_CLUSTERS_PER_GROUP(sb) / count; + for (i = 0; i < count; i++) { + start = get_random_u32() % max; + len = get_random_u32() % max; + len = min(len, max - start); + + ranges[i].start = start + i * max; + ranges[i].len = len; + } +} + +static void +validate_free_blocks_simple(struct kunit *test, struct super_block *sb, + ext4_group_t goal_group, ext4_grpblk_t start, + ext4_grpblk_t len) +{ + void *bitmap; + ext4_grpblk_t bit, max = EXT4_CLUSTERS_PER_GROUP(sb); + ext4_group_t i; + + for (i = 0; i < ext4_get_groups_count(sb); i++) { + if (i == goal_group) + continue; + + bitmap = mbt_ctx_bitmap(sb, i); + bit = mb_find_next_zero_bit(bitmap, max, 0); + KUNIT_ASSERT_EQ_MSG(test, bit, max, + "free block on unexpected group %d", i); + } + + bitmap = mbt_ctx_bitmap(sb, goal_group); + bit = mb_find_next_zero_bit(bitmap, max, 0); + KUNIT_ASSERT_EQ(test, bit, start); + + bit = mb_find_next_bit(bitmap, max, bit + 1); + KUNIT_ASSERT_EQ(test, bit, start + len); +} + +static void +test_free_blocks_simple_range(struct kunit *test, ext4_group_t goal_group, + ext4_grpblk_t start, ext4_grpblk_t len) +{ + struct super_block *sb = (struct super_block *)test->priv; + struct ext4_sb_info *sbi = EXT4_SB(sb); + struct inode *inode; + ext4_fsblk_t block; + + inode = kunit_kzalloc(test, sizeof(*inode), GFP_KERNEL); + if (!inode) + return; + inode->i_sb = sb; + + if (len == 0) + return; + + block = ext4_group_first_block_no(sb, goal_group) + + EXT4_C2B(sbi, start); + ext4_free_blocks_simple(inode, block, len); + validate_free_blocks_simple(test, sb, goal_group, start, len); + mbt_ctx_mark_used(sb, goal_group, 0, EXT4_CLUSTERS_PER_GROUP(sb)); +} + +static void test_free_blocks_simple(struct kunit *test) +{ + struct super_block *sb = (struct super_block *)test->priv; + ext4_grpblk_t max = EXT4_CLUSTERS_PER_GROUP(sb); + ext4_group_t i; + struct test_range ranges[TEST_RANGE_COUNT]; + + for (i = 0; i < ext4_get_groups_count(sb); i++) + mbt_ctx_mark_used(sb, i, 0, max); + + mbt_generate_test_ranges(sb, ranges, TEST_RANGE_COUNT); + for (i = 0; i < TEST_RANGE_COUNT; i++) + test_free_blocks_simple_range(test, TEST_GOAL_GROUP, + ranges[i].start, ranges[i].len); +} + +static void +test_mark_diskspace_used_range(struct kunit *test, + struct ext4_allocation_context *ac, + ext4_grpblk_t start, + ext4_grpblk_t len) +{ + struct super_block *sb = (struct super_block *)test->priv; + int ret; + void *bitmap; + ext4_grpblk_t i, max; + + /* ext4_mb_mark_diskspace_used will BUG if len is 0 */ + if (len == 0) + return; + + ac->ac_b_ex.fe_group = TEST_GOAL_GROUP; + ac->ac_b_ex.fe_start = start; + ac->ac_b_ex.fe_len = len; + + bitmap = mbt_ctx_bitmap(sb, TEST_GOAL_GROUP); + memset(bitmap, 0, sb->s_blocksize); + ret = ext4_mb_mark_diskspace_used(ac, NULL, 0); + KUNIT_ASSERT_EQ(test, ret, 0); + + max = EXT4_CLUSTERS_PER_GROUP(sb); + i = mb_find_next_bit(bitmap, max, 0); + KUNIT_ASSERT_EQ(test, i, start); + i = mb_find_next_zero_bit(bitmap, max, i + 1); + KUNIT_ASSERT_EQ(test, i, start + len); + i = mb_find_next_bit(bitmap, max, i + 1); + KUNIT_ASSERT_EQ(test, max, i); +} + +static void test_mark_diskspace_used(struct kunit *test) +{ + struct super_block *sb = (struct super_block *)test->priv; + struct inode *inode; + struct ext4_allocation_context ac; + struct test_range ranges[TEST_RANGE_COUNT]; + int i; + + mbt_generate_test_ranges(sb, ranges, TEST_RANGE_COUNT); + + inode = kunit_kzalloc(test, sizeof(*inode), GFP_KERNEL); + if (!inode) + return; + inode->i_sb = sb; + + ac.ac_status = AC_STATUS_FOUND; + ac.ac_sb = sb; + ac.ac_inode = inode; + for (i = 0; i < TEST_RANGE_COUNT; i++) + test_mark_diskspace_used_range(test, &ac, ranges[i].start, + ranges[i].len); +} + +static void mbt_generate_buddy(struct super_block *sb, void *buddy, + void *bitmap, struct ext4_group_info *grp) +{ + struct ext4_sb_info *sbi = EXT4_SB(sb); + uint32_t order, off; + void *bb, *bb_h; + int max; + + memset(buddy, 0xff, sb->s_blocksize); + memset(grp, 0, offsetof(struct ext4_group_info, + bb_counters[MB_NUM_ORDERS(sb)])); + + bb = bitmap; + max = EXT4_CLUSTERS_PER_GROUP(sb); + bb_h = buddy + sbi->s_mb_offsets[1]; + + off = mb_find_next_zero_bit(bb, max, 0); + grp->bb_first_free = off; + while (off < max) { + grp->bb_counters[0]++; + grp->bb_free++; + + if (!(off & 1) && !mb_test_bit(off + 1, bb)) { + grp->bb_free++; + grp->bb_counters[0]--; + mb_clear_bit(off >> 1, bb_h); + grp->bb_counters[1]++; + grp->bb_largest_free_order = 1; + off++; + } + + off = mb_find_next_zero_bit(bb, max, off + 1); + } + + for (order = 1; order < MB_NUM_ORDERS(sb) - 1; order++) { + bb = buddy + sbi->s_mb_offsets[order]; + bb_h = buddy + sbi->s_mb_offsets[order + 1]; + max = max >> 1; + off = mb_find_next_zero_bit(bb, max, 0); + + while (off < max) { + if (!(off & 1) && !mb_test_bit(off + 1, bb)) { + mb_set_bits(bb, off, 2); + grp->bb_counters[order] -= 2; + mb_clear_bit(off >> 1, bb_h); + grp->bb_counters[order + 1]++; + grp->bb_largest_free_order = order + 1; + off++; + } + + off = mb_find_next_zero_bit(bb, max, off + 1); + } + } + + max = EXT4_CLUSTERS_PER_GROUP(sb); + off = mb_find_next_zero_bit(bitmap, max, 0); + while (off < max) { + grp->bb_fragments++; + + off = mb_find_next_bit(bitmap, max, off + 1); + if (off + 1 >= max) + break; + + off = mb_find_next_zero_bit(bitmap, max, off + 1); + } +} + +static void +mbt_validate_group_info(struct kunit *test, struct ext4_group_info *grp1, + struct ext4_group_info *grp2) +{ + struct super_block *sb = (struct super_block *)test->priv; + int i; + + KUNIT_ASSERT_EQ(test, grp1->bb_first_free, + grp2->bb_first_free); + KUNIT_ASSERT_EQ(test, grp1->bb_fragments, + grp2->bb_fragments); + KUNIT_ASSERT_EQ(test, grp1->bb_free, grp2->bb_free); + KUNIT_ASSERT_EQ(test, grp1->bb_largest_free_order, + grp2->bb_largest_free_order); + + for (i = 1; i < MB_NUM_ORDERS(sb); i++) { + KUNIT_ASSERT_EQ_MSG(test, grp1->bb_counters[i], + grp2->bb_counters[i], + "bb_counters[%d] diffs, expected %d, generated %d", + i, grp1->bb_counters[i], + grp2->bb_counters[i]); + } +} + +static void +do_test_generate_buddy(struct kunit *test, struct super_block *sb, void *bitmap, + void *mbt_buddy, struct ext4_group_info *mbt_grp, + void *ext4_buddy, struct ext4_group_info *ext4_grp) +{ + int i; + + mbt_generate_buddy(sb, mbt_buddy, bitmap, mbt_grp); + + for (i = 0; i < MB_NUM_ORDERS(sb); i++) + ext4_grp->bb_counters[i] = 0; + /* needed by validation in ext4_mb_generate_buddy */ + ext4_grp->bb_free = mbt_grp->bb_free; + memset(ext4_buddy, 0xff, sb->s_blocksize); + ext4_mb_generate_buddy(sb, ext4_buddy, bitmap, TEST_GOAL_GROUP, + ext4_grp); + + KUNIT_ASSERT_EQ(test, memcmp(mbt_buddy, ext4_buddy, sb->s_blocksize), + 0); + mbt_validate_group_info(test, mbt_grp, ext4_grp); +} + +static void test_mb_generate_buddy(struct kunit *test) +{ + struct super_block *sb = (struct super_block *)test->priv; + void *bitmap, *expected_bb, *generate_bb; + struct ext4_group_info *expected_grp, *generate_grp; + struct test_range ranges[TEST_RANGE_COUNT]; + int i; + + bitmap = kunit_kzalloc(test, sb->s_blocksize, GFP_KERNEL); + KUNIT_ASSERT_NOT_ERR_OR_NULL(test, bitmap); + expected_bb = kunit_kzalloc(test, sb->s_blocksize, GFP_KERNEL); + KUNIT_ASSERT_NOT_ERR_OR_NULL(test, expected_bb); + generate_bb = kunit_kzalloc(test, sb->s_blocksize, GFP_KERNEL); + KUNIT_ASSERT_NOT_ERR_OR_NULL(test, generate_bb); + expected_grp = kunit_kzalloc(test, offsetof(struct ext4_group_info, + bb_counters[MB_NUM_ORDERS(sb)]), GFP_KERNEL); + KUNIT_ASSERT_NOT_ERR_OR_NULL(test, expected_grp); + generate_grp = ext4_get_group_info(sb, TEST_GOAL_GROUP); + KUNIT_ASSERT_NOT_NULL(test, generate_grp); + + mbt_generate_test_ranges(sb, ranges, TEST_RANGE_COUNT); + for (i = 0; i < TEST_RANGE_COUNT; i++) { + mb_set_bits(bitmap, ranges[i].start, ranges[i].len); + do_test_generate_buddy(test, sb, bitmap, expected_bb, + expected_grp, generate_bb, generate_grp); + } +} + +static void +test_mb_mark_used_range(struct kunit *test, struct ext4_buddy *e4b, + ext4_grpblk_t start, ext4_grpblk_t len, void *bitmap, + void *buddy, struct ext4_group_info *grp) +{ + struct super_block *sb = (struct super_block *)test->priv; + struct ext4_free_extent ex; + int i; + + /* mb_mark_used only accepts non-zero len */ + if (len == 0) + return; + + ex.fe_start = start; + ex.fe_len = len; + ex.fe_group = TEST_GOAL_GROUP; + + ext4_lock_group(sb, TEST_GOAL_GROUP); + mb_mark_used(e4b, &ex); + ext4_unlock_group(sb, TEST_GOAL_GROUP); + + mb_set_bits(bitmap, start, len); + /* bypass bb_free validatoin in ext4_mb_generate_buddy */ + grp->bb_free -= len; + memset(buddy, 0xff, sb->s_blocksize); + for (i = 0; i < MB_NUM_ORDERS(sb); i++) + grp->bb_counters[i] = 0; + ext4_mb_generate_buddy(sb, buddy, bitmap, 0, grp); + + KUNIT_ASSERT_EQ(test, memcmp(buddy, e4b->bd_buddy, sb->s_blocksize), + 0); + mbt_validate_group_info(test, grp, e4b->bd_info); +} + +static void test_mb_mark_used(struct kunit *test) +{ + struct ext4_buddy e4b; + struct super_block *sb = (struct super_block *)test->priv; + void *bitmap, *buddy; + struct ext4_group_info *grp; + int ret; + struct test_range ranges[TEST_RANGE_COUNT]; + int i; + + /* buddy cache assumes that each page contains at least one block */ + if (sb->s_blocksize > PAGE_SIZE) + kunit_skip(test, "blocksize exceeds pagesize"); + + bitmap = kunit_kzalloc(test, sb->s_blocksize, GFP_KERNEL); + KUNIT_ASSERT_NOT_ERR_OR_NULL(test, bitmap); + buddy = kunit_kzalloc(test, sb->s_blocksize, GFP_KERNEL); + KUNIT_ASSERT_NOT_ERR_OR_NULL(test, buddy); + grp = kunit_kzalloc(test, offsetof(struct ext4_group_info, + bb_counters[MB_NUM_ORDERS(sb)]), GFP_KERNEL); + + ret = ext4_mb_load_buddy(sb, TEST_GOAL_GROUP, &e4b); + KUNIT_ASSERT_EQ(test, ret, 0); + + grp->bb_free = EXT4_CLUSTERS_PER_GROUP(sb); + mbt_generate_test_ranges(sb, ranges, TEST_RANGE_COUNT); + for (i = 0; i < TEST_RANGE_COUNT; i++) + test_mb_mark_used_range(test, &e4b, ranges[i].start, + ranges[i].len, bitmap, buddy, grp); + + ext4_mb_unload_buddy(&e4b); +} + +static void +test_mb_free_blocks_range(struct kunit *test, struct ext4_buddy *e4b, + ext4_grpblk_t start, ext4_grpblk_t len, void *bitmap, + void *buddy, struct ext4_group_info *grp) +{ + struct super_block *sb = (struct super_block *)test->priv; + int i; + + /* mb_free_blocks will WARN if len is 0 */ + if (len == 0) + return; + + ext4_lock_group(sb, e4b->bd_group); + mb_free_blocks(NULL, e4b, start, len); + ext4_unlock_group(sb, e4b->bd_group); + + mb_clear_bits(bitmap, start, len); + /* bypass bb_free validatoin in ext4_mb_generate_buddy */ + grp->bb_free += len; + memset(buddy, 0xff, sb->s_blocksize); + for (i = 0; i < MB_NUM_ORDERS(sb); i++) + grp->bb_counters[i] = 0; + ext4_mb_generate_buddy(sb, buddy, bitmap, 0, grp); + + KUNIT_ASSERT_EQ(test, memcmp(buddy, e4b->bd_buddy, sb->s_blocksize), + 0); + mbt_validate_group_info(test, grp, e4b->bd_info); + +} + +static void test_mb_free_blocks(struct kunit *test) +{ + struct ext4_buddy e4b; + struct super_block *sb = (struct super_block *)test->priv; + void *bitmap, *buddy; + struct ext4_group_info *grp; + struct ext4_free_extent ex; + int ret; + int i; + struct test_range ranges[TEST_RANGE_COUNT]; + + /* buddy cache assumes that each page contains at least one block */ + if (sb->s_blocksize > PAGE_SIZE) + kunit_skip(test, "blocksize exceeds pagesize"); + + bitmap = kunit_kzalloc(test, sb->s_blocksize, GFP_KERNEL); + KUNIT_ASSERT_NOT_ERR_OR_NULL(test, bitmap); + buddy = kunit_kzalloc(test, sb->s_blocksize, GFP_KERNEL); + KUNIT_ASSERT_NOT_ERR_OR_NULL(test, buddy); + grp = kunit_kzalloc(test, offsetof(struct ext4_group_info, + bb_counters[MB_NUM_ORDERS(sb)]), GFP_KERNEL); + + ret = ext4_mb_load_buddy(sb, TEST_GOAL_GROUP, &e4b); + KUNIT_ASSERT_EQ(test, ret, 0); + + ex.fe_start = 0; + ex.fe_len = EXT4_CLUSTERS_PER_GROUP(sb); + ex.fe_group = TEST_GOAL_GROUP; + + ext4_lock_group(sb, TEST_GOAL_GROUP); + mb_mark_used(&e4b, &ex); + ext4_unlock_group(sb, TEST_GOAL_GROUP); + + grp->bb_free = 0; + memset(bitmap, 0xff, sb->s_blocksize); + + mbt_generate_test_ranges(sb, ranges, TEST_RANGE_COUNT); + for (i = 0; i < TEST_RANGE_COUNT; i++) + test_mb_free_blocks_range(test, &e4b, ranges[i].start, + ranges[i].len, bitmap, buddy, grp); + + ext4_mb_unload_buddy(&e4b); +} + static const struct mbt_ext4_block_layout mbt_test_layouts[] = { { .blocksize_bits = 10, @@ -334,6 +896,11 @@ KUNIT_ARRAY_PARAM(mbt_layouts, mbt_test_layouts, mbt_show_layout); static struct kunit_case mbt_test_cases[] = { KUNIT_CASE_PARAM(test_new_blocks_simple, mbt_layouts_gen_params), + KUNIT_CASE_PARAM(test_free_blocks_simple, mbt_layouts_gen_params), + KUNIT_CASE_PARAM(test_mb_generate_buddy, mbt_layouts_gen_params), + KUNIT_CASE_PARAM(test_mb_mark_used, mbt_layouts_gen_params), + KUNIT_CASE_PARAM(test_mb_free_blocks, mbt_layouts_gen_params), + KUNIT_CASE_PARAM(test_mark_diskspace_used, mbt_layouts_gen_params), {} }; diff --git a/fs/ext4/mballoc.c b/fs/ext4/mballoc.c index e4f7cf9d89c4..12b3f196010b 100644 --- a/fs/ext4/mballoc.c +++ b/fs/ext4/mballoc.c @@ -3015,8 +3015,8 @@ static int ext4_mb_seq_groups_show(struct seq_file *seq, void *v) { struct super_block *sb = pde_data(file_inode(seq->file)); ext4_group_t group = (ext4_group_t) ((unsigned long) v); - int i; - int err, buddy_loaded = 0; + int i, err; + char nbuf[16]; struct ext4_buddy e4b; struct ext4_group_info *grinfo; unsigned char blocksize_bits = min_t(unsigned char, @@ -3043,23 +3043,26 @@ static int ext4_mb_seq_groups_show(struct seq_file *seq, void *v) if (unlikely(EXT4_MB_GRP_NEED_INIT(grinfo))) { err = ext4_mb_load_buddy(sb, group, &e4b); if (err) { - seq_printf(seq, "#%-5u: I/O error\n", group); + seq_printf(seq, "#%-5u: %s\n", group, ext4_decode_error(NULL, err, nbuf)); return 0; } - buddy_loaded = 1; + ext4_mb_unload_buddy(&e4b); } + /* + * We care only about free space counters in the group info and + * these are safe to access even after the buddy has been unloaded + */ memcpy(&sg, grinfo, i); - - if (buddy_loaded) - ext4_mb_unload_buddy(&e4b); - seq_printf(seq, "#%-5u: %-5u %-5u %-5u [", group, sg.info.bb_free, sg.info.bb_fragments, sg.info.bb_first_free); for (i = 0; i <= 13; i++) seq_printf(seq, " %-5u", i <= blocksize_bits + 1 ? sg.info.bb_counters[i] : 0); - seq_puts(seq, " ]\n"); + seq_puts(seq, " ]"); + if (EXT4_MB_GRP_BBITMAP_CORRUPT(&sg.info)) + seq_puts(seq, " Block bitmap corrupted!"); + seq_puts(seq, "\n"); return 0; } @@ -3829,8 +3832,7 @@ void ext4_mb_release(struct super_block *sb) } static inline int ext4_issue_discard(struct super_block *sb, - ext4_group_t block_group, ext4_grpblk_t cluster, int count, - struct bio **biop) + ext4_group_t block_group, ext4_grpblk_t cluster, int count) { ext4_fsblk_t discard_block; @@ -3839,13 +3841,8 @@ static inline int ext4_issue_discard(struct super_block *sb, count = EXT4_C2B(EXT4_SB(sb), count); trace_ext4_discard_blocks(sb, (unsigned long long) discard_block, count); - if (biop) { - return __blkdev_issue_discard(sb->s_bdev, - (sector_t)discard_block << (sb->s_blocksize_bits - 9), - (sector_t)count << (sb->s_blocksize_bits - 9), - GFP_NOFS, biop); - } else - return sb_issue_discard(sb, discard_block, count, GFP_NOFS, 0); + + return sb_issue_discard(sb, discard_block, count, GFP_NOFS, 0); } static void ext4_free_data_in_buddy(struct super_block *sb, @@ -5169,10 +5166,16 @@ ext4_mb_new_inode_pa(struct ext4_allocation_context *ac) .fe_len = ac->ac_orig_goal_len, }; loff_t orig_goal_end = extent_logical_end(sbi, &ex); + loff_t o_ex_end = extent_logical_end(sbi, &ac->ac_o_ex); - /* we can't allocate as much as normalizer wants. - * so, found space must get proper lstart - * to cover original request */ + /* + * We can't allocate as much as normalizer wants, so we try + * to get proper lstart to cover the original request, except + * when the goal doesn't cover the original request as below: + * + * orig_ex:2045/2055(10), isize:8417280 -> normalized:0/2048 + * best_ex:0/200(200) -> adjusted: 1848/2048(200) + */ BUG_ON(ac->ac_g_ex.fe_logical > ac->ac_o_ex.fe_logical); BUG_ON(ac->ac_g_ex.fe_len < ac->ac_o_ex.fe_len); @@ -5184,7 +5187,7 @@ ext4_mb_new_inode_pa(struct ext4_allocation_context *ac) * 1. Check if best ex can be kept at end of goal (before * cr_best_avail trimmed it) and still cover original start * 2. Else, check if best ex can be kept at start of goal and - * still cover original start + * still cover original end * 3. Else, keep the best ex at start of original request. */ ex.fe_len = ac->ac_b_ex.fe_len; @@ -5194,7 +5197,7 @@ ext4_mb_new_inode_pa(struct ext4_allocation_context *ac) goto adjust_bex; ex.fe_logical = ac->ac_g_ex.fe_logical; - if (ac->ac_o_ex.fe_logical < extent_logical_end(sbi, &ex)) + if (o_ex_end <= extent_logical_end(sbi, &ex)) goto adjust_bex; ex.fe_logical = ac->ac_o_ex.fe_logical; @@ -5202,7 +5205,6 @@ adjust_bex: ac->ac_b_ex.fe_logical = ex.fe_logical; BUG_ON(ac->ac_o_ex.fe_logical < ac->ac_b_ex.fe_logical); - BUG_ON(ac->ac_o_ex.fe_len > ac->ac_b_ex.fe_len); BUG_ON(extent_logical_end(sbi, &ex) > orig_goal_end); } @@ -6487,8 +6489,14 @@ do_more: } else { if (test_opt(sb, DISCARD)) { err = ext4_issue_discard(sb, block_group, bit, - count_clusters, NULL); - if (err && err != -EOPNOTSUPP) + count_clusters); + /* + * Ignore EOPNOTSUPP error. This is consistent with + * what happens when using journal. + */ + if (err == -EOPNOTSUPP) + err = 0; + if (err) ext4_msg(sb, KERN_WARNING, "discard request in" " group:%u block:%d count:%lu failed" " with %d", block_group, bit, count, @@ -6738,7 +6746,7 @@ __acquires(bitlock) */ mb_mark_used(e4b, &ex); ext4_unlock_group(sb, group); - ret = ext4_issue_discard(sb, group, start, count, NULL); + ret = ext4_issue_discard(sb, group, start, count); ext4_lock_group(sb, group); mb_free_blocks(NULL, e4b, start, ex.fe_len); return ret; diff --git a/fs/ext4/resize.c b/fs/ext4/resize.c index 4d4a5a32e310..0ba9837d65ca 100644 --- a/fs/ext4/resize.c +++ b/fs/ext4/resize.c @@ -1602,7 +1602,8 @@ exit_journal: int gdb_num = group / EXT4_DESC_PER_BLOCK(sb); int gdb_num_end = ((group + flex_gd->count - 1) / EXT4_DESC_PER_BLOCK(sb)); - int meta_bg = ext4_has_feature_meta_bg(sb); + int meta_bg = ext4_has_feature_meta_bg(sb) && + gdb_num >= le32_to_cpu(es->s_first_meta_bg); sector_t padding_blocks = meta_bg ? 0 : sbi->s_sbh->b_blocknr - ext4_group_first_block_no(sb, 0); @@ -2084,7 +2085,7 @@ retry: } } - if ((!resize_inode && !meta_bg) || n_blocks_count == o_blocks_count) { + if ((!resize_inode && !meta_bg && n_desc_blocks > o_desc_blocks) || n_blocks_count == o_blocks_count) { err = ext4_convert_meta_bg(sb, resize_inode); if (err) goto out; diff --git a/fs/ext4/super.c b/fs/ext4/super.c index 59c72b6dd153..cfb8449c731f 100644 --- a/fs/ext4/super.c +++ b/fs/ext4/super.c @@ -1599,7 +1599,7 @@ static ssize_t ext4_quota_write(struct super_block *sb, int type, static int ext4_quota_enable(struct super_block *sb, int type, int format_id, unsigned int flags); -static struct dquot **ext4_get_dquots(struct inode *inode) +static struct dquot __rcu **ext4_get_dquots(struct inode *inode) { return EXT4_I(inode)->i_dquot; } @@ -4421,22 +4421,6 @@ static int ext4_handle_clustersize(struct super_block *sb) } sbi->s_cluster_bits = le32_to_cpu(es->s_log_cluster_size) - le32_to_cpu(es->s_log_block_size); - sbi->s_clusters_per_group = - le32_to_cpu(es->s_clusters_per_group); - if (sbi->s_clusters_per_group > sb->s_blocksize * 8) { - ext4_msg(sb, KERN_ERR, - "#clusters per group too big: %lu", - sbi->s_clusters_per_group); - return -EINVAL; - } - if (sbi->s_blocks_per_group != - (sbi->s_clusters_per_group * (clustersize / sb->s_blocksize))) { - ext4_msg(sb, KERN_ERR, "blocks per group (%lu) and " - "clusters per group (%lu) inconsistent", - sbi->s_blocks_per_group, - sbi->s_clusters_per_group); - return -EINVAL; - } } else { if (clustersize != sb->s_blocksize) { ext4_msg(sb, KERN_ERR, @@ -4450,9 +4434,21 @@ static int ext4_handle_clustersize(struct super_block *sb) sbi->s_blocks_per_group); return -EINVAL; } - sbi->s_clusters_per_group = sbi->s_blocks_per_group; sbi->s_cluster_bits = 0; } + sbi->s_clusters_per_group = le32_to_cpu(es->s_clusters_per_group); + if (sbi->s_clusters_per_group > sb->s_blocksize * 8) { + ext4_msg(sb, KERN_ERR, "#clusters per group too big: %lu", + sbi->s_clusters_per_group); + return -EINVAL; + } + if (sbi->s_blocks_per_group != + (sbi->s_clusters_per_group * (clustersize / sb->s_blocksize))) { + ext4_msg(sb, KERN_ERR, + "blocks per group (%lu) and clusters per group (%lu) inconsistent", + sbi->s_blocks_per_group, sbi->s_clusters_per_group); + return -EINVAL; + } sbi->s_cluster_ratio = clustersize / sb->s_blocksize; /* Do we have standard group size of clustersize * 8 blocks ? */ @@ -6864,6 +6860,10 @@ static int ext4_write_dquot(struct dquot *dquot) if (IS_ERR(handle)) return PTR_ERR(handle); ret = dquot_commit(dquot); + if (ret < 0) + ext4_error_err(dquot->dq_sb, -ret, + "Failed to commit dquot type %d", + dquot->dq_id.type); err = ext4_journal_stop(handle); if (!ret) ret = err; @@ -6880,6 +6880,10 @@ static int ext4_acquire_dquot(struct dquot *dquot) if (IS_ERR(handle)) return PTR_ERR(handle); ret = dquot_acquire(dquot); + if (ret < 0) + ext4_error_err(dquot->dq_sb, -ret, + "Failed to acquire dquot type %d", + dquot->dq_id.type); err = ext4_journal_stop(handle); if (!ret) ret = err; @@ -6899,6 +6903,10 @@ static int ext4_release_dquot(struct dquot *dquot) return PTR_ERR(handle); } ret = dquot_release(dquot); + if (ret < 0) + ext4_error_err(dquot->dq_sb, -ret, + "Failed to release dquot type %d", + dquot->dq_id.type); err = ext4_journal_stop(handle); if (!ret) ret = err; diff --git a/fs/ext4/xattr.c b/fs/ext4/xattr.c index 82dc5e673d5c..b67a176bfcf9 100644 --- a/fs/ext4/xattr.c +++ b/fs/ext4/xattr.c @@ -1565,46 +1565,49 @@ ext4_xattr_inode_cache_find(struct inode *inode, const void *value, /* * Add value of the EA in an inode. */ -static int ext4_xattr_inode_lookup_create(handle_t *handle, struct inode *inode, - const void *value, size_t value_len, - struct inode **ret_inode) +static struct inode *ext4_xattr_inode_lookup_create(handle_t *handle, + struct inode *inode, const void *value, size_t value_len) { struct inode *ea_inode; u32 hash; int err; + /* Account inode & space to quota even if sharing... */ + err = ext4_xattr_inode_alloc_quota(inode, value_len); + if (err) + return ERR_PTR(err); + hash = ext4_xattr_inode_hash(EXT4_SB(inode->i_sb), value, value_len); ea_inode = ext4_xattr_inode_cache_find(inode, value, value_len, hash); if (ea_inode) { err = ext4_xattr_inode_inc_ref(handle, ea_inode); - if (err) { - iput(ea_inode); - return err; - } - - *ret_inode = ea_inode; - return 0; + if (err) + goto out_err; + return ea_inode; } /* Create an inode for the EA value */ ea_inode = ext4_xattr_inode_create(handle, inode, hash); - if (IS_ERR(ea_inode)) - return PTR_ERR(ea_inode); + if (IS_ERR(ea_inode)) { + ext4_xattr_inode_free_quota(inode, NULL, value_len); + return ea_inode; + } err = ext4_xattr_inode_write(handle, ea_inode, value, value_len); if (err) { if (ext4_xattr_inode_dec_ref(handle, ea_inode)) ext4_warning_inode(ea_inode, "cleanup dec ref error %d", err); - iput(ea_inode); - return err; + goto out_err; } if (EA_INODE_CACHE(inode)) mb_cache_entry_create(EA_INODE_CACHE(inode), GFP_NOFS, hash, ea_inode->i_ino, true /* reusable */); - - *ret_inode = ea_inode; - return 0; + return ea_inode; +out_err: + iput(ea_inode); + ext4_xattr_inode_free_quota(inode, NULL, value_len); + return ERR_PTR(err); } /* @@ -1712,16 +1715,11 @@ static int ext4_xattr_set_entry(struct ext4_xattr_info *i, if (i->value && in_inode) { WARN_ON_ONCE(!i->value_len); - ret = ext4_xattr_inode_alloc_quota(inode, i->value_len); - if (ret) - goto out; - - ret = ext4_xattr_inode_lookup_create(handle, inode, i->value, - i->value_len, - &new_ea_inode); - if (ret) { + new_ea_inode = ext4_xattr_inode_lookup_create(handle, inode, + i->value, i->value_len); + if (IS_ERR(new_ea_inode)) { + ret = PTR_ERR(new_ea_inode); new_ea_inode = NULL; - ext4_xattr_inode_free_quota(inode, NULL, i->value_len); goto out; } } @@ -2160,17 +2158,6 @@ getblk_failed: ENTRY(header(s->base)+1)); if (error) goto getblk_failed; - if (ea_inode) { - /* Drop the extra ref on ea_inode. */ - error = ext4_xattr_inode_dec_ref(handle, - ea_inode); - if (error) - ext4_warning_inode(ea_inode, - "dec ref error=%d", - error); - iput(ea_inode); - ea_inode = NULL; - } lock_buffer(new_bh); error = ext4_journal_get_create_access(handle, sb, diff --git a/fs/f2fs/checkpoint.c b/fs/f2fs/checkpoint.c index b0597a539fc5..eac698b8dd38 100644 --- a/fs/f2fs/checkpoint.c +++ b/fs/f2fs/checkpoint.c @@ -154,49 +154,47 @@ static bool __is_bitmap_valid(struct f2fs_sb_info *sbi, block_t blkaddr, if (unlikely(f2fs_cp_error(sbi))) return exist; - if (exist && type == DATA_GENERIC_ENHANCE_UPDATE) { - f2fs_err(sbi, "Inconsistent error blkaddr:%u, sit bitmap:%d", - blkaddr, exist); - set_sbi_flag(sbi, SBI_NEED_FSCK); - return exist; - } + if ((exist && type == DATA_GENERIC_ENHANCE_UPDATE) || + (!exist && type == DATA_GENERIC_ENHANCE)) + goto out_err; + if (!exist && type != DATA_GENERIC_ENHANCE_UPDATE) + goto out_handle; + return exist; - if (!exist && type == DATA_GENERIC_ENHANCE) { - f2fs_err(sbi, "Inconsistent error blkaddr:%u, sit bitmap:%d", - blkaddr, exist); - set_sbi_flag(sbi, SBI_NEED_FSCK); - dump_stack(); - } +out_err: + f2fs_err(sbi, "Inconsistent error blkaddr:%u, sit bitmap:%d", + blkaddr, exist); + set_sbi_flag(sbi, SBI_NEED_FSCK); + dump_stack(); +out_handle: + f2fs_handle_error(sbi, ERROR_INVALID_BLKADDR); return exist; } -bool f2fs_is_valid_blkaddr(struct f2fs_sb_info *sbi, +static bool __f2fs_is_valid_blkaddr(struct f2fs_sb_info *sbi, block_t blkaddr, int type) { - if (time_to_inject(sbi, FAULT_BLKADDR)) - return false; - switch (type) { case META_NAT: break; case META_SIT: if (unlikely(blkaddr >= SIT_BLK_CNT(sbi))) - return false; + goto err; break; case META_SSA: if (unlikely(blkaddr >= MAIN_BLKADDR(sbi) || blkaddr < SM_I(sbi)->ssa_blkaddr)) - return false; + goto err; break; case META_CP: if (unlikely(blkaddr >= SIT_I(sbi)->sit_base_addr || blkaddr < __start_cp_addr(sbi))) - return false; + goto err; break; case META_POR: if (unlikely(blkaddr >= MAX_BLKADDR(sbi) || blkaddr < MAIN_BLKADDR(sbi))) - return false; + goto err; break; case DATA_GENERIC: case DATA_GENERIC_ENHANCE: @@ -213,7 +211,7 @@ bool f2fs_is_valid_blkaddr(struct f2fs_sb_info *sbi, blkaddr); set_sbi_flag(sbi, SBI_NEED_FSCK); dump_stack(); - return false; + goto err; } else { return __is_bitmap_valid(sbi, blkaddr, type); } @@ -221,13 +219,30 @@ bool f2fs_is_valid_blkaddr(struct f2fs_sb_info *sbi, case META_GENERIC: if (unlikely(blkaddr < SEG0_BLKADDR(sbi) || blkaddr >= MAIN_BLKADDR(sbi))) - return false; + goto err; break; default: BUG(); } return true; +err: + f2fs_handle_error(sbi, ERROR_INVALID_BLKADDR); + return false; +} + +bool f2fs_is_valid_blkaddr(struct f2fs_sb_info *sbi, + block_t blkaddr, int type) +{ + if (time_to_inject(sbi, FAULT_BLKADDR_VALIDITY)) + return false; + return __f2fs_is_valid_blkaddr(sbi, blkaddr, type); +} + +bool f2fs_is_valid_blkaddr_raw(struct f2fs_sb_info *sbi, + block_t blkaddr, int type) +{ + return __f2fs_is_valid_blkaddr(sbi, blkaddr, type); } /* @@ -889,7 +904,7 @@ static struct page *validate_checkpoint(struct f2fs_sb_info *sbi, cp_blocks = le32_to_cpu(cp_block->cp_pack_total_block_count); - if (cp_blocks > sbi->blocks_per_seg || cp_blocks <= F2FS_CP_PACKS) { + if (cp_blocks > BLKS_PER_SEG(sbi) || cp_blocks <= F2FS_CP_PACKS) { f2fs_warn(sbi, "invalid cp_pack_total_block_count:%u", le32_to_cpu(cp_block->cp_pack_total_block_count)); goto invalid_cp; @@ -1324,7 +1339,7 @@ static void update_ckpt_flags(struct f2fs_sb_info *sbi, struct cp_control *cpc) if (cpc->reason & CP_UMOUNT) { if (le32_to_cpu(ckpt->cp_pack_total_block_count) + - NM_I(sbi)->nat_bits_blocks > sbi->blocks_per_seg) { + NM_I(sbi)->nat_bits_blocks > BLKS_PER_SEG(sbi)) { clear_ckpt_flags(sbi, CP_NAT_BITS_FLAG); f2fs_notice(sbi, "Disable nat_bits due to no space"); } else if (!is_set_ckpt_flags(sbi, CP_NAT_BITS_FLAG) && @@ -1527,7 +1542,7 @@ static int do_checkpoint(struct f2fs_sb_info *sbi, struct cp_control *cpc) cp_ver |= ((__u64)crc32 << 32); *(__le64 *)nm_i->nat_bits = cpu_to_le64(cp_ver); - blk = start_blk + sbi->blocks_per_seg - nm_i->nat_bits_blocks; + blk = start_blk + BLKS_PER_SEG(sbi) - nm_i->nat_bits_blocks; for (i = 0; i < nm_i->nat_bits_blocks; i++) f2fs_update_meta_page(sbi, nm_i->nat_bits + (i << F2FS_BLKSIZE_BITS), blk + i); @@ -1587,8 +1602,9 @@ static int do_checkpoint(struct f2fs_sb_info *sbi, struct cp_control *cpc) */ if (f2fs_sb_has_encrypt(sbi) || f2fs_sb_has_verity(sbi) || f2fs_sb_has_compression(sbi)) - invalidate_mapping_pages(META_MAPPING(sbi), - MAIN_BLKADDR(sbi), MAX_BLKADDR(sbi) - 1); + f2fs_bug_on(sbi, + invalidate_inode_pages2_range(META_MAPPING(sbi), + MAIN_BLKADDR(sbi), MAX_BLKADDR(sbi) - 1)); f2fs_release_ino_entry(sbi, false); @@ -1730,9 +1746,9 @@ void f2fs_init_ino_entry_info(struct f2fs_sb_info *sbi) im->ino_num = 0; } - sbi->max_orphans = (sbi->blocks_per_seg - F2FS_CP_PACKS - + sbi->max_orphans = (BLKS_PER_SEG(sbi) - F2FS_CP_PACKS - NR_CURSEG_PERSIST_TYPE - __cp_payload(sbi)) * - F2FS_ORPHANS_PER_BLOCK; + F2FS_ORPHANS_PER_BLOCK; } int __init f2fs_create_checkpoint_caches(void) diff --git a/fs/f2fs/compress.c b/fs/f2fs/compress.c index 531517dac079..8892c8262141 100644 --- a/fs/f2fs/compress.c +++ b/fs/f2fs/compress.c @@ -512,8 +512,8 @@ static int lzorle_compress_pages(struct compress_ctx *cc) ret = lzorle1x_1_compress(cc->rbuf, cc->rlen, cc->cbuf->cdata, &cc->clen, cc->private); if (ret != LZO_E_OK) { - printk_ratelimited("%sF2FS-fs (%s): lzo-rle compress failed, ret:%d\n", - KERN_ERR, F2FS_I_SB(cc->inode)->sb->s_id, ret); + f2fs_err_ratelimited(F2FS_I_SB(cc->inode), + "lzo-rle compress failed, ret:%d", ret); return -EIO; } return 0; @@ -780,9 +780,9 @@ void f2fs_decompress_cluster(struct decompress_io_ctx *dic, bool in_task) if (provided != calculated) { if (!is_inode_flag_set(dic->inode, FI_COMPRESS_CORRUPT)) { set_inode_flag(dic->inode, FI_COMPRESS_CORRUPT); - printk_ratelimited( - "%sF2FS-fs (%s): checksum invalid, nid = %lu, %x vs %x", - KERN_INFO, sbi->sb->s_id, dic->inode->i_ino, + f2fs_info_ratelimited(sbi, + "checksum invalid, nid = %lu, %x vs %x", + dic->inode->i_ino, provided, calculated); } set_sbi_flag(sbi, SBI_NEED_FSCK); @@ -1418,6 +1418,8 @@ void f2fs_compress_write_end_io(struct bio *bio, struct page *page) struct f2fs_sb_info *sbi = bio->bi_private; struct compress_io_ctx *cic = (struct compress_io_ctx *)page_private(page); + enum count_type type = WB_DATA_TYPE(page, + f2fs_is_compressed_page(page)); int i; if (unlikely(bio->bi_status)) @@ -1425,7 +1427,7 @@ void f2fs_compress_write_end_io(struct bio *bio, struct page *page) f2fs_compress_free_page(page); - dec_page_count(sbi, F2FS_WB_DATA); + dec_page_count(sbi, type); if (atomic_dec_return(&cic->pending_pages)) return; @@ -1441,12 +1443,14 @@ void f2fs_compress_write_end_io(struct bio *bio, struct page *page) } static int f2fs_write_raw_pages(struct compress_ctx *cc, - int *submitted, + int *submitted_p, struct writeback_control *wbc, enum iostat_type io_type) { struct address_space *mapping = cc->inode->i_mapping; - int _submitted, compr_blocks, ret, i; + struct f2fs_sb_info *sbi = F2FS_M_SB(mapping); + int submitted, compr_blocks, i; + int ret = 0; compr_blocks = f2fs_compressed_blocks(cc); @@ -1461,6 +1465,10 @@ static int f2fs_write_raw_pages(struct compress_ctx *cc, if (compr_blocks < 0) return compr_blocks; + /* overwrite compressed cluster w/ normal cluster */ + if (compr_blocks > 0) + f2fs_lock_op(sbi); + for (i = 0; i < cc->cluster_size; i++) { if (!cc->rpages[i]) continue; @@ -1485,7 +1493,7 @@ continue_unlock: if (!clear_page_dirty_for_io(cc->rpages[i])) goto continue_unlock; - ret = f2fs_write_single_data_page(cc->rpages[i], &_submitted, + ret = f2fs_write_single_data_page(cc->rpages[i], &submitted, NULL, NULL, wbc, io_type, compr_blocks, false); if (ret) { @@ -1493,26 +1501,29 @@ continue_unlock: unlock_page(cc->rpages[i]); ret = 0; } else if (ret == -EAGAIN) { + ret = 0; /* * for quota file, just redirty left pages to * avoid deadlock caused by cluster update race * from foreground operation. */ if (IS_NOQUOTA(cc->inode)) - return 0; - ret = 0; + goto out; f2fs_io_schedule_timeout(DEFAULT_IO_TIMEOUT); goto retry_write; } - return ret; + goto out; } - *submitted += _submitted; + *submitted_p += submitted; } - f2fs_balance_fs(F2FS_M_SB(mapping), true); +out: + if (compr_blocks > 0) + f2fs_unlock_op(sbi); - return 0; + f2fs_balance_fs(sbi, true); + return ret; } int f2fs_write_multi_pages(struct compress_ctx *cc, @@ -1806,16 +1817,18 @@ void f2fs_put_page_dic(struct page *page, bool in_task) * check whether cluster blocks are contiguous, and add extent cache entry * only if cluster blocks are logically and physically contiguous. */ -unsigned int f2fs_cluster_blocks_are_contiguous(struct dnode_of_data *dn) +unsigned int f2fs_cluster_blocks_are_contiguous(struct dnode_of_data *dn, + unsigned int ofs_in_node) { - bool compressed = f2fs_data_blkaddr(dn) == COMPRESS_ADDR; + bool compressed = data_blkaddr(dn->inode, dn->node_page, + ofs_in_node) == COMPRESS_ADDR; int i = compressed ? 1 : 0; block_t first_blkaddr = data_blkaddr(dn->inode, dn->node_page, - dn->ofs_in_node + i); + ofs_in_node + i); for (i += 1; i < F2FS_I(dn->inode)->i_cluster_size; i++) { block_t blkaddr = data_blkaddr(dn->inode, dn->node_page, - dn->ofs_in_node + i); + ofs_in_node + i); if (!__is_valid_data_blkaddr(blkaddr)) break; @@ -1878,12 +1891,8 @@ void f2fs_cache_compressed_page(struct f2fs_sb_info *sbi, struct page *page, set_page_private_data(cpage, ino); - if (!f2fs_is_valid_blkaddr(sbi, blkaddr, DATA_GENERIC_ENHANCE_READ)) - goto out; - memcpy(page_address(cpage), page_address(page), PAGE_SIZE); SetPageUptodate(cpage); -out: f2fs_put_page(cpage, 1); } diff --git a/fs/f2fs/data.c b/fs/f2fs/data.c index 26e317696b33..d9494b5fc7c1 100644 --- a/fs/f2fs/data.c +++ b/fs/f2fs/data.c @@ -48,7 +48,7 @@ void f2fs_destroy_bioset(void) bioset_exit(&f2fs_bioset); } -static bool __is_cp_guaranteed(struct page *page) +bool f2fs_is_cp_guaranteed(struct page *page) { struct address_space *mapping = page->mapping; struct inode *inode; @@ -65,8 +65,6 @@ static bool __is_cp_guaranteed(struct page *page) S_ISDIR(inode->i_mode)) return true; - if (f2fs_is_compressed_page(page)) - return false; if ((S_ISREG(inode->i_mode) && IS_NOQUOTA(inode)) || page_private_gcing(page)) return true; @@ -338,18 +336,7 @@ static void f2fs_write_end_io(struct bio *bio) bio_for_each_segment_all(bvec, bio, iter_all) { struct page *page = bvec->bv_page; - enum count_type type = WB_DATA_TYPE(page); - - if (page_private_dummy(page)) { - clear_page_private_dummy(page); - unlock_page(page); - mempool_free(page, sbi->write_io_dummy); - - if (unlikely(bio->bi_status)) - f2fs_stop_checkpoint(sbi, true, - STOP_CP_REASON_WRITE_FAIL); - continue; - } + enum count_type type = WB_DATA_TYPE(page, false); fscrypt_finalize_bounce_page(&page); @@ -524,50 +511,13 @@ void f2fs_submit_read_bio(struct f2fs_sb_info *sbi, struct bio *bio, submit_bio(bio); } -static void f2fs_align_write_bio(struct f2fs_sb_info *sbi, struct bio *bio) -{ - unsigned int start = - (bio->bi_iter.bi_size >> F2FS_BLKSIZE_BITS) % F2FS_IO_SIZE(sbi); - - if (start == 0) - return; - - /* fill dummy pages */ - for (; start < F2FS_IO_SIZE(sbi); start++) { - struct page *page = - mempool_alloc(sbi->write_io_dummy, - GFP_NOIO | __GFP_NOFAIL); - f2fs_bug_on(sbi, !page); - - lock_page(page); - - zero_user_segment(page, 0, PAGE_SIZE); - set_page_private_dummy(page); - - if (bio_add_page(bio, page, PAGE_SIZE, 0) < PAGE_SIZE) - f2fs_bug_on(sbi, 1); - } -} - static void f2fs_submit_write_bio(struct f2fs_sb_info *sbi, struct bio *bio, enum page_type type) { WARN_ON_ONCE(is_read_io(bio_op(bio))); - if (type == DATA || type == NODE) { - if (f2fs_lfs_mode(sbi) && current->plug) - blk_finish_plug(current->plug); - - if (F2FS_IO_ALIGNED(sbi)) { - f2fs_align_write_bio(sbi, bio); - /* - * In the NODE case, we lose next block address chain. - * So, we need to do checkpoint in f2fs_sync_file. - */ - if (type == NODE) - set_sbi_flag(sbi, SBI_NEED_CP); - } - } + if (f2fs_lfs_mode(sbi) && current->plug && PAGE_TYPE_ON_MAIN(type)) + blk_finish_plug(current->plug); trace_f2fs_submit_write_bio(sbi->sb, type, bio); iostat_update_submit_ctx(bio, type); @@ -740,10 +690,8 @@ int f2fs_submit_page_bio(struct f2fs_io_info *fio) if (!f2fs_is_valid_blkaddr(fio->sbi, fio->new_blkaddr, fio->is_por ? META_POR : (__is_meta_io(fio) ? - META_GENERIC : DATA_GENERIC_ENHANCE))) { - f2fs_handle_error(fio->sbi, ERROR_INVALID_BLKADDR); + META_GENERIC : DATA_GENERIC_ENHANCE))) return -EFSCORRUPTED; - } trace_f2fs_submit_page_bio(page, fio); @@ -762,7 +710,7 @@ int f2fs_submit_page_bio(struct f2fs_io_info *fio) wbc_account_cgroup_owner(fio->io_wbc, fio->page, PAGE_SIZE); inc_page_count(fio->sbi, is_read_io(fio->op) ? - __read_io_type(page) : WB_DATA_TYPE(fio->page)); + __read_io_type(page) : WB_DATA_TYPE(fio->page, false)); if (is_read_io(bio_op(bio))) f2fs_submit_read_bio(fio->sbi, bio, fio->type); @@ -796,16 +744,6 @@ static bool io_is_mergeable(struct f2fs_sb_info *sbi, struct bio *bio, block_t last_blkaddr, block_t cur_blkaddr) { - if (F2FS_IO_ALIGNED(sbi) && (fio->type == DATA || fio->type == NODE)) { - unsigned int filled_blocks = - F2FS_BYTES_TO_BLK(bio->bi_iter.bi_size); - unsigned int io_size = F2FS_IO_SIZE(sbi); - unsigned int left_vecs = bio->bi_max_vecs - bio->bi_vcnt; - - /* IOs in bio is aligned and left space of vectors is not enough */ - if (!(filled_blocks % io_size) && left_vecs < io_size) - return false; - } if (!page_is_mergeable(sbi, bio, last_blkaddr, cur_blkaddr)) return false; return io_type_is_mergeable(io, fio); @@ -948,10 +886,8 @@ int f2fs_merge_page_bio(struct f2fs_io_info *fio) fio->encrypted_page : fio->page; if (!f2fs_is_valid_blkaddr(fio->sbi, fio->new_blkaddr, - __is_meta_io(fio) ? META_GENERIC : DATA_GENERIC)) { - f2fs_handle_error(fio->sbi, ERROR_INVALID_BLKADDR); + __is_meta_io(fio) ? META_GENERIC : DATA_GENERIC)) return -EFSCORRUPTED; - } trace_f2fs_submit_page_bio(page, fio); @@ -973,7 +909,7 @@ alloc_new: if (fio->io_wbc) wbc_account_cgroup_owner(fio->io_wbc, fio->page, PAGE_SIZE); - inc_page_count(fio->sbi, WB_DATA_TYPE(page)); + inc_page_count(fio->sbi, WB_DATA_TYPE(page, false)); *fio->last_block = fio->new_blkaddr; *fio->bio = bio; @@ -1007,11 +943,12 @@ void f2fs_submit_page_write(struct f2fs_io_info *fio) enum page_type btype = PAGE_TYPE_OF_BIO(fio->type); struct f2fs_bio_info *io = sbi->write_io[btype] + fio->temp; struct page *bio_page; + enum count_type type; f2fs_bug_on(sbi, is_read_io(fio->op)); f2fs_down_write(&io->io_rwsem); - +next: #ifdef CONFIG_BLK_DEV_ZONED if (f2fs_sb_has_blkzoned(sbi) && btype < META && io->zone_pending_bio) { wait_for_completion_io(&io->zone_wait); @@ -1021,7 +958,6 @@ void f2fs_submit_page_write(struct f2fs_io_info *fio) } #endif -next: if (fio->in_list) { spin_lock(&io->io_lock); if (list_empty(&io->io_list)) { @@ -1046,7 +982,8 @@ next: /* set submitted = true as a return value */ fio->submitted = 1; - inc_page_count(sbi, WB_DATA_TYPE(bio_page)); + type = WB_DATA_TYPE(bio_page, fio->compressed_page); + inc_page_count(sbi, type); if (io->bio && (!io_is_mergeable(sbi, io->bio, io, fio, io->last_block_in_bio, @@ -1056,13 +993,6 @@ next: __submit_merged_bio(io); alloc_new: if (io->bio == NULL) { - if (F2FS_IO_ALIGNED(sbi) && - (fio->type == DATA || fio->type == NODE) && - fio->new_blkaddr & F2FS_IO_SIZE_MASK(sbi)) { - dec_page_count(sbi, WB_DATA_TYPE(bio_page)); - fio->retry = 1; - goto skip; - } io->bio = __bio_alloc(fio, BIO_MAX_VECS); f2fs_set_bio_crypt_ctx(io->bio, fio->page->mapping->host, bio_page->index, fio, GFP_NOIO); @@ -1080,10 +1010,6 @@ alloc_new: io->last_block_in_bio = fio->new_blkaddr; trace_f2fs_submit_page_write(fio->page, fio); -skip: - if (fio->in_list) - goto next; -out: #ifdef CONFIG_BLK_DEV_ZONED if (f2fs_sb_has_blkzoned(sbi) && btype < META && is_end_zone_blkaddr(sbi, fio->new_blkaddr)) { @@ -1096,6 +1022,9 @@ out: __submit_merged_bio(io); } #endif + if (fio->in_list) + goto next; +out: if (is_sbi_flag_set(sbi, SBI_IS_SHUTDOWN) || !f2fs_is_checkpoint_ready(sbi)) __submit_merged_bio(io); @@ -1218,7 +1147,8 @@ int f2fs_reserve_new_blocks(struct dnode_of_data *dn, blkcnt_t count) if (unlikely(is_inode_flag_set(dn->inode, FI_NO_ALLOC))) return -EPERM; - if (unlikely((err = inc_valid_block_count(sbi, dn->inode, &count)))) + err = inc_valid_block_count(sbi, dn->inode, &count, true); + if (unlikely(err)) return err; trace_f2fs_reserve_new_blocks(dn->inode, dn->nid, @@ -1285,8 +1215,6 @@ struct page *f2fs_get_read_data_page(struct inode *inode, pgoff_t index, if (!f2fs_is_valid_blkaddr(F2FS_I_SB(inode), dn.data_blkaddr, DATA_GENERIC_ENHANCE_READ)) { err = -EFSCORRUPTED; - f2fs_handle_error(F2FS_I_SB(inode), - ERROR_INVALID_BLKADDR); goto put_err; } goto got_it; @@ -1312,8 +1240,6 @@ struct page *f2fs_get_read_data_page(struct inode *inode, pgoff_t index, dn.data_blkaddr, DATA_GENERIC_ENHANCE)) { err = -EFSCORRUPTED; - f2fs_handle_error(F2FS_I_SB(inode), - ERROR_INVALID_BLKADDR); goto put_err; } got_it: @@ -1475,15 +1401,18 @@ static int __allocate_data_block(struct dnode_of_data *dn, int seg_type) dn->data_blkaddr = f2fs_data_blkaddr(dn); if (dn->data_blkaddr == NULL_ADDR) { - err = inc_valid_block_count(sbi, dn->inode, &count); + err = inc_valid_block_count(sbi, dn->inode, &count, true); if (unlikely(err)) return err; } set_summary(&sum, dn->nid, dn->ofs_in_node, ni.version); old_blkaddr = dn->data_blkaddr; - f2fs_allocate_data_block(sbi, NULL, old_blkaddr, &dn->data_blkaddr, - &sum, seg_type, NULL); + err = f2fs_allocate_data_block(sbi, NULL, old_blkaddr, + &dn->data_blkaddr, &sum, seg_type, NULL); + if (err) + return err; + if (GET_SEGNO(sbi, old_blkaddr) != NULL_SEGNO) f2fs_invalidate_internal_cache(sbi, old_blkaddr); @@ -1641,7 +1570,6 @@ next_block: if (!is_hole && !f2fs_is_valid_blkaddr(sbi, blkaddr, DATA_GENERIC_ENHANCE)) { err = -EFSCORRUPTED; - f2fs_handle_error(sbi, ERROR_INVALID_BLKADDR); goto sync_out; } @@ -2165,8 +2093,6 @@ got_it: if (!f2fs_is_valid_blkaddr(F2FS_I_SB(inode), block_nr, DATA_GENERIC_ENHANCE_READ)) { ret = -EFSCORRUPTED; - f2fs_handle_error(F2FS_I_SB(inode), - ERROR_INVALID_BLKADDR); goto out; } } else { @@ -2668,8 +2594,6 @@ bool f2fs_should_update_outplace(struct inode *inode, struct f2fs_io_info *fio) if (fio) { if (page_private_gcing(fio->page)) return true; - if (page_private_dummy(fio->page)) - return true; if (unlikely(is_sbi_flag_set(sbi, SBI_CP_DISABLED) && f2fs_is_checkpointed_data(sbi, fio->old_blkaddr))) return true; @@ -2706,11 +2630,8 @@ int f2fs_do_write_data_page(struct f2fs_io_info *fio) f2fs_lookup_read_extent_cache_block(inode, page->index, &fio->old_blkaddr)) { if (!f2fs_is_valid_blkaddr(fio->sbi, fio->old_blkaddr, - DATA_GENERIC_ENHANCE)) { - f2fs_handle_error(fio->sbi, - ERROR_INVALID_BLKADDR); + DATA_GENERIC_ENHANCE)) return -EFSCORRUPTED; - } ipu_force = true; fio->need_lock = LOCK_DONE; @@ -2738,7 +2659,6 @@ got_it: !f2fs_is_valid_blkaddr(fio->sbi, fio->old_blkaddr, DATA_GENERIC_ENHANCE)) { err = -EFSCORRUPTED; - f2fs_handle_error(fio->sbi, ERROR_INVALID_BLKADDR); goto out_writepage; } @@ -2838,7 +2758,7 @@ int f2fs_write_single_data_page(struct page *page, int *submitted, .encrypted_page = NULL, .submitted = 0, .compr_blocks = compr_blocks, - .need_lock = LOCK_RETRY, + .need_lock = compr_blocks ? LOCK_DONE : LOCK_RETRY, .post_read = f2fs_post_read_required(inode) ? 1 : 0, .io_type = io_type, .io_wbc = wbc, @@ -2919,6 +2839,7 @@ write: if (err == -EAGAIN) { err = f2fs_do_write_data_page(&fio); if (err == -EAGAIN) { + f2fs_bug_on(sbi, compr_blocks); fio.need_lock = LOCK_REQ; err = f2fs_do_write_data_page(&fio); } @@ -3704,7 +3625,6 @@ repeat: if (!f2fs_is_valid_blkaddr(sbi, blkaddr, DATA_GENERIC_ENHANCE_READ)) { err = -EFSCORRUPTED; - f2fs_handle_error(sbi, ERROR_INVALID_BLKADDR); goto fail; } err = f2fs_submit_page_read(use_cow ? @@ -3905,26 +3825,36 @@ static int f2fs_migrate_blocks(struct inode *inode, block_t start_blk, struct f2fs_sb_info *sbi = F2FS_I_SB(inode); unsigned int blkofs; unsigned int blk_per_sec = BLKS_PER_SEC(sbi); + unsigned int end_blk = start_blk + blkcnt - 1; unsigned int secidx = start_blk / blk_per_sec; - unsigned int end_sec = secidx + blkcnt / blk_per_sec; + unsigned int end_sec; int ret = 0; + if (!blkcnt) + return 0; + end_sec = end_blk / blk_per_sec; + f2fs_down_write(&F2FS_I(inode)->i_gc_rwsem[WRITE]); filemap_invalidate_lock(inode->i_mapping); set_inode_flag(inode, FI_ALIGNED_WRITE); set_inode_flag(inode, FI_OPU_WRITE); - for (; secidx < end_sec; secidx++) { + for (; secidx <= end_sec; secidx++) { + unsigned int blkofs_end = secidx == end_sec ? + end_blk % blk_per_sec : blk_per_sec - 1; + f2fs_down_write(&sbi->pin_sem); - f2fs_lock_op(sbi); - f2fs_allocate_new_section(sbi, CURSEG_COLD_DATA_PINNED, false); - f2fs_unlock_op(sbi); + ret = f2fs_allocate_pinning_section(sbi); + if (ret) { + f2fs_up_write(&sbi->pin_sem); + break; + } set_inode_flag(inode, FI_SKIP_WRITES); - for (blkofs = 0; blkofs < blk_per_sec; blkofs++) { + for (blkofs = 0; blkofs <= blkofs_end; blkofs++) { struct page *page; unsigned int blkidx = secidx * blk_per_sec + blkofs; @@ -4013,27 +3943,34 @@ retry: nr_pblocks = map.m_len; if ((pblock - SM_I(sbi)->main_blkaddr) & sec_blks_mask || - nr_pblocks & sec_blks_mask) { + nr_pblocks & sec_blks_mask || + !f2fs_valid_pinned_area(sbi, pblock)) { + bool last_extent = false; + not_aligned++; nr_pblocks = roundup(nr_pblocks, blks_per_sec); if (cur_lblock + nr_pblocks > sis->max) nr_pblocks -= blks_per_sec; + /* this extent is last one */ if (!nr_pblocks) { - /* this extent is last one */ - nr_pblocks = map.m_len; - f2fs_warn(sbi, "Swapfile: last extent is not aligned to section"); - goto next; + nr_pblocks = last_lblock - cur_lblock; + last_extent = true; } ret = f2fs_migrate_blocks(inode, cur_lblock, nr_pblocks); - if (ret) + if (ret) { + if (ret == -ENOENT) + ret = -EINVAL; goto out; - goto retry; + } + + if (!last_extent) + goto retry; } -next: + if (cur_lblock + nr_pblocks >= sis->max) nr_pblocks = sis->max - cur_lblock; @@ -4071,17 +4008,17 @@ static int f2fs_swap_activate(struct swap_info_struct *sis, struct file *file, sector_t *span) { struct inode *inode = file_inode(file); + struct f2fs_sb_info *sbi = F2FS_I_SB(inode); int ret; if (!S_ISREG(inode->i_mode)) return -EINVAL; - if (f2fs_readonly(F2FS_I_SB(inode)->sb)) + if (f2fs_readonly(sbi->sb)) return -EROFS; - if (f2fs_lfs_mode(F2FS_I_SB(inode))) { - f2fs_err(F2FS_I_SB(inode), - "Swapfile not supported in LFS mode"); + if (f2fs_lfs_mode(sbi) && !f2fs_sb_has_blkzoned(sbi)) { + f2fs_err(sbi, "Swapfile not supported in LFS mode"); return -EINVAL; } @@ -4092,6 +4029,10 @@ static int f2fs_swap_activate(struct swap_info_struct *sis, struct file *file, if (!f2fs_disable_compressed_file(inode)) return -EINVAL; + ret = filemap_fdatawrite(inode->i_mapping); + if (ret < 0) + return ret; + f2fs_precache_extents(inode); ret = check_swap_activate(sis, file, span); @@ -4100,7 +4041,7 @@ static int f2fs_swap_activate(struct swap_info_struct *sis, struct file *file, stat_inc_swapfile_inode(inode); set_inode_flag(inode, FI_PIN_FILE); - f2fs_update_time(F2FS_I_SB(inode), REQ_TIME); + f2fs_update_time(sbi, REQ_TIME); return ret; } diff --git a/fs/f2fs/debug.c b/fs/f2fs/debug.c index fdbf994f1271..8b0e1e71b667 100644 --- a/fs/f2fs/debug.c +++ b/fs/f2fs/debug.c @@ -41,7 +41,7 @@ void f2fs_update_sit_info(struct f2fs_sb_info *sbi) total_vblocks = 0; blks_per_sec = CAP_BLKS_PER_SEC(sbi); hblks_per_sec = blks_per_sec / 2; - for (segno = 0; segno < MAIN_SEGS(sbi); segno += sbi->segs_per_sec) { + for (segno = 0; segno < MAIN_SEGS(sbi); segno += SEGS_PER_SEC(sbi)) { vblocks = get_valid_blocks(sbi, segno, true); dist = abs(vblocks - hblks_per_sec); bimodal += dist * dist; @@ -135,7 +135,7 @@ static void update_general_status(struct f2fs_sb_info *sbi) si->cur_ckpt_time = sbi->cprc_info.cur_time; si->peak_ckpt_time = sbi->cprc_info.peak_time; spin_unlock(&sbi->cprc_info.stat_lock); - si->total_count = (int)sbi->user_block_count / sbi->blocks_per_seg; + si->total_count = BLKS_TO_SEGS(sbi, (int)sbi->user_block_count); si->rsvd_segs = reserved_segments(sbi); si->overp_segs = overprovision_segments(sbi); si->valid_count = valid_user_blocks(sbi); @@ -176,11 +176,10 @@ static void update_general_status(struct f2fs_sb_info *sbi) si->alloc_nids = NM_I(sbi)->nid_cnt[PREALLOC_NID]; si->io_skip_bggc = sbi->io_skip_bggc; si->other_skip_bggc = sbi->other_skip_bggc; - si->util_free = (int)(free_user_blocks(sbi) >> sbi->log_blocks_per_seg) + si->util_free = (int)(BLKS_TO_SEGS(sbi, free_user_blocks(sbi))) * 100 / (int)(sbi->user_block_count >> sbi->log_blocks_per_seg) / 2; - si->util_valid = (int)(written_block_count(sbi) >> - sbi->log_blocks_per_seg) + si->util_valid = (int)(BLKS_TO_SEGS(sbi, written_block_count(sbi))) * 100 / (int)(sbi->user_block_count >> sbi->log_blocks_per_seg) / 2; si->util_invalid = 50 - si->util_free - si->util_valid; @@ -208,7 +207,7 @@ static void update_general_status(struct f2fs_sb_info *sbi) if (!blks) continue; - if (blks == sbi->blocks_per_seg) + if (blks == BLKS_PER_SEG(sbi)) si->full_seg[type]++; else si->dirty_seg[type]++; diff --git a/fs/f2fs/dir.c b/fs/f2fs/dir.c index 042593aed1ec..02c9355176d3 100644 --- a/fs/f2fs/dir.c +++ b/fs/f2fs/dir.c @@ -830,13 +830,14 @@ int f2fs_do_add_link(struct inode *dir, const struct qstr *name, return err; } -int f2fs_do_tmpfile(struct inode *inode, struct inode *dir) +int f2fs_do_tmpfile(struct inode *inode, struct inode *dir, + struct f2fs_filename *fname) { struct page *page; int err = 0; f2fs_down_write(&F2FS_I(inode)->i_sem); - page = f2fs_init_inode_metadata(inode, dir, NULL, NULL); + page = f2fs_init_inode_metadata(inode, dir, fname, NULL); if (IS_ERR(page)) { err = PTR_ERR(page); goto fail; @@ -995,9 +996,8 @@ int f2fs_fill_dentries(struct dir_context *ctx, struct f2fs_dentry_ptr *d, de = &d->dentry[bit_pos]; if (de->name_len == 0) { if (found_valid_dirent || !bit_pos) { - printk_ratelimited( - "%sF2FS-fs (%s): invalid namelen(0), ino:%u, run fsck to fix.", - KERN_WARNING, sbi->sb->s_id, + f2fs_warn_ratelimited(sbi, + "invalid namelen(0), ino:%u, run fsck to fix.", le32_to_cpu(de->ino)); set_sbi_flag(sbi, SBI_NEED_FSCK); } diff --git a/fs/f2fs/extent_cache.c b/fs/f2fs/extent_cache.c index ad8dfac73bd4..48048fa36427 100644 --- a/fs/f2fs/extent_cache.c +++ b/fs/f2fs/extent_cache.c @@ -43,7 +43,6 @@ bool sanity_check_extent_cache(struct inode *inode) if (!f2fs_is_valid_blkaddr(sbi, ei->blk, DATA_GENERIC_ENHANCE) || !f2fs_is_valid_blkaddr(sbi, ei->blk + ei->len - 1, DATA_GENERIC_ENHANCE)) { - set_sbi_flag(sbi, SBI_NEED_FSCK); f2fs_warn(sbi, "%s: inode (ino=%lx) extent info [%u, %u, %u] is incorrect, run fsck to fix", __func__, inode->i_ino, ei->blk, ei->fofs, ei->len); @@ -856,10 +855,8 @@ static int __get_new_block_age(struct inode *inode, struct extent_info *ei, goto out; if (__is_valid_data_blkaddr(blkaddr) && - !f2fs_is_valid_blkaddr(sbi, blkaddr, DATA_GENERIC_ENHANCE)) { - f2fs_bug_on(sbi, 1); + !f2fs_is_valid_blkaddr(sbi, blkaddr, DATA_GENERIC_ENHANCE)) return -EINVAL; - } out: /* * init block age with zero, this can happen when the block age extent diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h index 4c77e8ce5c75..fced2b7652f4 100644 --- a/fs/f2fs/f2fs.h +++ b/fs/f2fs/f2fs.h @@ -61,7 +61,9 @@ enum { FAULT_SLAB_ALLOC, FAULT_DQUOT_INIT, FAULT_LOCK_OP, - FAULT_BLKADDR, + FAULT_BLKADDR_VALIDITY, + FAULT_BLKADDR_CONSISTENCE, + FAULT_NO_SEGMENT, FAULT_MAX, }; @@ -76,6 +78,11 @@ struct f2fs_fault_info { extern const char *f2fs_fault_name[FAULT_MAX]; #define IS_FAULT_SET(fi, type) ((fi)->inject_type & BIT(type)) + +/* maximum retry count for injected failure */ +#define DEFAULT_FAILURE_RETRY_COUNT 8 +#else +#define DEFAULT_FAILURE_RETRY_COUNT 1 #endif /* @@ -143,7 +150,6 @@ struct f2fs_rwsem { struct f2fs_mount_info { unsigned int opt; - int write_io_size_bits; /* Write IO size bits */ block_t root_reserved_blocks; /* root reserved blocks */ kuid_t s_resuid; /* reserved blocks for uid */ kgid_t s_resgid; /* reserved blocks for gid */ @@ -830,7 +836,7 @@ struct f2fs_inode_info { spinlock_t i_size_lock; /* protect last_disk_size */ #ifdef CONFIG_QUOTA - struct dquot *i_dquot[MAXQUOTAS]; + struct dquot __rcu *i_dquot[MAXQUOTAS]; /* quota space reservation, managed internally by quota code */ qsize_t i_reserved_quota; @@ -1081,7 +1087,8 @@ struct f2fs_sm_info { * f2fs monitors the number of several block types such as on-writeback, * dirty dentry blocks, dirty node blocks, and dirty meta blocks. */ -#define WB_DATA_TYPE(p) (__is_cp_guaranteed(p) ? F2FS_WB_CP_DATA : F2FS_WB_DATA) +#define WB_DATA_TYPE(p, f) \ + (f || f2fs_is_cp_guaranteed(p) ? F2FS_WB_CP_DATA : F2FS_WB_DATA) enum count_type { F2FS_DIRTY_DENTS, F2FS_DIRTY_DATA, @@ -1111,6 +1118,7 @@ enum count_type { * ... Only can be used with META. */ #define PAGE_TYPE_OF_BIO(type) ((type) > META ? META : (type)) +#define PAGE_TYPE_ON_MAIN(type) ((type) == DATA || (type) == NODE) enum page_type { DATA = 0, NODE = 1, /* should not change this */ @@ -1205,7 +1213,6 @@ struct f2fs_io_info { unsigned int submitted:1; /* indicate IO submission */ unsigned int in_list:1; /* indicate fio is in io_list */ unsigned int is_por:1; /* indicate IO is from recovery or not */ - unsigned int retry:1; /* need to reallocate block address */ unsigned int encrypted:1; /* indicate file is encrypted */ unsigned int post_read:1; /* require post read */ enum iostat_type io_type; /* io type */ @@ -1407,18 +1414,16 @@ static inline void f2fs_clear_bit(unsigned int nr, char *addr); * Layout A: lowest bit should be 1 * | bit0 = 1 | bit1 | bit2 | ... | bit MAX | private data .... | * bit 0 PAGE_PRIVATE_NOT_POINTER - * bit 1 PAGE_PRIVATE_DUMMY_WRITE - * bit 2 PAGE_PRIVATE_ONGOING_MIGRATION - * bit 3 PAGE_PRIVATE_INLINE_INODE - * bit 4 PAGE_PRIVATE_REF_RESOURCE - * bit 5- f2fs private data + * bit 1 PAGE_PRIVATE_ONGOING_MIGRATION + * bit 2 PAGE_PRIVATE_INLINE_INODE + * bit 3 PAGE_PRIVATE_REF_RESOURCE + * bit 4- f2fs private data * * Layout B: lowest bit should be 0 * page.private is a wrapped pointer. */ enum { PAGE_PRIVATE_NOT_POINTER, /* private contains non-pointer data */ - PAGE_PRIVATE_DUMMY_WRITE, /* data page for padding aligned IO */ PAGE_PRIVATE_ONGOING_MIGRATION, /* data page which is on-going migrating */ PAGE_PRIVATE_INLINE_INODE, /* inode page contains inline data */ PAGE_PRIVATE_REF_RESOURCE, /* dirty page has referenced resources */ @@ -1565,7 +1570,6 @@ struct f2fs_sb_info { struct f2fs_bio_info *write_io[NR_PAGE_TYPE]; /* for write bios */ /* keep migration IO order for LFS mode */ struct f2fs_rwsem io_order_lock; - mempool_t *write_io_dummy; /* Dummy pages */ pgoff_t page_eio_ofs[NR_PAGE_TYPE]; /* EIO page offset */ int page_eio_cnt[NR_PAGE_TYPE]; /* EIO count */ @@ -1811,6 +1815,37 @@ struct f2fs_sb_info { #endif }; +/* Definitions to access f2fs_sb_info */ +#define SEGS_TO_BLKS(sbi, segs) \ + ((segs) << (sbi)->log_blocks_per_seg) +#define BLKS_TO_SEGS(sbi, blks) \ + ((blks) >> (sbi)->log_blocks_per_seg) + +#define BLKS_PER_SEG(sbi) ((sbi)->blocks_per_seg) +#define BLKS_PER_SEC(sbi) (SEGS_TO_BLKS(sbi, (sbi)->segs_per_sec)) +#define SEGS_PER_SEC(sbi) ((sbi)->segs_per_sec) + +__printf(3, 4) +void f2fs_printk(struct f2fs_sb_info *sbi, bool limit_rate, const char *fmt, ...); + +#define f2fs_err(sbi, fmt, ...) \ + f2fs_printk(sbi, false, KERN_ERR fmt, ##__VA_ARGS__) +#define f2fs_warn(sbi, fmt, ...) \ + f2fs_printk(sbi, false, KERN_WARNING fmt, ##__VA_ARGS__) +#define f2fs_notice(sbi, fmt, ...) \ + f2fs_printk(sbi, false, KERN_NOTICE fmt, ##__VA_ARGS__) +#define f2fs_info(sbi, fmt, ...) \ + f2fs_printk(sbi, false, KERN_INFO fmt, ##__VA_ARGS__) +#define f2fs_debug(sbi, fmt, ...) \ + f2fs_printk(sbi, false, KERN_DEBUG fmt, ##__VA_ARGS__) + +#define f2fs_err_ratelimited(sbi, fmt, ...) \ + f2fs_printk(sbi, true, KERN_ERR fmt, ##__VA_ARGS__) +#define f2fs_warn_ratelimited(sbi, fmt, ...) \ + f2fs_printk(sbi, true, KERN_WARNING fmt, ##__VA_ARGS__) +#define f2fs_info_ratelimited(sbi, fmt, ...) \ + f2fs_printk(sbi, true, KERN_INFO fmt, ##__VA_ARGS__) + #ifdef CONFIG_F2FS_FAULT_INJECTION #define time_to_inject(sbi, type) __time_to_inject(sbi, type, __func__, \ __builtin_return_address(0)) @@ -1828,9 +1863,8 @@ static inline bool __time_to_inject(struct f2fs_sb_info *sbi, int type, atomic_inc(&ffi->inject_ops); if (atomic_read(&ffi->inject_ops) >= ffi->inject_rate) { atomic_set(&ffi->inject_ops, 0); - printk_ratelimited("%sF2FS-fs (%s) : inject %s in %s of %pS\n", - KERN_INFO, sbi->sb->s_id, f2fs_fault_name[type], - func, parent_func); + f2fs_info_ratelimited(sbi, "inject %s in %s of %pS", + f2fs_fault_name[type], func, parent_func); return true; } return false; @@ -2250,9 +2284,30 @@ static inline bool __allow_reserved_blocks(struct f2fs_sb_info *sbi, return false; } +static inline unsigned int get_available_block_count(struct f2fs_sb_info *sbi, + struct inode *inode, bool cap) +{ + block_t avail_user_block_count; + + avail_user_block_count = sbi->user_block_count - + sbi->current_reserved_blocks; + + if (!__allow_reserved_blocks(sbi, inode, cap)) + avail_user_block_count -= F2FS_OPTION(sbi).root_reserved_blocks; + + if (unlikely(is_sbi_flag_set(sbi, SBI_CP_DISABLED))) { + if (avail_user_block_count > sbi->unusable_block_count) + avail_user_block_count -= sbi->unusable_block_count; + else + avail_user_block_count = 0; + } + + return avail_user_block_count; +} + static inline void f2fs_i_blocks_write(struct inode *, block_t, bool, bool); static inline int inc_valid_block_count(struct f2fs_sb_info *sbi, - struct inode *inode, blkcnt_t *count) + struct inode *inode, blkcnt_t *count, bool partial) { blkcnt_t diff = 0, release = 0; block_t avail_user_block_count; @@ -2275,23 +2330,14 @@ static inline int inc_valid_block_count(struct f2fs_sb_info *sbi, spin_lock(&sbi->stat_lock); sbi->total_valid_block_count += (block_t)(*count); - avail_user_block_count = sbi->user_block_count - - sbi->current_reserved_blocks; - - if (!__allow_reserved_blocks(sbi, inode, true)) - avail_user_block_count -= F2FS_OPTION(sbi).root_reserved_blocks; - - if (F2FS_IO_ALIGNED(sbi)) - avail_user_block_count -= sbi->blocks_per_seg * - SM_I(sbi)->additional_reserved_segments; + avail_user_block_count = get_available_block_count(sbi, inode, true); - if (unlikely(is_sbi_flag_set(sbi, SBI_CP_DISABLED))) { - if (avail_user_block_count > sbi->unusable_block_count) - avail_user_block_count -= sbi->unusable_block_count; - else - avail_user_block_count = 0; - } if (unlikely(sbi->total_valid_block_count > avail_user_block_count)) { + if (!partial) { + spin_unlock(&sbi->stat_lock); + goto enospc; + } + diff = sbi->total_valid_block_count - avail_user_block_count; if (diff > *count) diff = *count; @@ -2319,20 +2365,6 @@ release_quota: return -ENOSPC; } -__printf(2, 3) -void f2fs_printk(struct f2fs_sb_info *sbi, const char *fmt, ...); - -#define f2fs_err(sbi, fmt, ...) \ - f2fs_printk(sbi, KERN_ERR fmt, ##__VA_ARGS__) -#define f2fs_warn(sbi, fmt, ...) \ - f2fs_printk(sbi, KERN_WARNING fmt, ##__VA_ARGS__) -#define f2fs_notice(sbi, fmt, ...) \ - f2fs_printk(sbi, KERN_NOTICE fmt, ##__VA_ARGS__) -#define f2fs_info(sbi, fmt, ...) \ - f2fs_printk(sbi, KERN_INFO fmt, ##__VA_ARGS__) -#define f2fs_debug(sbi, fmt, ...) \ - f2fs_printk(sbi, KERN_DEBUG fmt, ##__VA_ARGS__) - #define PAGE_PRIVATE_GET_FUNC(name, flagname) \ static inline bool page_private_##name(struct page *page) \ { \ @@ -2361,17 +2393,14 @@ static inline void clear_page_private_##name(struct page *page) \ PAGE_PRIVATE_GET_FUNC(nonpointer, NOT_POINTER); PAGE_PRIVATE_GET_FUNC(inline, INLINE_INODE); PAGE_PRIVATE_GET_FUNC(gcing, ONGOING_MIGRATION); -PAGE_PRIVATE_GET_FUNC(dummy, DUMMY_WRITE); PAGE_PRIVATE_SET_FUNC(reference, REF_RESOURCE); PAGE_PRIVATE_SET_FUNC(inline, INLINE_INODE); PAGE_PRIVATE_SET_FUNC(gcing, ONGOING_MIGRATION); -PAGE_PRIVATE_SET_FUNC(dummy, DUMMY_WRITE); PAGE_PRIVATE_CLEAR_FUNC(reference, REF_RESOURCE); PAGE_PRIVATE_CLEAR_FUNC(inline, INLINE_INODE); PAGE_PRIVATE_CLEAR_FUNC(gcing, ONGOING_MIGRATION); -PAGE_PRIVATE_CLEAR_FUNC(dummy, DUMMY_WRITE); static inline unsigned long get_page_private_data(struct page *page) { @@ -2505,11 +2534,8 @@ static inline int get_dirty_pages(struct inode *inode) static inline int get_blocktype_secs(struct f2fs_sb_info *sbi, int block_type) { - unsigned int pages_per_sec = sbi->segs_per_sec * sbi->blocks_per_seg; - unsigned int segs = (get_pages(sbi, block_type) + pages_per_sec - 1) >> - sbi->log_blocks_per_seg; - - return segs / sbi->segs_per_sec; + return div_u64(get_pages(sbi, block_type) + BLKS_PER_SEC(sbi) - 1, + BLKS_PER_SEC(sbi)); } static inline block_t valid_user_blocks(struct f2fs_sb_info *sbi) @@ -2573,7 +2599,7 @@ static inline block_t __start_cp_addr(struct f2fs_sb_info *sbi) block_t start_addr = le32_to_cpu(F2FS_RAW_SUPER(sbi)->cp_blkaddr); if (sbi->cur_cp_pack == 2) - start_addr += sbi->blocks_per_seg; + start_addr += BLKS_PER_SEG(sbi); return start_addr; } @@ -2582,7 +2608,7 @@ static inline block_t __start_cp_next_addr(struct f2fs_sb_info *sbi) block_t start_addr = le32_to_cpu(F2FS_RAW_SUPER(sbi)->cp_blkaddr); if (sbi->cur_cp_pack == 1) - start_addr += sbi->blocks_per_seg; + start_addr += BLKS_PER_SEG(sbi); return start_addr; } @@ -2601,7 +2627,8 @@ static inline int inc_valid_node_count(struct f2fs_sb_info *sbi, struct inode *inode, bool is_inode) { block_t valid_block_count; - unsigned int valid_node_count, user_block_count; + unsigned int valid_node_count; + unsigned int avail_user_block_count; int err; if (is_inode) { @@ -2621,21 +2648,10 @@ static inline int inc_valid_node_count(struct f2fs_sb_info *sbi, spin_lock(&sbi->stat_lock); - valid_block_count = sbi->total_valid_block_count + - sbi->current_reserved_blocks + 1; - - if (!__allow_reserved_blocks(sbi, inode, false)) - valid_block_count += F2FS_OPTION(sbi).root_reserved_blocks; + valid_block_count = sbi->total_valid_block_count + 1; + avail_user_block_count = get_available_block_count(sbi, inode, false); - if (F2FS_IO_ALIGNED(sbi)) - valid_block_count += sbi->blocks_per_seg * - SM_I(sbi)->additional_reserved_segments; - - user_block_count = sbi->user_block_count; - if (unlikely(is_sbi_flag_set(sbi, SBI_CP_DISABLED))) - user_block_count -= sbi->unusable_block_count; - - if (unlikely(valid_block_count > user_block_count)) { + if (unlikely(valid_block_count > avail_user_block_count)) { spin_unlock(&sbi->stat_lock); goto enospc; } @@ -3022,6 +3038,7 @@ static inline void __mark_inode_dirty_flag(struct inode *inode, case FI_INLINE_DOTS: case FI_PIN_FILE: case FI_COMPRESS_RELEASED: + case FI_ATOMIC_COMMITTED: f2fs_mark_inode_dirty_sync(inode, true); } } @@ -3445,7 +3462,7 @@ static inline __le32 *get_dnode_addr(struct inode *inode, sizeof((f2fs_inode)->field)) \ <= (F2FS_OLD_ATTRIBUTE_SIZE + (extra_isize))) \ -#define __is_large_section(sbi) ((sbi)->segs_per_sec > 1) +#define __is_large_section(sbi) (SEGS_PER_SEC(sbi) > 1) #define __is_meta_io(fio) (PAGE_TYPE_OF_BIO((fio)->type) == META) @@ -3454,11 +3471,9 @@ bool f2fs_is_valid_blkaddr(struct f2fs_sb_info *sbi, static inline void verify_blkaddr(struct f2fs_sb_info *sbi, block_t blkaddr, int type) { - if (!f2fs_is_valid_blkaddr(sbi, blkaddr, type)) { + if (!f2fs_is_valid_blkaddr(sbi, blkaddr, type)) f2fs_err(sbi, "invalid blkaddr: %u, type: %d, run fsck to fix.", blkaddr, type); - f2fs_bug_on(sbi, 1); - } } static inline bool __is_valid_data_blkaddr(block_t blkaddr) @@ -3560,7 +3575,8 @@ int f2fs_do_add_link(struct inode *dir, const struct qstr *name, struct inode *inode, nid_t ino, umode_t mode); void f2fs_delete_entry(struct f2fs_dir_entry *dentry, struct page *page, struct inode *dir, struct inode *inode); -int f2fs_do_tmpfile(struct inode *inode, struct inode *dir); +int f2fs_do_tmpfile(struct inode *inode, struct inode *dir, + struct f2fs_filename *fname); bool f2fs_empty_dir(struct inode *dir); static inline int f2fs_add_link(struct dentry *dentry, struct inode *inode) @@ -3675,15 +3691,14 @@ int f2fs_disable_cp_again(struct f2fs_sb_info *sbi, block_t unusable); void f2fs_release_discard_addrs(struct f2fs_sb_info *sbi); int f2fs_npages_for_summary_flush(struct f2fs_sb_info *sbi, bool for_ra); bool f2fs_segment_has_free_slot(struct f2fs_sb_info *sbi, int segno); -void f2fs_init_inmem_curseg(struct f2fs_sb_info *sbi); +int f2fs_init_inmem_curseg(struct f2fs_sb_info *sbi); void f2fs_save_inmem_curseg(struct f2fs_sb_info *sbi); void f2fs_restore_inmem_curseg(struct f2fs_sb_info *sbi); -void f2fs_get_new_segment(struct f2fs_sb_info *sbi, - unsigned int *newseg, bool new_sec, int dir); -void f2fs_allocate_segment_for_resize(struct f2fs_sb_info *sbi, int type, +int f2fs_allocate_segment_for_resize(struct f2fs_sb_info *sbi, int type, unsigned int start, unsigned int end); -void f2fs_allocate_new_section(struct f2fs_sb_info *sbi, int type, bool force); -void f2fs_allocate_new_segments(struct f2fs_sb_info *sbi); +int f2fs_allocate_new_section(struct f2fs_sb_info *sbi, int type, bool force); +int f2fs_allocate_pinning_section(struct f2fs_sb_info *sbi); +int f2fs_allocate_new_segments(struct f2fs_sb_info *sbi); int f2fs_trim_fs(struct f2fs_sb_info *sbi, struct fstrim_range *range); bool f2fs_exist_trim_candidates(struct f2fs_sb_info *sbi, struct cp_control *cpc); @@ -3704,7 +3719,7 @@ void f2fs_replace_block(struct f2fs_sb_info *sbi, struct dnode_of_data *dn, block_t old_addr, block_t new_addr, unsigned char version, bool recover_curseg, bool recover_newaddr); -void f2fs_allocate_data_block(struct f2fs_sb_info *sbi, struct page *page, +int f2fs_allocate_data_block(struct f2fs_sb_info *sbi, struct page *page, block_t old_blkaddr, block_t *new_blkaddr, struct f2fs_summary *sum, int type, struct f2fs_io_info *fio); @@ -3754,6 +3769,8 @@ struct page *f2fs_get_meta_page_retry(struct f2fs_sb_info *sbi, pgoff_t index); struct page *f2fs_get_tmp_page(struct f2fs_sb_info *sbi, pgoff_t index); bool f2fs_is_valid_blkaddr(struct f2fs_sb_info *sbi, block_t blkaddr, int type); +bool f2fs_is_valid_blkaddr_raw(struct f2fs_sb_info *sbi, + block_t blkaddr, int type); int f2fs_ra_meta_pages(struct f2fs_sb_info *sbi, block_t start, int nrpages, int type, bool sync); void f2fs_ra_meta_pages_cond(struct f2fs_sb_info *sbi, pgoff_t index, @@ -3794,6 +3811,7 @@ void f2fs_init_ckpt_req_control(struct f2fs_sb_info *sbi); */ int __init f2fs_init_bioset(void); void f2fs_destroy_bioset(void); +bool f2fs_is_cp_guaranteed(struct page *page); int f2fs_init_bio_entry_cache(void); void f2fs_destroy_bio_entry_cache(void); void f2fs_submit_read_bio(struct f2fs_sb_info *sbi, struct bio *bio, @@ -3857,6 +3875,9 @@ void f2fs_stop_gc_thread(struct f2fs_sb_info *sbi); block_t f2fs_start_bidx_of_node(unsigned int node_ofs, struct inode *inode); int f2fs_gc(struct f2fs_sb_info *sbi, struct f2fs_gc_control *gc_control); void f2fs_build_gc_manager(struct f2fs_sb_info *sbi); +int f2fs_gc_range(struct f2fs_sb_info *sbi, + unsigned int start_seg, unsigned int end_seg, + bool dry_run, unsigned int dry_run_sections); int f2fs_resize_fs(struct file *filp, __u64 block_count); int __init f2fs_create_garbage_collection_cache(void); void f2fs_destroy_garbage_collection_cache(void); @@ -4277,7 +4298,8 @@ struct decompress_io_ctx *f2fs_alloc_dic(struct compress_ctx *cc); void f2fs_decompress_end_io(struct decompress_io_ctx *dic, bool failed, bool in_task); void f2fs_put_page_dic(struct page *page, bool in_task); -unsigned int f2fs_cluster_blocks_are_contiguous(struct dnode_of_data *dn); +unsigned int f2fs_cluster_blocks_are_contiguous(struct dnode_of_data *dn, + unsigned int ofs_in_node); int f2fs_init_compress_ctx(struct compress_ctx *cc); void f2fs_destroy_compress_ctx(struct compress_ctx *cc, bool reuse); void f2fs_init_compress_info(struct f2fs_sb_info *sbi); @@ -4334,7 +4356,8 @@ static inline void f2fs_put_page_dic(struct page *page, bool in_task) { WARN_ON_ONCE(1); } -static inline unsigned int f2fs_cluster_blocks_are_contiguous(struct dnode_of_data *dn) { return 0; } +static inline unsigned int f2fs_cluster_blocks_are_contiguous( + struct dnode_of_data *dn, unsigned int ofs_in_node) { return 0; } static inline bool f2fs_sanity_check_cluster(struct dnode_of_data *dn) { return false; } static inline int f2fs_init_compress_inode(struct f2fs_sb_info *sbi) { return 0; } static inline void f2fs_destroy_compress_inode(struct f2fs_sb_info *sbi) { } @@ -4391,15 +4414,24 @@ static inline bool f2fs_disable_compressed_file(struct inode *inode) { struct f2fs_inode_info *fi = F2FS_I(inode); - if (!f2fs_compressed_file(inode)) + f2fs_down_write(&F2FS_I(inode)->i_sem); + + if (!f2fs_compressed_file(inode)) { + f2fs_up_write(&F2FS_I(inode)->i_sem); return true; - if (S_ISREG(inode->i_mode) && F2FS_HAS_BLOCKS(inode)) + } + if (f2fs_is_mmap_file(inode) || + (S_ISREG(inode->i_mode) && F2FS_HAS_BLOCKS(inode))) { + f2fs_up_write(&F2FS_I(inode)->i_sem); return false; + } fi->i_flags &= ~F2FS_COMPR_FL; stat_dec_compr_inode(inode); clear_inode_flag(inode, FI_COMPRESSED_FILE); f2fs_mark_inode_dirty_sync(inode, true); + + f2fs_up_write(&F2FS_I(inode)->i_sem); return true; } @@ -4502,6 +4534,17 @@ static inline bool f2fs_lfs_mode(struct f2fs_sb_info *sbi) return F2FS_OPTION(sbi).fs_mode == FS_MODE_LFS; } +static inline bool f2fs_valid_pinned_area(struct f2fs_sb_info *sbi, + block_t blkaddr) +{ + if (f2fs_sb_has_blkzoned(sbi)) { + int devi = f2fs_target_device_index(sbi, blkaddr); + + return !bdev_is_zoned(FDEV(devi).bdev); + } + return true; +} + static inline bool f2fs_low_mem_mode(struct f2fs_sb_info *sbi) { return F2FS_OPTION(sbi).memory_mode == MEMORY_MODE_LOW; @@ -4603,10 +4646,36 @@ static inline bool f2fs_is_readonly(struct f2fs_sb_info *sbi) return f2fs_sb_has_readonly(sbi) || f2fs_readonly(sbi->sb); } +static inline void f2fs_truncate_meta_inode_pages(struct f2fs_sb_info *sbi, + block_t blkaddr, unsigned int cnt) +{ + bool need_submit = false; + int i = 0; + + do { + struct page *page; + + page = find_get_page(META_MAPPING(sbi), blkaddr + i); + if (page) { + if (PageWriteback(page)) + need_submit = true; + f2fs_put_page(page, 0); + } + } while (++i < cnt && !need_submit); + + if (need_submit) + f2fs_submit_merged_write_cond(sbi, sbi->meta_inode, + NULL, 0, DATA); + + truncate_inode_pages_range(META_MAPPING(sbi), + F2FS_BLK_TO_BYTES((loff_t)blkaddr), + F2FS_BLK_END_BYTES((loff_t)(blkaddr + cnt - 1))); +} + static inline void f2fs_invalidate_internal_cache(struct f2fs_sb_info *sbi, block_t blkaddr) { - invalidate_mapping_pages(META_MAPPING(sbi), blkaddr, blkaddr); + f2fs_truncate_meta_inode_pages(sbi, blkaddr, 1); f2fs_invalidate_compress_page(sbi, blkaddr); } diff --git a/fs/f2fs/file.c b/fs/f2fs/file.c index b58ab1157b7e..1761ad125f97 100644 --- a/fs/f2fs/file.c +++ b/fs/f2fs/file.c @@ -39,6 +39,7 @@ static vm_fault_t f2fs_filemap_fault(struct vm_fault *vmf) { struct inode *inode = file_inode(vmf->vma->vm_file); + vm_flags_t flags = vmf->vma->vm_flags; vm_fault_t ret; ret = filemap_fault(vmf); @@ -46,7 +47,7 @@ static vm_fault_t f2fs_filemap_fault(struct vm_fault *vmf) f2fs_update_iostat(F2FS_I_SB(inode), inode, APP_MAPPED_READ_IO, F2FS_BLKSIZE); - trace_f2fs_filemap_fault(inode, vmf->pgoff, vmf->vma->vm_flags, ret); + trace_f2fs_filemap_fault(inode, vmf->pgoff, flags, ret); return ret; } @@ -394,9 +395,20 @@ int f2fs_sync_file(struct file *file, loff_t start, loff_t end, int datasync) return f2fs_do_sync_file(file, start, end, datasync, false); } -static bool __found_offset(struct address_space *mapping, block_t blkaddr, - pgoff_t index, int whence) +static bool __found_offset(struct address_space *mapping, + struct dnode_of_data *dn, pgoff_t index, int whence) { + block_t blkaddr = f2fs_data_blkaddr(dn); + struct inode *inode = mapping->host; + bool compressed_cluster = false; + + if (f2fs_compressed_file(inode)) { + block_t first_blkaddr = data_blkaddr(dn->inode, dn->node_page, + ALIGN_DOWN(dn->ofs_in_node, F2FS_I(inode)->i_cluster_size)); + + compressed_cluster = first_blkaddr == COMPRESS_ADDR; + } + switch (whence) { case SEEK_DATA: if (__is_valid_data_blkaddr(blkaddr)) @@ -404,8 +416,12 @@ static bool __found_offset(struct address_space *mapping, block_t blkaddr, if (blkaddr == NEW_ADDR && xa_get_mark(&mapping->i_pages, index, PAGECACHE_TAG_DIRTY)) return true; + if (compressed_cluster) + return true; break; case SEEK_HOLE: + if (compressed_cluster) + return false; if (blkaddr == NULL_ADDR) return true; break; @@ -474,7 +490,7 @@ static loff_t f2fs_seek_block(struct file *file, loff_t offset, int whence) goto fail; } - if (__found_offset(file->f_mapping, blkaddr, + if (__found_offset(file->f_mapping, &dn, pgofs, whence)) { f2fs_put_dnode(&dn); goto found; @@ -590,8 +606,10 @@ void f2fs_truncate_data_blocks_range(struct dnode_of_data *dn, int count) f2fs_set_data_blkaddr(dn, NULL_ADDR); if (__is_valid_data_blkaddr(blkaddr)) { - if (!f2fs_is_valid_blkaddr(sbi, blkaddr, - DATA_GENERIC_ENHANCE)) + if (time_to_inject(sbi, FAULT_BLKADDR_CONSISTENCE)) + continue; + if (!f2fs_is_valid_blkaddr_raw(sbi, blkaddr, + DATA_GENERIC_ENHANCE)) continue; if (compressed_cluster) valid_blocks++; @@ -818,8 +836,6 @@ static bool f2fs_force_buffered_io(struct inode *inode, int rw) */ if (f2fs_sb_has_blkzoned(sbi) && (rw == WRITE)) return true; - if (f2fs_lfs_mode(sbi) && rw == WRITE && F2FS_IO_ALIGNED(sbi)) - return true; if (is_sbi_flag_set(sbi, SBI_CP_DISABLED)) return true; @@ -1192,7 +1208,6 @@ next_dnode: !f2fs_is_valid_blkaddr(sbi, *blkaddr, DATA_GENERIC_ENHANCE)) { f2fs_put_dnode(&dn); - f2fs_handle_error(sbi, ERROR_INVALID_BLKADDR); return -EFSCORRUPTED; } @@ -1478,7 +1493,6 @@ static int f2fs_do_zero_range(struct dnode_of_data *dn, pgoff_t start, if (!f2fs_is_valid_blkaddr(sbi, dn->data_blkaddr, DATA_GENERIC_ENHANCE)) { ret = -EFSCORRUPTED; - f2fs_handle_error(sbi, ERROR_INVALID_BLKADDR); break; } @@ -1662,10 +1676,12 @@ static int f2fs_insert_range(struct inode *inode, loff_t offset, loff_t len) } filemap_invalidate_unlock(mapping); f2fs_up_write(&F2FS_I(inode)->i_gc_rwsem[WRITE]); + if (ret) + return ret; /* write out all moved pages, if possible */ filemap_invalidate_lock(mapping); - filemap_write_and_wait_range(mapping, offset, LLONG_MAX); + ret = filemap_write_and_wait_range(mapping, offset, LLONG_MAX); truncate_pagecache(inode, offset); filemap_invalidate_unlock(mapping); @@ -1731,9 +1747,11 @@ next_alloc: f2fs_down_write(&sbi->pin_sem); - f2fs_lock_op(sbi); - f2fs_allocate_new_section(sbi, CURSEG_COLD_DATA_PINNED, false); - f2fs_unlock_op(sbi); + err = f2fs_allocate_pinning_section(sbi); + if (err) { + f2fs_up_write(&sbi->pin_sem); + goto out_err; + } map.m_seg_type = CURSEG_COLD_DATA_PINNED; err = f2fs_map_blocks(inode, &map, F2FS_GET_BLOCK_PRE_DIO); @@ -2066,7 +2084,8 @@ static int f2fs_ioc_start_atomic_write(struct file *filp, bool truncate) inode_lock(inode); - if (!f2fs_disable_compressed_file(inode)) { + if (!f2fs_disable_compressed_file(inode) || + f2fs_is_pinned_file(inode)) { ret = -EINVAL; goto out; } @@ -2243,8 +2262,11 @@ static int f2fs_ioc_shutdown(struct file *filp, unsigned long arg) case F2FS_GOING_DOWN_METASYNC: /* do checkpoint only */ ret = f2fs_sync_fs(sb, 1); - if (ret) + if (ret) { + if (ret == -EIO) + ret = 0; goto out; + } f2fs_stop_checkpoint(sbi, false, STOP_CP_REASON_SHUTDOWN); break; case F2FS_GOING_DOWN_NOSYNC: @@ -2260,6 +2282,8 @@ static int f2fs_ioc_shutdown(struct file *filp, unsigned long arg) set_sbi_flag(sbi, SBI_IS_DIRTY); /* do checkpoint only */ ret = f2fs_sync_fs(sb, 1); + if (ret == -EIO) + ret = 0; goto out; default: ret = -EINVAL; @@ -2578,7 +2602,6 @@ static int f2fs_defragment_range(struct f2fs_sb_info *sbi, .m_may_create = false }; struct extent_info ei = {}; pgoff_t pg_start, pg_end, next_pgofs; - unsigned int blk_per_seg = sbi->blocks_per_seg; unsigned int total = 0, sec_num; block_t blk_end = 0; bool fragmented = false; @@ -2687,7 +2710,8 @@ do_map: set_inode_flag(inode, FI_SKIP_WRITES); idx = map.m_lblk; - while (idx < map.m_lblk + map.m_len && cnt < blk_per_seg) { + while (idx < map.m_lblk + map.m_len && + cnt < BLKS_PER_SEG(sbi)) { struct page *page; page = f2fs_get_lock_data_page(inode, idx, true); @@ -2707,7 +2731,7 @@ do_map: map.m_lblk = idx; check: - if (map.m_lblk < pg_end && cnt < blk_per_seg) + if (map.m_lblk < pg_end && cnt < BLKS_PER_SEG(sbi)) goto do_map; clear_inode_flag(inode, FI_SKIP_WRITES); @@ -2976,8 +3000,8 @@ static int f2fs_ioc_flush_device(struct file *filp, unsigned long arg) if (!f2fs_is_multi_device(sbi) || sbi->s_ndevs - 1 <= range.dev_num || __is_large_section(sbi)) { - f2fs_warn(sbi, "Can't flush %u in %d for segs_per_sec %u != 1", - range.dev_num, sbi->s_ndevs, sbi->segs_per_sec); + f2fs_warn(sbi, "Can't flush %u in %d for SEGS_PER_SEC %u != 1", + range.dev_num, sbi->s_ndevs, SEGS_PER_SEC(sbi)); return -EINVAL; } @@ -3183,6 +3207,7 @@ int f2fs_pin_file_control(struct inode *inode, bool inc) static int f2fs_ioc_set_pin_file(struct file *filp, unsigned long arg) { struct inode *inode = file_inode(filp); + struct f2fs_sb_info *sbi = F2FS_I_SB(inode); __u32 pin; int ret = 0; @@ -3192,7 +3217,7 @@ static int f2fs_ioc_set_pin_file(struct file *filp, unsigned long arg) if (!S_ISREG(inode->i_mode)) return -EINVAL; - if (f2fs_readonly(F2FS_I_SB(inode)->sb)) + if (f2fs_readonly(sbi->sb)) return -EROFS; ret = mnt_want_write_file(filp); @@ -3205,9 +3230,18 @@ static int f2fs_ioc_set_pin_file(struct file *filp, unsigned long arg) clear_inode_flag(inode, FI_PIN_FILE); f2fs_i_gc_failures_write(inode, 0); goto done; + } else if (f2fs_is_pinned_file(inode)) { + goto done; } - if (f2fs_should_update_outplace(inode, NULL)) { + if (f2fs_sb_has_blkzoned(sbi) && F2FS_HAS_BLOCKS(inode)) { + ret = -EFBIG; + goto out; + } + + /* Let's allow file pinning on zoned device. */ + if (!f2fs_sb_has_blkzoned(sbi) && + f2fs_should_update_outplace(inode, NULL)) { ret = -EINVAL; goto out; } @@ -3229,7 +3263,7 @@ static int f2fs_ioc_set_pin_file(struct file *filp, unsigned long arg) set_inode_flag(inode, FI_PIN_FILE); ret = F2FS_I(inode)->i_gc_failures[GC_FAILURE_PIN]; done: - f2fs_update_time(F2FS_I_SB(inode), REQ_TIME); + f2fs_update_time(sbi, REQ_TIME); out: inode_unlock(inode); mnt_drop_write_file(filp); @@ -3438,10 +3472,8 @@ static int release_compress_blocks(struct dnode_of_data *dn, pgoff_t count) if (!__is_valid_data_blkaddr(blkaddr)) continue; if (unlikely(!f2fs_is_valid_blkaddr(sbi, blkaddr, - DATA_GENERIC_ENHANCE))) { - f2fs_handle_error(sbi, ERROR_INVALID_BLKADDR); + DATA_GENERIC_ENHANCE))) return -EFSCORRUPTED; - } } while (count) { @@ -3588,10 +3620,10 @@ out: return ret; } -static int reserve_compress_blocks(struct dnode_of_data *dn, pgoff_t count) +static int reserve_compress_blocks(struct dnode_of_data *dn, pgoff_t count, + unsigned int *reserved_blocks) { struct f2fs_sb_info *sbi = F2FS_I_SB(dn->inode); - unsigned int reserved_blocks = 0; int cluster_size = F2FS_I(dn->inode)->i_cluster_size; block_t blkaddr; int i; @@ -3603,10 +3635,8 @@ static int reserve_compress_blocks(struct dnode_of_data *dn, pgoff_t count) if (!__is_valid_data_blkaddr(blkaddr)) continue; if (unlikely(!f2fs_is_valid_blkaddr(sbi, blkaddr, - DATA_GENERIC_ENHANCE))) { - f2fs_handle_error(sbi, ERROR_INVALID_BLKADDR); + DATA_GENERIC_ENHANCE))) return -EFSCORRUPTED; - } } while (count) { @@ -3614,40 +3644,53 @@ static int reserve_compress_blocks(struct dnode_of_data *dn, pgoff_t count) blkcnt_t reserved; int ret; - for (i = 0; i < cluster_size; i++, dn->ofs_in_node++) { - blkaddr = f2fs_data_blkaddr(dn); + for (i = 0; i < cluster_size; i++) { + blkaddr = data_blkaddr(dn->inode, dn->node_page, + dn->ofs_in_node + i); if (i == 0) { - if (blkaddr == COMPRESS_ADDR) - continue; - dn->ofs_in_node += cluster_size; - goto next; + if (blkaddr != COMPRESS_ADDR) { + dn->ofs_in_node += cluster_size; + goto next; + } + continue; } - if (__is_valid_data_blkaddr(blkaddr)) { + /* + * compressed cluster was not released due to it + * fails in release_compress_blocks(), so NEW_ADDR + * is a possible case. + */ + if (blkaddr == NEW_ADDR || + __is_valid_data_blkaddr(blkaddr)) { compr_blocks++; continue; } - - f2fs_set_data_blkaddr(dn, NEW_ADDR); } reserved = cluster_size - compr_blocks; - ret = inc_valid_block_count(sbi, dn->inode, &reserved); - if (ret) + + /* for the case all blocks in cluster were reserved */ + if (reserved == 1) + goto next; + + ret = inc_valid_block_count(sbi, dn->inode, &reserved, false); + if (unlikely(ret)) return ret; - if (reserved != cluster_size - compr_blocks) - return -ENOSPC; + for (i = 0; i < cluster_size; i++, dn->ofs_in_node++) { + if (f2fs_data_blkaddr(dn) == NULL_ADDR) + f2fs_set_data_blkaddr(dn, NEW_ADDR); + } f2fs_i_compr_blocks_update(dn->inode, compr_blocks, true); - reserved_blocks += reserved; + *reserved_blocks += reserved; next: count -= cluster_size; } - return reserved_blocks; + return 0; } static int f2fs_reserve_compress_blocks(struct file *filp, unsigned long arg) @@ -3671,9 +3714,6 @@ static int f2fs_reserve_compress_blocks(struct file *filp, unsigned long arg) if (ret) return ret; - if (atomic_read(&F2FS_I(inode)->i_compr_blocks)) - goto out; - f2fs_balance_fs(sbi, true); inode_lock(inode); @@ -3683,6 +3723,9 @@ static int f2fs_reserve_compress_blocks(struct file *filp, unsigned long arg) goto unlock_inode; } + if (atomic_read(&F2FS_I(inode)->i_compr_blocks)) + goto unlock_inode; + f2fs_down_write(&F2FS_I(inode)->i_gc_rwsem[WRITE]); filemap_invalidate_lock(inode->i_mapping); @@ -3708,7 +3751,7 @@ static int f2fs_reserve_compress_blocks(struct file *filp, unsigned long arg) count = min(end_offset - dn.ofs_in_node, last_idx - page_idx); count = round_up(count, F2FS_I(inode)->i_cluster_size); - ret = reserve_compress_blocks(&dn, count); + ret = reserve_compress_blocks(&dn, count, &reserved_blocks); f2fs_put_dnode(&dn); @@ -3716,23 +3759,21 @@ static int f2fs_reserve_compress_blocks(struct file *filp, unsigned long arg) break; page_idx += count; - reserved_blocks += ret; } filemap_invalidate_unlock(inode->i_mapping); f2fs_up_write(&F2FS_I(inode)->i_gc_rwsem[WRITE]); - if (ret >= 0) { + if (!ret) { clear_inode_flag(inode, FI_COMPRESS_RELEASED); inode_set_ctime_current(inode); f2fs_mark_inode_dirty_sync(inode, true); } unlock_inode: inode_unlock(inode); -out: mnt_drop_write_file(filp); - if (ret >= 0) { + if (!ret) { ret = put_user(reserved_blocks, (u64 __user *)arg); } else if (reserved_blocks && atomic_read(&F2FS_I(inode)->i_compr_blocks)) { @@ -3877,8 +3918,6 @@ static int f2fs_sec_trim_file(struct file *filp, unsigned long arg) DATA_GENERIC_ENHANCE)) { ret = -EFSCORRUPTED; f2fs_put_dnode(&dn); - f2fs_handle_error(sbi, - ERROR_INVALID_BLKADDR); goto out; } @@ -3981,16 +4020,20 @@ static int f2fs_ioc_set_compress_option(struct file *filp, unsigned long arg) sizeof(option))) return -EFAULT; - if (!f2fs_compressed_file(inode) || - option.log_cluster_size < MIN_COMPRESS_LOG_SIZE || - option.log_cluster_size > MAX_COMPRESS_LOG_SIZE || - option.algorithm >= COMPRESS_MAX) + if (option.log_cluster_size < MIN_COMPRESS_LOG_SIZE || + option.log_cluster_size > MAX_COMPRESS_LOG_SIZE || + option.algorithm >= COMPRESS_MAX) return -EINVAL; file_start_write(filp); inode_lock(inode); f2fs_down_write(&F2FS_I(inode)->i_sem); + if (!f2fs_compressed_file(inode)) { + ret = -EINVAL; + goto out; + } + if (f2fs_is_mmap_file(inode) || get_dirty_pages(inode)) { ret = -EBUSY; goto out; @@ -4066,7 +4109,6 @@ static int f2fs_ioc_decompress_file(struct file *filp) struct f2fs_sb_info *sbi = F2FS_I_SB(inode); struct f2fs_inode_info *fi = F2FS_I(inode); pgoff_t page_idx = 0, last_idx; - unsigned int blk_per_seg = sbi->blocks_per_seg; int cluster_size = fi->i_cluster_size; int count, ret; @@ -4110,7 +4152,7 @@ static int f2fs_ioc_decompress_file(struct file *filp) if (ret < 0) break; - if (get_dirty_pages(inode) >= blk_per_seg) { + if (get_dirty_pages(inode) >= BLKS_PER_SEG(sbi)) { ret = filemap_fdatawrite(inode->i_mapping); if (ret < 0) break; @@ -4145,7 +4187,6 @@ static int f2fs_ioc_compress_file(struct file *filp) struct inode *inode = file_inode(filp); struct f2fs_sb_info *sbi = F2FS_I_SB(inode); pgoff_t page_idx = 0, last_idx; - unsigned int blk_per_seg = sbi->blocks_per_seg; int cluster_size = F2FS_I(inode)->i_cluster_size; int count, ret; @@ -4188,7 +4229,7 @@ static int f2fs_ioc_compress_file(struct file *filp) if (ret < 0) break; - if (get_dirty_pages(inode) >= blk_per_seg) { + if (get_dirty_pages(inode) >= BLKS_PER_SEG(sbi)) { ret = filemap_fdatawrite(inode->i_mapping); if (ret < 0) break; diff --git a/fs/f2fs/gc.c b/fs/f2fs/gc.c index a079eebfb080..8852814dab7f 100644 --- a/fs/f2fs/gc.c +++ b/fs/f2fs/gc.c @@ -259,7 +259,7 @@ static void select_policy(struct f2fs_sb_info *sbi, int gc_type, p->ofs_unit = 1; } else { p->gc_mode = select_gc_type(sbi, gc_type); - p->ofs_unit = sbi->segs_per_sec; + p->ofs_unit = SEGS_PER_SEC(sbi); if (__is_large_section(sbi)) { p->dirty_bitmap = dirty_i->dirty_secmap; p->max_search = count_bits(p->dirty_bitmap, @@ -280,11 +280,11 @@ static void select_policy(struct f2fs_sb_info *sbi, int gc_type, p->max_search > sbi->max_victim_search) p->max_search = sbi->max_victim_search; - /* let's select beginning hot/small space first in no_heap mode*/ + /* let's select beginning hot/small space first. */ if (f2fs_need_rand_seg(sbi)) - p->offset = get_random_u32_below(MAIN_SECS(sbi) * sbi->segs_per_sec); - else if (test_opt(sbi, NOHEAP) && - (type == CURSEG_HOT_DATA || IS_NODESEG(type))) + p->offset = get_random_u32_below(MAIN_SECS(sbi) * + SEGS_PER_SEC(sbi)); + else if (type == CURSEG_HOT_DATA || IS_NODESEG(type)) p->offset = 0; else p->offset = SIT_I(sbi)->last_victim[p->gc_mode]; @@ -295,13 +295,13 @@ static unsigned int get_max_cost(struct f2fs_sb_info *sbi, { /* SSR allocates in a segment unit */ if (p->alloc_mode == SSR) - return sbi->blocks_per_seg; + return BLKS_PER_SEG(sbi); else if (p->alloc_mode == AT_SSR) return UINT_MAX; /* LFS */ if (p->gc_mode == GC_GREEDY) - return 2 * sbi->blocks_per_seg * p->ofs_unit; + return SEGS_TO_BLKS(sbi, 2 * p->ofs_unit); else if (p->gc_mode == GC_CB) return UINT_MAX; else if (p->gc_mode == GC_AT) @@ -348,7 +348,7 @@ static unsigned int get_cb_cost(struct f2fs_sb_info *sbi, unsigned int segno) mtime = div_u64(mtime, usable_segs_per_sec); vblocks = div_u64(vblocks, usable_segs_per_sec); - u = (vblocks * 100) >> sbi->log_blocks_per_seg; + u = BLKS_TO_SEGS(sbi, vblocks * 100); /* Handle if the system time has changed by the user */ if (mtime < sit_i->min_mtime) @@ -496,9 +496,9 @@ static void add_victim_entry(struct f2fs_sb_info *sbi, return; } - for (i = 0; i < sbi->segs_per_sec; i++) + for (i = 0; i < SEGS_PER_SEC(sbi); i++) mtime += get_seg_entry(sbi, start + i)->mtime; - mtime = div_u64(mtime, sbi->segs_per_sec); + mtime = div_u64(mtime, SEGS_PER_SEC(sbi)); /* Handle if the system time has changed by the user */ if (mtime < sit_i->min_mtime) @@ -599,7 +599,6 @@ static void atssr_lookup_victim(struct f2fs_sb_info *sbi, unsigned long long age; unsigned long long max_mtime = sit_i->dirty_max_mtime; unsigned long long min_mtime = sit_i->dirty_min_mtime; - unsigned int seg_blocks = sbi->blocks_per_seg; unsigned int vblocks; unsigned int dirty_threshold = max(am->max_candidate_count, am->candidate_ratio * @@ -629,7 +628,7 @@ next_node: f2fs_bug_on(sbi, !vblocks); /* rare case */ - if (vblocks == seg_blocks) + if (vblocks == BLKS_PER_SEG(sbi)) goto skip_node; iter++; @@ -755,7 +754,7 @@ int f2fs_get_victim(struct f2fs_sb_info *sbi, unsigned int *result, int ret = 0; mutex_lock(&dirty_i->seglist_lock); - last_segment = MAIN_SECS(sbi) * sbi->segs_per_sec; + last_segment = MAIN_SECS(sbi) * SEGS_PER_SEC(sbi); p.alloc_mode = alloc_mode; p.age = age; @@ -896,7 +895,7 @@ next: else sm->last_victim[p.gc_mode] = segno + p.ofs_unit; sm->last_victim[p.gc_mode] %= - (MAIN_SECS(sbi) * sbi->segs_per_sec); + (MAIN_SECS(sbi) * SEGS_PER_SEC(sbi)); break; } } @@ -1184,7 +1183,6 @@ static int ra_data_block(struct inode *inode, pgoff_t index) .op_flags = 0, .encrypted_page = NULL, .in_list = 0, - .retry = 0, }; int err; @@ -1197,7 +1195,6 @@ static int ra_data_block(struct inode *inode, pgoff_t index) if (unlikely(!f2fs_is_valid_blkaddr(sbi, dn.data_blkaddr, DATA_GENERIC_ENHANCE_READ))) { err = -EFSCORRUPTED; - f2fs_handle_error(sbi, ERROR_INVALID_BLKADDR); goto put_page; } goto got_it; @@ -1216,7 +1213,6 @@ static int ra_data_block(struct inode *inode, pgoff_t index) if (unlikely(!f2fs_is_valid_blkaddr(sbi, dn.data_blkaddr, DATA_GENERIC_ENHANCE))) { err = -EFSCORRUPTED; - f2fs_handle_error(sbi, ERROR_INVALID_BLKADDR); goto put_page; } got_it: @@ -1273,7 +1269,6 @@ static int move_data_block(struct inode *inode, block_t bidx, .op_flags = 0, .encrypted_page = NULL, .in_list = 0, - .retry = 0, }; struct dnode_of_data dn; struct f2fs_summary sum; @@ -1364,8 +1359,13 @@ static int move_data_block(struct inode *inode, block_t bidx, set_summary(&sum, dn.nid, dn.ofs_in_node, ni.version); /* allocate block address */ - f2fs_allocate_data_block(fio.sbi, NULL, fio.old_blkaddr, &newaddr, + err = f2fs_allocate_data_block(fio.sbi, NULL, fio.old_blkaddr, &newaddr, &sum, type, NULL); + if (err) { + f2fs_put_page(mpage, 1); + /* filesystem should shutdown, no need to recovery block */ + goto up_out; + } fio.encrypted_page = f2fs_pagecache_get_page(META_MAPPING(fio.sbi), newaddr, FGP_LOCK | FGP_CREAT, GFP_NOFS); @@ -1393,18 +1393,12 @@ static int move_data_block(struct inode *inode, block_t bidx, fio.op_flags = REQ_SYNC; fio.new_blkaddr = newaddr; f2fs_submit_page_write(&fio); - if (fio.retry) { - err = -EAGAIN; - if (PageWriteback(fio.encrypted_page)) - end_page_writeback(fio.encrypted_page); - goto put_page_out; - } f2fs_update_iostat(fio.sbi, NULL, FS_GC_DATA_IO, F2FS_BLKSIZE); f2fs_update_data_blkaddr(&dn, newaddr); set_inode_flag(inode, FI_APPEND_WRITE); -put_page_out: + f2fs_put_page(fio.encrypted_page, 1); recover_block: if (err) @@ -1678,7 +1672,7 @@ static int do_garbage_collect(struct f2fs_sb_info *sbi, struct f2fs_summary_block *sum; struct blk_plug plug; unsigned int segno = start_segno; - unsigned int end_segno = start_segno + sbi->segs_per_sec; + unsigned int end_segno = start_segno + SEGS_PER_SEC(sbi); int seg_freed = 0, migrated = 0; unsigned char type = IS_DATASEG(get_seg_entry(sbi, segno)->type) ? SUM_TYPE_DATA : SUM_TYPE_NODE; @@ -1686,7 +1680,7 @@ static int do_garbage_collect(struct f2fs_sb_info *sbi, int submitted = 0; if (__is_large_section(sbi)) - end_segno = rounddown(end_segno, sbi->segs_per_sec); + end_segno = rounddown(end_segno, SEGS_PER_SEC(sbi)); /* * zone-capacity can be less than zone-size in zoned devices, @@ -1694,7 +1688,7 @@ static int do_garbage_collect(struct f2fs_sb_info *sbi, * calculate the end segno in the zone which can be garbage collected */ if (f2fs_sb_has_blkzoned(sbi)) - end_segno -= sbi->segs_per_sec - + end_segno -= SEGS_PER_SEC(sbi) - f2fs_usable_segs_in_sec(sbi, segno); sanity_check_seg_type(sbi, get_seg_entry(sbi, segno)->type); @@ -1983,10 +1977,43 @@ void f2fs_build_gc_manager(struct f2fs_sb_info *sbi) init_atgc_management(sbi); } +int f2fs_gc_range(struct f2fs_sb_info *sbi, + unsigned int start_seg, unsigned int end_seg, + bool dry_run, unsigned int dry_run_sections) +{ + unsigned int segno; + unsigned int gc_secs = dry_run_sections; + + if (unlikely(f2fs_cp_error(sbi))) + return -EIO; + + for (segno = start_seg; segno <= end_seg; segno += SEGS_PER_SEC(sbi)) { + struct gc_inode_list gc_list = { + .ilist = LIST_HEAD_INIT(gc_list.ilist), + .iroot = RADIX_TREE_INIT(gc_list.iroot, GFP_NOFS), + }; + + do_garbage_collect(sbi, segno, &gc_list, FG_GC, + dry_run_sections == 0); + put_gc_inode(&gc_list); + + if (!dry_run && get_valid_blocks(sbi, segno, true)) + return -EAGAIN; + if (dry_run && dry_run_sections && + !get_valid_blocks(sbi, segno, true) && --gc_secs == 0) + break; + + if (fatal_signal_pending(current)) + return -ERESTARTSYS; + } + + return 0; +} + static int free_segment_range(struct f2fs_sb_info *sbi, - unsigned int secs, bool gc_only) + unsigned int secs, bool dry_run) { - unsigned int segno, next_inuse, start, end; + unsigned int next_inuse, start, end; struct cp_control cpc = { CP_RESIZE, 0, 0, 0 }; int gc_mode, gc_type; int err = 0; @@ -1994,7 +2021,7 @@ static int free_segment_range(struct f2fs_sb_info *sbi, /* Force block allocation for GC */ MAIN_SECS(sbi) -= secs; - start = MAIN_SECS(sbi) * sbi->segs_per_sec; + start = MAIN_SECS(sbi) * SEGS_PER_SEC(sbi); end = MAIN_SEGS(sbi) - 1; mutex_lock(&DIRTY_I(sbi)->seglist_lock); @@ -2008,29 +2035,15 @@ static int free_segment_range(struct f2fs_sb_info *sbi, mutex_unlock(&DIRTY_I(sbi)->seglist_lock); /* Move out cursegs from the target range */ - for (type = CURSEG_HOT_DATA; type < NR_CURSEG_PERSIST_TYPE; type++) - f2fs_allocate_segment_for_resize(sbi, type, start, end); - - /* do GC to move out valid blocks in the range */ - for (segno = start; segno <= end; segno += sbi->segs_per_sec) { - struct gc_inode_list gc_list = { - .ilist = LIST_HEAD_INIT(gc_list.ilist), - .iroot = RADIX_TREE_INIT(gc_list.iroot, GFP_NOFS), - }; - - do_garbage_collect(sbi, segno, &gc_list, FG_GC, true); - put_gc_inode(&gc_list); - - if (!gc_only && get_valid_blocks(sbi, segno, true)) { - err = -EAGAIN; - goto out; - } - if (fatal_signal_pending(current)) { - err = -ERESTARTSYS; + for (type = CURSEG_HOT_DATA; type < NR_CURSEG_PERSIST_TYPE; type++) { + err = f2fs_allocate_segment_for_resize(sbi, type, start, end); + if (err) goto out; - } } - if (gc_only) + + /* do GC to move out valid blocks in the range */ + err = f2fs_gc_range(sbi, start, end, dry_run, 0); + if (err || dry_run) goto out; stat_inc_cp_call_count(sbi, TOTAL_CALL); @@ -2056,7 +2069,7 @@ static void update_sb_metadata(struct f2fs_sb_info *sbi, int secs) int segment_count; int segment_count_main; long long block_count; - int segs = secs * sbi->segs_per_sec; + int segs = secs * SEGS_PER_SEC(sbi); f2fs_down_write(&sbi->sb_lock); @@ -2069,7 +2082,7 @@ static void update_sb_metadata(struct f2fs_sb_info *sbi, int secs) raw_sb->segment_count = cpu_to_le32(segment_count + segs); raw_sb->segment_count_main = cpu_to_le32(segment_count_main + segs); raw_sb->block_count = cpu_to_le64(block_count + - (long long)segs * sbi->blocks_per_seg); + (long long)SEGS_TO_BLKS(sbi, segs)); if (f2fs_is_multi_device(sbi)) { int last_dev = sbi->s_ndevs - 1; int dev_segs = @@ -2084,8 +2097,8 @@ static void update_sb_metadata(struct f2fs_sb_info *sbi, int secs) static void update_fs_metadata(struct f2fs_sb_info *sbi, int secs) { - int segs = secs * sbi->segs_per_sec; - long long blks = (long long)segs * sbi->blocks_per_seg; + int segs = secs * SEGS_PER_SEC(sbi); + long long blks = SEGS_TO_BLKS(sbi, segs); long long user_block_count = le64_to_cpu(F2FS_CKPT(sbi)->user_block_count); @@ -2127,7 +2140,7 @@ int f2fs_resize_fs(struct file *filp, __u64 block_count) int last_dev = sbi->s_ndevs - 1; __u64 last_segs = FDEV(last_dev).total_segments; - if (block_count + last_segs * sbi->blocks_per_seg <= + if (block_count + SEGS_TO_BLKS(sbi, last_segs) <= old_block_count) return -EINVAL; } diff --git a/fs/f2fs/gc.h b/fs/f2fs/gc.h index 28a00942802c..9c0d06c4d19a 100644 --- a/fs/f2fs/gc.h +++ b/fs/f2fs/gc.h @@ -96,7 +96,7 @@ static inline block_t free_segs_blk_count(struct f2fs_sb_info *sbi) if (f2fs_sb_has_blkzoned(sbi)) return free_segs_blk_count_zoned(sbi); - return free_segments(sbi) << sbi->log_blocks_per_seg; + return SEGS_TO_BLKS(sbi, free_segments(sbi)); } static inline block_t free_user_blocks(struct f2fs_sb_info *sbi) @@ -104,7 +104,7 @@ static inline block_t free_user_blocks(struct f2fs_sb_info *sbi) block_t free_blks, ovp_blks; free_blks = free_segs_blk_count(sbi); - ovp_blks = overprovision_segments(sbi) << sbi->log_blocks_per_seg; + ovp_blks = SEGS_TO_BLKS(sbi, overprovision_segments(sbi)); if (free_blks < ovp_blks) return 0; diff --git a/fs/f2fs/namei.c b/fs/f2fs/namei.c index f7f63a567d86..e54f8c08bda8 100644 --- a/fs/f2fs/namei.c +++ b/fs/f2fs/namei.c @@ -851,7 +851,7 @@ out: static int __f2fs_tmpfile(struct mnt_idmap *idmap, struct inode *dir, struct file *file, umode_t mode, bool is_whiteout, - struct inode **new_inode) + struct inode **new_inode, struct f2fs_filename *fname) { struct f2fs_sb_info *sbi = F2FS_I_SB(dir); struct inode *inode; @@ -879,7 +879,7 @@ static int __f2fs_tmpfile(struct mnt_idmap *idmap, struct inode *dir, if (err) goto out; - err = f2fs_do_tmpfile(inode, dir); + err = f2fs_do_tmpfile(inode, dir, fname); if (err) goto release_out; @@ -930,22 +930,24 @@ static int f2fs_tmpfile(struct mnt_idmap *idmap, struct inode *dir, if (!f2fs_is_checkpoint_ready(sbi)) return -ENOSPC; - err = __f2fs_tmpfile(idmap, dir, file, mode, false, NULL); + err = __f2fs_tmpfile(idmap, dir, file, mode, false, NULL, NULL); return finish_open_simple(file, err); } static int f2fs_create_whiteout(struct mnt_idmap *idmap, - struct inode *dir, struct inode **whiteout) + struct inode *dir, struct inode **whiteout, + struct f2fs_filename *fname) { - return __f2fs_tmpfile(idmap, dir, NULL, - S_IFCHR | WHITEOUT_MODE, true, whiteout); + return __f2fs_tmpfile(idmap, dir, NULL, S_IFCHR | WHITEOUT_MODE, + true, whiteout, fname); } int f2fs_get_tmpfile(struct mnt_idmap *idmap, struct inode *dir, struct inode **new_inode) { - return __f2fs_tmpfile(idmap, dir, NULL, S_IFREG, false, new_inode); + return __f2fs_tmpfile(idmap, dir, NULL, S_IFREG, + false, new_inode, NULL); } static int f2fs_rename(struct mnt_idmap *idmap, struct inode *old_dir, @@ -989,7 +991,14 @@ static int f2fs_rename(struct mnt_idmap *idmap, struct inode *old_dir, } if (flags & RENAME_WHITEOUT) { - err = f2fs_create_whiteout(idmap, old_dir, &whiteout); + struct f2fs_filename fname; + + err = f2fs_setup_filename(old_dir, &old_dentry->d_name, + 0, &fname); + if (err) + return err; + + err = f2fs_create_whiteout(idmap, old_dir, &whiteout, &fname); if (err) return err; } @@ -1104,14 +1113,11 @@ static int f2fs_rename(struct mnt_idmap *idmap, struct inode *old_dir, iput(whiteout); } - if (old_is_dir) { - if (old_dir_entry) - f2fs_set_link(old_inode, old_dir_entry, - old_dir_page, new_dir); - else - f2fs_put_page(old_dir_page, 0); + if (old_dir_entry) + f2fs_set_link(old_inode, old_dir_entry, old_dir_page, new_dir); + if (old_is_dir) f2fs_i_links_write(old_dir, false); - } + if (F2FS_OPTION(sbi).fsync_mode == FSYNC_MODE_STRICT) { f2fs_add_ino_entry(sbi, new_dir->i_ino, TRANS_DIR_INO); if (S_ISDIR(old_inode->i_mode)) diff --git a/fs/f2fs/node.c b/fs/f2fs/node.c index 9b546fd21010..b3de6d6cdb02 100644 --- a/fs/f2fs/node.c +++ b/fs/f2fs/node.c @@ -852,21 +852,29 @@ int f2fs_get_dnode_of_data(struct dnode_of_data *dn, pgoff_t index, int mode) if (is_inode_flag_set(dn->inode, FI_COMPRESSED_FILE) && f2fs_sb_has_readonly(sbi)) { - unsigned int c_len = f2fs_cluster_blocks_are_contiguous(dn); + unsigned int cluster_size = F2FS_I(dn->inode)->i_cluster_size; + unsigned int ofs_in_node = dn->ofs_in_node; + pgoff_t fofs = index; + unsigned int c_len; block_t blkaddr; + /* should align fofs and ofs_in_node to cluster_size */ + if (fofs % cluster_size) { + fofs = round_down(fofs, cluster_size); + ofs_in_node = round_down(ofs_in_node, cluster_size); + } + + c_len = f2fs_cluster_blocks_are_contiguous(dn, ofs_in_node); if (!c_len) goto out; - blkaddr = f2fs_data_blkaddr(dn); + blkaddr = data_blkaddr(dn->inode, dn->node_page, ofs_in_node); if (blkaddr == COMPRESS_ADDR) blkaddr = data_blkaddr(dn->inode, dn->node_page, - dn->ofs_in_node + 1); + ofs_in_node + 1); f2fs_update_read_extent_tree_range_compressed(dn->inode, - index, blkaddr, - F2FS_I(dn->inode)->i_cluster_size, - c_len); + fofs, blkaddr, cluster_size, c_len); } out: return 0; @@ -1919,7 +1927,7 @@ void f2fs_flush_inline_data(struct f2fs_sb_info *sbi) for (i = 0; i < nr_folios; i++) { struct page *page = &fbatch.folios[i]->page; - if (!IS_DNODE(page)) + if (!IS_INODE(page)) continue; lock_page(page); @@ -2841,7 +2849,7 @@ int f2fs_restore_node_summary(struct f2fs_sb_info *sbi, int i, idx, last_offset, nrpages; /* scan the node segment */ - last_offset = sbi->blocks_per_seg; + last_offset = BLKS_PER_SEG(sbi); addr = START_BLOCK(sbi, segno); sum_entry = &sum->entries[0]; @@ -3158,7 +3166,7 @@ static int __get_nat_bitmaps(struct f2fs_sb_info *sbi) if (!is_set_ckpt_flags(sbi, CP_NAT_BITS_FLAG)) return 0; - nat_bits_addr = __start_cp_addr(sbi) + sbi->blocks_per_seg - + nat_bits_addr = __start_cp_addr(sbi) + BLKS_PER_SEG(sbi) - nm_i->nat_bits_blocks; for (i = 0; i < nm_i->nat_bits_blocks; i++) { struct page *page; diff --git a/fs/f2fs/node.h b/fs/f2fs/node.h index 5bd16a95eef8..6aea13024ac1 100644 --- a/fs/f2fs/node.h +++ b/fs/f2fs/node.h @@ -208,10 +208,10 @@ static inline pgoff_t current_nat_addr(struct f2fs_sb_info *sbi, nid_t start) block_addr = (pgoff_t)(nm_i->nat_blkaddr + (block_off << 1) - - (block_off & (sbi->blocks_per_seg - 1))); + (block_off & (BLKS_PER_SEG(sbi) - 1))); if (f2fs_test_bit(block_off, nm_i->nat_bitmap)) - block_addr += sbi->blocks_per_seg; + block_addr += BLKS_PER_SEG(sbi); return block_addr; } diff --git a/fs/f2fs/recovery.c b/fs/f2fs/recovery.c index d0f24ccbd1ac..e7bf15b8240a 100644 --- a/fs/f2fs/recovery.c +++ b/fs/f2fs/recovery.c @@ -354,7 +354,7 @@ static unsigned int adjust_por_ra_blocks(struct f2fs_sb_info *sbi, if (blkaddr + 1 == next_blkaddr) ra_blocks = min_t(unsigned int, RECOVERY_MAX_RA_BLOCKS, ra_blocks * 2); - else if (next_blkaddr % sbi->blocks_per_seg) + else if (next_blkaddr % BLKS_PER_SEG(sbi)) ra_blocks = max_t(unsigned int, RECOVERY_MIN_RA_BLOCKS, ra_blocks / 2); return ra_blocks; @@ -611,6 +611,19 @@ truncate_out: return 0; } +static int f2fs_reserve_new_block_retry(struct dnode_of_data *dn) +{ + int i, err = 0; + + for (i = DEFAULT_FAILURE_RETRY_COUNT; i > 0; i--) { + err = f2fs_reserve_new_block(dn); + if (!err) + break; + } + + return err; +} + static int do_recover_data(struct f2fs_sb_info *sbi, struct inode *inode, struct page *page) { @@ -680,14 +693,12 @@ retry_dn: if (__is_valid_data_blkaddr(src) && !f2fs_is_valid_blkaddr(sbi, src, META_POR)) { err = -EFSCORRUPTED; - f2fs_handle_error(sbi, ERROR_INVALID_BLKADDR); goto err; } if (__is_valid_data_blkaddr(dest) && !f2fs_is_valid_blkaddr(sbi, dest, META_POR)) { err = -EFSCORRUPTED; - f2fs_handle_error(sbi, ERROR_INVALID_BLKADDR); goto err; } @@ -712,14 +723,8 @@ retry_dn: */ if (dest == NEW_ADDR) { f2fs_truncate_data_blocks_range(&dn, 1); - do { - err = f2fs_reserve_new_block(&dn); - if (err == -ENOSPC) { - f2fs_bug_on(sbi, 1); - break; - } - } while (err && - IS_ENABLED(CONFIG_F2FS_FAULT_INJECTION)); + + err = f2fs_reserve_new_block_retry(&dn); if (err) goto err; continue; @@ -727,16 +732,8 @@ retry_dn: /* dest is valid block, try to recover from src to dest */ if (f2fs_is_valid_blkaddr(sbi, dest, META_POR)) { - if (src == NULL_ADDR) { - do { - err = f2fs_reserve_new_block(&dn); - if (err == -ENOSPC) { - f2fs_bug_on(sbi, 1); - break; - } - } while (err && - IS_ENABLED(CONFIG_F2FS_FAULT_INJECTION)); + err = f2fs_reserve_new_block_retry(&dn); if (err) goto err; } @@ -756,8 +753,6 @@ retry_prev: f2fs_err(sbi, "Inconsistent dest blkaddr:%u, ino:%lu, ofs:%u", dest, inode->i_ino, dn.ofs_in_node); err = -EFSCORRUPTED; - f2fs_handle_error(sbi, - ERROR_INVALID_BLKADDR); goto err; } @@ -852,7 +847,7 @@ next: f2fs_ra_meta_pages_cond(sbi, blkaddr, ra_blocks); } if (!err) - f2fs_allocate_new_segments(sbi); + err = f2fs_allocate_new_segments(sbi); return err; } @@ -864,7 +859,6 @@ int f2fs_recover_fsync_data(struct f2fs_sb_info *sbi, bool check_only) int ret = 0; unsigned long s_flags = sbi->sb->s_flags; bool need_writecp = false; - bool fix_curseg_write_pointer = false; if (is_sbi_flag_set(sbi, SBI_IS_WRITABLE)) f2fs_info(sbi, "recover fsync data on readonly fs"); @@ -895,8 +889,6 @@ int f2fs_recover_fsync_data(struct f2fs_sb_info *sbi, bool check_only) else f2fs_bug_on(sbi, sbi->sb->s_flags & SB_ACTIVE); skip: - fix_curseg_write_pointer = !check_only || list_empty(&inode_list); - destroy_fsync_dnodes(&inode_list, err); destroy_fsync_dnodes(&tmp_inode_list, err); @@ -914,11 +906,13 @@ skip: * and the f2fs is not read only, check and fix zoned block devices' * write pointer consistency. */ - if (!err && fix_curseg_write_pointer && !f2fs_readonly(sbi->sb) && - f2fs_sb_has_blkzoned(sbi)) { - err = f2fs_fix_curseg_write_pointer(sbi); - if (!err) - err = f2fs_check_write_pointer(sbi); + if (f2fs_sb_has_blkzoned(sbi) && !f2fs_readonly(sbi->sb)) { + int err2 = f2fs_fix_curseg_write_pointer(sbi); + + if (!err2) + err2 = f2fs_check_write_pointer(sbi); + if (err2) + err = err2; ret = err; } diff --git a/fs/f2fs/segment.c b/fs/f2fs/segment.c index e1065ba70207..4fd76e867e0a 100644 --- a/fs/f2fs/segment.c +++ b/fs/f2fs/segment.c @@ -192,6 +192,9 @@ void f2fs_abort_atomic_write(struct inode *inode, bool clean) if (!f2fs_is_atomic_file(inode)) return; + if (clean) + truncate_inode_pages_final(inode->i_mapping); + release_atomic_write_cnt(inode); clear_inode_flag(inode, FI_ATOMIC_COMMITTED); clear_inode_flag(inode, FI_ATOMIC_REPLACE); @@ -201,7 +204,6 @@ void f2fs_abort_atomic_write(struct inode *inode, bool clean) F2FS_I(inode)->atomic_write_task = NULL; if (clean) { - truncate_inode_pages_final(inode->i_mapping); f2fs_i_size_write(inode, fi->original_i_size); fi->original_i_size = 0; } @@ -248,7 +250,7 @@ retry: } else { blkcnt_t count = 1; - err = inc_valid_block_count(sbi, inode, &count); + err = inc_valid_block_count(sbi, inode, &count, true); if (err) { f2fs_put_dnode(&dn); return err; @@ -334,8 +336,6 @@ static int __f2fs_commit_atomic_write(struct inode *inode) DATA_GENERIC_ENHANCE)) { f2fs_put_dnode(&dn); ret = -EFSCORRUPTED; - f2fs_handle_error(sbi, - ERROR_INVALID_BLKADDR); goto out; } @@ -400,6 +400,9 @@ int f2fs_commit_atomic_write(struct inode *inode) */ void f2fs_balance_fs(struct f2fs_sb_info *sbi, bool need) { + if (f2fs_cp_error(sbi)) + return; + if (time_to_inject(sbi, FAULT_CHECKPOINT)) f2fs_stop_checkpoint(sbi, false, STOP_CP_REASON_FAULT_INJECT); @@ -448,8 +451,8 @@ static inline bool excess_dirty_threshold(struct f2fs_sb_info *sbi) unsigned int nodes = get_pages(sbi, F2FS_DIRTY_NODES); unsigned int meta = get_pages(sbi, F2FS_DIRTY_META); unsigned int imeta = get_pages(sbi, F2FS_DIRTY_IMETA); - unsigned int threshold = sbi->blocks_per_seg * factor * - DEFAULT_DIRTY_THRESHOLD; + unsigned int threshold = + SEGS_TO_BLKS(sbi, (factor * DEFAULT_DIRTY_THRESHOLD)); unsigned int global_threshold = threshold * 3 / 2; if (dents >= threshold || qdata >= threshold || @@ -872,7 +875,7 @@ block_t f2fs_get_unusable_blocks(struct f2fs_sb_info *sbi) { int ovp_hole_segs = (overprovision_segments(sbi) - reserved_segments(sbi)); - block_t ovp_holes = ovp_hole_segs << sbi->log_blocks_per_seg; + block_t ovp_holes = SEGS_TO_BLKS(sbi, ovp_hole_segs); struct dirty_seglist_info *dirty_i = DIRTY_I(sbi); block_t holes[2] = {0, 0}; /* DATA and NODE */ block_t unusable; @@ -901,11 +904,16 @@ int f2fs_disable_cp_again(struct f2fs_sb_info *sbi, block_t unusable) { int ovp_hole_segs = (overprovision_segments(sbi) - reserved_segments(sbi)); + + if (F2FS_OPTION(sbi).unusable_cap_perc == 100) + return 0; if (unusable > F2FS_OPTION(sbi).unusable_cap) return -EAGAIN; if (is_sbi_flag_set(sbi, SBI_CP_DISABLED_QUICK) && dirty_segments(sbi) > ovp_hole_segs) return -EAGAIN; + if (has_not_enough_free_secs(sbi, 0, 0)) + return -EAGAIN; return 0; } @@ -1132,8 +1140,7 @@ static void __check_sit_bitmap(struct f2fs_sb_info *sbi, struct seg_entry *sentry; unsigned int segno; block_t blk = start; - unsigned long offset, size, max_blocks = sbi->blocks_per_seg; - unsigned long *map; + unsigned long offset, size, *map; while (blk < end) { segno = GET_SEGNO(sbi, blk); @@ -1143,7 +1150,7 @@ static void __check_sit_bitmap(struct f2fs_sb_info *sbi, if (end < START_BLOCK(sbi, segno + 1)) size = GET_BLKOFF_FROM_SEG0(sbi, end); else - size = max_blocks; + size = BLKS_PER_SEG(sbi); map = (unsigned long *)(sentry->cur_valid_map); offset = __find_rev_next_bit(map, size, offset); f2fs_bug_on(sbi, offset != size); @@ -2048,7 +2055,6 @@ static bool add_discard_addrs(struct f2fs_sb_info *sbi, struct cp_control *cpc, bool check_only) { int entries = SIT_VBLOCK_MAP_SIZE / sizeof(unsigned long); - int max_blocks = sbi->blocks_per_seg; struct seg_entry *se = get_seg_entry(sbi, cpc->trim_start); unsigned long *cur_map = (unsigned long *)se->cur_valid_map; unsigned long *ckpt_map = (unsigned long *)se->ckpt_valid_map; @@ -2060,8 +2066,9 @@ static bool add_discard_addrs(struct f2fs_sb_info *sbi, struct cp_control *cpc, struct list_head *head = &SM_I(sbi)->dcc_info->entry_list; int i; - if (se->valid_blocks == max_blocks || !f2fs_hw_support_discard(sbi) || - !f2fs_block_unit_discard(sbi)) + if (se->valid_blocks == BLKS_PER_SEG(sbi) || + !f2fs_hw_support_discard(sbi) || + !f2fs_block_unit_discard(sbi)) return false; if (!force) { @@ -2078,13 +2085,14 @@ static bool add_discard_addrs(struct f2fs_sb_info *sbi, struct cp_control *cpc, while (force || SM_I(sbi)->dcc_info->nr_discards <= SM_I(sbi)->dcc_info->max_discards) { - start = __find_rev_next_bit(dmap, max_blocks, end + 1); - if (start >= max_blocks) + start = __find_rev_next_bit(dmap, BLKS_PER_SEG(sbi), end + 1); + if (start >= BLKS_PER_SEG(sbi)) break; - end = __find_rev_next_zero_bit(dmap, max_blocks, start + 1); - if (force && start && end != max_blocks - && (end - start) < cpc->trim_minlen) + end = __find_rev_next_zero_bit(dmap, + BLKS_PER_SEG(sbi), start + 1); + if (force && start && end != BLKS_PER_SEG(sbi) && + (end - start) < cpc->trim_minlen) continue; if (check_only) @@ -2166,8 +2174,8 @@ void f2fs_clear_prefree_segments(struct f2fs_sb_info *sbi, start + 1); if (section_alignment) { - start = rounddown(start, sbi->segs_per_sec); - end = roundup(end, sbi->segs_per_sec); + start = rounddown(start, SEGS_PER_SEC(sbi)); + end = roundup(end, SEGS_PER_SEC(sbi)); } for (i = start; i < end; i++) { @@ -2186,7 +2194,7 @@ void f2fs_clear_prefree_segments(struct f2fs_sb_info *sbi, if (!f2fs_sb_has_blkzoned(sbi) && (!f2fs_lfs_mode(sbi) || !__is_large_section(sbi))) { f2fs_issue_discard(sbi, START_BLOCK(sbi, start), - (end - start) << sbi->log_blocks_per_seg); + SEGS_TO_BLKS(sbi, end - start)); continue; } next: @@ -2195,9 +2203,9 @@ next: if (!IS_CURSEC(sbi, secno) && !get_valid_blocks(sbi, start, true)) f2fs_issue_discard(sbi, START_BLOCK(sbi, start_segno), - sbi->segs_per_sec << sbi->log_blocks_per_seg); + BLKS_PER_SEC(sbi)); - start = start_segno + sbi->segs_per_sec; + start = start_segno + SEGS_PER_SEC(sbi); if (start < end) goto next; else @@ -2216,7 +2224,7 @@ next: find_next: if (is_valid) { next_pos = find_next_zero_bit_le(entry->discard_map, - sbi->blocks_per_seg, cur_pos); + BLKS_PER_SEG(sbi), cur_pos); len = next_pos - cur_pos; if (f2fs_sb_has_blkzoned(sbi) || @@ -2228,13 +2236,13 @@ find_next: total_len += len; } else { next_pos = find_next_bit_le(entry->discard_map, - sbi->blocks_per_seg, cur_pos); + BLKS_PER_SEG(sbi), cur_pos); } skip: cur_pos = next_pos; is_valid = !is_valid; - if (cur_pos < sbi->blocks_per_seg) + if (cur_pos < BLKS_PER_SEG(sbi)) goto find_next; release_discard_addr(entry); @@ -2251,6 +2259,12 @@ int f2fs_start_discard_thread(struct f2fs_sb_info *sbi) struct discard_cmd_control *dcc = SM_I(sbi)->dcc_info; int err = 0; + if (f2fs_sb_has_readonly(sbi)) { + f2fs_info(sbi, + "Skip to start discard thread for readonly image"); + return 0; + } + if (!f2fs_realtime_discard_enable(sbi)) return 0; @@ -2283,7 +2297,7 @@ static int create_discard_cmd_control(struct f2fs_sb_info *sbi) dcc->max_ordered_discard = DEFAULT_MAX_ORDERED_DISCARD_GRANULARITY; dcc->discard_io_aware = DPOLICY_IO_AWARE_ENABLE; if (F2FS_OPTION(sbi).discard_unit == DISCARD_UNIT_SEGMENT) - dcc->discard_granularity = sbi->blocks_per_seg; + dcc->discard_granularity = BLKS_PER_SEG(sbi); else if (F2FS_OPTION(sbi).discard_unit == DISCARD_UNIT_SECTION) dcc->discard_granularity = BLKS_PER_SEC(sbi); @@ -2297,7 +2311,7 @@ static int create_discard_cmd_control(struct f2fs_sb_info *sbi) atomic_set(&dcc->queued_discard, 0); atomic_set(&dcc->discard_cmd_cnt, 0); dcc->nr_discards = 0; - dcc->max_discards = MAIN_SEGS(sbi) << sbi->log_blocks_per_seg; + dcc->max_discards = SEGS_TO_BLKS(sbi, MAIN_SEGS(sbi)); dcc->max_discard_request = DEF_MAX_DISCARD_REQUEST; dcc->min_discard_issue_time = DEF_MIN_DISCARD_ISSUE_TIME; dcc->mid_discard_issue_time = DEF_MID_DISCARD_ISSUE_TIME; @@ -2405,6 +2419,8 @@ static void update_sit_entry(struct f2fs_sb_info *sbi, block_t blkaddr, int del) #endif segno = GET_SEGNO(sbi, blkaddr); + if (segno == NULL_SEGNO) + return; se = get_seg_entry(sbi, segno); new_vblocks = se->valid_blocks + del; @@ -2546,7 +2562,7 @@ static unsigned short f2fs_curseg_valid_blocks(struct f2fs_sb_info *sbi, int typ struct curseg_info *curseg = CURSEG_I(sbi, type); if (sbi->ckpt->alloc_type[type] == SSR) - return sbi->blocks_per_seg; + return BLKS_PER_SEG(sbi); return curseg->next_blkoff; } @@ -2634,7 +2650,7 @@ static int is_next_segment_free(struct f2fs_sb_info *sbi, unsigned int segno = curseg->segno + 1; struct free_segmap_info *free_i = FREE_I(sbi); - if (segno < MAIN_SEGS(sbi) && segno % sbi->segs_per_sec) + if (segno < MAIN_SEGS(sbi) && segno % SEGS_PER_SEC(sbi)) return !test_bit(segno, free_i->free_segmap); return 0; } @@ -2643,54 +2659,51 @@ static int is_next_segment_free(struct f2fs_sb_info *sbi, * Find a new segment from the free segments bitmap to right order * This function should be returned with success, otherwise BUG */ -static void get_new_segment(struct f2fs_sb_info *sbi, - unsigned int *newseg, bool new_sec, int dir) +static int get_new_segment(struct f2fs_sb_info *sbi, + unsigned int *newseg, bool new_sec, bool pinning) { struct free_segmap_info *free_i = FREE_I(sbi); unsigned int segno, secno, zoneno; unsigned int total_zones = MAIN_SECS(sbi) / sbi->secs_per_zone; unsigned int hint = GET_SEC_FROM_SEG(sbi, *newseg); unsigned int old_zoneno = GET_ZONE_FROM_SEG(sbi, *newseg); - unsigned int left_start = hint; bool init = true; - int go_left = 0; int i; + int ret = 0; spin_lock(&free_i->segmap_lock); - if (!new_sec && ((*newseg + 1) % sbi->segs_per_sec)) { + if (time_to_inject(sbi, FAULT_NO_SEGMENT)) { + ret = -ENOSPC; + goto out_unlock; + } + + if (!new_sec && ((*newseg + 1) % SEGS_PER_SEC(sbi))) { segno = find_next_zero_bit(free_i->free_segmap, GET_SEG_FROM_SEC(sbi, hint + 1), *newseg + 1); if (segno < GET_SEG_FROM_SEC(sbi, hint + 1)) goto got_it; } + + /* + * If we format f2fs on zoned storage, let's try to get pinned sections + * from beginning of the storage, which should be a conventional one. + */ + if (f2fs_sb_has_blkzoned(sbi)) { + segno = pinning ? 0 : max(first_zoned_segno(sbi), *newseg); + hint = GET_SEC_FROM_SEG(sbi, segno); + } + find_other_zone: secno = find_next_zero_bit(free_i->free_secmap, MAIN_SECS(sbi), hint); if (secno >= MAIN_SECS(sbi)) { - if (dir == ALLOC_RIGHT) { - secno = find_first_zero_bit(free_i->free_secmap, + secno = find_first_zero_bit(free_i->free_secmap, MAIN_SECS(sbi)); - f2fs_bug_on(sbi, secno >= MAIN_SECS(sbi)); - } else { - go_left = 1; - left_start = hint - 1; + if (secno >= MAIN_SECS(sbi)) { + ret = -ENOSPC; + goto out_unlock; } } - if (go_left == 0) - goto skip_left; - - while (test_bit(left_start, free_i->free_secmap)) { - if (left_start > 0) { - left_start--; - continue; - } - left_start = find_first_zero_bit(free_i->free_secmap, - MAIN_SECS(sbi)); - f2fs_bug_on(sbi, left_start >= MAIN_SECS(sbi)); - break; - } - secno = left_start; -skip_left: segno = GET_SEG_FROM_SEC(sbi, secno); zoneno = GET_ZONE_FROM_SEC(sbi, secno); @@ -2701,21 +2714,13 @@ skip_left: goto got_it; if (zoneno == old_zoneno) goto got_it; - if (dir == ALLOC_LEFT) { - if (!go_left && zoneno + 1 >= total_zones) - goto got_it; - if (go_left && zoneno == 0) - goto got_it; - } for (i = 0; i < NR_CURSEG_TYPE; i++) if (CURSEG_I(sbi, i)->zone == zoneno) break; if (i < NR_CURSEG_TYPE) { /* zone is in user, try another */ - if (go_left) - hint = zoneno * sbi->secs_per_zone - 1; - else if (zoneno + 1 >= total_zones) + if (zoneno + 1 >= total_zones) hint = 0; else hint = (zoneno + 1) * sbi->secs_per_zone; @@ -2725,9 +2730,23 @@ skip_left: got_it: /* set it as dirty segment in free segmap */ f2fs_bug_on(sbi, test_bit(segno, free_i->free_segmap)); + + /* no free section in conventional zone */ + if (new_sec && pinning && + !f2fs_valid_pinned_area(sbi, START_BLOCK(sbi, segno))) { + ret = -EAGAIN; + goto out_unlock; + } __set_inuse(sbi, segno); *newseg = segno; +out_unlock: spin_unlock(&free_i->segmap_lock); + + if (ret == -ENOSPC) { + f2fs_stop_checkpoint(sbi, false, STOP_CP_REASON_NO_SEGMENT); + f2fs_bug_on(sbi, 1); + } + return ret; } static void reset_curseg(struct f2fs_sb_info *sbi, int type, int modified) @@ -2736,6 +2755,10 @@ static void reset_curseg(struct f2fs_sb_info *sbi, int type, int modified) struct summary_footer *sum_footer; unsigned short seg_type = curseg->seg_type; + /* only happen when get_new_segment() fails */ + if (curseg->next_segno == NULL_SEGNO) + return; + curseg->inited = true; curseg->segno = curseg->next_segno; curseg->zone = GET_ZONE_FROM_SEG(sbi, curseg->segno); @@ -2761,9 +2784,8 @@ static unsigned int __get_next_segno(struct f2fs_sb_info *sbi, int type) sanity_check_seg_type(sbi, seg_type); if (f2fs_need_rand_seg(sbi)) - return get_random_u32_below(MAIN_SECS(sbi) * sbi->segs_per_sec); + return get_random_u32_below(MAIN_SECS(sbi) * SEGS_PER_SEC(sbi)); - /* if segs_per_sec is large than 1, we need to keep original policy. */ if (__is_large_section(sbi)) return curseg->segno; @@ -2774,8 +2796,7 @@ static unsigned int __get_next_segno(struct f2fs_sb_info *sbi, int type) if (unlikely(is_sbi_flag_set(sbi, SBI_CP_DISABLED))) return 0; - if (test_opt(sbi, NOHEAP) && - (seg_type == CURSEG_HOT_DATA || IS_NODESEG(seg_type))) + if (seg_type == CURSEG_HOT_DATA || IS_NODESEG(seg_type)) return 0; if (SIT_I(sbi)->last_victim[ALLOC_NEXT]) @@ -2792,30 +2813,31 @@ static unsigned int __get_next_segno(struct f2fs_sb_info *sbi, int type) * Allocate a current working segment. * This function always allocates a free segment in LFS manner. */ -static void new_curseg(struct f2fs_sb_info *sbi, int type, bool new_sec) +static int new_curseg(struct f2fs_sb_info *sbi, int type, bool new_sec) { struct curseg_info *curseg = CURSEG_I(sbi, type); - unsigned short seg_type = curseg->seg_type; unsigned int segno = curseg->segno; - int dir = ALLOC_LEFT; + bool pinning = type == CURSEG_COLD_DATA_PINNED; + int ret; if (curseg->inited) - write_sum_page(sbi, curseg->sum_blk, - GET_SUM_BLOCK(sbi, segno)); - if (seg_type == CURSEG_WARM_DATA || seg_type == CURSEG_COLD_DATA) - dir = ALLOC_RIGHT; - - if (test_opt(sbi, NOHEAP)) - dir = ALLOC_RIGHT; + write_sum_page(sbi, curseg->sum_blk, GET_SUM_BLOCK(sbi, segno)); segno = __get_next_segno(sbi, type); - get_new_segment(sbi, &segno, new_sec, dir); + ret = get_new_segment(sbi, &segno, new_sec, pinning); + if (ret) { + if (ret == -ENOSPC) + curseg->segno = NULL_SEGNO; + return ret; + } + curseg->next_segno = segno; reset_curseg(sbi, type, 1); curseg->alloc_type = LFS; if (F2FS_OPTION(sbi).fs_mode == FS_MODE_FRAGMENT_BLK) curseg->fragment_remained_chunk = get_random_u32_inclusive(1, sbi->max_fragment_chunk); + return 0; } static int __next_free_blkoff(struct f2fs_sb_info *sbi, @@ -2831,7 +2853,7 @@ static int __next_free_blkoff(struct f2fs_sb_info *sbi, for (i = 0; i < entries; i++) target_map[i] = ckpt_map[i] | cur_map[i]; - return __find_rev_next_zero_bit(target_map, sbi->blocks_per_seg, start); + return __find_rev_next_zero_bit(target_map, BLKS_PER_SEG(sbi), start); } static int f2fs_find_next_ssr_block(struct f2fs_sb_info *sbi, @@ -2842,14 +2864,14 @@ static int f2fs_find_next_ssr_block(struct f2fs_sb_info *sbi, bool f2fs_segment_has_free_slot(struct f2fs_sb_info *sbi, int segno) { - return __next_free_blkoff(sbi, segno, 0) < sbi->blocks_per_seg; + return __next_free_blkoff(sbi, segno, 0) < BLKS_PER_SEG(sbi); } /* * This function always allocates a used segment(from dirty seglist) by SSR * manner, so it should recover the existing segment information of valid blocks */ -static void change_curseg(struct f2fs_sb_info *sbi, int type) +static int change_curseg(struct f2fs_sb_info *sbi, int type) { struct dirty_seglist_info *dirty_i = DIRTY_I(sbi); struct curseg_info *curseg = CURSEG_I(sbi, type); @@ -2874,21 +2896,23 @@ static void change_curseg(struct f2fs_sb_info *sbi, int type) if (IS_ERR(sum_page)) { /* GC won't be able to use stale summary pages by cp_error */ memset(curseg->sum_blk, 0, SUM_ENTRY_SIZE); - return; + return PTR_ERR(sum_page); } sum_node = (struct f2fs_summary_block *)page_address(sum_page); memcpy(curseg->sum_blk, sum_node, SUM_ENTRY_SIZE); f2fs_put_page(sum_page, 1); + return 0; } static int get_ssr_segment(struct f2fs_sb_info *sbi, int type, int alloc_mode, unsigned long long age); -static void get_atssr_segment(struct f2fs_sb_info *sbi, int type, +static int get_atssr_segment(struct f2fs_sb_info *sbi, int type, int target_type, int alloc_mode, unsigned long long age) { struct curseg_info *curseg = CURSEG_I(sbi, type); + int ret = 0; curseg->seg_type = target_type; @@ -2896,38 +2920,41 @@ static void get_atssr_segment(struct f2fs_sb_info *sbi, int type, struct seg_entry *se = get_seg_entry(sbi, curseg->next_segno); curseg->seg_type = se->type; - change_curseg(sbi, type); + ret = change_curseg(sbi, type); } else { /* allocate cold segment by default */ curseg->seg_type = CURSEG_COLD_DATA; - new_curseg(sbi, type, true); + ret = new_curseg(sbi, type, true); } stat_inc_seg_type(sbi, curseg); + return ret; } -static void __f2fs_init_atgc_curseg(struct f2fs_sb_info *sbi) +static int __f2fs_init_atgc_curseg(struct f2fs_sb_info *sbi) { struct curseg_info *curseg = CURSEG_I(sbi, CURSEG_ALL_DATA_ATGC); + int ret = 0; if (!sbi->am.atgc_enabled) - return; + return 0; f2fs_down_read(&SM_I(sbi)->curseg_lock); mutex_lock(&curseg->curseg_mutex); down_write(&SIT_I(sbi)->sentry_lock); - get_atssr_segment(sbi, CURSEG_ALL_DATA_ATGC, CURSEG_COLD_DATA, SSR, 0); + ret = get_atssr_segment(sbi, CURSEG_ALL_DATA_ATGC, + CURSEG_COLD_DATA, SSR, 0); up_write(&SIT_I(sbi)->sentry_lock); mutex_unlock(&curseg->curseg_mutex); f2fs_up_read(&SM_I(sbi)->curseg_lock); - + return ret; } -void f2fs_init_inmem_curseg(struct f2fs_sb_info *sbi) +int f2fs_init_inmem_curseg(struct f2fs_sb_info *sbi) { - __f2fs_init_atgc_curseg(sbi); + return __f2fs_init_atgc_curseg(sbi); } static void __f2fs_save_inmem_curseg(struct f2fs_sb_info *sbi, int type) @@ -3055,11 +3082,12 @@ static bool need_new_seg(struct f2fs_sb_info *sbi, int type) return false; } -void f2fs_allocate_segment_for_resize(struct f2fs_sb_info *sbi, int type, +int f2fs_allocate_segment_for_resize(struct f2fs_sb_info *sbi, int type, unsigned int start, unsigned int end) { struct curseg_info *curseg = CURSEG_I(sbi, type); unsigned int segno; + int ret = 0; f2fs_down_read(&SM_I(sbi)->curseg_lock); mutex_lock(&curseg->curseg_mutex); @@ -3070,9 +3098,9 @@ void f2fs_allocate_segment_for_resize(struct f2fs_sb_info *sbi, int type, goto unlock; if (f2fs_need_SSR(sbi) && get_ssr_segment(sbi, type, SSR, 0)) - change_curseg(sbi, type); + ret = change_curseg(sbi, type); else - new_curseg(sbi, type, true); + ret = new_curseg(sbi, type, true); stat_inc_seg_type(sbi, curseg); @@ -3086,45 +3114,84 @@ unlock: mutex_unlock(&curseg->curseg_mutex); f2fs_up_read(&SM_I(sbi)->curseg_lock); + return ret; } -static void __allocate_new_segment(struct f2fs_sb_info *sbi, int type, +static int __allocate_new_segment(struct f2fs_sb_info *sbi, int type, bool new_sec, bool force) { struct curseg_info *curseg = CURSEG_I(sbi, type); unsigned int old_segno; + int err = 0; + + if (type == CURSEG_COLD_DATA_PINNED && !curseg->inited) + goto allocate; if (!force && curseg->inited && !curseg->next_blkoff && !get_valid_blocks(sbi, curseg->segno, new_sec) && !get_ckpt_valid_blocks(sbi, curseg->segno, new_sec)) - return; + return 0; +allocate: old_segno = curseg->segno; - new_curseg(sbi, type, true); + err = new_curseg(sbi, type, true); + if (err) + return err; stat_inc_seg_type(sbi, curseg); locate_dirty_segment(sbi, old_segno); + return 0; } -void f2fs_allocate_new_section(struct f2fs_sb_info *sbi, int type, bool force) +int f2fs_allocate_new_section(struct f2fs_sb_info *sbi, int type, bool force) { + int ret; + f2fs_down_read(&SM_I(sbi)->curseg_lock); down_write(&SIT_I(sbi)->sentry_lock); - __allocate_new_segment(sbi, type, true, force); + ret = __allocate_new_segment(sbi, type, true, force); up_write(&SIT_I(sbi)->sentry_lock); f2fs_up_read(&SM_I(sbi)->curseg_lock); + + return ret; } -void f2fs_allocate_new_segments(struct f2fs_sb_info *sbi) +int f2fs_allocate_pinning_section(struct f2fs_sb_info *sbi) +{ + int err; + bool gc_required = true; + +retry: + f2fs_lock_op(sbi); + err = f2fs_allocate_new_section(sbi, CURSEG_COLD_DATA_PINNED, false); + f2fs_unlock_op(sbi); + + if (f2fs_sb_has_blkzoned(sbi) && err == -EAGAIN && gc_required) { + f2fs_down_write(&sbi->gc_lock); + err = f2fs_gc_range(sbi, 0, GET_SEGNO(sbi, FDEV(0).end_blk), true, 1); + f2fs_up_write(&sbi->gc_lock); + + gc_required = false; + if (!err) + goto retry; + } + + return err; +} + +int f2fs_allocate_new_segments(struct f2fs_sb_info *sbi) { int i; + int err = 0; f2fs_down_read(&SM_I(sbi)->curseg_lock); down_write(&SIT_I(sbi)->sentry_lock); for (i = CURSEG_HOT_DATA; i <= CURSEG_COLD_DATA; i++) - __allocate_new_segment(sbi, i, false, false); + err += __allocate_new_segment(sbi, i, false, false); up_write(&SIT_I(sbi)->sentry_lock); f2fs_up_read(&SM_I(sbi)->curseg_lock); + + return err; } bool f2fs_exist_trim_candidates(struct f2fs_sb_info *sbi, @@ -3242,8 +3309,8 @@ int f2fs_trim_fs(struct f2fs_sb_info *sbi, struct fstrim_range *range) end_segno = (end >= MAX_BLKADDR(sbi)) ? MAIN_SEGS(sbi) - 1 : GET_SEGNO(sbi, end); if (need_align) { - start_segno = rounddown(start_segno, sbi->segs_per_sec); - end_segno = roundup(end_segno + 1, sbi->segs_per_sec) - 1; + start_segno = rounddown(start_segno, SEGS_PER_SEC(sbi)); + end_segno = roundup(end_segno + 1, SEGS_PER_SEC(sbi)) - 1; } cpc.reason = CP_DISCARD; @@ -3416,7 +3483,14 @@ static void f2fs_randomize_chunk(struct f2fs_sb_info *sbi, get_random_u32_inclusive(1, sbi->max_fragment_hole); } -void f2fs_allocate_data_block(struct f2fs_sb_info *sbi, struct page *page, +static void reset_curseg_fields(struct curseg_info *curseg) +{ + curseg->inited = false; + curseg->segno = NULL_SEGNO; + curseg->next_segno = 0; +} + +int f2fs_allocate_data_block(struct f2fs_sb_info *sbi, struct page *page, block_t old_blkaddr, block_t *new_blkaddr, struct f2fs_summary *sum, int type, struct f2fs_io_info *fio) @@ -3427,12 +3501,18 @@ void f2fs_allocate_data_block(struct f2fs_sb_info *sbi, struct page *page, bool from_gc = (type == CURSEG_ALL_DATA_ATGC); struct seg_entry *se = NULL; bool segment_full = false; + int ret = 0; f2fs_down_read(&SM_I(sbi)->curseg_lock); mutex_lock(&curseg->curseg_mutex); down_write(&sit_i->sentry_lock); + if (curseg->segno == NULL_SEGNO) { + ret = -ENOSPC; + goto out_err; + } + if (from_gc) { f2fs_bug_on(sbi, GET_SEGNO(sbi, old_blkaddr) == NULL_SEGNO); se = get_seg_entry(sbi, GET_SEGNO(sbi, old_blkaddr)); @@ -3441,7 +3521,7 @@ void f2fs_allocate_data_block(struct f2fs_sb_info *sbi, struct page *page, } *new_blkaddr = NEXT_FREE_BLKADDR(sbi, curseg); - f2fs_bug_on(sbi, curseg->next_blkoff >= sbi->blocks_per_seg); + f2fs_bug_on(sbi, curseg->next_blkoff >= BLKS_PER_SEG(sbi)); f2fs_wait_discard_bio(sbi, *new_blkaddr); @@ -3470,25 +3550,35 @@ void f2fs_allocate_data_block(struct f2fs_sb_info *sbi, struct page *page, * since SSR needs latest valid block information. */ update_sit_entry(sbi, *new_blkaddr, 1); - if (GET_SEGNO(sbi, old_blkaddr) != NULL_SEGNO) - update_sit_entry(sbi, old_blkaddr, -1); + update_sit_entry(sbi, old_blkaddr, -1); /* * If the current segment is full, flush it out and replace it with a * new segment. */ if (segment_full) { + if (type == CURSEG_COLD_DATA_PINNED && + !((curseg->segno + 1) % sbi->segs_per_sec)) { + reset_curseg_fields(curseg); + goto skip_new_segment; + } + if (from_gc) { - get_atssr_segment(sbi, type, se->type, + ret = get_atssr_segment(sbi, type, se->type, AT_SSR, se->mtime); } else { if (need_new_seg(sbi, type)) - new_curseg(sbi, type, false); + ret = new_curseg(sbi, type, false); else - change_curseg(sbi, type); + ret = change_curseg(sbi, type); stat_inc_seg_type(sbi, curseg); } + + if (ret) + goto out_err; } + +skip_new_segment: /* * segment dirty status should be updated after segment allocation, * so we just need to update status only one time after previous @@ -3497,12 +3587,12 @@ void f2fs_allocate_data_block(struct f2fs_sb_info *sbi, struct page *page, locate_dirty_segment(sbi, GET_SEGNO(sbi, old_blkaddr)); locate_dirty_segment(sbi, GET_SEGNO(sbi, *new_blkaddr)); - if (IS_DATASEG(type)) + if (IS_DATASEG(curseg->seg_type)) atomic64_inc(&sbi->allocated_data_blocks); up_write(&sit_i->sentry_lock); - if (page && IS_NODESEG(type)) { + if (page && IS_NODESEG(curseg->seg_type)) { fill_node_footer_blkaddr(page, NEXT_FREE_BLKADDR(sbi, curseg)); f2fs_inode_chksum_set(sbi, page); @@ -3511,9 +3601,6 @@ void f2fs_allocate_data_block(struct f2fs_sb_info *sbi, struct page *page, if (fio) { struct f2fs_bio_info *io; - if (F2FS_IO_ALIGNED(sbi)) - fio->retry = 0; - INIT_LIST_HEAD(&fio->list); fio->in_list = 1; io = sbi->write_io[fio->type] + fio->temp; @@ -3523,8 +3610,15 @@ void f2fs_allocate_data_block(struct f2fs_sb_info *sbi, struct page *page, } mutex_unlock(&curseg->curseg_mutex); - f2fs_up_read(&SM_I(sbi)->curseg_lock); + return 0; +out_err: + *new_blkaddr = NULL_ADDR; + up_write(&sit_i->sentry_lock); + mutex_unlock(&curseg->curseg_mutex); + f2fs_up_read(&SM_I(sbi)->curseg_lock); + return ret; + } void f2fs_update_device_state(struct f2fs_sb_info *sbi, nid_t ino, @@ -3561,21 +3655,25 @@ static void do_write_page(struct f2fs_summary *sum, struct f2fs_io_info *fio) if (keep_order) f2fs_down_read(&fio->sbi->io_order_lock); -reallocate: - f2fs_allocate_data_block(fio->sbi, fio->page, fio->old_blkaddr, - &fio->new_blkaddr, sum, type, fio); + + if (f2fs_allocate_data_block(fio->sbi, fio->page, fio->old_blkaddr, + &fio->new_blkaddr, sum, type, fio)) { + if (fscrypt_inode_uses_fs_layer_crypto(fio->page->mapping->host)) + fscrypt_finalize_bounce_page(&fio->encrypted_page); + if (PageWriteback(fio->page)) + end_page_writeback(fio->page); + if (f2fs_in_warm_node_list(fio->sbi, fio->page)) + f2fs_del_fsync_node_entry(fio->sbi, fio->page); + goto out; + } if (GET_SEGNO(fio->sbi, fio->old_blkaddr) != NULL_SEGNO) f2fs_invalidate_internal_cache(fio->sbi, fio->old_blkaddr); /* writeout dirty page into bdev */ f2fs_submit_page_write(fio); - if (fio->retry) { - fio->old_blkaddr = fio->new_blkaddr; - goto reallocate; - } f2fs_update_device_state(fio->sbi, fio->ino, fio->new_blkaddr, 1); - +out: if (keep_order) f2fs_up_read(&fio->sbi->io_order_lock); } @@ -3659,8 +3757,7 @@ int f2fs_inplace_write_data(struct f2fs_io_info *fio) } if (fio->post_read) - invalidate_mapping_pages(META_MAPPING(sbi), - fio->new_blkaddr, fio->new_blkaddr); + f2fs_truncate_meta_inode_pages(sbi, fio->new_blkaddr, 1); stat_inc_inplace_blocks(fio->sbi); @@ -3749,7 +3846,8 @@ void f2fs_do_replace_block(struct f2fs_sb_info *sbi, struct f2fs_summary *sum, /* change the current segment */ if (segno != curseg->segno) { curseg->next_segno = segno; - change_curseg(sbi, type); + if (change_curseg(sbi, type)) + goto out_unlock; } curseg->next_blkoff = GET_BLKOFF_FROM_SEG0(sbi, new_blkaddr); @@ -3775,12 +3873,14 @@ void f2fs_do_replace_block(struct f2fs_sb_info *sbi, struct f2fs_summary *sum, if (recover_curseg) { if (old_cursegno != curseg->segno) { curseg->next_segno = old_cursegno; - change_curseg(sbi, type); + if (change_curseg(sbi, type)) + goto out_unlock; } curseg->next_blkoff = old_blkoff; curseg->alloc_type = old_alloc_type; } +out_unlock: up_write(&sit_i->sentry_lock); mutex_unlock(&curseg->curseg_mutex); f2fs_up_write(&SM_I(sbi)->curseg_lock); @@ -3850,7 +3950,7 @@ void f2fs_wait_on_block_writeback_range(struct inode *inode, block_t blkaddr, for (i = 0; i < len; i++) f2fs_wait_on_block_writeback(inode, blkaddr + i); - invalidate_mapping_pages(META_MAPPING(sbi), blkaddr, blkaddr + len - 1); + f2fs_truncate_meta_inode_pages(sbi, blkaddr, len); } static int read_compacted_summaries(struct f2fs_sb_info *sbi) @@ -3892,7 +3992,7 @@ static int read_compacted_summaries(struct f2fs_sb_info *sbi) seg_i->next_blkoff = blk_off; if (seg_i->alloc_type == SSR) - blk_off = sbi->blocks_per_seg; + blk_off = BLKS_PER_SEG(sbi); for (j = 0; j < blk_off; j++) { struct f2fs_summary *s; @@ -3960,7 +4060,7 @@ static int read_normal_summaries(struct f2fs_sb_info *sbi, int type) struct f2fs_summary *ns = &sum->entries[0]; int i; - for (i = 0; i < sbi->blocks_per_seg; i++, ns++) { + for (i = 0; i < BLKS_PER_SEG(sbi); i++, ns++) { ns->version = 0; ns->ofs_in_node = 0; } @@ -4466,7 +4566,7 @@ static int build_sit_info(struct f2fs_sb_info *sbi) #endif sit_i->sit_base_addr = le32_to_cpu(raw_super->sit_blkaddr); - sit_i->sit_blocks = sit_segs << sbi->log_blocks_per_seg; + sit_i->sit_blocks = SEGS_TO_BLKS(sbi, sit_segs); sit_i->written_valid_blocks = 0; sit_i->bitmap_size = sit_bitmap_size; sit_i->dirty_sentries = 0; @@ -4539,9 +4639,7 @@ static int build_curseg(struct f2fs_sb_info *sbi) array[i].seg_type = CURSEG_COLD_DATA; else if (i == CURSEG_ALL_DATA_ATGC) array[i].seg_type = CURSEG_COLD_DATA; - array[i].segno = NULL_SEGNO; - array[i].next_blkoff = 0; - array[i].inited = false; + reset_curseg_fields(&array[i]); } return restore_curseg_summaries(sbi); } @@ -4593,21 +4691,20 @@ static int build_sit_entries(struct f2fs_sb_info *sbi) sit_valid_blocks[SE_PAGETYPE(se)] += se->valid_blocks; - if (f2fs_block_unit_discard(sbi)) { - /* build discard map only one time */ - if (is_set_ckpt_flags(sbi, CP_TRIMMED_FLAG)) { - memset(se->discard_map, 0xff, + if (!f2fs_block_unit_discard(sbi)) + goto init_discard_map_done; + + /* build discard map only one time */ + if (is_set_ckpt_flags(sbi, CP_TRIMMED_FLAG)) { + memset(se->discard_map, 0xff, SIT_VBLOCK_MAP_SIZE); - } else { - memcpy(se->discard_map, - se->cur_valid_map, + goto init_discard_map_done; + } + memcpy(se->discard_map, se->cur_valid_map, SIT_VBLOCK_MAP_SIZE); - sbi->discard_blks += - sbi->blocks_per_seg - + sbi->discard_blks += BLKS_PER_SEG(sbi) - se->valid_blocks; - } - } - +init_discard_map_done: if (__is_large_section(sbi)) get_sec_entry(sbi, start)->valid_blocks += se->valid_blocks; @@ -4747,7 +4844,7 @@ static void init_dirty_segmap(struct f2fs_sb_info *sbi) return; mutex_lock(&dirty_i->seglist_lock); - for (segno = 0; segno < MAIN_SEGS(sbi); segno += sbi->segs_per_sec) { + for (segno = 0; segno < MAIN_SEGS(sbi); segno += SEGS_PER_SEC(sbi)) { valid_blocks = get_valid_blocks(sbi, segno, true); secno = GET_SEC_FROM_SEG(sbi, segno); @@ -4846,7 +4943,7 @@ static int sanity_check_curseg(struct f2fs_sb_info *sbi) if (curseg->alloc_type == SSR) continue; - for (blkofs += 1; blkofs < sbi->blocks_per_seg; blkofs++) { + for (blkofs += 1; blkofs < BLKS_PER_SEG(sbi); blkofs++) { if (!f2fs_test_bit(blkofs, se->cur_valid_map)) continue; out: @@ -4862,6 +4959,16 @@ out: } #ifdef CONFIG_BLK_DEV_ZONED +static const char *f2fs_zone_status[BLK_ZONE_COND_OFFLINE + 1] = { + [BLK_ZONE_COND_NOT_WP] = "NOT_WP", + [BLK_ZONE_COND_EMPTY] = "EMPTY", + [BLK_ZONE_COND_IMP_OPEN] = "IMPLICIT_OPEN", + [BLK_ZONE_COND_EXP_OPEN] = "EXPLICIT_OPEN", + [BLK_ZONE_COND_CLOSED] = "CLOSED", + [BLK_ZONE_COND_READONLY] = "READONLY", + [BLK_ZONE_COND_FULL] = "FULL", + [BLK_ZONE_COND_OFFLINE] = "OFFLINE", +}; static int check_zone_write_pointer(struct f2fs_sb_info *sbi, struct f2fs_dev_info *fdev, @@ -4883,14 +4990,19 @@ static int check_zone_write_pointer(struct f2fs_sb_info *sbi, * Skip check of zones cursegs point to, since * fix_curseg_write_pointer() checks them. */ - if (zone_segno >= MAIN_SEGS(sbi) || - IS_CURSEC(sbi, GET_SEC_FROM_SEG(sbi, zone_segno))) + if (zone_segno >= MAIN_SEGS(sbi)) return 0; /* * Get # of valid block of the zone. */ valid_block_cnt = get_valid_blocks(sbi, zone_segno, true); + if (IS_CURSEC(sbi, GET_SEC_FROM_SEG(sbi, zone_segno))) { + f2fs_notice(sbi, "Open zones: valid block[0x%x,0x%x] cond[%s]", + zone_segno, valid_block_cnt, + f2fs_zone_status[zone->cond]); + return 0; + } if ((!valid_block_cnt && zone->cond == BLK_ZONE_COND_EMPTY) || (valid_block_cnt && zone->cond == BLK_ZONE_COND_FULL)) @@ -4898,8 +5010,8 @@ static int check_zone_write_pointer(struct f2fs_sb_info *sbi, if (!valid_block_cnt) { f2fs_notice(sbi, "Zone without valid block has non-zero write " - "pointer. Reset the write pointer: cond[0x%x]", - zone->cond); + "pointer. Reset the write pointer: cond[%s]", + f2fs_zone_status[zone->cond]); ret = __f2fs_issue_discard_zone(sbi, fdev->bdev, zone_block, zone->len >> log_sectors_per_block); if (ret) @@ -4916,8 +5028,8 @@ static int check_zone_write_pointer(struct f2fs_sb_info *sbi, * selected for write operation until it get discarded. */ f2fs_notice(sbi, "Valid blocks are not aligned with write " - "pointer: valid block[0x%x,0x%x] cond[0x%x]", - zone_segno, valid_block_cnt, zone->cond); + "pointer: valid block[0x%x,0x%x] cond[%s]", + zone_segno, valid_block_cnt, f2fs_zone_status[zone->cond]); nofs_flags = memalloc_nofs_save(); ret = blkdev_zone_mgmt(fdev->bdev, REQ_OP_ZONE_FINISH, @@ -5128,7 +5240,7 @@ static inline unsigned int f2fs_usable_zone_blks_in_seg( unsigned int secno; if (!sbi->unusable_blocks_per_sec) - return sbi->blocks_per_seg; + return BLKS_PER_SEG(sbi); secno = GET_SEC_FROM_SEG(sbi, segno); seg_start = START_BLOCK(sbi, segno); @@ -5143,10 +5255,10 @@ static inline unsigned int f2fs_usable_zone_blks_in_seg( */ if (seg_start >= sec_cap_blkaddr) return 0; - if (seg_start + sbi->blocks_per_seg > sec_cap_blkaddr) + if (seg_start + BLKS_PER_SEG(sbi) > sec_cap_blkaddr) return sec_cap_blkaddr - seg_start; - return sbi->blocks_per_seg; + return BLKS_PER_SEG(sbi); } #else int f2fs_fix_curseg_write_pointer(struct f2fs_sb_info *sbi) @@ -5172,7 +5284,7 @@ unsigned int f2fs_usable_blks_in_seg(struct f2fs_sb_info *sbi, if (f2fs_sb_has_blkzoned(sbi)) return f2fs_usable_zone_blks_in_seg(sbi, segno); - return sbi->blocks_per_seg; + return BLKS_PER_SEG(sbi); } unsigned int f2fs_usable_segs_in_sec(struct f2fs_sb_info *sbi, @@ -5181,7 +5293,7 @@ unsigned int f2fs_usable_segs_in_sec(struct f2fs_sb_info *sbi, if (f2fs_sb_has_blkzoned(sbi)) return CAP_SEGS_PER_SEC(sbi); - return sbi->segs_per_sec; + return SEGS_PER_SEC(sbi); } /* @@ -5196,14 +5308,14 @@ static void init_min_max_mtime(struct f2fs_sb_info *sbi) sit_i->min_mtime = ULLONG_MAX; - for (segno = 0; segno < MAIN_SEGS(sbi); segno += sbi->segs_per_sec) { + for (segno = 0; segno < MAIN_SEGS(sbi); segno += SEGS_PER_SEC(sbi)) { unsigned int i; unsigned long long mtime = 0; - for (i = 0; i < sbi->segs_per_sec; i++) + for (i = 0; i < SEGS_PER_SEC(sbi); i++) mtime += get_seg_entry(sbi, segno + i)->mtime; - mtime = div_u64(mtime, sbi->segs_per_sec); + mtime = div_u64(mtime, SEGS_PER_SEC(sbi)); if (sit_i->min_mtime > mtime) sit_i->min_mtime = mtime; @@ -5242,7 +5354,7 @@ int f2fs_build_segment_manager(struct f2fs_sb_info *sbi) sm_info->ipu_policy = BIT(F2FS_IPU_FSYNC); sm_info->min_ipu_util = DEF_MIN_IPU_UTIL; sm_info->min_fsync_blocks = DEF_MIN_FSYNC_BLOCKS; - sm_info->min_seq_blocks = sbi->blocks_per_seg; + sm_info->min_seq_blocks = BLKS_PER_SEG(sbi); sm_info->min_hot_blocks = DEF_MIN_HOT_BLOCKS; sm_info->min_ssr_sections = reserved_sections(sbi); diff --git a/fs/f2fs/segment.h b/fs/f2fs/segment.h index 8129be788bd5..e1c0f418aa11 100644 --- a/fs/f2fs/segment.h +++ b/fs/f2fs/segment.h @@ -48,21 +48,21 @@ static inline void sanity_check_seg_type(struct f2fs_sb_info *sbi, #define IS_CURSEC(sbi, secno) \ (((secno) == CURSEG_I(sbi, CURSEG_HOT_DATA)->segno / \ - (sbi)->segs_per_sec) || \ + SEGS_PER_SEC(sbi)) || \ ((secno) == CURSEG_I(sbi, CURSEG_WARM_DATA)->segno / \ - (sbi)->segs_per_sec) || \ + SEGS_PER_SEC(sbi)) || \ ((secno) == CURSEG_I(sbi, CURSEG_COLD_DATA)->segno / \ - (sbi)->segs_per_sec) || \ + SEGS_PER_SEC(sbi)) || \ ((secno) == CURSEG_I(sbi, CURSEG_HOT_NODE)->segno / \ - (sbi)->segs_per_sec) || \ + SEGS_PER_SEC(sbi)) || \ ((secno) == CURSEG_I(sbi, CURSEG_WARM_NODE)->segno / \ - (sbi)->segs_per_sec) || \ + SEGS_PER_SEC(sbi)) || \ ((secno) == CURSEG_I(sbi, CURSEG_COLD_NODE)->segno / \ - (sbi)->segs_per_sec) || \ + SEGS_PER_SEC(sbi)) || \ ((secno) == CURSEG_I(sbi, CURSEG_COLD_DATA_PINNED)->segno / \ - (sbi)->segs_per_sec) || \ + SEGS_PER_SEC(sbi)) || \ ((secno) == CURSEG_I(sbi, CURSEG_ALL_DATA_ATGC)->segno / \ - (sbi)->segs_per_sec)) + SEGS_PER_SEC(sbi))) #define MAIN_BLKADDR(sbi) \ (SM_I(sbi) ? SM_I(sbi)->main_blkaddr : \ @@ -77,40 +77,37 @@ static inline void sanity_check_seg_type(struct f2fs_sb_info *sbi, #define TOTAL_SEGS(sbi) \ (SM_I(sbi) ? SM_I(sbi)->segment_count : \ le32_to_cpu(F2FS_RAW_SUPER(sbi)->segment_count)) -#define TOTAL_BLKS(sbi) (TOTAL_SEGS(sbi) << (sbi)->log_blocks_per_seg) +#define TOTAL_BLKS(sbi) (SEGS_TO_BLKS(sbi, TOTAL_SEGS(sbi))) #define MAX_BLKADDR(sbi) (SEG0_BLKADDR(sbi) + TOTAL_BLKS(sbi)) #define SEGMENT_SIZE(sbi) (1ULL << ((sbi)->log_blocksize + \ (sbi)->log_blocks_per_seg)) #define START_BLOCK(sbi, segno) (SEG0_BLKADDR(sbi) + \ - (GET_R2L_SEGNO(FREE_I(sbi), segno) << (sbi)->log_blocks_per_seg)) + (SEGS_TO_BLKS(sbi, GET_R2L_SEGNO(FREE_I(sbi), segno)))) #define NEXT_FREE_BLKADDR(sbi, curseg) \ (START_BLOCK(sbi, (curseg)->segno) + (curseg)->next_blkoff) #define GET_SEGOFF_FROM_SEG0(sbi, blk_addr) ((blk_addr) - SEG0_BLKADDR(sbi)) #define GET_SEGNO_FROM_SEG0(sbi, blk_addr) \ - (GET_SEGOFF_FROM_SEG0(sbi, blk_addr) >> (sbi)->log_blocks_per_seg) + (BLKS_TO_SEGS(sbi, GET_SEGOFF_FROM_SEG0(sbi, blk_addr))) #define GET_BLKOFF_FROM_SEG0(sbi, blk_addr) \ - (GET_SEGOFF_FROM_SEG0(sbi, blk_addr) & ((sbi)->blocks_per_seg - 1)) + (GET_SEGOFF_FROM_SEG0(sbi, blk_addr) & (BLKS_PER_SEG(sbi) - 1)) #define GET_SEGNO(sbi, blk_addr) \ ((!__is_valid_data_blkaddr(blk_addr)) ? \ NULL_SEGNO : GET_L2R_SEGNO(FREE_I(sbi), \ GET_SEGNO_FROM_SEG0(sbi, blk_addr))) -#define BLKS_PER_SEC(sbi) \ - ((sbi)->segs_per_sec * (sbi)->blocks_per_seg) #define CAP_BLKS_PER_SEC(sbi) \ - ((sbi)->segs_per_sec * (sbi)->blocks_per_seg - \ - (sbi)->unusable_blocks_per_sec) + (BLKS_PER_SEC(sbi) - (sbi)->unusable_blocks_per_sec) #define CAP_SEGS_PER_SEC(sbi) \ - ((sbi)->segs_per_sec - ((sbi)->unusable_blocks_per_sec >>\ - (sbi)->log_blocks_per_seg)) + (SEGS_PER_SEC(sbi) - \ + BLKS_TO_SEGS(sbi, (sbi)->unusable_blocks_per_sec)) #define GET_SEC_FROM_SEG(sbi, segno) \ - (((segno) == -1) ? -1 : (segno) / (sbi)->segs_per_sec) + (((segno) == -1) ? -1 : (segno) / SEGS_PER_SEC(sbi)) #define GET_SEG_FROM_SEC(sbi, secno) \ - ((secno) * (sbi)->segs_per_sec) + ((secno) * SEGS_PER_SEC(sbi)) #define GET_ZONE_FROM_SEC(sbi, secno) \ (((secno) == -1) ? -1 : (secno) / (sbi)->secs_per_zone) #define GET_ZONE_FROM_SEG(sbi, segno) \ @@ -139,16 +136,6 @@ static inline void sanity_check_seg_type(struct f2fs_sb_info *sbi, ((sectors) >> F2FS_LOG_SECTORS_PER_BLOCK) /* - * indicate a block allocation direction: RIGHT and LEFT. - * RIGHT means allocating new sections towards the end of volume. - * LEFT means the opposite direction. - */ -enum { - ALLOC_RIGHT = 0, - ALLOC_LEFT -}; - -/* * In the victim_sel_policy->alloc_mode, there are three block allocation modes. * LFS writes data sequentially with cleaning operations. * SSR (Slack Space Recycle) reuses obsolete space without cleaning operations. @@ -364,7 +351,7 @@ static inline unsigned int get_ckpt_valid_blocks(struct f2fs_sb_info *sbi, unsigned int blocks = 0; int i; - for (i = 0; i < sbi->segs_per_sec; i++, start_segno++) { + for (i = 0; i < SEGS_PER_SEC(sbi); i++, start_segno++) { struct seg_entry *se = get_seg_entry(sbi, start_segno); blocks += se->ckpt_valid_blocks; @@ -449,7 +436,7 @@ static inline void __set_free(struct f2fs_sb_info *sbi, unsigned int segno) free_i->free_segments++; next = find_next_bit(free_i->free_segmap, - start_segno + sbi->segs_per_sec, start_segno); + start_segno + SEGS_PER_SEC(sbi), start_segno); if (next >= start_segno + usable_segs) { clear_bit(secno, free_i->free_secmap); free_i->free_sections++; @@ -485,7 +472,7 @@ static inline void __set_test_and_free(struct f2fs_sb_info *sbi, if (!inmem && IS_CURSEC(sbi, secno)) goto skip_free; next = find_next_bit(free_i->free_segmap, - start_segno + sbi->segs_per_sec, start_segno); + start_segno + SEGS_PER_SEC(sbi), start_segno); if (next >= start_segno + usable_segs) { if (test_and_clear_bit(secno, free_i->free_secmap)) free_i->free_sections++; @@ -573,23 +560,22 @@ static inline bool has_curseg_enough_space(struct f2fs_sb_info *sbi, unsigned int node_blocks, unsigned int dent_blocks) { - unsigned int segno, left_blocks; + unsigned segno, left_blocks; int i; - /* check current node segment */ + /* check current node sections in the worst case. */ for (i = CURSEG_HOT_NODE; i <= CURSEG_COLD_NODE; i++) { segno = CURSEG_I(sbi, i)->segno; - left_blocks = f2fs_usable_blks_in_seg(sbi, segno) - - get_seg_entry(sbi, segno)->ckpt_valid_blocks; - + left_blocks = CAP_BLKS_PER_SEC(sbi) - + get_ckpt_valid_blocks(sbi, segno, true); if (node_blocks > left_blocks) return false; } - /* check current data segment */ + /* check current data section for dentry blocks. */ segno = CURSEG_I(sbi, CURSEG_HOT_DATA)->segno; - left_blocks = f2fs_usable_blks_in_seg(sbi, segno) - - get_seg_entry(sbi, segno)->ckpt_valid_blocks; + left_blocks = CAP_BLKS_PER_SEC(sbi) - + get_ckpt_valid_blocks(sbi, segno, true); if (dent_blocks > left_blocks) return false; return true; @@ -638,7 +624,7 @@ static inline bool has_not_enough_free_secs(struct f2fs_sb_info *sbi, if (free_secs > upper_secs) return false; - else if (free_secs <= lower_secs) + if (free_secs <= lower_secs) return true; return !curseg_space; } @@ -793,10 +779,10 @@ static inline int check_block_count(struct f2fs_sb_info *sbi, return -EFSCORRUPTED; } - if (usable_blks_per_seg < sbi->blocks_per_seg) + if (usable_blks_per_seg < BLKS_PER_SEG(sbi)) f2fs_bug_on(sbi, find_next_bit_le(&raw_sit->valid_map, - sbi->blocks_per_seg, - usable_blks_per_seg) != sbi->blocks_per_seg); + BLKS_PER_SEG(sbi), + usable_blks_per_seg) != BLKS_PER_SEG(sbi)); /* check segment usage, and check boundary of a given segment number */ if (unlikely(GET_SIT_VBLOCKS(raw_sit) > usable_blks_per_seg @@ -915,9 +901,9 @@ static inline int nr_pages_to_skip(struct f2fs_sb_info *sbi, int type) return 0; if (type == DATA) - return sbi->blocks_per_seg; + return BLKS_PER_SEG(sbi); else if (type == NODE) - return 8 * sbi->blocks_per_seg; + return SEGS_TO_BLKS(sbi, 8); else if (type == META) return 8 * BIO_MAX_VECS; else @@ -969,3 +955,13 @@ wake_up: dcc->discard_wake = true; wake_up_interruptible_all(&dcc->discard_wait_queue); } + +static inline unsigned int first_zoned_segno(struct f2fs_sb_info *sbi) +{ + int devi; + + for (devi = 0; devi < sbi->s_ndevs; devi++) + if (bdev_is_zoned(FDEV(devi).bdev)) + return GET_SEGNO(sbi, FDEV(devi).start_blk); + return 0; +} diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c index b880b746f226..a6867f26f141 100644 --- a/fs/f2fs/super.c +++ b/fs/f2fs/super.c @@ -44,24 +44,26 @@ static struct kmem_cache *f2fs_inode_cachep; #ifdef CONFIG_F2FS_FAULT_INJECTION const char *f2fs_fault_name[FAULT_MAX] = { - [FAULT_KMALLOC] = "kmalloc", - [FAULT_KVMALLOC] = "kvmalloc", - [FAULT_PAGE_ALLOC] = "page alloc", - [FAULT_PAGE_GET] = "page get", - [FAULT_ALLOC_NID] = "alloc nid", - [FAULT_ORPHAN] = "orphan", - [FAULT_BLOCK] = "no more block", - [FAULT_DIR_DEPTH] = "too big dir depth", - [FAULT_EVICT_INODE] = "evict_inode fail", - [FAULT_TRUNCATE] = "truncate fail", - [FAULT_READ_IO] = "read IO error", - [FAULT_CHECKPOINT] = "checkpoint error", - [FAULT_DISCARD] = "discard error", - [FAULT_WRITE_IO] = "write IO error", - [FAULT_SLAB_ALLOC] = "slab alloc", - [FAULT_DQUOT_INIT] = "dquot initialize", - [FAULT_LOCK_OP] = "lock_op", - [FAULT_BLKADDR] = "invalid blkaddr", + [FAULT_KMALLOC] = "kmalloc", + [FAULT_KVMALLOC] = "kvmalloc", + [FAULT_PAGE_ALLOC] = "page alloc", + [FAULT_PAGE_GET] = "page get", + [FAULT_ALLOC_NID] = "alloc nid", + [FAULT_ORPHAN] = "orphan", + [FAULT_BLOCK] = "no more block", + [FAULT_DIR_DEPTH] = "too big dir depth", + [FAULT_EVICT_INODE] = "evict_inode fail", + [FAULT_TRUNCATE] = "truncate fail", + [FAULT_READ_IO] = "read IO error", + [FAULT_CHECKPOINT] = "checkpoint error", + [FAULT_DISCARD] = "discard error", + [FAULT_WRITE_IO] = "write IO error", + [FAULT_SLAB_ALLOC] = "slab alloc", + [FAULT_DQUOT_INIT] = "dquot initialize", + [FAULT_LOCK_OP] = "lock_op", + [FAULT_BLKADDR_VALIDITY] = "invalid blkaddr", + [FAULT_BLKADDR_CONSISTENCE] = "inconsistent blkaddr", + [FAULT_NO_SEGMENT] = "no free segment", }; void f2fs_build_fault_attr(struct f2fs_sb_info *sbi, unsigned int rate, @@ -137,7 +139,6 @@ enum { Opt_resgid, Opt_resuid, Opt_mode, - Opt_io_size_bits, Opt_fault_injection, Opt_fault_type, Opt_lazytime, @@ -216,7 +217,6 @@ static match_table_t f2fs_tokens = { {Opt_resgid, "resgid=%u"}, {Opt_resuid, "resuid=%u"}, {Opt_mode, "mode=%s"}, - {Opt_io_size_bits, "io_bits=%u"}, {Opt_fault_injection, "fault_injection=%u"}, {Opt_fault_type, "fault_type=%u"}, {Opt_lazytime, "lazytime"}, @@ -263,7 +263,8 @@ static match_table_t f2fs_tokens = { {Opt_err, NULL}, }; -void f2fs_printk(struct f2fs_sb_info *sbi, const char *fmt, ...) +void f2fs_printk(struct f2fs_sb_info *sbi, bool limit_rate, + const char *fmt, ...) { struct va_format vaf; va_list args; @@ -274,8 +275,12 @@ void f2fs_printk(struct f2fs_sb_info *sbi, const char *fmt, ...) level = printk_get_level(fmt); vaf.fmt = printk_skip_level(fmt); vaf.va = &args; - printk("%c%cF2FS-fs (%s): %pV\n", - KERN_SOH_ASCII, level, sbi->sb->s_id, &vaf); + if (limit_rate) + printk_ratelimited("%c%cF2FS-fs (%s): %pV\n", + KERN_SOH_ASCII, level, sbi->sb->s_id, &vaf); + else + printk("%c%cF2FS-fs (%s): %pV\n", + KERN_SOH_ASCII, level, sbi->sb->s_id, &vaf); va_end(args); } @@ -343,46 +348,6 @@ static inline void limit_reserve_root(struct f2fs_sb_info *sbi) F2FS_OPTION(sbi).s_resgid)); } -static inline int adjust_reserved_segment(struct f2fs_sb_info *sbi) -{ - unsigned int sec_blks = sbi->blocks_per_seg * sbi->segs_per_sec; - unsigned int avg_vblocks; - unsigned int wanted_reserved_segments; - block_t avail_user_block_count; - - if (!F2FS_IO_ALIGNED(sbi)) - return 0; - - /* average valid block count in section in worst case */ - avg_vblocks = sec_blks / F2FS_IO_SIZE(sbi); - - /* - * we need enough free space when migrating one section in worst case - */ - wanted_reserved_segments = (F2FS_IO_SIZE(sbi) / avg_vblocks) * - reserved_segments(sbi); - wanted_reserved_segments -= reserved_segments(sbi); - - avail_user_block_count = sbi->user_block_count - - sbi->current_reserved_blocks - - F2FS_OPTION(sbi).root_reserved_blocks; - - if (wanted_reserved_segments * sbi->blocks_per_seg > - avail_user_block_count) { - f2fs_err(sbi, "IO align feature can't grab additional reserved segment: %u, available segments: %u", - wanted_reserved_segments, - avail_user_block_count >> sbi->log_blocks_per_seg); - return -ENOSPC; - } - - SM_I(sbi)->additional_reserved_segments = wanted_reserved_segments; - - f2fs_info(sbi, "IO align feature needs additional reserved segment: %u", - wanted_reserved_segments); - - return 0; -} - static inline void adjust_unusable_cap_perc(struct f2fs_sb_info *sbi) { if (!F2FS_OPTION(sbi).unusable_cap_perc) @@ -663,7 +628,7 @@ static int f2fs_set_lz4hc_level(struct f2fs_sb_info *sbi, const char *str) #ifdef CONFIG_F2FS_FS_ZSTD static int f2fs_set_zstd_level(struct f2fs_sb_info *sbi, const char *str) { - unsigned int level; + int level; int len = 4; if (strlen(str) == len) { @@ -677,9 +642,15 @@ static int f2fs_set_zstd_level(struct f2fs_sb_info *sbi, const char *str) f2fs_info(sbi, "wrong format, e.g. <alg_name>:<compr_level>"); return -EINVAL; } - if (kstrtouint(str + 1, 10, &level)) + if (kstrtoint(str + 1, 10, &level)) return -EINVAL; + /* f2fs does not support negative compress level now */ + if (level < 0) { + f2fs_info(sbi, "do not support negative compress level: %d", level); + return -ERANGE; + } + if (!f2fs_is_compress_level_valid(COMPRESS_ZSTD, level)) { f2fs_info(sbi, "invalid zstd compress level: %d", level); return -EINVAL; @@ -763,10 +734,8 @@ static int parse_options(struct super_block *sb, char *options, bool is_remount) clear_opt(sbi, DISCARD); break; case Opt_noheap: - set_opt(sbi, NOHEAP); - break; case Opt_heap: - clear_opt(sbi, NOHEAP); + f2fs_warn(sbi, "heap/no_heap options were deprecated"); break; #ifdef CONFIG_F2FS_FS_XATTR case Opt_user_xattr: @@ -913,16 +882,6 @@ static int parse_options(struct super_block *sb, char *options, bool is_remount) } kfree(name); break; - case Opt_io_size_bits: - if (args->from && match_int(args, &arg)) - return -EINVAL; - if (arg <= 0 || arg > __ilog2_u32(BIO_MAX_VECS)) { - f2fs_warn(sbi, "Not support %ld, larger than %d", - BIT(arg), BIO_MAX_VECS); - return -EINVAL; - } - F2FS_OPTION(sbi).write_io_size_bits = arg; - break; #ifdef CONFIG_F2FS_FAULT_INJECTION case Opt_fault_injection: if (args->from && match_int(args, &arg)) @@ -1392,12 +1351,6 @@ default_check: } #endif - if (F2FS_IO_SIZE_BITS(sbi) && !f2fs_lfs_mode(sbi)) { - f2fs_err(sbi, "Should set mode=lfs with %luKB-sized IO", - F2FS_IO_SIZE_KB(sbi)); - return -EINVAL; - } - if (test_opt(sbi, INLINE_XATTR_SIZE)) { int min_size, max_size; @@ -1718,7 +1671,6 @@ static void f2fs_put_super(struct super_block *sb) f2fs_destroy_page_array_cache(sbi); f2fs_destroy_xattr_caches(sbi); - mempool_destroy(sbi->write_io_dummy); #ifdef CONFIG_QUOTA for (i = 0; i < MAXQUOTAS; i++) kfree(F2FS_OPTION(sbi).s_qf_names[i]); @@ -2009,10 +1961,6 @@ static int f2fs_show_options(struct seq_file *seq, struct dentry *root) } else { seq_puts(seq, ",nodiscard"); } - if (test_opt(sbi, NOHEAP)) - seq_puts(seq, ",no_heap"); - else - seq_puts(seq, ",heap"); #ifdef CONFIG_F2FS_FS_XATTR if (test_opt(sbi, XATTR_USER)) seq_puts(seq, ",user_xattr"); @@ -2078,9 +2026,6 @@ static int f2fs_show_options(struct seq_file *seq, struct dentry *root) F2FS_OPTION(sbi).s_resuid), from_kgid_munged(&init_user_ns, F2FS_OPTION(sbi).s_resgid)); - if (F2FS_IO_SIZE_BITS(sbi)) - seq_printf(seq, ",io_bits=%u", - F2FS_OPTION(sbi).write_io_size_bits); #ifdef CONFIG_F2FS_FAULT_INJECTION if (test_opt(sbi, FAULT_INJECTION)) { seq_printf(seq, ",fault_injection=%u", @@ -2192,7 +2137,6 @@ static void default_options(struct f2fs_sb_info *sbi, bool remount) set_opt(sbi, INLINE_XATTR); set_opt(sbi, INLINE_DATA); set_opt(sbi, INLINE_DENTRY); - set_opt(sbi, NOHEAP); set_opt(sbi, MERGE_CHECKPOINT); F2FS_OPTION(sbi).unusable_cap = 0; sbi->sb->s_flags |= SB_LAZYTIME; @@ -2247,6 +2191,7 @@ static int f2fs_disable_checkpoint(struct f2fs_sb_info *sbi) .init_gc_type = FG_GC, .should_migrate_blocks = false, .err_gc_skipped = true, + .no_bg_gc = true, .nr_free_secs = 1 }; f2fs_down_write(&sbi->gc_lock); @@ -2332,7 +2277,6 @@ static int f2fs_remount(struct super_block *sb, int *flags, char *data) bool no_read_extent_cache = !test_opt(sbi, READ_EXTENT_CACHE); bool no_age_extent_cache = !test_opt(sbi, AGE_EXTENT_CACHE); bool enable_checkpoint = !test_opt(sbi, DISABLE_CHECKPOINT); - bool no_io_align = !F2FS_IO_ALIGNED(sbi); bool no_atgc = !test_opt(sbi, ATGC); bool no_discard = !test_opt(sbi, DISCARD); bool no_compress_cache = !test_opt(sbi, COMPRESS_CACHE); @@ -2440,12 +2384,6 @@ static int f2fs_remount(struct super_block *sb, int *flags, char *data) goto restore_opts; } - if (no_io_align == !!F2FS_IO_ALIGNED(sbi)) { - err = -EINVAL; - f2fs_warn(sbi, "switch io_bits option is not allowed"); - goto restore_opts; - } - if (no_compress_cache == !!test_opt(sbi, COMPRESS_CACHE)) { err = -EINVAL; f2fs_warn(sbi, "switch compress_cache option is not allowed"); @@ -2768,7 +2706,7 @@ int f2fs_dquot_initialize(struct inode *inode) return dquot_initialize(inode); } -static struct dquot **f2fs_get_dquots(struct inode *inode) +static struct dquot __rcu **f2fs_get_dquots(struct inode *inode) { return F2FS_I(inode)->i_dquot; } @@ -3706,7 +3644,7 @@ int f2fs_sanity_check_ckpt(struct f2fs_sb_info *sbi) } main_segs = le32_to_cpu(raw_super->segment_count_main); - blocks_per_seg = sbi->blocks_per_seg; + blocks_per_seg = BLKS_PER_SEG(sbi); for (i = 0; i < NR_CURSEG_NODE_TYPE; i++) { if (le32_to_cpu(ckpt->cur_node_segno[i]) >= main_segs || @@ -3818,9 +3756,9 @@ static void init_sb_info(struct f2fs_sb_info *sbi) sbi->segs_per_sec = le32_to_cpu(raw_super->segs_per_sec); sbi->secs_per_zone = le32_to_cpu(raw_super->secs_per_zone); sbi->total_sections = le32_to_cpu(raw_super->section_count); - sbi->total_node_count = - (le32_to_cpu(raw_super->segment_count_nat) / 2) - * sbi->blocks_per_seg * NAT_ENTRY_PER_BLOCK; + sbi->total_node_count = SEGS_TO_BLKS(sbi, + ((le32_to_cpu(raw_super->segment_count_nat) / 2) * + NAT_ENTRY_PER_BLOCK)); F2FS_ROOT_INO(sbi) = le32_to_cpu(raw_super->root_ino); F2FS_NODE_INO(sbi) = le32_to_cpu(raw_super->node_ino); F2FS_META_INO(sbi) = le32_to_cpu(raw_super->meta_ino); @@ -3829,7 +3767,7 @@ static void init_sb_info(struct f2fs_sb_info *sbi) sbi->next_victim_seg[BG_GC] = NULL_SEGNO; sbi->next_victim_seg[FG_GC] = NULL_SEGNO; sbi->max_victim_search = DEF_MAX_VICTIM_SEARCH; - sbi->migration_granularity = sbi->segs_per_sec; + sbi->migration_granularity = SEGS_PER_SEC(sbi); sbi->seq_file_ra_mul = MIN_RA_MUL; sbi->max_fragment_chunk = DEF_FRAGMENT_SIZE; sbi->max_fragment_hole = DEF_FRAGMENT_SIZE; @@ -3930,11 +3868,6 @@ static int init_blkz_info(struct f2fs_sb_info *sbi, int devi) return 0; zone_sectors = bdev_zone_sectors(bdev); - if (!is_power_of_2(zone_sectors)) { - f2fs_err(sbi, "F2FS does not support non power of 2 zone sizes\n"); - return -EINVAL; - } - if (sbi->blocks_per_blkz && sbi->blocks_per_blkz != SECTOR_TO_BLOCK(zone_sectors)) return -EINVAL; @@ -4090,7 +4023,9 @@ static void f2fs_record_stop_reason(struct f2fs_sb_info *sbi) f2fs_up_write(&sbi->sb_lock); if (err) - f2fs_err(sbi, "f2fs_commit_super fails to record err:%d", err); + f2fs_err_ratelimited(sbi, + "f2fs_commit_super fails to record stop_reason, err:%d", + err); } void f2fs_save_errors(struct f2fs_sb_info *sbi, unsigned char flag) @@ -4133,8 +4068,9 @@ static void f2fs_record_errors(struct f2fs_sb_info *sbi, unsigned char error) err = f2fs_commit_super(sbi, false); if (err) - f2fs_err(sbi, "f2fs_commit_super fails to record errors:%u, err:%d", - error, err); + f2fs_err_ratelimited(sbi, + "f2fs_commit_super fails to record errors:%u, err:%d", + error, err); out_unlock: f2fs_up_write(&sbi->sb_lock); } @@ -4259,14 +4195,14 @@ static int f2fs_scan_devices(struct f2fs_sb_info *sbi) if (i == 0) { FDEV(i).start_blk = 0; FDEV(i).end_blk = FDEV(i).start_blk + - (FDEV(i).total_segments << - sbi->log_blocks_per_seg) - 1 + - le32_to_cpu(raw_super->segment0_blkaddr); + SEGS_TO_BLKS(sbi, + FDEV(i).total_segments) - 1 + + le32_to_cpu(raw_super->segment0_blkaddr); } else { FDEV(i).start_blk = FDEV(i - 1).end_blk + 1; FDEV(i).end_blk = FDEV(i).start_blk + - (FDEV(i).total_segments << - sbi->log_blocks_per_seg) - 1; + SEGS_TO_BLKS(sbi, + FDEV(i).total_segments) - 1; FDEV(i).bdev_file = bdev_file_open_by_path( FDEV(i).path, mode, sbi->sb, NULL); } @@ -4305,8 +4241,6 @@ static int f2fs_scan_devices(struct f2fs_sb_info *sbi) FDEV(i).total_segments, FDEV(i).start_blk, FDEV(i).end_blk); } - f2fs_info(sbi, - "IO Block Size: %8ld KB", F2FS_IO_SIZE_KB(sbi)); return 0; } @@ -4519,19 +4453,10 @@ try_onemore: if (err) goto free_iostat; - if (F2FS_IO_ALIGNED(sbi)) { - sbi->write_io_dummy = - mempool_create_page_pool(2 * (F2FS_IO_SIZE(sbi) - 1), 0); - if (!sbi->write_io_dummy) { - err = -ENOMEM; - goto free_percpu; - } - } - /* init per sbi slab cache */ err = f2fs_init_xattr_caches(sbi); if (err) - goto free_io_dummy; + goto free_percpu; err = f2fs_init_page_array_cache(sbi); if (err) goto free_xattr_cache; @@ -4619,10 +4544,6 @@ try_onemore: goto free_nm; } - err = adjust_reserved_segment(sbi); - if (err) - goto free_nm; - /* For write statistics */ sbi->sectors_written_start = f2fs_get_sectors_written(sbi); @@ -4749,13 +4670,20 @@ reset_checkpoint: * If the f2fs is not readonly and fsync data recovery succeeds, * check zoned block devices' write pointer consistency. */ - if (!err && !f2fs_readonly(sb) && f2fs_sb_has_blkzoned(sbi)) { - err = f2fs_check_write_pointer(sbi); - if (err) - goto free_meta; + if (f2fs_sb_has_blkzoned(sbi) && !f2fs_readonly(sb)) { + int err2; + + f2fs_notice(sbi, "Checking entire write pointers"); + err2 = f2fs_check_write_pointer(sbi); + if (err2) + err = err2; } + if (err) + goto free_meta; - f2fs_init_inmem_curseg(sbi); + err = f2fs_init_inmem_curseg(sbi); + if (err) + goto sync_free_meta; /* f2fs_recover_fsync_data() cleared this already */ clear_sbi_flag(sbi, SBI_POR_DOING); @@ -4854,8 +4782,6 @@ free_page_array_cache: f2fs_destroy_page_array_cache(sbi); free_xattr_cache: f2fs_destroy_xattr_caches(sbi); -free_io_dummy: - mempool_destroy(sbi->write_io_dummy); free_percpu: destroy_percpu_info(sbi); free_iostat: diff --git a/fs/f2fs/sysfs.c b/fs/f2fs/sysfs.c index a7ec55c7bb20..a568ce96cf56 100644 --- a/fs/f2fs/sysfs.c +++ b/fs/f2fs/sysfs.c @@ -493,8 +493,8 @@ out: spin_lock(&sbi->stat_lock); if (t > (unsigned long)(sbi->user_block_count - F2FS_OPTION(sbi).root_reserved_blocks - - sbi->blocks_per_seg * - SM_I(sbi)->additional_reserved_segments)) { + SEGS_TO_BLKS(sbi, + SM_I(sbi)->additional_reserved_segments))) { spin_unlock(&sbi->stat_lock); return -EINVAL; } @@ -551,7 +551,7 @@ out: } if (!strcmp(a->attr.name, "migration_granularity")) { - if (t == 0 || t > sbi->segs_per_sec) + if (t == 0 || t > SEGS_PER_SEC(sbi)) return -EINVAL; } @@ -1492,6 +1492,50 @@ static int __maybe_unused discard_plist_seq_show(struct seq_file *seq, return 0; } +static int __maybe_unused disk_map_seq_show(struct seq_file *seq, + void *offset) +{ + struct super_block *sb = seq->private; + struct f2fs_sb_info *sbi = F2FS_SB(sb); + int i; + + seq_printf(seq, "Address Layout : %5luB Block address (# of Segments)\n", + F2FS_BLKSIZE); + seq_printf(seq, " SB : %12s\n", "0/1024B"); + seq_printf(seq, " seg0_blkaddr : 0x%010x\n", SEG0_BLKADDR(sbi)); + seq_printf(seq, " Checkpoint : 0x%010x (%10d)\n", + le32_to_cpu(F2FS_RAW_SUPER(sbi)->cp_blkaddr), 2); + seq_printf(seq, " SIT : 0x%010x (%10d)\n", + SIT_I(sbi)->sit_base_addr, + le32_to_cpu(F2FS_RAW_SUPER(sbi)->segment_count_sit)); + seq_printf(seq, " NAT : 0x%010x (%10d)\n", + NM_I(sbi)->nat_blkaddr, + le32_to_cpu(F2FS_RAW_SUPER(sbi)->segment_count_nat)); + seq_printf(seq, " SSA : 0x%010x (%10d)\n", + SM_I(sbi)->ssa_blkaddr, + le32_to_cpu(F2FS_RAW_SUPER(sbi)->segment_count_ssa)); + seq_printf(seq, " Main : 0x%010x (%10d)\n", + SM_I(sbi)->main_blkaddr, + le32_to_cpu(F2FS_RAW_SUPER(sbi)->segment_count_main)); + seq_printf(seq, " # of Sections : %12d\n", + le32_to_cpu(F2FS_RAW_SUPER(sbi)->section_count)); + seq_printf(seq, " Segs/Sections : %12d\n", + SEGS_PER_SEC(sbi)); + seq_printf(seq, " Section size : %12d MB\n", + SEGS_PER_SEC(sbi) << 1); + + if (!f2fs_is_multi_device(sbi)) + return 0; + + seq_puts(seq, "\nDisk Map for multi devices:\n"); + for (i = 0; i < sbi->s_ndevs; i++) + seq_printf(seq, "Disk:%2d (zoned=%d): 0x%010x - 0x%010x on %s\n", + i, bdev_is_zoned(FDEV(i).bdev), + FDEV(i).start_blk, FDEV(i).end_blk, + FDEV(i).path); + return 0; +} + int __init f2fs_init_sysfs(void) { int ret; @@ -1573,6 +1617,8 @@ int f2fs_register_sysfs(struct f2fs_sb_info *sbi) victim_bits_seq_show, sb); proc_create_single_data("discard_plist_info", 0444, sbi->s_proc, discard_plist_seq_show, sb); + proc_create_single_data("disk_map", 0444, sbi->s_proc, + disk_map_seq_show, sb); return 0; put_feature_list_kobj: kobject_put(&sbi->s_feature_list_kobj); diff --git a/fs/f2fs/verity.c b/fs/f2fs/verity.c index 4fc95f353a7a..f7bb0c54502c 100644 --- a/fs/f2fs/verity.c +++ b/fs/f2fs/verity.c @@ -258,21 +258,23 @@ static struct page *f2fs_read_merkle_tree_page(struct inode *inode, pgoff_t index, unsigned long num_ra_pages) { - struct page *page; + struct folio *folio; index += f2fs_verity_metadata_pos(inode) >> PAGE_SHIFT; - page = find_get_page_flags(inode->i_mapping, index, FGP_ACCESSED); - if (!page || !PageUptodate(page)) { + folio = __filemap_get_folio(inode->i_mapping, index, FGP_ACCESSED, 0); + if (IS_ERR(folio) || !folio_test_uptodate(folio)) { DEFINE_READAHEAD(ractl, NULL, NULL, inode->i_mapping, index); - if (page) - put_page(page); + if (!IS_ERR(folio)) + folio_put(folio); else if (num_ra_pages > 1) page_cache_ra_unbounded(&ractl, num_ra_pages, 0); - page = read_mapping_page(inode->i_mapping, index, NULL); + folio = read_mapping_folio(inode->i_mapping, index, NULL); + if (IS_ERR(folio)) + return ERR_CAST(folio); } - return page; + return folio_file_page(folio, index); } static int f2fs_write_merkle_tree_block(struct inode *inode, const void *buf, diff --git a/fs/fat/nfs.c b/fs/fat/nfs.c index c52e63e10d35..509eea96a457 100644 --- a/fs/fat/nfs.c +++ b/fs/fat/nfs.c @@ -130,6 +130,12 @@ fat_encode_fh_nostale(struct inode *inode, __u32 *fh, int *lenp, fid->parent_i_gen = parent->i_generation; type = FILEID_FAT_WITH_PARENT; *lenp = FAT_FID_SIZE_WITH_PARENT; + } else { + /* + * We need to initialize this field because the fh is actually + * 12 bytes long + */ + fid->parent_i_pos_hi = 0; } return type; diff --git a/fs/fuse/Kconfig b/fs/fuse/Kconfig index 038ed0b9aaa5..8674dbfbe59d 100644 --- a/fs/fuse/Kconfig +++ b/fs/fuse/Kconfig @@ -52,3 +52,14 @@ config FUSE_DAX If you want to allow mounting a Virtio Filesystem with the "dax" option, answer Y. + +config FUSE_PASSTHROUGH + bool "FUSE passthrough operations support" + default y + depends on FUSE_FS + select FS_STACK + help + This allows bypassing FUSE server by mapping specific FUSE operations + to be performed directly on a backing file. + + If you want to allow passthrough operations, answer Y. diff --git a/fs/fuse/Makefile b/fs/fuse/Makefile index 0c48b35c058d..6e0228c6d0cb 100644 --- a/fs/fuse/Makefile +++ b/fs/fuse/Makefile @@ -8,6 +8,8 @@ obj-$(CONFIG_CUSE) += cuse.o obj-$(CONFIG_VIRTIO_FS) += virtiofs.o fuse-y := dev.o dir.o file.o inode.o control.o xattr.o acl.o readdir.o ioctl.o +fuse-y += iomode.o fuse-$(CONFIG_FUSE_DAX) += dax.o +fuse-$(CONFIG_FUSE_PASSTHROUGH) += passthrough.o virtiofs-y := virtio_fs.o diff --git a/fs/fuse/control.c b/fs/fuse/control.c index 284a35006462..97ac994ff78f 100644 --- a/fs/fuse/control.c +++ b/fs/fuse/control.c @@ -174,11 +174,7 @@ static ssize_t fuse_conn_congestion_threshold_write(struct file *file, if (!fc) goto out; - down_read(&fc->killsb); - spin_lock(&fc->bg_lock); - fc->congestion_threshold = val; - spin_unlock(&fc->bg_lock); - up_read(&fc->killsb); + WRITE_ONCE(fc->congestion_threshold, val); fuse_conn_put(fc); out: return ret; diff --git a/fs/fuse/dev.c b/fs/fuse/dev.c index 1a8f82f478cb..3ec8bb5e68ff 100644 --- a/fs/fuse/dev.c +++ b/fs/fuse/dev.c @@ -1775,6 +1775,61 @@ copy_finish: return err; } +/* + * Resending all processing queue requests. + * + * During a FUSE daemon panics and failover, it is possible for some inflight + * requests to be lost and never returned. As a result, applications awaiting + * replies would become stuck forever. To address this, we can use notification + * to trigger resending of these pending requests to the FUSE daemon, ensuring + * they are properly processed again. + * + * Please note that this strategy is applicable only to idempotent requests or + * if the FUSE daemon takes careful measures to avoid processing duplicated + * non-idempotent requests. + */ +static void fuse_resend(struct fuse_conn *fc) +{ + struct fuse_dev *fud; + struct fuse_req *req, *next; + struct fuse_iqueue *fiq = &fc->iq; + LIST_HEAD(to_queue); + unsigned int i; + + spin_lock(&fc->lock); + if (!fc->connected) { + spin_unlock(&fc->lock); + return; + } + + list_for_each_entry(fud, &fc->devices, entry) { + struct fuse_pqueue *fpq = &fud->pq; + + spin_lock(&fpq->lock); + for (i = 0; i < FUSE_PQ_HASH_SIZE; i++) + list_splice_tail_init(&fpq->processing[i], &to_queue); + spin_unlock(&fpq->lock); + } + spin_unlock(&fc->lock); + + list_for_each_entry_safe(req, next, &to_queue, list) { + __set_bit(FR_PENDING, &req->flags); + /* mark the request as resend request */ + req->in.h.unique |= FUSE_UNIQUE_RESEND; + } + + spin_lock(&fiq->lock); + /* iq and pq requests are both oldest to newest */ + list_splice(&to_queue, &fiq->pending); + fiq->ops->wake_pending_and_unlock(fiq); +} + +static int fuse_notify_resend(struct fuse_conn *fc) +{ + fuse_resend(fc); + return 0; +} + static int fuse_notify(struct fuse_conn *fc, enum fuse_notify_code code, unsigned int size, struct fuse_copy_state *cs) { @@ -1800,6 +1855,9 @@ static int fuse_notify(struct fuse_conn *fc, enum fuse_notify_code code, case FUSE_NOTIFY_DELETE: return fuse_notify_delete(fc, size, cs); + case FUSE_NOTIFY_RESEND: + return fuse_notify_resend(fc); + default: fuse_copy_finish(cs); return -EINVAL; @@ -2251,43 +2309,91 @@ static int fuse_device_clone(struct fuse_conn *fc, struct file *new) return 0; } -static long fuse_dev_ioctl(struct file *file, unsigned int cmd, - unsigned long arg) +static long fuse_dev_ioctl_clone(struct file *file, __u32 __user *argp) { int res; int oldfd; struct fuse_dev *fud = NULL; struct fd f; + if (get_user(oldfd, argp)) + return -EFAULT; + + f = fdget(oldfd); + if (!f.file) + return -EINVAL; + + /* + * Check against file->f_op because CUSE + * uses the same ioctl handler. + */ + if (f.file->f_op == file->f_op) + fud = fuse_get_dev(f.file); + + res = -EINVAL; + if (fud) { + mutex_lock(&fuse_mutex); + res = fuse_device_clone(fud->fc, file); + mutex_unlock(&fuse_mutex); + } + + fdput(f); + return res; +} + +static long fuse_dev_ioctl_backing_open(struct file *file, + struct fuse_backing_map __user *argp) +{ + struct fuse_dev *fud = fuse_get_dev(file); + struct fuse_backing_map map; + + if (!fud) + return -EPERM; + + if (!IS_ENABLED(CONFIG_FUSE_PASSTHROUGH)) + return -EOPNOTSUPP; + + if (copy_from_user(&map, argp, sizeof(map))) + return -EFAULT; + + return fuse_backing_open(fud->fc, &map); +} + +static long fuse_dev_ioctl_backing_close(struct file *file, __u32 __user *argp) +{ + struct fuse_dev *fud = fuse_get_dev(file); + int backing_id; + + if (!fud) + return -EPERM; + + if (!IS_ENABLED(CONFIG_FUSE_PASSTHROUGH)) + return -EOPNOTSUPP; + + if (get_user(backing_id, argp)) + return -EFAULT; + + return fuse_backing_close(fud->fc, backing_id); +} + +static long fuse_dev_ioctl(struct file *file, unsigned int cmd, + unsigned long arg) +{ + void __user *argp = (void __user *)arg; + switch (cmd) { case FUSE_DEV_IOC_CLONE: - if (get_user(oldfd, (__u32 __user *)arg)) - return -EFAULT; + return fuse_dev_ioctl_clone(file, argp); - f = fdget(oldfd); - if (!f.file) - return -EINVAL; + case FUSE_DEV_IOC_BACKING_OPEN: + return fuse_dev_ioctl_backing_open(file, argp); + + case FUSE_DEV_IOC_BACKING_CLOSE: + return fuse_dev_ioctl_backing_close(file, argp); - /* - * Check against file->f_op because CUSE - * uses the same ioctl handler. - */ - if (f.file->f_op == file->f_op) - fud = fuse_get_dev(f.file); - - res = -EINVAL; - if (fud) { - mutex_lock(&fuse_mutex); - res = fuse_device_clone(fud->fc, file); - mutex_unlock(&fuse_mutex); - } - fdput(f); - break; default: - res = -ENOTTY; - break; + return -ENOTTY; } - return res; } const struct file_operations fuse_dev_operations = { diff --git a/fs/fuse/dir.c b/fs/fuse/dir.c index d19cbf34c634..4a6df591add6 100644 --- a/fs/fuse/dir.c +++ b/fs/fuse/dir.c @@ -391,6 +391,10 @@ int fuse_lookup_name(struct super_block *sb, u64 nodeid, const struct qstr *name err = -EIO; if (fuse_invalid_attr(&outarg->attr)) goto out_put_forget; + if (outarg->nodeid == FUSE_ROOT_ID && outarg->generation != 0) { + pr_warn_once("root generation should be zero\n"); + outarg->generation = 0; + } *inode = fuse_iget(sb, outarg->nodeid, outarg->generation, &outarg->attr, ATTR_TIMEOUT(outarg), @@ -615,7 +619,7 @@ static int fuse_create_open(struct inode *dir, struct dentry *entry, FUSE_ARGS(args); struct fuse_forget_link *forget; struct fuse_create_in inarg; - struct fuse_open_out outopen; + struct fuse_open_out *outopenp; struct fuse_entry_out outentry; struct fuse_inode *fi; struct fuse_file *ff; @@ -630,7 +634,7 @@ static int fuse_create_open(struct inode *dir, struct dentry *entry, goto out_err; err = -ENOMEM; - ff = fuse_file_alloc(fm); + ff = fuse_file_alloc(fm, true); if (!ff) goto out_put_forget_req; @@ -659,8 +663,10 @@ static int fuse_create_open(struct inode *dir, struct dentry *entry, args.out_numargs = 2; args.out_args[0].size = sizeof(outentry); args.out_args[0].value = &outentry; - args.out_args[1].size = sizeof(outopen); - args.out_args[1].value = &outopen; + /* Store outarg for fuse_finish_open() */ + outopenp = &ff->args->open_outarg; + args.out_args[1].size = sizeof(*outopenp); + args.out_args[1].value = outopenp; err = get_create_ext(&args, dir, entry, mode); if (err) @@ -676,9 +682,9 @@ static int fuse_create_open(struct inode *dir, struct dentry *entry, fuse_invalid_attr(&outentry.attr)) goto out_free_ff; - ff->fh = outopen.fh; + ff->fh = outopenp->fh; ff->nodeid = outentry.nodeid; - ff->open_flags = outopen.open_flags; + ff->open_flags = outopenp->open_flags; inode = fuse_iget(dir->i_sb, outentry.nodeid, outentry.generation, &outentry.attr, ATTR_TIMEOUT(&outentry), 0); if (!inode) { @@ -692,13 +698,15 @@ static int fuse_create_open(struct inode *dir, struct dentry *entry, d_instantiate(entry, inode); fuse_change_entry_timeout(entry, &outentry); fuse_dir_changed(dir); - err = finish_open(file, entry, generic_file_open); + err = generic_file_open(inode, file); + if (!err) { + file->private_data = ff; + err = finish_open(file, entry, fuse_finish_open); + } if (err) { fi = get_fuse_inode(inode); fuse_sync_release(fi, ff, flags); } else { - file->private_data = ff; - fuse_finish_open(inode, file); if (fm->fc->atomic_o_trunc && trunc) truncate_pagecache(inode, 0); else if (!(ff->open_flags & FOPEN_KEEP_CACHE)) @@ -1210,7 +1218,7 @@ static int fuse_do_statx(struct inode *inode, struct file *file, if (((sx->mask & STATX_SIZE) && !fuse_valid_size(sx->size)) || ((sx->mask & STATX_TYPE) && (!fuse_valid_type(sx->mode) || inode_wrong_type(inode, sx->mode)))) { - make_bad_inode(inode); + fuse_make_bad(inode); return -EIO; } @@ -1485,7 +1493,7 @@ static int fuse_perm_getattr(struct inode *inode, int mask) * * 1) Local access checking ('default_permissions' mount option) based * on file mode. This is the plain old disk filesystem permission - * modell. + * model. * * 2) "Remote" access checking, where server is responsible for * checking permission in each inode operation. An exception to this @@ -1630,7 +1638,30 @@ out_err: static int fuse_dir_open(struct inode *inode, struct file *file) { - return fuse_open_common(inode, file, true); + struct fuse_mount *fm = get_fuse_mount(inode); + int err; + + if (fuse_is_bad(inode)) + return -EIO; + + err = generic_file_open(inode, file); + if (err) + return err; + + err = fuse_do_open(fm, get_node_id(inode), file, true); + if (!err) { + struct fuse_file *ff = file->private_data; + + /* + * Keep handling FOPEN_STREAM and FOPEN_NONSEEKABLE for + * directories for backward compatibility, though it's unlikely + * to be useful. + */ + if (ff->open_flags & (FOPEN_STREAM | FOPEN_NONSEEKABLE)) + nonseekable_open(inode, file); + } + + return err; } static int fuse_dir_release(struct inode *inode, struct file *file) diff --git a/fs/fuse/file.c b/fs/fuse/file.c index c007b0f0c3a7..a56e7bffd000 100644 --- a/fs/fuse/file.c +++ b/fs/fuse/file.c @@ -20,6 +20,7 @@ #include <linux/fs.h> #include <linux/filelock.h> #include <linux/splice.h> +#include <linux/task_io_accounting_ops.h> static int fuse_send_open(struct fuse_mount *fm, u64 nodeid, unsigned int open_flags, int opcode, @@ -50,13 +51,7 @@ static int fuse_send_open(struct fuse_mount *fm, u64 nodeid, return fuse_simple_request(fm, &args); } -struct fuse_release_args { - struct fuse_args args; - struct fuse_release_in inarg; - struct inode *inode; -}; - -struct fuse_file *fuse_file_alloc(struct fuse_mount *fm) +struct fuse_file *fuse_file_alloc(struct fuse_mount *fm, bool release) { struct fuse_file *ff; @@ -65,15 +60,15 @@ struct fuse_file *fuse_file_alloc(struct fuse_mount *fm) return NULL; ff->fm = fm; - ff->release_args = kzalloc(sizeof(*ff->release_args), - GFP_KERNEL_ACCOUNT); - if (!ff->release_args) { - kfree(ff); - return NULL; + if (release) { + ff->args = kzalloc(sizeof(*ff->args), GFP_KERNEL_ACCOUNT); + if (!ff->args) { + kfree(ff); + return NULL; + } } INIT_LIST_HEAD(&ff->write_entry); - mutex_init(&ff->readdir.lock); refcount_set(&ff->count, 1); RB_CLEAR_NODE(&ff->polled_node); init_waitqueue_head(&ff->poll_wait); @@ -85,8 +80,7 @@ struct fuse_file *fuse_file_alloc(struct fuse_mount *fm) void fuse_file_free(struct fuse_file *ff) { - kfree(ff->release_args); - mutex_destroy(&ff->readdir.lock); + kfree(ff->args); kfree(ff); } @@ -105,14 +99,17 @@ static void fuse_release_end(struct fuse_mount *fm, struct fuse_args *args, kfree(ra); } -static void fuse_file_put(struct fuse_file *ff, bool sync, bool isdir) +static void fuse_file_put(struct fuse_file *ff, bool sync) { if (refcount_dec_and_test(&ff->count)) { - struct fuse_args *args = &ff->release_args->args; + struct fuse_release_args *ra = &ff->args->release_args; + struct fuse_args *args = (ra ? &ra->args : NULL); - if (isdir ? ff->fm->fc->no_opendir : ff->fm->fc->no_open) { - /* Do nothing when client does not implement 'open' */ - fuse_release_end(ff->fm, args, 0); + if (ra && ra->inode) + fuse_file_io_release(ff, ra->inode); + + if (!args) { + /* Do nothing when server does not implement 'open' */ } else if (sync) { fuse_simple_request(ff->fm, args); fuse_release_end(ff->fm, args, 0); @@ -132,27 +129,31 @@ struct fuse_file *fuse_file_open(struct fuse_mount *fm, u64 nodeid, struct fuse_conn *fc = fm->fc; struct fuse_file *ff; int opcode = isdir ? FUSE_OPENDIR : FUSE_OPEN; + bool open = isdir ? !fc->no_opendir : !fc->no_open; - ff = fuse_file_alloc(fm); + ff = fuse_file_alloc(fm, open); if (!ff) return ERR_PTR(-ENOMEM); ff->fh = 0; /* Default for no-open */ ff->open_flags = FOPEN_KEEP_CACHE | (isdir ? FOPEN_CACHE_DIR : 0); - if (isdir ? !fc->no_opendir : !fc->no_open) { - struct fuse_open_out outarg; + if (open) { + /* Store outarg for fuse_finish_open() */ + struct fuse_open_out *outargp = &ff->args->open_outarg; int err; - err = fuse_send_open(fm, nodeid, open_flags, opcode, &outarg); + err = fuse_send_open(fm, nodeid, open_flags, opcode, outargp); if (!err) { - ff->fh = outarg.fh; - ff->open_flags = outarg.open_flags; - + ff->fh = outargp->fh; + ff->open_flags = outargp->open_flags; } else if (err != -ENOSYS) { fuse_file_free(ff); return ERR_PTR(err); } else { + /* No release needed */ + kfree(ff->args); + ff->args = NULL; if (isdir) fc->no_opendir = 1; else @@ -195,40 +196,50 @@ static void fuse_link_write_file(struct file *file) spin_unlock(&fi->lock); } -void fuse_finish_open(struct inode *inode, struct file *file) +int fuse_finish_open(struct inode *inode, struct file *file) { struct fuse_file *ff = file->private_data; struct fuse_conn *fc = get_fuse_conn(inode); + int err; + + err = fuse_file_io_open(file, inode); + if (err) + return err; if (ff->open_flags & FOPEN_STREAM) stream_open(inode, file); else if (ff->open_flags & FOPEN_NONSEEKABLE) nonseekable_open(inode, file); - if (fc->atomic_o_trunc && (file->f_flags & O_TRUNC)) { - struct fuse_inode *fi = get_fuse_inode(inode); - - spin_lock(&fi->lock); - fi->attr_version = atomic64_inc_return(&fc->attr_version); - i_size_write(inode, 0); - spin_unlock(&fi->lock); - file_update_time(file); - fuse_invalidate_attr_mask(inode, FUSE_STATX_MODSIZE); - } if ((file->f_mode & FMODE_WRITE) && fc->writeback_cache) fuse_link_write_file(file); + + return 0; } -int fuse_open_common(struct inode *inode, struct file *file, bool isdir) +static void fuse_truncate_update_attr(struct inode *inode, struct file *file) +{ + struct fuse_conn *fc = get_fuse_conn(inode); + struct fuse_inode *fi = get_fuse_inode(inode); + + spin_lock(&fi->lock); + fi->attr_version = atomic64_inc_return(&fc->attr_version); + i_size_write(inode, 0); + spin_unlock(&fi->lock); + file_update_time(file); + fuse_invalidate_attr_mask(inode, FUSE_STATX_MODSIZE); +} + +static int fuse_open(struct inode *inode, struct file *file) { struct fuse_mount *fm = get_fuse_mount(inode); + struct fuse_inode *fi = get_fuse_inode(inode); struct fuse_conn *fc = fm->fc; + struct fuse_file *ff; int err; - bool is_wb_truncate = (file->f_flags & O_TRUNC) && - fc->atomic_o_trunc && - fc->writeback_cache; - bool dax_truncate = (file->f_flags & O_TRUNC) && - fc->atomic_o_trunc && FUSE_IS_DAX(inode); + bool is_truncate = (file->f_flags & O_TRUNC) && fc->atomic_o_trunc; + bool is_wb_truncate = is_truncate && fc->writeback_cache; + bool dax_truncate = is_truncate && FUSE_IS_DAX(inode); if (fuse_is_bad(inode)) return -EIO; @@ -250,16 +261,20 @@ int fuse_open_common(struct inode *inode, struct file *file, bool isdir) if (is_wb_truncate || dax_truncate) fuse_set_nowrite(inode); - err = fuse_do_open(fm, get_node_id(inode), file, isdir); - if (!err) - fuse_finish_open(inode, file); + err = fuse_do_open(fm, get_node_id(inode), file, false); + if (!err) { + ff = file->private_data; + err = fuse_finish_open(inode, file); + if (err) + fuse_sync_release(fi, ff, file->f_flags); + else if (is_truncate) + fuse_truncate_update_attr(inode, file); + } if (is_wb_truncate || dax_truncate) fuse_release_nowrite(inode); if (!err) { - struct fuse_file *ff = file->private_data; - - if (fc->atomic_o_trunc && (file->f_flags & O_TRUNC)) + if (is_truncate) truncate_pagecache(inode, 0); else if (!(ff->open_flags & FOPEN_KEEP_CACHE)) invalidate_inode_pages2(inode->i_mapping); @@ -274,10 +289,13 @@ out_inode_unlock: } static void fuse_prepare_release(struct fuse_inode *fi, struct fuse_file *ff, - unsigned int flags, int opcode) + unsigned int flags, int opcode, bool sync) { struct fuse_conn *fc = ff->fm->fc; - struct fuse_release_args *ra = ff->release_args; + struct fuse_release_args *ra = &ff->args->release_args; + + if (fuse_file_passthrough(ff)) + fuse_passthrough_release(ff, fuse_inode_backing(fi)); /* Inode is NULL on error path of fuse_create_open() */ if (likely(fi)) { @@ -292,6 +310,11 @@ static void fuse_prepare_release(struct fuse_inode *fi, struct fuse_file *ff, wake_up_interruptible_all(&ff->poll_wait); + if (!ra) + return; + + /* ff->args was used for open outarg */ + memset(ff->args, 0, sizeof(*ff->args)); ra->inarg.fh = ff->fh; ra->inarg.flags = flags; ra->args.in_numargs = 1; @@ -301,23 +324,28 @@ static void fuse_prepare_release(struct fuse_inode *fi, struct fuse_file *ff, ra->args.nodeid = ff->nodeid; ra->args.force = true; ra->args.nocreds = true; + + /* + * Hold inode until release is finished. + * From fuse_sync_release() the refcount is 1 and everything's + * synchronous, so we are fine with not doing igrab() here. + */ + ra->inode = sync ? NULL : igrab(&fi->inode); } void fuse_file_release(struct inode *inode, struct fuse_file *ff, unsigned int open_flags, fl_owner_t id, bool isdir) { struct fuse_inode *fi = get_fuse_inode(inode); - struct fuse_release_args *ra = ff->release_args; + struct fuse_release_args *ra = &ff->args->release_args; int opcode = isdir ? FUSE_RELEASEDIR : FUSE_RELEASE; - fuse_prepare_release(fi, ff, open_flags, opcode); + fuse_prepare_release(fi, ff, open_flags, opcode, false); - if (ff->flock) { + if (ra && ff->flock) { ra->inarg.release_flags |= FUSE_RELEASE_FLOCK_UNLOCK; ra->inarg.lock_owner = fuse_lock_owner_id(ff->fm->fc, id); } - /* Hold inode until release is finished */ - ra->inode = igrab(inode); /* * Normally this will send the RELEASE request, however if @@ -328,7 +356,7 @@ void fuse_file_release(struct inode *inode, struct fuse_file *ff, * synchronous RELEASE is allowed (and desirable) in this case * because the server can be trusted not to screw up. */ - fuse_file_put(ff, ff->fm->fc->destroy, isdir); + fuse_file_put(ff, ff->fm->fc->destroy); } void fuse_release_common(struct file *file, bool isdir) @@ -337,11 +365,6 @@ void fuse_release_common(struct file *file, bool isdir) (fl_owner_t) file, isdir); } -static int fuse_open(struct inode *inode, struct file *file) -{ - return fuse_open_common(inode, file, false); -} - static int fuse_release(struct inode *inode, struct file *file) { struct fuse_conn *fc = get_fuse_conn(inode); @@ -363,12 +386,8 @@ void fuse_sync_release(struct fuse_inode *fi, struct fuse_file *ff, unsigned int flags) { WARN_ON(refcount_read(&ff->count) > 1); - fuse_prepare_release(fi, ff, flags, FUSE_RELEASE); - /* - * iput(NULL) is a no-op and since the refcount is 1 and everything's - * synchronous, we are fine with not doing igrab() here" - */ - fuse_file_put(ff, true, false); + fuse_prepare_release(fi, ff, flags, FUSE_RELEASE, true); + fuse_file_put(ff, true); } EXPORT_SYMBOL_GPL(fuse_sync_release); @@ -634,7 +653,8 @@ static void fuse_release_user_pages(struct fuse_args_pages *ap, for (i = 0; i < ap->num_pages; i++) { if (should_dirty) set_page_dirty_lock(ap->pages[i]); - put_page(ap->pages[i]); + if (ap->args.is_pinned) + unpin_user_page(ap->pages[i]); } } @@ -925,7 +945,7 @@ static void fuse_readpages_end(struct fuse_mount *fm, struct fuse_args *args, put_page(page); } if (ia->ff) - fuse_file_put(ia->ff, false, false); + fuse_file_put(ia->ff, false); fuse_io_free(ia); } @@ -1299,13 +1319,93 @@ static ssize_t fuse_perform_write(struct kiocb *iocb, struct iov_iter *ii) return res; } +static bool fuse_io_past_eof(struct kiocb *iocb, struct iov_iter *iter) +{ + struct inode *inode = file_inode(iocb->ki_filp); + + return iocb->ki_pos + iov_iter_count(iter) > i_size_read(inode); +} + +/* + * @return true if an exclusive lock for direct IO writes is needed + */ +static bool fuse_dio_wr_exclusive_lock(struct kiocb *iocb, struct iov_iter *from) +{ + struct file *file = iocb->ki_filp; + struct fuse_file *ff = file->private_data; + struct inode *inode = file_inode(iocb->ki_filp); + struct fuse_inode *fi = get_fuse_inode(inode); + + /* Server side has to advise that it supports parallel dio writes. */ + if (!(ff->open_flags & FOPEN_PARALLEL_DIRECT_WRITES)) + return true; + + /* + * Append will need to know the eventual EOF - always needs an + * exclusive lock. + */ + if (iocb->ki_flags & IOCB_APPEND) + return true; + + /* shared locks are not allowed with parallel page cache IO */ + if (test_bit(FUSE_I_CACHE_IO_MODE, &fi->state)) + return false; + + /* Parallel dio beyond EOF is not supported, at least for now. */ + if (fuse_io_past_eof(iocb, from)) + return true; + + return false; +} + +static void fuse_dio_lock(struct kiocb *iocb, struct iov_iter *from, + bool *exclusive) +{ + struct inode *inode = file_inode(iocb->ki_filp); + struct fuse_file *ff = iocb->ki_filp->private_data; + + *exclusive = fuse_dio_wr_exclusive_lock(iocb, from); + if (*exclusive) { + inode_lock(inode); + } else { + inode_lock_shared(inode); + /* + * New parallal dio allowed only if inode is not in caching + * mode and denies new opens in caching mode. This check + * should be performed only after taking shared inode lock. + * Previous past eof check was without inode lock and might + * have raced, so check it again. + */ + if (fuse_io_past_eof(iocb, from) || + fuse_file_uncached_io_start(inode, ff, NULL) != 0) { + inode_unlock_shared(inode); + inode_lock(inode); + *exclusive = true; + } + } +} + +static void fuse_dio_unlock(struct kiocb *iocb, bool exclusive) +{ + struct inode *inode = file_inode(iocb->ki_filp); + struct fuse_file *ff = iocb->ki_filp->private_data; + + if (exclusive) { + inode_unlock(inode); + } else { + /* Allow opens in caching mode after last parallel dio end */ + fuse_file_uncached_io_end(inode, ff); + inode_unlock_shared(inode); + } +} + static ssize_t fuse_cache_write_iter(struct kiocb *iocb, struct iov_iter *from) { struct file *file = iocb->ki_filp; struct address_space *mapping = file->f_mapping; ssize_t written = 0; struct inode *inode = mapping->host; - ssize_t err; + ssize_t err, count; struct fuse_conn *fc = get_fuse_conn(inode); if (fc->writeback_cache) { @@ -1327,10 +1427,12 @@ static ssize_t fuse_cache_write_iter(struct kiocb *iocb, struct iov_iter *from) writethrough: inode_lock(inode); - err = generic_write_checks(iocb, from); + err = count = generic_write_checks(iocb, from); if (err <= 0) goto out; + task_io_account_write(count); + err = file_remove_privs(file); if (err) goto out; @@ -1392,10 +1494,13 @@ static int fuse_get_user_pages(struct fuse_args_pages *ap, struct iov_iter *ii, while (nbytes < *nbytesp && ap->num_pages < max_pages) { unsigned npages; size_t start; - ret = iov_iter_get_pages2(ii, &ap->pages[ap->num_pages], - *nbytesp - nbytes, - max_pages - ap->num_pages, - &start); + struct page **pt_pages; + + pt_pages = &ap->pages[ap->num_pages]; + ret = iov_iter_extract_pages(ii, &pt_pages, + *nbytesp - nbytes, + max_pages - ap->num_pages, + 0, &start); if (ret < 0) break; @@ -1412,6 +1517,7 @@ static int fuse_get_user_pages(struct fuse_args_pages *ap, struct iov_iter *ii, (PAGE_SIZE - ret) & (PAGE_SIZE - 1); } + ap->args.is_pinned = iov_iter_extract_will_pin(ii); ap->args.user_pages = true; if (write) ap->args.in_pages = true; @@ -1558,51 +1664,17 @@ static ssize_t fuse_direct_read_iter(struct kiocb *iocb, struct iov_iter *to) return res; } -static bool fuse_direct_write_extending_i_size(struct kiocb *iocb, - struct iov_iter *iter) -{ - struct inode *inode = file_inode(iocb->ki_filp); - - return iocb->ki_pos + iov_iter_count(iter) > i_size_read(inode); -} - static ssize_t fuse_direct_write_iter(struct kiocb *iocb, struct iov_iter *from) { struct inode *inode = file_inode(iocb->ki_filp); - struct file *file = iocb->ki_filp; - struct fuse_file *ff = file->private_data; struct fuse_io_priv io = FUSE_IO_PRIV_SYNC(iocb); ssize_t res; - bool exclusive_lock = - !(ff->open_flags & FOPEN_PARALLEL_DIRECT_WRITES) || - get_fuse_conn(inode)->direct_io_allow_mmap || - iocb->ki_flags & IOCB_APPEND || - fuse_direct_write_extending_i_size(iocb, from); - - /* - * Take exclusive lock if - * - Parallel direct writes are disabled - a user space decision - * - Parallel direct writes are enabled and i_size is being extended. - * - Shared mmap on direct_io file is supported (FUSE_DIRECT_IO_ALLOW_MMAP). - * This might not be needed at all, but needs further investigation. - */ - if (exclusive_lock) - inode_lock(inode); - else { - inode_lock_shared(inode); - - /* A race with truncate might have come up as the decision for - * the lock type was done without holding the lock, check again. - */ - if (fuse_direct_write_extending_i_size(iocb, from)) { - inode_unlock_shared(inode); - inode_lock(inode); - exclusive_lock = true; - } - } + bool exclusive; + fuse_dio_lock(iocb, from, &exclusive); res = generic_write_checks(iocb, from); if (res > 0) { + task_io_account_write(res); if (!is_sync_kiocb(iocb) && iocb->ki_flags & IOCB_DIRECT) { res = fuse_direct_IO(iocb, from); } else { @@ -1611,10 +1683,7 @@ static ssize_t fuse_direct_write_iter(struct kiocb *iocb, struct iov_iter *from) fuse_write_update_attr(inode, iocb->ki_pos, res); } } - if (exclusive_lock) - inode_unlock(inode); - else - inode_unlock_shared(inode); + fuse_dio_unlock(iocb, exclusive); return res; } @@ -1631,10 +1700,13 @@ static ssize_t fuse_file_read_iter(struct kiocb *iocb, struct iov_iter *to) if (FUSE_IS_DAX(inode)) return fuse_dax_read_iter(iocb, to); - if (!(ff->open_flags & FOPEN_DIRECT_IO)) - return fuse_cache_read_iter(iocb, to); - else + /* FOPEN_DIRECT_IO overrides FOPEN_PASSTHROUGH */ + if (ff->open_flags & FOPEN_DIRECT_IO) return fuse_direct_read_iter(iocb, to); + else if (fuse_file_passthrough(ff)) + return fuse_passthrough_read_iter(iocb, to); + else + return fuse_cache_read_iter(iocb, to); } static ssize_t fuse_file_write_iter(struct kiocb *iocb, struct iov_iter *from) @@ -1649,10 +1721,38 @@ static ssize_t fuse_file_write_iter(struct kiocb *iocb, struct iov_iter *from) if (FUSE_IS_DAX(inode)) return fuse_dax_write_iter(iocb, from); - if (!(ff->open_flags & FOPEN_DIRECT_IO)) + /* FOPEN_DIRECT_IO overrides FOPEN_PASSTHROUGH */ + if (ff->open_flags & FOPEN_DIRECT_IO) + return fuse_direct_write_iter(iocb, from); + else if (fuse_file_passthrough(ff)) + return fuse_passthrough_write_iter(iocb, from); + else return fuse_cache_write_iter(iocb, from); +} + +static ssize_t fuse_splice_read(struct file *in, loff_t *ppos, + struct pipe_inode_info *pipe, size_t len, + unsigned int flags) +{ + struct fuse_file *ff = in->private_data; + + /* FOPEN_DIRECT_IO overrides FOPEN_PASSTHROUGH */ + if (fuse_file_passthrough(ff) && !(ff->open_flags & FOPEN_DIRECT_IO)) + return fuse_passthrough_splice_read(in, ppos, pipe, len, flags); else - return fuse_direct_write_iter(iocb, from); + return filemap_splice_read(in, ppos, pipe, len, flags); +} + +static ssize_t fuse_splice_write(struct pipe_inode_info *pipe, struct file *out, + loff_t *ppos, size_t len, unsigned int flags) +{ + struct fuse_file *ff = out->private_data; + + /* FOPEN_DIRECT_IO overrides FOPEN_PASSTHROUGH */ + if (fuse_file_passthrough(ff) && !(ff->open_flags & FOPEN_DIRECT_IO)) + return fuse_passthrough_splice_write(pipe, out, ppos, len, flags); + else + return iter_file_splice_write(pipe, out, ppos, len, flags); } static void fuse_writepage_free(struct fuse_writepage_args *wpa) @@ -1667,7 +1767,7 @@ static void fuse_writepage_free(struct fuse_writepage_args *wpa) __free_page(ap->pages[i]); if (wpa->ia.ff) - fuse_file_put(wpa->ia.ff, false, false); + fuse_file_put(wpa->ia.ff, false); kfree(ap->pages); kfree(wpa); @@ -1909,7 +2009,7 @@ int fuse_write_inode(struct inode *inode, struct writeback_control *wbc) ff = __fuse_write_file_get(fi); err = fuse_flush_times(inode, ff); if (ff) - fuse_file_put(ff, false, false); + fuse_file_put(ff, false); return err; } @@ -1947,26 +2047,26 @@ static void fuse_writepage_add_to_bucket(struct fuse_conn *fc, rcu_read_unlock(); } -static int fuse_writepage_locked(struct page *page) +static int fuse_writepage_locked(struct folio *folio) { - struct address_space *mapping = page->mapping; + struct address_space *mapping = folio->mapping; struct inode *inode = mapping->host; struct fuse_conn *fc = get_fuse_conn(inode); struct fuse_inode *fi = get_fuse_inode(inode); struct fuse_writepage_args *wpa; struct fuse_args_pages *ap; - struct page *tmp_page; + struct folio *tmp_folio; int error = -ENOMEM; - set_page_writeback(page); + folio_start_writeback(folio); wpa = fuse_writepage_args_alloc(); if (!wpa) goto err; ap = &wpa->ia.ap; - tmp_page = alloc_page(GFP_NOFS | __GFP_HIGHMEM); - if (!tmp_page) + tmp_folio = folio_alloc(GFP_NOFS | __GFP_HIGHMEM, 0); + if (!tmp_folio) goto err_free; error = -EIO; @@ -1975,21 +2075,21 @@ static int fuse_writepage_locked(struct page *page) goto err_nofile; fuse_writepage_add_to_bucket(fc, wpa); - fuse_write_args_fill(&wpa->ia, wpa->ia.ff, page_offset(page), 0); + fuse_write_args_fill(&wpa->ia, wpa->ia.ff, folio_pos(folio), 0); - copy_highpage(tmp_page, page); + folio_copy(tmp_folio, folio); wpa->ia.write.in.write_flags |= FUSE_WRITE_CACHE; wpa->next = NULL; ap->args.in_pages = true; ap->num_pages = 1; - ap->pages[0] = tmp_page; + ap->pages[0] = &tmp_folio->page; ap->descs[0].offset = 0; ap->descs[0].length = PAGE_SIZE; ap->args.end = fuse_writepage_end; wpa->inode = inode; inc_wb_stat(&inode_to_bdi(inode)->wb, WB_WRITEBACK); - inc_node_page_state(tmp_page, NR_WRITEBACK_TEMP); + node_stat_add_folio(tmp_folio, NR_WRITEBACK_TEMP); spin_lock(&fi->lock); tree_insert(&fi->writepages, wpa); @@ -1997,48 +2097,20 @@ static int fuse_writepage_locked(struct page *page) fuse_flush_writepages(inode); spin_unlock(&fi->lock); - end_page_writeback(page); + folio_end_writeback(folio); return 0; err_nofile: - __free_page(tmp_page); + folio_put(tmp_folio); err_free: kfree(wpa); err: - mapping_set_error(page->mapping, error); - end_page_writeback(page); + mapping_set_error(folio->mapping, error); + folio_end_writeback(folio); return error; } -static int fuse_writepage(struct page *page, struct writeback_control *wbc) -{ - struct fuse_conn *fc = get_fuse_conn(page->mapping->host); - int err; - - if (fuse_page_is_writeback(page->mapping->host, page->index)) { - /* - * ->writepages() should be called for sync() and friends. We - * should only get here on direct reclaim and then we are - * allowed to skip a page which is already in flight - */ - WARN_ON(wbc->sync_mode == WB_SYNC_ALL); - - redirty_page_for_writepage(wbc, page); - unlock_page(page); - return 0; - } - - if (wbc->sync_mode == WB_SYNC_NONE && - fc->num_background >= fc->congestion_threshold) - return AOP_WRITEPAGE_ACTIVATE; - - err = fuse_writepage_locked(page); - unlock_page(page); - - return err; -} - struct fuse_fill_wb_data { struct fuse_writepage_args *wpa; struct fuse_file *ff; @@ -2307,7 +2379,7 @@ static int fuse_writepages(struct address_space *mapping, fuse_writepages_send(&data); } if (data.ff) - fuse_file_put(data.ff, false, false); + fuse_file_put(data.ff, false); kfree(data.orig_pages); out: @@ -2401,7 +2473,7 @@ static int fuse_launder_folio(struct folio *folio) /* Serialize with pending writeback for the same page */ fuse_wait_on_page_writeback(inode, folio->index); - err = fuse_writepage_locked(&folio->page); + err = fuse_writepage_locked(folio); if (!err) fuse_wait_on_page_writeback(inode, folio->index); } @@ -2462,13 +2534,30 @@ static int fuse_file_mmap(struct file *file, struct vm_area_struct *vma) { struct fuse_file *ff = file->private_data; struct fuse_conn *fc = ff->fm->fc; + struct inode *inode = file_inode(file); + int rc; /* DAX mmap is superior to direct_io mmap */ - if (FUSE_IS_DAX(file_inode(file))) + if (FUSE_IS_DAX(inode)) return fuse_dax_mmap(file, vma); + /* + * If inode is in passthrough io mode, because it has some file open + * in passthrough mode, either mmap to backing file or fail mmap, + * because mixing cached mmap and passthrough io mode is not allowed. + */ + if (fuse_file_passthrough(ff)) + return fuse_passthrough_mmap(file, vma); + else if (fuse_inode_backing(get_fuse_inode(inode))) + return -ENODEV; + + /* + * FOPEN_DIRECT_IO handling is special compared to O_DIRECT, + * as does not allow MAP_SHARED mmap without FUSE_DIRECT_IO_ALLOW_MMAP. + */ if (ff->open_flags & FOPEN_DIRECT_IO) { - /* Can't provide the coherency needed for MAP_SHARED + /* + * Can't provide the coherency needed for MAP_SHARED * if FUSE_DIRECT_IO_ALLOW_MMAP isn't set. */ if ((vma->vm_flags & VM_MAYSHARE) && !fc->direct_io_allow_mmap) @@ -2476,7 +2565,19 @@ static int fuse_file_mmap(struct file *file, struct vm_area_struct *vma) invalidate_inode_pages2(file->f_mapping); - return generic_file_mmap(file, vma); + if (!(vma->vm_flags & VM_MAYSHARE)) { + /* MAP_PRIVATE */ + return generic_file_mmap(file, vma); + } + + /* + * First mmap of direct_io file enters caching inode io mode. + * Also waits for parallel dio writers to go into serial mode + * (exclusive instead of shared lock). + */ + rc = fuse_file_cached_io_start(inode, ff); + if (rc) + return rc; } if ((vma->vm_flags & VM_SHARED) && (vma->vm_flags & VM_MAYWRITE)) @@ -2580,10 +2681,6 @@ static int fuse_setlk(struct file *file, struct file_lock *fl, int flock) return -ENOLCK; } - /* Unlock on close is handled by the flush method */ - if ((fl->c.flc_flags & FL_CLOSE_POSIX) == FL_CLOSE_POSIX) - return 0; - fuse_lk_fill(&args, file, fl, opcode, pid_nr, flock, &inarg); err = fuse_simple_request(fm, &args); @@ -3213,8 +3310,8 @@ static const struct file_operations fuse_file_operations = { .lock = fuse_file_lock, .get_unmapped_area = thp_get_unmapped_area, .flock = fuse_file_flock, - .splice_read = filemap_splice_read, - .splice_write = iter_file_splice_write, + .splice_read = fuse_splice_read, + .splice_write = fuse_splice_write, .unlocked_ioctl = fuse_file_ioctl, .compat_ioctl = fuse_file_compat_ioctl, .poll = fuse_file_poll, @@ -3225,10 +3322,10 @@ static const struct file_operations fuse_file_operations = { static const struct address_space_operations fuse_file_aops = { .read_folio = fuse_read_folio, .readahead = fuse_readahead, - .writepage = fuse_writepage, .writepages = fuse_writepages, .launder_folio = fuse_launder_folio, .dirty_folio = filemap_dirty_folio, + .migrate_folio = filemap_migrate_folio, .bmap = fuse_bmap, .direct_IO = fuse_direct_IO, .write_begin = fuse_write_begin, @@ -3245,7 +3342,9 @@ void fuse_init_file_inode(struct inode *inode, unsigned int flags) INIT_LIST_HEAD(&fi->write_files); INIT_LIST_HEAD(&fi->queued_writes); fi->writectr = 0; + fi->iocachectr = 0; init_waitqueue_head(&fi->page_waitq); + init_waitqueue_head(&fi->direct_io_waitq); fi->writepages = RB_ROOT; if (IS_ENABLED(CONFIG_FUSE_DAX)) diff --git a/fs/fuse/fuse_i.h b/fs/fuse/fuse_i.h index bcbe34488862..b24084b60864 100644 --- a/fs/fuse/fuse_i.h +++ b/fs/fuse/fuse_i.h @@ -76,6 +76,16 @@ struct fuse_submount_lookup { struct fuse_forget_link *forget; }; +/** Container for data related to mapping to backing file */ +struct fuse_backing { + struct file *file; + struct cred *cred; + + /** refcount */ + refcount_t count; + struct rcu_head rcu; +}; + /** FUSE inode */ struct fuse_inode { /** Inode data */ @@ -111,7 +121,7 @@ struct fuse_inode { u64 attr_version; union { - /* Write related fields (regular file only) */ + /* read/write io cache (regular file only) */ struct { /* Files usable in writepage. Protected by fi->lock */ struct list_head write_files; @@ -123,9 +133,15 @@ struct fuse_inode { * (FUSE_NOWRITE) means more writes are blocked */ int writectr; + /** Number of files/maps using page cache */ + int iocachectr; + /* Waitq for writepage completion */ wait_queue_head_t page_waitq; + /* waitq for direct-io completion */ + wait_queue_head_t direct_io_waitq; + /* List of writepage requestst (pending or sent) */ struct rb_root writepages; }; @@ -173,6 +189,10 @@ struct fuse_inode { #endif /** Submount specific lookup tracking */ struct fuse_submount_lookup *submount_lookup; +#ifdef CONFIG_FUSE_PASSTHROUGH + /** Reference to backing file in passthrough mode */ + struct fuse_backing *fb; +#endif }; /** FUSE inode state bits */ @@ -187,19 +207,21 @@ enum { FUSE_I_BAD, /* Has btime */ FUSE_I_BTIME, + /* Wants or already has page cache IO */ + FUSE_I_CACHE_IO_MODE, }; struct fuse_conn; struct fuse_mount; -struct fuse_release_args; +union fuse_file_args; /** FUSE specific file data */ struct fuse_file { /** Fuse connection for this file */ struct fuse_mount *fm; - /* Argument space reserved for release */ - struct fuse_release_args *release_args; + /* Argument space reserved for open/release */ + union fuse_file_args *args; /** Kernel file handle guaranteed to be unique */ u64 kh; @@ -221,12 +243,6 @@ struct fuse_file { /* Readdir related */ struct { - /* - * Protects below fields against (crazy) parallel readdir on - * same open file. Uncontended in the normal case. - */ - struct mutex lock; - /* Dir stream position */ loff_t pos; @@ -244,6 +260,15 @@ struct fuse_file { /** Wait queue head for poll */ wait_queue_head_t poll_wait; + /** Does file hold a fi->iocachectr refcount? */ + enum { IOM_NONE, IOM_CACHED, IOM_UNCACHED } iomode; + +#ifdef CONFIG_FUSE_PASSTHROUGH + /** Reference to backing file in passthrough mode */ + struct file *passthrough; + const struct cred *cred; +#endif + /** Has flock been performed on this file? */ bool flock:1; }; @@ -283,6 +308,7 @@ struct fuse_args { bool page_replace:1; bool may_block:1; bool is_ext:1; + bool is_pinned:1; struct fuse_in_arg in_args[3]; struct fuse_arg out_args[2]; void (*end)(struct fuse_mount *fm, struct fuse_args *args, int error); @@ -295,6 +321,19 @@ struct fuse_args_pages { unsigned int num_pages; }; +struct fuse_release_args { + struct fuse_args args; + struct fuse_release_in inarg; + struct inode *inode; +}; + +union fuse_file_args { + /* Used during open() */ + struct fuse_open_out open_outarg; + /* Used during release() */ + struct fuse_release_args release_args; +}; + #define FUSE_ARGS(args) struct fuse_args args = {} /** The request IO state (for asynchronous processing) */ @@ -818,6 +857,12 @@ struct fuse_conn { /* Is statx not implemented by fs? */ unsigned int no_statx:1; + /** Passthrough support for read/write IO */ + unsigned int passthrough:1; + + /** Maximum stack depth for passthrough backing files */ + int max_stack_depth; + /** The number of requests waiting for completion */ atomic_t num_waiting; @@ -867,6 +912,11 @@ struct fuse_conn { /* New writepages go into this bucket */ struct fuse_sync_bucket __rcu *curr_bucket; + +#ifdef CONFIG_FUSE_PASSTHROUGH + /** IDR for backing files ids */ + struct idr backing_files_map; +#endif }; /* @@ -940,7 +990,6 @@ static inline bool fuse_stale_inode(const struct inode *inode, int generation, static inline void fuse_make_bad(struct inode *inode) { - remove_inode_hash(inode); set_bit(FUSE_I_BAD, &get_fuse_inode(inode)->state); } @@ -1032,14 +1081,9 @@ void fuse_read_args_fill(struct fuse_io_args *ia, struct file *file, loff_t pos, size_t count, int opcode); -/** - * Send OPEN or OPENDIR request - */ -int fuse_open_common(struct inode *inode, struct file *file, bool isdir); - -struct fuse_file *fuse_file_alloc(struct fuse_mount *fm); +struct fuse_file *fuse_file_alloc(struct fuse_mount *fm, bool release); void fuse_file_free(struct fuse_file *ff); -void fuse_finish_open(struct inode *inode, struct file *file); +int fuse_finish_open(struct inode *inode, struct file *file); void fuse_sync_release(struct fuse_inode *fi, struct fuse_file *ff, unsigned int flags); @@ -1349,11 +1393,82 @@ int fuse_fileattr_get(struct dentry *dentry, struct fileattr *fa); int fuse_fileattr_set(struct mnt_idmap *idmap, struct dentry *dentry, struct fileattr *fa); -/* file.c */ +/* iomode.c */ +int fuse_file_cached_io_start(struct inode *inode, struct fuse_file *ff); +int fuse_file_uncached_io_start(struct inode *inode, struct fuse_file *ff, struct fuse_backing *fb); +void fuse_file_uncached_io_end(struct inode *inode, struct fuse_file *ff); + +int fuse_file_io_open(struct file *file, struct inode *inode); +void fuse_file_io_release(struct fuse_file *ff, struct inode *inode); +/* file.c */ struct fuse_file *fuse_file_open(struct fuse_mount *fm, u64 nodeid, unsigned int open_flags, bool isdir); void fuse_file_release(struct inode *inode, struct fuse_file *ff, unsigned int open_flags, fl_owner_t id, bool isdir); +/* passthrough.c */ +static inline struct fuse_backing *fuse_inode_backing(struct fuse_inode *fi) +{ +#ifdef CONFIG_FUSE_PASSTHROUGH + return READ_ONCE(fi->fb); +#else + return NULL; +#endif +} + +static inline struct fuse_backing *fuse_inode_backing_set(struct fuse_inode *fi, + struct fuse_backing *fb) +{ +#ifdef CONFIG_FUSE_PASSTHROUGH + return xchg(&fi->fb, fb); +#else + return NULL; +#endif +} + +#ifdef CONFIG_FUSE_PASSTHROUGH +struct fuse_backing *fuse_backing_get(struct fuse_backing *fb); +void fuse_backing_put(struct fuse_backing *fb); +#else + +static inline struct fuse_backing *fuse_backing_get(struct fuse_backing *fb) +{ + return NULL; +} + +static inline void fuse_backing_put(struct fuse_backing *fb) +{ +} +#endif + +void fuse_backing_files_init(struct fuse_conn *fc); +void fuse_backing_files_free(struct fuse_conn *fc); +int fuse_backing_open(struct fuse_conn *fc, struct fuse_backing_map *map); +int fuse_backing_close(struct fuse_conn *fc, int backing_id); + +struct fuse_backing *fuse_passthrough_open(struct file *file, + struct inode *inode, + int backing_id); +void fuse_passthrough_release(struct fuse_file *ff, struct fuse_backing *fb); + +static inline struct file *fuse_file_passthrough(struct fuse_file *ff) +{ +#ifdef CONFIG_FUSE_PASSTHROUGH + return ff->passthrough; +#else + return NULL; +#endif +} + +ssize_t fuse_passthrough_read_iter(struct kiocb *iocb, struct iov_iter *iter); +ssize_t fuse_passthrough_write_iter(struct kiocb *iocb, struct iov_iter *iter); +ssize_t fuse_passthrough_splice_read(struct file *in, loff_t *ppos, + struct pipe_inode_info *pipe, + size_t len, unsigned int flags); +ssize_t fuse_passthrough_splice_write(struct pipe_inode_info *pipe, + struct file *out, loff_t *ppos, + size_t len, unsigned int flags); +ssize_t fuse_passthrough_mmap(struct file *file, struct vm_area_struct *vma); + #endif /* _FS_FUSE_I_H */ diff --git a/fs/fuse/inode.c b/fs/fuse/inode.c index 516ea2979a90..3a5d88878335 100644 --- a/fs/fuse/inode.c +++ b/fs/fuse/inode.c @@ -111,6 +111,9 @@ static struct inode *fuse_alloc_inode(struct super_block *sb) if (IS_ENABLED(CONFIG_FUSE_DAX) && !fuse_dax_inode_alloc(sb, fi)) goto out_free_forget; + if (IS_ENABLED(CONFIG_FUSE_PASSTHROUGH)) + fuse_inode_backing_set(fi, NULL); + return &fi->inode; out_free_forget: @@ -129,6 +132,9 @@ static void fuse_free_inode(struct inode *inode) #ifdef CONFIG_FUSE_DAX kfree(fi->dax); #endif + if (IS_ENABLED(CONFIG_FUSE_PASSTHROUGH)) + fuse_backing_put(fuse_inode_backing(fi)); + kmem_cache_free(fuse_inode_cachep, fi); } @@ -469,8 +475,11 @@ retry: } else if (fuse_stale_inode(inode, generation, attr)) { /* nodeid was reused, any I/O on the old inode should fail */ fuse_make_bad(inode); - iput(inode); - goto retry; + if (inode != d_inode(sb->s_root)) { + remove_inode_hash(inode); + iput(inode); + goto retry; + } } fi = get_fuse_inode(inode); spin_lock(&fi->lock); @@ -924,6 +933,9 @@ void fuse_conn_init(struct fuse_conn *fc, struct fuse_mount *fm, fc->max_pages = FUSE_DEFAULT_MAX_PAGES_PER_REQ; fc->max_pages_limit = FUSE_MAX_MAX_PAGES; + if (IS_ENABLED(CONFIG_FUSE_PASSTHROUGH)) + fuse_backing_files_init(fc); + INIT_LIST_HEAD(&fc->mounts); list_add(&fm->fc_entry, &fc->mounts); fm->fc = fc; @@ -954,6 +966,8 @@ void fuse_conn_put(struct fuse_conn *fc) WARN_ON(atomic_read(&bucket->count) != 1); kfree(bucket); } + if (IS_ENABLED(CONFIG_FUSE_PASSTHROUGH)) + fuse_backing_files_free(fc); call_rcu(&fc->rcu, delayed_release); } } @@ -974,7 +988,7 @@ static struct inode *fuse_get_root_inode(struct super_block *sb, unsigned mode) attr.mode = mode; attr.ino = FUSE_ROOT_ID; attr.nlink = 1; - return fuse_iget(sb, 1, 0, &attr, 0, 0); + return fuse_iget(sb, FUSE_ROOT_ID, 0, &attr, 0, 0); } struct fuse_inode_handle { @@ -1117,6 +1131,11 @@ static struct dentry *fuse_get_parent(struct dentry *child) return parent; } +/* only for fid encoding; no support for file handle */ +static const struct export_operations fuse_export_fid_operations = { + .encode_fh = fuse_encode_fh, +}; + static const struct export_operations fuse_export_operations = { .fh_to_dentry = fuse_fh_to_dentry, .fh_to_parent = fuse_fh_to_parent, @@ -1291,6 +1310,26 @@ static void process_init_reply(struct fuse_mount *fm, struct fuse_args *args, fc->create_supp_group = 1; if (flags & FUSE_DIRECT_IO_ALLOW_MMAP) fc->direct_io_allow_mmap = 1; + /* + * max_stack_depth is the max stack depth of FUSE fs, + * so it has to be at least 1 to support passthrough + * to backing files. + * + * with max_stack_depth > 1, the backing files can be + * on a stacked fs (e.g. overlayfs) themselves and with + * max_stack_depth == 1, FUSE fs can be stacked as the + * underlying fs of a stacked fs (e.g. overlayfs). + */ + if (IS_ENABLED(CONFIG_FUSE_PASSTHROUGH) && + (flags & FUSE_PASSTHROUGH) && + arg->max_stack_depth > 0 && + arg->max_stack_depth <= FILESYSTEM_MAX_STACK_DEPTH) { + fc->passthrough = 1; + fc->max_stack_depth = arg->max_stack_depth; + fm->sb->s_stack_depth = arg->max_stack_depth; + } + if (flags & FUSE_NO_EXPORT_SUPPORT) + fm->sb->s_export_op = &fuse_export_fid_operations; } else { ra_pages = fc->max_read / PAGE_SIZE; fc->no_lock = 1; @@ -1337,7 +1376,8 @@ void fuse_send_init(struct fuse_mount *fm) FUSE_NO_OPENDIR_SUPPORT | FUSE_EXPLICIT_INVAL_DATA | FUSE_HANDLE_KILLPRIV_V2 | FUSE_SETXATTR_EXT | FUSE_INIT_EXT | FUSE_SECURITY_CTX | FUSE_CREATE_SUPP_GROUP | - FUSE_HAS_EXPIRE_ONLY | FUSE_DIRECT_IO_ALLOW_MMAP; + FUSE_HAS_EXPIRE_ONLY | FUSE_DIRECT_IO_ALLOW_MMAP | + FUSE_NO_EXPORT_SUPPORT | FUSE_HAS_RESEND; #ifdef CONFIG_FUSE_DAX if (fm->fc->dax) flags |= FUSE_MAP_ALIGNMENT; @@ -1346,6 +1386,8 @@ void fuse_send_init(struct fuse_mount *fm) #endif if (fm->fc->auto_submounts) flags |= FUSE_SUBMOUNTS; + if (IS_ENABLED(CONFIG_FUSE_PASSTHROUGH)) + flags |= FUSE_PASSTHROUGH; ia->in.flags = flags; ia->in.flags2 = flags >> 32; @@ -1496,8 +1538,8 @@ static void fuse_fill_attr_from_inode(struct fuse_attr *attr, .ctimensec = ctime.tv_nsec, .mode = fi->inode.i_mode, .nlink = fi->inode.i_nlink, - .uid = fi->inode.i_uid.val, - .gid = fi->inode.i_gid.val, + .uid = __kuid_val(fi->inode.i_uid), + .gid = __kgid_val(fi->inode.i_gid), .rdev = fi->inode.i_rdev, .blksize = 1u << fi->inode.i_blkbits, }; @@ -1534,6 +1576,7 @@ static int fuse_fill_super_submount(struct super_block *sb, sb->s_bdi = bdi_get(parent_sb->s_bdi); sb->s_xattr = parent_sb->s_xattr; + sb->s_export_op = parent_sb->s_export_op; sb->s_time_gran = parent_sb->s_time_gran; sb->s_blocksize = parent_sb->s_blocksize; sb->s_blocksize_bits = parent_sb->s_blocksize_bits; diff --git a/fs/fuse/iomode.c b/fs/fuse/iomode.c new file mode 100644 index 000000000000..c653ddcf0578 --- /dev/null +++ b/fs/fuse/iomode.c @@ -0,0 +1,254 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * FUSE inode io modes. + * + * Copyright (c) 2024 CTERA Networks. + */ + +#include "fuse_i.h" + +#include <linux/kernel.h> +#include <linux/sched.h> +#include <linux/file.h> +#include <linux/fs.h> + +/* + * Return true if need to wait for new opens in caching mode. + */ +static inline bool fuse_is_io_cache_wait(struct fuse_inode *fi) +{ + return READ_ONCE(fi->iocachectr) < 0 && !fuse_inode_backing(fi); +} + +/* + * Start cached io mode. + * + * Blocks new parallel dio writes and waits for the in-progress parallel dio + * writes to complete. + */ +int fuse_file_cached_io_start(struct inode *inode, struct fuse_file *ff) +{ + struct fuse_inode *fi = get_fuse_inode(inode); + + /* There are no io modes if server does not implement open */ + if (!ff->args) + return 0; + + spin_lock(&fi->lock); + /* + * Setting the bit advises new direct-io writes to use an exclusive + * lock - without it the wait below might be forever. + */ + while (fuse_is_io_cache_wait(fi)) { + set_bit(FUSE_I_CACHE_IO_MODE, &fi->state); + spin_unlock(&fi->lock); + wait_event(fi->direct_io_waitq, !fuse_is_io_cache_wait(fi)); + spin_lock(&fi->lock); + } + + /* + * Check if inode entered passthrough io mode while waiting for parallel + * dio write completion. + */ + if (fuse_inode_backing(fi)) { + clear_bit(FUSE_I_CACHE_IO_MODE, &fi->state); + spin_unlock(&fi->lock); + return -ETXTBSY; + } + + WARN_ON(ff->iomode == IOM_UNCACHED); + if (ff->iomode == IOM_NONE) { + ff->iomode = IOM_CACHED; + if (fi->iocachectr == 0) + set_bit(FUSE_I_CACHE_IO_MODE, &fi->state); + fi->iocachectr++; + } + spin_unlock(&fi->lock); + return 0; +} + +static void fuse_file_cached_io_end(struct inode *inode, struct fuse_file *ff) +{ + struct fuse_inode *fi = get_fuse_inode(inode); + + spin_lock(&fi->lock); + WARN_ON(fi->iocachectr <= 0); + WARN_ON(ff->iomode != IOM_CACHED); + ff->iomode = IOM_NONE; + fi->iocachectr--; + if (fi->iocachectr == 0) + clear_bit(FUSE_I_CACHE_IO_MODE, &fi->state); + spin_unlock(&fi->lock); +} + +/* Start strictly uncached io mode where cache access is not allowed */ +int fuse_file_uncached_io_start(struct inode *inode, struct fuse_file *ff, struct fuse_backing *fb) +{ + struct fuse_inode *fi = get_fuse_inode(inode); + struct fuse_backing *oldfb; + int err = 0; + + spin_lock(&fi->lock); + /* deny conflicting backing files on same fuse inode */ + oldfb = fuse_inode_backing(fi); + if (oldfb && oldfb != fb) { + err = -EBUSY; + goto unlock; + } + if (fi->iocachectr > 0) { + err = -ETXTBSY; + goto unlock; + } + WARN_ON(ff->iomode != IOM_NONE); + fi->iocachectr--; + ff->iomode = IOM_UNCACHED; + + /* fuse inode holds a single refcount of backing file */ + if (!oldfb) { + oldfb = fuse_inode_backing_set(fi, fb); + WARN_ON_ONCE(oldfb != NULL); + } else { + fuse_backing_put(fb); + } +unlock: + spin_unlock(&fi->lock); + return err; +} + +void fuse_file_uncached_io_end(struct inode *inode, struct fuse_file *ff) +{ + struct fuse_inode *fi = get_fuse_inode(inode); + struct fuse_backing *oldfb = NULL; + + spin_lock(&fi->lock); + WARN_ON(fi->iocachectr >= 0); + WARN_ON(ff->iomode != IOM_UNCACHED); + ff->iomode = IOM_NONE; + fi->iocachectr++; + if (!fi->iocachectr) { + wake_up(&fi->direct_io_waitq); + oldfb = fuse_inode_backing_set(fi, NULL); + } + spin_unlock(&fi->lock); + if (oldfb) + fuse_backing_put(oldfb); +} + +/* + * Open flags that are allowed in combination with FOPEN_PASSTHROUGH. + * A combination of FOPEN_PASSTHROUGH and FOPEN_DIRECT_IO means that read/write + * operations go directly to the server, but mmap is done on the backing file. + * FOPEN_PASSTHROUGH mode should not co-exist with any users of the fuse inode + * page cache, so FOPEN_KEEP_CACHE is a strange and undesired combination. + */ +#define FOPEN_PASSTHROUGH_MASK \ + (FOPEN_PASSTHROUGH | FOPEN_DIRECT_IO | FOPEN_PARALLEL_DIRECT_WRITES | \ + FOPEN_NOFLUSH) + +static int fuse_file_passthrough_open(struct inode *inode, struct file *file) +{ + struct fuse_file *ff = file->private_data; + struct fuse_conn *fc = get_fuse_conn(inode); + struct fuse_backing *fb; + int err; + + /* Check allowed conditions for file open in passthrough mode */ + if (!IS_ENABLED(CONFIG_FUSE_PASSTHROUGH) || !fc->passthrough || + (ff->open_flags & ~FOPEN_PASSTHROUGH_MASK)) + return -EINVAL; + + fb = fuse_passthrough_open(file, inode, + ff->args->open_outarg.backing_id); + if (IS_ERR(fb)) + return PTR_ERR(fb); + + /* First passthrough file open denies caching inode io mode */ + err = fuse_file_uncached_io_start(inode, ff, fb); + if (!err) + return 0; + + fuse_passthrough_release(ff, fb); + fuse_backing_put(fb); + + return err; +} + +/* Request access to submit new io to inode via open file */ +int fuse_file_io_open(struct file *file, struct inode *inode) +{ + struct fuse_file *ff = file->private_data; + struct fuse_inode *fi = get_fuse_inode(inode); + int err; + + /* + * io modes are not relevant with DAX and with server that does not + * implement open. + */ + if (FUSE_IS_DAX(inode) || !ff->args) + return 0; + + /* + * Server is expected to use FOPEN_PASSTHROUGH for all opens of an inode + * which is already open for passthrough. + */ + err = -EINVAL; + if (fuse_inode_backing(fi) && !(ff->open_flags & FOPEN_PASSTHROUGH)) + goto fail; + + /* + * FOPEN_PARALLEL_DIRECT_WRITES requires FOPEN_DIRECT_IO. + */ + if (!(ff->open_flags & FOPEN_DIRECT_IO)) + ff->open_flags &= ~FOPEN_PARALLEL_DIRECT_WRITES; + + /* + * First passthrough file open denies caching inode io mode. + * First caching file open enters caching inode io mode. + * + * Note that if user opens a file open with O_DIRECT, but server did + * not specify FOPEN_DIRECT_IO, a later fcntl() could remove O_DIRECT, + * so we put the inode in caching mode to prevent parallel dio. + */ + if ((ff->open_flags & FOPEN_DIRECT_IO) && + !(ff->open_flags & FOPEN_PASSTHROUGH)) + return 0; + + if (ff->open_flags & FOPEN_PASSTHROUGH) + err = fuse_file_passthrough_open(inode, file); + else + err = fuse_file_cached_io_start(inode, ff); + if (err) + goto fail; + + return 0; + +fail: + pr_debug("failed to open file in requested io mode (open_flags=0x%x, err=%i).\n", + ff->open_flags, err); + /* + * The file open mode determines the inode io mode. + * Using incorrect open mode is a server mistake, which results in + * user visible failure of open() with EIO error. + */ + return -EIO; +} + +/* No more pending io and no new io possible to inode via open/mmapped file */ +void fuse_file_io_release(struct fuse_file *ff, struct inode *inode) +{ + /* + * Last parallel dio close allows caching inode io mode. + * Last caching file close exits caching inode io mode. + */ + switch (ff->iomode) { + case IOM_NONE: + /* Nothing to do */ + break; + case IOM_UNCACHED: + fuse_file_uncached_io_end(inode, ff); + break; + case IOM_CACHED: + fuse_file_cached_io_end(inode, ff); + break; + } +} diff --git a/fs/fuse/passthrough.c b/fs/fuse/passthrough.c new file mode 100644 index 000000000000..1567f0323858 --- /dev/null +++ b/fs/fuse/passthrough.c @@ -0,0 +1,355 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * FUSE passthrough to backing file. + * + * Copyright (c) 2023 CTERA Networks. + */ + +#include "fuse_i.h" + +#include <linux/file.h> +#include <linux/backing-file.h> +#include <linux/splice.h> + +static void fuse_file_accessed(struct file *file) +{ + struct inode *inode = file_inode(file); + + fuse_invalidate_atime(inode); +} + +static void fuse_file_modified(struct file *file) +{ + struct inode *inode = file_inode(file); + + fuse_invalidate_attr_mask(inode, FUSE_STATX_MODSIZE); +} + +ssize_t fuse_passthrough_read_iter(struct kiocb *iocb, struct iov_iter *iter) +{ + struct file *file = iocb->ki_filp; + struct fuse_file *ff = file->private_data; + struct file *backing_file = fuse_file_passthrough(ff); + size_t count = iov_iter_count(iter); + ssize_t ret; + struct backing_file_ctx ctx = { + .cred = ff->cred, + .user_file = file, + .accessed = fuse_file_accessed, + }; + + + pr_debug("%s: backing_file=0x%p, pos=%lld, len=%zu\n", __func__, + backing_file, iocb->ki_pos, count); + + if (!count) + return 0; + + ret = backing_file_read_iter(backing_file, iter, iocb, iocb->ki_flags, + &ctx); + + return ret; +} + +ssize_t fuse_passthrough_write_iter(struct kiocb *iocb, + struct iov_iter *iter) +{ + struct file *file = iocb->ki_filp; + struct inode *inode = file_inode(file); + struct fuse_file *ff = file->private_data; + struct file *backing_file = fuse_file_passthrough(ff); + size_t count = iov_iter_count(iter); + ssize_t ret; + struct backing_file_ctx ctx = { + .cred = ff->cred, + .user_file = file, + .end_write = fuse_file_modified, + }; + + pr_debug("%s: backing_file=0x%p, pos=%lld, len=%zu\n", __func__, + backing_file, iocb->ki_pos, count); + + if (!count) + return 0; + + inode_lock(inode); + ret = backing_file_write_iter(backing_file, iter, iocb, iocb->ki_flags, + &ctx); + inode_unlock(inode); + + return ret; +} + +ssize_t fuse_passthrough_splice_read(struct file *in, loff_t *ppos, + struct pipe_inode_info *pipe, + size_t len, unsigned int flags) +{ + struct fuse_file *ff = in->private_data; + struct file *backing_file = fuse_file_passthrough(ff); + struct backing_file_ctx ctx = { + .cred = ff->cred, + .user_file = in, + .accessed = fuse_file_accessed, + }; + + pr_debug("%s: backing_file=0x%p, pos=%lld, len=%zu, flags=0x%x\n", __func__, + backing_file, ppos ? *ppos : 0, len, flags); + + return backing_file_splice_read(backing_file, ppos, pipe, len, flags, + &ctx); +} + +ssize_t fuse_passthrough_splice_write(struct pipe_inode_info *pipe, + struct file *out, loff_t *ppos, + size_t len, unsigned int flags) +{ + struct fuse_file *ff = out->private_data; + struct file *backing_file = fuse_file_passthrough(ff); + struct inode *inode = file_inode(out); + ssize_t ret; + struct backing_file_ctx ctx = { + .cred = ff->cred, + .user_file = out, + .end_write = fuse_file_modified, + }; + + pr_debug("%s: backing_file=0x%p, pos=%lld, len=%zu, flags=0x%x\n", __func__, + backing_file, ppos ? *ppos : 0, len, flags); + + inode_lock(inode); + ret = backing_file_splice_write(pipe, backing_file, ppos, len, flags, + &ctx); + inode_unlock(inode); + + return ret; +} + +ssize_t fuse_passthrough_mmap(struct file *file, struct vm_area_struct *vma) +{ + struct fuse_file *ff = file->private_data; + struct file *backing_file = fuse_file_passthrough(ff); + struct backing_file_ctx ctx = { + .cred = ff->cred, + .user_file = file, + .accessed = fuse_file_accessed, + }; + + pr_debug("%s: backing_file=0x%p, start=%lu, end=%lu\n", __func__, + backing_file, vma->vm_start, vma->vm_end); + + return backing_file_mmap(backing_file, vma, &ctx); +} + +struct fuse_backing *fuse_backing_get(struct fuse_backing *fb) +{ + if (fb && refcount_inc_not_zero(&fb->count)) + return fb; + return NULL; +} + +static void fuse_backing_free(struct fuse_backing *fb) +{ + pr_debug("%s: fb=0x%p\n", __func__, fb); + + if (fb->file) + fput(fb->file); + put_cred(fb->cred); + kfree_rcu(fb, rcu); +} + +void fuse_backing_put(struct fuse_backing *fb) +{ + if (fb && refcount_dec_and_test(&fb->count)) + fuse_backing_free(fb); +} + +void fuse_backing_files_init(struct fuse_conn *fc) +{ + idr_init(&fc->backing_files_map); +} + +static int fuse_backing_id_alloc(struct fuse_conn *fc, struct fuse_backing *fb) +{ + int id; + + idr_preload(GFP_KERNEL); + spin_lock(&fc->lock); + /* FIXME: xarray might be space inefficient */ + id = idr_alloc_cyclic(&fc->backing_files_map, fb, 1, 0, GFP_ATOMIC); + spin_unlock(&fc->lock); + idr_preload_end(); + + WARN_ON_ONCE(id == 0); + return id; +} + +static struct fuse_backing *fuse_backing_id_remove(struct fuse_conn *fc, + int id) +{ + struct fuse_backing *fb; + + spin_lock(&fc->lock); + fb = idr_remove(&fc->backing_files_map, id); + spin_unlock(&fc->lock); + + return fb; +} + +static int fuse_backing_id_free(int id, void *p, void *data) +{ + struct fuse_backing *fb = p; + + WARN_ON_ONCE(refcount_read(&fb->count) != 1); + fuse_backing_free(fb); + return 0; +} + +void fuse_backing_files_free(struct fuse_conn *fc) +{ + idr_for_each(&fc->backing_files_map, fuse_backing_id_free, NULL); + idr_destroy(&fc->backing_files_map); +} + +int fuse_backing_open(struct fuse_conn *fc, struct fuse_backing_map *map) +{ + struct file *file; + struct super_block *backing_sb; + struct fuse_backing *fb = NULL; + int res; + + pr_debug("%s: fd=%d flags=0x%x\n", __func__, map->fd, map->flags); + + /* TODO: relax CAP_SYS_ADMIN once backing files are visible to lsof */ + res = -EPERM; + if (!fc->passthrough || !capable(CAP_SYS_ADMIN)) + goto out; + + res = -EINVAL; + if (map->flags) + goto out; + + file = fget(map->fd); + res = -EBADF; + if (!file) + goto out; + + res = -EOPNOTSUPP; + if (!file->f_op->read_iter || !file->f_op->write_iter) + goto out_fput; + + backing_sb = file_inode(file)->i_sb; + res = -ELOOP; + if (backing_sb->s_stack_depth >= fc->max_stack_depth) + goto out_fput; + + fb = kmalloc(sizeof(struct fuse_backing), GFP_KERNEL); + res = -ENOMEM; + if (!fb) + goto out_fput; + + fb->file = file; + fb->cred = prepare_creds(); + refcount_set(&fb->count, 1); + + res = fuse_backing_id_alloc(fc, fb); + if (res < 0) { + fuse_backing_free(fb); + fb = NULL; + } + +out: + pr_debug("%s: fb=0x%p, ret=%i\n", __func__, fb, res); + + return res; + +out_fput: + fput(file); + goto out; +} + +int fuse_backing_close(struct fuse_conn *fc, int backing_id) +{ + struct fuse_backing *fb = NULL; + int err; + + pr_debug("%s: backing_id=%d\n", __func__, backing_id); + + /* TODO: relax CAP_SYS_ADMIN once backing files are visible to lsof */ + err = -EPERM; + if (!fc->passthrough || !capable(CAP_SYS_ADMIN)) + goto out; + + err = -EINVAL; + if (backing_id <= 0) + goto out; + + err = -ENOENT; + fb = fuse_backing_id_remove(fc, backing_id); + if (!fb) + goto out; + + fuse_backing_put(fb); + err = 0; +out: + pr_debug("%s: fb=0x%p, err=%i\n", __func__, fb, err); + + return err; +} + +/* + * Setup passthrough to a backing file. + * + * Returns an fb object with elevated refcount to be stored in fuse inode. + */ +struct fuse_backing *fuse_passthrough_open(struct file *file, + struct inode *inode, + int backing_id) +{ + struct fuse_file *ff = file->private_data; + struct fuse_conn *fc = ff->fm->fc; + struct fuse_backing *fb = NULL; + struct file *backing_file; + int err; + + err = -EINVAL; + if (backing_id <= 0) + goto out; + + rcu_read_lock(); + fb = idr_find(&fc->backing_files_map, backing_id); + fb = fuse_backing_get(fb); + rcu_read_unlock(); + + err = -ENOENT; + if (!fb) + goto out; + + /* Allocate backing file per fuse file to store fuse path */ + backing_file = backing_file_open(&file->f_path, file->f_flags, + &fb->file->f_path, fb->cred); + err = PTR_ERR(backing_file); + if (IS_ERR(backing_file)) { + fuse_backing_put(fb); + goto out; + } + + err = 0; + ff->passthrough = backing_file; + ff->cred = get_cred(fb->cred); +out: + pr_debug("%s: backing_id=%d, fb=0x%p, backing_file=0x%p, err=%i\n", __func__, + backing_id, fb, ff->passthrough, err); + + return err ? ERR_PTR(err) : fb; +} + +void fuse_passthrough_release(struct fuse_file *ff, struct fuse_backing *fb) +{ + pr_debug("%s: fb=0x%p, backing_file=0x%p\n", __func__, + fb, ff->passthrough); + + fput(ff->passthrough); + ff->passthrough = NULL; + put_cred(ff->cred); + ff->cred = NULL; +} diff --git a/fs/fuse/readdir.c b/fs/fuse/readdir.c index c66a54d6c7d3..0377b6dc24c8 100644 --- a/fs/fuse/readdir.c +++ b/fs/fuse/readdir.c @@ -592,15 +592,11 @@ int fuse_readdir(struct file *file, struct dir_context *ctx) if (fuse_is_bad(inode)) return -EIO; - mutex_lock(&ff->readdir.lock); - err = UNCACHED; if (ff->open_flags & FOPEN_CACHE_DIR) err = fuse_readdir_cached(file, ctx); if (err == UNCACHED) err = fuse_readdir_uncached(file, ctx); - mutex_unlock(&ff->readdir.lock); - return err; } diff --git a/fs/fuse/virtio_fs.c b/fs/fuse/virtio_fs.c index 5f1be1da92ce..322af827a232 100644 --- a/fs/fuse/virtio_fs.c +++ b/fs/fuse/virtio_fs.c @@ -16,6 +16,7 @@ #include <linux/fs_context.h> #include <linux/fs_parser.h> #include <linux/highmem.h> +#include <linux/cleanup.h> #include <linux/uio.h> #include "fuse_i.h" @@ -31,6 +32,9 @@ static DEFINE_MUTEX(virtio_fs_mutex); static LIST_HEAD(virtio_fs_instances); +/* The /sys/fs/virtio_fs/ kset */ +static struct kset *virtio_fs_kset; + enum { VQ_HIPRIO, VQ_REQUEST @@ -55,7 +59,7 @@ struct virtio_fs_vq { /* A virtio-fs device instance */ struct virtio_fs { - struct kref refcount; + struct kobject kobj; struct list_head list; /* on virtio_fs_instances */ char *tag; struct virtio_fs_vq *vqs; @@ -161,18 +165,40 @@ static inline void dec_in_flight_req(struct virtio_fs_vq *fsvq) complete(&fsvq->in_flight_zero); } -static void release_virtio_fs_obj(struct kref *ref) +static ssize_t tag_show(struct kobject *kobj, + struct kobj_attribute *attr, char *buf) +{ + struct virtio_fs *fs = container_of(kobj, struct virtio_fs, kobj); + + return sysfs_emit(buf, fs->tag); +} + +static struct kobj_attribute virtio_fs_tag_attr = __ATTR_RO(tag); + +static struct attribute *virtio_fs_attrs[] = { + &virtio_fs_tag_attr.attr, + NULL +}; +ATTRIBUTE_GROUPS(virtio_fs); + +static void virtio_fs_ktype_release(struct kobject *kobj) { - struct virtio_fs *vfs = container_of(ref, struct virtio_fs, refcount); + struct virtio_fs *vfs = container_of(kobj, struct virtio_fs, kobj); kfree(vfs->vqs); kfree(vfs); } +static const struct kobj_type virtio_fs_ktype = { + .release = virtio_fs_ktype_release, + .sysfs_ops = &kobj_sysfs_ops, + .default_groups = virtio_fs_groups, +}; + /* Make sure virtiofs_mutex is held */ static void virtio_fs_put(struct virtio_fs *fs) { - kref_put(&fs->refcount, release_virtio_fs_obj); + kobject_put(&fs->kobj); } static void virtio_fs_fiq_release(struct fuse_iqueue *fiq) @@ -243,25 +269,46 @@ static void virtio_fs_start_all_queues(struct virtio_fs *fs) } /* Add a new instance to the list or return -EEXIST if tag name exists*/ -static int virtio_fs_add_instance(struct virtio_fs *fs) +static int virtio_fs_add_instance(struct virtio_device *vdev, + struct virtio_fs *fs) { struct virtio_fs *fs2; - bool duplicate = false; + int ret; mutex_lock(&virtio_fs_mutex); list_for_each_entry(fs2, &virtio_fs_instances, list) { - if (strcmp(fs->tag, fs2->tag) == 0) - duplicate = true; + if (strcmp(fs->tag, fs2->tag) == 0) { + mutex_unlock(&virtio_fs_mutex); + return -EEXIST; + } } - if (!duplicate) - list_add_tail(&fs->list, &virtio_fs_instances); + /* Use the virtio_device's index as a unique identifier, there is no + * need to allocate our own identifiers because the virtio_fs instance + * is only visible to userspace as long as the underlying virtio_device + * exists. + */ + fs->kobj.kset = virtio_fs_kset; + ret = kobject_add(&fs->kobj, NULL, "%d", vdev->index); + if (ret < 0) { + mutex_unlock(&virtio_fs_mutex); + return ret; + } + + ret = sysfs_create_link(&fs->kobj, &vdev->dev.kobj, "device"); + if (ret < 0) { + kobject_del(&fs->kobj); + mutex_unlock(&virtio_fs_mutex); + return ret; + } + + list_add_tail(&fs->list, &virtio_fs_instances); mutex_unlock(&virtio_fs_mutex); - if (duplicate) - return -EEXIST; + kobject_uevent(&fs->kobj, KOBJ_ADD); + return 0; } @@ -274,7 +321,7 @@ static struct virtio_fs *virtio_fs_find_instance(const char *tag) list_for_each_entry(fs, &virtio_fs_instances, list) { if (strcmp(fs->tag, tag) == 0) { - kref_get(&fs->refcount); + kobject_get(&fs->kobj); goto found; } } @@ -323,6 +370,16 @@ static int virtio_fs_read_tag(struct virtio_device *vdev, struct virtio_fs *fs) return -ENOMEM; memcpy(fs->tag, tag_buf, len); fs->tag[len] = '\0'; + + /* While the VIRTIO specification allows any character, newlines are + * awkward on mount(8) command-lines and cause problems in the sysfs + * "tag" attr and uevent TAG= properties. Forbid them. + */ + if (strchr(fs->tag, '\n')) { + dev_dbg(&vdev->dev, "refusing virtiofs tag with newline character\n"); + return -EINVAL; + } + return 0; } @@ -345,7 +402,7 @@ static void virtio_fs_hiprio_done_work(struct work_struct *work) kfree(req); dec_in_flight_req(fsvq); } - } while (!virtqueue_enable_cb(vq) && likely(!virtqueue_is_broken(vq))); + } while (!virtqueue_enable_cb(vq)); spin_unlock(&fsvq->lock); } @@ -627,7 +684,7 @@ static void virtio_fs_requests_done_work(struct work_struct *work) list_move_tail(&req->list, &reqs); spin_unlock(&fpq->lock); } - } while (!virtqueue_enable_cb(vq) && likely(!virtqueue_is_broken(vq))); + } while (!virtqueue_enable_cb(vq)); spin_unlock(&fsvq->lock); /* End requests */ @@ -795,8 +852,11 @@ static void virtio_fs_cleanup_dax(void *data) put_dax(dax_dev); } +DEFINE_FREE(cleanup_dax, struct dax_dev *, if (!IS_ERR_OR_NULL(_T)) virtio_fs_cleanup_dax(_T)) + static int virtio_fs_setup_dax(struct virtio_device *vdev, struct virtio_fs *fs) { + struct dax_device *dax_dev __free(cleanup_dax) = NULL; struct virtio_shm_region cache_reg; struct dev_pagemap *pgmap; bool have_cache; @@ -804,6 +864,12 @@ static int virtio_fs_setup_dax(struct virtio_device *vdev, struct virtio_fs *fs) if (!IS_ENABLED(CONFIG_FUSE_DAX)) return 0; + dax_dev = alloc_dax(fs, &virtio_fs_dax_ops); + if (IS_ERR(dax_dev)) { + int rc = PTR_ERR(dax_dev); + return rc == -EOPNOTSUPP ? 0 : rc; + } + /* Get cache region */ have_cache = virtio_get_shm_region(vdev, &cache_reg, (u8)VIRTIO_FS_SHMCAP_ID_CACHE); @@ -849,10 +915,7 @@ static int virtio_fs_setup_dax(struct virtio_device *vdev, struct virtio_fs *fs) dev_dbg(&vdev->dev, "%s: window kaddr 0x%px phys_addr 0x%llx len 0x%llx\n", __func__, fs->window_kaddr, cache_reg.addr, cache_reg.len); - fs->dax_dev = alloc_dax(fs, &virtio_fs_dax_ops); - if (IS_ERR(fs->dax_dev)) - return PTR_ERR(fs->dax_dev); - + fs->dax_dev = no_free_ptr(dax_dev); return devm_add_action_or_reset(&vdev->dev, virtio_fs_cleanup_dax, fs->dax_dev); } @@ -865,7 +928,7 @@ static int virtio_fs_probe(struct virtio_device *vdev) fs = kzalloc(sizeof(*fs), GFP_KERNEL); if (!fs) return -ENOMEM; - kref_init(&fs->refcount); + kobject_init(&fs->kobj, &virtio_fs_ktype); vdev->priv = fs; ret = virtio_fs_read_tag(vdev, fs); @@ -887,7 +950,7 @@ static int virtio_fs_probe(struct virtio_device *vdev) */ virtio_device_ready(vdev); - ret = virtio_fs_add_instance(fs); + ret = virtio_fs_add_instance(vdev, fs); if (ret < 0) goto out_vqs; @@ -896,11 +959,10 @@ static int virtio_fs_probe(struct virtio_device *vdev) out_vqs: virtio_reset_device(vdev); virtio_fs_cleanup_vqs(vdev); - kfree(fs->vqs); out: vdev->priv = NULL; - kfree(fs); + kobject_put(&fs->kobj); return ret; } @@ -924,6 +986,8 @@ static void virtio_fs_remove(struct virtio_device *vdev) mutex_lock(&virtio_fs_mutex); /* This device is going away. No one should get new reference */ list_del_init(&fs->list); + sysfs_remove_link(&fs->kobj, "device"); + kobject_del(&fs->kobj); virtio_fs_stop_all_queues(fs); virtio_fs_drain_all_queues_locked(fs); virtio_reset_device(vdev); @@ -1510,21 +1574,56 @@ static struct file_system_type virtio_fs_type = { .kill_sb = virtio_kill_sb, }; +static int virtio_fs_uevent(const struct kobject *kobj, struct kobj_uevent_env *env) +{ + const struct virtio_fs *fs = container_of(kobj, struct virtio_fs, kobj); + + add_uevent_var(env, "TAG=%s", fs->tag); + return 0; +} + +static const struct kset_uevent_ops virtio_fs_uevent_ops = { + .uevent = virtio_fs_uevent, +}; + +static int __init virtio_fs_sysfs_init(void) +{ + virtio_fs_kset = kset_create_and_add("virtiofs", &virtio_fs_uevent_ops, + fs_kobj); + if (!virtio_fs_kset) + return -ENOMEM; + return 0; +} + +static void virtio_fs_sysfs_exit(void) +{ + kset_unregister(virtio_fs_kset); + virtio_fs_kset = NULL; +} + static int __init virtio_fs_init(void) { int ret; - ret = register_virtio_driver(&virtio_fs_driver); + ret = virtio_fs_sysfs_init(); if (ret < 0) return ret; + ret = register_virtio_driver(&virtio_fs_driver); + if (ret < 0) + goto sysfs_exit; + ret = register_filesystem(&virtio_fs_type); - if (ret < 0) { - unregister_virtio_driver(&virtio_fs_driver); - return ret; - } + if (ret < 0) + goto unregister_virtio_driver; return 0; + +unregister_virtio_driver: + unregister_virtio_driver(&virtio_fs_driver); +sysfs_exit: + virtio_fs_sysfs_exit(); + return ret; } module_init(virtio_fs_init); @@ -1532,6 +1631,7 @@ static void __exit virtio_fs_exit(void) { unregister_filesystem(&virtio_fs_type); unregister_virtio_driver(&virtio_fs_driver); + virtio_fs_sysfs_exit(); } module_exit(virtio_fs_exit); diff --git a/fs/inode.c b/fs/inode.c index d290f007b3d1..3a41f83a4ba5 100644 --- a/fs/inode.c +++ b/fs/inode.c @@ -2033,7 +2033,7 @@ static int __remove_privs(struct mnt_idmap *idmap, return notify_change(idmap, dentry, &newattrs, NULL); } -static int __file_remove_privs(struct file *file, unsigned int flags) +int file_remove_privs_flags(struct file *file, unsigned int flags) { struct dentry *dentry = file_dentry(file); struct inode *inode = file_inode(file); @@ -2058,6 +2058,7 @@ static int __file_remove_privs(struct file *file, unsigned int flags) inode_has_no_xattr(inode); return error; } +EXPORT_SYMBOL_GPL(file_remove_privs_flags); /** * file_remove_privs - remove special file privileges (suid, capabilities) @@ -2070,7 +2071,7 @@ static int __file_remove_privs(struct file *file, unsigned int flags) */ int file_remove_privs(struct file *file) { - return __file_remove_privs(file, 0); + return file_remove_privs_flags(file, 0); } EXPORT_SYMBOL(file_remove_privs); @@ -2163,7 +2164,7 @@ static int file_modified_flags(struct file *file, int flags) * Clear the security bits if the process is not being run by root. * This keeps people from modifying setuid and setgid binaries. */ - ret = __file_remove_privs(file, flags); + ret = file_remove_privs_flags(file, flags); if (ret) return ret; diff --git a/fs/isofs/inode.c b/fs/isofs/inode.c index 25fca44149dd..2a616a9f289d 100644 --- a/fs/isofs/inode.c +++ b/fs/isofs/inode.c @@ -908,8 +908,22 @@ root_found: * we then decide whether to use the Joliet descriptor. */ inode = isofs_iget(s, sbi->s_firstdatazone, 0); - if (IS_ERR(inode)) - goto out_no_root; + + /* + * Fix for broken CDs with a corrupt root inode but a correct Joliet + * root directory. + */ + if (IS_ERR(inode)) { + if (joliet_level && sbi->s_firstdatazone != first_data_zone) { + printk(KERN_NOTICE + "ISOFS: root inode is unusable. " + "Disabling Rock Ridge and switching to Joliet."); + sbi->s_rock = 0; + inode = NULL; + } else { + goto out_no_root; + } + } /* * Fix for broken CDs with Rock Ridge and empty ISO root directory but diff --git a/fs/jfs/jfs_incore.h b/fs/jfs/jfs_incore.h index dd4264aa9bed..10934f9a11be 100644 --- a/fs/jfs/jfs_incore.h +++ b/fs/jfs/jfs_incore.h @@ -92,7 +92,7 @@ struct jfs_inode_info { } link; } u; #ifdef CONFIG_QUOTA - struct dquot *i_dquot[MAXQUOTAS]; + struct dquot __rcu *i_dquot[MAXQUOTAS]; #endif u32 dev; /* will die when we get wide dev_t */ struct inode vfs_inode; diff --git a/fs/jfs/super.c b/fs/jfs/super.c index 73f09a762b79..e1be21ca5d6e 100644 --- a/fs/jfs/super.c +++ b/fs/jfs/super.c @@ -824,7 +824,7 @@ out: return len - towrite; } -static struct dquot **jfs_get_dquots(struct inode *inode) +static struct dquot __rcu **jfs_get_dquots(struct inode *inode) { return JFS_IP(inode)->i_dquot; } diff --git a/fs/kernfs/dir.c b/fs/kernfs/dir.c index bce1d7ac95ca..458519e416fe 100644 --- a/fs/kernfs/dir.c +++ b/fs/kernfs/dir.c @@ -529,6 +529,20 @@ void kernfs_get(struct kernfs_node *kn) } EXPORT_SYMBOL_GPL(kernfs_get); +static void kernfs_free_rcu(struct rcu_head *rcu) +{ + struct kernfs_node *kn = container_of(rcu, struct kernfs_node, rcu); + + kfree_const(kn->name); + + if (kn->iattr) { + simple_xattrs_free(&kn->iattr->xattrs, NULL); + kmem_cache_free(kernfs_iattrs_cache, kn->iattr); + } + + kmem_cache_free(kernfs_node_cache, kn); +} + /** * kernfs_put - put a reference count on a kernfs_node * @kn: the target kernfs_node @@ -557,16 +571,11 @@ void kernfs_put(struct kernfs_node *kn) if (kernfs_type(kn) == KERNFS_LINK) kernfs_put(kn->symlink.target_kn); - kfree_const(kn->name); - - if (kn->iattr) { - simple_xattrs_free(&kn->iattr->xattrs, NULL); - kmem_cache_free(kernfs_iattrs_cache, kn->iattr); - } spin_lock(&kernfs_idr_lock); idr_remove(&root->ino_idr, (u32)kernfs_ino(kn)); spin_unlock(&kernfs_idr_lock); - kmem_cache_free(kernfs_node_cache, kn); + + call_rcu(&kn->rcu, kernfs_free_rcu); kn = parent; if (kn) { @@ -575,7 +584,7 @@ void kernfs_put(struct kernfs_node *kn) } else { /* just released the root kn, free @root too */ idr_destroy(&root->ino_idr); - kfree(root); + kfree_rcu(root, rcu); } } EXPORT_SYMBOL_GPL(kernfs_put); @@ -715,7 +724,7 @@ struct kernfs_node *kernfs_find_and_get_node_by_id(struct kernfs_root *root, ino_t ino = kernfs_id_ino(id); u32 gen = kernfs_id_gen(id); - spin_lock(&kernfs_idr_lock); + rcu_read_lock(); kn = idr_find(&root->ino_idr, (u32)ino); if (!kn) @@ -739,10 +748,10 @@ struct kernfs_node *kernfs_find_and_get_node_by_id(struct kernfs_root *root, if (unlikely(!__kernfs_active(kn) || !atomic_inc_not_zero(&kn->count))) goto err_unlock; - spin_unlock(&kernfs_idr_lock); + rcu_read_unlock(); return kn; err_unlock: - spin_unlock(&kernfs_idr_lock); + rcu_read_unlock(); return NULL; } diff --git a/fs/kernfs/file.c b/fs/kernfs/file.c index ffa4565c275a..e9df2f87072c 100644 --- a/fs/kernfs/file.c +++ b/fs/kernfs/file.c @@ -483,9 +483,11 @@ static int kernfs_fop_mmap(struct file *file, struct vm_area_struct *vma) goto out_put; rc = 0; - of->mmapped = true; - of_on(of)->nr_mmapped++; - of->vm_ops = vma->vm_ops; + if (!of->mmapped) { + of->mmapped = true; + of_on(of)->nr_mmapped++; + of->vm_ops = vma->vm_ops; + } vma->vm_ops = &kernfs_vm_ops; out_put: kernfs_put_active(of->kn); diff --git a/fs/kernfs/kernfs-internal.h b/fs/kernfs/kernfs-internal.h index 237f2764b941..b42ee6547cdc 100644 --- a/fs/kernfs/kernfs-internal.h +++ b/fs/kernfs/kernfs-internal.h @@ -49,6 +49,8 @@ struct kernfs_root { struct rw_semaphore kernfs_rwsem; struct rw_semaphore kernfs_iattr_rwsem; struct rw_semaphore kernfs_supers_rwsem; + + struct rcu_head rcu; }; /* +1 to avoid triggering overflow warning when negating it */ diff --git a/fs/netfs/fscache_io.c b/fs/netfs/fscache_io.c index ad572f7ee897..43a651ed8264 100644 --- a/fs/netfs/fscache_io.c +++ b/fs/netfs/fscache_io.c @@ -83,8 +83,10 @@ static int fscache_begin_operation(struct netfs_cache_resources *cres, cres->debug_id = cookie->debug_id; cres->inval_counter = cookie->inval_counter; - if (!fscache_begin_cookie_access(cookie, why)) + if (!fscache_begin_cookie_access(cookie, why)) { + cres->cache_priv = NULL; return -ENOBUFS; + } again: spin_lock(&cookie->lock); diff --git a/fs/nfs/client.c b/fs/nfs/client.c index fbdc9ca80f71..de77848ae654 100644 --- a/fs/nfs/client.c +++ b/fs/nfs/client.c @@ -73,14 +73,9 @@ const struct rpc_program nfs_program = { .number = NFS_PROGRAM, .nrvers = ARRAY_SIZE(nfs_version), .version = nfs_version, - .stats = &nfs_rpcstat, .pipe_dir_name = NFS_PIPE_DIRNAME, }; -struct rpc_stat nfs_rpcstat = { - .program = &nfs_program -}; - static struct nfs_subversion *find_nfs_version(unsigned int version) { struct nfs_subversion *nfs; @@ -502,6 +497,7 @@ int nfs_create_rpc_client(struct nfs_client *clp, const struct nfs_client_initdata *cl_init, rpc_authflavor_t flavor) { + struct nfs_net *nn = net_generic(clp->cl_net, nfs_net_id); struct rpc_clnt *clnt = NULL; struct rpc_create_args args = { .net = clp->cl_net, @@ -513,6 +509,7 @@ int nfs_create_rpc_client(struct nfs_client *clp, .servername = clp->cl_hostname, .nodename = cl_init->nodename, .program = &nfs_program, + .stats = &nn->rpcstats, .version = clp->rpc_ops->version, .authflavor = flavor, .cred = cl_init->cred, @@ -1182,6 +1179,8 @@ void nfs_clients_init(struct net *net) #endif spin_lock_init(&nn->nfs_client_lock); nn->boot_time = ktime_get_real(); + memset(&nn->rpcstats, 0, sizeof(nn->rpcstats)); + nn->rpcstats.program = &nfs_program; nfs_netns_sysfs_setup(nn, net); } diff --git a/fs/nfs/delegation.c b/fs/nfs/delegation.c index d4a42ce0c7e3..6bace5fece04 100644 --- a/fs/nfs/delegation.c +++ b/fs/nfs/delegation.c @@ -181,7 +181,6 @@ static int nfs_delegation_claim_opens(struct inode *inode, struct nfs_open_context *ctx; struct nfs4_state_owner *sp; struct nfs4_state *state; - unsigned int seq; int err; again: @@ -202,12 +201,9 @@ again: sp = state->owner; /* Block nfs4_proc_unlck */ mutex_lock(&sp->so_delegreturn_mutex); - seq = raw_seqcount_begin(&sp->so_reclaim_seqcount); err = nfs4_open_delegation_recall(ctx, state, stateid); if (!err) err = nfs_delegation_claim_locks(state, stateid); - if (!err && read_seqcount_retry(&sp->so_reclaim_seqcount, seq)) - err = -EAGAIN; mutex_unlock(&sp->so_delegreturn_mutex); put_nfs_open_context(ctx); if (err != 0) diff --git a/fs/nfs/direct.c b/fs/nfs/direct.c index 7af5d270de28..bb2f583eb28b 100644 --- a/fs/nfs/direct.c +++ b/fs/nfs/direct.c @@ -606,6 +606,7 @@ static void nfs_direct_commit_complete(struct nfs_commit_data *data) trace_nfs_direct_commit_complete(dreq); + spin_lock(&dreq->lock); if (status < 0) { /* Errors in commit are fatal */ dreq->error = status; @@ -613,6 +614,7 @@ static void nfs_direct_commit_complete(struct nfs_commit_data *data) } else { status = dreq->error; } + spin_unlock(&dreq->lock); nfs_init_cinfo_from_dreq(&cinfo, dreq); @@ -625,7 +627,10 @@ static void nfs_direct_commit_complete(struct nfs_commit_data *data) spin_unlock(&dreq->lock); nfs_release_request(req); } else if (!nfs_write_match_verf(verf, req)) { - dreq->flags = NFS_ODIRECT_RESCHED_WRITES; + spin_lock(&dreq->lock); + if (dreq->flags == 0) + dreq->flags = NFS_ODIRECT_RESCHED_WRITES; + spin_unlock(&dreq->lock); /* * Despite the reboot, the write was successful, * so reset wb_nio. @@ -667,10 +672,17 @@ static void nfs_direct_commit_schedule(struct nfs_direct_req *dreq) LIST_HEAD(mds_list); nfs_init_cinfo_from_dreq(&cinfo, dreq); + nfs_commit_begin(cinfo.mds); nfs_scan_commit(dreq->inode, &mds_list, &cinfo); res = nfs_generic_commit_list(dreq->inode, &mds_list, 0, &cinfo); - if (res < 0) /* res == -ENOMEM */ - nfs_direct_write_reschedule(dreq); + if (res < 0) { /* res == -ENOMEM */ + spin_lock(&dreq->lock); + if (dreq->flags == 0) + dreq->flags = NFS_ODIRECT_RESCHED_WRITES; + spin_unlock(&dreq->lock); + } + if (nfs_commit_end(cinfo.mds)) + nfs_direct_write_complete(dreq); } static void nfs_direct_write_clear_reqs(struct nfs_direct_req *dreq) diff --git a/fs/nfs/filelayout/filelayoutdev.c b/fs/nfs/filelayout/filelayoutdev.c index acf4b88889dc..4fa304fa5bc4 100644 --- a/fs/nfs/filelayout/filelayoutdev.c +++ b/fs/nfs/filelayout/filelayoutdev.c @@ -35,6 +35,7 @@ #include "../internal.h" #include "../nfs4session.h" #include "filelayout.h" +#include "../nfs4trace.h" #define NFSDBG_FACILITY NFSDBG_PNFS_LD @@ -172,6 +173,7 @@ nfs4_fl_alloc_deviceid_node(struct nfs_server *server, struct pnfs_device *pdev, dsaddr->ds_list[i] = nfs4_pnfs_ds_add(&dsaddrs, gfp_flags); if (!dsaddr->ds_list[i]) goto out_err_drain_dsaddrs; + trace_fl_getdevinfo(server, &pdev->dev_id, dsaddr->ds_list[i]->ds_remotestr); /* If DS was already in cache, free ds addrs */ while (!list_empty(&dsaddrs)) { diff --git a/fs/nfs/flexfilelayout/flexfilelayout.c b/fs/nfs/flexfilelayout/flexfilelayout.c index ef817a0475ff..3e724cb7ef01 100644 --- a/fs/nfs/flexfilelayout/flexfilelayout.c +++ b/fs/nfs/flexfilelayout/flexfilelayout.c @@ -2016,7 +2016,7 @@ static void ff_layout_cancel_io(struct pnfs_layout_segment *lseg) for (idx = 0; idx < flseg->mirror_array_cnt; idx++) { mirror = flseg->mirror_array[idx]; mirror_ds = mirror->mirror_ds; - if (!mirror_ds) + if (IS_ERR_OR_NULL(mirror_ds)) continue; ds = mirror->mirror_ds->ds; if (!ds) diff --git a/fs/nfs/fs_context.c b/fs/nfs/fs_context.c index 853e8d609bb3..d0a0956f8a13 100644 --- a/fs/nfs/fs_context.c +++ b/fs/nfs/fs_context.c @@ -652,6 +652,7 @@ static int nfs_fs_context_parse_param(struct fs_context *fc, ctx->fscache_uniq = NULL; break; case Opt_fscache: + trace_nfs_mount_assign(param->key, param->string); ctx->options |= NFS_OPTION_FSCACHE; kfree(ctx->fscache_uniq); ctx->fscache_uniq = param->string; diff --git a/fs/nfs/fscache.c b/fs/nfs/fscache.c index 2d1bfee225c3..ddc1ee031955 100644 --- a/fs/nfs/fscache.c +++ b/fs/nfs/fscache.c @@ -301,11 +301,11 @@ static void nfs_netfs_issue_read(struct netfs_io_subrequest *sreq) struct inode *inode = sreq->rreq->inode; struct nfs_open_context *ctx = sreq->rreq->netfs_priv; struct page *page; + unsigned long idx; int err; pgoff_t start = (sreq->start + sreq->transferred) >> PAGE_SHIFT; pgoff_t last = ((sreq->start + sreq->len - sreq->transferred - 1) >> PAGE_SHIFT); - XA_STATE(xas, &sreq->rreq->mapping->i_pages, start); nfs_pageio_init_read(&pgio, inode, false, &nfs_async_read_completion_ops); @@ -316,19 +316,14 @@ static void nfs_netfs_issue_read(struct netfs_io_subrequest *sreq) pgio.pg_netfs = netfs; /* used in completion */ - xas_lock(&xas); - xas_for_each(&xas, page, last) { + xa_for_each_range(&sreq->rreq->mapping->i_pages, idx, page, start, last) { /* nfs_read_add_folio() may schedule() due to pNFS layout and other RPCs */ - xas_pause(&xas); - xas_unlock(&xas); err = nfs_read_add_folio(&pgio, ctx, page_folio(page)); if (err < 0) { netfs->error = err; goto out; } - xas_lock(&xas); } - xas_unlock(&xas); out: nfs_pageio_complete_read(&pgio); nfs_netfs_put(netfs); diff --git a/fs/nfs/inode.c b/fs/nfs/inode.c index 93ea49a7eb61..c709c296ea9a 100644 --- a/fs/nfs/inode.c +++ b/fs/nfs/inode.c @@ -2426,12 +2426,16 @@ EXPORT_SYMBOL_GPL(nfs_net_id); static int nfs_net_init(struct net *net) { + struct nfs_net *nn = net_generic(net, nfs_net_id); + nfs_clients_init(net); + rpc_proc_register(net, &nn->rpcstats); return nfs_fs_proc_net_init(net); } static void nfs_net_exit(struct net *net) { + rpc_proc_unregister(net, "nfs"); nfs_fs_proc_net_exit(net); nfs_clients_exit(net); } @@ -2486,15 +2490,12 @@ static int __init init_nfs_fs(void) if (err) goto out1; - rpc_proc_register(&init_net, &nfs_rpcstat); - err = register_nfs_fs(); if (err) goto out0; return 0; out0: - rpc_proc_unregister(&init_net, "nfs"); nfs_destroy_directcache(); out1: nfs_destroy_writepagecache(); @@ -2524,7 +2525,6 @@ static void __exit exit_nfs_fs(void) nfs_destroy_inodecache(); nfs_destroy_nfspagecache(); unregister_pernet_subsys(&nfs_net_ops); - rpc_proc_unregister(&init_net, "nfs"); unregister_nfs_fs(); nfs_fs_proc_exit(); nfsiod_stop(); diff --git a/fs/nfs/internal.h b/fs/nfs/internal.h index e3722ce6722e..06253695fe53 100644 --- a/fs/nfs/internal.h +++ b/fs/nfs/internal.h @@ -449,8 +449,6 @@ int nfs_try_get_tree(struct fs_context *); int nfs_get_tree_common(struct fs_context *); void nfs_kill_super(struct super_block *); -extern struct rpc_stat nfs_rpcstat; - extern int __init register_nfs_fs(void); extern void __exit unregister_nfs_fs(void); extern bool nfs_sb_active(struct super_block *sb); diff --git a/fs/nfs/netns.h b/fs/nfs/netns.h index c8374f74dce1..a68b21603ea9 100644 --- a/fs/nfs/netns.h +++ b/fs/nfs/netns.h @@ -9,6 +9,7 @@ #include <linux/nfs4.h> #include <net/net_namespace.h> #include <net/netns/generic.h> +#include <linux/sunrpc/stats.h> struct bl_dev_msg { int32_t status; @@ -34,6 +35,7 @@ struct nfs_net { struct nfs_netns_client *nfs_client; spinlock_t nfs_client_lock; ktime_t boot_time; + struct rpc_stat rpcstats; #ifdef CONFIG_PROC_FS struct proc_dir_entry *proc_nfsfs; #endif diff --git a/fs/nfs/nfs3client.c b/fs/nfs/nfs3client.c index 674c012868b1..b0c8a39c2bbd 100644 --- a/fs/nfs/nfs3client.c +++ b/fs/nfs/nfs3client.c @@ -111,6 +111,7 @@ struct nfs_client *nfs3_set_ds_client(struct nfs_server *mds_srv, cl_init.hostname = buf; switch (ds_proto) { + case XPRT_TRANSPORT_RDMA: case XPRT_TRANSPORT_TCP: case XPRT_TRANSPORT_TCP_TLS: if (mds_clp->cl_nconnect > 1) diff --git a/fs/nfs/nfs42.h b/fs/nfs/nfs42.h index b59876b01a1e..0282d93c8bcc 100644 --- a/fs/nfs/nfs42.h +++ b/fs/nfs/nfs42.h @@ -55,11 +55,14 @@ int nfs42_proc_removexattr(struct inode *inode, const char *name); * They would be 7 bytes long in the eventual buffer ("user.x\0"), and * 8 bytes long XDR-encoded. * - * Include the trailing eof word as well. + * Include the trailing eof word as well and make the result a multiple + * of 4 bytes. */ static inline u32 nfs42_listxattr_xdrsize(u32 buflen) { - return ((buflen / (XATTR_USER_PREFIX_LEN + 2)) * 8) + 4; + u32 size = 8 * buflen / (XATTR_USER_PREFIX_LEN + 2) + 4; + + return (size + 3) & ~3; } #endif /* CONFIG_NFS_V4_2 */ #endif /* __LINUX_FS_NFS_NFS4_2_H */ diff --git a/fs/nfs/nfs4_fs.h b/fs/nfs/nfs4_fs.h index 6ff41ceb9f1c..7024230f0d1d 100644 --- a/fs/nfs/nfs4_fs.h +++ b/fs/nfs/nfs4_fs.h @@ -120,7 +120,6 @@ struct nfs4_state_owner { unsigned long so_flags; struct list_head so_states; struct nfs_seqid_counter so_seqid; - seqcount_spinlock_t so_reclaim_seqcount; struct mutex so_delegreturn_mutex; }; diff --git a/fs/nfs/nfs4client.c b/fs/nfs/nfs4client.c index 11e3a285594c..84573df5cf5a 100644 --- a/fs/nfs/nfs4client.c +++ b/fs/nfs/nfs4client.c @@ -924,6 +924,7 @@ static int nfs4_set_client(struct nfs_server *server, else cl_init.max_connect = max_connect; switch (proto) { + case XPRT_TRANSPORT_RDMA: case XPRT_TRANSPORT_TCP: case XPRT_TRANSPORT_TCP_TLS: cl_init.nconnect = nconnect; @@ -1000,6 +1001,7 @@ struct nfs_client *nfs4_set_ds_client(struct nfs_server *mds_srv, cl_init.hostname = buf; switch (ds_proto) { + case XPRT_TRANSPORT_RDMA: case XPRT_TRANSPORT_TCP: case XPRT_TRANSPORT_TCP_TLS: if (mds_clp->cl_nconnect > 1) { diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c index 815996cb27fc..ea390db94b62 100644 --- a/fs/nfs/nfs4proc.c +++ b/fs/nfs/nfs4proc.c @@ -3069,10 +3069,8 @@ static int _nfs4_open_and_get_state(struct nfs4_opendata *opendata, fmode_t acc_mode = _nfs4_ctx_to_accessmode(ctx); struct inode *dir = d_inode(opendata->dir); unsigned long dir_verifier; - unsigned int seq; int ret; - seq = raw_seqcount_begin(&sp->so_reclaim_seqcount); dir_verifier = nfs_save_change_attribute(dir); ret = _nfs4_proc_open(opendata, ctx); @@ -3125,11 +3123,8 @@ static int _nfs4_open_and_get_state(struct nfs4_opendata *opendata, if (ret != 0) goto out; - if (d_inode(dentry) == state->inode) { + if (d_inode(dentry) == state->inode) nfs_inode_attach_open_context(ctx); - if (read_seqcount_retry(&sp->so_reclaim_seqcount, seq)) - nfs4_schedule_stateid_recovery(server, state); - } out: if (!opendata->cancelled) { @@ -8973,10 +8968,12 @@ try_again: return; status = task->tk_status; - if (status == 0) + if (status == 0) { status = nfs4_detect_session_trunking(adata->clp, task->tk_msg.rpc_resp, xprt); - + trace_nfs4_trunked_exchange_id(adata->clp, + xprt->address_strings[RPC_DISPLAY_ADDR], status); + } if (status == 0) rpc_clnt_xprt_switch_add_xprt(clnt, xprt); else if (status != -NFS4ERR_DELAY && rpc_clnt_xprt_switch_has_addr(clnt, @@ -10618,29 +10615,33 @@ const struct nfs4_minor_version_ops *nfs_v4_minor_ops[] = { static ssize_t nfs4_listxattr(struct dentry *dentry, char *list, size_t size) { ssize_t error, error2, error3; + size_t left = size; - error = generic_listxattr(dentry, list, size); + error = generic_listxattr(dentry, list, left); if (error < 0) return error; if (list) { list += error; - size -= error; + left -= error; } - error2 = nfs4_listxattr_nfs4_label(d_inode(dentry), list, size); + error2 = nfs4_listxattr_nfs4_label(d_inode(dentry), list, left); if (error2 < 0) return error2; if (list) { list += error2; - size -= error2; + left -= error2; } - error3 = nfs4_listxattr_nfs4_user(d_inode(dentry), list, size); + error3 = nfs4_listxattr_nfs4_user(d_inode(dentry), list, left); if (error3 < 0) return error3; - return error + error2 + error3; + error += error2 + error3; + if (size && error > size) + return -ERANGE; + return error; } static void nfs4_enable_swap(struct inode *inode) diff --git a/fs/nfs/nfs4state.c b/fs/nfs/nfs4state.c index 8cfabdbda336..662e86ea3a2d 100644 --- a/fs/nfs/nfs4state.c +++ b/fs/nfs/nfs4state.c @@ -513,7 +513,6 @@ nfs4_alloc_state_owner(struct nfs_server *server, nfs4_init_seqid_counter(&sp->so_seqid); atomic_set(&sp->so_count, 1); INIT_LIST_HEAD(&sp->so_lru); - seqcount_spinlock_init(&sp->so_reclaim_seqcount, &sp->so_lock); mutex_init(&sp->so_delegreturn_mutex); return sp; } @@ -1667,7 +1666,6 @@ static int nfs4_reclaim_open_state(struct nfs4_state_owner *sp, * server that doesn't support a grace period. */ spin_lock(&sp->so_lock); - raw_write_seqcount_begin(&sp->so_reclaim_seqcount); restart: list_for_each_entry(state, &sp->so_states, open_states) { if (!test_and_clear_bit(ops->state_flag_bit, &state->flags)) @@ -1735,7 +1733,6 @@ restart: spin_lock(&sp->so_lock); goto restart; } - raw_write_seqcount_end(&sp->so_reclaim_seqcount); spin_unlock(&sp->so_lock); #ifdef CONFIG_NFS_V4_2 if (found_ssc_copy_state) @@ -1745,7 +1742,6 @@ restart: out_err: nfs4_put_open_state(state); spin_lock(&sp->so_lock); - raw_write_seqcount_end(&sp->so_reclaim_seqcount); spin_unlock(&sp->so_lock); return status; } @@ -1928,9 +1924,12 @@ static int nfs4_do_reclaim(struct nfs_client *clp, const struct nfs4_state_recov struct nfs_server *server; struct rb_node *pos; LIST_HEAD(freeme); - int status = 0; int lost_locks = 0; + int status; + status = nfs4_begin_drain_session(clp); + if (status < 0) + return status; restart: rcu_read_lock(); list_for_each_entry_rcu(server, &clp->cl_superblocks, client_link) { @@ -2694,6 +2693,9 @@ static void nfs4_state_manager(struct nfs_client *clp) /* Detect expired delegations... */ if (test_and_clear_bit(NFS4CLNT_DELEGATION_EXPIRED, &clp->cl_state)) { section = "detect expired delegations"; + status = nfs4_begin_drain_session(clp); + if (status < 0) + goto out_error; nfs_reap_expired_delegations(clp); continue; } diff --git a/fs/nfs/nfs4super.c b/fs/nfs/nfs4super.c index d09bcfd7db89..8da5a9c000f4 100644 --- a/fs/nfs/nfs4super.c +++ b/fs/nfs/nfs4super.c @@ -145,6 +145,7 @@ static int do_nfs4_mount(struct nfs_server *server, const char *export_path) { struct nfs_fs_context *root_ctx; + struct nfs_fs_context *ctx; struct fs_context *root_fc; struct vfsmount *root_mnt; struct dentry *dentry; @@ -157,6 +158,12 @@ static int do_nfs4_mount(struct nfs_server *server, .dirfd = -1, }; + struct fs_parameter param_fsc = { + .key = "fsc", + .type = fs_value_is_string, + .dirfd = -1, + }; + if (IS_ERR(server)) return PTR_ERR(server); @@ -168,9 +175,26 @@ static int do_nfs4_mount(struct nfs_server *server, kfree(root_fc->source); root_fc->source = NULL; + ctx = nfs_fc2context(fc); root_ctx = nfs_fc2context(root_fc); root_ctx->internal = true; root_ctx->server = server; + + if (ctx->fscache_uniq) { + len = strlen(ctx->fscache_uniq); + param_fsc.size = len; + param_fsc.string = kmemdup_nul(ctx->fscache_uniq, len, GFP_KERNEL); + if (param_fsc.string == NULL) { + put_fs_context(root_fc); + return -ENOMEM; + } + ret = vfs_parse_fs_param(root_fc, ¶m_fsc); + kfree(param_fsc.string); + if (ret < 0) { + put_fs_context(root_fc); + return ret; + } + } /* We leave export_path unset as it's not used to find the root. */ len = strlen(hostname) + 5; diff --git a/fs/nfs/nfs4trace.c b/fs/nfs/nfs4trace.c index d9ac556bebcf..d22c6670f770 100644 --- a/fs/nfs/nfs4trace.c +++ b/fs/nfs/nfs4trace.c @@ -28,4 +28,6 @@ EXPORT_TRACEPOINT_SYMBOL_GPL(pnfs_mds_fallback_write_pagelist); EXPORT_TRACEPOINT_SYMBOL_GPL(ff_layout_read_error); EXPORT_TRACEPOINT_SYMBOL_GPL(ff_layout_write_error); EXPORT_TRACEPOINT_SYMBOL_GPL(ff_layout_commit_error); + +EXPORT_TRACEPOINT_SYMBOL_GPL(fl_getdevinfo); #endif diff --git a/fs/nfs/nfs4trace.h b/fs/nfs/nfs4trace.h index fd7cb15b08b2..10985a4b8259 100644 --- a/fs/nfs/nfs4trace.h +++ b/fs/nfs/nfs4trace.h @@ -77,6 +77,36 @@ DEFINE_NFS4_CLIENTID_EVENT(nfs4_bind_conn_to_session); DEFINE_NFS4_CLIENTID_EVENT(nfs4_sequence); DEFINE_NFS4_CLIENTID_EVENT(nfs4_reclaim_complete); +TRACE_EVENT(nfs4_trunked_exchange_id, + TP_PROTO( + const struct nfs_client *clp, + const char *addr, + int error + ), + + TP_ARGS(clp, addr, error), + + TP_STRUCT__entry( + __string(main_addr, clp->cl_hostname) + __string(trunk_addr, addr) + __field(unsigned long, error) + ), + + TP_fast_assign( + __entry->error = error < 0 ? -error : 0; + __assign_str(main_addr, clp->cl_hostname); + __assign_str(trunk_addr, addr); + ), + + TP_printk( + "error=%ld (%s) main_addr=%s trunk_addr=%s", + -__entry->error, + show_nfs4_status(__entry->error), + __get_str(main_addr), + __get_str(trunk_addr) + ) +); + TRACE_EVENT(nfs4_sequence_done, TP_PROTO( const struct nfs4_session *session, @@ -1991,6 +2021,34 @@ DECLARE_EVENT_CLASS(nfs4_deviceid_status, DEFINE_PNFS_DEVICEID_STATUS(nfs4_getdeviceinfo); DEFINE_PNFS_DEVICEID_STATUS(nfs4_find_deviceid); +TRACE_EVENT(fl_getdevinfo, + TP_PROTO( + const struct nfs_server *server, + const struct nfs4_deviceid *deviceid, + char *ds_remotestr + ), + TP_ARGS(server, deviceid, ds_remotestr), + + TP_STRUCT__entry( + __string(mds_addr, server->nfs_client->cl_hostname) + __array(unsigned char, deviceid, NFS4_DEVICEID4_SIZE) + __string(ds_ips, ds_remotestr) + ), + + TP_fast_assign( + __assign_str(mds_addr, server->nfs_client->cl_hostname); + __assign_str(ds_ips, ds_remotestr); + memcpy(__entry->deviceid, deviceid->data, + NFS4_DEVICEID4_SIZE); + ), + TP_printk( + "deviceid=%s, mds_addr=%s, ds_ips=%s", + __print_hex(__entry->deviceid, NFS4_DEVICEID4_SIZE), + __get_str(mds_addr), + __get_str(ds_ips) + ) +); + DECLARE_EVENT_CLASS(nfs4_flexfiles_io_event, TP_PROTO( const struct nfs_pgio_header *hdr diff --git a/fs/nfs/nfsroot.c b/fs/nfs/nfsroot.c index 7600100ba26f..432612d22437 100644 --- a/fs/nfs/nfsroot.c +++ b/fs/nfs/nfsroot.c @@ -175,10 +175,10 @@ static int __init root_nfs_cat(char *dest, const char *src, size_t len = strlen(dest); if (len && dest[len - 1] != ',') - if (strlcat(dest, ",", destlen) > destlen) + if (strlcat(dest, ",", destlen) >= destlen) return -1; - if (strlcat(dest, src, destlen) > destlen) + if (strlcat(dest, src, destlen) >= destlen) return -1; return 0; } diff --git a/fs/nfs/pnfs.c b/fs/nfs/pnfs.c index 0c0fed1ecd0b..a5cc6199127f 100644 --- a/fs/nfs/pnfs.c +++ b/fs/nfs/pnfs.c @@ -1999,6 +1999,14 @@ pnfs_update_layout(struct inode *ino, } lookup_again: + if (!nfs4_valid_open_stateid(ctx->state)) { + trace_pnfs_update_layout(ino, pos, count, + iomode, lo, lseg, + PNFS_UPDATE_LAYOUT_INVALID_OPEN); + lseg = ERR_PTR(-EIO); + goto out; + } + lseg = ERR_PTR(nfs4_client_recover_expired_lease(clp)); if (IS_ERR(lseg)) goto out; diff --git a/fs/nfs/pnfs_nfs.c b/fs/nfs/pnfs_nfs.c index afd23910f3bf..88e061bd711b 100644 --- a/fs/nfs/pnfs_nfs.c +++ b/fs/nfs/pnfs_nfs.c @@ -919,6 +919,8 @@ static int _nfs4_pnfs_v4_ds_connect(struct nfs_server *mds_srv, dprintk("--> %s DS %s\n", __func__, ds->ds_remotestr); list_for_each_entry(da, &ds->ds_addrs, da_node) { + char servername[48]; + dprintk("%s: DS %s: trying address %s\n", __func__, ds->ds_remotestr, da->da_remotestr); @@ -929,6 +931,7 @@ static int _nfs4_pnfs_v4_ds_connect(struct nfs_server *mds_srv, .dstaddr = (struct sockaddr *)&da->da_addr, .addrlen = da->da_addrlen, .servername = clp->cl_hostname, + .xprtsec = clp->cl_xprtsec, }; struct nfs4_add_xprt_data xprtdata = { .clp = clp, @@ -938,10 +941,45 @@ static int _nfs4_pnfs_v4_ds_connect(struct nfs_server *mds_srv, .data = &xprtdata, }; - if (da->da_transport != clp->cl_proto) + if (da->da_transport != clp->cl_proto && + clp->cl_proto != XPRT_TRANSPORT_TCP_TLS) continue; + if (da->da_transport == XPRT_TRANSPORT_TCP && + mds_srv->nfs_client->cl_proto == + XPRT_TRANSPORT_TCP_TLS) { + struct sockaddr *addr = + (struct sockaddr *)&da->da_addr; + struct sockaddr_in *sin = + (struct sockaddr_in *)&da->da_addr; + struct sockaddr_in6 *sin6 = + (struct sockaddr_in6 *)&da->da_addr; + + /* for NFS with TLS we need to supply a correct + * servername of the trunked transport, not the + * servername of the main transport stored in + * clp->cl_hostname. And set the protocol to + * indicate to use TLS + */ + servername[0] = '\0'; + switch(addr->sa_family) { + case AF_INET: + snprintf(servername, sizeof(servername), + "%pI4", &sin->sin_addr.s_addr); + break; + case AF_INET6: + snprintf(servername, sizeof(servername), + "%pI6", &sin6->sin6_addr); + break; + default: + /* do not consider this address */ + continue; + } + xprt_args.ident = XPRT_TRANSPORT_TCP_TLS; + xprt_args.servername = servername; + } if (da->da_addr.ss_family != clp->cl_addr.ss_family) continue; + /** * Test this address for session trunking and * add as an alias @@ -953,6 +991,10 @@ static int _nfs4_pnfs_v4_ds_connect(struct nfs_server *mds_srv, if (xprtdata.cred) put_cred(xprtdata.cred); } else { + if (da->da_transport == XPRT_TRANSPORT_TCP && + mds_srv->nfs_client->cl_proto == + XPRT_TRANSPORT_TCP_TLS) + da->da_transport = XPRT_TRANSPORT_TCP_TLS; clp = nfs4_set_ds_client(mds_srv, &da->da_addr, da->da_addrlen, diff --git a/fs/nfs/read.c b/fs/nfs/read.c index 7dc21a48e3e7..a142287d86f6 100644 --- a/fs/nfs/read.c +++ b/fs/nfs/read.c @@ -305,6 +305,8 @@ int nfs_read_add_folio(struct nfs_pageio_descriptor *pgio, new = nfs_page_create_from_folio(ctx, folio, 0, aligned_len); if (IS_ERR(new)) { error = PTR_ERR(new); + if (nfs_netfs_folio_unlock(folio)) + folio_unlock(folio); goto out; } diff --git a/fs/nfs/super.c b/fs/nfs/super.c index 075b31c93f87..dc03f98f7616 100644 --- a/fs/nfs/super.c +++ b/fs/nfs/super.c @@ -516,8 +516,16 @@ static void nfs_show_mount_options(struct seq_file *m, struct nfs_server *nfss, else nfs_show_nfsv4_options(m, nfss, showdefaults); - if (nfss->options & NFS_OPTION_FSCACHE) + if (nfss->options & NFS_OPTION_FSCACHE) { +#ifdef CONFIG_NFS_FSCACHE + if (nfss->fscache_uniq) + seq_printf(m, ",fsc=%s", nfss->fscache_uniq); + else + seq_puts(m, ",fsc"); +#else seq_puts(m, ",fsc"); +#endif + } if (nfss->options & NFS_OPTION_MIGRATION) seq_puts(m, ",migration"); diff --git a/fs/nfs/write.c b/fs/nfs/write.c index 84bb85264572..5de85d725fb9 100644 --- a/fs/nfs/write.c +++ b/fs/nfs/write.c @@ -667,10 +667,6 @@ static int nfs_writepage_locked(struct folio *folio, struct inode *inode = folio_file_mapping(folio)->host; int err; - if (wbc->sync_mode == WB_SYNC_NONE && - NFS_SERVER(inode)->write_congested) - return AOP_WRITEPAGE_ACTIVATE; - nfs_inc_stats(inode, NFSIOS_VFSWRITEPAGE); nfs_pageio_init_write(&pgio, inode, 0, false, &nfs_async_write_completion_ops); @@ -1650,7 +1646,7 @@ static int wait_on_commit(struct nfs_mds_commit_info *cinfo) !atomic_read(&cinfo->rpcs_out)); } -static void nfs_commit_begin(struct nfs_mds_commit_info *cinfo) +void nfs_commit_begin(struct nfs_mds_commit_info *cinfo) { atomic_inc(&cinfo->rpcs_out); } diff --git a/fs/nfsd/trace.h b/fs/nfsd/trace.h index e545e92c4408..1cd2076210b1 100644 --- a/fs/nfsd/trace.h +++ b/fs/nfsd/trace.h @@ -104,7 +104,7 @@ TRACE_EVENT(nfsd_compound, TP_fast_assign( __entry->xid = be32_to_cpu(rqst->rq_xid); __entry->opcnt = opcnt; - __assign_str_len(tag, tag, taglen); + __assign_str(tag, tag); ), TP_printk("xid=0x%08x opcnt=%u tag=%s", __entry->xid, __entry->opcnt, __get_str(tag) @@ -485,7 +485,7 @@ TRACE_EVENT(nfsd_dirent, TP_fast_assign( __entry->fh_hash = fhp ? knfsd_fh_hash(&fhp->fh_handle) : 0; __entry->ino = ino; - __assign_str_len(name, name, namlen) + __assign_str(name, name); ), TP_printk("fh_hash=0x%08x ino=%llu name=%s", __entry->fh_hash, __entry->ino, __get_str(name) @@ -896,7 +896,7 @@ DECLARE_EVENT_CLASS(nfsd_clid_class, __array(unsigned char, addr, sizeof(struct sockaddr_in6)) __field(unsigned long, flavor) __array(unsigned char, verifier, NFS4_VERIFIER_SIZE) - __string_len(name, name, clp->cl_name.len) + __string_len(name, clp->cl_name.data, clp->cl_name.len) ), TP_fast_assign( __entry->cl_boot = clp->cl_clientid.cl_boot; @@ -906,7 +906,7 @@ DECLARE_EVENT_CLASS(nfsd_clid_class, __entry->flavor = clp->cl_cred.cr_flavor; memcpy(__entry->verifier, (void *)&clp->cl_verifier, NFS4_VERIFIER_SIZE); - __assign_str_len(name, clp->cl_name.data, clp->cl_name.len); + __assign_str(name, clp->cl_name.data); ), TP_printk("addr=%pISpc name='%s' verifier=0x%s flavor=%s client=%08x:%08x", __entry->addr, __get_str(name), @@ -1976,7 +1976,7 @@ TRACE_EVENT(nfsd_ctl_time, TP_fast_assign( __entry->netns_ino = net->ns.inum; __entry->time = time; - __assign_str_len(name, name, namelen); + __assign_str(name, name); ), TP_printk("file=%s time=%d\n", __get_str(name), __entry->time diff --git a/fs/nilfs2/alloc.c b/fs/nilfs2/alloc.c index 7342de296ec3..89caef7513db 100644 --- a/fs/nilfs2/alloc.c +++ b/fs/nilfs2/alloc.c @@ -525,54 +525,55 @@ int nilfs_palloc_prepare_alloc_entry(struct inode *inode, ret = nilfs_palloc_get_desc_block(inode, group, 1, &desc_bh); if (ret < 0) return ret; - desc_kaddr = kmap(desc_bh->b_page); + desc_kaddr = kmap_local_page(desc_bh->b_page); desc = nilfs_palloc_block_get_group_desc( inode, group, desc_bh, desc_kaddr); n = nilfs_palloc_rest_groups_in_desc_block(inode, group, maxgroup); - for (j = 0; j < n; j++, desc++, group++) { + for (j = 0; j < n; j++, desc++, group++, group_offset = 0) { lock = nilfs_mdt_bgl_lock(inode, group); - if (nilfs_palloc_group_desc_nfrees(desc, lock) > 0) { - ret = nilfs_palloc_get_bitmap_block( - inode, group, 1, &bitmap_bh); - if (ret < 0) - goto out_desc; - bitmap_kaddr = kmap(bitmap_bh->b_page); - bitmap = bitmap_kaddr + bh_offset(bitmap_bh); - pos = nilfs_palloc_find_available_slot( - bitmap, group_offset, - entries_per_group, lock); - if (pos >= 0) { - /* found a free entry */ - nilfs_palloc_group_desc_add_entries( - desc, lock, -1); - req->pr_entry_nr = - entries_per_group * group + pos; - kunmap(desc_bh->b_page); - kunmap(bitmap_bh->b_page); - - req->pr_desc_bh = desc_bh; - req->pr_bitmap_bh = bitmap_bh; - return 0; - } - kunmap(bitmap_bh->b_page); - brelse(bitmap_bh); + if (nilfs_palloc_group_desc_nfrees(desc, lock) == 0) + continue; + + kunmap_local(desc_kaddr); + ret = nilfs_palloc_get_bitmap_block(inode, group, 1, + &bitmap_bh); + if (unlikely(ret < 0)) { + brelse(desc_bh); + return ret; } - group_offset = 0; + desc_kaddr = kmap_local_page(desc_bh->b_page); + desc = nilfs_palloc_block_get_group_desc( + inode, group, desc_bh, desc_kaddr); + + bitmap_kaddr = kmap_local_page(bitmap_bh->b_page); + bitmap = bitmap_kaddr + bh_offset(bitmap_bh); + pos = nilfs_palloc_find_available_slot( + bitmap, group_offset, entries_per_group, lock); + kunmap_local(bitmap_kaddr); + if (pos >= 0) + goto found; + + brelse(bitmap_bh); } - kunmap(desc_bh->b_page); + kunmap_local(desc_kaddr); brelse(desc_bh); } /* no entries left */ return -ENOSPC; - out_desc: - kunmap(desc_bh->b_page); - brelse(desc_bh); - return ret; +found: + /* found a free entry */ + nilfs_palloc_group_desc_add_entries(desc, lock, -1); + req->pr_entry_nr = entries_per_group * group + pos; + kunmap_local(desc_kaddr); + + req->pr_desc_bh = desc_bh; + req->pr_bitmap_bh = bitmap_bh; + return 0; } /** @@ -606,10 +607,10 @@ void nilfs_palloc_commit_free_entry(struct inode *inode, spinlock_t *lock; group = nilfs_palloc_group(inode, req->pr_entry_nr, &group_offset); - desc_kaddr = kmap(req->pr_desc_bh->b_page); + desc_kaddr = kmap_local_page(req->pr_desc_bh->b_page); desc = nilfs_palloc_block_get_group_desc(inode, group, req->pr_desc_bh, desc_kaddr); - bitmap_kaddr = kmap(req->pr_bitmap_bh->b_page); + bitmap_kaddr = kmap_local_page(req->pr_bitmap_bh->b_page); bitmap = bitmap_kaddr + bh_offset(req->pr_bitmap_bh); lock = nilfs_mdt_bgl_lock(inode, group); @@ -621,8 +622,8 @@ void nilfs_palloc_commit_free_entry(struct inode *inode, else nilfs_palloc_group_desc_add_entries(desc, lock, 1); - kunmap(req->pr_bitmap_bh->b_page); - kunmap(req->pr_desc_bh->b_page); + kunmap_local(bitmap_kaddr); + kunmap_local(desc_kaddr); mark_buffer_dirty(req->pr_desc_bh); mark_buffer_dirty(req->pr_bitmap_bh); @@ -647,10 +648,10 @@ void nilfs_palloc_abort_alloc_entry(struct inode *inode, spinlock_t *lock; group = nilfs_palloc_group(inode, req->pr_entry_nr, &group_offset); - desc_kaddr = kmap(req->pr_desc_bh->b_page); + desc_kaddr = kmap_local_page(req->pr_desc_bh->b_page); desc = nilfs_palloc_block_get_group_desc(inode, group, req->pr_desc_bh, desc_kaddr); - bitmap_kaddr = kmap(req->pr_bitmap_bh->b_page); + bitmap_kaddr = kmap_local_page(req->pr_bitmap_bh->b_page); bitmap = bitmap_kaddr + bh_offset(req->pr_bitmap_bh); lock = nilfs_mdt_bgl_lock(inode, group); @@ -662,8 +663,8 @@ void nilfs_palloc_abort_alloc_entry(struct inode *inode, else nilfs_palloc_group_desc_add_entries(desc, lock, 1); - kunmap(req->pr_bitmap_bh->b_page); - kunmap(req->pr_desc_bh->b_page); + kunmap_local(bitmap_kaddr); + kunmap_local(desc_kaddr); brelse(req->pr_bitmap_bh); brelse(req->pr_desc_bh); @@ -755,7 +756,7 @@ int nilfs_palloc_freev(struct inode *inode, __u64 *entry_nrs, size_t nitems) /* Get the first entry number of the group */ group_min_nr = (__u64)group * epg; - bitmap_kaddr = kmap(bitmap_bh->b_page); + bitmap_kaddr = kmap_local_page(bitmap_bh->b_page); bitmap = bitmap_kaddr + bh_offset(bitmap_bh); lock = nilfs_mdt_bgl_lock(inode, group); @@ -801,7 +802,7 @@ int nilfs_palloc_freev(struct inode *inode, __u64 *entry_nrs, size_t nitems) entry_start = rounddown(group_offset, epb); } while (true); - kunmap(bitmap_bh->b_page); + kunmap_local(bitmap_kaddr); mark_buffer_dirty(bitmap_bh); brelse(bitmap_bh); @@ -815,11 +816,11 @@ int nilfs_palloc_freev(struct inode *inode, __u64 *entry_nrs, size_t nitems) inode->i_ino); } - desc_kaddr = kmap_atomic(desc_bh->b_page); + desc_kaddr = kmap_local_page(desc_bh->b_page); desc = nilfs_palloc_block_get_group_desc( inode, group, desc_bh, desc_kaddr); nfree = nilfs_palloc_group_desc_add_entries(desc, lock, n); - kunmap_atomic(desc_kaddr); + kunmap_local(desc_kaddr); mark_buffer_dirty(desc_bh); nilfs_mdt_mark_dirty(inode); brelse(desc_bh); diff --git a/fs/nilfs2/bmap.c b/fs/nilfs2/bmap.c index 7a8f166f2c8d..383f0afa2cea 100644 --- a/fs/nilfs2/bmap.c +++ b/fs/nilfs2/bmap.c @@ -548,13 +548,10 @@ int nilfs_bmap_read(struct nilfs_bmap *bmap, struct nilfs_inode *raw_inode) */ void nilfs_bmap_write(struct nilfs_bmap *bmap, struct nilfs_inode *raw_inode) { - down_write(&bmap->b_sem); memcpy(raw_inode->i_bmap, bmap->b_u.u_data, NILFS_INODE_BMAP_SIZE * sizeof(__le64)); if (bmap->b_inode->i_ino == NILFS_DAT_INO) bmap->b_last_allocated_ptr = NILFS_BMAP_NEW_PTR_INIT; - - up_write(&bmap->b_sem); } void nilfs_bmap_init_gc(struct nilfs_bmap *bmap) diff --git a/fs/nilfs2/btree.c b/fs/nilfs2/btree.c index 13592e82eaf6..65659fa0372e 100644 --- a/fs/nilfs2/btree.c +++ b/fs/nilfs2/btree.c @@ -724,7 +724,7 @@ static int nilfs_btree_lookup_contig(const struct nilfs_bmap *btree, dat = nilfs_bmap_get_dat(btree); ret = nilfs_dat_translate(dat, ptr, &blocknr); if (ret < 0) - goto out; + goto dat_error; ptr = blocknr; } cnt = 1; @@ -743,7 +743,7 @@ static int nilfs_btree_lookup_contig(const struct nilfs_bmap *btree, if (dat) { ret = nilfs_dat_translate(dat, ptr2, &blocknr); if (ret < 0) - goto out; + goto dat_error; ptr2 = blocknr; } if (ptr2 != ptr + cnt || ++cnt == maxblocks) @@ -781,6 +781,11 @@ static int nilfs_btree_lookup_contig(const struct nilfs_bmap *btree, out: nilfs_btree_free_path(path); return ret; + + dat_error: + if (ret == -ENOENT) + ret = -EINVAL; /* Notify bmap layer of metadata corruption */ + goto out; } static void nilfs_btree_promote_key(struct nilfs_bmap *btree, diff --git a/fs/nilfs2/cpfile.c b/fs/nilfs2/cpfile.c index 39136637f715..69a5cced1e84 100644 --- a/fs/nilfs2/cpfile.c +++ b/fs/nilfs2/cpfile.c @@ -28,7 +28,7 @@ nilfs_cpfile_get_blkoff(const struct inode *cpfile, __u64 cno) { __u64 tcno = cno + NILFS_MDT(cpfile)->mi_first_entry_offset - 1; - do_div(tcno, nilfs_cpfile_checkpoints_per_block(cpfile)); + tcno = div64_ul(tcno, nilfs_cpfile_checkpoints_per_block(cpfile)); return (unsigned long)tcno; } @@ -187,35 +187,90 @@ static inline int nilfs_cpfile_delete_checkpoint_block(struct inode *cpfile, } /** - * nilfs_cpfile_get_checkpoint - get a checkpoint - * @cpfile: inode of checkpoint file - * @cno: checkpoint number - * @create: create flag - * @cpp: pointer to a checkpoint - * @bhp: pointer to a buffer head - * - * Description: nilfs_cpfile_get_checkpoint() acquires the checkpoint - * specified by @cno. A new checkpoint will be created if @cno is the current - * checkpoint number and @create is nonzero. - * - * Return Value: On success, 0 is returned, and the checkpoint and the - * buffer head of the buffer on which the checkpoint is located are stored in - * the place pointed by @cpp and @bhp, respectively. On error, one of the - * following negative error codes is returned. + * nilfs_cpfile_read_checkpoint - read a checkpoint entry in cpfile + * @cpfile: checkpoint file inode + * @cno: number of checkpoint entry to read + * @root: nilfs root object + * @ifile: ifile's inode to read and attach to @root * - * %-EIO - I/O error. + * This function imports checkpoint information from the checkpoint file and + * stores it to the inode file given by @ifile and the nilfs root object + * given by @root. * - * %-ENOMEM - Insufficient amount of memory available. + * Return: 0 on success, or the following negative error code on failure. + * * %-EINVAL - Invalid checkpoint. + * * %-ENOMEM - Insufficient memory available. + * * %-EIO - I/O error (including metadata corruption). + */ +int nilfs_cpfile_read_checkpoint(struct inode *cpfile, __u64 cno, + struct nilfs_root *root, struct inode *ifile) +{ + struct buffer_head *cp_bh; + struct nilfs_checkpoint *cp; + void *kaddr; + int ret; + + if (cno < 1 || cno > nilfs_mdt_cno(cpfile)) + return -EINVAL; + + down_read(&NILFS_MDT(cpfile)->mi_sem); + ret = nilfs_cpfile_get_checkpoint_block(cpfile, cno, 0, &cp_bh); + if (unlikely(ret < 0)) { + if (ret == -ENOENT) + ret = -EINVAL; + goto out_sem; + } + + kaddr = kmap_local_page(cp_bh->b_page); + cp = nilfs_cpfile_block_get_checkpoint(cpfile, cno, cp_bh, kaddr); + if (nilfs_checkpoint_invalid(cp)) { + ret = -EINVAL; + goto put_cp; + } + + ret = nilfs_read_inode_common(ifile, &cp->cp_ifile_inode); + if (unlikely(ret)) { + /* + * Since this inode is on a checkpoint entry, treat errors + * as metadata corruption. + */ + nilfs_err(cpfile->i_sb, + "ifile inode (checkpoint number=%llu) corrupted", + (unsigned long long)cno); + ret = -EIO; + goto put_cp; + } + + /* Configure the nilfs root object */ + atomic64_set(&root->inodes_count, le64_to_cpu(cp->cp_inodes_count)); + atomic64_set(&root->blocks_count, le64_to_cpu(cp->cp_blocks_count)); + root->ifile = ifile; + +put_cp: + kunmap_local(kaddr); + brelse(cp_bh); +out_sem: + up_read(&NILFS_MDT(cpfile)->mi_sem); + return ret; +} + +/** + * nilfs_cpfile_create_checkpoint - create a checkpoint entry on cpfile + * @cpfile: checkpoint file inode + * @cno: number of checkpoint to set up * - * %-ENOENT - No such checkpoint. + * This function creates a checkpoint with the number specified by @cno on + * cpfile. If the specified checkpoint entry already exists due to a past + * failure, it will be reused without returning an error. + * In either case, the buffer of the block containing the checkpoint entry + * and the cpfile inode are made dirty for inclusion in the write log. * - * %-EINVAL - invalid checkpoint. + * Return: 0 on success, or the following negative error code on failure. + * * %-ENOMEM - Insufficient memory available. + * * %-EIO - I/O error (including metadata corruption). + * * %-EROFS - Read only filesystem */ -int nilfs_cpfile_get_checkpoint(struct inode *cpfile, - __u64 cno, - int create, - struct nilfs_checkpoint **cpp, - struct buffer_head **bhp) +int nilfs_cpfile_create_checkpoint(struct inode *cpfile, __u64 cno) { struct buffer_head *header_bh, *cp_bh; struct nilfs_cpfile_header *header; @@ -223,70 +278,128 @@ int nilfs_cpfile_get_checkpoint(struct inode *cpfile, void *kaddr; int ret; - if (unlikely(cno < 1 || cno > nilfs_mdt_cno(cpfile) || - (cno < nilfs_mdt_cno(cpfile) && create))) - return -EINVAL; + if (WARN_ON_ONCE(cno < 1)) + return -EIO; down_write(&NILFS_MDT(cpfile)->mi_sem); - ret = nilfs_cpfile_get_header_block(cpfile, &header_bh); - if (ret < 0) + if (unlikely(ret < 0)) { + if (ret == -ENOENT) { + nilfs_error(cpfile->i_sb, + "checkpoint creation failed due to metadata corruption."); + ret = -EIO; + } goto out_sem; - ret = nilfs_cpfile_get_checkpoint_block(cpfile, cno, create, &cp_bh); - if (ret < 0) + } + ret = nilfs_cpfile_get_checkpoint_block(cpfile, cno, 1, &cp_bh); + if (unlikely(ret < 0)) goto out_header; - kaddr = kmap(cp_bh->b_page); + + kaddr = kmap_local_page(cp_bh->b_page); cp = nilfs_cpfile_block_get_checkpoint(cpfile, cno, cp_bh, kaddr); if (nilfs_checkpoint_invalid(cp)) { - if (!create) { - kunmap(cp_bh->b_page); - brelse(cp_bh); - ret = -ENOENT; - goto out_header; - } /* a newly-created checkpoint */ nilfs_checkpoint_clear_invalid(cp); if (!nilfs_cpfile_is_in_first(cpfile, cno)) nilfs_cpfile_block_add_valid_checkpoints(cpfile, cp_bh, kaddr, 1); - mark_buffer_dirty(cp_bh); + kunmap_local(kaddr); - kaddr = kmap_atomic(header_bh->b_page); + kaddr = kmap_local_page(header_bh->b_page); header = nilfs_cpfile_block_get_header(cpfile, header_bh, kaddr); le64_add_cpu(&header->ch_ncheckpoints, 1); - kunmap_atomic(kaddr); + kunmap_local(kaddr); mark_buffer_dirty(header_bh); - nilfs_mdt_mark_dirty(cpfile); + } else { + kunmap_local(kaddr); } - if (cpp != NULL) - *cpp = cp; - *bhp = cp_bh; + /* Force the buffer and the inode to become dirty */ + mark_buffer_dirty(cp_bh); + brelse(cp_bh); + nilfs_mdt_mark_dirty(cpfile); - out_header: +out_header: brelse(header_bh); - out_sem: +out_sem: up_write(&NILFS_MDT(cpfile)->mi_sem); return ret; } /** - * nilfs_cpfile_put_checkpoint - put a checkpoint - * @cpfile: inode of checkpoint file - * @cno: checkpoint number - * @bh: buffer head + * nilfs_cpfile_finalize_checkpoint - fill in a checkpoint entry in cpfile + * @cpfile: checkpoint file inode + * @cno: checkpoint number + * @root: nilfs root object + * @blkinc: number of blocks added by this checkpoint + * @ctime: checkpoint creation time + * @minor: minor checkpoint flag + * + * This function completes the checkpoint entry numbered by @cno in the + * cpfile with the data given by the arguments @root, @blkinc, @ctime, and + * @minor. * - * Description: nilfs_cpfile_put_checkpoint() releases the checkpoint - * specified by @cno. @bh must be the buffer head which has been returned by - * a previous call to nilfs_cpfile_get_checkpoint() with @cno. + * Return: 0 on success, or the following negative error code on failure. + * * %-ENOMEM - Insufficient memory available. + * * %-EIO - I/O error (including metadata corruption). */ -void nilfs_cpfile_put_checkpoint(struct inode *cpfile, __u64 cno, - struct buffer_head *bh) +int nilfs_cpfile_finalize_checkpoint(struct inode *cpfile, __u64 cno, + struct nilfs_root *root, __u64 blkinc, + time64_t ctime, bool minor) { - kunmap(bh->b_page); - brelse(bh); + struct buffer_head *cp_bh; + struct nilfs_checkpoint *cp; + void *kaddr; + int ret; + + if (WARN_ON_ONCE(cno < 1)) + return -EIO; + + down_write(&NILFS_MDT(cpfile)->mi_sem); + ret = nilfs_cpfile_get_checkpoint_block(cpfile, cno, 0, &cp_bh); + if (unlikely(ret < 0)) { + if (ret == -ENOENT) + goto error; + goto out_sem; + } + + kaddr = kmap_local_page(cp_bh->b_page); + cp = nilfs_cpfile_block_get_checkpoint(cpfile, cno, cp_bh, kaddr); + if (unlikely(nilfs_checkpoint_invalid(cp))) { + kunmap_local(kaddr); + brelse(cp_bh); + goto error; + } + + cp->cp_snapshot_list.ssl_next = 0; + cp->cp_snapshot_list.ssl_prev = 0; + cp->cp_inodes_count = cpu_to_le64(atomic64_read(&root->inodes_count)); + cp->cp_blocks_count = cpu_to_le64(atomic64_read(&root->blocks_count)); + cp->cp_nblk_inc = cpu_to_le64(blkinc); + cp->cp_create = cpu_to_le64(ctime); + cp->cp_cno = cpu_to_le64(cno); + + if (minor) + nilfs_checkpoint_set_minor(cp); + else + nilfs_checkpoint_clear_minor(cp); + + nilfs_write_inode_common(root->ifile, &cp->cp_ifile_inode); + nilfs_bmap_write(NILFS_I(root->ifile)->i_bmap, &cp->cp_ifile_inode); + + kunmap_local(kaddr); + brelse(cp_bh); +out_sem: + up_write(&NILFS_MDT(cpfile)->mi_sem); + return ret; + +error: + nilfs_error(cpfile->i_sb, + "checkpoint finalization failed due to metadata corruption."); + ret = -EIO; + goto out_sem; } /** @@ -347,7 +460,7 @@ int nilfs_cpfile_delete_checkpoints(struct inode *cpfile, continue; } - kaddr = kmap_atomic(cp_bh->b_page); + kaddr = kmap_local_page(cp_bh->b_page); cp = nilfs_cpfile_block_get_checkpoint( cpfile, cno, cp_bh, kaddr); nicps = 0; @@ -369,7 +482,7 @@ int nilfs_cpfile_delete_checkpoints(struct inode *cpfile, cpfile, cp_bh, kaddr, nicps); if (count == 0) { /* make hole */ - kunmap_atomic(kaddr); + kunmap_local(kaddr); brelse(cp_bh); ret = nilfs_cpfile_delete_checkpoint_block( @@ -384,18 +497,18 @@ int nilfs_cpfile_delete_checkpoints(struct inode *cpfile, } } - kunmap_atomic(kaddr); + kunmap_local(kaddr); brelse(cp_bh); } if (tnicps > 0) { - kaddr = kmap_atomic(header_bh->b_page); + kaddr = kmap_local_page(header_bh->b_page); header = nilfs_cpfile_block_get_header(cpfile, header_bh, kaddr); le64_add_cpu(&header->ch_ncheckpoints, -(u64)tnicps); mark_buffer_dirty(header_bh); nilfs_mdt_mark_dirty(cpfile); - kunmap_atomic(kaddr); + kunmap_local(kaddr); } brelse(header_bh); @@ -447,7 +560,7 @@ static ssize_t nilfs_cpfile_do_get_cpinfo(struct inode *cpfile, __u64 *cnop, } ncps = nilfs_cpfile_checkpoints_in_block(cpfile, cno, cur_cno); - kaddr = kmap_atomic(bh->b_page); + kaddr = kmap_local_page(bh->b_page); cp = nilfs_cpfile_block_get_checkpoint(cpfile, cno, bh, kaddr); for (i = 0; i < ncps && n < nci; i++, cp = (void *)cp + cpsz) { if (!nilfs_checkpoint_invalid(cp)) { @@ -457,7 +570,7 @@ static ssize_t nilfs_cpfile_do_get_cpinfo(struct inode *cpfile, __u64 *cnop, n++; } } - kunmap_atomic(kaddr); + kunmap_local(kaddr); brelse(bh); } @@ -491,10 +604,10 @@ static ssize_t nilfs_cpfile_do_get_ssinfo(struct inode *cpfile, __u64 *cnop, ret = nilfs_cpfile_get_header_block(cpfile, &bh); if (ret < 0) goto out; - kaddr = kmap_atomic(bh->b_page); + kaddr = kmap_local_page(bh->b_page); header = nilfs_cpfile_block_get_header(cpfile, bh, kaddr); curr = le64_to_cpu(header->ch_snapshot_list.ssl_next); - kunmap_atomic(kaddr); + kunmap_local(kaddr); brelse(bh); if (curr == 0) { ret = 0; @@ -512,7 +625,7 @@ static ssize_t nilfs_cpfile_do_get_ssinfo(struct inode *cpfile, __u64 *cnop, ret = 0; /* No snapshots (started from a hole block) */ goto out; } - kaddr = kmap_atomic(bh->b_page); + kaddr = kmap_local_page(bh->b_page); while (n < nci) { cp = nilfs_cpfile_block_get_checkpoint(cpfile, curr, bh, kaddr); curr = ~(__u64)0; /* Terminator */ @@ -528,7 +641,7 @@ static ssize_t nilfs_cpfile_do_get_ssinfo(struct inode *cpfile, __u64 *cnop, next_blkoff = nilfs_cpfile_get_blkoff(cpfile, next); if (curr_blkoff != next_blkoff) { - kunmap_atomic(kaddr); + kunmap_local(kaddr); brelse(bh); ret = nilfs_cpfile_get_checkpoint_block(cpfile, next, 0, &bh); @@ -536,12 +649,12 @@ static ssize_t nilfs_cpfile_do_get_ssinfo(struct inode *cpfile, __u64 *cnop, WARN_ON(ret == -ENOENT); goto out; } - kaddr = kmap_atomic(bh->b_page); + kaddr = kmap_local_page(bh->b_page); } curr = next; curr_blkoff = next_blkoff; } - kunmap_atomic(kaddr); + kunmap_local(kaddr); brelse(bh); *cnop = curr; ret = n; @@ -650,24 +763,24 @@ static int nilfs_cpfile_set_snapshot(struct inode *cpfile, __u64 cno) ret = nilfs_cpfile_get_checkpoint_block(cpfile, cno, 0, &cp_bh); if (ret < 0) goto out_sem; - kaddr = kmap_atomic(cp_bh->b_page); + kaddr = kmap_local_page(cp_bh->b_page); cp = nilfs_cpfile_block_get_checkpoint(cpfile, cno, cp_bh, kaddr); if (nilfs_checkpoint_invalid(cp)) { ret = -ENOENT; - kunmap_atomic(kaddr); + kunmap_local(kaddr); goto out_cp; } if (nilfs_checkpoint_snapshot(cp)) { ret = 0; - kunmap_atomic(kaddr); + kunmap_local(kaddr); goto out_cp; } - kunmap_atomic(kaddr); + kunmap_local(kaddr); ret = nilfs_cpfile_get_header_block(cpfile, &header_bh); if (ret < 0) goto out_cp; - kaddr = kmap_atomic(header_bh->b_page); + kaddr = kmap_local_page(header_bh->b_page); header = nilfs_cpfile_block_get_header(cpfile, header_bh, kaddr); list = &header->ch_snapshot_list; curr_bh = header_bh; @@ -679,13 +792,13 @@ static int nilfs_cpfile_set_snapshot(struct inode *cpfile, __u64 cno) prev_blkoff = nilfs_cpfile_get_blkoff(cpfile, prev); curr = prev; if (curr_blkoff != prev_blkoff) { - kunmap_atomic(kaddr); + kunmap_local(kaddr); brelse(curr_bh); ret = nilfs_cpfile_get_checkpoint_block(cpfile, curr, 0, &curr_bh); if (ret < 0) goto out_header; - kaddr = kmap_atomic(curr_bh->b_page); + kaddr = kmap_local_page(curr_bh->b_page); } curr_blkoff = prev_blkoff; cp = nilfs_cpfile_block_get_checkpoint( @@ -693,7 +806,7 @@ static int nilfs_cpfile_set_snapshot(struct inode *cpfile, __u64 cno) list = &cp->cp_snapshot_list; prev = le64_to_cpu(list->ssl_prev); } - kunmap_atomic(kaddr); + kunmap_local(kaddr); if (prev != 0) { ret = nilfs_cpfile_get_checkpoint_block(cpfile, prev, 0, @@ -705,29 +818,29 @@ static int nilfs_cpfile_set_snapshot(struct inode *cpfile, __u64 cno) get_bh(prev_bh); } - kaddr = kmap_atomic(curr_bh->b_page); + kaddr = kmap_local_page(curr_bh->b_page); list = nilfs_cpfile_block_get_snapshot_list( cpfile, curr, curr_bh, kaddr); list->ssl_prev = cpu_to_le64(cno); - kunmap_atomic(kaddr); + kunmap_local(kaddr); - kaddr = kmap_atomic(cp_bh->b_page); + kaddr = kmap_local_page(cp_bh->b_page); cp = nilfs_cpfile_block_get_checkpoint(cpfile, cno, cp_bh, kaddr); cp->cp_snapshot_list.ssl_next = cpu_to_le64(curr); cp->cp_snapshot_list.ssl_prev = cpu_to_le64(prev); nilfs_checkpoint_set_snapshot(cp); - kunmap_atomic(kaddr); + kunmap_local(kaddr); - kaddr = kmap_atomic(prev_bh->b_page); + kaddr = kmap_local_page(prev_bh->b_page); list = nilfs_cpfile_block_get_snapshot_list( cpfile, prev, prev_bh, kaddr); list->ssl_next = cpu_to_le64(cno); - kunmap_atomic(kaddr); + kunmap_local(kaddr); - kaddr = kmap_atomic(header_bh->b_page); + kaddr = kmap_local_page(header_bh->b_page); header = nilfs_cpfile_block_get_header(cpfile, header_bh, kaddr); le64_add_cpu(&header->ch_nsnapshots, 1); - kunmap_atomic(kaddr); + kunmap_local(kaddr); mark_buffer_dirty(prev_bh); mark_buffer_dirty(curr_bh); @@ -768,23 +881,23 @@ static int nilfs_cpfile_clear_snapshot(struct inode *cpfile, __u64 cno) ret = nilfs_cpfile_get_checkpoint_block(cpfile, cno, 0, &cp_bh); if (ret < 0) goto out_sem; - kaddr = kmap_atomic(cp_bh->b_page); + kaddr = kmap_local_page(cp_bh->b_page); cp = nilfs_cpfile_block_get_checkpoint(cpfile, cno, cp_bh, kaddr); if (nilfs_checkpoint_invalid(cp)) { ret = -ENOENT; - kunmap_atomic(kaddr); + kunmap_local(kaddr); goto out_cp; } if (!nilfs_checkpoint_snapshot(cp)) { ret = 0; - kunmap_atomic(kaddr); + kunmap_local(kaddr); goto out_cp; } list = &cp->cp_snapshot_list; next = le64_to_cpu(list->ssl_next); prev = le64_to_cpu(list->ssl_prev); - kunmap_atomic(kaddr); + kunmap_local(kaddr); ret = nilfs_cpfile_get_header_block(cpfile, &header_bh); if (ret < 0) @@ -808,29 +921,29 @@ static int nilfs_cpfile_clear_snapshot(struct inode *cpfile, __u64 cno) get_bh(prev_bh); } - kaddr = kmap_atomic(next_bh->b_page); + kaddr = kmap_local_page(next_bh->b_page); list = nilfs_cpfile_block_get_snapshot_list( cpfile, next, next_bh, kaddr); list->ssl_prev = cpu_to_le64(prev); - kunmap_atomic(kaddr); + kunmap_local(kaddr); - kaddr = kmap_atomic(prev_bh->b_page); + kaddr = kmap_local_page(prev_bh->b_page); list = nilfs_cpfile_block_get_snapshot_list( cpfile, prev, prev_bh, kaddr); list->ssl_next = cpu_to_le64(next); - kunmap_atomic(kaddr); + kunmap_local(kaddr); - kaddr = kmap_atomic(cp_bh->b_page); + kaddr = kmap_local_page(cp_bh->b_page); cp = nilfs_cpfile_block_get_checkpoint(cpfile, cno, cp_bh, kaddr); cp->cp_snapshot_list.ssl_next = cpu_to_le64(0); cp->cp_snapshot_list.ssl_prev = cpu_to_le64(0); nilfs_checkpoint_clear_snapshot(cp); - kunmap_atomic(kaddr); + kunmap_local(kaddr); - kaddr = kmap_atomic(header_bh->b_page); + kaddr = kmap_local_page(header_bh->b_page); header = nilfs_cpfile_block_get_header(cpfile, header_bh, kaddr); le64_add_cpu(&header->ch_nsnapshots, -1); - kunmap_atomic(kaddr); + kunmap_local(kaddr); mark_buffer_dirty(next_bh); mark_buffer_dirty(prev_bh); @@ -889,13 +1002,13 @@ int nilfs_cpfile_is_snapshot(struct inode *cpfile, __u64 cno) ret = nilfs_cpfile_get_checkpoint_block(cpfile, cno, 0, &bh); if (ret < 0) goto out; - kaddr = kmap_atomic(bh->b_page); + kaddr = kmap_local_page(bh->b_page); cp = nilfs_cpfile_block_get_checkpoint(cpfile, cno, bh, kaddr); if (nilfs_checkpoint_invalid(cp)) ret = -ENOENT; else ret = nilfs_checkpoint_snapshot(cp); - kunmap_atomic(kaddr); + kunmap_local(kaddr); brelse(bh); out: @@ -972,12 +1085,12 @@ int nilfs_cpfile_get_stat(struct inode *cpfile, struct nilfs_cpstat *cpstat) ret = nilfs_cpfile_get_header_block(cpfile, &bh); if (ret < 0) goto out_sem; - kaddr = kmap_atomic(bh->b_page); + kaddr = kmap_local_page(bh->b_page); header = nilfs_cpfile_block_get_header(cpfile, bh, kaddr); cpstat->cs_cno = nilfs_mdt_cno(cpfile); cpstat->cs_ncps = le64_to_cpu(header->ch_ncheckpoints); cpstat->cs_nsss = le64_to_cpu(header->ch_nsnapshots); - kunmap_atomic(kaddr); + kunmap_local(kaddr); brelse(bh); out_sem: diff --git a/fs/nilfs2/cpfile.h b/fs/nilfs2/cpfile.h index edabb2dc5756..f5b1d59289eb 100644 --- a/fs/nilfs2/cpfile.h +++ b/fs/nilfs2/cpfile.h @@ -16,10 +16,12 @@ #include <linux/nilfs2_ondisk.h> /* nilfs_inode, nilfs_checkpoint */ -int nilfs_cpfile_get_checkpoint(struct inode *, __u64, int, - struct nilfs_checkpoint **, - struct buffer_head **); -void nilfs_cpfile_put_checkpoint(struct inode *, __u64, struct buffer_head *); +int nilfs_cpfile_read_checkpoint(struct inode *cpfile, __u64 cno, + struct nilfs_root *root, struct inode *ifile); +int nilfs_cpfile_create_checkpoint(struct inode *cpfile, __u64 cno); +int nilfs_cpfile_finalize_checkpoint(struct inode *cpfile, __u64 cno, + struct nilfs_root *root, __u64 blkinc, + time64_t ctime, bool minor); int nilfs_cpfile_delete_checkpoints(struct inode *, __u64, __u64); int nilfs_cpfile_delete_checkpoint(struct inode *, __u64); int nilfs_cpfile_change_cpmode(struct inode *, __u64, int); diff --git a/fs/nilfs2/dat.c b/fs/nilfs2/dat.c index 9cf6ba58f585..180fc8d36213 100644 --- a/fs/nilfs2/dat.c +++ b/fs/nilfs2/dat.c @@ -91,13 +91,13 @@ void nilfs_dat_commit_alloc(struct inode *dat, struct nilfs_palloc_req *req) struct nilfs_dat_entry *entry; void *kaddr; - kaddr = kmap_atomic(req->pr_entry_bh->b_page); + kaddr = kmap_local_page(req->pr_entry_bh->b_page); entry = nilfs_palloc_block_get_entry(dat, req->pr_entry_nr, req->pr_entry_bh, kaddr); entry->de_start = cpu_to_le64(NILFS_CNO_MIN); entry->de_end = cpu_to_le64(NILFS_CNO_MAX); entry->de_blocknr = cpu_to_le64(0); - kunmap_atomic(kaddr); + kunmap_local(kaddr); nilfs_palloc_commit_alloc_entry(dat, req); nilfs_dat_commit_entry(dat, req); @@ -115,13 +115,13 @@ static void nilfs_dat_commit_free(struct inode *dat, struct nilfs_dat_entry *entry; void *kaddr; - kaddr = kmap_atomic(req->pr_entry_bh->b_page); + kaddr = kmap_local_page(req->pr_entry_bh->b_page); entry = nilfs_palloc_block_get_entry(dat, req->pr_entry_nr, req->pr_entry_bh, kaddr); entry->de_start = cpu_to_le64(NILFS_CNO_MIN); entry->de_end = cpu_to_le64(NILFS_CNO_MIN); entry->de_blocknr = cpu_to_le64(0); - kunmap_atomic(kaddr); + kunmap_local(kaddr); nilfs_dat_commit_entry(dat, req); @@ -145,12 +145,12 @@ void nilfs_dat_commit_start(struct inode *dat, struct nilfs_palloc_req *req, struct nilfs_dat_entry *entry; void *kaddr; - kaddr = kmap_atomic(req->pr_entry_bh->b_page); + kaddr = kmap_local_page(req->pr_entry_bh->b_page); entry = nilfs_palloc_block_get_entry(dat, req->pr_entry_nr, req->pr_entry_bh, kaddr); entry->de_start = cpu_to_le64(nilfs_mdt_cno(dat)); entry->de_blocknr = cpu_to_le64(blocknr); - kunmap_atomic(kaddr); + kunmap_local(kaddr); nilfs_dat_commit_entry(dat, req); } @@ -167,12 +167,12 @@ int nilfs_dat_prepare_end(struct inode *dat, struct nilfs_palloc_req *req) if (ret < 0) return ret; - kaddr = kmap_atomic(req->pr_entry_bh->b_page); + kaddr = kmap_local_page(req->pr_entry_bh->b_page); entry = nilfs_palloc_block_get_entry(dat, req->pr_entry_nr, req->pr_entry_bh, kaddr); start = le64_to_cpu(entry->de_start); blocknr = le64_to_cpu(entry->de_blocknr); - kunmap_atomic(kaddr); + kunmap_local(kaddr); if (blocknr == 0) { ret = nilfs_palloc_prepare_free_entry(dat, req); @@ -202,7 +202,7 @@ void nilfs_dat_commit_end(struct inode *dat, struct nilfs_palloc_req *req, sector_t blocknr; void *kaddr; - kaddr = kmap_atomic(req->pr_entry_bh->b_page); + kaddr = kmap_local_page(req->pr_entry_bh->b_page); entry = nilfs_palloc_block_get_entry(dat, req->pr_entry_nr, req->pr_entry_bh, kaddr); end = start = le64_to_cpu(entry->de_start); @@ -212,7 +212,7 @@ void nilfs_dat_commit_end(struct inode *dat, struct nilfs_palloc_req *req, } entry->de_end = cpu_to_le64(end); blocknr = le64_to_cpu(entry->de_blocknr); - kunmap_atomic(kaddr); + kunmap_local(kaddr); if (blocknr == 0) nilfs_dat_commit_free(dat, req); @@ -227,12 +227,12 @@ void nilfs_dat_abort_end(struct inode *dat, struct nilfs_palloc_req *req) sector_t blocknr; void *kaddr; - kaddr = kmap_atomic(req->pr_entry_bh->b_page); + kaddr = kmap_local_page(req->pr_entry_bh->b_page); entry = nilfs_palloc_block_get_entry(dat, req->pr_entry_nr, req->pr_entry_bh, kaddr); start = le64_to_cpu(entry->de_start); blocknr = le64_to_cpu(entry->de_blocknr); - kunmap_atomic(kaddr); + kunmap_local(kaddr); if (start == nilfs_mdt_cno(dat) && blocknr == 0) nilfs_palloc_abort_free_entry(dat, req); @@ -362,7 +362,7 @@ int nilfs_dat_move(struct inode *dat, __u64 vblocknr, sector_t blocknr) } } - kaddr = kmap_atomic(entry_bh->b_page); + kaddr = kmap_local_page(entry_bh->b_page); entry = nilfs_palloc_block_get_entry(dat, vblocknr, entry_bh, kaddr); if (unlikely(entry->de_blocknr == cpu_to_le64(0))) { nilfs_crit(dat->i_sb, @@ -370,13 +370,13 @@ int nilfs_dat_move(struct inode *dat, __u64 vblocknr, sector_t blocknr) __func__, (unsigned long long)vblocknr, (unsigned long long)le64_to_cpu(entry->de_start), (unsigned long long)le64_to_cpu(entry->de_end)); - kunmap_atomic(kaddr); + kunmap_local(kaddr); brelse(entry_bh); return -EINVAL; } WARN_ON(blocknr == 0); entry->de_blocknr = cpu_to_le64(blocknr); - kunmap_atomic(kaddr); + kunmap_local(kaddr); mark_buffer_dirty(entry_bh); nilfs_mdt_mark_dirty(dat); @@ -426,7 +426,7 @@ int nilfs_dat_translate(struct inode *dat, __u64 vblocknr, sector_t *blocknrp) } } - kaddr = kmap_atomic(entry_bh->b_page); + kaddr = kmap_local_page(entry_bh->b_page); entry = nilfs_palloc_block_get_entry(dat, vblocknr, entry_bh, kaddr); blocknr = le64_to_cpu(entry->de_blocknr); if (blocknr == 0) { @@ -436,7 +436,7 @@ int nilfs_dat_translate(struct inode *dat, __u64 vblocknr, sector_t *blocknrp) *blocknrp = blocknr; out: - kunmap_atomic(kaddr); + kunmap_local(kaddr); brelse(entry_bh); return ret; } @@ -457,10 +457,10 @@ ssize_t nilfs_dat_get_vinfo(struct inode *dat, void *buf, unsigned int visz, 0, &entry_bh); if (ret < 0) return ret; - kaddr = kmap_atomic(entry_bh->b_page); + kaddr = kmap_local_page(entry_bh->b_page); /* last virtual block number in this block */ first = vinfo->vi_vblocknr; - do_div(first, entries_per_block); + first = div64_ul(first, entries_per_block); first *= entries_per_block; last = first + entries_per_block - 1; for (j = i, n = 0; @@ -473,7 +473,7 @@ ssize_t nilfs_dat_get_vinfo(struct inode *dat, void *buf, unsigned int visz, vinfo->vi_end = le64_to_cpu(entry->de_end); vinfo->vi_blocknr = le64_to_cpu(entry->de_blocknr); } - kunmap_atomic(kaddr); + kunmap_local(kaddr); brelse(entry_bh); } diff --git a/fs/nilfs2/direct.c b/fs/nilfs2/direct.c index 4c85914f2abc..893ab36824cc 100644 --- a/fs/nilfs2/direct.c +++ b/fs/nilfs2/direct.c @@ -66,7 +66,7 @@ static int nilfs_direct_lookup_contig(const struct nilfs_bmap *direct, dat = nilfs_bmap_get_dat(direct); ret = nilfs_dat_translate(dat, ptr, &blocknr); if (ret < 0) - return ret; + goto dat_error; ptr = blocknr; } @@ -79,7 +79,7 @@ static int nilfs_direct_lookup_contig(const struct nilfs_bmap *direct, if (dat) { ret = nilfs_dat_translate(dat, ptr2, &blocknr); if (ret < 0) - return ret; + goto dat_error; ptr2 = blocknr; } if (ptr2 != ptr + cnt) @@ -87,6 +87,11 @@ static int nilfs_direct_lookup_contig(const struct nilfs_bmap *direct, } *ptrp = ptr; return cnt; + + dat_error: + if (ret == -ENOENT) + ret = -EINVAL; /* Notify bmap layer of metadata corruption */ + return ret; } static __u64 diff --git a/fs/nilfs2/ifile.c b/fs/nilfs2/ifile.c index a8a4bc8490b4..612e609158b5 100644 --- a/fs/nilfs2/ifile.c +++ b/fs/nilfs2/ifile.c @@ -15,6 +15,7 @@ #include "mdt.h" #include "alloc.h" #include "ifile.h" +#include "cpfile.h" /** * struct nilfs_ifile_info - on-memory private data of ifile @@ -115,11 +116,11 @@ int nilfs_ifile_delete_inode(struct inode *ifile, ino_t ino) return ret; } - kaddr = kmap_atomic(req.pr_entry_bh->b_page); + kaddr = kmap_local_page(req.pr_entry_bh->b_page); raw_inode = nilfs_palloc_block_get_entry(ifile, req.pr_entry_nr, req.pr_entry_bh, kaddr); raw_inode->i_flags = 0; - kunmap_atomic(kaddr); + kunmap_local(kaddr); mark_buffer_dirty(req.pr_entry_bh); brelse(req.pr_entry_bh); @@ -173,14 +174,18 @@ int nilfs_ifile_count_free_inodes(struct inode *ifile, * nilfs_ifile_read - read or get ifile inode * @sb: super block instance * @root: root object + * @cno: number of checkpoint entry to read * @inode_size: size of an inode - * @raw_inode: on-disk ifile inode - * @inodep: buffer to store the inode + * + * Return: 0 on success, or the following negative error code on failure. + * * %-EINVAL - Invalid checkpoint. + * * %-ENOMEM - Insufficient memory available. + * * %-EIO - I/O error (including metadata corruption). */ int nilfs_ifile_read(struct super_block *sb, struct nilfs_root *root, - size_t inode_size, struct nilfs_inode *raw_inode, - struct inode **inodep) + __u64 cno, size_t inode_size) { + struct the_nilfs *nilfs; struct inode *ifile; int err; @@ -201,13 +206,13 @@ int nilfs_ifile_read(struct super_block *sb, struct nilfs_root *root, nilfs_palloc_setup_cache(ifile, &NILFS_IFILE_I(ifile)->palloc_cache); - err = nilfs_read_inode_common(ifile, raw_inode); + nilfs = sb->s_fs_info; + err = nilfs_cpfile_read_checkpoint(nilfs->ns_cpfile, cno, root, ifile); if (err) goto failed; unlock_new_inode(ifile); out: - *inodep = ifile; return 0; failed: iget_failed(ifile); diff --git a/fs/nilfs2/ifile.h b/fs/nilfs2/ifile.h index 35c5273f4821..625545cc2a98 100644 --- a/fs/nilfs2/ifile.h +++ b/fs/nilfs2/ifile.h @@ -21,15 +21,14 @@ static inline struct nilfs_inode * nilfs_ifile_map_inode(struct inode *ifile, ino_t ino, struct buffer_head *ibh) { - void *kaddr = kmap(ibh->b_page); + void *kaddr = kmap_local_page(ibh->b_page); return nilfs_palloc_block_get_entry(ifile, ino, ibh, kaddr); } -static inline void nilfs_ifile_unmap_inode(struct inode *ifile, ino_t ino, - struct buffer_head *ibh) +static inline void nilfs_ifile_unmap_inode(struct nilfs_inode *raw_inode) { - kunmap(ibh->b_page); + kunmap_local(raw_inode); } int nilfs_ifile_create_inode(struct inode *, ino_t *, struct buffer_head **); @@ -39,7 +38,6 @@ int nilfs_ifile_get_inode_block(struct inode *, ino_t, struct buffer_head **); int nilfs_ifile_count_free_inodes(struct inode *, u64 *, u64 *); int nilfs_ifile_read(struct super_block *sb, struct nilfs_root *root, - size_t inode_size, struct nilfs_inode *raw_inode, - struct inode **inodep); + __u64 cno, size_t inode_size); #endif /* _NILFS_IFILE_H */ diff --git a/fs/nilfs2/inode.c b/fs/nilfs2/inode.c index 9c334c722fc1..7340a01d80e1 100644 --- a/fs/nilfs2/inode.c +++ b/fs/nilfs2/inode.c @@ -112,7 +112,7 @@ int nilfs_get_block(struct inode *inode, sector_t blkoff, "%s (ino=%lu): a race condition while inserting a data block at offset=%llu", __func__, inode->i_ino, (unsigned long long)blkoff); - err = 0; + err = -EAGAIN; } nilfs_transaction_abort(inode->i_sb); goto out; @@ -520,7 +520,7 @@ static int __nilfs_read_inode(struct super_block *sb, inode, inode->i_mode, huge_decode_dev(le64_to_cpu(raw_inode->i_device_code))); } - nilfs_ifile_unmap_inode(root->ifile, ino, bh); + nilfs_ifile_unmap_inode(raw_inode); brelse(bh); up_read(&NILFS_MDT(nilfs->ns_dat)->mi_sem); nilfs_set_inode_flags(inode); @@ -529,7 +529,7 @@ static int __nilfs_read_inode(struct super_block *sb, return 0; failed_unmap: - nilfs_ifile_unmap_inode(root->ifile, ino, bh); + nilfs_ifile_unmap_inode(raw_inode); brelse(bh); bad_inode: @@ -759,8 +759,18 @@ struct inode *nilfs_iget_for_shadow(struct inode *inode) return s_inode; } +/** + * nilfs_write_inode_common - export common inode information to on-disk inode + * @inode: inode object + * @raw_inode: on-disk inode + * + * This function writes standard information from the on-memory inode @inode + * to @raw_inode on ifile, cpfile or a super root block. Since inode bmap + * data is not exported, nilfs_bmap_write() must be called separately during + * log writing. + */ void nilfs_write_inode_common(struct inode *inode, - struct nilfs_inode *raw_inode, int has_bmap) + struct nilfs_inode *raw_inode) { struct nilfs_inode_info *ii = NILFS_I(inode); @@ -778,21 +788,6 @@ void nilfs_write_inode_common(struct inode *inode, raw_inode->i_flags = cpu_to_le32(ii->i_flags); raw_inode->i_generation = cpu_to_le32(inode->i_generation); - if (NILFS_ROOT_METADATA_FILE(inode->i_ino)) { - struct the_nilfs *nilfs = inode->i_sb->s_fs_info; - - /* zero-fill unused portion in the case of super root block */ - raw_inode->i_xattr = 0; - raw_inode->i_pad = 0; - memset((void *)raw_inode + sizeof(*raw_inode), 0, - nilfs->ns_inode_size - sizeof(*raw_inode)); - } - - if (has_bmap) - nilfs_bmap_write(ii->i_bmap, raw_inode); - else if (S_ISCHR(inode->i_mode) || S_ISBLK(inode->i_mode)) - raw_inode->i_device_code = - cpu_to_le64(huge_encode_dev(inode->i_rdev)); /* * When extending inode, nilfs->ns_inode_size should be checked * for substitutions of appended fields. @@ -813,14 +808,13 @@ void nilfs_update_inode(struct inode *inode, struct buffer_head *ibh, int flags) if (flags & I_DIRTY_DATASYNC) set_bit(NILFS_I_INODE_SYNC, &ii->i_state); - nilfs_write_inode_common(inode, raw_inode, 0); - /* - * XXX: call with has_bmap = 0 is a workaround to avoid - * deadlock of bmap. This delays update of i_bmap to just - * before writing. - */ + nilfs_write_inode_common(inode, raw_inode); + + if (S_ISCHR(inode->i_mode) || S_ISBLK(inode->i_mode)) + raw_inode->i_device_code = + cpu_to_le64(huge_encode_dev(inode->i_rdev)); - nilfs_ifile_unmap_inode(ifile, ino, ibh); + nilfs_ifile_unmap_inode(raw_inode); } #define NILFS_MAX_TRUNCATE_BLOCKS 16384 /* 64MB for 4KB block */ diff --git a/fs/nilfs2/ioctl.c b/fs/nilfs2/ioctl.c index cfb6aca5ec38..f1a01c191cf5 100644 --- a/fs/nilfs2/ioctl.c +++ b/fs/nilfs2/ioctl.c @@ -1111,7 +1111,7 @@ static int nilfs_ioctl_set_alloc_range(struct inode *inode, void __user *argp) segbytes = nilfs->ns_blocks_per_segment * nilfs->ns_blocksize; minseg = range[0] + segbytes - 1; - do_div(minseg, segbytes); + minseg = div64_ul(minseg, segbytes); if (range[1] < 4096) goto out; @@ -1120,7 +1120,7 @@ static int nilfs_ioctl_set_alloc_range(struct inode *inode, void __user *argp) if (maxseg < segbytes) goto out; - do_div(maxseg, segbytes); + maxseg = div64_ul(maxseg, segbytes); maxseg--; ret = nilfs_sufile_set_alloc_range(nilfs->ns_sufile, minseg, maxseg); diff --git a/fs/nilfs2/mdt.c b/fs/nilfs2/mdt.c index e45c01a559c0..4f792a0ad0f0 100644 --- a/fs/nilfs2/mdt.c +++ b/fs/nilfs2/mdt.c @@ -47,12 +47,12 @@ nilfs_mdt_insert_new_block(struct inode *inode, unsigned long block, set_buffer_mapped(bh); - kaddr = kmap_atomic(bh->b_page); + kaddr = kmap_local_page(bh->b_page); memset(kaddr + bh_offset(bh), 0, i_blocksize(inode)); if (init_block) init_block(inode, bh, kaddr); flush_dcache_page(bh->b_page); - kunmap_atomic(kaddr); + kunmap_local(kaddr); set_buffer_uptodate(bh); mark_buffer_dirty(bh); diff --git a/fs/nilfs2/nilfs.h b/fs/nilfs2/nilfs.h index 98cffaf0ac12..2e29b98ba8ba 100644 --- a/fs/nilfs2/nilfs.h +++ b/fs/nilfs2/nilfs.h @@ -256,7 +256,8 @@ extern struct inode *nilfs_new_inode(struct inode *, umode_t); extern int nilfs_get_block(struct inode *, sector_t, struct buffer_head *, int); extern void nilfs_set_inode_flags(struct inode *); extern int nilfs_read_inode_common(struct inode *, struct nilfs_inode *); -extern void nilfs_write_inode_common(struct inode *, struct nilfs_inode *, int); +void nilfs_write_inode_common(struct inode *inode, + struct nilfs_inode *raw_inode); struct inode *nilfs_ilookup(struct super_block *sb, struct nilfs_root *root, unsigned long ino); struct inode *nilfs_iget_locked(struct super_block *sb, struct nilfs_root *root, diff --git a/fs/nilfs2/page.c b/fs/nilfs2/page.c index 5c2eba1987bd..14e470fb8870 100644 --- a/fs/nilfs2/page.c +++ b/fs/nilfs2/page.c @@ -103,11 +103,11 @@ void nilfs_copy_buffer(struct buffer_head *dbh, struct buffer_head *sbh) struct page *spage = sbh->b_page, *dpage = dbh->b_page; struct buffer_head *bh; - kaddr0 = kmap_atomic(spage); - kaddr1 = kmap_atomic(dpage); + kaddr0 = kmap_local_page(spage); + kaddr1 = kmap_local_page(dpage); memcpy(kaddr1 + bh_offset(dbh), kaddr0 + bh_offset(sbh), sbh->b_size); - kunmap_atomic(kaddr1); - kunmap_atomic(kaddr0); + kunmap_local(kaddr1); + kunmap_local(kaddr0); dbh->b_state = sbh->b_state & NILFS_BUFFER_INHERENT_BITS; dbh->b_blocknr = sbh->b_blocknr; diff --git a/fs/nilfs2/recovery.c b/fs/nilfs2/recovery.c index a9b8d77c8c1d..49a70c68bf3c 100644 --- a/fs/nilfs2/recovery.c +++ b/fs/nilfs2/recovery.c @@ -482,9 +482,9 @@ static int nilfs_recovery_copy_block(struct the_nilfs *nilfs, if (unlikely(!bh_org)) return -EIO; - kaddr = kmap_atomic(page); + kaddr = kmap_local_page(page); memcpy(kaddr + from, bh_org->b_data, bh_org->b_size); - kunmap_atomic(kaddr); + kunmap_local(kaddr); brelse(bh_org); return 0; } diff --git a/fs/nilfs2/segbuf.c b/fs/nilfs2/segbuf.c index 6e59dc19a732..dc431b4c34c9 100644 --- a/fs/nilfs2/segbuf.c +++ b/fs/nilfs2/segbuf.c @@ -220,9 +220,9 @@ static void nilfs_segbuf_fill_in_data_crc(struct nilfs_segment_buffer *segbuf, crc = crc32_le(crc, bh->b_data, bh->b_size); } list_for_each_entry(bh, &segbuf->sb_payload_buffers, b_assoc_buffers) { - kaddr = kmap_atomic(bh->b_page); + kaddr = kmap_local_page(bh->b_page); crc = crc32_le(crc, kaddr + bh_offset(bh), bh->b_size); - kunmap_atomic(kaddr); + kunmap_local(kaddr); } raw_sum->ss_datasum = cpu_to_le32(crc); } diff --git a/fs/nilfs2/segment.c b/fs/nilfs2/segment.c index 2bfb08052d39..aa5290cb7467 100644 --- a/fs/nilfs2/segment.c +++ b/fs/nilfs2/segment.c @@ -880,76 +880,6 @@ static void nilfs_segctor_clear_metadata_dirty(struct nilfs_sc_info *sci) nilfs_mdt_clear_dirty(nilfs->ns_dat); } -static int nilfs_segctor_create_checkpoint(struct nilfs_sc_info *sci) -{ - struct the_nilfs *nilfs = sci->sc_super->s_fs_info; - struct buffer_head *bh_cp; - struct nilfs_checkpoint *raw_cp; - int err; - - /* XXX: this interface will be changed */ - err = nilfs_cpfile_get_checkpoint(nilfs->ns_cpfile, nilfs->ns_cno, 1, - &raw_cp, &bh_cp); - if (likely(!err)) { - /* - * The following code is duplicated with cpfile. But, it is - * needed to collect the checkpoint even if it was not newly - * created. - */ - mark_buffer_dirty(bh_cp); - nilfs_mdt_mark_dirty(nilfs->ns_cpfile); - nilfs_cpfile_put_checkpoint( - nilfs->ns_cpfile, nilfs->ns_cno, bh_cp); - } else if (err == -EINVAL || err == -ENOENT) { - nilfs_error(sci->sc_super, - "checkpoint creation failed due to metadata corruption."); - err = -EIO; - } - return err; -} - -static int nilfs_segctor_fill_in_checkpoint(struct nilfs_sc_info *sci) -{ - struct the_nilfs *nilfs = sci->sc_super->s_fs_info; - struct buffer_head *bh_cp; - struct nilfs_checkpoint *raw_cp; - int err; - - err = nilfs_cpfile_get_checkpoint(nilfs->ns_cpfile, nilfs->ns_cno, 0, - &raw_cp, &bh_cp); - if (unlikely(err)) { - if (err == -EINVAL || err == -ENOENT) { - nilfs_error(sci->sc_super, - "checkpoint finalization failed due to metadata corruption."); - err = -EIO; - } - goto failed_ibh; - } - raw_cp->cp_snapshot_list.ssl_next = 0; - raw_cp->cp_snapshot_list.ssl_prev = 0; - raw_cp->cp_inodes_count = - cpu_to_le64(atomic64_read(&sci->sc_root->inodes_count)); - raw_cp->cp_blocks_count = - cpu_to_le64(atomic64_read(&sci->sc_root->blocks_count)); - raw_cp->cp_nblk_inc = - cpu_to_le64(sci->sc_nblk_inc + sci->sc_nblk_this_inc); - raw_cp->cp_create = cpu_to_le64(sci->sc_seg_ctime); - raw_cp->cp_cno = cpu_to_le64(nilfs->ns_cno); - - if (test_bit(NILFS_SC_HAVE_DELTA, &sci->sc_flags)) - nilfs_checkpoint_clear_minor(raw_cp); - else - nilfs_checkpoint_set_minor(raw_cp); - - nilfs_write_inode_common(sci->sc_root->ifile, - &raw_cp->cp_ifile_inode, 1); - nilfs_cpfile_put_checkpoint(nilfs->ns_cpfile, nilfs->ns_cno, bh_cp); - return 0; - - failed_ibh: - return err; -} - static void nilfs_fill_in_file_bmap(struct inode *ifile, struct nilfs_inode_info *ii) @@ -963,7 +893,7 @@ static void nilfs_fill_in_file_bmap(struct inode *ifile, raw_inode = nilfs_ifile_map_inode(ifile, ii->vfs_inode.i_ino, ibh); nilfs_bmap_write(ii->i_bmap, raw_inode); - nilfs_ifile_unmap_inode(ifile, ii->vfs_inode.i_ino, ibh); + nilfs_ifile_unmap_inode(raw_inode); } } @@ -977,6 +907,33 @@ static void nilfs_segctor_fill_in_file_bmap(struct nilfs_sc_info *sci) } } +/** + * nilfs_write_root_mdt_inode - export root metadata inode information to + * the on-disk inode + * @inode: inode object of the root metadata file + * @raw_inode: on-disk inode + * + * nilfs_write_root_mdt_inode() writes inode information and bmap data of + * @inode to the inode area of the metadata file allocated on the super root + * block created to finalize the log. Since super root blocks are configured + * each time, this function zero-fills the unused area of @raw_inode. + */ +static void nilfs_write_root_mdt_inode(struct inode *inode, + struct nilfs_inode *raw_inode) +{ + struct the_nilfs *nilfs = inode->i_sb->s_fs_info; + + nilfs_write_inode_common(inode, raw_inode); + + /* zero-fill unused portion of raw_inode */ + raw_inode->i_xattr = 0; + raw_inode->i_pad = 0; + memset((void *)raw_inode + sizeof(*raw_inode), 0, + nilfs->ns_inode_size - sizeof(*raw_inode)); + + nilfs_bmap_write(NILFS_I(inode)->i_bmap, raw_inode); +} + static void nilfs_segctor_fill_in_super_root(struct nilfs_sc_info *sci, struct the_nilfs *nilfs) { @@ -998,12 +955,13 @@ static void nilfs_segctor_fill_in_super_root(struct nilfs_sc_info *sci, nilfs->ns_nongc_ctime : sci->sc_seg_ctime); raw_sr->sr_flags = 0; - nilfs_write_inode_common(nilfs->ns_dat, (void *)raw_sr + - NILFS_SR_DAT_OFFSET(isz), 1); - nilfs_write_inode_common(nilfs->ns_cpfile, (void *)raw_sr + - NILFS_SR_CPFILE_OFFSET(isz), 1); - nilfs_write_inode_common(nilfs->ns_sufile, (void *)raw_sr + - NILFS_SR_SUFILE_OFFSET(isz), 1); + nilfs_write_root_mdt_inode(nilfs->ns_dat, (void *)raw_sr + + NILFS_SR_DAT_OFFSET(isz)); + nilfs_write_root_mdt_inode(nilfs->ns_cpfile, (void *)raw_sr + + NILFS_SR_CPFILE_OFFSET(isz)); + nilfs_write_root_mdt_inode(nilfs->ns_sufile, (void *)raw_sr + + NILFS_SR_SUFILE_OFFSET(isz)); + memset((void *)raw_sr + srsz, 0, nilfs->ns_blocksize - srsz); set_buffer_uptodate(bh_sr); unlock_buffer(bh_sr); @@ -1230,7 +1188,8 @@ static int nilfs_segctor_collect_blocks(struct nilfs_sc_info *sci, int mode) break; nilfs_sc_cstage_inc(sci); /* Creating a checkpoint */ - err = nilfs_segctor_create_checkpoint(sci); + err = nilfs_cpfile_create_checkpoint(nilfs->ns_cpfile, + nilfs->ns_cno); if (unlikely(err)) break; fallthrough; @@ -2101,7 +2060,11 @@ static int nilfs_segctor_do_construct(struct nilfs_sc_info *sci, int mode) if (mode == SC_LSEG_SR && nilfs_sc_cstage_get(sci) >= NILFS_ST_CPFILE) { - err = nilfs_segctor_fill_in_checkpoint(sci); + err = nilfs_cpfile_finalize_checkpoint( + nilfs->ns_cpfile, nilfs->ns_cno, sci->sc_root, + sci->sc_nblk_inc + sci->sc_nblk_this_inc, + sci->sc_seg_ctime, + !test_bit(NILFS_SC_HAVE_DELTA, &sci->sc_flags)); if (unlikely(err)) goto failed_to_write; diff --git a/fs/nilfs2/sufile.c b/fs/nilfs2/sufile.c index 0a8119456c21..6748218be7c5 100644 --- a/fs/nilfs2/sufile.c +++ b/fs/nilfs2/sufile.c @@ -48,7 +48,7 @@ nilfs_sufile_get_blkoff(const struct inode *sufile, __u64 segnum) { __u64 t = segnum + NILFS_MDT(sufile)->mi_first_entry_offset; - do_div(t, nilfs_sufile_segment_usages_per_block(sufile)); + t = div64_ul(t, nilfs_sufile_segment_usages_per_block(sufile)); return (unsigned long)t; } @@ -107,11 +107,11 @@ static void nilfs_sufile_mod_counter(struct buffer_head *header_bh, struct nilfs_sufile_header *header; void *kaddr; - kaddr = kmap_atomic(header_bh->b_page); + kaddr = kmap_local_page(header_bh->b_page); header = kaddr + bh_offset(header_bh); le64_add_cpu(&header->sh_ncleansegs, ncleanadd); le64_add_cpu(&header->sh_ndirtysegs, ndirtyadd); - kunmap_atomic(kaddr); + kunmap_local(kaddr); mark_buffer_dirty(header_bh); } @@ -315,10 +315,10 @@ int nilfs_sufile_alloc(struct inode *sufile, __u64 *segnump) ret = nilfs_sufile_get_header_block(sufile, &header_bh); if (ret < 0) goto out_sem; - kaddr = kmap_atomic(header_bh->b_page); + kaddr = kmap_local_page(header_bh->b_page); header = kaddr + bh_offset(header_bh); last_alloc = le64_to_cpu(header->sh_last_alloc); - kunmap_atomic(kaddr); + kunmap_local(kaddr); nsegments = nilfs_sufile_get_nsegments(sufile); maxsegnum = sui->allocmax; @@ -352,7 +352,7 @@ int nilfs_sufile_alloc(struct inode *sufile, __u64 *segnump) &su_bh); if (ret < 0) goto out_header; - kaddr = kmap_atomic(su_bh->b_page); + kaddr = kmap_local_page(su_bh->b_page); su = nilfs_sufile_block_get_segment_usage( sufile, segnum, su_bh, kaddr); @@ -363,14 +363,14 @@ int nilfs_sufile_alloc(struct inode *sufile, __u64 *segnump) continue; /* found a clean segment */ nilfs_segment_usage_set_dirty(su); - kunmap_atomic(kaddr); + kunmap_local(kaddr); - kaddr = kmap_atomic(header_bh->b_page); + kaddr = kmap_local_page(header_bh->b_page); header = kaddr + bh_offset(header_bh); le64_add_cpu(&header->sh_ncleansegs, -1); le64_add_cpu(&header->sh_ndirtysegs, 1); header->sh_last_alloc = cpu_to_le64(segnum); - kunmap_atomic(kaddr); + kunmap_local(kaddr); sui->ncleansegs--; mark_buffer_dirty(header_bh); @@ -384,7 +384,7 @@ int nilfs_sufile_alloc(struct inode *sufile, __u64 *segnump) goto out_header; } - kunmap_atomic(kaddr); + kunmap_local(kaddr); brelse(su_bh); } @@ -406,16 +406,16 @@ void nilfs_sufile_do_cancel_free(struct inode *sufile, __u64 segnum, struct nilfs_segment_usage *su; void *kaddr; - kaddr = kmap_atomic(su_bh->b_page); + kaddr = kmap_local_page(su_bh->b_page); su = nilfs_sufile_block_get_segment_usage(sufile, segnum, su_bh, kaddr); if (unlikely(!nilfs_segment_usage_clean(su))) { nilfs_warn(sufile->i_sb, "%s: segment %llu must be clean", __func__, (unsigned long long)segnum); - kunmap_atomic(kaddr); + kunmap_local(kaddr); return; } nilfs_segment_usage_set_dirty(su); - kunmap_atomic(kaddr); + kunmap_local(kaddr); nilfs_sufile_mod_counter(header_bh, -1, 1); NILFS_SUI(sufile)->ncleansegs--; @@ -432,11 +432,11 @@ void nilfs_sufile_do_scrap(struct inode *sufile, __u64 segnum, void *kaddr; int clean, dirty; - kaddr = kmap_atomic(su_bh->b_page); + kaddr = kmap_local_page(su_bh->b_page); su = nilfs_sufile_block_get_segment_usage(sufile, segnum, su_bh, kaddr); if (su->su_flags == cpu_to_le32(BIT(NILFS_SEGMENT_USAGE_DIRTY)) && su->su_nblocks == cpu_to_le32(0)) { - kunmap_atomic(kaddr); + kunmap_local(kaddr); return; } clean = nilfs_segment_usage_clean(su); @@ -446,7 +446,7 @@ void nilfs_sufile_do_scrap(struct inode *sufile, __u64 segnum, su->su_lastmod = cpu_to_le64(0); su->su_nblocks = cpu_to_le32(0); su->su_flags = cpu_to_le32(BIT(NILFS_SEGMENT_USAGE_DIRTY)); - kunmap_atomic(kaddr); + kunmap_local(kaddr); nilfs_sufile_mod_counter(header_bh, clean ? (u64)-1 : 0, dirty ? 0 : 1); NILFS_SUI(sufile)->ncleansegs -= clean; @@ -463,12 +463,12 @@ void nilfs_sufile_do_free(struct inode *sufile, __u64 segnum, void *kaddr; int sudirty; - kaddr = kmap_atomic(su_bh->b_page); + kaddr = kmap_local_page(su_bh->b_page); su = nilfs_sufile_block_get_segment_usage(sufile, segnum, su_bh, kaddr); if (nilfs_segment_usage_clean(su)) { nilfs_warn(sufile->i_sb, "%s: segment %llu is already clean", __func__, (unsigned long long)segnum); - kunmap_atomic(kaddr); + kunmap_local(kaddr); return; } if (unlikely(nilfs_segment_usage_error(su))) @@ -481,7 +481,7 @@ void nilfs_sufile_do_free(struct inode *sufile, __u64 segnum, (unsigned long long)segnum); nilfs_segment_usage_set_clean(su); - kunmap_atomic(kaddr); + kunmap_local(kaddr); mark_buffer_dirty(su_bh); nilfs_sufile_mod_counter(header_bh, 1, sudirty ? (u64)-1 : 0); @@ -509,12 +509,12 @@ int nilfs_sufile_mark_dirty(struct inode *sufile, __u64 segnum) if (ret) goto out_sem; - kaddr = kmap_atomic(bh->b_page); + kaddr = kmap_local_page(bh->b_page); su = nilfs_sufile_block_get_segment_usage(sufile, segnum, bh, kaddr); if (unlikely(nilfs_segment_usage_error(su))) { struct the_nilfs *nilfs = sufile->i_sb->s_fs_info; - kunmap_atomic(kaddr); + kunmap_local(kaddr); brelse(bh); if (nilfs_segment_is_active(nilfs, segnum)) { nilfs_error(sufile->i_sb, @@ -532,7 +532,7 @@ int nilfs_sufile_mark_dirty(struct inode *sufile, __u64 segnum) ret = -EIO; } else { nilfs_segment_usage_set_dirty(su); - kunmap_atomic(kaddr); + kunmap_local(kaddr); mark_buffer_dirty(bh); nilfs_mdt_mark_dirty(sufile); brelse(bh); @@ -562,7 +562,7 @@ int nilfs_sufile_set_segment_usage(struct inode *sufile, __u64 segnum, if (ret < 0) goto out_sem; - kaddr = kmap_atomic(bh->b_page); + kaddr = kmap_local_page(bh->b_page); su = nilfs_sufile_block_get_segment_usage(sufile, segnum, bh, kaddr); if (modtime) { /* @@ -573,7 +573,7 @@ int nilfs_sufile_set_segment_usage(struct inode *sufile, __u64 segnum, su->su_lastmod = cpu_to_le64(modtime); } su->su_nblocks = cpu_to_le32(nblocks); - kunmap_atomic(kaddr); + kunmap_local(kaddr); mark_buffer_dirty(bh); nilfs_mdt_mark_dirty(sufile); @@ -614,7 +614,7 @@ int nilfs_sufile_get_stat(struct inode *sufile, struct nilfs_sustat *sustat) if (ret < 0) goto out_sem; - kaddr = kmap_atomic(header_bh->b_page); + kaddr = kmap_local_page(header_bh->b_page); header = kaddr + bh_offset(header_bh); sustat->ss_nsegs = nilfs_sufile_get_nsegments(sufile); sustat->ss_ncleansegs = le64_to_cpu(header->sh_ncleansegs); @@ -624,7 +624,7 @@ int nilfs_sufile_get_stat(struct inode *sufile, struct nilfs_sustat *sustat) spin_lock(&nilfs->ns_last_segment_lock); sustat->ss_prot_seq = nilfs->ns_prot_seq; spin_unlock(&nilfs->ns_last_segment_lock); - kunmap_atomic(kaddr); + kunmap_local(kaddr); brelse(header_bh); out_sem: @@ -640,15 +640,15 @@ void nilfs_sufile_do_set_error(struct inode *sufile, __u64 segnum, void *kaddr; int suclean; - kaddr = kmap_atomic(su_bh->b_page); + kaddr = kmap_local_page(su_bh->b_page); su = nilfs_sufile_block_get_segment_usage(sufile, segnum, su_bh, kaddr); if (nilfs_segment_usage_error(su)) { - kunmap_atomic(kaddr); + kunmap_local(kaddr); return; } suclean = nilfs_segment_usage_clean(su); nilfs_segment_usage_set_error(su); - kunmap_atomic(kaddr); + kunmap_local(kaddr); if (suclean) { nilfs_sufile_mod_counter(header_bh, -1, 0); @@ -717,7 +717,7 @@ static int nilfs_sufile_truncate_range(struct inode *sufile, /* hole */ continue; } - kaddr = kmap_atomic(su_bh->b_page); + kaddr = kmap_local_page(su_bh->b_page); su = nilfs_sufile_block_get_segment_usage( sufile, segnum, su_bh, kaddr); su2 = su; @@ -726,7 +726,7 @@ static int nilfs_sufile_truncate_range(struct inode *sufile, ~BIT(NILFS_SEGMENT_USAGE_ERROR)) || nilfs_segment_is_active(nilfs, segnum + j)) { ret = -EBUSY; - kunmap_atomic(kaddr); + kunmap_local(kaddr); brelse(su_bh); goto out_header; } @@ -738,7 +738,7 @@ static int nilfs_sufile_truncate_range(struct inode *sufile, nc++; } } - kunmap_atomic(kaddr); + kunmap_local(kaddr); if (nc > 0) { mark_buffer_dirty(su_bh); ncleaned += nc; @@ -823,10 +823,10 @@ int nilfs_sufile_resize(struct inode *sufile, __u64 newnsegs) sui->allocmin = 0; } - kaddr = kmap_atomic(header_bh->b_page); + kaddr = kmap_local_page(header_bh->b_page); header = kaddr + bh_offset(header_bh); header->sh_ncleansegs = cpu_to_le64(sui->ncleansegs); - kunmap_atomic(kaddr); + kunmap_local(kaddr); mark_buffer_dirty(header_bh); nilfs_mdt_mark_dirty(sufile); @@ -891,7 +891,7 @@ ssize_t nilfs_sufile_get_suinfo(struct inode *sufile, __u64 segnum, void *buf, continue; } - kaddr = kmap_atomic(su_bh->b_page); + kaddr = kmap_local_page(su_bh->b_page); su = nilfs_sufile_block_get_segment_usage( sufile, segnum, su_bh, kaddr); for (j = 0; j < n; @@ -904,7 +904,7 @@ ssize_t nilfs_sufile_get_suinfo(struct inode *sufile, __u64 segnum, void *buf, si->sui_flags |= BIT(NILFS_SEGMENT_USAGE_ACTIVE); } - kunmap_atomic(kaddr); + kunmap_local(kaddr); brelse(su_bh); } ret = nsegs; @@ -973,7 +973,7 @@ ssize_t nilfs_sufile_set_suinfo(struct inode *sufile, void *buf, goto out_header; for (;;) { - kaddr = kmap_atomic(bh->b_page); + kaddr = kmap_local_page(bh->b_page); su = nilfs_sufile_block_get_segment_usage( sufile, sup->sup_segnum, bh, kaddr); @@ -1010,7 +1010,7 @@ ssize_t nilfs_sufile_set_suinfo(struct inode *sufile, void *buf, su->su_flags = cpu_to_le32(sup->sup_sui.sui_flags); } - kunmap_atomic(kaddr); + kunmap_local(kaddr); sup = (void *)sup + supsz; if (sup >= supend) @@ -1115,7 +1115,7 @@ int nilfs_sufile_trim_fs(struct inode *sufile, struct fstrim_range *range) continue; } - kaddr = kmap_atomic(su_bh->b_page); + kaddr = kmap_local_page(su_bh->b_page); su = nilfs_sufile_block_get_segment_usage(sufile, segnum, su_bh, kaddr); for (i = 0; i < n; ++i, ++segnum, su = (void *)su + susz) { @@ -1145,7 +1145,7 @@ int nilfs_sufile_trim_fs(struct inode *sufile, struct fstrim_range *range) } if (nblocks >= minlen) { - kunmap_atomic(kaddr); + kunmap_local(kaddr); ret = blkdev_issue_discard(nilfs->ns_bdev, start * sects_per_block, @@ -1157,7 +1157,7 @@ int nilfs_sufile_trim_fs(struct inode *sufile, struct fstrim_range *range) } ndiscarded += nblocks; - kaddr = kmap_atomic(su_bh->b_page); + kaddr = kmap_local_page(su_bh->b_page); su = nilfs_sufile_block_get_segment_usage( sufile, segnum, su_bh, kaddr); } @@ -1166,7 +1166,7 @@ int nilfs_sufile_trim_fs(struct inode *sufile, struct fstrim_range *range) start = seg_start; nblocks = seg_end - seg_start + 1; } - kunmap_atomic(kaddr); + kunmap_local(kaddr); put_bh(su_bh); } @@ -1246,10 +1246,10 @@ int nilfs_sufile_read(struct super_block *sb, size_t susize, goto failed; sui = NILFS_SUI(sufile); - kaddr = kmap_atomic(header_bh->b_page); + kaddr = kmap_local_page(header_bh->b_page); header = kaddr + bh_offset(header_bh); sui->ncleansegs = le64_to_cpu(header->sh_ncleansegs); - kunmap_atomic(kaddr); + kunmap_local(kaddr); brelse(header_bh); sui->allocmax = nilfs_sufile_get_nsegments(sufile) - 1; diff --git a/fs/nilfs2/super.c b/fs/nilfs2/super.c index df8674173b22..ac24ed109ce9 100644 --- a/fs/nilfs2/super.c +++ b/fs/nilfs2/super.c @@ -448,7 +448,7 @@ int nilfs_resize_fs(struct super_block *sb, __u64 newsize) sb2off = NILFS_SB2_OFFSET_BYTES(newsize); newnsegs = sb2off >> nilfs->ns_blocksize_bits; - do_div(newnsegs, nilfs->ns_blocks_per_segment); + newnsegs = div64_ul(newnsegs, nilfs->ns_blocks_per_segment); ret = nilfs_sufile_resize(nilfs->ns_sufile, newnsegs); up_write(&nilfs->ns_segctor_sem); @@ -544,8 +544,6 @@ int nilfs_attach_checkpoint(struct super_block *sb, __u64 cno, int curr_mnt, { struct the_nilfs *nilfs = sb->s_fs_info; struct nilfs_root *root; - struct nilfs_checkpoint *raw_cp; - struct buffer_head *bh_cp; int err = -ENOMEM; root = nilfs_find_or_create_root( @@ -557,38 +555,19 @@ int nilfs_attach_checkpoint(struct super_block *sb, __u64 cno, int curr_mnt, goto reuse; /* already attached checkpoint */ down_read(&nilfs->ns_segctor_sem); - err = nilfs_cpfile_get_checkpoint(nilfs->ns_cpfile, cno, 0, &raw_cp, - &bh_cp); + err = nilfs_ifile_read(sb, root, cno, nilfs->ns_inode_size); up_read(&nilfs->ns_segctor_sem); - if (unlikely(err)) { - if (err == -ENOENT || err == -EINVAL) { - nilfs_err(sb, - "Invalid checkpoint (checkpoint number=%llu)", - (unsigned long long)cno); - err = -EINVAL; - } + if (unlikely(err)) goto failed; - } - - err = nilfs_ifile_read(sb, root, nilfs->ns_inode_size, - &raw_cp->cp_ifile_inode, &root->ifile); - if (err) - goto failed_bh; - - atomic64_set(&root->inodes_count, - le64_to_cpu(raw_cp->cp_inodes_count)); - atomic64_set(&root->blocks_count, - le64_to_cpu(raw_cp->cp_blocks_count)); - - nilfs_cpfile_put_checkpoint(nilfs->ns_cpfile, cno, bh_cp); reuse: *rootp = root; return 0; - failed_bh: - nilfs_cpfile_put_checkpoint(nilfs->ns_cpfile, cno, bh_cp); failed: + if (err == -EINVAL) + nilfs_err(sb, "Invalid checkpoint (checkpoint number=%llu)", + (unsigned long long)cno); nilfs_put_root(root); return err; diff --git a/fs/nilfs2/the_nilfs.c b/fs/nilfs2/the_nilfs.c index 71400496ed36..2ae2c1bbf6d1 100644 --- a/fs/nilfs2/the_nilfs.c +++ b/fs/nilfs2/the_nilfs.c @@ -413,7 +413,7 @@ static u64 nilfs_max_segment_count(struct the_nilfs *nilfs) { u64 max_count = U64_MAX; - do_div(max_count, nilfs->ns_blocks_per_segment); + max_count = div64_ul(max_count, nilfs->ns_blocks_per_segment); return min_t(u64, max_count, ULONG_MAX); } diff --git a/fs/notify/fanotify/fanotify.c b/fs/notify/fanotify/fanotify.c index 1e4def21811e..224bccaab4cc 100644 --- a/fs/notify/fanotify/fanotify.c +++ b/fs/notify/fanotify/fanotify.c @@ -228,8 +228,10 @@ static int fanotify_get_response(struct fsnotify_group *group, pr_debug("%s: group=%p event=%p\n", __func__, group, event); - ret = wait_event_killable(group->fanotify_data.access_waitq, - event->state == FAN_EVENT_ANSWERED); + ret = wait_event_state(group->fanotify_data.access_waitq, + event->state == FAN_EVENT_ANSWERED, + (TASK_KILLABLE|TASK_FREEZABLE)); + /* Signal pending? */ if (ret < 0) { spin_lock(&group->notification_lock); diff --git a/fs/notify/fsnotify.c b/fs/notify/fsnotify.c index 8bfd690e9f10..2fc105a72a8f 100644 --- a/fs/notify/fsnotify.c +++ b/fs/notify/fsnotify.c @@ -141,7 +141,7 @@ void __fsnotify_update_child_dentry_flags(struct inode *inode) } /* Are inode/sb/mount interested in parent and name info with this event? */ -static bool fsnotify_event_needs_parent(struct inode *inode, struct mount *mnt, +static bool fsnotify_event_needs_parent(struct inode *inode, __u32 mnt_mask, __u32 mask) { __u32 marks_mask = 0; @@ -160,13 +160,22 @@ static bool fsnotify_event_needs_parent(struct inode *inode, struct mount *mnt, /* Did either inode/sb/mount subscribe for events with parent/name? */ marks_mask |= fsnotify_parent_needed_mask(inode->i_fsnotify_mask); marks_mask |= fsnotify_parent_needed_mask(inode->i_sb->s_fsnotify_mask); - if (mnt) - marks_mask |= fsnotify_parent_needed_mask(mnt->mnt_fsnotify_mask); + marks_mask |= fsnotify_parent_needed_mask(mnt_mask); /* Did they subscribe for this event with parent/name info? */ return mask & marks_mask; } +/* Are there any inode/mount/sb objects that are interested in this event? */ +static inline bool fsnotify_object_watched(struct inode *inode, __u32 mnt_mask, + __u32 mask) +{ + __u32 marks_mask = inode->i_fsnotify_mask | mnt_mask | + inode->i_sb->s_fsnotify_mask; + + return mask & marks_mask & ALL_FSNOTIFY_EVENTS; +} + /* * Notify this dentry's parent about a child's events with child name info * if parent is watching or if inode/sb/mount are interested in events with @@ -179,7 +188,7 @@ int __fsnotify_parent(struct dentry *dentry, __u32 mask, const void *data, int data_type) { const struct path *path = fsnotify_data_path(data, data_type); - struct mount *mnt = path ? real_mount(path->mnt) : NULL; + __u32 mnt_mask = path ? real_mount(path->mnt)->mnt_fsnotify_mask : 0; struct inode *inode = d_inode(dentry); struct dentry *parent; bool parent_watched = dentry->d_flags & DCACHE_FSNOTIFY_PARENT_WATCHED; @@ -190,16 +199,13 @@ int __fsnotify_parent(struct dentry *dentry, __u32 mask, const void *data, struct qstr *file_name = NULL; int ret = 0; - /* - * Do inode/sb/mount care about parent and name info on non-dir? - * Do they care about any event at all? - */ - if (!inode->i_fsnotify_marks && !inode->i_sb->s_fsnotify_marks && - (!mnt || !mnt->mnt_fsnotify_marks) && !parent_watched) + /* Optimize the likely case of nobody watching this path */ + if (likely(!parent_watched && + !fsnotify_object_watched(inode, mnt_mask, mask))) return 0; parent = NULL; - parent_needed = fsnotify_event_needs_parent(inode, mnt, mask); + parent_needed = fsnotify_event_needs_parent(inode, mnt_mask, mask); if (!parent_watched && !parent_needed) goto notify; diff --git a/fs/ocfs2/dlmglue.c b/fs/ocfs2/dlmglue.c index 64a6ef638495..cb40cafbc062 100644 --- a/fs/ocfs2/dlmglue.c +++ b/fs/ocfs2/dlmglue.c @@ -1615,7 +1615,7 @@ update_holders: unlock: lockres_clear_flags(lockres, OCFS2_LOCK_UPCONVERT_FINISHING); - /* ocfs2_unblock_lock reques on seeing OCFS2_LOCK_UPCONVERT_FINISHING */ + /* ocfs2_unblock_lock request on seeing OCFS2_LOCK_UPCONVERT_FINISHING */ kick_dc = (lockres->l_flags & OCFS2_LOCK_BLOCKED); spin_unlock_irqrestore(&lockres->l_lock, flags); diff --git a/fs/ocfs2/file.c b/fs/ocfs2/file.c index 8b6d15010703..0da8e7bd3261 100644 --- a/fs/ocfs2/file.c +++ b/fs/ocfs2/file.c @@ -2763,6 +2763,7 @@ const struct inode_operations ocfs2_file_iops = { const struct inode_operations ocfs2_special_file_iops = { .setattr = ocfs2_setattr, .getattr = ocfs2_getattr, + .listxattr = ocfs2_listxattr, .permission = ocfs2_permission, .get_inode_acl = ocfs2_iop_get_acl, .set_acl = ocfs2_iop_set_acl, diff --git a/fs/ocfs2/inode.h b/fs/ocfs2/inode.h index 82b28fdacc7e..accf03d4765e 100644 --- a/fs/ocfs2/inode.h +++ b/fs/ocfs2/inode.h @@ -65,7 +65,7 @@ struct ocfs2_inode_info tid_t i_sync_tid; tid_t i_datasync_tid; - struct dquot *i_dquot[MAXQUOTAS]; + struct dquot __rcu *i_dquot[MAXQUOTAS]; }; /* diff --git a/fs/ocfs2/quota_global.c b/fs/ocfs2/quota_global.c index dc9f76ab7e13..0575c2d060eb 100644 --- a/fs/ocfs2/quota_global.c +++ b/fs/ocfs2/quota_global.c @@ -447,14 +447,17 @@ int ocfs2_global_write_info(struct super_block *sb, int type) int err; struct quota_info *dqopt = sb_dqopt(sb); struct ocfs2_mem_dqinfo *info = dqopt->info[type].dqi_priv; + unsigned int memalloc; down_write(&dqopt->dqio_sem); + memalloc = memalloc_nofs_save(); err = ocfs2_qinfo_lock(info, 1); if (err < 0) goto out_sem; err = __ocfs2_global_write_info(sb, type); ocfs2_qinfo_unlock(info, 1); out_sem: + memalloc_nofs_restore(memalloc); up_write(&dqopt->dqio_sem); return err; } @@ -601,6 +604,7 @@ static int ocfs2_sync_dquot_helper(struct dquot *dquot, unsigned long type) struct ocfs2_mem_dqinfo *oinfo = sb_dqinfo(sb, type)->dqi_priv; struct ocfs2_super *osb = OCFS2_SB(sb); int status = 0; + unsigned int memalloc; trace_ocfs2_sync_dquot_helper(from_kqid(&init_user_ns, dquot->dq_id), dquot->dq_id.type, @@ -618,6 +622,7 @@ static int ocfs2_sync_dquot_helper(struct dquot *dquot, unsigned long type) goto out_ilock; } down_write(&sb_dqopt(sb)->dqio_sem); + memalloc = memalloc_nofs_save(); status = ocfs2_sync_dquot(dquot); if (status < 0) mlog_errno(status); @@ -625,6 +630,7 @@ static int ocfs2_sync_dquot_helper(struct dquot *dquot, unsigned long type) status = ocfs2_local_write_dquot(dquot); if (status < 0) mlog_errno(status); + memalloc_nofs_restore(memalloc); up_write(&sb_dqopt(sb)->dqio_sem); ocfs2_commit_trans(osb, handle); out_ilock: @@ -662,6 +668,7 @@ static int ocfs2_write_dquot(struct dquot *dquot) handle_t *handle; struct ocfs2_super *osb = OCFS2_SB(dquot->dq_sb); int status = 0; + unsigned int memalloc; trace_ocfs2_write_dquot(from_kqid(&init_user_ns, dquot->dq_id), dquot->dq_id.type); @@ -673,7 +680,9 @@ static int ocfs2_write_dquot(struct dquot *dquot) goto out; } down_write(&sb_dqopt(dquot->dq_sb)->dqio_sem); + memalloc = memalloc_nofs_save(); status = ocfs2_local_write_dquot(dquot); + memalloc_nofs_restore(memalloc); up_write(&sb_dqopt(dquot->dq_sb)->dqio_sem); ocfs2_commit_trans(osb, handle); out: @@ -920,6 +929,7 @@ static int ocfs2_mark_dquot_dirty(struct dquot *dquot) struct ocfs2_mem_dqinfo *oinfo = sb_dqinfo(sb, type)->dqi_priv; handle_t *handle; struct ocfs2_super *osb = OCFS2_SB(sb); + unsigned int memalloc; trace_ocfs2_mark_dquot_dirty(from_kqid(&init_user_ns, dquot->dq_id), type); @@ -946,6 +956,7 @@ static int ocfs2_mark_dquot_dirty(struct dquot *dquot) goto out_ilock; } down_write(&sb_dqopt(sb)->dqio_sem); + memalloc = memalloc_nofs_save(); status = ocfs2_sync_dquot(dquot); if (status < 0) { mlog_errno(status); @@ -954,6 +965,7 @@ static int ocfs2_mark_dquot_dirty(struct dquot *dquot) /* Now write updated local dquot structure */ status = ocfs2_local_write_dquot(dquot); out_dlock: + memalloc_nofs_restore(memalloc); up_write(&sb_dqopt(sb)->dqio_sem); ocfs2_commit_trans(osb, handle); out_ilock: diff --git a/fs/ocfs2/quota_local.c b/fs/ocfs2/quota_local.c index e09842fc9d4d..8ce462c64c51 100644 --- a/fs/ocfs2/quota_local.c +++ b/fs/ocfs2/quota_local.c @@ -470,6 +470,7 @@ static int ocfs2_recover_local_quota_file(struct inode *lqinode, int bit, chunk; struct ocfs2_recovery_chunk *rchunk, *next; qsize_t spacechange, inodechange; + unsigned int memalloc; trace_ocfs2_recover_local_quota_file((unsigned long)lqinode->i_ino, type); @@ -521,6 +522,7 @@ static int ocfs2_recover_local_quota_file(struct inode *lqinode, goto out_drop_lock; } down_write(&sb_dqopt(sb)->dqio_sem); + memalloc = memalloc_nofs_save(); spin_lock(&dquot->dq_dqb_lock); /* Add usage from quota entry into quota changes * of our node. Auxiliary variables are important @@ -553,6 +555,7 @@ static int ocfs2_recover_local_quota_file(struct inode *lqinode, unlock_buffer(qbh); ocfs2_journal_dirty(handle, qbh); out_commit: + memalloc_nofs_restore(memalloc); up_write(&sb_dqopt(sb)->dqio_sem); ocfs2_commit_trans(OCFS2_SB(sb), handle); out_drop_lock: diff --git a/fs/ocfs2/super.c b/fs/ocfs2/super.c index b3f860888e93..8aabaed2c1cb 100644 --- a/fs/ocfs2/super.c +++ b/fs/ocfs2/super.c @@ -122,7 +122,7 @@ static int ocfs2_susp_quotas(struct ocfs2_super *osb, int unsuspend); static int ocfs2_enable_quotas(struct ocfs2_super *osb); static void ocfs2_disable_quotas(struct ocfs2_super *osb); -static struct dquot **ocfs2_get_dquots(struct inode *inode) +static struct dquot __rcu **ocfs2_get_dquots(struct inode *inode) { return OCFS2_I(inode)->i_dquot; } @@ -1711,12 +1711,12 @@ static int ocfs2_initialize_mem_caches(void) ocfs2_dquot_cachep = kmem_cache_create("ocfs2_dquot_cache", sizeof(struct ocfs2_dquot), 0, - (SLAB_HWCACHE_ALIGN|SLAB_RECLAIM_ACCOUNT), + SLAB_HWCACHE_ALIGN|SLAB_RECLAIM_ACCOUNT, NULL); ocfs2_qf_chunk_cachep = kmem_cache_create("ocfs2_qf_chunk_cache", sizeof(struct ocfs2_quota_chunk), 0, - (SLAB_RECLAIM_ACCOUNT), + SLAB_RECLAIM_ACCOUNT, NULL); if (!ocfs2_inode_cachep || !ocfs2_dquot_cachep || !ocfs2_qf_chunk_cachep) { diff --git a/fs/orangefs/orangefs-cache.c b/fs/orangefs/orangefs-cache.c index 3b6982bf6bcf..e75e173a9186 100644 --- a/fs/orangefs/orangefs-cache.c +++ b/fs/orangefs/orangefs-cache.c @@ -22,7 +22,7 @@ int op_cache_initialize(void) op_cache = kmem_cache_create("orangefs_op_cache", sizeof(struct orangefs_kernel_op_s), 0, - ORANGEFS_CACHE_CREATE_FLAGS, + 0, NULL); if (!op_cache) { diff --git a/fs/orangefs/orangefs-kernel.h b/fs/orangefs/orangefs-kernel.h index 926d9c0a428a..e2df7eeadc7a 100644 --- a/fs/orangefs/orangefs-kernel.h +++ b/fs/orangefs/orangefs-kernel.h @@ -93,16 +93,6 @@ enum orangefs_vfs_op_states { OP_VFS_STATE_GIVEN_UP = 16, }; -/* - * orangefs kernel memory related flags - */ - -#if (defined CONFIG_DEBUG_SLAB) -#define ORANGEFS_CACHE_CREATE_FLAGS SLAB_RED_ZONE -#else -#define ORANGEFS_CACHE_CREATE_FLAGS 0 -#endif - extern const struct xattr_handler * const orangefs_xattr_handlers[]; extern struct posix_acl *orangefs_get_acl(struct inode *inode, int type, bool rcu); diff --git a/fs/orangefs/super.c b/fs/orangefs/super.c index 5254256a224d..34849b4a3243 100644 --- a/fs/orangefs/super.c +++ b/fs/orangefs/super.c @@ -527,7 +527,7 @@ struct dentry *orangefs_mount(struct file_system_type *fst, sb->s_fs_info = kzalloc(sizeof(struct orangefs_sb_info_s), GFP_KERNEL); if (!ORANGEFS_SB(sb)) { d = ERR_PTR(-ENOMEM); - goto free_sb_and_op; + goto free_op; } ret = orangefs_fill_sb(sb, @@ -644,7 +644,7 @@ int orangefs_inode_cache_initialize(void) "orangefs_inode_cache", sizeof(struct orangefs_inode_s), 0, - ORANGEFS_CACHE_CREATE_FLAGS, + 0, offsetof(struct orangefs_inode_s, link_target), sizeof_field(struct orangefs_inode_s, diff --git a/fs/overlayfs/copy_up.c b/fs/overlayfs/copy_up.c index 8586e2f5d243..0762575a1e70 100644 --- a/fs/overlayfs/copy_up.c +++ b/fs/overlayfs/copy_up.c @@ -234,11 +234,11 @@ static int ovl_verify_area(loff_t pos, loff_t pos2, loff_t len, loff_t totlen) { loff_t tmp; - if (WARN_ON_ONCE(pos != pos2)) + if (pos != pos2) return -EIO; - if (WARN_ON_ONCE(pos < 0 || len < 0 || totlen < 0)) + if (pos < 0 || len < 0 || totlen < 0) return -EIO; - if (WARN_ON_ONCE(check_add_overflow(pos, len, &tmp))) + if (check_add_overflow(pos, len, &tmp)) return -EIO; return 0; } diff --git a/fs/proc/Kconfig b/fs/proc/Kconfig index 32b1116ae137..d80a1431ef7b 100644 --- a/fs/proc/Kconfig +++ b/fs/proc/Kconfig @@ -32,7 +32,7 @@ config PROC_FS config PROC_KCORE bool "/proc/kcore support" if !ARM depends on PROC_FS && MMU - select CRASH_CORE + select VMCORE_INFO help Provides a virtual ELF core file of the live kernel. This can be read with gdb and other ELF tools. No modifications can be diff --git a/fs/proc/kcore.c b/fs/proc/kcore.c index 6422e569b080..8e08a9a1b7ed 100644 --- a/fs/proc/kcore.c +++ b/fs/proc/kcore.c @@ -10,7 +10,7 @@ * Safe accesses to vmalloc/direct-mapped discontiguous areas, Kanoj Sarcar <kanoj@sgi.com> */ -#include <linux/crash_core.h> +#include <linux/vmcore_info.h> #include <linux/mm.h> #include <linux/proc_fs.h> #include <linux/kcore.h> diff --git a/fs/proc/task_mmu.c b/fs/proc/task_mmu.c index 3f78ebbb795f..23fbab954c20 100644 --- a/fs/proc/task_mmu.c +++ b/fs/proc/task_mmu.c @@ -1352,8 +1352,7 @@ static inline pagemap_entry_t make_pme(u64 frame, u64 flags) return (pagemap_entry_t) { .pme = (frame & PM_PFRAME_MASK) | flags }; } -static int add_to_pagemap(unsigned long addr, pagemap_entry_t *pme, - struct pagemapread *pm) +static int add_to_pagemap(pagemap_entry_t *pme, struct pagemapread *pm) { pm->buffer[pm->pos++] = *pme; if (pm->pos >= pm->len) @@ -1380,7 +1379,7 @@ static int pagemap_pte_hole(unsigned long start, unsigned long end, hole_end = end; for (; addr < hole_end; addr += PAGE_SIZE) { - err = add_to_pagemap(addr, &pme, pm); + err = add_to_pagemap(&pme, pm); if (err) goto out; } @@ -1392,7 +1391,7 @@ static int pagemap_pte_hole(unsigned long start, unsigned long end, if (vma->vm_flags & VM_SOFTDIRTY) pme = make_pme(0, PM_SOFT_DIRTY); for (; addr < min(end, vma->vm_end); addr += PAGE_SIZE) { - err = add_to_pagemap(addr, &pme, pm); + err = add_to_pagemap(&pme, pm); if (err) goto out; } @@ -1519,7 +1518,7 @@ static int pagemap_pmd_range(pmd_t *pmdp, unsigned long addr, unsigned long end, for (; addr != end; addr += PAGE_SIZE) { pagemap_entry_t pme = make_pme(frame, flags); - err = add_to_pagemap(addr, &pme, pm); + err = add_to_pagemap(&pme, pm); if (err) break; if (pm->show_pfn) { @@ -1547,7 +1546,7 @@ static int pagemap_pmd_range(pmd_t *pmdp, unsigned long addr, unsigned long end, pagemap_entry_t pme; pme = pte_to_pagemap_entry(pm, vma, addr, ptep_get(pte)); - err = add_to_pagemap(addr, &pme, pm); + err = add_to_pagemap(&pme, pm); if (err) break; } @@ -1597,7 +1596,7 @@ static int pagemap_hugetlb_range(pte_t *ptep, unsigned long hmask, for (; addr != end; addr += PAGE_SIZE) { pagemap_entry_t pme = make_pme(frame, flags); - err = add_to_pagemap(addr, &pme, pm); + err = add_to_pagemap(&pme, pm); if (err) return err; if (pm->show_pfn && (flags & PM_PRESENT)) @@ -1807,7 +1806,7 @@ static unsigned long pagemap_page_category(struct pagemap_scan_private *p, if (p->masks_of_interest & PAGE_IS_FILE) { swp = pte_to_swp_entry(pte); if (is_pfn_swap_entry(swp) && - !PageAnon(pfn_swap_entry_to_page(swp))) + !folio_test_anon(pfn_swap_entry_folio(swp))) categories |= PAGE_IS_FILE; } if (pte_swp_soft_dirty(pte)) @@ -1873,7 +1872,7 @@ static unsigned long pagemap_thp_category(struct pagemap_scan_private *p, if (p->masks_of_interest & PAGE_IS_FILE) { swp = pmd_to_swp_entry(pmd); if (is_pfn_swap_entry(swp) && - !PageAnon(pfn_swap_entry_to_page(swp))) + !folio_test_anon(pfn_swap_entry_folio(swp))) categories |= PAGE_IS_FILE; } } diff --git a/fs/quota/dquot.c b/fs/quota/dquot.c index eb6e9d95dea1..dacbee455c03 100644 --- a/fs/quota/dquot.c +++ b/fs/quota/dquot.c @@ -399,15 +399,17 @@ int dquot_mark_dquot_dirty(struct dquot *dquot) EXPORT_SYMBOL(dquot_mark_dquot_dirty); /* Dirtify all the dquots - this can block when journalling */ -static inline int mark_all_dquot_dirty(struct dquot * const *dquot) +static inline int mark_all_dquot_dirty(struct dquot __rcu * const *dquots) { int ret, err, cnt; + struct dquot *dquot; ret = err = 0; for (cnt = 0; cnt < MAXQUOTAS; cnt++) { - if (dquot[cnt]) + dquot = srcu_dereference(dquots[cnt], &dquot_srcu); + if (dquot) /* Even in case of error we have to continue */ - ret = mark_dquot_dirty(dquot[cnt]); + ret = mark_dquot_dirty(dquot); if (!err) err = ret; } @@ -875,10 +877,7 @@ void dqput(struct dquot *dquot) } /* Need to release dquot? */ -#ifdef CONFIG_QUOTA_DEBUG - /* sanity check */ - BUG_ON(!list_empty(&dquot->dq_free)); -#endif + WARN_ON_ONCE(!list_empty(&dquot->dq_free)); put_releasing_dquots(dquot); atomic_dec(&dquot->dq_count); spin_unlock(&dq_list_lock); @@ -987,9 +986,8 @@ we_slept: * smp_mb__before_atomic() in dquot_acquire(). */ smp_rmb(); -#ifdef CONFIG_QUOTA_DEBUG - BUG_ON(!dquot->dq_sb); /* Has somebody invalidated entry under us? */ -#endif + /* Has somebody invalidated entry under us? */ + WARN_ON_ONCE(hlist_unhashed(&dquot->dq_hash)); out: if (empty) do_destroy_dquot(empty); @@ -998,14 +996,14 @@ out: } EXPORT_SYMBOL(dqget); -static inline struct dquot **i_dquot(struct inode *inode) +static inline struct dquot __rcu **i_dquot(struct inode *inode) { return inode->i_sb->s_op->get_dquots(inode); } static int dqinit_needed(struct inode *inode, int type) { - struct dquot * const *dquots; + struct dquot __rcu * const *dquots; int cnt; if (IS_NOQUOTA(inode)) @@ -1095,14 +1093,16 @@ static void remove_dquot_ref(struct super_block *sb, int type) */ spin_lock(&dq_data_lock); if (!IS_NOQUOTA(inode)) { - struct dquot **dquots = i_dquot(inode); - struct dquot *dquot = dquots[type]; + struct dquot __rcu **dquots = i_dquot(inode); + struct dquot *dquot = srcu_dereference_check( + dquots[type], &dquot_srcu, + lockdep_is_held(&dq_data_lock)); #ifdef CONFIG_QUOTA_DEBUG if (unlikely(inode_get_rsv_space(inode) > 0)) reserved = 1; #endif - dquots[type] = NULL; + rcu_assign_pointer(dquots[type], NULL); if (dquot) dqput(dquot); } @@ -1455,7 +1455,8 @@ static int inode_quota_active(const struct inode *inode) static int __dquot_initialize(struct inode *inode, int type) { int cnt, init_needed = 0; - struct dquot **dquots, *got[MAXQUOTAS] = {}; + struct dquot __rcu **dquots; + struct dquot *got[MAXQUOTAS] = {}; struct super_block *sb = inode->i_sb; qsize_t rsv; int ret = 0; @@ -1530,7 +1531,7 @@ static int __dquot_initialize(struct inode *inode, int type) if (!got[cnt]) continue; if (!dquots[cnt]) { - dquots[cnt] = got[cnt]; + rcu_assign_pointer(dquots[cnt], got[cnt]); got[cnt] = NULL; /* * Make quota reservation system happy if someone @@ -1538,12 +1539,16 @@ static int __dquot_initialize(struct inode *inode, int type) */ rsv = inode_get_rsv_space(inode); if (unlikely(rsv)) { + struct dquot *dquot = srcu_dereference_check( + dquots[cnt], &dquot_srcu, + lockdep_is_held(&dq_data_lock)); + spin_lock(&inode->i_lock); /* Get reservation again under proper lock */ rsv = __inode_get_rsv_space(inode); - spin_lock(&dquots[cnt]->dq_dqb_lock); - dquots[cnt]->dq_dqb.dqb_rsvspace += rsv; - spin_unlock(&dquots[cnt]->dq_dqb_lock); + spin_lock(&dquot->dq_dqb_lock); + dquot->dq_dqb.dqb_rsvspace += rsv; + spin_unlock(&dquot->dq_dqb_lock); spin_unlock(&inode->i_lock); } } @@ -1565,7 +1570,7 @@ EXPORT_SYMBOL(dquot_initialize); bool dquot_initialize_needed(struct inode *inode) { - struct dquot **dquots; + struct dquot __rcu **dquots; int i; if (!inode_quota_active(inode)) @@ -1590,13 +1595,14 @@ EXPORT_SYMBOL(dquot_initialize_needed); static void __dquot_drop(struct inode *inode) { int cnt; - struct dquot **dquots = i_dquot(inode); + struct dquot __rcu **dquots = i_dquot(inode); struct dquot *put[MAXQUOTAS]; spin_lock(&dq_data_lock); for (cnt = 0; cnt < MAXQUOTAS; cnt++) { - put[cnt] = dquots[cnt]; - dquots[cnt] = NULL; + put[cnt] = srcu_dereference_check(dquots[cnt], &dquot_srcu, + lockdep_is_held(&dq_data_lock)); + rcu_assign_pointer(dquots[cnt], NULL); } spin_unlock(&dq_data_lock); dqput_all(put); @@ -1604,7 +1610,7 @@ static void __dquot_drop(struct inode *inode) void dquot_drop(struct inode *inode) { - struct dquot * const *dquots; + struct dquot __rcu * const *dquots; int cnt; if (IS_NOQUOTA(inode)) @@ -1677,7 +1683,8 @@ int __dquot_alloc_space(struct inode *inode, qsize_t number, int flags) int cnt, ret = 0, index; struct dquot_warn warn[MAXQUOTAS]; int reserve = flags & DQUOT_SPACE_RESERVE; - struct dquot **dquots; + struct dquot __rcu **dquots; + struct dquot *dquot; if (!inode_quota_active(inode)) { if (reserve) { @@ -1697,27 +1704,26 @@ int __dquot_alloc_space(struct inode *inode, qsize_t number, int flags) index = srcu_read_lock(&dquot_srcu); spin_lock(&inode->i_lock); for (cnt = 0; cnt < MAXQUOTAS; cnt++) { - if (!dquots[cnt]) + dquot = srcu_dereference(dquots[cnt], &dquot_srcu); + if (!dquot) continue; if (reserve) { - ret = dquot_add_space(dquots[cnt], 0, number, flags, - &warn[cnt]); + ret = dquot_add_space(dquot, 0, number, flags, &warn[cnt]); } else { - ret = dquot_add_space(dquots[cnt], number, 0, flags, - &warn[cnt]); + ret = dquot_add_space(dquot, number, 0, flags, &warn[cnt]); } if (ret) { /* Back out changes we already did */ for (cnt--; cnt >= 0; cnt--) { - if (!dquots[cnt]) + dquot = srcu_dereference(dquots[cnt], &dquot_srcu); + if (!dquot) continue; - spin_lock(&dquots[cnt]->dq_dqb_lock); + spin_lock(&dquot->dq_dqb_lock); if (reserve) - dquot_free_reserved_space(dquots[cnt], - number); + dquot_free_reserved_space(dquot, number); else - dquot_decr_space(dquots[cnt], number); - spin_unlock(&dquots[cnt]->dq_dqb_lock); + dquot_decr_space(dquot, number); + spin_unlock(&dquot->dq_dqb_lock); } spin_unlock(&inode->i_lock); goto out_flush_warn; @@ -1747,7 +1753,8 @@ int dquot_alloc_inode(struct inode *inode) { int cnt, ret = 0, index; struct dquot_warn warn[MAXQUOTAS]; - struct dquot * const *dquots; + struct dquot __rcu * const *dquots; + struct dquot *dquot; if (!inode_quota_active(inode)) return 0; @@ -1758,17 +1765,19 @@ int dquot_alloc_inode(struct inode *inode) index = srcu_read_lock(&dquot_srcu); spin_lock(&inode->i_lock); for (cnt = 0; cnt < MAXQUOTAS; cnt++) { - if (!dquots[cnt]) + dquot = srcu_dereference(dquots[cnt], &dquot_srcu); + if (!dquot) continue; - ret = dquot_add_inodes(dquots[cnt], 1, &warn[cnt]); + ret = dquot_add_inodes(dquot, 1, &warn[cnt]); if (ret) { for (cnt--; cnt >= 0; cnt--) { - if (!dquots[cnt]) + dquot = srcu_dereference(dquots[cnt], &dquot_srcu); + if (!dquot) continue; /* Back out changes we already did */ - spin_lock(&dquots[cnt]->dq_dqb_lock); - dquot_decr_inodes(dquots[cnt], 1); - spin_unlock(&dquots[cnt]->dq_dqb_lock); + spin_lock(&dquot->dq_dqb_lock); + dquot_decr_inodes(dquot, 1); + spin_unlock(&dquot->dq_dqb_lock); } goto warn_put_all; } @@ -1789,7 +1798,8 @@ EXPORT_SYMBOL(dquot_alloc_inode); */ void dquot_claim_space_nodirty(struct inode *inode, qsize_t number) { - struct dquot **dquots; + struct dquot __rcu **dquots; + struct dquot *dquot; int cnt, index; if (!inode_quota_active(inode)) { @@ -1805,9 +1815,8 @@ void dquot_claim_space_nodirty(struct inode *inode, qsize_t number) spin_lock(&inode->i_lock); /* Claim reserved quotas to allocated quotas */ for (cnt = 0; cnt < MAXQUOTAS; cnt++) { - if (dquots[cnt]) { - struct dquot *dquot = dquots[cnt]; - + dquot = srcu_dereference(dquots[cnt], &dquot_srcu); + if (dquot) { spin_lock(&dquot->dq_dqb_lock); if (WARN_ON_ONCE(dquot->dq_dqb.dqb_rsvspace < number)) number = dquot->dq_dqb.dqb_rsvspace; @@ -1831,7 +1840,8 @@ EXPORT_SYMBOL(dquot_claim_space_nodirty); */ void dquot_reclaim_space_nodirty(struct inode *inode, qsize_t number) { - struct dquot **dquots; + struct dquot __rcu **dquots; + struct dquot *dquot; int cnt, index; if (!inode_quota_active(inode)) { @@ -1847,9 +1857,8 @@ void dquot_reclaim_space_nodirty(struct inode *inode, qsize_t number) spin_lock(&inode->i_lock); /* Claim reserved quotas to allocated quotas */ for (cnt = 0; cnt < MAXQUOTAS; cnt++) { - if (dquots[cnt]) { - struct dquot *dquot = dquots[cnt]; - + dquot = srcu_dereference(dquots[cnt], &dquot_srcu); + if (dquot) { spin_lock(&dquot->dq_dqb_lock); if (WARN_ON_ONCE(dquot->dq_dqb.dqb_curspace < number)) number = dquot->dq_dqb.dqb_curspace; @@ -1875,7 +1884,8 @@ void __dquot_free_space(struct inode *inode, qsize_t number, int flags) { unsigned int cnt; struct dquot_warn warn[MAXQUOTAS]; - struct dquot **dquots; + struct dquot __rcu **dquots; + struct dquot *dquot; int reserve = flags & DQUOT_SPACE_RESERVE, index; if (!inode_quota_active(inode)) { @@ -1896,17 +1906,18 @@ void __dquot_free_space(struct inode *inode, qsize_t number, int flags) int wtype; warn[cnt].w_type = QUOTA_NL_NOWARN; - if (!dquots[cnt]) + dquot = srcu_dereference(dquots[cnt], &dquot_srcu); + if (!dquot) continue; - spin_lock(&dquots[cnt]->dq_dqb_lock); - wtype = info_bdq_free(dquots[cnt], number); + spin_lock(&dquot->dq_dqb_lock); + wtype = info_bdq_free(dquot, number); if (wtype != QUOTA_NL_NOWARN) - prepare_warning(&warn[cnt], dquots[cnt], wtype); + prepare_warning(&warn[cnt], dquot, wtype); if (reserve) - dquot_free_reserved_space(dquots[cnt], number); + dquot_free_reserved_space(dquot, number); else - dquot_decr_space(dquots[cnt], number); - spin_unlock(&dquots[cnt]->dq_dqb_lock); + dquot_decr_space(dquot, number); + spin_unlock(&dquot->dq_dqb_lock); } if (reserve) *inode_reserved_space(inode) -= number; @@ -1930,7 +1941,8 @@ void dquot_free_inode(struct inode *inode) { unsigned int cnt; struct dquot_warn warn[MAXQUOTAS]; - struct dquot * const *dquots; + struct dquot __rcu * const *dquots; + struct dquot *dquot; int index; if (!inode_quota_active(inode)) @@ -1941,16 +1953,16 @@ void dquot_free_inode(struct inode *inode) spin_lock(&inode->i_lock); for (cnt = 0; cnt < MAXQUOTAS; cnt++) { int wtype; - warn[cnt].w_type = QUOTA_NL_NOWARN; - if (!dquots[cnt]) + dquot = srcu_dereference(dquots[cnt], &dquot_srcu); + if (!dquot) continue; - spin_lock(&dquots[cnt]->dq_dqb_lock); - wtype = info_idq_free(dquots[cnt], 1); + spin_lock(&dquot->dq_dqb_lock); + wtype = info_idq_free(dquot, 1); if (wtype != QUOTA_NL_NOWARN) - prepare_warning(&warn[cnt], dquots[cnt], wtype); - dquot_decr_inodes(dquots[cnt], 1); - spin_unlock(&dquots[cnt]->dq_dqb_lock); + prepare_warning(&warn[cnt], dquot, wtype); + dquot_decr_inodes(dquot, 1); + spin_unlock(&dquot->dq_dqb_lock); } spin_unlock(&inode->i_lock); mark_all_dquot_dirty(dquots); @@ -1976,8 +1988,9 @@ int __dquot_transfer(struct inode *inode, struct dquot **transfer_to) qsize_t cur_space; qsize_t rsv_space = 0; qsize_t inode_usage = 1; + struct dquot __rcu **dquots; struct dquot *transfer_from[MAXQUOTAS] = {}; - int cnt, ret = 0; + int cnt, index, ret = 0; char is_valid[MAXQUOTAS] = {}; struct dquot_warn warn_to[MAXQUOTAS]; struct dquot_warn warn_from_inodes[MAXQUOTAS]; @@ -2008,6 +2021,7 @@ int __dquot_transfer(struct inode *inode, struct dquot **transfer_to) } cur_space = __inode_get_bytes(inode); rsv_space = __inode_get_rsv_space(inode); + dquots = i_dquot(inode); /* * Build the transfer_from list, check limits, and update usage in * the target structures. @@ -2022,7 +2036,8 @@ int __dquot_transfer(struct inode *inode, struct dquot **transfer_to) if (!sb_has_quota_active(inode->i_sb, cnt)) continue; is_valid[cnt] = 1; - transfer_from[cnt] = i_dquot(inode)[cnt]; + transfer_from[cnt] = srcu_dereference_check(dquots[cnt], + &dquot_srcu, lockdep_is_held(&dq_data_lock)); ret = dquot_add_inodes(transfer_to[cnt], inode_usage, &warn_to[cnt]); if (ret) @@ -2061,13 +2076,21 @@ int __dquot_transfer(struct inode *inode, struct dquot **transfer_to) rsv_space); spin_unlock(&transfer_from[cnt]->dq_dqb_lock); } - i_dquot(inode)[cnt] = transfer_to[cnt]; + rcu_assign_pointer(dquots[cnt], transfer_to[cnt]); } spin_unlock(&inode->i_lock); spin_unlock(&dq_data_lock); - mark_all_dquot_dirty(transfer_from); - mark_all_dquot_dirty(transfer_to); + /* + * These arrays are local and we hold dquot references so we don't need + * the srcu protection but still take dquot_srcu to avoid warning in + * mark_all_dquot_dirty(). + */ + index = srcu_read_lock(&dquot_srcu); + mark_all_dquot_dirty((struct dquot __rcu **)transfer_from); + mark_all_dquot_dirty((struct dquot __rcu **)transfer_to); + srcu_read_unlock(&dquot_srcu, index); + flush_warnings(warn_to); flush_warnings(warn_from_inodes); flush_warnings(warn_from_space); @@ -2388,7 +2411,8 @@ int dquot_load_quota_sb(struct super_block *sb, int type, int format_id, lockdep_assert_held_write(&sb->s_umount); /* Just unsuspend quotas? */ - BUG_ON(flags & DQUOT_SUSPENDED); + if (WARN_ON_ONCE(flags & DQUOT_SUSPENDED)) + return -EINVAL; if (!fmt) return -ESRCH; diff --git a/fs/quota/quota_tree.c b/fs/quota/quota_tree.c index 0f1493e0f6d0..afceef3ddfaa 100644 --- a/fs/quota/quota_tree.c +++ b/fs/quota/quota_tree.c @@ -21,6 +21,12 @@ MODULE_AUTHOR("Jan Kara"); MODULE_DESCRIPTION("Quota trie support"); MODULE_LICENSE("GPL"); +/* + * Maximum quota tree depth we support. Only to limit recursion when working + * with the tree. + */ +#define MAX_QTREE_DEPTH 6 + #define __QUOTA_QT_PARANOIA static int __get_index(struct qtree_mem_dqinfo *info, qid_t id, int depth) @@ -108,7 +114,7 @@ static int check_dquot_block_header(struct qtree_mem_dqinfo *info, /* Remove empty block from list and return it */ static int get_free_dqblk(struct qtree_mem_dqinfo *info) { - char *buf = kmalloc(info->dqi_usable_bs, GFP_NOFS); + char *buf = kmalloc(info->dqi_usable_bs, GFP_KERNEL); struct qt_disk_dqdbheader *dh = (struct qt_disk_dqdbheader *)buf; int ret, blk; @@ -160,7 +166,7 @@ static int put_free_dqblk(struct qtree_mem_dqinfo *info, char *buf, uint blk) static int remove_free_dqentry(struct qtree_mem_dqinfo *info, char *buf, uint blk) { - char *tmpbuf = kmalloc(info->dqi_usable_bs, GFP_NOFS); + char *tmpbuf = kmalloc(info->dqi_usable_bs, GFP_KERNEL); struct qt_disk_dqdbheader *dh = (struct qt_disk_dqdbheader *)buf; uint nextblk = le32_to_cpu(dh->dqdh_next_free); uint prevblk = le32_to_cpu(dh->dqdh_prev_free); @@ -207,7 +213,7 @@ out_buf: static int insert_free_dqentry(struct qtree_mem_dqinfo *info, char *buf, uint blk) { - char *tmpbuf = kmalloc(info->dqi_usable_bs, GFP_NOFS); + char *tmpbuf = kmalloc(info->dqi_usable_bs, GFP_KERNEL); struct qt_disk_dqdbheader *dh = (struct qt_disk_dqdbheader *)buf; int err; @@ -255,7 +261,7 @@ static uint find_free_dqentry(struct qtree_mem_dqinfo *info, { uint blk, i; struct qt_disk_dqdbheader *dh; - char *buf = kmalloc(info->dqi_usable_bs, GFP_NOFS); + char *buf = kmalloc(info->dqi_usable_bs, GFP_KERNEL); char *ddquot; *err = 0; @@ -327,27 +333,36 @@ out_buf: /* Insert reference to structure into the trie */ static int do_insert_tree(struct qtree_mem_dqinfo *info, struct dquot *dquot, - uint *treeblk, int depth) + uint *blks, int depth) { - char *buf = kmalloc(info->dqi_usable_bs, GFP_NOFS); + char *buf = kmalloc(info->dqi_usable_bs, GFP_KERNEL); int ret = 0, newson = 0, newact = 0; __le32 *ref; uint newblk; + int i; if (!buf) return -ENOMEM; - if (!*treeblk) { + if (!blks[depth]) { ret = get_free_dqblk(info); if (ret < 0) goto out_buf; - *treeblk = ret; + for (i = 0; i < depth; i++) + if (ret == blks[i]) { + quota_error(dquot->dq_sb, + "Free block already used in tree: block %u", + ret); + ret = -EIO; + goto out_buf; + } + blks[depth] = ret; memset(buf, 0, info->dqi_usable_bs); newact = 1; } else { - ret = read_blk(info, *treeblk, buf); + ret = read_blk(info, blks[depth], buf); if (ret < 0) { quota_error(dquot->dq_sb, "Can't read tree quota " - "block %u", *treeblk); + "block %u", blks[depth]); goto out_buf; } } @@ -357,8 +372,20 @@ static int do_insert_tree(struct qtree_mem_dqinfo *info, struct dquot *dquot, info->dqi_blocks - 1); if (ret) goto out_buf; - if (!newblk) + if (!newblk) { newson = 1; + } else { + for (i = 0; i <= depth; i++) + if (newblk == blks[i]) { + quota_error(dquot->dq_sb, + "Cycle in quota tree detected: block %u index %u", + blks[depth], + get_index(info, dquot->dq_id, depth)); + ret = -EIO; + goto out_buf; + } + } + blks[depth + 1] = newblk; if (depth == info->dqi_qtree_depth - 1) { #ifdef __QUOTA_QT_PARANOIA if (newblk) { @@ -370,16 +397,16 @@ static int do_insert_tree(struct qtree_mem_dqinfo *info, struct dquot *dquot, goto out_buf; } #endif - newblk = find_free_dqentry(info, dquot, &ret); + blks[depth + 1] = find_free_dqentry(info, dquot, &ret); } else { - ret = do_insert_tree(info, dquot, &newblk, depth+1); + ret = do_insert_tree(info, dquot, blks, depth + 1); } if (newson && ret >= 0) { ref[get_index(info, dquot->dq_id, depth)] = - cpu_to_le32(newblk); - ret = write_blk(info, *treeblk, buf); + cpu_to_le32(blks[depth + 1]); + ret = write_blk(info, blks[depth], buf); } else if (newact && ret < 0) { - put_free_dqblk(info, buf, *treeblk); + put_free_dqblk(info, buf, blks[depth]); } out_buf: kfree(buf); @@ -390,7 +417,7 @@ out_buf: static inline int dq_insert_tree(struct qtree_mem_dqinfo *info, struct dquot *dquot) { - int tmp = QT_TREEOFF; + uint blks[MAX_QTREE_DEPTH] = { QT_TREEOFF }; #ifdef __QUOTA_QT_PARANOIA if (info->dqi_blocks <= QT_TREEOFF) { @@ -398,7 +425,11 @@ static inline int dq_insert_tree(struct qtree_mem_dqinfo *info, return -EIO; } #endif - return do_insert_tree(info, dquot, &tmp, 0); + if (info->dqi_qtree_depth >= MAX_QTREE_DEPTH) { + quota_error(dquot->dq_sb, "Quota tree depth too big!"); + return -EIO; + } + return do_insert_tree(info, dquot, blks, 0); } /* @@ -410,7 +441,7 @@ int qtree_write_dquot(struct qtree_mem_dqinfo *info, struct dquot *dquot) int type = dquot->dq_id.type; struct super_block *sb = dquot->dq_sb; ssize_t ret; - char *ddquot = kmalloc(info->dqi_entry_size, GFP_NOFS); + char *ddquot = kmalloc(info->dqi_entry_size, GFP_KERNEL); if (!ddquot) return -ENOMEM; @@ -449,7 +480,7 @@ static int free_dqentry(struct qtree_mem_dqinfo *info, struct dquot *dquot, uint blk) { struct qt_disk_dqdbheader *dh; - char *buf = kmalloc(info->dqi_usable_bs, GFP_NOFS); + char *buf = kmalloc(info->dqi_usable_bs, GFP_KERNEL); int ret = 0; if (!buf) @@ -511,19 +542,20 @@ out_buf: /* Remove reference to dquot from tree */ static int remove_tree(struct qtree_mem_dqinfo *info, struct dquot *dquot, - uint *blk, int depth) + uint *blks, int depth) { - char *buf = kmalloc(info->dqi_usable_bs, GFP_NOFS); + char *buf = kmalloc(info->dqi_usable_bs, GFP_KERNEL); int ret = 0; uint newblk; __le32 *ref = (__le32 *)buf; + int i; if (!buf) return -ENOMEM; - ret = read_blk(info, *blk, buf); + ret = read_blk(info, blks[depth], buf); if (ret < 0) { quota_error(dquot->dq_sb, "Can't read quota data block %u", - *blk); + blks[depth]); goto out_buf; } newblk = le32_to_cpu(ref[get_index(info, dquot->dq_id, depth)]); @@ -532,29 +564,38 @@ static int remove_tree(struct qtree_mem_dqinfo *info, struct dquot *dquot, if (ret) goto out_buf; + for (i = 0; i <= depth; i++) + if (newblk == blks[i]) { + quota_error(dquot->dq_sb, + "Cycle in quota tree detected: block %u index %u", + blks[depth], + get_index(info, dquot->dq_id, depth)); + ret = -EIO; + goto out_buf; + } if (depth == info->dqi_qtree_depth - 1) { ret = free_dqentry(info, dquot, newblk); - newblk = 0; + blks[depth + 1] = 0; } else { - ret = remove_tree(info, dquot, &newblk, depth+1); + blks[depth + 1] = newblk; + ret = remove_tree(info, dquot, blks, depth + 1); } - if (ret >= 0 && !newblk) { - int i; + if (ret >= 0 && !blks[depth + 1]) { ref[get_index(info, dquot->dq_id, depth)] = cpu_to_le32(0); /* Block got empty? */ for (i = 0; i < (info->dqi_usable_bs >> 2) && !ref[i]; i++) ; /* Don't put the root block into the free block list */ if (i == (info->dqi_usable_bs >> 2) - && *blk != QT_TREEOFF) { - put_free_dqblk(info, buf, *blk); - *blk = 0; + && blks[depth] != QT_TREEOFF) { + put_free_dqblk(info, buf, blks[depth]); + blks[depth] = 0; } else { - ret = write_blk(info, *blk, buf); + ret = write_blk(info, blks[depth], buf); if (ret < 0) quota_error(dquot->dq_sb, "Can't write quota tree block %u", - *blk); + blks[depth]); } } out_buf: @@ -565,11 +606,15 @@ out_buf: /* Delete dquot from tree */ int qtree_delete_dquot(struct qtree_mem_dqinfo *info, struct dquot *dquot) { - uint tmp = QT_TREEOFF; + uint blks[MAX_QTREE_DEPTH] = { QT_TREEOFF }; if (!dquot->dq_off) /* Even not allocated? */ return 0; - return remove_tree(info, dquot, &tmp, 0); + if (info->dqi_qtree_depth >= MAX_QTREE_DEPTH) { + quota_error(dquot->dq_sb, "Quota tree depth too big!"); + return -EIO; + } + return remove_tree(info, dquot, blks, 0); } EXPORT_SYMBOL(qtree_delete_dquot); @@ -577,7 +622,7 @@ EXPORT_SYMBOL(qtree_delete_dquot); static loff_t find_block_dqentry(struct qtree_mem_dqinfo *info, struct dquot *dquot, uint blk) { - char *buf = kmalloc(info->dqi_usable_bs, GFP_NOFS); + char *buf = kmalloc(info->dqi_usable_bs, GFP_KERNEL); loff_t ret = 0; int i; char *ddquot; @@ -613,18 +658,20 @@ out_buf: /* Find entry for given id in the tree */ static loff_t find_tree_dqentry(struct qtree_mem_dqinfo *info, - struct dquot *dquot, uint blk, int depth) + struct dquot *dquot, uint *blks, int depth) { - char *buf = kmalloc(info->dqi_usable_bs, GFP_NOFS); + char *buf = kmalloc(info->dqi_usable_bs, GFP_KERNEL); loff_t ret = 0; __le32 *ref = (__le32 *)buf; + uint blk; + int i; if (!buf) return -ENOMEM; - ret = read_blk(info, blk, buf); + ret = read_blk(info, blks[depth], buf); if (ret < 0) { quota_error(dquot->dq_sb, "Can't read quota tree block %u", - blk); + blks[depth]); goto out_buf; } ret = 0; @@ -636,8 +683,19 @@ static loff_t find_tree_dqentry(struct qtree_mem_dqinfo *info, if (ret) goto out_buf; + /* Check for cycles in the tree */ + for (i = 0; i <= depth; i++) + if (blk == blks[i]) { + quota_error(dquot->dq_sb, + "Cycle in quota tree detected: block %u index %u", + blks[depth], + get_index(info, dquot->dq_id, depth)); + ret = -EIO; + goto out_buf; + } + blks[depth + 1] = blk; if (depth < info->dqi_qtree_depth - 1) - ret = find_tree_dqentry(info, dquot, blk, depth+1); + ret = find_tree_dqentry(info, dquot, blks, depth + 1); else ret = find_block_dqentry(info, dquot, blk); out_buf: @@ -649,7 +707,13 @@ out_buf: static inline loff_t find_dqentry(struct qtree_mem_dqinfo *info, struct dquot *dquot) { - return find_tree_dqentry(info, dquot, QT_TREEOFF, 0); + uint blks[MAX_QTREE_DEPTH] = { QT_TREEOFF }; + + if (info->dqi_qtree_depth >= MAX_QTREE_DEPTH) { + quota_error(dquot->dq_sb, "Quota tree depth too big!"); + return -EIO; + } + return find_tree_dqentry(info, dquot, blks, 0); } int qtree_read_dquot(struct qtree_mem_dqinfo *info, struct dquot *dquot) @@ -684,7 +748,7 @@ int qtree_read_dquot(struct qtree_mem_dqinfo *info, struct dquot *dquot) } dquot->dq_off = offset; } - ddquot = kmalloc(info->dqi_entry_size, GFP_NOFS); + ddquot = kmalloc(info->dqi_entry_size, GFP_KERNEL); if (!ddquot) return -ENOMEM; ret = sb->s_op->quota_read(sb, type, ddquot, info->dqi_entry_size, @@ -728,7 +792,7 @@ EXPORT_SYMBOL(qtree_release_dquot); static int find_next_id(struct qtree_mem_dqinfo *info, qid_t *id, unsigned int blk, int depth) { - char *buf = kmalloc(info->dqi_usable_bs, GFP_NOFS); + char *buf = kmalloc(info->dqi_usable_bs, GFP_KERNEL); __le32 *ref = (__le32 *)buf; ssize_t ret; unsigned int epb = info->dqi_usable_bs >> 2; diff --git a/fs/quota/quota_v1.c b/fs/quota/quota_v1.c index a0db3f195e95..3f3e8acc05db 100644 --- a/fs/quota/quota_v1.c +++ b/fs/quota/quota_v1.c @@ -160,9 +160,11 @@ static int v1_read_file_info(struct super_block *sb, int type) { struct quota_info *dqopt = sb_dqopt(sb); struct v1_disk_dqblk dqblk; + unsigned int memalloc; int ret; down_read(&dqopt->dqio_sem); + memalloc = memalloc_nofs_save(); ret = sb->s_op->quota_read(sb, type, (char *)&dqblk, sizeof(struct v1_disk_dqblk), v1_dqoff(0)); if (ret != sizeof(struct v1_disk_dqblk)) { @@ -179,6 +181,7 @@ static int v1_read_file_info(struct super_block *sb, int type) dqopt->info[type].dqi_bgrace = dqblk.dqb_btime ? dqblk.dqb_btime : MAX_DQ_TIME; out: + memalloc_nofs_restore(memalloc); up_read(&dqopt->dqio_sem); return ret; } @@ -187,9 +190,11 @@ static int v1_write_file_info(struct super_block *sb, int type) { struct quota_info *dqopt = sb_dqopt(sb); struct v1_disk_dqblk dqblk; + unsigned int memalloc; int ret; down_write(&dqopt->dqio_sem); + memalloc = memalloc_nofs_save(); ret = sb->s_op->quota_read(sb, type, (char *)&dqblk, sizeof(struct v1_disk_dqblk), v1_dqoff(0)); if (ret != sizeof(struct v1_disk_dqblk)) { @@ -209,6 +214,7 @@ static int v1_write_file_info(struct super_block *sb, int type) else if (ret >= 0) ret = -EIO; out: + memalloc_nofs_restore(memalloc); up_write(&dqopt->dqio_sem); return ret; } diff --git a/fs/quota/quota_v2.c b/fs/quota/quota_v2.c index ae99e7b88205..c48c233f3bef 100644 --- a/fs/quota/quota_v2.c +++ b/fs/quota/quota_v2.c @@ -96,9 +96,11 @@ static int v2_read_file_info(struct super_block *sb, int type) struct qtree_mem_dqinfo *qinfo; ssize_t size; unsigned int version; + unsigned int memalloc; int ret; down_read(&dqopt->dqio_sem); + memalloc = memalloc_nofs_save(); ret = v2_read_header(sb, type, &dqhead); if (ret < 0) goto out; @@ -119,7 +121,7 @@ static int v2_read_file_info(struct super_block *sb, int type) ret = -EIO; goto out; } - info->dqi_priv = kmalloc(sizeof(struct qtree_mem_dqinfo), GFP_NOFS); + info->dqi_priv = kmalloc(sizeof(struct qtree_mem_dqinfo), GFP_KERNEL); if (!info->dqi_priv) { ret = -ENOMEM; goto out; @@ -166,14 +168,17 @@ static int v2_read_file_info(struct super_block *sb, int type) i_size_read(sb_dqopt(sb)->files[type])); goto out_free; } - if (qinfo->dqi_free_blk >= qinfo->dqi_blocks) { - quota_error(sb, "Free block number too big (%u >= %u).", - qinfo->dqi_free_blk, qinfo->dqi_blocks); + if (qinfo->dqi_free_blk && (qinfo->dqi_free_blk <= QT_TREEOFF || + qinfo->dqi_free_blk >= qinfo->dqi_blocks)) { + quota_error(sb, "Free block number %u out of range (%u, %u).", + qinfo->dqi_free_blk, QT_TREEOFF, qinfo->dqi_blocks); goto out_free; } - if (qinfo->dqi_free_entry >= qinfo->dqi_blocks) { - quota_error(sb, "Block with free entry too big (%u >= %u).", - qinfo->dqi_free_entry, qinfo->dqi_blocks); + if (qinfo->dqi_free_entry && (qinfo->dqi_free_entry <= QT_TREEOFF || + qinfo->dqi_free_entry >= qinfo->dqi_blocks)) { + quota_error(sb, "Block with free entry %u out of range (%u, %u).", + qinfo->dqi_free_entry, QT_TREEOFF, + qinfo->dqi_blocks); goto out_free; } ret = 0; @@ -183,6 +188,7 @@ out_free: info->dqi_priv = NULL; } out: + memalloc_nofs_restore(memalloc); up_read(&dqopt->dqio_sem); return ret; } @@ -195,8 +201,10 @@ static int v2_write_file_info(struct super_block *sb, int type) struct mem_dqinfo *info = &dqopt->info[type]; struct qtree_mem_dqinfo *qinfo = info->dqi_priv; ssize_t size; + unsigned int memalloc; down_write(&dqopt->dqio_sem); + memalloc = memalloc_nofs_save(); spin_lock(&dq_data_lock); info->dqi_flags &= ~DQF_INFO_DIRTY; dinfo.dqi_bgrace = cpu_to_le32(info->dqi_bgrace); @@ -209,6 +217,7 @@ static int v2_write_file_info(struct super_block *sb, int type) dinfo.dqi_free_entry = cpu_to_le32(qinfo->dqi_free_entry); size = sb->s_op->quota_write(sb, type, (char *)&dinfo, sizeof(struct v2_disk_dqinfo), V2_DQINFOOFF); + memalloc_nofs_restore(memalloc); up_write(&dqopt->dqio_sem); if (size != sizeof(struct v2_disk_dqinfo)) { quota_error(sb, "Can't write info structure"); @@ -328,11 +337,14 @@ static int v2_read_dquot(struct dquot *dquot) { struct quota_info *dqopt = sb_dqopt(dquot->dq_sb); int ret; + unsigned int memalloc; down_read(&dqopt->dqio_sem); + memalloc = memalloc_nofs_save(); ret = qtree_read_dquot( sb_dqinfo(dquot->dq_sb, dquot->dq_id.type)->dqi_priv, dquot); + memalloc_nofs_restore(memalloc); up_read(&dqopt->dqio_sem); return ret; } @@ -342,6 +354,7 @@ static int v2_write_dquot(struct dquot *dquot) struct quota_info *dqopt = sb_dqopt(dquot->dq_sb); int ret; bool alloc = false; + unsigned int memalloc; /* * If space for dquot is already allocated, we don't need any @@ -355,9 +368,11 @@ static int v2_write_dquot(struct dquot *dquot) } else { down_read(&dqopt->dqio_sem); } + memalloc = memalloc_nofs_save(); ret = qtree_write_dquot( sb_dqinfo(dquot->dq_sb, dquot->dq_id.type)->dqi_priv, dquot); + memalloc_nofs_restore(memalloc); if (alloc) up_write(&dqopt->dqio_sem); else @@ -368,10 +383,13 @@ static int v2_write_dquot(struct dquot *dquot) static int v2_release_dquot(struct dquot *dquot) { struct quota_info *dqopt = sb_dqopt(dquot->dq_sb); + unsigned int memalloc; int ret; down_write(&dqopt->dqio_sem); + memalloc = memalloc_nofs_save(); ret = qtree_release_dquot(sb_dqinfo(dquot->dq_sb, dquot->dq_id.type)->dqi_priv, dquot); + memalloc_nofs_restore(memalloc); up_write(&dqopt->dqio_sem); return ret; @@ -386,10 +404,13 @@ static int v2_free_file_info(struct super_block *sb, int type) static int v2_get_next_id(struct super_block *sb, struct kqid *qid) { struct quota_info *dqopt = sb_dqopt(sb); + unsigned int memalloc; int ret; down_read(&dqopt->dqio_sem); + memalloc = memalloc_nofs_save(); ret = qtree_get_next_id(sb_dqinfo(sb, qid->type)->dqi_priv, qid); + memalloc_nofs_restore(memalloc); up_read(&dqopt->dqio_sem); return ret; } diff --git a/fs/reiserfs/reiserfs.h b/fs/reiserfs/reiserfs.h index 0554903f42a9..f0e1f29f20ee 100644 --- a/fs/reiserfs/reiserfs.h +++ b/fs/reiserfs/reiserfs.h @@ -97,7 +97,7 @@ struct reiserfs_inode_info { struct rw_semaphore i_xattr_sem; #endif #ifdef CONFIG_QUOTA - struct dquot *i_dquot[MAXQUOTAS]; + struct dquot __rcu *i_dquot[MAXQUOTAS]; #endif struct inode vfs_inode; diff --git a/fs/reiserfs/super.c b/fs/reiserfs/super.c index 2cc469d481a2..ab76468da02d 100644 --- a/fs/reiserfs/super.c +++ b/fs/reiserfs/super.c @@ -801,7 +801,7 @@ static ssize_t reiserfs_quota_write(struct super_block *, int, const char *, static ssize_t reiserfs_quota_read(struct super_block *, int, char *, size_t, loff_t); -static struct dquot **reiserfs_get_dquots(struct inode *inode) +static struct dquot __rcu **reiserfs_get_dquots(struct inode *inode) { return REISERFS_I(inode)->i_dquot; } diff --git a/fs/smb/server/glob.h b/fs/smb/server/glob.h index 5b8f3e0ebdb3..d528b20b37a8 100644 --- a/fs/smb/server/glob.h +++ b/fs/smb/server/glob.h @@ -12,8 +12,6 @@ #include "unicode.h" #include "vfs_cache.h" -#define KSMBD_VERSION "3.4.2" - extern int ksmbd_debug_types; #define KSMBD_DEBUG_SMB BIT(0) diff --git a/fs/smb/server/ksmbd_netlink.h b/fs/smb/server/ksmbd_netlink.h index 0ebf91ffa236..8ca8a45c4c62 100644 --- a/fs/smb/server/ksmbd_netlink.h +++ b/fs/smb/server/ksmbd_netlink.h @@ -75,6 +75,7 @@ struct ksmbd_heartbeat { #define KSMBD_GLOBAL_FLAG_SMB2_ENCRYPTION BIT(1) #define KSMBD_GLOBAL_FLAG_SMB3_MULTICHANNEL BIT(2) #define KSMBD_GLOBAL_FLAG_SMB2_ENCRYPTION_OFF BIT(3) +#define KSMBD_GLOBAL_FLAG_DURABLE_HANDLE BIT(4) /* * IPC request for ksmbd server startup diff --git a/fs/smb/server/mgmt/user_session.c b/fs/smb/server/mgmt/user_session.c index 15f68ee05089..aec0a7a12405 100644 --- a/fs/smb/server/mgmt/user_session.c +++ b/fs/smb/server/mgmt/user_session.c @@ -156,7 +156,7 @@ void ksmbd_session_destroy(struct ksmbd_session *sess) kfree(sess); } -static struct ksmbd_session *__session_lookup(unsigned long long id) +struct ksmbd_session *__session_lookup(unsigned long long id) { struct ksmbd_session *sess; @@ -305,6 +305,32 @@ struct preauth_session *ksmbd_preauth_session_alloc(struct ksmbd_conn *conn, return sess; } +void destroy_previous_session(struct ksmbd_conn *conn, + struct ksmbd_user *user, u64 id) +{ + struct ksmbd_session *prev_sess; + struct ksmbd_user *prev_user; + + down_write(&sessions_table_lock); + down_write(&conn->session_lock); + prev_sess = __session_lookup(id); + if (!prev_sess || prev_sess->state == SMB2_SESSION_EXPIRED) + goto out; + + prev_user = prev_sess->user; + if (!prev_user || + strcmp(user->name, prev_user->name) || + user->passkey_sz != prev_user->passkey_sz || + memcmp(user->passkey, prev_user->passkey, user->passkey_sz)) + goto out; + + ksmbd_destroy_file_table(&prev_sess->file_table); + prev_sess->state = SMB2_SESSION_EXPIRED; +out: + up_write(&conn->session_lock); + up_write(&sessions_table_lock); +} + static bool ksmbd_preauth_session_id_match(struct preauth_session *sess, unsigned long long id) { diff --git a/fs/smb/server/mgmt/user_session.h b/fs/smb/server/mgmt/user_session.h index 63cb08fffde8..dc9fded2cd43 100644 --- a/fs/smb/server/mgmt/user_session.h +++ b/fs/smb/server/mgmt/user_session.h @@ -88,8 +88,11 @@ struct ksmbd_session *ksmbd_session_lookup(struct ksmbd_conn *conn, int ksmbd_session_register(struct ksmbd_conn *conn, struct ksmbd_session *sess); void ksmbd_sessions_deregister(struct ksmbd_conn *conn); +struct ksmbd_session *__session_lookup(unsigned long long id); struct ksmbd_session *ksmbd_session_lookup_all(struct ksmbd_conn *conn, unsigned long long id); +void destroy_previous_session(struct ksmbd_conn *conn, + struct ksmbd_user *user, u64 id); struct preauth_session *ksmbd_preauth_session_alloc(struct ksmbd_conn *conn, u64 sess_id); struct preauth_session *ksmbd_preauth_session_lookup(struct ksmbd_conn *conn, diff --git a/fs/smb/server/oplock.c b/fs/smb/server/oplock.c index 53dfaac425c6..4978edfb15f9 100644 --- a/fs/smb/server/oplock.c +++ b/fs/smb/server/oplock.c @@ -159,7 +159,8 @@ static struct oplock_info *opinfo_get_list(struct ksmbd_inode *ci) opinfo = list_first_or_null_rcu(&ci->m_op_list, struct oplock_info, op_entry); if (opinfo) { - if (!atomic_inc_not_zero(&opinfo->refcount)) + if (opinfo->conn == NULL || + !atomic_inc_not_zero(&opinfo->refcount)) opinfo = NULL; else { atomic_inc(&opinfo->conn->r_count); @@ -527,7 +528,7 @@ static struct oplock_info *same_client_has_lease(struct ksmbd_inode *ci, */ read_lock(&ci->m_lock); list_for_each_entry(opinfo, &ci->m_op_list, op_entry) { - if (!opinfo->is_lease) + if (!opinfo->is_lease || !opinfo->conn) continue; read_unlock(&ci->m_lock); lease = opinfo->o_lease; @@ -641,7 +642,7 @@ static void __smb2_oplock_break_noti(struct work_struct *wk) struct smb2_hdr *rsp_hdr; struct ksmbd_file *fp; - fp = ksmbd_lookup_durable_fd(br_info->fid); + fp = ksmbd_lookup_global_fd(br_info->fid); if (!fp) goto out; @@ -1106,7 +1107,7 @@ void smb_send_parent_lease_break_noti(struct ksmbd_file *fp, read_lock(&p_ci->m_lock); list_for_each_entry(opinfo, &p_ci->m_op_list, op_entry) { - if (!opinfo->is_lease) + if (opinfo->conn == NULL || !opinfo->is_lease) continue; if (opinfo->o_lease->state != SMB2_OPLOCK_LEVEL_NONE && @@ -1142,7 +1143,7 @@ void smb_lazy_parent_lease_break_close(struct ksmbd_file *fp) opinfo = rcu_dereference(fp->f_opinfo); rcu_read_unlock(); - if (!opinfo->is_lease || opinfo->o_lease->version != 2) + if (!opinfo || !opinfo->is_lease || opinfo->o_lease->version != 2) return; p_ci = ksmbd_inode_lookup_lock(fp->filp->f_path.dentry->d_parent); @@ -1151,7 +1152,7 @@ void smb_lazy_parent_lease_break_close(struct ksmbd_file *fp) read_lock(&p_ci->m_lock); list_for_each_entry(opinfo, &p_ci->m_op_list, op_entry) { - if (!opinfo->is_lease) + if (opinfo->conn == NULL || !opinfo->is_lease) continue; if (opinfo->o_lease->state != SMB2_OPLOCK_LEVEL_NONE) { @@ -1361,6 +1362,9 @@ void smb_break_all_levII_oplock(struct ksmbd_work *work, struct ksmbd_file *fp, rcu_read_lock(); list_for_each_entry_rcu(brk_op, &ci->m_op_list, op_entry) { + if (brk_op->conn == NULL) + continue; + if (!atomic_inc_not_zero(&brk_op->refcount)) continue; @@ -1496,11 +1500,10 @@ void create_lease_buf(u8 *rbuf, struct lease *lease) /** * parse_lease_state() - parse lease context containted in file open request * @open_req: buffer containing smb2 file open(create) request - * @is_dir: whether leasing file is directory * * Return: oplock state, -ENOENT if create lease context not found */ -struct lease_ctx_info *parse_lease_state(void *open_req, bool is_dir) +struct lease_ctx_info *parse_lease_state(void *open_req) { struct create_context *cc; struct smb2_create_req *req = (struct smb2_create_req *)open_req; @@ -1518,12 +1521,7 @@ struct lease_ctx_info *parse_lease_state(void *open_req, bool is_dir) struct create_lease_v2 *lc = (struct create_lease_v2 *)cc; memcpy(lreq->lease_key, lc->lcontext.LeaseKey, SMB2_LEASE_KEY_SIZE); - if (is_dir) { - lreq->req_state = lc->lcontext.LeaseState & - ~SMB2_LEASE_WRITE_CACHING_LE; - lreq->is_dir = true; - } else - lreq->req_state = lc->lcontext.LeaseState; + lreq->req_state = lc->lcontext.LeaseState; lreq->flags = lc->lcontext.LeaseFlags; lreq->epoch = lc->lcontext.Epoch; lreq->duration = lc->lcontext.LeaseDuration; @@ -1646,6 +1644,8 @@ void create_durable_v2_rsp_buf(char *cc, struct ksmbd_file *fp) buf->Name[3] = 'Q'; buf->Timeout = cpu_to_le32(fp->durable_timeout); + if (fp->is_persistent) + buf->Flags = cpu_to_le32(SMB2_DHANDLE_FLAG_PERSISTENT); } /** @@ -1813,3 +1813,71 @@ out: read_unlock(&lease_list_lock); return ret_op; } + +int smb2_check_durable_oplock(struct ksmbd_conn *conn, + struct ksmbd_share_config *share, + struct ksmbd_file *fp, + struct lease_ctx_info *lctx, + char *name) +{ + struct oplock_info *opinfo = opinfo_get(fp); + int ret = 0; + + if (!opinfo) + return 0; + + if (opinfo->is_lease == false) { + if (lctx) { + pr_err("create context include lease\n"); + ret = -EBADF; + goto out; + } + + if (opinfo->level != SMB2_OPLOCK_LEVEL_BATCH) { + pr_err("oplock level is not equal to SMB2_OPLOCK_LEVEL_BATCH\n"); + ret = -EBADF; + } + + goto out; + } + + if (memcmp(conn->ClientGUID, fp->client_guid, + SMB2_CLIENT_GUID_SIZE)) { + ksmbd_debug(SMB, "Client guid of fp is not equal to the one of connection\n"); + ret = -EBADF; + goto out; + } + + if (!lctx) { + ksmbd_debug(SMB, "create context does not include lease\n"); + ret = -EBADF; + goto out; + } + + if (memcmp(opinfo->o_lease->lease_key, lctx->lease_key, + SMB2_LEASE_KEY_SIZE)) { + ksmbd_debug(SMB, + "lease key of fp does not match lease key in create context\n"); + ret = -EBADF; + goto out; + } + + if (!(opinfo->o_lease->state & SMB2_LEASE_HANDLE_CACHING_LE)) { + ksmbd_debug(SMB, "lease state does not contain SMB2_LEASE_HANDLE_CACHING\n"); + ret = -EBADF; + goto out; + } + + if (opinfo->o_lease->version != lctx->version) { + ksmbd_debug(SMB, + "lease version of fp does not match the one in create context\n"); + ret = -EBADF; + goto out; + } + + if (!ksmbd_inode_pending_delete(fp)) + ret = ksmbd_validate_name_reconnect(share, fp, name); +out: + opinfo_put(opinfo); + return ret; +} diff --git a/fs/smb/server/oplock.h b/fs/smb/server/oplock.h index 5b93ea9196c0..e9da63f25b20 100644 --- a/fs/smb/server/oplock.h +++ b/fs/smb/server/oplock.h @@ -111,7 +111,7 @@ void opinfo_put(struct oplock_info *opinfo); /* Lease related functions */ void create_lease_buf(u8 *rbuf, struct lease *lease); -struct lease_ctx_info *parse_lease_state(void *open_req, bool is_dir); +struct lease_ctx_info *parse_lease_state(void *open_req); __u8 smb2_map_lease_to_oplock(__le32 lease_state); int lease_read_to_write(struct oplock_info *opinfo); @@ -130,4 +130,9 @@ void destroy_lease_table(struct ksmbd_conn *conn); void smb_send_parent_lease_break_noti(struct ksmbd_file *fp, struct lease_ctx_info *lctx); void smb_lazy_parent_lease_break_close(struct ksmbd_file *fp); +int smb2_check_durable_oplock(struct ksmbd_conn *conn, + struct ksmbd_share_config *share, + struct ksmbd_file *fp, + struct lease_ctx_info *lctx, + char *name); #endif /* __KSMBD_OPLOCK_H */ diff --git a/fs/smb/server/server.c b/fs/smb/server/server.c index 3079e607c5fe..c0788188aa82 100644 --- a/fs/smb/server/server.c +++ b/fs/smb/server/server.c @@ -625,7 +625,6 @@ static void __exit ksmbd_server_exit(void) } MODULE_AUTHOR("Namjae Jeon <linkinjeon@kernel.org>"); -MODULE_VERSION(KSMBD_VERSION); MODULE_DESCRIPTION("Linux kernel CIFS/SMB SERVER"); MODULE_LICENSE("GPL"); MODULE_SOFTDEP("pre: ecb"); diff --git a/fs/smb/server/smb2misc.c b/fs/smb/server/smb2misc.c index 03dded29a980..727cb49926ee 100644 --- a/fs/smb/server/smb2misc.c +++ b/fs/smb/server/smb2misc.c @@ -101,13 +101,17 @@ static int smb2_get_data_area_len(unsigned int *off, unsigned int *len, *len = le16_to_cpu(((struct smb2_sess_setup_req *)hdr)->SecurityBufferLength); break; case SMB2_TREE_CONNECT: - *off = le16_to_cpu(((struct smb2_tree_connect_req *)hdr)->PathOffset); + *off = max_t(unsigned short int, + le16_to_cpu(((struct smb2_tree_connect_req *)hdr)->PathOffset), + offsetof(struct smb2_tree_connect_req, Buffer)); *len = le16_to_cpu(((struct smb2_tree_connect_req *)hdr)->PathLength); break; case SMB2_CREATE: { unsigned short int name_off = - le16_to_cpu(((struct smb2_create_req *)hdr)->NameOffset); + max_t(unsigned short int, + le16_to_cpu(((struct smb2_create_req *)hdr)->NameOffset), + offsetof(struct smb2_create_req, Buffer)); unsigned short int name_len = le16_to_cpu(((struct smb2_create_req *)hdr)->NameLength); @@ -128,11 +132,15 @@ static int smb2_get_data_area_len(unsigned int *off, unsigned int *len, break; } case SMB2_QUERY_INFO: - *off = le16_to_cpu(((struct smb2_query_info_req *)hdr)->InputBufferOffset); + *off = max_t(unsigned int, + le16_to_cpu(((struct smb2_query_info_req *)hdr)->InputBufferOffset), + offsetof(struct smb2_query_info_req, Buffer)); *len = le32_to_cpu(((struct smb2_query_info_req *)hdr)->InputBufferLength); break; case SMB2_SET_INFO: - *off = le16_to_cpu(((struct smb2_set_info_req *)hdr)->BufferOffset); + *off = max_t(unsigned int, + le16_to_cpu(((struct smb2_set_info_req *)hdr)->BufferOffset), + offsetof(struct smb2_set_info_req, Buffer)); *len = le32_to_cpu(((struct smb2_set_info_req *)hdr)->BufferLength); break; case SMB2_READ: @@ -142,7 +150,7 @@ static int smb2_get_data_area_len(unsigned int *off, unsigned int *len, case SMB2_WRITE: if (((struct smb2_write_req *)hdr)->DataOffset || ((struct smb2_write_req *)hdr)->Length) { - *off = max_t(unsigned int, + *off = max_t(unsigned short int, le16_to_cpu(((struct smb2_write_req *)hdr)->DataOffset), offsetof(struct smb2_write_req, Buffer)); *len = le32_to_cpu(((struct smb2_write_req *)hdr)->Length); @@ -153,7 +161,9 @@ static int smb2_get_data_area_len(unsigned int *off, unsigned int *len, *len = le16_to_cpu(((struct smb2_write_req *)hdr)->WriteChannelInfoLength); break; case SMB2_QUERY_DIRECTORY: - *off = le16_to_cpu(((struct smb2_query_directory_req *)hdr)->FileNameOffset); + *off = max_t(unsigned short int, + le16_to_cpu(((struct smb2_query_directory_req *)hdr)->FileNameOffset), + offsetof(struct smb2_query_directory_req, Buffer)); *len = le16_to_cpu(((struct smb2_query_directory_req *)hdr)->FileNameLength); break; case SMB2_LOCK: @@ -168,7 +178,9 @@ static int smb2_get_data_area_len(unsigned int *off, unsigned int *len, break; } case SMB2_IOCTL: - *off = le32_to_cpu(((struct smb2_ioctl_req *)hdr)->InputOffset); + *off = max_t(unsigned int, + le32_to_cpu(((struct smb2_ioctl_req *)hdr)->InputOffset), + offsetof(struct smb2_ioctl_req, Buffer)); *len = le32_to_cpu(((struct smb2_ioctl_req *)hdr)->InputCount); break; default: diff --git a/fs/smb/server/smb2ops.c b/fs/smb/server/smb2ops.c index 27a9dce3e03a..a45f7dca482e 100644 --- a/fs/smb/server/smb2ops.c +++ b/fs/smb/server/smb2ops.c @@ -256,6 +256,9 @@ void init_smb3_02_server(struct ksmbd_conn *conn) if (server_conf.flags & KSMBD_GLOBAL_FLAG_SMB3_MULTICHANNEL) conn->vals->capabilities |= SMB2_GLOBAL_CAP_MULTI_CHANNEL; + + if (server_conf.flags & KSMBD_GLOBAL_FLAG_DURABLE_HANDLE) + conn->vals->capabilities |= SMB2_GLOBAL_CAP_PERSISTENT_HANDLES; } /** @@ -283,6 +286,9 @@ int init_smb3_11_server(struct ksmbd_conn *conn) if (server_conf.flags & KSMBD_GLOBAL_FLAG_SMB3_MULTICHANNEL) conn->vals->capabilities |= SMB2_GLOBAL_CAP_MULTI_CHANNEL; + if (server_conf.flags & KSMBD_GLOBAL_FLAG_DURABLE_HANDLE) + conn->vals->capabilities |= SMB2_GLOBAL_CAP_PERSISTENT_HANDLES; + INIT_LIST_HEAD(&conn->preauth_sess_table); return 0; } diff --git a/fs/smb/server/smb2pdu.c b/fs/smb/server/smb2pdu.c index 089527a8b4ff..d478fa0c57ab 100644 --- a/fs/smb/server/smb2pdu.c +++ b/fs/smb/server/smb2pdu.c @@ -607,30 +607,6 @@ int smb2_check_user_session(struct ksmbd_work *work) return -ENOENT; } -static void destroy_previous_session(struct ksmbd_conn *conn, - struct ksmbd_user *user, u64 id) -{ - struct ksmbd_session *prev_sess = ksmbd_session_lookup_slowpath(id); - struct ksmbd_user *prev_user; - struct channel *chann; - long index; - - if (!prev_sess) - return; - - prev_user = prev_sess->user; - - if (!prev_user || - strcmp(user->name, prev_user->name) || - user->passkey_sz != prev_user->passkey_sz || - memcmp(user->passkey, prev_user->passkey, user->passkey_sz)) - return; - - prev_sess->state = SMB2_SESSION_EXPIRED; - xa_for_each(&prev_sess->ksmbd_chann_list, index, chann) - ksmbd_conn_set_exiting(chann->conn); -} - /** * smb2_get_name() - get filename string from on the wire smb format * @src: source buffer @@ -1951,7 +1927,7 @@ int smb2_tree_connect(struct ksmbd_work *work) WORK_BUFFERS(work, req, rsp); - treename = smb_strndup_from_utf16(req->Buffer, + treename = smb_strndup_from_utf16((char *)req + le16_to_cpu(req->PathOffset), le16_to_cpu(req->PathLength), true, conn->local_nls); if (IS_ERR(treename)) { @@ -2642,6 +2618,165 @@ static void ksmbd_acls_fattr(struct smb_fattr *fattr, } } +enum { + DURABLE_RECONN_V2 = 1, + DURABLE_RECONN, + DURABLE_REQ_V2, + DURABLE_REQ, +}; + +struct durable_info { + struct ksmbd_file *fp; + unsigned short int type; + bool persistent; + bool reconnected; + unsigned int timeout; + char *CreateGuid; +}; + +static int parse_durable_handle_context(struct ksmbd_work *work, + struct smb2_create_req *req, + struct lease_ctx_info *lc, + struct durable_info *dh_info) +{ + struct ksmbd_conn *conn = work->conn; + struct create_context *context; + int dh_idx, err = 0; + u64 persistent_id = 0; + int req_op_level; + static const char * const durable_arr[] = {"DH2C", "DHnC", "DH2Q", "DHnQ"}; + + req_op_level = req->RequestedOplockLevel; + for (dh_idx = DURABLE_RECONN_V2; dh_idx <= ARRAY_SIZE(durable_arr); + dh_idx++) { + context = smb2_find_context_vals(req, durable_arr[dh_idx - 1], 4); + if (IS_ERR(context)) { + err = PTR_ERR(context); + goto out; + } + if (!context) + continue; + + switch (dh_idx) { + case DURABLE_RECONN_V2: + { + struct create_durable_reconn_v2_req *recon_v2; + + if (dh_info->type == DURABLE_RECONN || + dh_info->type == DURABLE_REQ_V2) { + err = -EINVAL; + goto out; + } + + recon_v2 = (struct create_durable_reconn_v2_req *)context; + persistent_id = recon_v2->Fid.PersistentFileId; + dh_info->fp = ksmbd_lookup_durable_fd(persistent_id); + if (!dh_info->fp) { + ksmbd_debug(SMB, "Failed to get durable handle state\n"); + err = -EBADF; + goto out; + } + + if (memcmp(dh_info->fp->create_guid, recon_v2->CreateGuid, + SMB2_CREATE_GUID_SIZE)) { + err = -EBADF; + ksmbd_put_durable_fd(dh_info->fp); + goto out; + } + + dh_info->type = dh_idx; + dh_info->reconnected = true; + ksmbd_debug(SMB, + "reconnect v2 Persistent-id from reconnect = %llu\n", + persistent_id); + break; + } + case DURABLE_RECONN: + { + struct create_durable_reconn_req *recon; + + if (dh_info->type == DURABLE_RECONN_V2 || + dh_info->type == DURABLE_REQ_V2) { + err = -EINVAL; + goto out; + } + + recon = (struct create_durable_reconn_req *)context; + persistent_id = recon->Data.Fid.PersistentFileId; + dh_info->fp = ksmbd_lookup_durable_fd(persistent_id); + if (!dh_info->fp) { + ksmbd_debug(SMB, "Failed to get durable handle state\n"); + err = -EBADF; + goto out; + } + + dh_info->type = dh_idx; + dh_info->reconnected = true; + ksmbd_debug(SMB, "reconnect Persistent-id from reconnect = %llu\n", + persistent_id); + break; + } + case DURABLE_REQ_V2: + { + struct create_durable_req_v2 *durable_v2_blob; + + if (dh_info->type == DURABLE_RECONN || + dh_info->type == DURABLE_RECONN_V2) { + err = -EINVAL; + goto out; + } + + durable_v2_blob = + (struct create_durable_req_v2 *)context; + ksmbd_debug(SMB, "Request for durable v2 open\n"); + dh_info->fp = ksmbd_lookup_fd_cguid(durable_v2_blob->CreateGuid); + if (dh_info->fp) { + if (!memcmp(conn->ClientGUID, dh_info->fp->client_guid, + SMB2_CLIENT_GUID_SIZE)) { + if (!(req->hdr.Flags & SMB2_FLAGS_REPLAY_OPERATION)) { + err = -ENOEXEC; + goto out; + } + + dh_info->fp->conn = conn; + dh_info->reconnected = true; + goto out; + } + } + + if (((lc && (lc->req_state & SMB2_LEASE_HANDLE_CACHING_LE)) || + req_op_level == SMB2_OPLOCK_LEVEL_BATCH)) { + dh_info->CreateGuid = + durable_v2_blob->CreateGuid; + dh_info->persistent = + le32_to_cpu(durable_v2_blob->Flags); + dh_info->timeout = + le32_to_cpu(durable_v2_blob->Timeout); + dh_info->type = dh_idx; + } + break; + } + case DURABLE_REQ: + if (dh_info->type == DURABLE_RECONN) + goto out; + if (dh_info->type == DURABLE_RECONN_V2 || + dh_info->type == DURABLE_REQ_V2) { + err = -EINVAL; + goto out; + } + + if (((lc && (lc->req_state & SMB2_LEASE_HANDLE_CACHING_LE)) || + req_op_level == SMB2_OPLOCK_LEVEL_BATCH)) { + ksmbd_debug(SMB, "Request for durable open\n"); + dh_info->type = dh_idx; + } + } + } + +out: + return err; +} + /** * smb2_open() - handler for smb file open request * @work: smb work containing request buffer @@ -2665,6 +2800,7 @@ int smb2_open(struct ksmbd_work *work) struct lease_ctx_info *lc = NULL; struct create_ea_buf_req *ea_buf = NULL; struct oplock_info *opinfo; + struct durable_info dh_info = {0}; __le32 *next_ptr = NULL; int req_op_level = 0, open_flags = 0, may_flags = 0, file_info = 0; int rc = 0; @@ -2704,7 +2840,7 @@ int smb2_open(struct ksmbd_work *work) goto err_out2; } - name = smb2_get_name(req->Buffer, + name = smb2_get_name((char *)req + le16_to_cpu(req->NameOffset), le16_to_cpu(req->NameLength), work->conn->local_nls); if (IS_ERR(name)) { @@ -2745,6 +2881,49 @@ int smb2_open(struct ksmbd_work *work) } } + req_op_level = req->RequestedOplockLevel; + + if (server_conf.flags & KSMBD_GLOBAL_FLAG_DURABLE_HANDLE && + req->CreateContextsOffset) { + lc = parse_lease_state(req); + rc = parse_durable_handle_context(work, req, lc, &dh_info); + if (rc) { + ksmbd_debug(SMB, "error parsing durable handle context\n"); + goto err_out2; + } + + if (dh_info.reconnected == true) { + rc = smb2_check_durable_oplock(conn, share, dh_info.fp, lc, name); + if (rc) { + ksmbd_put_durable_fd(dh_info.fp); + goto err_out2; + } + + rc = ksmbd_reopen_durable_fd(work, dh_info.fp); + if (rc) { + ksmbd_put_durable_fd(dh_info.fp); + goto err_out2; + } + + if (ksmbd_override_fsids(work)) { + rc = -ENOMEM; + ksmbd_put_durable_fd(dh_info.fp); + goto err_out2; + } + + fp = dh_info.fp; + file_info = FILE_OPENED; + + rc = ksmbd_vfs_getattr(&fp->filp->f_path, &stat); + if (rc) + goto err_out2; + + ksmbd_put_durable_fd(fp); + goto reconnected_fp; + } + } else if (req_op_level == SMB2_OPLOCK_LEVEL_LEASE) + lc = parse_lease_state(req); + if (le32_to_cpu(req->ImpersonationLevel) > le32_to_cpu(IL_DELEGATE)) { pr_err("Invalid impersonationlevel : 0x%x\n", le32_to_cpu(req->ImpersonationLevel)); @@ -3207,10 +3386,6 @@ int smb2_open(struct ksmbd_work *work) need_truncate = 1; } - req_op_level = req->RequestedOplockLevel; - if (req_op_level == SMB2_OPLOCK_LEVEL_LEASE) - lc = parse_lease_state(req, S_ISDIR(file_inode(filp)->i_mode)); - share_ret = ksmbd_smb_check_shared_mode(fp->filp, fp); if (!test_share_config_flag(work->tcon->share_conf, KSMBD_SHARE_FLAG_OPLOCKS) || (req_op_level == SMB2_OPLOCK_LEVEL_LEASE && @@ -3221,6 +3396,11 @@ int smb2_open(struct ksmbd_work *work) } } else { if (req_op_level == SMB2_OPLOCK_LEVEL_LEASE) { + if (S_ISDIR(file_inode(filp)->i_mode)) { + lc->req_state &= ~SMB2_LEASE_WRITE_CACHING_LE; + lc->is_dir = true; + } + /* * Compare parent lease using parent key. If there is no * a lease that has same parent key, Send lease break @@ -3317,6 +3497,24 @@ int smb2_open(struct ksmbd_work *work) memcpy(fp->client_guid, conn->ClientGUID, SMB2_CLIENT_GUID_SIZE); + if (dh_info.type == DURABLE_REQ_V2 || dh_info.type == DURABLE_REQ) { + if (dh_info.type == DURABLE_REQ_V2 && dh_info.persistent) + fp->is_persistent = true; + else + fp->is_durable = true; + + if (dh_info.type == DURABLE_REQ_V2) { + memcpy(fp->create_guid, dh_info.CreateGuid, + SMB2_CREATE_GUID_SIZE); + if (dh_info.timeout) + fp->durable_timeout = min(dh_info.timeout, + 300000); + else + fp->durable_timeout = 60; + } + } + +reconnected_fp: rsp->StructureSize = cpu_to_le16(89); rcu_read_lock(); opinfo = rcu_dereference(fp->f_opinfo); @@ -3403,6 +3601,33 @@ int smb2_open(struct ksmbd_work *work) next_off = conn->vals->create_disk_id_size; } + if (dh_info.type == DURABLE_REQ || dh_info.type == DURABLE_REQ_V2) { + struct create_context *durable_ccontext; + + durable_ccontext = (struct create_context *)(rsp->Buffer + + le32_to_cpu(rsp->CreateContextsLength)); + contxt_cnt++; + if (dh_info.type == DURABLE_REQ) { + create_durable_rsp_buf(rsp->Buffer + + le32_to_cpu(rsp->CreateContextsLength)); + le32_add_cpu(&rsp->CreateContextsLength, + conn->vals->create_durable_size); + iov_len += conn->vals->create_durable_size; + } else { + create_durable_v2_rsp_buf(rsp->Buffer + + le32_to_cpu(rsp->CreateContextsLength), + fp); + le32_add_cpu(&rsp->CreateContextsLength, + conn->vals->create_durable_v2_size); + iov_len += conn->vals->create_durable_v2_size; + } + + if (next_ptr) + *next_ptr = cpu_to_le32(next_off); + next_ptr = &durable_ccontext->Next; + next_off = conn->vals->create_durable_size; + } + if (posix_ctxt) { contxt_cnt++; create_posix_rsp_buf(rsp->Buffer + @@ -3828,11 +4053,16 @@ static int process_query_dir_entries(struct smb2_query_dir_private *priv) } ksmbd_kstat.kstat = &kstat; - if (priv->info_level != FILE_NAMES_INFORMATION) - ksmbd_vfs_fill_dentry_attrs(priv->work, - idmap, - dent, - &ksmbd_kstat); + if (priv->info_level != FILE_NAMES_INFORMATION) { + rc = ksmbd_vfs_fill_dentry_attrs(priv->work, + idmap, + dent, + &ksmbd_kstat); + if (rc) { + dput(dent); + continue; + } + } rc = smb2_populate_readdir_entry(priv->work->conn, priv->info_level, @@ -4075,7 +4305,7 @@ int smb2_query_dir(struct ksmbd_work *work) } srch_flag = req->Flags; - srch_ptr = smb_strndup_from_utf16(req->Buffer, + srch_ptr = smb_strndup_from_utf16((char *)req + le16_to_cpu(req->FileNameOffset), le16_to_cpu(req->FileNameLength), 1, conn->local_nls); if (IS_ERR(srch_ptr)) { @@ -4335,7 +4565,8 @@ static int smb2_get_ea(struct ksmbd_work *work, struct ksmbd_file *fp, sizeof(struct smb2_ea_info_req)) return -EINVAL; - ea_req = (struct smb2_ea_info_req *)req->Buffer; + ea_req = (struct smb2_ea_info_req *)((char *)req + + le16_to_cpu(req->InputBufferOffset)); } else { /* need to send all EAs, if no specific EA is requested*/ if (le32_to_cpu(req->Flags) & SL_RETURN_SINGLE_ENTRY) @@ -4480,6 +4711,7 @@ static int get_file_basic_info(struct smb2_query_info_rsp *rsp, struct smb2_file_basic_info *basic_info; struct kstat stat; u64 time; + int ret; if (!(fp->daccess & FILE_READ_ATTRIBUTES_LE)) { pr_err("no right to read the attributes : 0x%x\n", @@ -4487,9 +4719,12 @@ static int get_file_basic_info(struct smb2_query_info_rsp *rsp, return -EACCES; } + ret = vfs_getattr(&fp->filp->f_path, &stat, STATX_BASIC_STATS, + AT_STATX_SYNC_AS_STAT); + if (ret) + return ret; + basic_info = (struct smb2_file_basic_info *)rsp->Buffer; - generic_fillattr(file_mnt_idmap(fp->filp), STATX_BASIC_STATS, - file_inode(fp->filp), &stat); basic_info->CreationTime = cpu_to_le64(fp->create_time); time = ksmbd_UnixTimeToNT(stat.atime); basic_info->LastAccessTime = cpu_to_le64(time); @@ -4504,27 +4739,31 @@ static int get_file_basic_info(struct smb2_query_info_rsp *rsp, return 0; } -static void get_file_standard_info(struct smb2_query_info_rsp *rsp, - struct ksmbd_file *fp, void *rsp_org) +static int get_file_standard_info(struct smb2_query_info_rsp *rsp, + struct ksmbd_file *fp, void *rsp_org) { struct smb2_file_standard_info *sinfo; unsigned int delete_pending; - struct inode *inode; struct kstat stat; + int ret; - inode = file_inode(fp->filp); - generic_fillattr(file_mnt_idmap(fp->filp), STATX_BASIC_STATS, inode, &stat); + ret = vfs_getattr(&fp->filp->f_path, &stat, STATX_BASIC_STATS, + AT_STATX_SYNC_AS_STAT); + if (ret) + return ret; sinfo = (struct smb2_file_standard_info *)rsp->Buffer; delete_pending = ksmbd_inode_pending_delete(fp); - sinfo->AllocationSize = cpu_to_le64(inode->i_blocks << 9); + sinfo->AllocationSize = cpu_to_le64(stat.blocks << 9); sinfo->EndOfFile = S_ISDIR(stat.mode) ? 0 : cpu_to_le64(stat.size); sinfo->NumberOfLinks = cpu_to_le32(get_nlink(&stat) - delete_pending); sinfo->DeletePending = delete_pending; sinfo->Directory = S_ISDIR(stat.mode) ? 1 : 0; rsp->OutputBufferLength = cpu_to_le32(sizeof(struct smb2_file_standard_info)); + + return 0; } static void get_file_alignment_info(struct smb2_query_info_rsp *rsp, @@ -4546,11 +4785,11 @@ static int get_file_all_info(struct ksmbd_work *work, struct ksmbd_conn *conn = work->conn; struct smb2_file_all_info *file_info; unsigned int delete_pending; - struct inode *inode; struct kstat stat; int conv_len; char *filename; u64 time; + int ret; if (!(fp->daccess & FILE_READ_ATTRIBUTES_LE)) { ksmbd_debug(SMB, "no right to read the attributes : 0x%x\n", @@ -4562,8 +4801,10 @@ static int get_file_all_info(struct ksmbd_work *work, if (IS_ERR(filename)) return PTR_ERR(filename); - inode = file_inode(fp->filp); - generic_fillattr(file_mnt_idmap(fp->filp), STATX_BASIC_STATS, inode, &stat); + ret = vfs_getattr(&fp->filp->f_path, &stat, STATX_BASIC_STATS, + AT_STATX_SYNC_AS_STAT); + if (ret) + return ret; ksmbd_debug(SMB, "filename = %s\n", filename); delete_pending = ksmbd_inode_pending_delete(fp); @@ -4579,7 +4820,7 @@ static int get_file_all_info(struct ksmbd_work *work, file_info->Attributes = fp->f_ci->m_fattr; file_info->Pad1 = 0; file_info->AllocationSize = - cpu_to_le64(inode->i_blocks << 9); + cpu_to_le64(stat.blocks << 9); file_info->EndOfFile = S_ISDIR(stat.mode) ? 0 : cpu_to_le64(stat.size); file_info->NumberOfLinks = cpu_to_le32(get_nlink(&stat) - delete_pending); @@ -4623,10 +4864,10 @@ static void get_file_alternate_info(struct ksmbd_work *work, cpu_to_le32(sizeof(struct smb2_file_alt_name_info) + conv_len); } -static void get_file_stream_info(struct ksmbd_work *work, - struct smb2_query_info_rsp *rsp, - struct ksmbd_file *fp, - void *rsp_org) +static int get_file_stream_info(struct ksmbd_work *work, + struct smb2_query_info_rsp *rsp, + struct ksmbd_file *fp, + void *rsp_org) { struct ksmbd_conn *conn = work->conn; struct smb2_file_stream_info *file_info; @@ -4637,9 +4878,13 @@ static void get_file_stream_info(struct ksmbd_work *work, int nbytes = 0, streamlen, stream_name_len, next, idx = 0; int buf_free_len; struct smb2_query_info_req *req = ksmbd_req_buf_next(work); + int ret; + + ret = vfs_getattr(&fp->filp->f_path, &stat, STATX_BASIC_STATS, + AT_STATX_SYNC_AS_STAT); + if (ret) + return ret; - generic_fillattr(file_mnt_idmap(fp->filp), STATX_BASIC_STATS, - file_inode(fp->filp), &stat); file_info = (struct smb2_file_stream_info *)rsp->Buffer; buf_free_len = @@ -4720,29 +4965,37 @@ out: kvfree(xattr_list); rsp->OutputBufferLength = cpu_to_le32(nbytes); + + return 0; } -static void get_file_internal_info(struct smb2_query_info_rsp *rsp, - struct ksmbd_file *fp, void *rsp_org) +static int get_file_internal_info(struct smb2_query_info_rsp *rsp, + struct ksmbd_file *fp, void *rsp_org) { struct smb2_file_internal_info *file_info; struct kstat stat; + int ret; + + ret = vfs_getattr(&fp->filp->f_path, &stat, STATX_BASIC_STATS, + AT_STATX_SYNC_AS_STAT); + if (ret) + return ret; - generic_fillattr(file_mnt_idmap(fp->filp), STATX_BASIC_STATS, - file_inode(fp->filp), &stat); file_info = (struct smb2_file_internal_info *)rsp->Buffer; file_info->IndexNumber = cpu_to_le64(stat.ino); rsp->OutputBufferLength = cpu_to_le32(sizeof(struct smb2_file_internal_info)); + + return 0; } static int get_file_network_open_info(struct smb2_query_info_rsp *rsp, struct ksmbd_file *fp, void *rsp_org) { struct smb2_file_ntwrk_info *file_info; - struct inode *inode; struct kstat stat; u64 time; + int ret; if (!(fp->daccess & FILE_READ_ATTRIBUTES_LE)) { pr_err("no right to read the attributes : 0x%x\n", @@ -4750,10 +5003,12 @@ static int get_file_network_open_info(struct smb2_query_info_rsp *rsp, return -EACCES; } - file_info = (struct smb2_file_ntwrk_info *)rsp->Buffer; + ret = vfs_getattr(&fp->filp->f_path, &stat, STATX_BASIC_STATS, + AT_STATX_SYNC_AS_STAT); + if (ret) + return ret; - inode = file_inode(fp->filp); - generic_fillattr(file_mnt_idmap(fp->filp), STATX_BASIC_STATS, inode, &stat); + file_info = (struct smb2_file_ntwrk_info *)rsp->Buffer; file_info->CreationTime = cpu_to_le64(fp->create_time); time = ksmbd_UnixTimeToNT(stat.atime); @@ -4763,8 +5018,7 @@ static int get_file_network_open_info(struct smb2_query_info_rsp *rsp, time = ksmbd_UnixTimeToNT(stat.ctime); file_info->ChangeTime = cpu_to_le64(time); file_info->Attributes = fp->f_ci->m_fattr; - file_info->AllocationSize = - cpu_to_le64(inode->i_blocks << 9); + file_info->AllocationSize = cpu_to_le64(stat.blocks << 9); file_info->EndOfFile = S_ISDIR(stat.mode) ? 0 : cpu_to_le64(stat.size); file_info->Reserved = cpu_to_le32(0); rsp->OutputBufferLength = @@ -4804,14 +5058,17 @@ static void get_file_mode_info(struct smb2_query_info_rsp *rsp, cpu_to_le32(sizeof(struct smb2_file_mode_info)); } -static void get_file_compression_info(struct smb2_query_info_rsp *rsp, - struct ksmbd_file *fp, void *rsp_org) +static int get_file_compression_info(struct smb2_query_info_rsp *rsp, + struct ksmbd_file *fp, void *rsp_org) { struct smb2_file_comp_info *file_info; struct kstat stat; + int ret; - generic_fillattr(file_mnt_idmap(fp->filp), STATX_BASIC_STATS, - file_inode(fp->filp), &stat); + ret = vfs_getattr(&fp->filp->f_path, &stat, STATX_BASIC_STATS, + AT_STATX_SYNC_AS_STAT); + if (ret) + return ret; file_info = (struct smb2_file_comp_info *)rsp->Buffer; file_info->CompressedFileSize = cpu_to_le64(stat.blocks << 9); @@ -4823,6 +5080,8 @@ static void get_file_compression_info(struct smb2_query_info_rsp *rsp, rsp->OutputBufferLength = cpu_to_le32(sizeof(struct smb2_file_comp_info)); + + return 0; } static int get_file_attribute_tag_info(struct smb2_query_info_rsp *rsp, @@ -4844,7 +5103,7 @@ static int get_file_attribute_tag_info(struct smb2_query_info_rsp *rsp, return 0; } -static void find_file_posix_info(struct smb2_query_info_rsp *rsp, +static int find_file_posix_info(struct smb2_query_info_rsp *rsp, struct ksmbd_file *fp, void *rsp_org) { struct smb311_posix_qinfo *file_info; @@ -4852,24 +5111,31 @@ static void find_file_posix_info(struct smb2_query_info_rsp *rsp, struct mnt_idmap *idmap = file_mnt_idmap(fp->filp); vfsuid_t vfsuid = i_uid_into_vfsuid(idmap, inode); vfsgid_t vfsgid = i_gid_into_vfsgid(idmap, inode); + struct kstat stat; u64 time; int out_buf_len = sizeof(struct smb311_posix_qinfo) + 32; + int ret; + + ret = vfs_getattr(&fp->filp->f_path, &stat, STATX_BASIC_STATS, + AT_STATX_SYNC_AS_STAT); + if (ret) + return ret; file_info = (struct smb311_posix_qinfo *)rsp->Buffer; file_info->CreationTime = cpu_to_le64(fp->create_time); - time = ksmbd_UnixTimeToNT(inode_get_atime(inode)); + time = ksmbd_UnixTimeToNT(stat.atime); file_info->LastAccessTime = cpu_to_le64(time); - time = ksmbd_UnixTimeToNT(inode_get_mtime(inode)); + time = ksmbd_UnixTimeToNT(stat.mtime); file_info->LastWriteTime = cpu_to_le64(time); - time = ksmbd_UnixTimeToNT(inode_get_ctime(inode)); + time = ksmbd_UnixTimeToNT(stat.ctime); file_info->ChangeTime = cpu_to_le64(time); file_info->DosAttributes = fp->f_ci->m_fattr; - file_info->Inode = cpu_to_le64(inode->i_ino); - file_info->EndOfFile = cpu_to_le64(inode->i_size); - file_info->AllocationSize = cpu_to_le64(inode->i_blocks << 9); - file_info->HardLinks = cpu_to_le32(inode->i_nlink); - file_info->Mode = cpu_to_le32(inode->i_mode & 0777); - file_info->DeviceId = cpu_to_le32(inode->i_rdev); + file_info->Inode = cpu_to_le64(stat.ino); + file_info->EndOfFile = cpu_to_le64(stat.size); + file_info->AllocationSize = cpu_to_le64(stat.blocks << 9); + file_info->HardLinks = cpu_to_le32(stat.nlink); + file_info->Mode = cpu_to_le32(stat.mode & 0777); + file_info->DeviceId = cpu_to_le32(stat.rdev); /* * Sids(32) contain two sids(Domain sid(16), UNIX group sid(16)). @@ -4882,6 +5148,8 @@ static void find_file_posix_info(struct smb2_query_info_rsp *rsp, SIDUNIX_GROUP, (struct smb_sid *)&file_info->Sids[16]); rsp->OutputBufferLength = cpu_to_le32(out_buf_len); + + return 0; } static int smb2_get_info_file(struct ksmbd_work *work, @@ -4930,7 +5198,7 @@ static int smb2_get_info_file(struct ksmbd_work *work, break; case FILE_STANDARD_INFORMATION: - get_file_standard_info(rsp, fp, work->response_buf); + rc = get_file_standard_info(rsp, fp, work->response_buf); break; case FILE_ALIGNMENT_INFORMATION: @@ -4946,11 +5214,11 @@ static int smb2_get_info_file(struct ksmbd_work *work, break; case FILE_STREAM_INFORMATION: - get_file_stream_info(work, rsp, fp, work->response_buf); + rc = get_file_stream_info(work, rsp, fp, work->response_buf); break; case FILE_INTERNAL_INFORMATION: - get_file_internal_info(rsp, fp, work->response_buf); + rc = get_file_internal_info(rsp, fp, work->response_buf); break; case FILE_NETWORK_OPEN_INFORMATION: @@ -4974,7 +5242,7 @@ static int smb2_get_info_file(struct ksmbd_work *work, break; case FILE_COMPRESSION_INFORMATION: - get_file_compression_info(rsp, fp, work->response_buf); + rc = get_file_compression_info(rsp, fp, work->response_buf); break; case FILE_ATTRIBUTE_TAG_INFORMATION: @@ -4985,7 +5253,7 @@ static int smb2_get_info_file(struct ksmbd_work *work, pr_err("client doesn't negotiate with SMB3.1.1 POSIX Extensions\n"); rc = -EOPNOTSUPP; } else { - find_file_posix_info(rsp, fp, work->response_buf); + rc = find_file_posix_info(rsp, fp, work->response_buf); } break; default: @@ -5398,7 +5666,6 @@ int smb2_close(struct ksmbd_work *work) struct smb2_close_rsp *rsp; struct ksmbd_conn *conn = work->conn; struct ksmbd_file *fp; - struct inode *inode; u64 time; int err = 0; @@ -5453,24 +5720,33 @@ int smb2_close(struct ksmbd_work *work) rsp->Reserved = 0; if (req->Flags == SMB2_CLOSE_FLAG_POSTQUERY_ATTRIB) { + struct kstat stat; + int ret; + fp = ksmbd_lookup_fd_fast(work, volatile_id); if (!fp) { err = -ENOENT; goto out; } - inode = file_inode(fp->filp); + ret = vfs_getattr(&fp->filp->f_path, &stat, STATX_BASIC_STATS, + AT_STATX_SYNC_AS_STAT); + if (ret) { + ksmbd_fd_put(work, fp); + goto out; + } + rsp->Flags = SMB2_CLOSE_FLAG_POSTQUERY_ATTRIB; - rsp->AllocationSize = S_ISDIR(inode->i_mode) ? 0 : - cpu_to_le64(inode->i_blocks << 9); - rsp->EndOfFile = cpu_to_le64(inode->i_size); + rsp->AllocationSize = S_ISDIR(stat.mode) ? 0 : + cpu_to_le64(stat.blocks << 9); + rsp->EndOfFile = cpu_to_le64(stat.size); rsp->Attributes = fp->f_ci->m_fattr; rsp->CreationTime = cpu_to_le64(fp->create_time); - time = ksmbd_UnixTimeToNT(inode_get_atime(inode)); + time = ksmbd_UnixTimeToNT(stat.atime); rsp->LastAccessTime = cpu_to_le64(time); - time = ksmbd_UnixTimeToNT(inode_get_mtime(inode)); + time = ksmbd_UnixTimeToNT(stat.mtime); rsp->LastWriteTime = cpu_to_le64(time); - time = ksmbd_UnixTimeToNT(inode_get_ctime(inode)); + time = ksmbd_UnixTimeToNT(stat.ctime); rsp->ChangeTime = cpu_to_le64(time); ksmbd_fd_put(work, fp); } else { @@ -5759,15 +6035,21 @@ static int set_file_allocation_info(struct ksmbd_work *work, loff_t alloc_blks; struct inode *inode; + struct kstat stat; int rc; if (!(fp->daccess & FILE_WRITE_DATA_LE)) return -EACCES; + rc = vfs_getattr(&fp->filp->f_path, &stat, STATX_BASIC_STATS, + AT_STATX_SYNC_AS_STAT); + if (rc) + return rc; + alloc_blks = (le64_to_cpu(file_alloc_info->AllocationSize) + 511) >> 9; inode = file_inode(fp->filp); - if (alloc_blks > inode->i_blocks) { + if (alloc_blks > stat.blocks) { smb_break_all_levII_oplock(work, fp, 1); rc = vfs_fallocate(fp->filp, FALLOC_FL_KEEP_SIZE, 0, alloc_blks * 512); @@ -5775,7 +6057,7 @@ static int set_file_allocation_info(struct ksmbd_work *work, pr_err("vfs_fallocate is failed : %d\n", rc); return rc; } - } else if (alloc_blks < inode->i_blocks) { + } else if (alloc_blks < stat.blocks) { loff_t size; /* @@ -5930,6 +6212,7 @@ static int smb2_set_info_file(struct ksmbd_work *work, struct ksmbd_file *fp, struct ksmbd_share_config *share) { unsigned int buf_len = le32_to_cpu(req->BufferLength); + char *buffer = (char *)req + le16_to_cpu(req->BufferOffset); switch (req->FileInfoClass) { case FILE_BASIC_INFORMATION: @@ -5937,7 +6220,7 @@ static int smb2_set_info_file(struct ksmbd_work *work, struct ksmbd_file *fp, if (buf_len < sizeof(struct smb2_file_basic_info)) return -EINVAL; - return set_file_basic_info(fp, (struct smb2_file_basic_info *)req->Buffer, share); + return set_file_basic_info(fp, (struct smb2_file_basic_info *)buffer, share); } case FILE_ALLOCATION_INFORMATION: { @@ -5945,7 +6228,7 @@ static int smb2_set_info_file(struct ksmbd_work *work, struct ksmbd_file *fp, return -EINVAL; return set_file_allocation_info(work, fp, - (struct smb2_file_alloc_info *)req->Buffer); + (struct smb2_file_alloc_info *)buffer); } case FILE_END_OF_FILE_INFORMATION: { @@ -5953,7 +6236,7 @@ static int smb2_set_info_file(struct ksmbd_work *work, struct ksmbd_file *fp, return -EINVAL; return set_end_of_file_info(work, fp, - (struct smb2_file_eof_info *)req->Buffer); + (struct smb2_file_eof_info *)buffer); } case FILE_RENAME_INFORMATION: { @@ -5961,7 +6244,7 @@ static int smb2_set_info_file(struct ksmbd_work *work, struct ksmbd_file *fp, return -EINVAL; return set_rename_info(work, fp, - (struct smb2_file_rename_info *)req->Buffer, + (struct smb2_file_rename_info *)buffer, buf_len); } case FILE_LINK_INFORMATION: @@ -5970,7 +6253,7 @@ static int smb2_set_info_file(struct ksmbd_work *work, struct ksmbd_file *fp, return -EINVAL; return smb2_create_link(work, work->tcon->share_conf, - (struct smb2_file_link_info *)req->Buffer, + (struct smb2_file_link_info *)buffer, buf_len, fp->filp, work->conn->local_nls); } @@ -5980,7 +6263,7 @@ static int smb2_set_info_file(struct ksmbd_work *work, struct ksmbd_file *fp, return -EINVAL; return set_file_disposition_info(fp, - (struct smb2_file_disposition_info *)req->Buffer); + (struct smb2_file_disposition_info *)buffer); } case FILE_FULL_EA_INFORMATION: { @@ -5993,7 +6276,7 @@ static int smb2_set_info_file(struct ksmbd_work *work, struct ksmbd_file *fp, if (buf_len < sizeof(struct smb2_ea_info)) return -EINVAL; - return smb2_set_ea((struct smb2_ea_info *)req->Buffer, + return smb2_set_ea((struct smb2_ea_info *)buffer, buf_len, &fp->filp->f_path, true); } case FILE_POSITION_INFORMATION: @@ -6001,14 +6284,14 @@ static int smb2_set_info_file(struct ksmbd_work *work, struct ksmbd_file *fp, if (buf_len < sizeof(struct smb2_file_pos_info)) return -EINVAL; - return set_file_position_info(fp, (struct smb2_file_pos_info *)req->Buffer); + return set_file_position_info(fp, (struct smb2_file_pos_info *)buffer); } case FILE_MODE_INFORMATION: { if (buf_len < sizeof(struct smb2_file_mode_info)) return -EINVAL; - return set_file_mode_info(fp, (struct smb2_file_mode_info *)req->Buffer); + return set_file_mode_info(fp, (struct smb2_file_mode_info *)buffer); } } @@ -6089,7 +6372,7 @@ int smb2_set_info(struct ksmbd_work *work) } rc = smb2_set_info_sec(fp, le32_to_cpu(req->AdditionalInformation), - req->Buffer, + (char *)req + le16_to_cpu(req->BufferOffset), le32_to_cpu(req->BufferLength)); ksmbd_revert_fsids(work); break; @@ -7535,7 +7818,7 @@ static int fsctl_pipe_transceive(struct ksmbd_work *work, u64 id, struct smb2_ioctl_rsp *rsp) { struct ksmbd_rpc_command *rpc_resp; - char *data_buf = (char *)&req->Buffer[0]; + char *data_buf = (char *)req + le32_to_cpu(req->InputOffset); int nbytes = 0; rpc_resp = ksmbd_rpc_ioctl(work->sess, id, data_buf, @@ -7648,6 +7931,7 @@ int smb2_ioctl(struct ksmbd_work *work) u64 id = KSMBD_NO_FID; struct ksmbd_conn *conn = work->conn; int ret = 0; + char *buffer; if (work->next_smb2_rcv_hdr_off) { req = ksmbd_req_buf_next(work); @@ -7670,6 +7954,8 @@ int smb2_ioctl(struct ksmbd_work *work) goto out; } + buffer = (char *)req + le32_to_cpu(req->InputOffset); + cnt_code = le32_to_cpu(req->CtlCode); ret = smb2_calc_max_out_buf_len(work, 48, le32_to_cpu(req->MaxOutputResponse)); @@ -7727,7 +8013,7 @@ int smb2_ioctl(struct ksmbd_work *work) } ret = fsctl_validate_negotiate_info(conn, - (struct validate_negotiate_info_req *)&req->Buffer[0], + (struct validate_negotiate_info_req *)buffer, (struct validate_negotiate_info_rsp *)&rsp->Buffer[0], in_buf_len); if (ret < 0) @@ -7780,7 +8066,7 @@ int smb2_ioctl(struct ksmbd_work *work) rsp->VolatileFileId = req->VolatileFileId; rsp->PersistentFileId = req->PersistentFileId; fsctl_copychunk(work, - (struct copychunk_ioctl_req *)&req->Buffer[0], + (struct copychunk_ioctl_req *)buffer, le32_to_cpu(req->CtlCode), le32_to_cpu(req->InputCount), req->VolatileFileId, @@ -7793,8 +8079,7 @@ int smb2_ioctl(struct ksmbd_work *work) goto out; } - ret = fsctl_set_sparse(work, id, - (struct file_sparse *)&req->Buffer[0]); + ret = fsctl_set_sparse(work, id, (struct file_sparse *)buffer); if (ret < 0) goto out; break; @@ -7817,7 +8102,7 @@ int smb2_ioctl(struct ksmbd_work *work) } zero_data = - (struct file_zero_data_information *)&req->Buffer[0]; + (struct file_zero_data_information *)buffer; off = le64_to_cpu(zero_data->FileOffset); bfz = le64_to_cpu(zero_data->BeyondFinalZero); @@ -7848,7 +8133,7 @@ int smb2_ioctl(struct ksmbd_work *work) } ret = fsctl_query_allocated_ranges(work, id, - (struct file_allocated_range_buffer *)&req->Buffer[0], + (struct file_allocated_range_buffer *)buffer, (struct file_allocated_range_buffer *)&rsp->Buffer[0], out_buf_len / sizeof(struct file_allocated_range_buffer), &nbytes); @@ -7892,7 +8177,7 @@ int smb2_ioctl(struct ksmbd_work *work) goto out; } - dup_ext = (struct duplicate_extents_to_file *)&req->Buffer[0]; + dup_ext = (struct duplicate_extents_to_file *)buffer; fp_in = ksmbd_lookup_fd_slow(work, dup_ext->VolatileFileHandle, dup_ext->PersistentFileHandle); diff --git a/fs/smb/server/smb2pdu.h b/fs/smb/server/smb2pdu.h index d12cfd3b0927..bd1d2a0e9203 100644 --- a/fs/smb/server/smb2pdu.h +++ b/fs/smb/server/smb2pdu.h @@ -72,6 +72,18 @@ struct create_durable_req_v2 { __u8 CreateGuid[16]; } __packed; +struct create_durable_reconn_req { + struct create_context ccontext; + __u8 Name[8]; + union { + __u8 Reserved[16]; + struct { + __u64 PersistentFileId; + __u64 VolatileFileId; + } Fid; + } Data; +} __packed; + struct create_durable_reconn_v2_req { struct create_context ccontext; __u8 Name[8]; @@ -98,6 +110,9 @@ struct create_durable_rsp { } Data; } __packed; +/* See MS-SMB2 2.2.13.2.11 */ +/* Flags */ +#define SMB2_DHANDLE_FLAG_PERSISTENT 0x00000002 struct create_durable_v2_rsp { struct create_context ccontext; __u8 Name[8]; diff --git a/fs/smb/server/smb_common.c b/fs/smb/server/smb_common.c index 7c98bf699772..fcaf373cc008 100644 --- a/fs/smb/server/smb_common.c +++ b/fs/smb/server/smb_common.c @@ -457,10 +457,13 @@ int ksmbd_populate_dot_dotdot_entries(struct ksmbd_work *work, int info_level, } ksmbd_kstat.kstat = &kstat; - ksmbd_vfs_fill_dentry_attrs(work, - idmap, - dentry, - &ksmbd_kstat); + rc = ksmbd_vfs_fill_dentry_attrs(work, + idmap, + dentry, + &ksmbd_kstat); + if (rc) + break; + rc = fn(conn, info_level, d_info, &ksmbd_kstat); if (rc) break; diff --git a/fs/smb/server/vfs.c b/fs/smb/server/vfs.c index c487e834331a..22f0f3db3ac9 100644 --- a/fs/smb/server/vfs.c +++ b/fs/smb/server/vfs.c @@ -1682,11 +1682,19 @@ int ksmbd_vfs_fill_dentry_attrs(struct ksmbd_work *work, struct dentry *dentry, struct ksmbd_kstat *ksmbd_kstat) { + struct ksmbd_share_config *share_conf = work->tcon->share_conf; u64 time; int rc; + struct path path = { + .mnt = share_conf->vfs_path.mnt, + .dentry = dentry, + }; - generic_fillattr(idmap, STATX_BASIC_STATS, d_inode(dentry), - ksmbd_kstat->kstat); + rc = vfs_getattr(&path, ksmbd_kstat->kstat, + STATX_BASIC_STATS | STATX_BTIME, + AT_STATX_SYNC_AS_STAT); + if (rc) + return rc; time = ksmbd_UnixTimeToNT(ksmbd_kstat->kstat->ctime); ksmbd_kstat->create_time = time; diff --git a/fs/smb/server/vfs_cache.c b/fs/smb/server/vfs_cache.c index 4e82ff627d12..030f70700036 100644 --- a/fs/smb/server/vfs_cache.c +++ b/fs/smb/server/vfs_cache.c @@ -305,7 +305,8 @@ static void __ksmbd_close_fd(struct ksmbd_file_table *ft, struct ksmbd_file *fp) fd_limit_close(); __ksmbd_remove_durable_fd(fp); - __ksmbd_remove_fd(ft, fp); + if (ft) + __ksmbd_remove_fd(ft, fp); close_id_del_oplock(fp); filp = fp->filp; @@ -465,11 +466,32 @@ struct ksmbd_file *ksmbd_lookup_fd_slow(struct ksmbd_work *work, u64 id, return fp; } -struct ksmbd_file *ksmbd_lookup_durable_fd(unsigned long long id) +struct ksmbd_file *ksmbd_lookup_global_fd(unsigned long long id) { return __ksmbd_lookup_fd(&global_ft, id); } +struct ksmbd_file *ksmbd_lookup_durable_fd(unsigned long long id) +{ + struct ksmbd_file *fp; + + fp = __ksmbd_lookup_fd(&global_ft, id); + if (fp && fp->conn) { + ksmbd_put_durable_fd(fp); + fp = NULL; + } + + return fp; +} + +void ksmbd_put_durable_fd(struct ksmbd_file *fp) +{ + if (!atomic_dec_and_test(&fp->refcount)) + return; + + __ksmbd_close_fd(NULL, fp); +} + struct ksmbd_file *ksmbd_lookup_fd_cguid(char *cguid) { struct ksmbd_file *fp = NULL; @@ -639,6 +661,32 @@ __close_file_table_ids(struct ksmbd_file_table *ft, return num; } +static inline bool is_reconnectable(struct ksmbd_file *fp) +{ + struct oplock_info *opinfo = opinfo_get(fp); + bool reconn = false; + + if (!opinfo) + return false; + + if (opinfo->op_state != OPLOCK_STATE_NONE) { + opinfo_put(opinfo); + return false; + } + + if (fp->is_resilient || fp->is_persistent) + reconn = true; + else if (fp->is_durable && opinfo->is_lease && + opinfo->o_lease->state & SMB2_LEASE_HANDLE_CACHING_LE) + reconn = true; + + else if (fp->is_durable && opinfo->level == SMB2_OPLOCK_LEVEL_BATCH) + reconn = true; + + opinfo_put(opinfo); + return reconn; +} + static bool tree_conn_fd_check(struct ksmbd_tree_connect *tcon, struct ksmbd_file *fp) { @@ -648,7 +696,28 @@ static bool tree_conn_fd_check(struct ksmbd_tree_connect *tcon, static bool session_fd_check(struct ksmbd_tree_connect *tcon, struct ksmbd_file *fp) { - return false; + struct ksmbd_inode *ci; + struct oplock_info *op; + struct ksmbd_conn *conn; + + if (!is_reconnectable(fp)) + return false; + + conn = fp->conn; + ci = fp->f_ci; + write_lock(&ci->m_lock); + list_for_each_entry_rcu(op, &ci->m_op_list, op_entry) { + if (op->conn != conn) + continue; + op->conn = NULL; + } + write_unlock(&ci->m_lock); + + fp->conn = NULL; + fp->tcon = NULL; + fp->volatile_id = KSMBD_NO_FID; + + return true; } void ksmbd_close_tree_conn_fds(struct ksmbd_work *work) @@ -687,6 +756,68 @@ void ksmbd_free_global_file_table(void) ksmbd_destroy_file_table(&global_ft); } +int ksmbd_validate_name_reconnect(struct ksmbd_share_config *share, + struct ksmbd_file *fp, char *name) +{ + char *pathname, *ab_pathname; + int ret = 0; + + pathname = kmalloc(PATH_MAX, GFP_KERNEL); + if (!pathname) + return -EACCES; + + ab_pathname = d_path(&fp->filp->f_path, pathname, PATH_MAX); + if (IS_ERR(ab_pathname)) { + kfree(pathname); + return -EACCES; + } + + if (name && strcmp(&ab_pathname[share->path_sz + 1], name)) { + ksmbd_debug(SMB, "invalid name reconnect %s\n", name); + ret = -EINVAL; + } + + kfree(pathname); + + return ret; +} + +int ksmbd_reopen_durable_fd(struct ksmbd_work *work, struct ksmbd_file *fp) +{ + struct ksmbd_inode *ci; + struct oplock_info *op; + + if (!fp->is_durable || fp->conn || fp->tcon) { + pr_err("Invalid durable fd [%p:%p]\n", fp->conn, fp->tcon); + return -EBADF; + } + + if (has_file_id(fp->volatile_id)) { + pr_err("Still in use durable fd: %llu\n", fp->volatile_id); + return -EBADF; + } + + fp->conn = work->conn; + fp->tcon = work->tcon; + + ci = fp->f_ci; + write_lock(&ci->m_lock); + list_for_each_entry_rcu(op, &ci->m_op_list, op_entry) { + if (op->conn) + continue; + op->conn = fp->conn; + } + write_unlock(&ci->m_lock); + + __open_id(&work->sess->file_table, fp, OPEN_ID_TYPE_VOLATILE_ID); + if (!has_file_id(fp->volatile_id)) { + fp->conn = NULL; + fp->tcon = NULL; + return -EBADF; + } + return 0; +} + int ksmbd_init_file_table(struct ksmbd_file_table *ft) { ft->idr = kzalloc(sizeof(struct idr), GFP_KERNEL); diff --git a/fs/smb/server/vfs_cache.h b/fs/smb/server/vfs_cache.h index a528f0cc775a..ed44fb4e18e7 100644 --- a/fs/smb/server/vfs_cache.h +++ b/fs/smb/server/vfs_cache.h @@ -14,6 +14,7 @@ #include <linux/workqueue.h> #include "vfs.h" +#include "mgmt/share_config.h" /* Windows style file permissions for extended response */ #define FILE_GENERIC_ALL 0x1F01FF @@ -106,6 +107,9 @@ struct ksmbd_file { int dot_dotdot[2]; unsigned int f_state; bool reserve_lease_break; + bool is_durable; + bool is_persistent; + bool is_resilient; }; static inline void set_ctx_actor(struct dir_context *ctx, @@ -141,7 +145,9 @@ struct ksmbd_file *ksmbd_lookup_fd_slow(struct ksmbd_work *work, u64 id, void ksmbd_fd_put(struct ksmbd_work *work, struct ksmbd_file *fp); struct ksmbd_inode *ksmbd_inode_lookup_lock(struct dentry *d); void ksmbd_inode_put(struct ksmbd_inode *ci); +struct ksmbd_file *ksmbd_lookup_global_fd(unsigned long long id); struct ksmbd_file *ksmbd_lookup_durable_fd(unsigned long long id); +void ksmbd_put_durable_fd(struct ksmbd_file *fp); struct ksmbd_file *ksmbd_lookup_fd_cguid(char *cguid); struct ksmbd_file *ksmbd_lookup_fd_inode(struct dentry *dentry); unsigned int ksmbd_open_durable_fd(struct ksmbd_file *fp); @@ -173,6 +179,9 @@ void ksmbd_set_inode_pending_delete(struct ksmbd_file *fp); void ksmbd_clear_inode_pending_delete(struct ksmbd_file *fp); void ksmbd_fd_set_delete_on_close(struct ksmbd_file *fp, int file_info); +int ksmbd_reopen_durable_fd(struct ksmbd_work *work, struct ksmbd_file *fp); +int ksmbd_validate_name_reconnect(struct ksmbd_share_config *share, + struct ksmbd_file *fp, char *name); int ksmbd_init_file_cache(void); void ksmbd_exit_file_cache(void); #endif /* __VFS_CACHE_H__ */ diff --git a/fs/super.c b/fs/super.c index ee05ab6b37e7..71d9779c42b1 100644 --- a/fs/super.c +++ b/fs/super.c @@ -1515,11 +1515,29 @@ static int fs_bdev_thaw(struct block_device *bdev) return error; } +static void fs_bdev_super_get(void *data) +{ + struct super_block *sb = data; + + spin_lock(&sb_lock); + sb->s_count++; + spin_unlock(&sb_lock); +} + +static void fs_bdev_super_put(void *data) +{ + struct super_block *sb = data; + + put_super(sb); +} + const struct blk_holder_ops fs_holder_ops = { .mark_dead = fs_bdev_mark_dead, .sync = fs_bdev_sync, .freeze = fs_bdev_freeze, .thaw = fs_bdev_thaw, + .get_holder = fs_bdev_super_get, + .put_holder = fs_bdev_super_put, }; EXPORT_SYMBOL_GPL(fs_holder_ops); diff --git a/fs/sysfs/group.c b/fs/sysfs/group.c index 138676463336..d22ad67a0f32 100644 --- a/fs/sysfs/group.c +++ b/fs/sysfs/group.c @@ -31,6 +31,17 @@ static void remove_files(struct kernfs_node *parent, kernfs_remove_by_name(parent, (*bin_attr)->attr.name); } +static umode_t __first_visible(const struct attribute_group *grp, struct kobject *kobj) +{ + if (grp->attrs && grp->attrs[0] && grp->is_visible) + return grp->is_visible(kobj, grp->attrs[0], 0); + + if (grp->bin_attrs && grp->bin_attrs[0] && grp->is_bin_visible) + return grp->is_bin_visible(kobj, grp->bin_attrs[0], 0); + + return 0; +} + static int create_files(struct kernfs_node *parent, struct kobject *kobj, kuid_t uid, kgid_t gid, const struct attribute_group *grp, int update) @@ -52,6 +63,7 @@ static int create_files(struct kernfs_node *parent, struct kobject *kobj, kernfs_remove_by_name(parent, (*attr)->name); if (grp->is_visible) { mode = grp->is_visible(kobj, *attr, i); + mode &= ~SYSFS_GROUP_INVISIBLE; if (!mode) continue; } @@ -81,6 +93,7 @@ static int create_files(struct kernfs_node *parent, struct kobject *kobj, (*bin_attr)->attr.name); if (grp->is_bin_visible) { mode = grp->is_bin_visible(kobj, *bin_attr, i); + mode &= ~SYSFS_GROUP_INVISIBLE; if (!mode) continue; } @@ -127,16 +140,31 @@ static int internal_create_group(struct kobject *kobj, int update, kobject_get_ownership(kobj, &uid, &gid); if (grp->name) { + umode_t mode = __first_visible(grp, kobj); + + if (mode & SYSFS_GROUP_INVISIBLE) + mode = 0; + else + mode = S_IRWXU | S_IRUGO | S_IXUGO; + if (update) { kn = kernfs_find_and_get(kobj->sd, grp->name); if (!kn) { - pr_warn("Can't update unknown attr grp name: %s/%s\n", - kobj->name, grp->name); - return -EINVAL; + pr_debug("attr grp %s/%s not created yet\n", + kobj->name, grp->name); + /* may have been invisible prior to this update */ + update = 0; + } else if (!mode) { + sysfs_remove_group(kobj, grp); + kernfs_put(kn); + return 0; } - } else { - kn = kernfs_create_dir_ns(kobj->sd, grp->name, - S_IRWXU | S_IRUGO | S_IXUGO, + } + + if (!update) { + if (!mode) + return 0; + kn = kernfs_create_dir_ns(kobj->sd, grp->name, mode, uid, gid, kobj, NULL); if (IS_ERR(kn)) { if (PTR_ERR(kn) == -EEXIST) @@ -279,9 +307,8 @@ void sysfs_remove_group(struct kobject *kobj, if (grp->name) { kn = kernfs_find_and_get(parent, grp->name); if (!kn) { - WARN(!kn, KERN_WARNING - "sysfs group '%s' not found for kobject '%s'\n", - grp->name, kobject_name(kobj)); + pr_debug("sysfs group '%s' not found for kobject '%s'\n", + grp->name, kobject_name(kobj)); return; } } else { @@ -318,13 +345,13 @@ void sysfs_remove_groups(struct kobject *kobj, EXPORT_SYMBOL_GPL(sysfs_remove_groups); /** - * sysfs_merge_group - merge files into a pre-existing attribute group. + * sysfs_merge_group - merge files into a pre-existing named attribute group. * @kobj: The kobject containing the group. * @grp: The files to create and the attribute group they belong to. * - * This function returns an error if the group doesn't exist or any of the - * files already exist in that group, in which case none of the new files - * are created. + * This function returns an error if the group doesn't exist, the .name field is + * NULL or any of the files already exist in that group, in which case none of + * the new files are created. */ int sysfs_merge_group(struct kobject *kobj, const struct attribute_group *grp) @@ -356,7 +383,7 @@ int sysfs_merge_group(struct kobject *kobj, EXPORT_SYMBOL_GPL(sysfs_merge_group); /** - * sysfs_unmerge_group - remove files from a pre-existing attribute group. + * sysfs_unmerge_group - remove files from a pre-existing named attribute group. * @kobj: The kobject containing the group. * @grp: The files to remove and the attribute group they belong to. */ diff --git a/fs/tracefs/event_inode.c b/fs/tracefs/event_inode.c index 110e8a272189..dc067eeb6387 100644 --- a/fs/tracefs/event_inode.c +++ b/fs/tracefs/event_inode.c @@ -35,6 +35,17 @@ static DEFINE_MUTEX(eventfs_mutex); /* Choose something "unique" ;-) */ #define EVENTFS_FILE_INODE_INO 0x12c4e37 +struct eventfs_root_inode { + struct eventfs_inode ei; + struct dentry *events_dir; +}; + +static struct eventfs_root_inode *get_root_inode(struct eventfs_inode *ei) +{ + WARN_ON_ONCE(!ei->is_events); + return container_of(ei, struct eventfs_root_inode, ei); +} + /* Just try to make something consistent and unique */ static int eventfs_dir_ino(struct eventfs_inode *ei) { @@ -73,12 +84,18 @@ enum { static void release_ei(struct kref *ref) { struct eventfs_inode *ei = container_of(ref, struct eventfs_inode, kref); + struct eventfs_root_inode *rei; WARN_ON_ONCE(!ei->is_freed); kfree(ei->entry_attrs); kfree_const(ei->name); - kfree_rcu(ei, rcu); + if (ei->is_events) { + rei = get_root_inode(ei); + kfree_rcu(rei, ei.rcu); + } else { + kfree_rcu(ei, rcu); + } } static inline void put_ei(struct eventfs_inode *ei) @@ -408,19 +425,43 @@ static struct dentry *lookup_dir_entry(struct dentry *dentry, return NULL; } +static inline struct eventfs_inode *init_ei(struct eventfs_inode *ei, const char *name) +{ + ei->name = kstrdup_const(name, GFP_KERNEL); + if (!ei->name) + return NULL; + kref_init(&ei->kref); + return ei; +} + static inline struct eventfs_inode *alloc_ei(const char *name) { struct eventfs_inode *ei = kzalloc(sizeof(*ei), GFP_KERNEL); + struct eventfs_inode *result; if (!ei) return NULL; - ei->name = kstrdup_const(name, GFP_KERNEL); - if (!ei->name) { + result = init_ei(ei, name); + if (!result) kfree(ei); + + return result; +} + +static inline struct eventfs_inode *alloc_root_ei(const char *name) +{ + struct eventfs_root_inode *rei = kzalloc(sizeof(*rei), GFP_KERNEL); + struct eventfs_inode *ei; + + if (!rei) return NULL; - } - kref_init(&ei->kref); + + rei->ei.is_events = 1; + ei = init_ei(&rei->ei, name); + if (!ei) + kfree(rei); + return ei; } @@ -483,7 +524,7 @@ static struct dentry *eventfs_root_lookup(struct inode *dir, struct dentry *result = NULL; ti = get_tracefs(dir); - if (!(ti->flags & TRACEFS_EVENT_INODE)) + if (WARN_ON_ONCE(!(ti->flags & TRACEFS_EVENT_INODE))) return ERR_PTR(-EIO); mutex_lock(&eventfs_mutex); @@ -495,7 +536,8 @@ static struct dentry *eventfs_root_lookup(struct inode *dir, list_for_each_entry(ei_child, &ei->children, list) { if (strcmp(ei_child->name, name) != 0) continue; - if (ei_child->is_freed) + /* A child is freed and removed from the list at the same time */ + if (WARN_ON_ONCE(ei_child->is_freed)) goto out; result = lookup_dir_entry(dentry, ei, ei_child); goto out; @@ -709,6 +751,7 @@ struct eventfs_inode *eventfs_create_events_dir(const char *name, struct dentry int size, void *data) { struct dentry *dentry = tracefs_start_creating(name, parent); + struct eventfs_root_inode *rei; struct eventfs_inode *ei; struct tracefs_inode *ti; struct inode *inode; @@ -721,7 +764,7 @@ struct eventfs_inode *eventfs_create_events_dir(const char *name, struct dentry if (IS_ERR(dentry)) return ERR_CAST(dentry); - ei = alloc_ei(name); + ei = alloc_root_ei(name); if (!ei) goto fail; @@ -730,10 +773,11 @@ struct eventfs_inode *eventfs_create_events_dir(const char *name, struct dentry goto fail; // Note: we have a ref to the dentry from tracefs_start_creating() - ei->events_dir = dentry; + rei = get_root_inode(ei); + rei->events_dir = dentry; + ei->entries = entries; ei->nr_entries = size; - ei->is_events = 1; ei->data = data; /* Save the ownership of this directory */ @@ -844,13 +888,15 @@ void eventfs_remove_dir(struct eventfs_inode *ei) */ void eventfs_remove_events_dir(struct eventfs_inode *ei) { + struct eventfs_root_inode *rei; struct dentry *dentry; - dentry = ei->events_dir; + rei = get_root_inode(ei); + dentry = rei->events_dir; if (!dentry) return; - ei->events_dir = NULL; + rei->events_dir = NULL; eventfs_remove_dir(ei); /* diff --git a/fs/tracefs/internal.h b/fs/tracefs/internal.h index beb3dcd0e434..15c26f9aaad4 100644 --- a/fs/tracefs/internal.h +++ b/fs/tracefs/internal.h @@ -36,7 +36,6 @@ struct eventfs_attr { * @children: link list into the child eventfs_inode * @entries: the array of entries representing the files in the directory * @name: the name of the directory to create - * @events_dir: the dentry of the events directory * @entry_attrs: Saved mode and ownership of the @d_children * @data: The private data to pass to the callbacks * @attr: Saved mode and ownership of eventfs_inode itself @@ -54,7 +53,6 @@ struct eventfs_inode { struct list_head children; const struct eventfs_entry *entries; const char *name; - struct dentry *events_dir; struct eventfs_attr *entry_attrs; void *data; struct eventfs_attr attr; diff --git a/fs/ubifs/debug.c b/fs/ubifs/debug.c index d013c5b3f1ed..ac77ac1fd73e 100644 --- a/fs/ubifs/debug.c +++ b/fs/ubifs/debug.c @@ -1742,17 +1742,22 @@ int dbg_check_idx_size(struct ubifs_info *c, long long idx_size) err = dbg_walk_index(c, NULL, add_size, &calc); if (err) { ubifs_err(c, "error %d while walking the index", err); - return err; + goto out_err; } if (calc != idx_size) { ubifs_err(c, "index size check failed: calculated size is %lld, should be %lld", calc, idx_size); dump_stack(); - return -EINVAL; + err = -EINVAL; + goto out_err; } return 0; + +out_err: + ubifs_destroy_tnc_tree(c); + return err; } /** diff --git a/fs/ubifs/dir.c b/fs/ubifs/dir.c index 551148de66cd..eac0fef801f1 100644 --- a/fs/ubifs/dir.c +++ b/fs/ubifs/dir.c @@ -1133,6 +1133,8 @@ out_cancel: dir_ui->ui_size = dir->i_size; mutex_unlock(&dir_ui->ui_mutex); out_inode: + /* Free inode->i_link before inode is marked as bad. */ + fscrypt_free_inode(inode); make_bad_inode(inode); iput(inode); out_fname: diff --git a/fs/ubifs/file.c b/fs/ubifs/file.c index 5029eb3390a5..a1f46919934c 100644 --- a/fs/ubifs/file.c +++ b/fs/ubifs/file.c @@ -96,36 +96,36 @@ dump: return -EINVAL; } -static int do_readpage(struct page *page) +static int do_readpage(struct folio *folio) { void *addr; int err = 0, i; unsigned int block, beyond; - struct ubifs_data_node *dn; - struct inode *inode = page->mapping->host; + struct ubifs_data_node *dn = NULL; + struct inode *inode = folio->mapping->host; struct ubifs_info *c = inode->i_sb->s_fs_info; loff_t i_size = i_size_read(inode); dbg_gen("ino %lu, pg %lu, i_size %lld, flags %#lx", - inode->i_ino, page->index, i_size, page->flags); - ubifs_assert(c, !PageChecked(page)); - ubifs_assert(c, !PagePrivate(page)); + inode->i_ino, folio->index, i_size, folio->flags); + ubifs_assert(c, !folio_test_checked(folio)); + ubifs_assert(c, !folio->private); - addr = kmap(page); + addr = kmap_local_folio(folio, 0); - block = page->index << UBIFS_BLOCKS_PER_PAGE_SHIFT; + block = folio->index << UBIFS_BLOCKS_PER_PAGE_SHIFT; beyond = (i_size + UBIFS_BLOCK_SIZE - 1) >> UBIFS_BLOCK_SHIFT; if (block >= beyond) { /* Reading beyond inode */ - SetPageChecked(page); - memset(addr, 0, PAGE_SIZE); + folio_set_checked(folio); + addr = folio_zero_tail(folio, 0, addr); goto out; } dn = kmalloc(UBIFS_MAX_DATA_NODE_SZ, GFP_NOFS); if (!dn) { err = -ENOMEM; - goto error; + goto out; } i = 0; @@ -150,39 +150,35 @@ static int do_readpage(struct page *page) memset(addr + ilen, 0, dlen - ilen); } } - if (++i >= UBIFS_BLOCKS_PER_PAGE) + if (++i >= (UBIFS_BLOCKS_PER_PAGE << folio_order(folio))) break; block += 1; addr += UBIFS_BLOCK_SIZE; + if (folio_test_highmem(folio) && (offset_in_page(addr) == 0)) { + kunmap_local(addr - UBIFS_BLOCK_SIZE); + addr = kmap_local_folio(folio, i * UBIFS_BLOCK_SIZE); + } } + if (err) { struct ubifs_info *c = inode->i_sb->s_fs_info; if (err == -ENOENT) { /* Not found, so it must be a hole */ - SetPageChecked(page); + folio_set_checked(folio); dbg_gen("hole"); - goto out_free; + err = 0; + } else { + ubifs_err(c, "cannot read page %lu of inode %lu, error %d", + folio->index, inode->i_ino, err); } - ubifs_err(c, "cannot read page %lu of inode %lu, error %d", - page->index, inode->i_ino, err); - goto error; } -out_free: - kfree(dn); out: - SetPageUptodate(page); - ClearPageError(page); - flush_dcache_page(page); - kunmap(page); - return 0; - -error: kfree(dn); - ClearPageUptodate(page); - SetPageError(page); - flush_dcache_page(page); - kunmap(page); + if (!err) + folio_mark_uptodate(folio); + flush_dcache_folio(folio); + kunmap_local(addr); return err; } @@ -222,16 +218,16 @@ static int write_begin_slow(struct address_space *mapping, pgoff_t index = pos >> PAGE_SHIFT; struct ubifs_budget_req req = { .new_page = 1 }; int err, appending = !!(pos + len > inode->i_size); - struct page *page; + struct folio *folio; dbg_gen("ino %lu, pos %llu, len %u, i_size %lld", inode->i_ino, pos, len, inode->i_size); /* - * At the slow path we have to budget before locking the page, because - * budgeting may force write-back, which would wait on locked pages and - * deadlock if we had the page locked. At this point we do not know - * anything about the page, so assume that this is a new page which is + * At the slow path we have to budget before locking the folio, because + * budgeting may force write-back, which would wait on locked folios and + * deadlock if we had the folio locked. At this point we do not know + * anything about the folio, so assume that this is a new folio which is * written to a hole. This corresponds to largest budget. Later the * budget will be amended if this is not true. */ @@ -243,45 +239,43 @@ static int write_begin_slow(struct address_space *mapping, if (unlikely(err)) return err; - page = grab_cache_page_write_begin(mapping, index); - if (unlikely(!page)) { + folio = __filemap_get_folio(mapping, index, FGP_WRITEBEGIN, + mapping_gfp_mask(mapping)); + if (IS_ERR(folio)) { ubifs_release_budget(c, &req); - return -ENOMEM; + return PTR_ERR(folio); } - if (!PageUptodate(page)) { - if (!(pos & ~PAGE_MASK) && len == PAGE_SIZE) - SetPageChecked(page); + if (!folio_test_uptodate(folio)) { + if (pos == folio_pos(folio) && len >= folio_size(folio)) + folio_set_checked(folio); else { - err = do_readpage(page); + err = do_readpage(folio); if (err) { - unlock_page(page); - put_page(page); + folio_unlock(folio); + folio_put(folio); ubifs_release_budget(c, &req); return err; } } - - SetPageUptodate(page); - ClearPageError(page); } - if (PagePrivate(page)) + if (folio->private) /* - * The page is dirty, which means it was budgeted twice: + * The folio is dirty, which means it was budgeted twice: * o first time the budget was allocated by the task which - * made the page dirty and set the PG_private flag; + * made the folio dirty and set the private field; * o and then we budgeted for it for the second time at the * very beginning of this function. * - * So what we have to do is to release the page budget we + * So what we have to do is to release the folio budget we * allocated. */ release_new_page_budget(c); - else if (!PageChecked(page)) + else if (!folio_test_checked(folio)) /* - * We are changing a page which already exists on the media. - * This means that changing the page does not make the amount + * We are changing a folio which already exists on the media. + * This means that changing the folio does not make the amount * of indexing information larger, and this part of the budget * which we have already acquired may be released. */ @@ -304,14 +298,14 @@ static int write_begin_slow(struct address_space *mapping, ubifs_release_dirty_inode_budget(c, ui); } - *pagep = page; + *pagep = &folio->page; return 0; } /** * allocate_budget - allocate budget for 'ubifs_write_begin()'. * @c: UBIFS file-system description object - * @page: page to allocate budget for + * @folio: folio to allocate budget for * @ui: UBIFS inode object the page belongs to * @appending: non-zero if the page is appended * @@ -322,15 +316,15 @@ static int write_begin_slow(struct address_space *mapping, * * Returns: %0 in case of success and %-ENOSPC in case of failure. */ -static int allocate_budget(struct ubifs_info *c, struct page *page, +static int allocate_budget(struct ubifs_info *c, struct folio *folio, struct ubifs_inode *ui, int appending) { struct ubifs_budget_req req = { .fast = 1 }; - if (PagePrivate(page)) { + if (folio->private) { if (!appending) /* - * The page is dirty and we are not appending, which + * The folio is dirty and we are not appending, which * means no budget is needed at all. */ return 0; @@ -354,11 +348,11 @@ static int allocate_budget(struct ubifs_info *c, struct page *page, */ req.dirtied_ino = 1; } else { - if (PageChecked(page)) + if (folio_test_checked(folio)) /* * The page corresponds to a hole and does not * exist on the media. So changing it makes - * make the amount of indexing information + * the amount of indexing information * larger, and we have to budget for a new * page. */ @@ -428,7 +422,7 @@ static int ubifs_write_begin(struct file *file, struct address_space *mapping, pgoff_t index = pos >> PAGE_SHIFT; int err, appending = !!(pos + len > inode->i_size); int skipped_read = 0; - struct page *page; + struct folio *folio; ubifs_assert(c, ubifs_inode(inode)->ui_size == inode->i_size); ubifs_assert(c, !c->ro_media && !c->ro_mount); @@ -437,13 +431,14 @@ static int ubifs_write_begin(struct file *file, struct address_space *mapping, return -EROFS; /* Try out the fast-path part first */ - page = grab_cache_page_write_begin(mapping, index); - if (unlikely(!page)) - return -ENOMEM; + folio = __filemap_get_folio(mapping, index, FGP_WRITEBEGIN, + mapping_gfp_mask(mapping)); + if (IS_ERR(folio)) + return PTR_ERR(folio); - if (!PageUptodate(page)) { + if (!folio_test_uptodate(folio)) { /* The page is not loaded from the flash */ - if (!(pos & ~PAGE_MASK) && len == PAGE_SIZE) { + if (pos == folio_pos(folio) && len >= folio_size(folio)) { /* * We change whole page so no need to load it. But we * do not know whether this page exists on the media or @@ -453,32 +448,27 @@ static int ubifs_write_begin(struct file *file, struct address_space *mapping, * media. Thus, we are setting the @PG_checked flag * here. */ - SetPageChecked(page); + folio_set_checked(folio); skipped_read = 1; } else { - err = do_readpage(page); + err = do_readpage(folio); if (err) { - unlock_page(page); - put_page(page); + folio_unlock(folio); + folio_put(folio); return err; } } - - SetPageUptodate(page); - ClearPageError(page); } - err = allocate_budget(c, page, ui, appending); + err = allocate_budget(c, folio, ui, appending); if (unlikely(err)) { ubifs_assert(c, err == -ENOSPC); /* * If we skipped reading the page because we were going to * write all of it, then it is not up to date. */ - if (skipped_read) { - ClearPageChecked(page); - ClearPageUptodate(page); - } + if (skipped_read) + folio_clear_checked(folio); /* * Budgeting failed which means it would have to force * write-back but didn't, because we set the @fast flag in the @@ -490,8 +480,8 @@ static int ubifs_write_begin(struct file *file, struct address_space *mapping, ubifs_assert(c, mutex_is_locked(&ui->ui_mutex)); mutex_unlock(&ui->ui_mutex); } - unlock_page(page); - put_page(page); + folio_unlock(folio); + folio_put(folio); return write_begin_slow(mapping, pos, len, pagep); } @@ -502,22 +492,21 @@ static int ubifs_write_begin(struct file *file, struct address_space *mapping, * with @ui->ui_mutex locked if we are appending pages, and unlocked * otherwise. This is an optimization (slightly hacky though). */ - *pagep = page; + *pagep = &folio->page; return 0; - } /** * cancel_budget - cancel budget. * @c: UBIFS file-system description object - * @page: page to cancel budget for + * @folio: folio to cancel budget for * @ui: UBIFS inode object the page belongs to * @appending: non-zero if the page is appended * * This is a helper function for a page write operation. It unlocks the * @ui->ui_mutex in case of appending. */ -static void cancel_budget(struct ubifs_info *c, struct page *page, +static void cancel_budget(struct ubifs_info *c, struct folio *folio, struct ubifs_inode *ui, int appending) { if (appending) { @@ -525,8 +514,8 @@ static void cancel_budget(struct ubifs_info *c, struct page *page, ubifs_release_dirty_inode_budget(c, ui); mutex_unlock(&ui->ui_mutex); } - if (!PagePrivate(page)) { - if (PageChecked(page)) + if (!folio->private) { + if (folio_test_checked(folio)) release_new_page_budget(c); else release_existing_page_budget(c); @@ -537,6 +526,7 @@ static int ubifs_write_end(struct file *file, struct address_space *mapping, loff_t pos, unsigned len, unsigned copied, struct page *page, void *fsdata) { + struct folio *folio = page_folio(page); struct inode *inode = mapping->host; struct ubifs_inode *ui = ubifs_inode(inode); struct ubifs_info *c = inode->i_sb->s_fs_info; @@ -544,44 +534,47 @@ static int ubifs_write_end(struct file *file, struct address_space *mapping, int appending = !!(end_pos > inode->i_size); dbg_gen("ino %lu, pos %llu, pg %lu, len %u, copied %d, i_size %lld", - inode->i_ino, pos, page->index, len, copied, inode->i_size); + inode->i_ino, pos, folio->index, len, copied, inode->i_size); - if (unlikely(copied < len && len == PAGE_SIZE)) { + if (unlikely(copied < len && !folio_test_uptodate(folio))) { /* - * VFS copied less data to the page that it intended and + * VFS copied less data to the folio than it intended and * declared in its '->write_begin()' call via the @len - * argument. If the page was not up-to-date, and @len was - * @PAGE_SIZE, the 'ubifs_write_begin()' function did + * argument. If the folio was not up-to-date, + * the 'ubifs_write_begin()' function did * not load it from the media (for optimization reasons). This - * means that part of the page contains garbage. So read the - * page now. + * means that part of the folio contains garbage. So read the + * folio now. */ dbg_gen("copied %d instead of %d, read page and repeat", copied, len); - cancel_budget(c, page, ui, appending); - ClearPageChecked(page); + cancel_budget(c, folio, ui, appending); + folio_clear_checked(folio); /* * Return 0 to force VFS to repeat the whole operation, or the * error code if 'do_readpage()' fails. */ - copied = do_readpage(page); + copied = do_readpage(folio); goto out; } - if (!PagePrivate(page)) { - attach_page_private(page, (void *)1); + if (len == folio_size(folio)) + folio_mark_uptodate(folio); + + if (!folio->private) { + folio_attach_private(folio, (void *)1); atomic_long_inc(&c->dirty_pg_cnt); - __set_page_dirty_nobuffers(page); + filemap_dirty_folio(mapping, folio); } if (appending) { i_size_write(inode, end_pos); ui->ui_size = end_pos; /* - * Note, we do not set @I_DIRTY_PAGES (which means that the - * inode has dirty pages), this has been done in - * '__set_page_dirty_nobuffers()'. + * We do not set @I_DIRTY_PAGES (which means that + * the inode has dirty pages), this was done in + * filemap_dirty_folio(). */ __mark_inode_dirty(inode, I_DIRTY_DATASYNC); ubifs_assert(c, mutex_is_locked(&ui->ui_mutex)); @@ -589,43 +582,43 @@ static int ubifs_write_end(struct file *file, struct address_space *mapping, } out: - unlock_page(page); - put_page(page); + folio_unlock(folio); + folio_put(folio); return copied; } /** * populate_page - copy data nodes into a page for bulk-read. * @c: UBIFS file-system description object - * @page: page + * @folio: folio * @bu: bulk-read information * @n: next zbranch slot * * Returns: %0 on success and a negative error code on failure. */ -static int populate_page(struct ubifs_info *c, struct page *page, +static int populate_page(struct ubifs_info *c, struct folio *folio, struct bu_info *bu, int *n) { int i = 0, nn = *n, offs = bu->zbranch[0].offs, hole = 0, read = 0; - struct inode *inode = page->mapping->host; + struct inode *inode = folio->mapping->host; loff_t i_size = i_size_read(inode); unsigned int page_block; void *addr, *zaddr; pgoff_t end_index; dbg_gen("ino %lu, pg %lu, i_size %lld, flags %#lx", - inode->i_ino, page->index, i_size, page->flags); + inode->i_ino, folio->index, i_size, folio->flags); - addr = zaddr = kmap(page); + addr = zaddr = kmap_local_folio(folio, 0); end_index = (i_size - 1) >> PAGE_SHIFT; - if (!i_size || page->index > end_index) { + if (!i_size || folio->index > end_index) { hole = 1; - memset(addr, 0, PAGE_SIZE); + addr = folio_zero_tail(folio, 0, addr); goto out_hole; } - page_block = page->index << UBIFS_BLOCKS_PER_PAGE_SHIFT; + page_block = folio->index << UBIFS_BLOCKS_PER_PAGE_SHIFT; while (1) { int err, len, out_len, dlen; @@ -674,9 +667,13 @@ static int populate_page(struct ubifs_info *c, struct page *page, break; addr += UBIFS_BLOCK_SIZE; page_block += 1; + if (folio_test_highmem(folio) && (offset_in_page(addr) == 0)) { + kunmap_local(addr - UBIFS_BLOCK_SIZE); + addr = kmap_local_folio(folio, i * UBIFS_BLOCK_SIZE); + } } - if (end_index == page->index) { + if (end_index == folio->index) { int len = i_size & (PAGE_SIZE - 1); if (len && len < read) @@ -685,22 +682,19 @@ static int populate_page(struct ubifs_info *c, struct page *page, out_hole: if (hole) { - SetPageChecked(page); + folio_set_checked(folio); dbg_gen("hole"); } - SetPageUptodate(page); - ClearPageError(page); - flush_dcache_page(page); - kunmap(page); + folio_mark_uptodate(folio); + flush_dcache_folio(folio); + kunmap_local(addr); *n = nn; return 0; out_err: - ClearPageUptodate(page); - SetPageError(page); - flush_dcache_page(page); - kunmap(page); + flush_dcache_folio(folio); + kunmap_local(addr); ubifs_err(c, "bad data node (block %u, inode %lu)", page_block, inode->i_ino); return -EINVAL; @@ -710,15 +704,15 @@ out_err: * ubifs_do_bulk_read - do bulk-read. * @c: UBIFS file-system description object * @bu: bulk-read information - * @page1: first page to read + * @folio1: first folio to read * * Returns: %1 if the bulk-read is done, otherwise %0 is returned. */ static int ubifs_do_bulk_read(struct ubifs_info *c, struct bu_info *bu, - struct page *page1) + struct folio *folio1) { - pgoff_t offset = page1->index, end_index; - struct address_space *mapping = page1->mapping; + pgoff_t offset = folio1->index, end_index; + struct address_space *mapping = folio1->mapping; struct inode *inode = mapping->host; struct ubifs_inode *ui = ubifs_inode(inode); int err, page_idx, page_cnt, ret = 0, n = 0; @@ -768,11 +762,11 @@ static int ubifs_do_bulk_read(struct ubifs_info *c, struct bu_info *bu, goto out_warn; } - err = populate_page(c, page1, bu, &n); + err = populate_page(c, folio1, bu, &n); if (err) goto out_warn; - unlock_page(page1); + folio_unlock(folio1); ret = 1; isize = i_size_read(inode); @@ -782,19 +776,19 @@ static int ubifs_do_bulk_read(struct ubifs_info *c, struct bu_info *bu, for (page_idx = 1; page_idx < page_cnt; page_idx++) { pgoff_t page_offset = offset + page_idx; - struct page *page; + struct folio *folio; if (page_offset > end_index) break; - page = pagecache_get_page(mapping, page_offset, + folio = __filemap_get_folio(mapping, page_offset, FGP_LOCK|FGP_ACCESSED|FGP_CREAT|FGP_NOWAIT, ra_gfp_mask); - if (!page) + if (IS_ERR(folio)) break; - if (!PageUptodate(page)) - err = populate_page(c, page, bu, &n); - unlock_page(page); - put_page(page); + if (!folio_test_uptodate(folio)) + err = populate_page(c, folio, bu, &n); + folio_unlock(folio); + folio_put(folio); if (err) break; } @@ -817,7 +811,7 @@ out_bu_off: /** * ubifs_bulk_read - determine whether to bulk-read and, if so, do it. - * @page: page from which to start bulk-read. + * @folio: folio from which to start bulk-read. * * Some flash media are capable of reading sequentially at faster rates. UBIFS * bulk-read facility is designed to take advantage of that, by reading in one @@ -826,12 +820,12 @@ out_bu_off: * * Returns: %1 if a bulk-read is done and %0 otherwise. */ -static int ubifs_bulk_read(struct page *page) +static int ubifs_bulk_read(struct folio *folio) { - struct inode *inode = page->mapping->host; + struct inode *inode = folio->mapping->host; struct ubifs_info *c = inode->i_sb->s_fs_info; struct ubifs_inode *ui = ubifs_inode(inode); - pgoff_t index = page->index, last_page_read = ui->last_page_read; + pgoff_t index = folio->index, last_page_read = ui->last_page_read; struct bu_info *bu; int err = 0, allocated = 0; @@ -879,8 +873,8 @@ static int ubifs_bulk_read(struct page *page) bu->buf_len = c->max_bu_buf_len; data_key_init(c, &bu->key, inode->i_ino, - page->index << UBIFS_BLOCKS_PER_PAGE_SHIFT); - err = ubifs_do_bulk_read(c, bu, page); + folio->index << UBIFS_BLOCKS_PER_PAGE_SHIFT); + err = ubifs_do_bulk_read(c, bu, folio); if (!allocated) mutex_unlock(&c->bu_mutex); @@ -894,69 +888,71 @@ out_unlock: static int ubifs_read_folio(struct file *file, struct folio *folio) { - struct page *page = &folio->page; - - if (ubifs_bulk_read(page)) + if (ubifs_bulk_read(folio)) return 0; - do_readpage(page); + do_readpage(folio); folio_unlock(folio); return 0; } -static int do_writepage(struct page *page, int len) +static int do_writepage(struct folio *folio, size_t len) { - int err = 0, i, blen; + int err = 0, blen; unsigned int block; void *addr; + size_t offset = 0; union ubifs_key key; - struct inode *inode = page->mapping->host; + struct inode *inode = folio->mapping->host; struct ubifs_info *c = inode->i_sb->s_fs_info; #ifdef UBIFS_DEBUG struct ubifs_inode *ui = ubifs_inode(inode); spin_lock(&ui->ui_lock); - ubifs_assert(c, page->index <= ui->synced_i_size >> PAGE_SHIFT); + ubifs_assert(c, folio->index <= ui->synced_i_size >> PAGE_SHIFT); spin_unlock(&ui->ui_lock); #endif - /* Update radix tree tags */ - set_page_writeback(page); + folio_start_writeback(folio); - addr = kmap(page); - block = page->index << UBIFS_BLOCKS_PER_PAGE_SHIFT; - i = 0; - while (len) { - blen = min_t(int, len, UBIFS_BLOCK_SIZE); + addr = kmap_local_folio(folio, offset); + block = folio->index << UBIFS_BLOCKS_PER_PAGE_SHIFT; + for (;;) { + blen = min_t(size_t, len, UBIFS_BLOCK_SIZE); data_key_init(c, &key, inode->i_ino, block); err = ubifs_jnl_write_data(c, inode, &key, addr, blen); if (err) break; - if (++i >= UBIFS_BLOCKS_PER_PAGE) + len -= blen; + if (!len) break; block += 1; addr += blen; - len -= blen; + if (folio_test_highmem(folio) && !offset_in_page(addr)) { + kunmap_local(addr - blen); + offset += PAGE_SIZE; + addr = kmap_local_folio(folio, offset); + } } + kunmap_local(addr); if (err) { - SetPageError(page); - ubifs_err(c, "cannot write page %lu of inode %lu, error %d", - page->index, inode->i_ino, err); + mapping_set_error(folio->mapping, err); + ubifs_err(c, "cannot write folio %lu of inode %lu, error %d", + folio->index, inode->i_ino, err); ubifs_ro_mode(c, err); } - ubifs_assert(c, PagePrivate(page)); - if (PageChecked(page)) + ubifs_assert(c, folio->private != NULL); + if (folio_test_checked(folio)) release_new_page_budget(c); else release_existing_page_budget(c); atomic_long_dec(&c->dirty_pg_cnt); - detach_page_private(page); - ClearPageChecked(page); + folio_detach_private(folio); + folio_clear_checked(folio); - kunmap(page); - unlock_page(page); - end_page_writeback(page); + folio_unlock(folio); + folio_end_writeback(folio); return err; } @@ -1006,22 +1002,21 @@ static int do_writepage(struct page *page, int len) * on the page lock and it would not write the truncated inode node to the * journal before we have finished. */ -static int ubifs_writepage(struct page *page, struct writeback_control *wbc) +static int ubifs_writepage(struct folio *folio, struct writeback_control *wbc, + void *data) { - struct inode *inode = page->mapping->host; + struct inode *inode = folio->mapping->host; struct ubifs_info *c = inode->i_sb->s_fs_info; struct ubifs_inode *ui = ubifs_inode(inode); loff_t i_size = i_size_read(inode), synced_i_size; - pgoff_t end_index = i_size >> PAGE_SHIFT; - int err, len = i_size & (PAGE_SIZE - 1); - void *kaddr; + int err, len = folio_size(folio); dbg_gen("ino %lu, pg %lu, pg flags %#lx", - inode->i_ino, page->index, page->flags); - ubifs_assert(c, PagePrivate(page)); + inode->i_ino, folio->index, folio->flags); + ubifs_assert(c, folio->private != NULL); - /* Is the page fully outside @i_size? (truncate in progress) */ - if (page->index > end_index || (page->index == end_index && !len)) { + /* Is the folio fully outside @i_size? (truncate in progress) */ + if (folio_pos(folio) >= i_size) { err = 0; goto out_unlock; } @@ -1030,9 +1025,9 @@ static int ubifs_writepage(struct page *page, struct writeback_control *wbc) synced_i_size = ui->synced_i_size; spin_unlock(&ui->ui_lock); - /* Is the page fully inside @i_size? */ - if (page->index < end_index) { - if (page->index >= synced_i_size >> PAGE_SHIFT) { + /* Is the folio fully inside i_size? */ + if (folio_pos(folio) + len <= i_size) { + if (folio_pos(folio) >= synced_i_size) { err = inode->i_sb->s_op->write_inode(inode, NULL); if (err) goto out_redirty; @@ -1045,20 +1040,18 @@ static int ubifs_writepage(struct page *page, struct writeback_control *wbc) * with this. */ } - return do_writepage(page, PAGE_SIZE); + return do_writepage(folio, len); } /* - * The page straddles @i_size. It must be zeroed out on each and every + * The folio straddles @i_size. It must be zeroed out on each and every * writepage invocation because it may be mmapped. "A file is mapped * in multiples of the page size. For a file that is not a multiple of * the page size, the remaining memory is zeroed when mapped, and * writes to that region are not written out to the file." */ - kaddr = kmap_atomic(page); - memset(kaddr + len, 0, PAGE_SIZE - len); - flush_dcache_page(page); - kunmap_atomic(kaddr); + len = i_size - folio_pos(folio); + folio_zero_segment(folio, len, folio_size(folio)); if (i_size > synced_i_size) { err = inode->i_sb->s_op->write_inode(inode, NULL); @@ -1066,19 +1059,25 @@ static int ubifs_writepage(struct page *page, struct writeback_control *wbc) goto out_redirty; } - return do_writepage(page, len); + return do_writepage(folio, len); out_redirty: /* - * redirty_page_for_writepage() won't call ubifs_dirty_inode() because + * folio_redirty_for_writepage() won't call ubifs_dirty_inode() because * it passes I_DIRTY_PAGES flag while calling __mark_inode_dirty(), so * there is no need to do space budget for dirty inode. */ - redirty_page_for_writepage(wbc, page); + folio_redirty_for_writepage(wbc, folio); out_unlock: - unlock_page(page); + folio_unlock(folio); return err; } +static int ubifs_writepages(struct address_space *mapping, + struct writeback_control *wbc) +{ + return write_cache_pages(mapping, wbc, ubifs_writepage, NULL); +} + /** * do_attr_changes - change inode attributes. * @inode: inode to change attributes for @@ -1155,11 +1154,11 @@ static int do_truncation(struct ubifs_info *c, struct inode *inode, if (offset) { pgoff_t index = new_size >> PAGE_SHIFT; - struct page *page; + struct folio *folio; - page = find_lock_page(inode->i_mapping, index); - if (page) { - if (PageDirty(page)) { + folio = filemap_lock_folio(inode->i_mapping, index); + if (!IS_ERR(folio)) { + if (folio_test_dirty(folio)) { /* * 'ubifs_jnl_truncate()' will try to truncate * the last data node, but it contains @@ -1168,14 +1167,14 @@ static int do_truncation(struct ubifs_info *c, struct inode *inode, * 'ubifs_jnl_truncate()' will see an already * truncated (and up to date) data node. */ - ubifs_assert(c, PagePrivate(page)); + ubifs_assert(c, folio->private != NULL); - clear_page_dirty_for_io(page); + folio_clear_dirty_for_io(folio); if (UBIFS_BLOCKS_PER_PAGE_SHIFT) - offset = new_size & - (PAGE_SIZE - 1); - err = do_writepage(page, offset); - put_page(page); + offset = offset_in_folio(folio, + new_size); + err = do_writepage(folio, offset); + folio_put(folio); if (err) goto out_budg; /* @@ -1188,8 +1187,8 @@ static int do_truncation(struct ubifs_info *c, struct inode *inode, * to 'ubifs_jnl_truncate()' to save it from * having to read it. */ - unlock_page(page); - put_page(page); + folio_unlock(folio); + folio_put(folio); } } } @@ -1512,14 +1511,14 @@ static bool ubifs_release_folio(struct folio *folio, gfp_t unused_gfp_flags) */ static vm_fault_t ubifs_vm_page_mkwrite(struct vm_fault *vmf) { - struct page *page = vmf->page; + struct folio *folio = page_folio(vmf->page); struct inode *inode = file_inode(vmf->vma->vm_file); struct ubifs_info *c = inode->i_sb->s_fs_info; struct timespec64 now = current_time(inode); struct ubifs_budget_req req = { .new_page = 1 }; int err, update_time; - dbg_gen("ino %lu, pg %lu, i_size %lld", inode->i_ino, page->index, + dbg_gen("ino %lu, pg %lu, i_size %lld", inode->i_ino, folio->index, i_size_read(inode)); ubifs_assert(c, !c->ro_media && !c->ro_mount); @@ -1527,17 +1526,17 @@ static vm_fault_t ubifs_vm_page_mkwrite(struct vm_fault *vmf) return VM_FAULT_SIGBUS; /* -EROFS */ /* - * We have not locked @page so far so we may budget for changing the - * page. Note, we cannot do this after we locked the page, because + * We have not locked @folio so far so we may budget for changing the + * folio. Note, we cannot do this after we locked the folio, because * budgeting may cause write-back which would cause deadlock. * - * At the moment we do not know whether the page is dirty or not, so we - * assume that it is not and budget for a new page. We could look at + * At the moment we do not know whether the folio is dirty or not, so we + * assume that it is not and budget for a new folio. We could look at * the @PG_private flag and figure this out, but we may race with write - * back and the page state may change by the time we lock it, so this + * back and the folio state may change by the time we lock it, so this * would need additional care. We do not bother with this at the * moment, although it might be good idea to do. Instead, we allocate - * budget for a new page and amend it later on if the page was in fact + * budget for a new folio and amend it later on if the folio was in fact * dirty. * * The budgeting-related logic of this function is similar to what we @@ -1560,21 +1559,21 @@ static vm_fault_t ubifs_vm_page_mkwrite(struct vm_fault *vmf) return VM_FAULT_SIGBUS; } - lock_page(page); - if (unlikely(page->mapping != inode->i_mapping || - page_offset(page) > i_size_read(inode))) { - /* Page got truncated out from underneath us */ + folio_lock(folio); + if (unlikely(folio->mapping != inode->i_mapping || + folio_pos(folio) >= i_size_read(inode))) { + /* Folio got truncated out from underneath us */ goto sigbus; } - if (PagePrivate(page)) + if (folio->private) release_new_page_budget(c); else { - if (!PageChecked(page)) + if (!folio_test_checked(folio)) ubifs_convert_page_budget(c); - attach_page_private(page, (void *)1); + folio_attach_private(folio, (void *)1); atomic_long_inc(&c->dirty_pg_cnt); - __set_page_dirty_nobuffers(page); + filemap_dirty_folio(folio->mapping, folio); } if (update_time) { @@ -1590,11 +1589,11 @@ static vm_fault_t ubifs_vm_page_mkwrite(struct vm_fault *vmf) ubifs_release_dirty_inode_budget(c, ui); } - wait_for_stable_page(page); + folio_wait_stable(folio); return VM_FAULT_LOCKED; sigbus: - unlock_page(page); + folio_unlock(folio); ubifs_release_budget(c, &req); return VM_FAULT_SIGBUS; } @@ -1648,7 +1647,7 @@ static int ubifs_symlink_getattr(struct mnt_idmap *idmap, const struct address_space_operations ubifs_file_address_operations = { .read_folio = ubifs_read_folio, - .writepage = ubifs_writepage, + .writepages = ubifs_writepages, .write_begin = ubifs_write_begin, .write_end = ubifs_write_end, .invalidate_folio = ubifs_invalidate_folio, diff --git a/fs/ubifs/find.c b/fs/ubifs/find.c index 873e6e1c92b5..6ebf3c04ac5f 100644 --- a/fs/ubifs/find.c +++ b/fs/ubifs/find.c @@ -82,8 +82,9 @@ static int valuable(struct ubifs_info *c, const struct ubifs_lprops *lprops) */ static int scan_for_dirty_cb(struct ubifs_info *c, const struct ubifs_lprops *lprops, int in_tree, - struct scan_data *data) + void *arg) { + struct scan_data *data = arg; int ret = LPT_SCAN_CONTINUE; /* Exclude LEBs that are currently in use */ @@ -166,8 +167,7 @@ static const struct ubifs_lprops *scan_for_dirty(struct ubifs_info *c, data.pick_free = pick_free; data.lnum = -1; data.exclude_index = exclude_index; - err = ubifs_lpt_scan_nolock(c, -1, c->lscan_lnum, - (ubifs_lpt_scan_callback)scan_for_dirty_cb, + err = ubifs_lpt_scan_nolock(c, -1, c->lscan_lnum, scan_for_dirty_cb, &data); if (err) return ERR_PTR(err); @@ -349,8 +349,9 @@ out: */ static int scan_for_free_cb(struct ubifs_info *c, const struct ubifs_lprops *lprops, int in_tree, - struct scan_data *data) + void *arg) { + struct scan_data *data = arg; int ret = LPT_SCAN_CONTINUE; /* Exclude LEBs that are currently in use */ @@ -446,7 +447,7 @@ const struct ubifs_lprops *do_find_free_space(struct ubifs_info *c, data.pick_free = pick_free; data.lnum = -1; err = ubifs_lpt_scan_nolock(c, -1, c->lscan_lnum, - (ubifs_lpt_scan_callback)scan_for_free_cb, + scan_for_free_cb, &data); if (err) return ERR_PTR(err); @@ -589,8 +590,9 @@ out: */ static int scan_for_idx_cb(struct ubifs_info *c, const struct ubifs_lprops *lprops, int in_tree, - struct scan_data *data) + void *arg) { + struct scan_data *data = arg; int ret = LPT_SCAN_CONTINUE; /* Exclude LEBs that are currently in use */ @@ -625,8 +627,7 @@ static const struct ubifs_lprops *scan_for_leb_for_idx(struct ubifs_info *c) int err; data.lnum = -1; - err = ubifs_lpt_scan_nolock(c, -1, c->lscan_lnum, - (ubifs_lpt_scan_callback)scan_for_idx_cb, + err = ubifs_lpt_scan_nolock(c, -1, c->lscan_lnum, scan_for_idx_cb, &data); if (err) return ERR_PTR(err); @@ -726,11 +727,10 @@ out: return err; } -static int cmp_dirty_idx(const struct ubifs_lprops **a, - const struct ubifs_lprops **b) +static int cmp_dirty_idx(const void *a, const void *b) { - const struct ubifs_lprops *lpa = *a; - const struct ubifs_lprops *lpb = *b; + const struct ubifs_lprops *lpa = *(const struct ubifs_lprops **)a; + const struct ubifs_lprops *lpb = *(const struct ubifs_lprops **)b; return lpa->dirty + lpa->free - lpb->dirty - lpb->free; } @@ -754,7 +754,7 @@ int ubifs_save_dirty_idx_lnums(struct ubifs_info *c) sizeof(void *) * c->dirty_idx.cnt); /* Sort it so that the dirtiest is now at the end */ sort(c->dirty_idx.arr, c->dirty_idx.cnt, sizeof(void *), - (int (*)(const void *, const void *))cmp_dirty_idx, NULL); + cmp_dirty_idx, NULL); dbg_find("found %d dirty index LEBs", c->dirty_idx.cnt); if (c->dirty_idx.cnt) dbg_find("dirtiest index LEB is %d with dirty %d and free %d", @@ -782,8 +782,9 @@ int ubifs_save_dirty_idx_lnums(struct ubifs_info *c) */ static int scan_dirty_idx_cb(struct ubifs_info *c, const struct ubifs_lprops *lprops, int in_tree, - struct scan_data *data) + void *arg) { + struct scan_data *data = arg; int ret = LPT_SCAN_CONTINUE; /* Exclude LEBs that are currently in use */ @@ -842,8 +843,7 @@ static int find_dirty_idx_leb(struct ubifs_info *c) if (c->pnodes_have >= c->pnode_cnt) /* All pnodes are in memory, so skip scan */ return -ENOSPC; - err = ubifs_lpt_scan_nolock(c, -1, c->lscan_lnum, - (ubifs_lpt_scan_callback)scan_dirty_idx_cb, + err = ubifs_lpt_scan_nolock(c, -1, c->lscan_lnum, scan_dirty_idx_cb, &data); if (err) return err; diff --git a/fs/ubifs/journal.c b/fs/ubifs/journal.c index f0a5538c84b0..74aee92433d7 100644 --- a/fs/ubifs/journal.c +++ b/fs/ubifs/journal.c @@ -293,6 +293,96 @@ static int write_head(struct ubifs_info *c, int jhead, void *buf, int len, } /** + * __queue_and_wait - queue a task and wait until the task is waked up. + * @c: UBIFS file-system description object + * + * This function adds current task in queue and waits until the task is waked + * up. This function should be called with @c->reserve_space_wq locked. + */ +static void __queue_and_wait(struct ubifs_info *c) +{ + DEFINE_WAIT(wait); + + __add_wait_queue_entry_tail_exclusive(&c->reserve_space_wq, &wait); + set_current_state(TASK_UNINTERRUPTIBLE); + spin_unlock(&c->reserve_space_wq.lock); + + schedule(); + finish_wait(&c->reserve_space_wq, &wait); +} + +/** + * wait_for_reservation - try queuing current task to wait until waked up. + * @c: UBIFS file-system description object + * + * This function queues current task to wait until waked up, if queuing is + * started(@c->need_wait_space is not %0). Returns %true if current task is + * added in queue, otherwise %false is returned. + */ +static bool wait_for_reservation(struct ubifs_info *c) +{ + if (likely(atomic_read(&c->need_wait_space) == 0)) + /* Quick path to check whether queuing is started. */ + return false; + + spin_lock(&c->reserve_space_wq.lock); + if (atomic_read(&c->need_wait_space) == 0) { + /* Queuing is not started, don't queue current task. */ + spin_unlock(&c->reserve_space_wq.lock); + return false; + } + + __queue_and_wait(c); + return true; +} + +/** + * wake_up_reservation - wake up first task in queue or stop queuing. + * @c: UBIFS file-system description object + * + * This function wakes up the first task in queue if it exists, or stops + * queuing if no tasks in queue. + */ +static void wake_up_reservation(struct ubifs_info *c) +{ + spin_lock(&c->reserve_space_wq.lock); + if (waitqueue_active(&c->reserve_space_wq)) + wake_up_locked(&c->reserve_space_wq); + else + /* + * Compared with wait_for_reservation(), set @c->need_wait_space + * under the protection of wait queue lock, which can avoid that + * @c->need_wait_space is set to 0 after new task queued. + */ + atomic_set(&c->need_wait_space, 0); + spin_unlock(&c->reserve_space_wq.lock); +} + +/** + * wake_up_reservation - add current task in queue or start queuing. + * @c: UBIFS file-system description object + * + * This function starts queuing if queuing is not started, otherwise adds + * current task in queue. + */ +static void add_or_start_queue(struct ubifs_info *c) +{ + spin_lock(&c->reserve_space_wq.lock); + if (atomic_cmpxchg(&c->need_wait_space, 0, 1) == 0) { + /* Starts queuing, task can go on directly. */ + spin_unlock(&c->reserve_space_wq.lock); + return; + } + + /* + * There are at least two tasks have retried more than 32 times + * at certain point, first task has started queuing, just queue + * the left tasks. + */ + __queue_and_wait(c); +} + +/** * make_reservation - reserve journal space. * @c: UBIFS file-system description object * @jhead: journal head @@ -311,33 +401,27 @@ static int write_head(struct ubifs_info *c, int jhead, void *buf, int len, static int make_reservation(struct ubifs_info *c, int jhead, int len) { int err, cmt_retries = 0, nospc_retries = 0; + bool blocked = wait_for_reservation(c); again: down_read(&c->commit_sem); err = reserve_space(c, jhead, len); - if (!err) + if (!err) { /* c->commit_sem will get released via finish_reservation(). */ - return 0; + goto out_wake_up; + } up_read(&c->commit_sem); if (err == -ENOSPC) { /* * GC could not make any progress. We should try to commit - * once because it could make some dirty space and GC would - * make progress, so make the error -EAGAIN so that the below + * because it could make some dirty space and GC would make + * progress, so make the error -EAGAIN so that the below * will commit and re-try. */ - if (nospc_retries++ < 2) { - dbg_jnl("no space, retry"); - err = -EAGAIN; - } - - /* - * This means that the budgeting is incorrect. We always have - * to be able to write to the media, because all operations are - * budgeted. Deletions are not budgeted, though, but we reserve - * an extra LEB for them. - */ + nospc_retries++; + dbg_jnl("no space, retry"); + err = -EAGAIN; } if (err != -EAGAIN) @@ -349,15 +433,37 @@ again: */ if (cmt_retries > 128) { /* - * This should not happen unless the journal size limitations - * are too tough. + * This should not happen unless: + * 1. The journal size limitations are too tough. + * 2. The budgeting is incorrect. We always have to be able to + * write to the media, because all operations are budgeted. + * Deletions are not budgeted, though, but we reserve an + * extra LEB for them. */ - ubifs_err(c, "stuck in space allocation"); + ubifs_err(c, "stuck in space allocation, nospc_retries %d", + nospc_retries); err = -ENOSPC; goto out; - } else if (cmt_retries > 32) - ubifs_warn(c, "too many space allocation re-tries (%d)", - cmt_retries); + } else if (cmt_retries > 32) { + /* + * It's almost impossible to happen, unless there are many tasks + * making reservation concurrently and someone task has retried + * gc + commit for many times, generated available space during + * this period are grabbed by other tasks. + * But if it happens, start queuing up all tasks that will make + * space reservation, then there is only one task making space + * reservation at any time, and it can always make success under + * the premise of correct budgeting. + */ + ubifs_warn(c, "too many space allocation cmt_retries (%d) " + "nospc_retries (%d), start queuing tasks", + cmt_retries, nospc_retries); + + if (!blocked) { + blocked = true; + add_or_start_queue(c); + } + } dbg_jnl("-EAGAIN, commit and retry (retried %d times)", cmt_retries); @@ -365,7 +471,7 @@ again: err = ubifs_run_commit(c); if (err) - return err; + goto out_wake_up; goto again; out: @@ -380,6 +486,27 @@ out: cmt_retries = dbg_check_lprops(c); up_write(&c->commit_sem); } +out_wake_up: + if (blocked) { + /* + * Only tasks that have ever started queuing or ever been queued + * can wake up other queued tasks, which can make sure that + * there is only one task waked up to make space reservation. + * For example: + * task A task B task C + * make_reservation make_reservation + * reserve_space // 0 + * wake_up_reservation + * atomic_cmpxchg // 0, start queuing + * reserve_space + * wait_for_reservation + * __queue_and_wait + * add_wait_queue + * if (blocked) // false + * // So that task C won't be waked up to race with task B + */ + wake_up_reservation(c); + } return err; } diff --git a/fs/ubifs/lprops.c b/fs/ubifs/lprops.c index 6d6cd85c2b4c..a11c3dab7e16 100644 --- a/fs/ubifs/lprops.c +++ b/fs/ubifs/lprops.c @@ -1014,8 +1014,9 @@ out: */ static int scan_check_cb(struct ubifs_info *c, const struct ubifs_lprops *lp, int in_tree, - struct ubifs_lp_stats *lst) + void *arg) { + struct ubifs_lp_stats *lst = arg; struct ubifs_scan_leb *sleb; struct ubifs_scan_node *snod; int cat, lnum = lp->lnum, is_idx = 0, used = 0, free, dirty, ret; @@ -1269,8 +1270,7 @@ int dbg_check_lprops(struct ubifs_info *c) memset(&lst, 0, sizeof(struct ubifs_lp_stats)); err = ubifs_lpt_scan_nolock(c, c->main_first, c->leb_cnt - 1, - (ubifs_lpt_scan_callback)scan_check_cb, - &lst); + scan_check_cb, &lst); if (err && err != -ENOSPC) goto out; diff --git a/fs/ubifs/lpt_commit.c b/fs/ubifs/lpt_commit.c index c4d079328b92..07351fdce722 100644 --- a/fs/ubifs/lpt_commit.c +++ b/fs/ubifs/lpt_commit.c @@ -1646,7 +1646,6 @@ static int dbg_check_ltab_lnum(struct ubifs_info *c, int lnum) len -= node_len; } - err = 0; out: vfree(buf); return err; diff --git a/fs/ubifs/super.c b/fs/ubifs/super.c index 7f4031a15f4d..291583005dd1 100644 --- a/fs/ubifs/super.c +++ b/fs/ubifs/super.c @@ -2151,6 +2151,8 @@ static struct ubifs_info *alloc_ubifs_info(struct ubi_volume_desc *ubi) mutex_init(&c->bu_mutex); mutex_init(&c->write_reserve_mutex); init_waitqueue_head(&c->cmt_wq); + init_waitqueue_head(&c->reserve_space_wq); + atomic_set(&c->need_wait_space, 0); c->buds = RB_ROOT; c->old_idx = RB_ROOT; c->size_tree = RB_ROOT; diff --git a/fs/ubifs/tnc.c b/fs/ubifs/tnc.c index f4728e65d1bd..45cacdcd4746 100644 --- a/fs/ubifs/tnc.c +++ b/fs/ubifs/tnc.c @@ -3116,14 +3116,7 @@ static void tnc_destroy_cnext(struct ubifs_info *c) void ubifs_tnc_close(struct ubifs_info *c) { tnc_destroy_cnext(c); - if (c->zroot.znode) { - long n, freed; - - n = atomic_long_read(&c->clean_zn_cnt); - freed = ubifs_destroy_tnc_subtree(c, c->zroot.znode); - ubifs_assert(c, freed == n); - atomic_long_sub(n, &ubifs_clean_zn_cnt); - } + ubifs_destroy_tnc_tree(c); kfree(c->gap_lebs); kfree(c->ilebs); destroy_old_idx(c); diff --git a/fs/ubifs/tnc_misc.c b/fs/ubifs/tnc_misc.c index 4d686e34e64d..d3f8a6aa1f49 100644 --- a/fs/ubifs/tnc_misc.c +++ b/fs/ubifs/tnc_misc.c @@ -251,6 +251,28 @@ long ubifs_destroy_tnc_subtree(const struct ubifs_info *c, } /** + * ubifs_destroy_tnc_tree - destroy all znodes connected to the TNC tree. + * @c: UBIFS file-system description object + * + * This function destroys the whole TNC tree and updates clean global znode + * count. + */ +void ubifs_destroy_tnc_tree(struct ubifs_info *c) +{ + long n, freed; + + if (!c->zroot.znode) + return; + + n = atomic_long_read(&c->clean_zn_cnt); + freed = ubifs_destroy_tnc_subtree(c, c->zroot.znode); + ubifs_assert(c, freed == n); + atomic_long_sub(n, &ubifs_clean_zn_cnt); + + c->zroot.znode = NULL; +} + +/** * read_znode - read an indexing node from flash and fill znode. * @c: UBIFS file-system description object * @zzbr: the zbranch describing the node to read diff --git a/fs/ubifs/ubifs.h b/fs/ubifs/ubifs.h index 3916dc4f30ca..1f3ea879d93a 100644 --- a/fs/ubifs/ubifs.h +++ b/fs/ubifs/ubifs.h @@ -1047,6 +1047,8 @@ struct ubifs_debug_info; * @bg_bud_bytes: number of bud bytes when background commit is initiated * @old_buds: buds to be released after commit ends * @max_bud_cnt: maximum number of buds + * @need_wait_space: Non %0 means space reservation tasks need to wait in queue + * @reserve_space_wq: wait queue to sleep on if @need_wait_space is not %0 * * @commit_sem: synchronizes committer with other processes * @cmt_state: commit state @@ -1305,6 +1307,8 @@ struct ubifs_info { long long bg_bud_bytes; struct list_head old_buds; int max_bud_cnt; + atomic_t need_wait_space; + wait_queue_head_t reserve_space_wq; struct rw_semaphore commit_sem; int cmt_state; @@ -1903,6 +1907,7 @@ struct ubifs_znode *ubifs_tnc_postorder_next(const struct ubifs_info *c, struct ubifs_znode *znode); long ubifs_destroy_tnc_subtree(const struct ubifs_info *c, struct ubifs_znode *zr); +void ubifs_destroy_tnc_tree(struct ubifs_info *c); struct ubifs_znode *ubifs_load_znode(struct ubifs_info *c, struct ubifs_zbranch *zbr, struct ubifs_znode *parent, int iip); diff --git a/fs/udf/dir.c b/fs/udf/dir.c index f6533f93851b..f94f45fe2c91 100644 --- a/fs/udf/dir.c +++ b/fs/udf/dir.c @@ -67,7 +67,7 @@ static int udf_readdir(struct file *file, struct dir_context *ctx) pos_valid = true; } - fname = kmalloc(UDF_NAME_LEN, GFP_NOFS); + fname = kmalloc(UDF_NAME_LEN, GFP_KERNEL); if (!fname) { ret = -ENOMEM; goto out; diff --git a/fs/udf/inode.c b/fs/udf/inode.c index d8493449d4c5..2f831a3a91af 100644 --- a/fs/udf/inode.c +++ b/fs/udf/inode.c @@ -357,7 +357,7 @@ int udf_expand_file_adinicb(struct inode *inode) return 0; } - page = find_or_create_page(inode->i_mapping, 0, GFP_NOFS); + page = find_or_create_page(inode->i_mapping, 0, GFP_KERNEL); if (!page) return -ENOMEM; diff --git a/fs/udf/namei.c b/fs/udf/namei.c index 1bb6ed948927..1308109fd42d 100644 --- a/fs/udf/namei.c +++ b/fs/udf/namei.c @@ -59,7 +59,7 @@ static int udf_fiiter_find_entry(struct inode *dir, const struct qstr *child, child->name[0] == '.' && child->name[1] == '.'; int ret; - fname = kmalloc(UDF_NAME_LEN, GFP_NOFS); + fname = kmalloc(UDF_NAME_LEN, GFP_KERNEL); if (!fname) return -ENOMEM; @@ -566,7 +566,7 @@ out: static int udf_symlink(struct mnt_idmap *idmap, struct inode *dir, struct dentry *dentry, const char *symname) { - struct inode *inode = udf_new_inode(dir, S_IFLNK | 0777); + struct inode *inode; struct pathComponent *pc; const char *compstart; struct extent_position epos = {}; @@ -579,17 +579,20 @@ static int udf_symlink(struct mnt_idmap *idmap, struct inode *dir, struct udf_inode_info *iinfo; struct super_block *sb = dir->i_sb; - if (IS_ERR(inode)) - return PTR_ERR(inode); - - iinfo = UDF_I(inode); - down_write(&iinfo->i_data_sem); - name = kmalloc(UDF_NAME_LEN_CS0, GFP_NOFS); + name = kmalloc(UDF_NAME_LEN_CS0, GFP_KERNEL); if (!name) { err = -ENOMEM; - goto out_no_entry; + goto out; + } + + inode = udf_new_inode(dir, S_IFLNK | 0777); + if (IS_ERR(inode)) { + err = PTR_ERR(inode); + goto out; } + iinfo = UDF_I(inode); + down_write(&iinfo->i_data_sem); inode->i_data.a_ops = &udf_symlink_aops; inode->i_op = &udf_symlink_inode_operations; inode_nohighmem(inode); diff --git a/fs/udf/super.c b/fs/udf/super.c index 6f420f4ca005..2217f7ed7a49 100644 --- a/fs/udf/super.c +++ b/fs/udf/super.c @@ -40,20 +40,20 @@ #include <linux/slab.h> #include <linux/kernel.h> #include <linux/module.h> -#include <linux/parser.h> #include <linux/stat.h> #include <linux/cdrom.h> #include <linux/nls.h> #include <linux/vfs.h> #include <linux/vmalloc.h> #include <linux/errno.h> -#include <linux/mount.h> #include <linux/seq_file.h> #include <linux/bitmap.h> #include <linux/crc-itu-t.h> #include <linux/log2.h> #include <asm/byteorder.h> #include <linux/iversion.h> +#include <linux/fs_context.h> +#include <linux/fs_parser.h> #include "udf_sb.h" #include "udf_i.h" @@ -91,16 +91,20 @@ enum { UDF_MAX_LINKS = 0xffff }; #define UDF_MAX_FILESIZE (1ULL << 42) /* These are the "meat" - everything else is stuffing */ -static int udf_fill_super(struct super_block *, void *, int); +static int udf_fill_super(struct super_block *sb, struct fs_context *fc); static void udf_put_super(struct super_block *); static int udf_sync_fs(struct super_block *, int); -static int udf_remount_fs(struct super_block *, int *, char *); static void udf_load_logicalvolint(struct super_block *, struct kernel_extent_ad); static void udf_open_lvid(struct super_block *); static void udf_close_lvid(struct super_block *); static unsigned int udf_count_free(struct super_block *); static int udf_statfs(struct dentry *, struct kstatfs *); static int udf_show_options(struct seq_file *, struct dentry *); +static int udf_init_fs_context(struct fs_context *fc); +static int udf_parse_param(struct fs_context *fc, struct fs_parameter *param); +static int udf_reconfigure(struct fs_context *fc); +static void udf_free_fc(struct fs_context *fc); +static const struct fs_parameter_spec udf_param_spec[]; struct logicalVolIntegrityDescImpUse *udf_sb_lvidiu(struct super_block *sb) { @@ -119,18 +123,25 @@ struct logicalVolIntegrityDescImpUse *udf_sb_lvidiu(struct super_block *sb) } /* UDF filesystem type */ -static struct dentry *udf_mount(struct file_system_type *fs_type, - int flags, const char *dev_name, void *data) +static int udf_get_tree(struct fs_context *fc) { - return mount_bdev(fs_type, flags, dev_name, data, udf_fill_super); + return get_tree_bdev(fc, udf_fill_super); } +static const struct fs_context_operations udf_context_ops = { + .parse_param = udf_parse_param, + .get_tree = udf_get_tree, + .reconfigure = udf_reconfigure, + .free = udf_free_fc, +}; + static struct file_system_type udf_fstype = { .owner = THIS_MODULE, .name = "udf", - .mount = udf_mount, .kill_sb = kill_block_super, .fs_flags = FS_REQUIRES_DEV, + .init_fs_context = udf_init_fs_context, + .parameters = udf_param_spec, }; MODULE_ALIAS_FS("udf"); @@ -203,12 +214,10 @@ static const struct super_operations udf_sb_ops = { .put_super = udf_put_super, .sync_fs = udf_sync_fs, .statfs = udf_statfs, - .remount_fs = udf_remount_fs, .show_options = udf_show_options, }; struct udf_options { - unsigned char novrs; unsigned int blocksize; unsigned int session; unsigned int lastblock; @@ -222,6 +231,65 @@ struct udf_options { struct nls_table *nls_map; }; +/* + * UDF has historically preserved prior mount options across + * a remount, so copy those here if remounting, otherwise set + * initial mount defaults. + */ +static void udf_init_options(struct fs_context *fc, struct udf_options *uopt) +{ + if (fc->purpose == FS_CONTEXT_FOR_RECONFIGURE) { + struct super_block *sb = fc->root->d_sb; + struct udf_sb_info *sbi = UDF_SB(sb); + + uopt->flags = sbi->s_flags; + uopt->uid = sbi->s_uid; + uopt->gid = sbi->s_gid; + uopt->umask = sbi->s_umask; + uopt->fmode = sbi->s_fmode; + uopt->dmode = sbi->s_dmode; + uopt->nls_map = NULL; + } else { + uopt->flags = (1 << UDF_FLAG_USE_AD_IN_ICB) | + (1 << UDF_FLAG_STRICT); + /* + * By default we'll use overflow[ug]id when UDF + * inode [ug]id == -1 + */ + uopt->uid = make_kuid(current_user_ns(), overflowuid); + uopt->gid = make_kgid(current_user_ns(), overflowgid); + uopt->umask = 0; + uopt->fmode = UDF_INVALID_MODE; + uopt->dmode = UDF_INVALID_MODE; + uopt->nls_map = NULL; + uopt->session = 0xFFFFFFFF; + } +} + +static int udf_init_fs_context(struct fs_context *fc) +{ + struct udf_options *uopt; + + uopt = kzalloc(sizeof(*uopt), GFP_KERNEL); + if (!uopt) + return -ENOMEM; + + udf_init_options(fc, uopt); + + fc->fs_private = uopt; + fc->ops = &udf_context_ops; + + return 0; +} + +static void udf_free_fc(struct fs_context *fc) +{ + struct udf_options *uopt = fc->fs_private; + + unload_nls(uopt->nls_map); + kfree(fc->fs_private); +} + static int __init init_udf_fs(void) { int err; @@ -357,7 +425,7 @@ static int udf_show_options(struct seq_file *seq, struct dentry *root) } /* - * udf_parse_options + * udf_parse_param * * PURPOSE * Parse mount options. @@ -400,12 +468,12 @@ static int udf_show_options(struct seq_file *seq, struct dentry *root) * yield highly unpredictable results. * * PRE-CONDITIONS - * options Pointer to mount options string. - * uopts Pointer to mount options variable. + * fc fs_context with pointer to mount options variable. + * param Pointer to fs_parameter being parsed. * * POST-CONDITIONS - * <return> 1 Mount options parsed okay. - * <return> 0 Error parsing mount options. + * <return> 0 Mount options parsed okay. + * <return> errno Error parsing mount options. * * HISTORY * July 1, 1997 - Andrew E. Mileski @@ -417,229 +485,193 @@ enum { Opt_noadinicb, Opt_adinicb, Opt_shortad, Opt_longad, Opt_gid, Opt_uid, Opt_umask, Opt_session, Opt_lastblock, Opt_anchor, Opt_volume, Opt_partition, Opt_fileset, - Opt_rootdir, Opt_utf8, Opt_iocharset, - Opt_err, Opt_uforget, Opt_uignore, Opt_gforget, Opt_gignore, - Opt_fmode, Opt_dmode -}; - -static const match_table_t tokens = { - {Opt_novrs, "novrs"}, - {Opt_nostrict, "nostrict"}, - {Opt_bs, "bs=%u"}, - {Opt_unhide, "unhide"}, - {Opt_undelete, "undelete"}, - {Opt_noadinicb, "noadinicb"}, - {Opt_adinicb, "adinicb"}, - {Opt_shortad, "shortad"}, - {Opt_longad, "longad"}, - {Opt_uforget, "uid=forget"}, - {Opt_uignore, "uid=ignore"}, - {Opt_gforget, "gid=forget"}, - {Opt_gignore, "gid=ignore"}, - {Opt_gid, "gid=%u"}, - {Opt_uid, "uid=%u"}, - {Opt_umask, "umask=%o"}, - {Opt_session, "session=%u"}, - {Opt_lastblock, "lastblock=%u"}, - {Opt_anchor, "anchor=%u"}, - {Opt_volume, "volume=%u"}, - {Opt_partition, "partition=%u"}, - {Opt_fileset, "fileset=%u"}, - {Opt_rootdir, "rootdir=%u"}, - {Opt_utf8, "utf8"}, - {Opt_iocharset, "iocharset=%s"}, - {Opt_fmode, "mode=%o"}, - {Opt_dmode, "dmode=%o"}, - {Opt_err, NULL} + Opt_rootdir, Opt_utf8, Opt_iocharset, Opt_err, Opt_fmode, Opt_dmode }; -static int udf_parse_options(char *options, struct udf_options *uopt, - bool remount) +static const struct fs_parameter_spec udf_param_spec[] = { + fsparam_flag ("novrs", Opt_novrs), + fsparam_flag ("nostrict", Opt_nostrict), + fsparam_u32 ("bs", Opt_bs), + fsparam_flag ("unhide", Opt_unhide), + fsparam_flag ("undelete", Opt_undelete), + fsparam_flag_no ("adinicb", Opt_adinicb), + fsparam_flag ("shortad", Opt_shortad), + fsparam_flag ("longad", Opt_longad), + fsparam_string ("gid", Opt_gid), + fsparam_string ("uid", Opt_uid), + fsparam_u32 ("umask", Opt_umask), + fsparam_u32 ("session", Opt_session), + fsparam_u32 ("lastblock", Opt_lastblock), + fsparam_u32 ("anchor", Opt_anchor), + fsparam_u32 ("volume", Opt_volume), + fsparam_u32 ("partition", Opt_partition), + fsparam_u32 ("fileset", Opt_fileset), + fsparam_u32 ("rootdir", Opt_rootdir), + fsparam_flag ("utf8", Opt_utf8), + fsparam_string ("iocharset", Opt_iocharset), + fsparam_u32 ("mode", Opt_fmode), + fsparam_u32 ("dmode", Opt_dmode), + {} + }; + +static int udf_parse_param(struct fs_context *fc, struct fs_parameter *param) { - char *p; - int option; unsigned int uv; - - uopt->novrs = 0; - uopt->session = 0xFFFFFFFF; - uopt->lastblock = 0; - uopt->anchor = 0; - - if (!options) - return 1; - - while ((p = strsep(&options, ",")) != NULL) { - substring_t args[MAX_OPT_ARGS]; - int token; - unsigned n; - if (!*p) - continue; - - token = match_token(p, tokens, args); - switch (token) { - case Opt_novrs: - uopt->novrs = 1; - break; - case Opt_bs: - if (match_int(&args[0], &option)) - return 0; - n = option; - if (n != 512 && n != 1024 && n != 2048 && n != 4096) - return 0; - uopt->blocksize = n; - uopt->flags |= (1 << UDF_FLAG_BLOCKSIZE_SET); - break; - case Opt_unhide: - uopt->flags |= (1 << UDF_FLAG_UNHIDE); - break; - case Opt_undelete: - uopt->flags |= (1 << UDF_FLAG_UNDELETE); - break; - case Opt_noadinicb: + unsigned int n; + struct udf_options *uopt = fc->fs_private; + struct fs_parse_result result; + int token; + bool remount = (fc->purpose & FS_CONTEXT_FOR_RECONFIGURE); + + token = fs_parse(fc, udf_param_spec, param, &result); + if (token < 0) + return token; + + switch (token) { + case Opt_novrs: + uopt->flags |= (1 << UDF_FLAG_NOVRS); + break; + case Opt_bs: + n = result.uint_32; + if (n != 512 && n != 1024 && n != 2048 && n != 4096) + return -EINVAL; + uopt->blocksize = n; + uopt->flags |= (1 << UDF_FLAG_BLOCKSIZE_SET); + break; + case Opt_unhide: + uopt->flags |= (1 << UDF_FLAG_UNHIDE); + break; + case Opt_undelete: + uopt->flags |= (1 << UDF_FLAG_UNDELETE); + break; + case Opt_adinicb: + if (result.negated) uopt->flags &= ~(1 << UDF_FLAG_USE_AD_IN_ICB); - break; - case Opt_adinicb: + else uopt->flags |= (1 << UDF_FLAG_USE_AD_IN_ICB); - break; - case Opt_shortad: - uopt->flags |= (1 << UDF_FLAG_USE_SHORT_AD); - break; - case Opt_longad: - uopt->flags &= ~(1 << UDF_FLAG_USE_SHORT_AD); - break; - case Opt_gid: - if (match_uint(args, &uv)) - return 0; - uopt->gid = make_kgid(current_user_ns(), uv); - if (!gid_valid(uopt->gid)) - return 0; + break; + case Opt_shortad: + uopt->flags |= (1 << UDF_FLAG_USE_SHORT_AD); + break; + case Opt_longad: + uopt->flags &= ~(1 << UDF_FLAG_USE_SHORT_AD); + break; + case Opt_gid: + if (kstrtoint(param->string, 10, &uv) == 0) { + kgid_t gid = make_kgid(current_user_ns(), uv); + if (!gid_valid(gid)) + return -EINVAL; + uopt->gid = gid; uopt->flags |= (1 << UDF_FLAG_GID_SET); - break; - case Opt_uid: - if (match_uint(args, &uv)) - return 0; - uopt->uid = make_kuid(current_user_ns(), uv); - if (!uid_valid(uopt->uid)) - return 0; + } else if (!strcmp(param->string, "forget")) { + uopt->flags |= (1 << UDF_FLAG_GID_FORGET); + } else if (!strcmp(param->string, "ignore")) { + /* this option is superseded by gid=<number> */ + ; + } else { + return -EINVAL; + } + break; + case Opt_uid: + if (kstrtoint(param->string, 10, &uv) == 0) { + kuid_t uid = make_kuid(current_user_ns(), uv); + if (!uid_valid(uid)) + return -EINVAL; + uopt->uid = uid; uopt->flags |= (1 << UDF_FLAG_UID_SET); - break; - case Opt_umask: - if (match_octal(args, &option)) - return 0; - uopt->umask = option; - break; - case Opt_nostrict: - uopt->flags &= ~(1 << UDF_FLAG_STRICT); - break; - case Opt_session: - if (match_int(args, &option)) - return 0; - uopt->session = option; - if (!remount) - uopt->flags |= (1 << UDF_FLAG_SESSION_SET); - break; - case Opt_lastblock: - if (match_int(args, &option)) - return 0; - uopt->lastblock = option; - if (!remount) - uopt->flags |= (1 << UDF_FLAG_LASTBLOCK_SET); - break; - case Opt_anchor: - if (match_int(args, &option)) - return 0; - uopt->anchor = option; - break; - case Opt_volume: - case Opt_partition: - case Opt_fileset: - case Opt_rootdir: - /* Ignored (never implemented properly) */ - break; - case Opt_utf8: - if (!remount) { - unload_nls(uopt->nls_map); - uopt->nls_map = NULL; - } - break; - case Opt_iocharset: - if (!remount) { - unload_nls(uopt->nls_map); - uopt->nls_map = NULL; - } - /* When nls_map is not loaded then UTF-8 is used */ - if (!remount && strcmp(args[0].from, "utf8") != 0) { - uopt->nls_map = load_nls(args[0].from); - if (!uopt->nls_map) { - pr_err("iocharset %s not found\n", - args[0].from); - return 0; - } - } - break; - case Opt_uforget: + } else if (!strcmp(param->string, "forget")) { uopt->flags |= (1 << UDF_FLAG_UID_FORGET); - break; - case Opt_uignore: - case Opt_gignore: - /* These options are superseeded by uid=<number> */ - break; - case Opt_gforget: - uopt->flags |= (1 << UDF_FLAG_GID_FORGET); - break; - case Opt_fmode: - if (match_octal(args, &option)) - return 0; - uopt->fmode = option & 0777; - break; - case Opt_dmode: - if (match_octal(args, &option)) - return 0; - uopt->dmode = option & 0777; - break; - default: - pr_err("bad mount option \"%s\" or missing value\n", p); - return 0; + } else if (!strcmp(param->string, "ignore")) { + /* this option is superseded by uid=<number> */ + ; + } else { + return -EINVAL; + } + break; + case Opt_umask: + uopt->umask = result.uint_32; + break; + case Opt_nostrict: + uopt->flags &= ~(1 << UDF_FLAG_STRICT); + break; + case Opt_session: + uopt->session = result.uint_32; + if (!remount) + uopt->flags |= (1 << UDF_FLAG_SESSION_SET); + break; + case Opt_lastblock: + uopt->lastblock = result.uint_32; + if (!remount) + uopt->flags |= (1 << UDF_FLAG_LASTBLOCK_SET); + break; + case Opt_anchor: + uopt->anchor = result.uint_32; + break; + case Opt_volume: + case Opt_partition: + case Opt_fileset: + case Opt_rootdir: + /* Ignored (never implemented properly) */ + break; + case Opt_utf8: + if (!remount) { + unload_nls(uopt->nls_map); + uopt->nls_map = NULL; } + break; + case Opt_iocharset: + if (!remount) { + unload_nls(uopt->nls_map); + uopt->nls_map = NULL; + } + /* When nls_map is not loaded then UTF-8 is used */ + if (!remount && strcmp(param->string, "utf8") != 0) { + uopt->nls_map = load_nls(param->string); + if (!uopt->nls_map) { + errorf(fc, "iocharset %s not found", + param->string); + return -EINVAL;; + } + } + break; + case Opt_fmode: + uopt->fmode = result.uint_32 & 0777; + break; + case Opt_dmode: + uopt->dmode = result.uint_32 & 0777; + break; + default: + return -EINVAL; } - return 1; + return 0; } -static int udf_remount_fs(struct super_block *sb, int *flags, char *options) +static int udf_reconfigure(struct fs_context *fc) { - struct udf_options uopt; + struct udf_options *uopt = fc->fs_private; + struct super_block *sb = fc->root->d_sb; struct udf_sb_info *sbi = UDF_SB(sb); + int readonly = fc->sb_flags & SB_RDONLY; int error = 0; - if (!(*flags & SB_RDONLY) && UDF_QUERY_FLAG(sb, UDF_FLAG_RW_INCOMPAT)) + if (!readonly && UDF_QUERY_FLAG(sb, UDF_FLAG_RW_INCOMPAT)) return -EACCES; sync_filesystem(sb); - uopt.flags = sbi->s_flags; - uopt.uid = sbi->s_uid; - uopt.gid = sbi->s_gid; - uopt.umask = sbi->s_umask; - uopt.fmode = sbi->s_fmode; - uopt.dmode = sbi->s_dmode; - uopt.nls_map = NULL; - - if (!udf_parse_options(options, &uopt, true)) - return -EINVAL; - write_lock(&sbi->s_cred_lock); - sbi->s_flags = uopt.flags; - sbi->s_uid = uopt.uid; - sbi->s_gid = uopt.gid; - sbi->s_umask = uopt.umask; - sbi->s_fmode = uopt.fmode; - sbi->s_dmode = uopt.dmode; + sbi->s_flags = uopt->flags; + sbi->s_uid = uopt->uid; + sbi->s_gid = uopt->gid; + sbi->s_umask = uopt->umask; + sbi->s_fmode = uopt->fmode; + sbi->s_dmode = uopt->dmode; write_unlock(&sbi->s_cred_lock); - if ((bool)(*flags & SB_RDONLY) == sb_rdonly(sb)) + if (readonly == sb_rdonly(sb)) goto out_unlock; - if (*flags & SB_RDONLY) + if (readonly) udf_close_lvid(sb); else udf_open_lvid(sb); @@ -863,7 +895,7 @@ static int udf_load_pvoldesc(struct super_block *sb, sector_t block) int ret; struct timestamp *ts; - outstr = kmalloc(128, GFP_NOFS); + outstr = kmalloc(128, GFP_KERNEL); if (!outstr) return -ENOMEM; @@ -1538,6 +1570,20 @@ out_bh: return ret; } +static bool udf_lvid_valid(struct super_block *sb, + struct logicalVolIntegrityDesc *lvid) +{ + u32 parts, impuselen; + + parts = le32_to_cpu(lvid->numOfPartitions); + impuselen = le32_to_cpu(lvid->lengthOfImpUse); + if (parts >= sb->s_blocksize || impuselen >= sb->s_blocksize || + sizeof(struct logicalVolIntegrityDesc) + impuselen + + 2 * parts * sizeof(u32) > sb->s_blocksize) + return false; + return true; +} + /* * Find the prevailing Logical Volume Integrity Descriptor. */ @@ -1548,7 +1594,6 @@ static void udf_load_logicalvolint(struct super_block *sb, struct kernel_extent_ struct udf_sb_info *sbi = UDF_SB(sb); struct logicalVolIntegrityDesc *lvid; int indirections = 0; - u32 parts, impuselen; while (++indirections <= UDF_MAX_LVID_NESTING) { final_bh = NULL; @@ -1570,32 +1615,27 @@ static void udf_load_logicalvolint(struct super_block *sb, struct kernel_extent_ if (!final_bh) return; - brelse(sbi->s_lvid_bh); - sbi->s_lvid_bh = final_bh; - lvid = (struct logicalVolIntegrityDesc *)final_bh->b_data; + if (udf_lvid_valid(sb, lvid)) { + brelse(sbi->s_lvid_bh); + sbi->s_lvid_bh = final_bh; + } else { + udf_warn(sb, "Corrupted LVID (parts=%u, impuselen=%u), " + "ignoring.\n", + le32_to_cpu(lvid->numOfPartitions), + le32_to_cpu(lvid->lengthOfImpUse)); + } + if (lvid->nextIntegrityExt.extLength == 0) - goto check; + return; loc = leea_to_cpu(lvid->nextIntegrityExt); } udf_warn(sb, "Too many LVID indirections (max %u), ignoring.\n", UDF_MAX_LVID_NESTING); -out_err: brelse(sbi->s_lvid_bh); sbi->s_lvid_bh = NULL; - return; -check: - parts = le32_to_cpu(lvid->numOfPartitions); - impuselen = le32_to_cpu(lvid->lengthOfImpUse); - if (parts >= sb->s_blocksize || impuselen >= sb->s_blocksize || - sizeof(struct logicalVolIntegrityDesc) + impuselen + - 2 * parts * sizeof(u32) > sb->s_blocksize) { - udf_warn(sb, "Corrupted LVID (parts=%u, impuselen=%u), " - "ignoring.\n", parts, impuselen); - goto out_err; - } } /* @@ -1945,7 +1985,7 @@ static int udf_load_vrs(struct super_block *sb, struct udf_options *uopt, return -EINVAL; } sbi->s_last_block = uopt->lastblock; - if (!uopt->novrs) { + if (!UDF_QUERY_FLAG(sb, UDF_FLAG_NOVRS)) { /* Check that it is NSR02 compliant */ nsr = udf_check_vsd(sb); if (!nsr) { @@ -2083,23 +2123,15 @@ u64 lvid_get_unique_id(struct super_block *sb) return ret; } -static int udf_fill_super(struct super_block *sb, void *options, int silent) +static int udf_fill_super(struct super_block *sb, struct fs_context *fc) { int ret = -EINVAL; struct inode *inode = NULL; - struct udf_options uopt; + struct udf_options *uopt = fc->fs_private; struct kernel_lb_addr rootdir, fileset; struct udf_sb_info *sbi; bool lvid_open = false; - - uopt.flags = (1 << UDF_FLAG_USE_AD_IN_ICB) | (1 << UDF_FLAG_STRICT); - /* By default we'll use overflow[ug]id when UDF inode [ug]id == -1 */ - uopt.uid = make_kuid(current_user_ns(), overflowuid); - uopt.gid = make_kgid(current_user_ns(), overflowgid); - uopt.umask = 0; - uopt.fmode = UDF_INVALID_MODE; - uopt.dmode = UDF_INVALID_MODE; - uopt.nls_map = NULL; + int silent = fc->sb_flags & SB_SILENT; sbi = kzalloc(sizeof(*sbi), GFP_KERNEL); if (!sbi) @@ -2109,25 +2141,23 @@ static int udf_fill_super(struct super_block *sb, void *options, int silent) mutex_init(&sbi->s_alloc_mutex); - if (!udf_parse_options((char *)options, &uopt, false)) - goto parse_options_failure; - fileset.logicalBlockNum = 0xFFFFFFFF; fileset.partitionReferenceNum = 0xFFFF; - sbi->s_flags = uopt.flags; - sbi->s_uid = uopt.uid; - sbi->s_gid = uopt.gid; - sbi->s_umask = uopt.umask; - sbi->s_fmode = uopt.fmode; - sbi->s_dmode = uopt.dmode; - sbi->s_nls_map = uopt.nls_map; + sbi->s_flags = uopt->flags; + sbi->s_uid = uopt->uid; + sbi->s_gid = uopt->gid; + sbi->s_umask = uopt->umask; + sbi->s_fmode = uopt->fmode; + sbi->s_dmode = uopt->dmode; + sbi->s_nls_map = uopt->nls_map; + uopt->nls_map = NULL; rwlock_init(&sbi->s_cred_lock); - if (uopt.session == 0xFFFFFFFF) + if (uopt->session == 0xFFFFFFFF) sbi->s_session = udf_get_last_session(sb); else - sbi->s_session = uopt.session; + sbi->s_session = uopt->session; udf_debug("Multi-session=%d\n", sbi->s_session); @@ -2138,16 +2168,16 @@ static int udf_fill_super(struct super_block *sb, void *options, int silent) sb->s_magic = UDF_SUPER_MAGIC; sb->s_time_gran = 1000; - if (uopt.flags & (1 << UDF_FLAG_BLOCKSIZE_SET)) { - ret = udf_load_vrs(sb, &uopt, silent, &fileset); + if (uopt->flags & (1 << UDF_FLAG_BLOCKSIZE_SET)) { + ret = udf_load_vrs(sb, uopt, silent, &fileset); } else { - uopt.blocksize = bdev_logical_block_size(sb->s_bdev); - while (uopt.blocksize <= 4096) { - ret = udf_load_vrs(sb, &uopt, silent, &fileset); + uopt->blocksize = bdev_logical_block_size(sb->s_bdev); + while (uopt->blocksize <= 4096) { + ret = udf_load_vrs(sb, uopt, silent, &fileset); if (ret < 0) { if (!silent && ret != -EACCES) { pr_notice("Scanning with blocksize %u failed\n", - uopt.blocksize); + uopt->blocksize); } brelse(sbi->s_lvid_bh); sbi->s_lvid_bh = NULL; @@ -2160,7 +2190,7 @@ static int udf_fill_super(struct super_block *sb, void *options, int silent) } else break; - uopt.blocksize <<= 1; + uopt->blocksize <<= 1; } } if (ret < 0) { @@ -2265,8 +2295,7 @@ static int udf_fill_super(struct super_block *sb, void *options, int silent) error_out: iput(sbi->s_vat_inode); -parse_options_failure: - unload_nls(uopt.nls_map); + unload_nls(uopt->nls_map); if (lvid_open) udf_close_lvid(sb); brelse(sbi->s_lvid_bh); diff --git a/fs/udf/udf_sb.h b/fs/udf/udf_sb.h index f9a60bc1abcf..08ec8756b948 100644 --- a/fs/udf/udf_sb.h +++ b/fs/udf/udf_sb.h @@ -23,6 +23,7 @@ #define UDF_FLAG_STRICT 5 #define UDF_FLAG_UNDELETE 6 #define UDF_FLAG_UNHIDE 7 +#define UDF_FLAG_NOVRS 8 #define UDF_FLAG_UID_FORGET 11 /* save -1 for uid to disk */ #define UDF_FLAG_GID_FORGET 12 #define UDF_FLAG_UID_SET 13 diff --git a/fs/userfaultfd.c b/fs/userfaultfd.c index 959551ff9a95..60dcfafdc11a 100644 --- a/fs/userfaultfd.c +++ b/fs/userfaultfd.c @@ -50,45 +50,6 @@ static struct ctl_table vm_userfaultfd_table[] = { static struct kmem_cache *userfaultfd_ctx_cachep __ro_after_init; -/* - * Start with fault_pending_wqh and fault_wqh so they're more likely - * to be in the same cacheline. - * - * Locking order: - * fd_wqh.lock - * fault_pending_wqh.lock - * fault_wqh.lock - * event_wqh.lock - * - * To avoid deadlocks, IRQs must be disabled when taking any of the above locks, - * since fd_wqh.lock is taken by aio_poll() while it's holding a lock that's - * also taken in IRQ context. - */ -struct userfaultfd_ctx { - /* waitqueue head for the pending (i.e. not read) userfaults */ - wait_queue_head_t fault_pending_wqh; - /* waitqueue head for the userfaults */ - wait_queue_head_t fault_wqh; - /* waitqueue head for the pseudo fd to wakeup poll/read */ - wait_queue_head_t fd_wqh; - /* waitqueue head for events */ - wait_queue_head_t event_wqh; - /* a refile sequence protected by fault_pending_wqh lock */ - seqcount_spinlock_t refile_seq; - /* pseudo fd refcounting */ - refcount_t refcount; - /* userfaultfd syscall flags */ - unsigned int flags; - /* features requested from the userspace */ - unsigned int features; - /* released */ - bool released; - /* memory mappings are changing because of non-cooperative event */ - atomic_t mmap_changing; - /* mm with one ore more vmas attached to this userfaultfd_ctx */ - struct mm_struct *mm; -}; - struct userfaultfd_fork_ctx { struct userfaultfd_ctx *orig; struct userfaultfd_ctx *new; @@ -724,12 +685,15 @@ int dup_userfaultfd(struct vm_area_struct *vma, struct list_head *fcs) ctx->flags = octx->flags; ctx->features = octx->features; ctx->released = false; + init_rwsem(&ctx->map_changing_lock); atomic_set(&ctx->mmap_changing, 0); ctx->mm = vma->vm_mm; mmgrab(ctx->mm); userfaultfd_ctx_get(octx); + down_write(&octx->map_changing_lock); atomic_inc(&octx->mmap_changing); + up_write(&octx->map_changing_lock); fctx->orig = octx; fctx->new = ctx; list_add_tail(&fctx->list, fcs); @@ -776,7 +740,9 @@ void mremap_userfaultfd_prep(struct vm_area_struct *vma, if (ctx->features & UFFD_FEATURE_EVENT_REMAP) { vm_ctx->ctx = ctx; userfaultfd_ctx_get(ctx); + down_write(&ctx->map_changing_lock); atomic_inc(&ctx->mmap_changing); + up_write(&ctx->map_changing_lock); } else { /* Drop uffd context if remap feature not enabled */ vma_start_write(vma); @@ -822,7 +788,9 @@ bool userfaultfd_remove(struct vm_area_struct *vma, return true; userfaultfd_ctx_get(ctx); + down_write(&ctx->map_changing_lock); atomic_inc(&ctx->mmap_changing); + up_write(&ctx->map_changing_lock); mmap_read_unlock(mm); msg_init(&ewq.msg); @@ -864,7 +832,9 @@ int userfaultfd_unmap_prep(struct vm_area_struct *vma, unsigned long start, return -ENOMEM; userfaultfd_ctx_get(ctx); + down_write(&ctx->map_changing_lock); atomic_inc(&ctx->mmap_changing); + up_write(&ctx->map_changing_lock); unmap_ctx->ctx = ctx; unmap_ctx->start = start; unmap_ctx->end = end; @@ -1748,9 +1718,8 @@ static int userfaultfd_copy(struct userfaultfd_ctx *ctx, if (uffdio_copy.mode & UFFDIO_COPY_MODE_WP) flags |= MFILL_ATOMIC_WP; if (mmget_not_zero(ctx->mm)) { - ret = mfill_atomic_copy(ctx->mm, uffdio_copy.dst, uffdio_copy.src, - uffdio_copy.len, &ctx->mmap_changing, - flags); + ret = mfill_atomic_copy(ctx, uffdio_copy.dst, uffdio_copy.src, + uffdio_copy.len, flags); mmput(ctx->mm); } else { return -ESRCH; @@ -1800,9 +1769,8 @@ static int userfaultfd_zeropage(struct userfaultfd_ctx *ctx, goto out; if (mmget_not_zero(ctx->mm)) { - ret = mfill_atomic_zeropage(ctx->mm, uffdio_zeropage.range.start, - uffdio_zeropage.range.len, - &ctx->mmap_changing); + ret = mfill_atomic_zeropage(ctx, uffdio_zeropage.range.start, + uffdio_zeropage.range.len); mmput(ctx->mm); } else { return -ESRCH; @@ -1857,9 +1825,8 @@ static int userfaultfd_writeprotect(struct userfaultfd_ctx *ctx, return -EINVAL; if (mmget_not_zero(ctx->mm)) { - ret = mwriteprotect_range(ctx->mm, uffdio_wp.range.start, - uffdio_wp.range.len, mode_wp, - &ctx->mmap_changing); + ret = mwriteprotect_range(ctx, uffdio_wp.range.start, + uffdio_wp.range.len, mode_wp); mmput(ctx->mm); } else { return -ESRCH; @@ -1909,9 +1876,8 @@ static int userfaultfd_continue(struct userfaultfd_ctx *ctx, unsigned long arg) flags |= MFILL_ATOMIC_WP; if (mmget_not_zero(ctx->mm)) { - ret = mfill_atomic_continue(ctx->mm, uffdio_continue.range.start, - uffdio_continue.range.len, - &ctx->mmap_changing, flags); + ret = mfill_atomic_continue(ctx, uffdio_continue.range.start, + uffdio_continue.range.len, flags); mmput(ctx->mm); } else { return -ESRCH; @@ -1964,9 +1930,8 @@ static inline int userfaultfd_poison(struct userfaultfd_ctx *ctx, unsigned long goto out; if (mmget_not_zero(ctx->mm)) { - ret = mfill_atomic_poison(ctx->mm, uffdio_poison.range.start, - uffdio_poison.range.len, - &ctx->mmap_changing, 0); + ret = mfill_atomic_poison(ctx, uffdio_poison.range.start, + uffdio_poison.range.len, 0); mmput(ctx->mm); } else { return -ESRCH; @@ -2040,16 +2005,8 @@ static int userfaultfd_move(struct userfaultfd_ctx *ctx, return -EINVAL; if (mmget_not_zero(mm)) { - mmap_read_lock(mm); - - /* Re-check after taking mmap_lock */ - if (likely(!atomic_read(&ctx->mmap_changing))) - ret = move_pages(ctx, mm, uffdio_move.dst, uffdio_move.src, - uffdio_move.len, uffdio_move.mode); - else - ret = -EINVAL; - - mmap_read_unlock(mm); + ret = move_pages(ctx, uffdio_move.dst, uffdio_move.src, + uffdio_move.len, uffdio_move.mode); mmput(mm); } else { return -ESRCH; @@ -2255,6 +2212,7 @@ static int new_userfaultfd(int flags) ctx->flags = flags; ctx->features = 0; ctx->released = false; + init_rwsem(&ctx->map_changing_lock); atomic_set(&ctx->mmap_changing, 0); ctx->mm = current->mm; /* prevent the mm struct to be freed */ diff --git a/fs/xfs/Kconfig b/fs/xfs/Kconfig index 567fb37274d3..d41edd30388b 100644 --- a/fs/xfs/Kconfig +++ b/fs/xfs/Kconfig @@ -124,12 +124,24 @@ config XFS_DRAIN_INTENTS bool select JUMP_LABEL if HAVE_ARCH_JUMP_LABEL +config XFS_LIVE_HOOKS + bool + select JUMP_LABEL if HAVE_ARCH_JUMP_LABEL + +config XFS_MEMORY_BUFS + bool + +config XFS_BTREE_IN_MEM + bool + config XFS_ONLINE_SCRUB bool "XFS online metadata check support" default n depends on XFS_FS depends on TMPFS && SHMEM + select XFS_LIVE_HOOKS select XFS_DRAIN_INTENTS + select XFS_MEMORY_BUFS help If you say Y here you will be able to check metadata on a mounted XFS filesystem. This feature is intended to reduce @@ -164,6 +176,7 @@ config XFS_ONLINE_REPAIR bool "XFS online metadata repair support" default n depends on XFS_FS && XFS_ONLINE_SCRUB + select XFS_BTREE_IN_MEM help If you say Y here you will be able to repair metadata on a mounted XFS filesystem. This feature is intended to reduce diff --git a/fs/xfs/Makefile b/fs/xfs/Makefile index fbe3cdc79036..76674ad5833e 100644 --- a/fs/xfs/Makefile +++ b/fs/xfs/Makefile @@ -92,8 +92,7 @@ xfs-y += xfs_aops.o \ xfs_symlink.o \ xfs_sysfs.o \ xfs_trans.o \ - xfs_xattr.o \ - kmem.o + xfs_xattr.o # low-level transaction/log code xfs-y += xfs_log.o \ @@ -137,6 +136,9 @@ xfs-$(CONFIG_FS_DAX) += xfs_notify_failure.o endif xfs-$(CONFIG_XFS_DRAIN_INTENTS) += xfs_drain.o +xfs-$(CONFIG_XFS_LIVE_HOOKS) += xfs_hooks.o +xfs-$(CONFIG_XFS_MEMORY_BUFS) += xfs_buf_mem.o +xfs-$(CONFIG_XFS_BTREE_IN_MEM) += libxfs/xfs_btree_mem.o # online scrub/repair ifeq ($(CONFIG_XFS_ONLINE_SCRUB),y) @@ -159,6 +161,8 @@ xfs-y += $(addprefix scrub/, \ health.o \ ialloc.o \ inode.o \ + iscan.o \ + nlinks.o \ parent.o \ readdir.o \ refcount.o \ @@ -179,6 +183,7 @@ xfs-$(CONFIG_XFS_RT) += $(addprefix scrub/, \ xfs-$(CONFIG_XFS_QUOTA) += $(addprefix scrub/, \ dqiterate.o \ quota.o \ + quotacheck.o \ ) # online repair @@ -188,12 +193,17 @@ xfs-y += $(addprefix scrub/, \ alloc_repair.o \ bmap_repair.o \ cow_repair.o \ + fscounters_repair.o \ ialloc_repair.o \ inode_repair.o \ newbt.o \ + nlinks_repair.o \ + rcbag_btree.o \ + rcbag.o \ reap.o \ refcount_repair.o \ repair.o \ + rmap_repair.o \ ) xfs-$(CONFIG_XFS_RT) += $(addprefix scrub/, \ @@ -202,6 +212,7 @@ xfs-$(CONFIG_XFS_RT) += $(addprefix scrub/, \ xfs-$(CONFIG_XFS_QUOTA) += $(addprefix scrub/, \ quota_repair.o \ + quotacheck_repair.o \ ) endif endif diff --git a/fs/xfs/kmem.c b/fs/xfs/kmem.c deleted file mode 100644 index c557a030acfe..000000000000 --- a/fs/xfs/kmem.c +++ /dev/null @@ -1,30 +0,0 @@ -// SPDX-License-Identifier: GPL-2.0 -/* - * Copyright (c) 2000-2005 Silicon Graphics, Inc. - * All Rights Reserved. - */ -#include "xfs.h" -#include "xfs_message.h" -#include "xfs_trace.h" - -void * -kmem_alloc(size_t size, xfs_km_flags_t flags) -{ - int retries = 0; - gfp_t lflags = kmem_flags_convert(flags); - void *ptr; - - trace_kmem_alloc(size, flags, _RET_IP_); - - do { - ptr = kmalloc(size, lflags); - if (ptr || (flags & KM_MAYFAIL)) - return ptr; - if (!(++retries % 100)) - xfs_err(NULL, - "%s(%u) possible memory allocation deadlock size %u in %s (mode:0x%x)", - current->comm, current->pid, - (unsigned int)size, __func__, lflags); - memalloc_retry_wait(lflags); - } while (1); -} diff --git a/fs/xfs/kmem.h b/fs/xfs/kmem.h deleted file mode 100644 index b987dc2c6851..000000000000 --- a/fs/xfs/kmem.h +++ /dev/null @@ -1,83 +0,0 @@ -/* SPDX-License-Identifier: GPL-2.0 */ -/* - * Copyright (c) 2000-2005 Silicon Graphics, Inc. - * All Rights Reserved. - */ -#ifndef __XFS_SUPPORT_KMEM_H__ -#define __XFS_SUPPORT_KMEM_H__ - -#include <linux/slab.h> -#include <linux/sched.h> -#include <linux/mm.h> -#include <linux/vmalloc.h> - -/* - * General memory allocation interfaces - */ - -typedef unsigned __bitwise xfs_km_flags_t; -#define KM_NOFS ((__force xfs_km_flags_t)0x0004u) -#define KM_MAYFAIL ((__force xfs_km_flags_t)0x0008u) -#define KM_ZERO ((__force xfs_km_flags_t)0x0010u) -#define KM_NOLOCKDEP ((__force xfs_km_flags_t)0x0020u) - -/* - * We use a special process flag to avoid recursive callbacks into - * the filesystem during transactions. We will also issue our own - * warnings, so we explicitly skip any generic ones (silly of us). - */ -static inline gfp_t -kmem_flags_convert(xfs_km_flags_t flags) -{ - gfp_t lflags; - - BUG_ON(flags & ~(KM_NOFS | KM_MAYFAIL | KM_ZERO | KM_NOLOCKDEP)); - - lflags = GFP_KERNEL | __GFP_NOWARN; - if (flags & KM_NOFS) - lflags &= ~__GFP_FS; - - /* - * Default page/slab allocator behavior is to retry for ever - * for small allocations. We can override this behavior by using - * __GFP_RETRY_MAYFAIL which will tell the allocator to retry as long - * as it is feasible but rather fail than retry forever for all - * request sizes. - */ - if (flags & KM_MAYFAIL) - lflags |= __GFP_RETRY_MAYFAIL; - - if (flags & KM_ZERO) - lflags |= __GFP_ZERO; - - if (flags & KM_NOLOCKDEP) - lflags |= __GFP_NOLOCKDEP; - - return lflags; -} - -extern void *kmem_alloc(size_t, xfs_km_flags_t); -static inline void kmem_free(const void *ptr) -{ - kvfree(ptr); -} - - -static inline void * -kmem_zalloc(size_t size, xfs_km_flags_t flags) -{ - return kmem_alloc(size, flags | KM_ZERO); -} - -/* - * Zone interfaces - */ -static inline struct page * -kmem_to_page(void *addr) -{ - if (is_vmalloc_addr(addr)) - return vmalloc_to_page(addr); - return virt_to_page(addr); -} - -#endif /* __XFS_SUPPORT_KMEM_H__ */ diff --git a/fs/xfs/libxfs/xfs_ag.c b/fs/xfs/libxfs/xfs_ag.c index 39d9525270b7..dc1873f76bff 100644 --- a/fs/xfs/libxfs/xfs_ag.c +++ b/fs/xfs/libxfs/xfs_ag.c @@ -217,6 +217,7 @@ xfs_initialize_perag_data( */ if (fdblocks > sbp->sb_dblocks || ifree > ialloc) { xfs_alert(mp, "AGF corruption. Please run xfs_repair."); + xfs_fs_mark_sick(mp, XFS_SICK_FS_COUNTERS); error = -EFSCORRUPTED; goto out; } @@ -241,7 +242,7 @@ __xfs_free_perag( struct xfs_perag *pag = container_of(head, struct xfs_perag, rcu_head); ASSERT(!delayed_work_pending(&pag->pag_blockgc_work)); - kmem_free(pag); + kfree(pag); } /* @@ -263,7 +264,7 @@ xfs_free_perag( xfs_defer_drain_free(&pag->pag_intents_drain); cancel_delayed_work_sync(&pag->pag_blockgc_work); - xfs_buf_hash_destroy(pag); + xfs_buf_cache_destroy(&pag->pag_bcache); /* drop the mount's active reference */ xfs_perag_rele(pag); @@ -351,9 +352,9 @@ xfs_free_unused_perag_range( spin_unlock(&mp->m_perag_lock); if (!pag) break; - xfs_buf_hash_destroy(pag); + xfs_buf_cache_destroy(&pag->pag_bcache); xfs_defer_drain_free(&pag->pag_intents_drain); - kmem_free(pag); + kfree(pag); } } @@ -381,7 +382,7 @@ xfs_initialize_perag( continue; } - pag = kmem_zalloc(sizeof(*pag), KM_MAYFAIL); + pag = kzalloc(sizeof(*pag), GFP_KERNEL | __GFP_RETRY_MAYFAIL); if (!pag) { error = -ENOMEM; goto out_unwind_new_pags; @@ -389,7 +390,7 @@ xfs_initialize_perag( pag->pag_agno = index; pag->pag_mount = mp; - error = radix_tree_preload(GFP_NOFS); + error = radix_tree_preload(GFP_KERNEL | __GFP_RETRY_MAYFAIL); if (error) goto out_free_pag; @@ -416,9 +417,10 @@ xfs_initialize_perag( init_waitqueue_head(&pag->pag_active_wq); pag->pagb_count = 0; pag->pagb_tree = RB_ROOT; + xfs_hooks_init(&pag->pag_rmap_update_hooks); #endif /* __KERNEL__ */ - error = xfs_buf_hash_init(pag); + error = xfs_buf_cache_init(&pag->pag_bcache); if (error) goto out_remove_pag; @@ -453,7 +455,7 @@ out_remove_pag: radix_tree_delete(&mp->m_perag_tree, index); spin_unlock(&mp->m_perag_lock); out_free_pag: - kmem_free(pag); + kfree(pag); out_unwind_new_pags: /* unwind any prior newly initialized pags */ xfs_free_unused_perag_range(mp, first_initialised, agcount); @@ -491,7 +493,7 @@ xfs_btroot_init( struct xfs_buf *bp, struct aghdr_init_data *id) { - xfs_btree_init_block(mp, bp, id->type, 0, 0, id->agno); + xfs_btree_init_buf(mp, bp, id->bc_ops, 0, 0, id->agno); } /* Finish initializing a free space btree. */ @@ -549,7 +551,7 @@ xfs_freesp_init_recs( } /* - * Alloc btree root block init functions + * bnobt/cntbt btree root block init functions */ static void xfs_bnoroot_init( @@ -557,17 +559,7 @@ xfs_bnoroot_init( struct xfs_buf *bp, struct aghdr_init_data *id) { - xfs_btree_init_block(mp, bp, XFS_BTNUM_BNO, 0, 0, id->agno); - xfs_freesp_init_recs(mp, bp, id); -} - -static void -xfs_cntroot_init( - struct xfs_mount *mp, - struct xfs_buf *bp, - struct aghdr_init_data *id) -{ - xfs_btree_init_block(mp, bp, XFS_BTNUM_CNT, 0, 0, id->agno); + xfs_btree_init_buf(mp, bp, id->bc_ops, 0, 0, id->agno); xfs_freesp_init_recs(mp, bp, id); } @@ -583,7 +575,7 @@ xfs_rmaproot_init( struct xfs_btree_block *block = XFS_BUF_TO_BLOCK(bp); struct xfs_rmap_rec *rrec; - xfs_btree_init_block(mp, bp, XFS_BTNUM_RMAP, 0, 4, id->agno); + xfs_btree_init_buf(mp, bp, id->bc_ops, 0, 4, id->agno); /* * mark the AG header regions as static metadata The BNO @@ -678,14 +670,13 @@ xfs_agfblock_init( agf->agf_versionnum = cpu_to_be32(XFS_AGF_VERSION); agf->agf_seqno = cpu_to_be32(id->agno); agf->agf_length = cpu_to_be32(id->agsize); - agf->agf_roots[XFS_BTNUM_BNOi] = cpu_to_be32(XFS_BNO_BLOCK(mp)); - agf->agf_roots[XFS_BTNUM_CNTi] = cpu_to_be32(XFS_CNT_BLOCK(mp)); - agf->agf_levels[XFS_BTNUM_BNOi] = cpu_to_be32(1); - agf->agf_levels[XFS_BTNUM_CNTi] = cpu_to_be32(1); + agf->agf_bno_root = cpu_to_be32(XFS_BNO_BLOCK(mp)); + agf->agf_cnt_root = cpu_to_be32(XFS_CNT_BLOCK(mp)); + agf->agf_bno_level = cpu_to_be32(1); + agf->agf_cnt_level = cpu_to_be32(1); if (xfs_has_rmapbt(mp)) { - agf->agf_roots[XFS_BTNUM_RMAPi] = - cpu_to_be32(XFS_RMAP_BLOCK(mp)); - agf->agf_levels[XFS_BTNUM_RMAPi] = cpu_to_be32(1); + agf->agf_rmap_root = cpu_to_be32(XFS_RMAP_BLOCK(mp)); + agf->agf_rmap_level = cpu_to_be32(1); agf->agf_rmap_blocks = cpu_to_be32(1); } @@ -796,7 +787,7 @@ struct xfs_aghdr_grow_data { size_t numblks; const struct xfs_buf_ops *ops; aghdr_init_work_f work; - xfs_btnum_t type; + const struct xfs_btree_ops *bc_ops; bool need_init; }; @@ -850,13 +841,15 @@ xfs_ag_init_headers( .numblks = BTOBB(mp->m_sb.sb_blocksize), .ops = &xfs_bnobt_buf_ops, .work = &xfs_bnoroot_init, + .bc_ops = &xfs_bnobt_ops, .need_init = true }, { /* CNT root block */ .daddr = XFS_AGB_TO_DADDR(mp, id->agno, XFS_CNT_BLOCK(mp)), .numblks = BTOBB(mp->m_sb.sb_blocksize), .ops = &xfs_cntbt_buf_ops, - .work = &xfs_cntroot_init, + .work = &xfs_bnoroot_init, + .bc_ops = &xfs_cntbt_ops, .need_init = true }, { /* INO root block */ @@ -864,7 +857,7 @@ xfs_ag_init_headers( .numblks = BTOBB(mp->m_sb.sb_blocksize), .ops = &xfs_inobt_buf_ops, .work = &xfs_btroot_init, - .type = XFS_BTNUM_INO, + .bc_ops = &xfs_inobt_ops, .need_init = true }, { /* FINO root block */ @@ -872,7 +865,7 @@ xfs_ag_init_headers( .numblks = BTOBB(mp->m_sb.sb_blocksize), .ops = &xfs_finobt_buf_ops, .work = &xfs_btroot_init, - .type = XFS_BTNUM_FINO, + .bc_ops = &xfs_finobt_ops, .need_init = xfs_has_finobt(mp) }, { /* RMAP root block */ @@ -880,6 +873,7 @@ xfs_ag_init_headers( .numblks = BTOBB(mp->m_sb.sb_blocksize), .ops = &xfs_rmapbt_buf_ops, .work = &xfs_rmaproot_init, + .bc_ops = &xfs_rmapbt_ops, .need_init = xfs_has_rmapbt(mp) }, { /* REFC root block */ @@ -887,7 +881,7 @@ xfs_ag_init_headers( .numblks = BTOBB(mp->m_sb.sb_blocksize), .ops = &xfs_refcountbt_buf_ops, .work = &xfs_btroot_init, - .type = XFS_BTNUM_REFC, + .bc_ops = &xfs_refcountbt_ops, .need_init = xfs_has_reflink(mp) }, { /* NULL terminating block */ @@ -905,7 +899,7 @@ xfs_ag_init_headers( id->daddr = dp->daddr; id->numblks = dp->numblks; - id->type = dp->type; + id->bc_ops = dp->bc_ops; error = xfs_ag_init_hdr(mp, id, dp->work, dp->ops); if (error) break; @@ -950,8 +944,10 @@ xfs_ag_shrink_space( agf = agfbp->b_addr; aglen = be32_to_cpu(agi->agi_length); /* some extra paranoid checks before we shrink the ag */ - if (XFS_IS_CORRUPT(mp, agf->agf_length != agi->agi_length)) + if (XFS_IS_CORRUPT(mp, agf->agf_length != agi->agi_length)) { + xfs_ag_mark_sick(pag, XFS_SICK_AG_AGF); return -EFSCORRUPTED; + } if (delta >= aglen) return -EINVAL; @@ -979,14 +975,23 @@ xfs_ag_shrink_space( if (error) { /* - * if extent allocation fails, need to roll the transaction to + * If extent allocation fails, need to roll the transaction to * ensure that the AGFL fixup has been committed anyway. + * + * We need to hold the AGF across the roll to ensure nothing can + * access the AG for allocation until the shrink is fully + * cleaned up. And due to the resetting of the AG block + * reservation space needing to lock the AGI, we also have to + * hold that so we don't get AGI/AGF lock order inversions in + * the error handling path. */ xfs_trans_bhold(*tpp, agfbp); + xfs_trans_bhold(*tpp, agibp); err2 = xfs_trans_roll(tpp); if (err2) return err2; xfs_trans_bjoin(*tpp, agfbp); + xfs_trans_bjoin(*tpp, agibp); goto resv_init_out; } diff --git a/fs/xfs/libxfs/xfs_ag.h b/fs/xfs/libxfs/xfs_ag.h index 4b343c4fac28..35de09a2516c 100644 --- a/fs/xfs/libxfs/xfs_ag.h +++ b/fs/xfs/libxfs/xfs_ag.h @@ -36,8 +36,9 @@ struct xfs_perag { atomic_t pag_active_ref; /* active reference count */ wait_queue_head_t pag_active_wq;/* woken active_ref falls to zero */ unsigned long pag_opstate; - uint8_t pagf_levels[XFS_BTNUM_AGF]; - /* # of levels in bno & cnt btree */ + uint8_t pagf_bno_level; /* # of levels in bno btree */ + uint8_t pagf_cnt_level; /* # of levels in cnt btree */ + uint8_t pagf_rmap_level;/* # of levels in rmap btree */ uint32_t pagf_flcount; /* count of blocks in freelist */ xfs_extlen_t pagf_freeblks; /* total free blocks */ xfs_extlen_t pagf_longest; /* longest free space */ @@ -86,8 +87,10 @@ struct xfs_perag { * Alternate btree heights so that online repair won't trip the write * verifiers while rebuilding the AG btrees. */ - uint8_t pagf_repair_levels[XFS_BTNUM_AGF]; + uint8_t pagf_repair_bno_level; + uint8_t pagf_repair_cnt_level; uint8_t pagf_repair_refcount_level; + uint8_t pagf_repair_rmap_level; #endif spinlock_t pag_state_lock; @@ -104,9 +107,7 @@ struct xfs_perag { int pag_ici_reclaimable; /* reclaimable inodes */ unsigned long pag_ici_reclaim_cursor; /* reclaim restart point */ - /* buffer cache index */ - spinlock_t pag_buf_lock; /* lock for pag_buf_hash */ - struct rhashtable pag_buf_hash; + struct xfs_buf_cache pag_bcache; /* background prealloc block trimming */ struct delayed_work pag_blockgc_work; @@ -119,6 +120,9 @@ struct xfs_perag { * inconsistencies. */ struct xfs_defer_drain pag_intents_drain; + + /* Hook to feed rmapbt updates to an active online repair. */ + struct xfs_hooks pag_rmap_update_hooks; #endif /* __KERNEL__ */ }; @@ -331,7 +335,7 @@ struct aghdr_init_data { /* per header data */ xfs_daddr_t daddr; /* header location */ size_t numblks; /* size of header */ - xfs_btnum_t type; /* type of btree root block */ + const struct xfs_btree_ops *bc_ops; /* btree ops */ }; int xfs_ag_init_headers(struct xfs_mount *mp, struct aghdr_init_data *id); diff --git a/fs/xfs/libxfs/xfs_alloc.c b/fs/xfs/libxfs/xfs_alloc.c index 3bd0a33fee0a..9da52e92172a 100644 --- a/fs/xfs/libxfs/xfs_alloc.c +++ b/fs/xfs/libxfs/xfs_alloc.c @@ -26,6 +26,7 @@ #include "xfs_ag.h" #include "xfs_ag_resv.h" #include "xfs_bmap.h" +#include "xfs_health.h" struct kmem_cache *xfs_extfree_item_cache; @@ -150,23 +151,38 @@ xfs_alloc_ag_max_usable( return mp->m_sb.sb_agblocks - blocks; } + +static int +xfs_alloc_lookup( + struct xfs_btree_cur *cur, + xfs_lookup_t dir, + xfs_agblock_t bno, + xfs_extlen_t len, + int *stat) +{ + int error; + + cur->bc_rec.a.ar_startblock = bno; + cur->bc_rec.a.ar_blockcount = len; + error = xfs_btree_lookup(cur, dir, stat); + if (*stat == 1) + cur->bc_flags |= XFS_BTREE_ALLOCBT_ACTIVE; + else + cur->bc_flags &= ~XFS_BTREE_ALLOCBT_ACTIVE; + return error; +} + /* * Lookup the record equal to [bno, len] in the btree given by cur. */ -STATIC int /* error */ +static inline int /* error */ xfs_alloc_lookup_eq( struct xfs_btree_cur *cur, /* btree cursor */ xfs_agblock_t bno, /* starting block of extent */ xfs_extlen_t len, /* length of extent */ int *stat) /* success/failure */ { - int error; - - cur->bc_rec.a.ar_startblock = bno; - cur->bc_rec.a.ar_blockcount = len; - error = xfs_btree_lookup(cur, XFS_LOOKUP_EQ, stat); - cur->bc_ag.abt.active = (*stat == 1); - return error; + return xfs_alloc_lookup(cur, XFS_LOOKUP_EQ, bno, len, stat); } /* @@ -180,13 +196,7 @@ xfs_alloc_lookup_ge( xfs_extlen_t len, /* length of extent */ int *stat) /* success/failure */ { - int error; - - cur->bc_rec.a.ar_startblock = bno; - cur->bc_rec.a.ar_blockcount = len; - error = xfs_btree_lookup(cur, XFS_LOOKUP_GE, stat); - cur->bc_ag.abt.active = (*stat == 1); - return error; + return xfs_alloc_lookup(cur, XFS_LOOKUP_GE, bno, len, stat); } /* @@ -200,19 +210,14 @@ xfs_alloc_lookup_le( xfs_extlen_t len, /* length of extent */ int *stat) /* success/failure */ { - int error; - cur->bc_rec.a.ar_startblock = bno; - cur->bc_rec.a.ar_blockcount = len; - error = xfs_btree_lookup(cur, XFS_LOOKUP_LE, stat); - cur->bc_ag.abt.active = (*stat == 1); - return error; + return xfs_alloc_lookup(cur, XFS_LOOKUP_LE, bno, len, stat); } static inline bool xfs_alloc_cur_active( struct xfs_btree_cur *cur) { - return cur && cur->bc_ag.abt.active; + return cur && (cur->bc_flags & XFS_BTREE_ALLOCBT_ACTIVE); } /* @@ -268,12 +273,12 @@ xfs_alloc_complain_bad_rec( struct xfs_mount *mp = cur->bc_mp; xfs_warn(mp, - "%s Freespace BTree record corruption in AG %d detected at %pS!", - cur->bc_btnum == XFS_BTNUM_BNO ? "Block" : "Size", - cur->bc_ag.pag->pag_agno, fa); + "%sbt record corruption in AG %d detected at %pS!", + cur->bc_ops->name, cur->bc_ag.pag->pag_agno, fa); xfs_warn(mp, "start block 0x%x block count 0x%x", irec->ar_startblock, irec->ar_blockcount); + xfs_btree_mark_sick(cur); return -EFSCORRUPTED; } @@ -497,14 +502,18 @@ xfs_alloc_fixup_trees( if (XFS_IS_CORRUPT(mp, i != 1 || nfbno1 != fbno || - nflen1 != flen)) + nflen1 != flen)) { + xfs_btree_mark_sick(cnt_cur); return -EFSCORRUPTED; + } #endif } else { if ((error = xfs_alloc_lookup_eq(cnt_cur, fbno, flen, &i))) return error; - if (XFS_IS_CORRUPT(mp, i != 1)) + if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cnt_cur); return -EFSCORRUPTED; + } } /* * Look up the record in the by-block tree if necessary. @@ -516,14 +525,18 @@ xfs_alloc_fixup_trees( if (XFS_IS_CORRUPT(mp, i != 1 || nfbno1 != fbno || - nflen1 != flen)) + nflen1 != flen)) { + xfs_btree_mark_sick(bno_cur); return -EFSCORRUPTED; + } #endif } else { if ((error = xfs_alloc_lookup_eq(bno_cur, fbno, flen, &i))) return error; - if (XFS_IS_CORRUPT(mp, i != 1)) + if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(bno_cur); return -EFSCORRUPTED; + } } #ifdef DEBUG @@ -536,8 +549,10 @@ xfs_alloc_fixup_trees( if (XFS_IS_CORRUPT(mp, bnoblock->bb_numrecs != - cntblock->bb_numrecs)) + cntblock->bb_numrecs)) { + xfs_btree_mark_sick(bno_cur); return -EFSCORRUPTED; + } } #endif @@ -567,30 +582,40 @@ xfs_alloc_fixup_trees( */ if ((error = xfs_btree_delete(cnt_cur, &i))) return error; - if (XFS_IS_CORRUPT(mp, i != 1)) + if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cnt_cur); return -EFSCORRUPTED; + } /* * Add new by-size btree entry(s). */ if (nfbno1 != NULLAGBLOCK) { if ((error = xfs_alloc_lookup_eq(cnt_cur, nfbno1, nflen1, &i))) return error; - if (XFS_IS_CORRUPT(mp, i != 0)) + if (XFS_IS_CORRUPT(mp, i != 0)) { + xfs_btree_mark_sick(cnt_cur); return -EFSCORRUPTED; + } if ((error = xfs_btree_insert(cnt_cur, &i))) return error; - if (XFS_IS_CORRUPT(mp, i != 1)) + if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cnt_cur); return -EFSCORRUPTED; + } } if (nfbno2 != NULLAGBLOCK) { if ((error = xfs_alloc_lookup_eq(cnt_cur, nfbno2, nflen2, &i))) return error; - if (XFS_IS_CORRUPT(mp, i != 0)) + if (XFS_IS_CORRUPT(mp, i != 0)) { + xfs_btree_mark_sick(cnt_cur); return -EFSCORRUPTED; + } if ((error = xfs_btree_insert(cnt_cur, &i))) return error; - if (XFS_IS_CORRUPT(mp, i != 1)) + if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cnt_cur); return -EFSCORRUPTED; + } } /* * Fix up the by-block btree entry(s). @@ -601,8 +626,10 @@ xfs_alloc_fixup_trees( */ if ((error = xfs_btree_delete(bno_cur, &i))) return error; - if (XFS_IS_CORRUPT(mp, i != 1)) + if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(bno_cur); return -EFSCORRUPTED; + } } else { /* * Update the by-block entry to start later|be shorter. @@ -616,12 +643,16 @@ xfs_alloc_fixup_trees( */ if ((error = xfs_alloc_lookup_eq(bno_cur, nfbno2, nflen2, &i))) return error; - if (XFS_IS_CORRUPT(mp, i != 0)) + if (XFS_IS_CORRUPT(mp, i != 0)) { + xfs_btree_mark_sick(bno_cur); return -EFSCORRUPTED; + } if ((error = xfs_btree_insert(bno_cur, &i))) return error; - if (XFS_IS_CORRUPT(mp, i != 1)) + if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(bno_cur); return -EFSCORRUPTED; + } } return 0; } @@ -755,6 +786,8 @@ xfs_alloc_read_agfl( mp, tp, mp->m_ddev_targp, XFS_AG_DADDR(mp, pag->pag_agno, XFS_AGFL_DADDR(mp)), XFS_FSS_TO_BB(mp, 1), 0, &bp, &xfs_agfl_buf_ops); + if (xfs_metadata_is_sick(error)) + xfs_ag_mark_sick(pag, XFS_SICK_AG_AGFL); if (error) return error; xfs_buf_set_ref(bp, XFS_AGFL_REF); @@ -776,6 +809,7 @@ xfs_alloc_update_counters( if (unlikely(be32_to_cpu(agf->agf_freeblks) > be32_to_cpu(agf->agf_length))) { xfs_buf_mark_corrupt(agbp); + xfs_ag_mark_sick(agbp->b_pag, XFS_SICK_AG_AGF); return -EFSCORRUPTED; } @@ -828,8 +862,8 @@ xfs_alloc_cur_setup( * attempt a small allocation. */ if (!acur->cnt) - acur->cnt = xfs_allocbt_init_cursor(args->mp, args->tp, - args->agbp, args->pag, XFS_BTNUM_CNT); + acur->cnt = xfs_cntbt_init_cursor(args->mp, args->tp, + args->agbp, args->pag); error = xfs_alloc_lookup_ge(acur->cnt, 0, args->maxlen, &i); if (error) return error; @@ -838,11 +872,11 @@ xfs_alloc_cur_setup( * Allocate the bnobt left and right search cursors. */ if (!acur->bnolt) - acur->bnolt = xfs_allocbt_init_cursor(args->mp, args->tp, - args->agbp, args->pag, XFS_BTNUM_BNO); + acur->bnolt = xfs_bnobt_init_cursor(args->mp, args->tp, + args->agbp, args->pag); if (!acur->bnogt) - acur->bnogt = xfs_allocbt_init_cursor(args->mp, args->tp, - args->agbp, args->pag, XFS_BTNUM_BNO); + acur->bnogt = xfs_bnobt_init_cursor(args->mp, args->tp, + args->agbp, args->pag); return i == 1 ? 0 : -ENOSPC; } @@ -884,15 +918,17 @@ xfs_alloc_cur_check( bool busy; unsigned busy_gen = 0; bool deactivate = false; - bool isbnobt = cur->bc_btnum == XFS_BTNUM_BNO; + bool isbnobt = xfs_btree_is_bno(cur->bc_ops); *new = 0; error = xfs_alloc_get_rec(cur, &bno, &len, &i); if (error) return error; - if (XFS_IS_CORRUPT(args->mp, i != 1)) + if (XFS_IS_CORRUPT(args->mp, i != 1)) { + xfs_btree_mark_sick(cur); return -EFSCORRUPTED; + } /* * Check minlen and deactivate a cntbt cursor if out of acceptable size @@ -958,9 +994,8 @@ xfs_alloc_cur_check( deactivate = true; out: if (deactivate) - cur->bc_ag.abt.active = false; - trace_xfs_alloc_cur_check(args->mp, cur->bc_btnum, bno, len, diff, - *new); + cur->bc_flags &= ~XFS_BTREE_ALLOCBT_ACTIVE; + trace_xfs_alloc_cur_check(cur, bno, len, diff, *new); return 0; } @@ -1098,6 +1133,7 @@ xfs_alloc_ag_vextent_small( if (error) goto error; if (XFS_IS_CORRUPT(args->mp, i != 1)) { + xfs_btree_mark_sick(ccur); error = -EFSCORRUPTED; goto error; } @@ -1132,6 +1168,7 @@ xfs_alloc_ag_vextent_small( *fbnop = args->agbno = fbno; *flenp = args->len = 1; if (XFS_IS_CORRUPT(args->mp, fbno >= be32_to_cpu(agf->agf_length))) { + xfs_btree_mark_sick(ccur); error = -EFSCORRUPTED; goto error; } @@ -1197,8 +1234,8 @@ xfs_alloc_ag_vextent_exact( /* * Allocate/initialize a cursor for the by-number freespace btree. */ - bno_cur = xfs_allocbt_init_cursor(args->mp, args->tp, args->agbp, - args->pag, XFS_BTNUM_BNO); + bno_cur = xfs_bnobt_init_cursor(args->mp, args->tp, args->agbp, + args->pag); /* * Lookup bno and minlen in the btree (minlen is irrelevant, really). @@ -1218,6 +1255,7 @@ xfs_alloc_ag_vextent_exact( if (error) goto error0; if (XFS_IS_CORRUPT(args->mp, i != 1)) { + xfs_btree_mark_sick(bno_cur); error = -EFSCORRUPTED; goto error0; } @@ -1257,8 +1295,8 @@ xfs_alloc_ag_vextent_exact( * We are allocating agbno for args->len * Allocate/initialize a cursor for the by-size btree. */ - cnt_cur = xfs_allocbt_init_cursor(args->mp, args->tp, args->agbp, - args->pag, XFS_BTNUM_CNT); + cnt_cur = xfs_cntbt_init_cursor(args->mp, args->tp, args->agbp, + args->pag); ASSERT(args->agbno + args->len <= be32_to_cpu(agf->agf_length)); error = xfs_alloc_fixup_trees(cnt_cur, bno_cur, fbno, flen, args->agbno, args->len, XFSA_FIXUP_BNO_OK); @@ -1330,7 +1368,7 @@ xfs_alloc_walk_iter( if (error) return error; if (i == 0) - cur->bc_ag.abt.active = false; + cur->bc_flags &= ~XFS_BTREE_ALLOCBT_ACTIVE; if (count > 0) count--; @@ -1444,7 +1482,7 @@ xfs_alloc_ag_vextent_locality( if (error) return error; if (i) { - acur->cnt->bc_ag.abt.active = true; + acur->cnt->bc_flags |= XFS_BTREE_ALLOCBT_ACTIVE; fbcur = acur->cnt; fbinc = false; } @@ -1497,8 +1535,10 @@ xfs_alloc_ag_vextent_lastblock( error = xfs_alloc_get_rec(acur->cnt, bno, len, &i); if (error) return error; - if (XFS_IS_CORRUPT(args->mp, i != 1)) + if (XFS_IS_CORRUPT(args->mp, i != 1)) { + xfs_btree_mark_sick(acur->cnt); return -EFSCORRUPTED; + } if (*len >= args->minlen) break; error = xfs_btree_increment(acur->cnt, 0, &i); @@ -1670,8 +1710,8 @@ restart: /* * Allocate and initialize a cursor for the by-size btree. */ - cnt_cur = xfs_allocbt_init_cursor(args->mp, args->tp, args->agbp, - args->pag, XFS_BTNUM_CNT); + cnt_cur = xfs_cntbt_init_cursor(args->mp, args->tp, args->agbp, + args->pag); bno_cur = NULL; /* @@ -1710,6 +1750,7 @@ restart: if (error) goto error0; if (XFS_IS_CORRUPT(args->mp, i != 1)) { + xfs_btree_mark_sick(cnt_cur); error = -EFSCORRUPTED; goto error0; } @@ -1756,6 +1797,7 @@ restart: rlen != 0 && (rlen > flen || rbno + rlen > fbno + flen))) { + xfs_btree_mark_sick(cnt_cur); error = -EFSCORRUPTED; goto error0; } @@ -1778,6 +1820,7 @@ restart: &i))) goto error0; if (XFS_IS_CORRUPT(args->mp, i != 1)) { + xfs_btree_mark_sick(cnt_cur); error = -EFSCORRUPTED; goto error0; } @@ -1790,6 +1833,7 @@ restart: rlen != 0 && (rlen > flen || rbno + rlen > fbno + flen))) { + xfs_btree_mark_sick(cnt_cur); error = -EFSCORRUPTED; goto error0; } @@ -1806,6 +1850,7 @@ restart: &i))) goto error0; if (XFS_IS_CORRUPT(args->mp, i != 1)) { + xfs_btree_mark_sick(cnt_cur); error = -EFSCORRUPTED; goto error0; } @@ -1844,14 +1889,15 @@ restart: rlen = args->len; if (XFS_IS_CORRUPT(args->mp, rlen > flen)) { + xfs_btree_mark_sick(cnt_cur); error = -EFSCORRUPTED; goto error0; } /* * Allocate and initialize a cursor for the by-block tree. */ - bno_cur = xfs_allocbt_init_cursor(args->mp, args->tp, args->agbp, - args->pag, XFS_BTNUM_BNO); + bno_cur = xfs_bnobt_init_cursor(args->mp, args->tp, args->agbp, + args->pag); if ((error = xfs_alloc_fixup_trees(cnt_cur, bno_cur, fbno, flen, rbno, rlen, XFSA_FIXUP_CNT_OK))) goto error0; @@ -1863,6 +1909,7 @@ restart: if (XFS_IS_CORRUPT(args->mp, args->agbno + args->len > be32_to_cpu(agf->agf_length))) { + xfs_ag_mark_sick(args->pag, XFS_SICK_AG_BNOBT); error = -EFSCORRUPTED; goto error0; } @@ -1924,7 +1971,7 @@ xfs_free_ag_extent( /* * Allocate and initialize a cursor for the by-block btree. */ - bno_cur = xfs_allocbt_init_cursor(mp, tp, agbp, pag, XFS_BTNUM_BNO); + bno_cur = xfs_bnobt_init_cursor(mp, tp, agbp, pag); /* * Look for a neighboring block on the left (lower block numbers) * that is contiguous with this space. @@ -1938,6 +1985,7 @@ xfs_free_ag_extent( if ((error = xfs_alloc_get_rec(bno_cur, <bno, <len, &i))) goto error0; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(bno_cur); error = -EFSCORRUPTED; goto error0; } @@ -1953,6 +2001,7 @@ xfs_free_ag_extent( * Very bad. */ if (XFS_IS_CORRUPT(mp, ltbno + ltlen > bno)) { + xfs_btree_mark_sick(bno_cur); error = -EFSCORRUPTED; goto error0; } @@ -1971,6 +2020,7 @@ xfs_free_ag_extent( if ((error = xfs_alloc_get_rec(bno_cur, >bno, >len, &i))) goto error0; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(bno_cur); error = -EFSCORRUPTED; goto error0; } @@ -1986,6 +2036,7 @@ xfs_free_ag_extent( * Very bad. */ if (XFS_IS_CORRUPT(mp, bno + len > gtbno)) { + xfs_btree_mark_sick(bno_cur); error = -EFSCORRUPTED; goto error0; } @@ -1994,7 +2045,7 @@ xfs_free_ag_extent( /* * Now allocate and initialize a cursor for the by-size tree. */ - cnt_cur = xfs_allocbt_init_cursor(mp, tp, agbp, pag, XFS_BTNUM_CNT); + cnt_cur = xfs_cntbt_init_cursor(mp, tp, agbp, pag); /* * Have both left and right contiguous neighbors. * Merge all three into a single free block. @@ -2006,12 +2057,14 @@ xfs_free_ag_extent( if ((error = xfs_alloc_lookup_eq(cnt_cur, ltbno, ltlen, &i))) goto error0; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cnt_cur); error = -EFSCORRUPTED; goto error0; } if ((error = xfs_btree_delete(cnt_cur, &i))) goto error0; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cnt_cur); error = -EFSCORRUPTED; goto error0; } @@ -2021,12 +2074,14 @@ xfs_free_ag_extent( if ((error = xfs_alloc_lookup_eq(cnt_cur, gtbno, gtlen, &i))) goto error0; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cnt_cur); error = -EFSCORRUPTED; goto error0; } if ((error = xfs_btree_delete(cnt_cur, &i))) goto error0; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cnt_cur); error = -EFSCORRUPTED; goto error0; } @@ -2036,6 +2091,7 @@ xfs_free_ag_extent( if ((error = xfs_btree_delete(bno_cur, &i))) goto error0; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(bno_cur); error = -EFSCORRUPTED; goto error0; } @@ -2045,6 +2101,7 @@ xfs_free_ag_extent( if ((error = xfs_btree_decrement(bno_cur, 0, &i))) goto error0; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(bno_cur); error = -EFSCORRUPTED; goto error0; } @@ -2064,6 +2121,7 @@ xfs_free_ag_extent( i != 1 || xxbno != ltbno || xxlen != ltlen)) { + xfs_btree_mark_sick(bno_cur); error = -EFSCORRUPTED; goto error0; } @@ -2088,12 +2146,14 @@ xfs_free_ag_extent( if ((error = xfs_alloc_lookup_eq(cnt_cur, ltbno, ltlen, &i))) goto error0; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cnt_cur); error = -EFSCORRUPTED; goto error0; } if ((error = xfs_btree_delete(cnt_cur, &i))) goto error0; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cnt_cur); error = -EFSCORRUPTED; goto error0; } @@ -2104,6 +2164,7 @@ xfs_free_ag_extent( if ((error = xfs_btree_decrement(bno_cur, 0, &i))) goto error0; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(bno_cur); error = -EFSCORRUPTED; goto error0; } @@ -2123,12 +2184,14 @@ xfs_free_ag_extent( if ((error = xfs_alloc_lookup_eq(cnt_cur, gtbno, gtlen, &i))) goto error0; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cnt_cur); error = -EFSCORRUPTED; goto error0; } if ((error = xfs_btree_delete(cnt_cur, &i))) goto error0; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cnt_cur); error = -EFSCORRUPTED; goto error0; } @@ -2151,6 +2214,7 @@ xfs_free_ag_extent( if ((error = xfs_btree_insert(bno_cur, &i))) goto error0; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(bno_cur); error = -EFSCORRUPTED; goto error0; } @@ -2163,12 +2227,14 @@ xfs_free_ag_extent( if ((error = xfs_alloc_lookup_eq(cnt_cur, nbno, nlen, &i))) goto error0; if (XFS_IS_CORRUPT(mp, i != 0)) { + xfs_btree_mark_sick(cnt_cur); error = -EFSCORRUPTED; goto error0; } if ((error = xfs_btree_insert(cnt_cur, &i))) goto error0; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cnt_cur); error = -EFSCORRUPTED; goto error0; } @@ -2267,8 +2333,9 @@ xfs_alloc_min_freelist( struct xfs_perag *pag) { /* AG btrees have at least 1 level. */ - static const uint8_t fake_levels[XFS_BTNUM_AGF] = {1, 1, 1}; - const uint8_t *levels = pag ? pag->pagf_levels : fake_levels; + const unsigned int bno_level = pag ? pag->pagf_bno_level : 1; + const unsigned int cnt_level = pag ? pag->pagf_cnt_level : 1; + const unsigned int rmap_level = pag ? pag->pagf_rmap_level : 1; unsigned int min_free; ASSERT(mp->m_alloc_maxlevels > 0); @@ -2295,16 +2362,12 @@ xfs_alloc_min_freelist( */ /* space needed by-bno freespace btree */ - min_free = min_t(unsigned int, levels[XFS_BTNUM_BNOi] + 1, - mp->m_alloc_maxlevels) * 2 - 2; + min_free = min(bno_level + 1, mp->m_alloc_maxlevels) * 2 - 2; /* space needed by-size freespace btree */ - min_free += min_t(unsigned int, levels[XFS_BTNUM_CNTi] + 1, - mp->m_alloc_maxlevels) * 2 - 2; + min_free += min(cnt_level + 1, mp->m_alloc_maxlevels) * 2 - 2; /* space needed reverse mapping used space btree */ if (xfs_has_rmapbt(mp)) - min_free += min_t(unsigned int, levels[XFS_BTNUM_RMAPi] + 1, - mp->m_rmap_maxlevels) * 2 - 2; - + min_free += min(rmap_level + 1, mp->m_rmap_maxlevels) * 2 - 2; return min_free; } @@ -2691,13 +2754,14 @@ xfs_exact_minlen_extent_available( xfs_extlen_t flen; int error = 0; - cnt_cur = xfs_allocbt_init_cursor(args->mp, args->tp, agbp, - args->pag, XFS_BTNUM_CNT); + cnt_cur = xfs_cntbt_init_cursor(args->mp, args->tp, agbp, + args->pag); error = xfs_alloc_lookup_ge(cnt_cur, 0, args->minlen, stat); if (error) goto out; if (*stat == 0) { + xfs_btree_mark_sick(cnt_cur); error = -EFSCORRUPTED; goto out; } @@ -2987,8 +3051,8 @@ xfs_alloc_log_agf( offsetof(xfs_agf_t, agf_versionnum), offsetof(xfs_agf_t, agf_seqno), offsetof(xfs_agf_t, agf_length), - offsetof(xfs_agf_t, agf_roots[0]), - offsetof(xfs_agf_t, agf_levels[0]), + offsetof(xfs_agf_t, agf_bno_root), /* also cnt/rmap root */ + offsetof(xfs_agf_t, agf_bno_level), /* also cnt/rmap levels */ offsetof(xfs_agf_t, agf_flfirst), offsetof(xfs_agf_t, agf_fllast), offsetof(xfs_agf_t, agf_flcount), @@ -3167,12 +3231,10 @@ xfs_agf_verify( be32_to_cpu(agf->agf_freeblks) > agf_length) return __this_address; - if (be32_to_cpu(agf->agf_levels[XFS_BTNUM_BNO]) < 1 || - be32_to_cpu(agf->agf_levels[XFS_BTNUM_CNT]) < 1 || - be32_to_cpu(agf->agf_levels[XFS_BTNUM_BNO]) > - mp->m_alloc_maxlevels || - be32_to_cpu(agf->agf_levels[XFS_BTNUM_CNT]) > - mp->m_alloc_maxlevels) + if (be32_to_cpu(agf->agf_bno_level) < 1 || + be32_to_cpu(agf->agf_cnt_level) < 1 || + be32_to_cpu(agf->agf_bno_level) > mp->m_alloc_maxlevels || + be32_to_cpu(agf->agf_cnt_level) > mp->m_alloc_maxlevels) return __this_address; if (xfs_has_lazysbcount(mp) && @@ -3183,9 +3245,8 @@ xfs_agf_verify( if (be32_to_cpu(agf->agf_rmap_blocks) > agf_length) return __this_address; - if (be32_to_cpu(agf->agf_levels[XFS_BTNUM_RMAP]) < 1 || - be32_to_cpu(agf->agf_levels[XFS_BTNUM_RMAP]) > - mp->m_rmap_maxlevels) + if (be32_to_cpu(agf->agf_rmap_level) < 1 || + be32_to_cpu(agf->agf_rmap_level) > mp->m_rmap_maxlevels) return __this_address; } @@ -3268,6 +3329,8 @@ xfs_read_agf( error = xfs_trans_read_buf(mp, tp, mp->m_ddev_targp, XFS_AG_DADDR(mp, pag->pag_agno, XFS_AGF_DADDR(mp)), XFS_FSS_TO_BB(mp, 1), flags, agfbpp, &xfs_agf_buf_ops); + if (xfs_metadata_is_sick(error)) + xfs_ag_mark_sick(pag, XFS_SICK_AG_AGF); if (error) return error; @@ -3309,12 +3372,9 @@ xfs_alloc_read_agf( pag->pagf_btreeblks = be32_to_cpu(agf->agf_btreeblks); pag->pagf_flcount = be32_to_cpu(agf->agf_flcount); pag->pagf_longest = be32_to_cpu(agf->agf_longest); - pag->pagf_levels[XFS_BTNUM_BNOi] = - be32_to_cpu(agf->agf_levels[XFS_BTNUM_BNOi]); - pag->pagf_levels[XFS_BTNUM_CNTi] = - be32_to_cpu(agf->agf_levels[XFS_BTNUM_CNTi]); - pag->pagf_levels[XFS_BTNUM_RMAPi] = - be32_to_cpu(agf->agf_levels[XFS_BTNUM_RMAPi]); + pag->pagf_bno_level = be32_to_cpu(agf->agf_bno_level); + pag->pagf_cnt_level = be32_to_cpu(agf->agf_cnt_level); + pag->pagf_rmap_level = be32_to_cpu(agf->agf_rmap_level); pag->pagf_refcount_level = be32_to_cpu(agf->agf_refcount_level); if (xfs_agfl_needs_reset(pag->pag_mount, agf)) set_bit(XFS_AGSTATE_AGFL_NEEDS_RESET, &pag->pag_opstate); @@ -3343,10 +3403,8 @@ xfs_alloc_read_agf( ASSERT(pag->pagf_btreeblks == be32_to_cpu(agf->agf_btreeblks)); ASSERT(pag->pagf_flcount == be32_to_cpu(agf->agf_flcount)); ASSERT(pag->pagf_longest == be32_to_cpu(agf->agf_longest)); - ASSERT(pag->pagf_levels[XFS_BTNUM_BNOi] == - be32_to_cpu(agf->agf_levels[XFS_BTNUM_BNOi])); - ASSERT(pag->pagf_levels[XFS_BTNUM_CNTi] == - be32_to_cpu(agf->agf_levels[XFS_BTNUM_CNTi])); + ASSERT(pag->pagf_bno_level == be32_to_cpu(agf->agf_bno_level)); + ASSERT(pag->pagf_cnt_level == be32_to_cpu(agf->agf_cnt_level)); } #endif if (agfbpp) @@ -3895,17 +3953,23 @@ __xfs_free_extent( return -EIO; error = xfs_free_extent_fix_freelist(tp, pag, &agbp); - if (error) + if (error) { + if (xfs_metadata_is_sick(error)) + xfs_ag_mark_sick(pag, XFS_SICK_AG_BNOBT); return error; + } + agf = agbp->b_addr; if (XFS_IS_CORRUPT(mp, agbno >= mp->m_sb.sb_agblocks)) { + xfs_ag_mark_sick(pag, XFS_SICK_AG_BNOBT); error = -EFSCORRUPTED; goto err_release; } /* validate the extent size is legal now we have the agf locked */ if (XFS_IS_CORRUPT(mp, agbno + len > be32_to_cpu(agf->agf_length))) { + xfs_ag_mark_sick(pag, XFS_SICK_AG_BNOBT); error = -EFSCORRUPTED; goto err_release; } @@ -3962,7 +4026,7 @@ xfs_alloc_query_range( union xfs_btree_irec high_brec = { .a = *high_rec }; struct xfs_alloc_query_range_info query = { .priv = priv, .fn = fn }; - ASSERT(cur->bc_btnum == XFS_BTNUM_BNO); + ASSERT(xfs_btree_is_bno(cur->bc_ops)); return xfs_btree_query_range(cur, &low_brec, &high_brec, xfs_alloc_query_range_helper, &query); } @@ -3976,7 +4040,7 @@ xfs_alloc_query_all( { struct xfs_alloc_query_range_info query; - ASSERT(cur->bc_btnum == XFS_BTNUM_BNO); + ASSERT(xfs_btree_is_bno(cur->bc_ops)); query.priv = priv; query.fn = fn; return xfs_btree_query_all(cur, xfs_alloc_query_range_helper, &query); diff --git a/fs/xfs/libxfs/xfs_alloc_btree.c b/fs/xfs/libxfs/xfs_alloc_btree.c index a7032bf0cd37..6ef5ddd89600 100644 --- a/fs/xfs/libxfs/xfs_alloc_btree.c +++ b/fs/xfs/libxfs/xfs_alloc_btree.c @@ -16,6 +16,7 @@ #include "xfs_alloc.h" #include "xfs_extent_busy.h" #include "xfs_error.h" +#include "xfs_health.h" #include "xfs_trace.h" #include "xfs_trans.h" #include "xfs_ag.h" @@ -23,13 +24,22 @@ static struct kmem_cache *xfs_allocbt_cur_cache; STATIC struct xfs_btree_cur * -xfs_allocbt_dup_cursor( +xfs_bnobt_dup_cursor( struct xfs_btree_cur *cur) { - return xfs_allocbt_init_cursor(cur->bc_mp, cur->bc_tp, - cur->bc_ag.agbp, cur->bc_ag.pag, cur->bc_btnum); + return xfs_bnobt_init_cursor(cur->bc_mp, cur->bc_tp, cur->bc_ag.agbp, + cur->bc_ag.pag); } +STATIC struct xfs_btree_cur * +xfs_cntbt_dup_cursor( + struct xfs_btree_cur *cur) +{ + return xfs_cntbt_init_cursor(cur->bc_mp, cur->bc_tp, cur->bc_ag.agbp, + cur->bc_ag.pag); +} + + STATIC void xfs_allocbt_set_root( struct xfs_btree_cur *cur, @@ -38,13 +48,18 @@ xfs_allocbt_set_root( { struct xfs_buf *agbp = cur->bc_ag.agbp; struct xfs_agf *agf = agbp->b_addr; - int btnum = cur->bc_btnum; ASSERT(ptr->s != 0); - agf->agf_roots[btnum] = ptr->s; - be32_add_cpu(&agf->agf_levels[btnum], inc); - cur->bc_ag.pag->pagf_levels[btnum] += inc; + if (xfs_btree_is_bno(cur->bc_ops)) { + agf->agf_bno_root = ptr->s; + be32_add_cpu(&agf->agf_bno_level, inc); + cur->bc_ag.pag->pagf_bno_level += inc; + } else { + agf->agf_cnt_root = ptr->s; + be32_add_cpu(&agf->agf_cnt_level, inc); + cur->bc_ag.pag->pagf_cnt_level += inc; + } xfs_alloc_log_agf(cur->bc_tp, agbp, XFS_AGF_ROOTS | XFS_AGF_LEVELS); } @@ -116,7 +131,7 @@ xfs_allocbt_update_lastrec( __be32 len; int numrecs; - ASSERT(cur->bc_btnum == XFS_BTNUM_CNT); + ASSERT(!xfs_btree_is_bno(cur->bc_ops)); switch (reason) { case LASTREC_UPDATE: @@ -226,7 +241,10 @@ xfs_allocbt_init_ptr_from_cur( ASSERT(cur->bc_ag.pag->pag_agno == be32_to_cpu(agf->agf_seqno)); - ptr->s = agf->agf_roots[cur->bc_btnum]; + if (xfs_btree_is_bno(cur->bc_ops)) + ptr->s = agf->agf_bno_root; + else + ptr->s = agf->agf_cnt_root; } STATIC int64_t @@ -299,13 +317,12 @@ xfs_allocbt_verify( struct xfs_perag *pag = bp->b_pag; xfs_failaddr_t fa; unsigned int level; - xfs_btnum_t btnum = XFS_BTNUM_BNOi; if (!xfs_verify_magic(bp, block->bb_magic)) return __this_address; if (xfs_has_crc(mp)) { - fa = xfs_btree_sblock_v5hdr_verify(bp); + fa = xfs_btree_agblock_v5hdr_verify(bp); if (fa) return fa; } @@ -320,26 +337,32 @@ xfs_allocbt_verify( * against. */ level = be16_to_cpu(block->bb_level); - if (bp->b_ops->magic[0] == cpu_to_be32(XFS_ABTC_MAGIC)) - btnum = XFS_BTNUM_CNTi; if (pag && xfs_perag_initialised_agf(pag)) { - unsigned int maxlevel = pag->pagf_levels[btnum]; + unsigned int maxlevel, repair_maxlevel = 0; -#ifdef CONFIG_XFS_ONLINE_REPAIR /* * Online repair could be rewriting the free space btrees, so * we'll validate against the larger of either tree while this * is going on. */ - maxlevel = max_t(unsigned int, maxlevel, - pag->pagf_repair_levels[btnum]); + if (bp->b_ops->magic[0] == cpu_to_be32(XFS_ABTC_MAGIC)) { + maxlevel = pag->pagf_cnt_level; +#ifdef CONFIG_XFS_ONLINE_REPAIR + repair_maxlevel = pag->pagf_repair_cnt_level; +#endif + } else { + maxlevel = pag->pagf_bno_level; +#ifdef CONFIG_XFS_ONLINE_REPAIR + repair_maxlevel = pag->pagf_repair_bno_level; #endif - if (level >= maxlevel) + } + + if (level >= max(maxlevel, repair_maxlevel)) return __this_address; } else if (level >= mp->m_alloc_maxlevels) return __this_address; - return xfs_btree_sblock_verify(bp, mp->m_alloc_mxr[level != 0]); + return xfs_btree_agblock_verify(bp, mp->m_alloc_mxr[level != 0]); } static void @@ -348,7 +371,7 @@ xfs_allocbt_read_verify( { xfs_failaddr_t fa; - if (!xfs_btree_sblock_verify_crc(bp)) + if (!xfs_btree_agblock_verify_crc(bp)) xfs_verifier_error(bp, -EFSBADCRC, __this_address); else { fa = xfs_allocbt_verify(bp); @@ -372,7 +395,7 @@ xfs_allocbt_write_verify( xfs_verifier_error(bp, -EFSCORRUPTED, fa); return; } - xfs_btree_sblock_calc_crc(bp); + xfs_btree_agblock_calc_crc(bp); } @@ -454,11 +477,19 @@ xfs_allocbt_keys_contiguous( be32_to_cpu(key2->alloc.ar_startblock)); } -static const struct xfs_btree_ops xfs_bnobt_ops = { +const struct xfs_btree_ops xfs_bnobt_ops = { + .name = "bno", + .type = XFS_BTREE_TYPE_AG, + .rec_len = sizeof(xfs_alloc_rec_t), .key_len = sizeof(xfs_alloc_key_t), + .ptr_len = XFS_BTREE_SHORT_PTR_LEN, - .dup_cursor = xfs_allocbt_dup_cursor, + .lru_refs = XFS_ALLOC_BTREE_REF, + .statoff = XFS_STATS_CALC_INDEX(xs_abtb_2), + .sick_mask = XFS_SICK_AG_BNOBT, + + .dup_cursor = xfs_bnobt_dup_cursor, .set_root = xfs_allocbt_set_root, .alloc_block = xfs_allocbt_alloc_block, .free_block = xfs_allocbt_free_block, @@ -477,11 +508,20 @@ static const struct xfs_btree_ops xfs_bnobt_ops = { .keys_contiguous = xfs_allocbt_keys_contiguous, }; -static const struct xfs_btree_ops xfs_cntbt_ops = { +const struct xfs_btree_ops xfs_cntbt_ops = { + .name = "cnt", + .type = XFS_BTREE_TYPE_AG, + .geom_flags = XFS_BTGEO_LASTREC_UPDATE, + .rec_len = sizeof(xfs_alloc_rec_t), .key_len = sizeof(xfs_alloc_key_t), + .ptr_len = XFS_BTREE_SHORT_PTR_LEN, + + .lru_refs = XFS_ALLOC_BTREE_REF, + .statoff = XFS_STATS_CALC_INDEX(xs_abtc_2), + .sick_mask = XFS_SICK_AG_CNTBT, - .dup_cursor = xfs_allocbt_dup_cursor, + .dup_cursor = xfs_cntbt_dup_cursor, .set_root = xfs_allocbt_set_root, .alloc_block = xfs_allocbt_alloc_block, .free_block = xfs_allocbt_free_block, @@ -500,76 +540,55 @@ static const struct xfs_btree_ops xfs_cntbt_ops = { .keys_contiguous = NULL, /* not needed right now */ }; -/* Allocate most of a new allocation btree cursor. */ -STATIC struct xfs_btree_cur * -xfs_allocbt_init_common( +/* + * Allocate a new bnobt cursor. + * + * For staging cursors tp and agbp are NULL. + */ +struct xfs_btree_cur * +xfs_bnobt_init_cursor( struct xfs_mount *mp, struct xfs_trans *tp, - struct xfs_perag *pag, - xfs_btnum_t btnum) + struct xfs_buf *agbp, + struct xfs_perag *pag) { struct xfs_btree_cur *cur; - ASSERT(btnum == XFS_BTNUM_BNO || btnum == XFS_BTNUM_CNT); - - cur = xfs_btree_alloc_cursor(mp, tp, btnum, mp->m_alloc_maxlevels, - xfs_allocbt_cur_cache); - cur->bc_ag.abt.active = false; - - if (btnum == XFS_BTNUM_CNT) { - cur->bc_ops = &xfs_cntbt_ops; - cur->bc_statoff = XFS_STATS_CALC_INDEX(xs_abtc_2); - cur->bc_flags = XFS_BTREE_LASTREC_UPDATE; - } else { - cur->bc_ops = &xfs_bnobt_ops; - cur->bc_statoff = XFS_STATS_CALC_INDEX(xs_abtb_2); - } - + cur = xfs_btree_alloc_cursor(mp, tp, &xfs_bnobt_ops, + mp->m_alloc_maxlevels, xfs_allocbt_cur_cache); cur->bc_ag.pag = xfs_perag_hold(pag); + cur->bc_ag.agbp = agbp; + if (agbp) { + struct xfs_agf *agf = agbp->b_addr; - if (xfs_has_crc(mp)) - cur->bc_flags |= XFS_BTREE_CRC_BLOCKS; - + cur->bc_nlevels = be32_to_cpu(agf->agf_bno_level); + } return cur; } /* - * Allocate a new allocation btree cursor. + * Allocate a new cntbt cursor. + * + * For staging cursors tp and agbp are NULL. */ -struct xfs_btree_cur * /* new alloc btree cursor */ -xfs_allocbt_init_cursor( - struct xfs_mount *mp, /* file system mount point */ - struct xfs_trans *tp, /* transaction pointer */ - struct xfs_buf *agbp, /* buffer for agf structure */ - struct xfs_perag *pag, - xfs_btnum_t btnum) /* btree identifier */ -{ - struct xfs_agf *agf = agbp->b_addr; - struct xfs_btree_cur *cur; - - cur = xfs_allocbt_init_common(mp, tp, pag, btnum); - if (btnum == XFS_BTNUM_CNT) - cur->bc_nlevels = be32_to_cpu(agf->agf_levels[XFS_BTNUM_CNT]); - else - cur->bc_nlevels = be32_to_cpu(agf->agf_levels[XFS_BTNUM_BNO]); - - cur->bc_ag.agbp = agbp; - - return cur; -} - -/* Create a free space btree cursor with a fake root for staging. */ struct xfs_btree_cur * -xfs_allocbt_stage_cursor( +xfs_cntbt_init_cursor( struct xfs_mount *mp, - struct xbtree_afakeroot *afake, - struct xfs_perag *pag, - xfs_btnum_t btnum) + struct xfs_trans *tp, + struct xfs_buf *agbp, + struct xfs_perag *pag) { struct xfs_btree_cur *cur; - cur = xfs_allocbt_init_common(mp, NULL, pag, btnum); - xfs_btree_stage_afakeroot(cur, afake); + cur = xfs_btree_alloc_cursor(mp, tp, &xfs_cntbt_ops, + mp->m_alloc_maxlevels, xfs_allocbt_cur_cache); + cur->bc_ag.pag = xfs_perag_hold(pag); + cur->bc_ag.agbp = agbp; + if (agbp) { + struct xfs_agf *agf = agbp->b_addr; + + cur->bc_nlevels = be32_to_cpu(agf->agf_cnt_level); + } return cur; } @@ -588,16 +607,16 @@ xfs_allocbt_commit_staged_btree( ASSERT(cur->bc_flags & XFS_BTREE_STAGING); - agf->agf_roots[cur->bc_btnum] = cpu_to_be32(afake->af_root); - agf->agf_levels[cur->bc_btnum] = cpu_to_be32(afake->af_levels); - xfs_alloc_log_agf(tp, agbp, XFS_AGF_ROOTS | XFS_AGF_LEVELS); - - if (cur->bc_btnum == XFS_BTNUM_BNO) { - xfs_btree_commit_afakeroot(cur, tp, agbp, &xfs_bnobt_ops); + if (xfs_btree_is_bno(cur->bc_ops)) { + agf->agf_bno_root = cpu_to_be32(afake->af_root); + agf->agf_bno_level = cpu_to_be32(afake->af_levels); } else { - cur->bc_flags |= XFS_BTREE_LASTREC_UPDATE; - xfs_btree_commit_afakeroot(cur, tp, agbp, &xfs_cntbt_ops); + agf->agf_cnt_root = cpu_to_be32(afake->af_root); + agf->agf_cnt_level = cpu_to_be32(afake->af_levels); } + xfs_alloc_log_agf(tp, agbp, XFS_AGF_ROOTS | XFS_AGF_LEVELS); + + xfs_btree_commit_afakeroot(cur, tp, agbp); } /* Calculate number of records in an alloc btree block. */ diff --git a/fs/xfs/libxfs/xfs_alloc_btree.h b/fs/xfs/libxfs/xfs_alloc_btree.h index 45df893ef6bb..155b47f231ab 100644 --- a/fs/xfs/libxfs/xfs_alloc_btree.h +++ b/fs/xfs/libxfs/xfs_alloc_btree.h @@ -47,12 +47,12 @@ struct xbtree_afakeroot; (maxrecs) * sizeof(xfs_alloc_key_t) + \ ((index) - 1) * sizeof(xfs_alloc_ptr_t))) -extern struct xfs_btree_cur *xfs_allocbt_init_cursor(struct xfs_mount *mp, +struct xfs_btree_cur *xfs_bnobt_init_cursor(struct xfs_mount *mp, struct xfs_trans *tp, struct xfs_buf *bp, - struct xfs_perag *pag, xfs_btnum_t btnum); -struct xfs_btree_cur *xfs_allocbt_stage_cursor(struct xfs_mount *mp, - struct xbtree_afakeroot *afake, struct xfs_perag *pag, - xfs_btnum_t btnum); + struct xfs_perag *pag); +struct xfs_btree_cur *xfs_cntbt_init_cursor(struct xfs_mount *mp, + struct xfs_trans *tp, struct xfs_buf *bp, + struct xfs_perag *pag); extern int xfs_allocbt_maxrecs(struct xfs_mount *, int, int); extern xfs_extlen_t xfs_allocbt_calc_size(struct xfs_mount *mp, unsigned long long len); diff --git a/fs/xfs/libxfs/xfs_attr.c b/fs/xfs/libxfs/xfs_attr.c index e965a48e7db9..673a4b6d2e8d 100644 --- a/fs/xfs/libxfs/xfs_attr.c +++ b/fs/xfs/libxfs/xfs_attr.c @@ -224,7 +224,7 @@ int xfs_attr_get_ilocked( struct xfs_da_args *args) { - ASSERT(xfs_isilocked(args->dp, XFS_ILOCK_SHARED | XFS_ILOCK_EXCL)); + xfs_assert_ilocked(args->dp, XFS_ILOCK_SHARED | XFS_ILOCK_EXCL); if (!xfs_inode_hasattr(args->dp)) return -ENOATTR; @@ -891,7 +891,8 @@ xfs_attr_defer_add( struct xfs_attr_intent *new; - new = kmem_cache_zalloc(xfs_attr_intent_cache, GFP_NOFS | __GFP_NOFAIL); + new = kmem_cache_zalloc(xfs_attr_intent_cache, + GFP_KERNEL | __GFP_NOFAIL); new->xattri_op_flags = op_flags; new->xattri_da_args = args; diff --git a/fs/xfs/libxfs/xfs_attr_leaf.c b/fs/xfs/libxfs/xfs_attr_leaf.c index 6374bf107242..ac904cc1a97b 100644 --- a/fs/xfs/libxfs/xfs_attr_leaf.c +++ b/fs/xfs/libxfs/xfs_attr_leaf.c @@ -29,6 +29,7 @@ #include "xfs_log.h" #include "xfs_ag.h" #include "xfs_errortag.h" +#include "xfs_health.h" /* @@ -879,8 +880,7 @@ xfs_attr_shortform_to_leaf( trace_xfs_attr_sf_to_leaf(args); - tmpbuffer = kmem_alloc(size, 0); - ASSERT(tmpbuffer != NULL); + tmpbuffer = kmalloc(size, GFP_KERNEL | __GFP_NOFAIL); memcpy(tmpbuffer, ifp->if_data, size); sf = (struct xfs_attr_sf_hdr *)tmpbuffer; @@ -924,7 +924,7 @@ xfs_attr_shortform_to_leaf( } error = 0; out: - kmem_free(tmpbuffer); + kfree(tmpbuffer); return error; } @@ -1059,7 +1059,7 @@ xfs_attr3_leaf_to_shortform( trace_xfs_attr_leaf_to_sf(args); - tmpbuffer = kmem_alloc(args->geo->blksize, 0); + tmpbuffer = kmalloc(args->geo->blksize, GFP_KERNEL | __GFP_NOFAIL); if (!tmpbuffer) return -ENOMEM; @@ -1125,7 +1125,7 @@ xfs_attr3_leaf_to_shortform( error = 0; out: - kmem_free(tmpbuffer); + kfree(tmpbuffer); return error; } @@ -1533,7 +1533,7 @@ xfs_attr3_leaf_compact( trace_xfs_attr_leaf_compact(args); - tmpbuffer = kmem_alloc(args->geo->blksize, 0); + tmpbuffer = kmalloc(args->geo->blksize, GFP_KERNEL | __GFP_NOFAIL); memcpy(tmpbuffer, bp->b_addr, args->geo->blksize); memset(bp->b_addr, 0, args->geo->blksize); leaf_src = (xfs_attr_leafblock_t *)tmpbuffer; @@ -1571,7 +1571,7 @@ xfs_attr3_leaf_compact( */ xfs_trans_log_buf(trans, bp, 0, args->geo->blksize - 1); - kmem_free(tmpbuffer); + kfree(tmpbuffer); } /* @@ -2250,7 +2250,8 @@ xfs_attr3_leaf_unbalance( struct xfs_attr_leafblock *tmp_leaf; struct xfs_attr3_icleaf_hdr tmphdr; - tmp_leaf = kmem_zalloc(state->args->geo->blksize, 0); + tmp_leaf = kzalloc(state->args->geo->blksize, + GFP_KERNEL | __GFP_NOFAIL); /* * Copy the header into the temp leaf so that all the stuff @@ -2290,7 +2291,7 @@ xfs_attr3_leaf_unbalance( } memcpy(save_leaf, tmp_leaf, state->args->geo->blksize); savehdr = tmphdr; /* struct copy */ - kmem_free(tmp_leaf); + kfree(tmp_leaf); } xfs_attr3_leaf_hdr_to_disk(state->args->geo, save_leaf, &savehdr); @@ -2343,6 +2344,7 @@ xfs_attr3_leaf_lookup_int( entries = xfs_attr3_leaf_entryp(leaf); if (ichdr.count >= args->geo->blksize / 8) { xfs_buf_mark_corrupt(bp); + xfs_da_mark_sick(args); return -EFSCORRUPTED; } @@ -2362,10 +2364,12 @@ xfs_attr3_leaf_lookup_int( } if (!(probe >= 0 && (!ichdr.count || probe < ichdr.count))) { xfs_buf_mark_corrupt(bp); + xfs_da_mark_sick(args); return -EFSCORRUPTED; } if (!(span <= 4 || be32_to_cpu(entry->hashval) == hashval)) { xfs_buf_mark_corrupt(bp); + xfs_da_mark_sick(args); return -EFSCORRUPTED; } diff --git a/fs/xfs/libxfs/xfs_attr_remote.c b/fs/xfs/libxfs/xfs_attr_remote.c index d440393b40eb..ff0412828772 100644 --- a/fs/xfs/libxfs/xfs_attr_remote.c +++ b/fs/xfs/libxfs/xfs_attr_remote.c @@ -22,6 +22,7 @@ #include "xfs_attr_remote.h" #include "xfs_trace.h" #include "xfs_error.h" +#include "xfs_health.h" #define ATTR_RMTVALUE_MAPSIZE 1 /* # of map entries at once */ @@ -276,17 +277,18 @@ xfs_attr3_rmt_hdr_set( */ STATIC int xfs_attr_rmtval_copyout( - struct xfs_mount *mp, - struct xfs_buf *bp, - xfs_ino_t ino, - int *offset, - int *valuelen, - uint8_t **dst) + struct xfs_mount *mp, + struct xfs_buf *bp, + struct xfs_inode *dp, + int *offset, + int *valuelen, + uint8_t **dst) { - char *src = bp->b_addr; - xfs_daddr_t bno = xfs_buf_daddr(bp); - int len = BBTOB(bp->b_length); - int blksize = mp->m_attr_geo->blksize; + char *src = bp->b_addr; + xfs_ino_t ino = dp->i_ino; + xfs_daddr_t bno = xfs_buf_daddr(bp); + int len = BBTOB(bp->b_length); + int blksize = mp->m_attr_geo->blksize; ASSERT(len >= blksize); @@ -302,6 +304,7 @@ xfs_attr_rmtval_copyout( xfs_alert(mp, "remote attribute header mismatch bno/off/len/owner (0x%llx/0x%x/Ox%x/0x%llx)", bno, *offset, byte_cnt, ino); + xfs_dirattr_mark_sick(dp, XFS_ATTR_FORK); return -EFSCORRUPTED; } hdr_size = sizeof(struct xfs_attr3_rmt_hdr); @@ -418,10 +421,12 @@ xfs_attr_rmtval_get( dblkcnt = XFS_FSB_TO_BB(mp, map[i].br_blockcount); error = xfs_buf_read(mp->m_ddev_targp, dblkno, dblkcnt, 0, &bp, &xfs_attr3_rmt_buf_ops); + if (xfs_metadata_is_sick(error)) + xfs_dirattr_mark_sick(args->dp, XFS_ATTR_FORK); if (error) return error; - error = xfs_attr_rmtval_copyout(mp, bp, args->dp->i_ino, + error = xfs_attr_rmtval_copyout(mp, bp, args->dp, &offset, &valuelen, &dst); xfs_buf_relse(bp); @@ -545,11 +550,13 @@ xfs_attr_rmtval_stale( struct xfs_buf *bp; int error; - ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL)); + xfs_assert_ilocked(ip, XFS_ILOCK_EXCL); if (XFS_IS_CORRUPT(mp, map->br_startblock == DELAYSTARTBLOCK) || - XFS_IS_CORRUPT(mp, map->br_startblock == HOLESTARTBLOCK)) + XFS_IS_CORRUPT(mp, map->br_startblock == HOLESTARTBLOCK)) { + xfs_bmap_mark_sick(ip, XFS_ATTR_FORK); return -EFSCORRUPTED; + } error = xfs_buf_incore(mp->m_ddev_targp, XFS_FSB_TO_DADDR(mp, map->br_startblock), @@ -659,8 +666,10 @@ xfs_attr_rmtval_invalidate( blkcnt, &map, &nmap, XFS_BMAPI_ATTRFORK); if (error) return error; - if (XFS_IS_CORRUPT(args->dp->i_mount, nmap != 1)) + if (XFS_IS_CORRUPT(args->dp->i_mount, nmap != 1)) { + xfs_bmap_mark_sick(args->dp, XFS_ATTR_FORK); return -EFSCORRUPTED; + } error = xfs_attr_rmtval_stale(args->dp, &map, XBF_TRYLOCK); if (error) return error; diff --git a/fs/xfs/libxfs/xfs_bmap.c b/fs/xfs/libxfs/xfs_bmap.c index f362345467fa..656c95a22f2e 100644 --- a/fs/xfs/libxfs/xfs_bmap.c +++ b/fs/xfs/libxfs/xfs_bmap.c @@ -36,6 +36,9 @@ #include "xfs_refcount.h" #include "xfs_icache.h" #include "xfs_iomap.h" +#include "xfs_health.h" +#include "xfs_bmap_item.h" +#include "xfs_symlink_remote.h" struct kmem_cache *xfs_bmap_intent_cache; @@ -225,6 +228,28 @@ xfs_bmap_forkoff_reset( } } +static int +xfs_bmap_read_buf( + struct xfs_mount *mp, /* file system mount point */ + struct xfs_trans *tp, /* transaction pointer */ + xfs_fsblock_t fsbno, /* file system block number */ + struct xfs_buf **bpp) /* buffer for fsbno */ +{ + struct xfs_buf *bp; /* return value */ + int error; + + if (!xfs_verify_fsbno(mp, fsbno)) + return -EFSCORRUPTED; + error = xfs_trans_read_buf(mp, tp, mp->m_ddev_targp, + XFS_FSB_TO_DADDR(mp, fsbno), mp->m_bsize, 0, &bp, + &xfs_bmbt_buf_ops); + if (!error) { + xfs_buf_set_ref(bp, XFS_BMAP_BTREE_REF); + *bpp = bp; + } + return error; +} + #ifdef DEBUG STATIC struct xfs_buf * xfs_bmap_get_bp( @@ -364,9 +389,9 @@ xfs_bmap_check_leaf_extents( bp = xfs_bmap_get_bp(cur, XFS_FSB_TO_DADDR(mp, bno)); if (!bp) { bp_release = 1; - error = xfs_btree_read_bufl(mp, NULL, bno, &bp, - XFS_BMAP_BTREE_REF, - &xfs_bmbt_buf_ops); + error = xfs_bmap_read_buf(mp, NULL, bno, &bp); + if (xfs_metadata_is_sick(error)) + xfs_btree_mark_sick(cur); if (error) goto error_norelse; } @@ -383,6 +408,7 @@ xfs_bmap_check_leaf_extents( pp = XFS_BMBT_PTR_ADDR(mp, block, 1, mp->m_bmap_dmxr[1]); bno = be64_to_cpu(*pp); if (XFS_IS_CORRUPT(mp, !xfs_verify_fsbno(mp, bno))) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto error0; } @@ -450,9 +476,9 @@ xfs_bmap_check_leaf_extents( bp = xfs_bmap_get_bp(cur, XFS_FSB_TO_DADDR(mp, bno)); if (!bp) { bp_release = 1; - error = xfs_btree_read_bufl(mp, NULL, bno, &bp, - XFS_BMAP_BTREE_REF, - &xfs_bmbt_buf_ops); + error = xfs_bmap_read_buf(mp, NULL, bno, &bp); + if (xfs_metadata_is_sick(error)) + xfs_btree_mark_sick(cur); if (error) goto error_norelse; } @@ -562,11 +588,14 @@ xfs_bmap_btree_to_extents( pp = XFS_BMAP_BROOT_PTR_ADDR(mp, rblock, 1, ifp->if_broot_bytes); cbno = be64_to_cpu(*pp); #ifdef DEBUG - if (XFS_IS_CORRUPT(cur->bc_mp, !xfs_btree_check_lptr(cur, cbno, 1))) + if (XFS_IS_CORRUPT(cur->bc_mp, !xfs_verify_fsbno(mp, cbno))) { + xfs_btree_mark_sick(cur); return -EFSCORRUPTED; + } #endif - error = xfs_btree_read_bufl(mp, tp, cbno, &cbp, XFS_BMAP_BTREE_REF, - &xfs_bmbt_buf_ops); + error = xfs_bmap_read_buf(mp, tp, cbno, &cbp); + if (xfs_metadata_is_sick(error)) + xfs_btree_mark_sick(cur); if (error) return error; cblock = XFS_BUF_TO_BLOCK(cbp); @@ -634,14 +663,13 @@ xfs_bmap_extents_to_btree( * Fill in the root. */ block = ifp->if_broot; - xfs_btree_init_block_int(mp, block, XFS_BUF_DADDR_NULL, - XFS_BTNUM_BMAP, 1, 1, ip->i_ino, - XFS_BTREE_LONG_PTRS); + xfs_bmbt_init_block(ip, block, NULL, 1, 1); /* * Need a cursor. Can't allocate until bb_level is filled in. */ cur = xfs_bmbt_init_cursor(mp, tp, ip, whichfork); - cur->bc_ino.flags = wasdel ? XFS_BTCUR_BMBT_WASDEL : 0; + if (wasdel) + cur->bc_flags |= XFS_BTREE_BMBT_WASDEL; /* * Convert to a btree with two levels, one record in root. */ @@ -667,7 +695,7 @@ xfs_bmap_extents_to_btree( goto out_root_realloc; } - cur->bc_ino.allocated++; + cur->bc_bmap.allocated++; ip->i_nblocks++; xfs_trans_mod_dquot_byino(tp, ip, XFS_TRANS_DQ_BCOUNT, 1L); error = xfs_trans_get_buf(tp, mp->m_ddev_targp, @@ -679,11 +707,8 @@ xfs_bmap_extents_to_btree( /* * Fill in the child block. */ - abp->b_ops = &xfs_bmbt_buf_ops; ablock = XFS_BUF_TO_BLOCK(abp); - xfs_btree_init_block_int(mp, ablock, xfs_buf_daddr(abp), - XFS_BTNUM_BMAP, 0, 0, ip->i_ino, - XFS_BTREE_LONG_PTRS); + xfs_bmbt_init_block(ip, ablock, abp, 0, 0); for_each_xfs_iext(ifp, &icur, &rec) { if (isnullstartblock(rec.br_startblock)) @@ -878,6 +903,7 @@ xfs_bmap_add_attrfork_btree( goto error0; /* must be at least one entry */ if (XFS_IS_CORRUPT(mp, stat != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto error0; } @@ -887,7 +913,7 @@ xfs_bmap_add_attrfork_btree( xfs_btree_del_cursor(cur, XFS_BTREE_NOERROR); return -ENOSPC; } - cur->bc_ino.allocated = 0; + cur->bc_bmap.allocated = 0; xfs_btree_del_cursor(cur, XFS_BTREE_NOERROR); } return 0; @@ -915,7 +941,7 @@ xfs_bmap_add_attrfork_extents( error = xfs_bmap_extents_to_btree(tp, ip, &cur, 0, flags, XFS_DATA_FORK); if (cur) { - cur->bc_ino.allocated = 0; + cur->bc_bmap.allocated = 0; xfs_btree_del_cursor(cur, error); } return error; @@ -960,6 +986,7 @@ xfs_bmap_add_attrfork_local( /* should only be called for types that support local format data */ ASSERT(0); + xfs_bmap_mark_sick(ip, XFS_ATTR_FORK); return -EFSCORRUPTED; } @@ -1143,6 +1170,7 @@ xfs_iread_bmbt_block( (unsigned long long)ip->i_ino); xfs_inode_verifier_error(ip, -EFSCORRUPTED, __func__, block, sizeof(*block), __this_address); + xfs_bmap_mark_sick(ip, whichfork); return -EFSCORRUPTED; } @@ -1158,6 +1186,7 @@ xfs_iread_bmbt_block( xfs_inode_verifier_error(ip, -EFSCORRUPTED, "xfs_iread_extents(2)", frp, sizeof(*frp), fa); + xfs_bmap_mark_sick(ip, whichfork); return xfs_bmap_complain_bad_rec(ip, whichfork, fa, &new); } @@ -1189,7 +1218,7 @@ xfs_iread_extents( if (!xfs_need_iread_extents(ifp)) return 0; - ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL)); + xfs_assert_ilocked(ip, XFS_ILOCK_EXCL); ir.loaded = 0; xfs_iext_first(ifp, &ir.icur); @@ -1201,6 +1230,7 @@ xfs_iread_extents( goto out; if (XFS_IS_CORRUPT(mp, ir.loaded != ifp->if_nextents)) { + xfs_bmap_mark_sick(ip, whichfork); error = -EFSCORRUPTED; goto out; } @@ -1213,6 +1243,8 @@ xfs_iread_extents( smp_store_release(&ifp->if_needextents, 0); return 0; out: + if (xfs_metadata_is_sick(error)) + xfs_bmap_mark_sick(ip, whichfork); xfs_iext_destroy(ifp); return error; } @@ -1292,6 +1324,7 @@ xfs_bmap_last_before( break; default: ASSERT(0); + xfs_bmap_mark_sick(ip, whichfork); return -EFSCORRUPTED; } @@ -1388,8 +1421,10 @@ xfs_bmap_last_offset( if (ifp->if_format == XFS_DINODE_FMT_LOCAL) return 0; - if (XFS_IS_CORRUPT(ip->i_mount, !xfs_ifork_has_extents(ifp))) + if (XFS_IS_CORRUPT(ip->i_mount, !xfs_ifork_has_extents(ifp))) { + xfs_bmap_mark_sick(ip, whichfork); return -EFSCORRUPTED; + } error = xfs_bmap_last_extent(NULL, ip, whichfork, &rec, &is_empty); if (error || is_empty) @@ -1429,8 +1464,7 @@ xfs_bmap_add_extent_delay_real( ASSERT(whichfork != XFS_ATTR_FORK); ASSERT(!isnullstartblock(new->br_startblock)); - ASSERT(!bma->cur || - (bma->cur->bc_ino.flags & XFS_BTCUR_BMBT_WASDEL)); + ASSERT(!bma->cur || (bma->cur->bc_flags & XFS_BTREE_BMBT_WASDEL)); XFS_STATS_INC(mp, xs_add_exlist); @@ -1528,6 +1562,7 @@ xfs_bmap_add_extent_delay_real( if (error) goto done; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(bma->cur); error = -EFSCORRUPTED; goto done; } @@ -1535,6 +1570,7 @@ xfs_bmap_add_extent_delay_real( if (error) goto done; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(bma->cur); error = -EFSCORRUPTED; goto done; } @@ -1542,6 +1578,7 @@ xfs_bmap_add_extent_delay_real( if (error) goto done; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(bma->cur); error = -EFSCORRUPTED; goto done; } @@ -1571,6 +1608,7 @@ xfs_bmap_add_extent_delay_real( if (error) goto done; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(bma->cur); error = -EFSCORRUPTED; goto done; } @@ -1604,6 +1642,7 @@ xfs_bmap_add_extent_delay_real( if (error) goto done; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(bma->cur); error = -EFSCORRUPTED; goto done; } @@ -1632,6 +1671,7 @@ xfs_bmap_add_extent_delay_real( if (error) goto done; if (XFS_IS_CORRUPT(mp, i != 0)) { + xfs_btree_mark_sick(bma->cur); error = -EFSCORRUPTED; goto done; } @@ -1639,6 +1679,7 @@ xfs_bmap_add_extent_delay_real( if (error) goto done; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(bma->cur); error = -EFSCORRUPTED; goto done; } @@ -1673,6 +1714,7 @@ xfs_bmap_add_extent_delay_real( if (error) goto done; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(bma->cur); error = -EFSCORRUPTED; goto done; } @@ -1698,6 +1740,7 @@ xfs_bmap_add_extent_delay_real( if (error) goto done; if (XFS_IS_CORRUPT(mp, i != 0)) { + xfs_btree_mark_sick(bma->cur); error = -EFSCORRUPTED; goto done; } @@ -1705,6 +1748,7 @@ xfs_bmap_add_extent_delay_real( if (error) goto done; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(bma->cur); error = -EFSCORRUPTED; goto done; } @@ -1721,7 +1765,7 @@ xfs_bmap_add_extent_delay_real( temp = PREV.br_blockcount - new->br_blockcount; da_new = XFS_FILBLKS_MIN(xfs_bmap_worst_indlen(bma->ip, temp), startblockval(PREV.br_startblock) - - (bma->cur ? bma->cur->bc_ino.allocated : 0)); + (bma->cur ? bma->cur->bc_bmap.allocated : 0)); PREV.br_startoff = new_endoff; PREV.br_blockcount = temp; @@ -1749,6 +1793,7 @@ xfs_bmap_add_extent_delay_real( if (error) goto done; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(bma->cur); error = -EFSCORRUPTED; goto done; } @@ -1785,6 +1830,7 @@ xfs_bmap_add_extent_delay_real( if (error) goto done; if (XFS_IS_CORRUPT(mp, i != 0)) { + xfs_btree_mark_sick(bma->cur); error = -EFSCORRUPTED; goto done; } @@ -1792,6 +1838,7 @@ xfs_bmap_add_extent_delay_real( if (error) goto done; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(bma->cur); error = -EFSCORRUPTED; goto done; } @@ -1808,7 +1855,7 @@ xfs_bmap_add_extent_delay_real( temp = PREV.br_blockcount - new->br_blockcount; da_new = XFS_FILBLKS_MIN(xfs_bmap_worst_indlen(bma->ip, temp), startblockval(PREV.br_startblock) - - (bma->cur ? bma->cur->bc_ino.allocated : 0)); + (bma->cur ? bma->cur->bc_bmap.allocated : 0)); PREV.br_startblock = nullstartblock(da_new); PREV.br_blockcount = temp; @@ -1871,6 +1918,7 @@ xfs_bmap_add_extent_delay_real( if (error) goto done; if (XFS_IS_CORRUPT(mp, i != 0)) { + xfs_btree_mark_sick(bma->cur); error = -EFSCORRUPTED; goto done; } @@ -1878,6 +1926,7 @@ xfs_bmap_add_extent_delay_real( if (error) goto done; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(bma->cur); error = -EFSCORRUPTED; goto done; } @@ -1929,8 +1978,8 @@ xfs_bmap_add_extent_delay_real( xfs_mod_delalloc(mp, (int64_t)da_new - da_old); if (bma->cur) { - da_new += bma->cur->bc_ino.allocated; - bma->cur->bc_ino.allocated = 0; + da_new += bma->cur->bc_bmap.allocated; + bma->cur->bc_bmap.allocated = 0; } /* adjust for changes in reserved delayed indirect blocks */ @@ -2074,30 +2123,35 @@ xfs_bmap_add_extent_unwritten_real( if (error) goto done; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto done; } if ((error = xfs_btree_delete(cur, &i))) goto done; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto done; } if ((error = xfs_btree_decrement(cur, 0, &i))) goto done; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto done; } if ((error = xfs_btree_delete(cur, &i))) goto done; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto done; } if ((error = xfs_btree_decrement(cur, 0, &i))) goto done; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto done; } @@ -2126,18 +2180,21 @@ xfs_bmap_add_extent_unwritten_real( if (error) goto done; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto done; } if ((error = xfs_btree_delete(cur, &i))) goto done; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto done; } if ((error = xfs_btree_decrement(cur, 0, &i))) goto done; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto done; } @@ -2169,18 +2226,21 @@ xfs_bmap_add_extent_unwritten_real( if (error) goto done; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto done; } if ((error = xfs_btree_delete(cur, &i))) goto done; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto done; } if ((error = xfs_btree_decrement(cur, 0, &i))) goto done; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto done; } @@ -2207,6 +2267,7 @@ xfs_bmap_add_extent_unwritten_real( if (error) goto done; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto done; } @@ -2240,6 +2301,7 @@ xfs_bmap_add_extent_unwritten_real( if (error) goto done; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto done; } @@ -2277,6 +2339,7 @@ xfs_bmap_add_extent_unwritten_real( if (error) goto done; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto done; } @@ -2287,6 +2350,7 @@ xfs_bmap_add_extent_unwritten_real( if ((error = xfs_btree_insert(cur, &i))) goto done; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto done; } @@ -2317,6 +2381,7 @@ xfs_bmap_add_extent_unwritten_real( if (error) goto done; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto done; } @@ -2353,6 +2418,7 @@ xfs_bmap_add_extent_unwritten_real( if (error) goto done; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto done; } @@ -2363,12 +2429,14 @@ xfs_bmap_add_extent_unwritten_real( if (error) goto done; if (XFS_IS_CORRUPT(mp, i != 0)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto done; } if ((error = xfs_btree_insert(cur, &i))) goto done; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto done; } @@ -2405,6 +2473,7 @@ xfs_bmap_add_extent_unwritten_real( if (error) goto done; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto done; } @@ -2417,6 +2486,7 @@ xfs_bmap_add_extent_unwritten_real( if ((error = xfs_btree_insert(cur, &i))) goto done; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto done; } @@ -2429,6 +2499,7 @@ xfs_bmap_add_extent_unwritten_real( if (error) goto done; if (XFS_IS_CORRUPT(mp, i != 0)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto done; } @@ -2436,6 +2507,7 @@ xfs_bmap_add_extent_unwritten_real( if ((error = xfs_btree_insert(cur, &i))) goto done; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto done; } @@ -2472,7 +2544,7 @@ xfs_bmap_add_extent_unwritten_real( /* clear out the allocated field, done with it now in any case. */ if (cur) { - cur->bc_ino.allocated = 0; + cur->bc_bmap.allocated = 0; *curp = cur; } @@ -2651,7 +2723,7 @@ xfs_bmap_add_extent_hole_real( struct xfs_bmbt_irec old; ASSERT(!isnullstartblock(new->br_startblock)); - ASSERT(!cur || !(cur->bc_ino.flags & XFS_BTCUR_BMBT_WASDEL)); + ASSERT(!cur || !(cur->bc_flags & XFS_BTREE_BMBT_WASDEL)); XFS_STATS_INC(mp, xs_add_exlist); @@ -2721,6 +2793,7 @@ xfs_bmap_add_extent_hole_real( if (error) goto done; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto done; } @@ -2728,6 +2801,7 @@ xfs_bmap_add_extent_hole_real( if (error) goto done; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto done; } @@ -2735,6 +2809,7 @@ xfs_bmap_add_extent_hole_real( if (error) goto done; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto done; } @@ -2764,6 +2839,7 @@ xfs_bmap_add_extent_hole_real( if (error) goto done; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto done; } @@ -2794,6 +2870,7 @@ xfs_bmap_add_extent_hole_real( if (error) goto done; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto done; } @@ -2820,6 +2897,7 @@ xfs_bmap_add_extent_hole_real( if (error) goto done; if (XFS_IS_CORRUPT(mp, i != 0)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto done; } @@ -2827,6 +2905,7 @@ xfs_bmap_add_extent_hole_real( if (error) goto done; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto done; } @@ -2853,7 +2932,7 @@ xfs_bmap_add_extent_hole_real( /* clear out the allocated field, done with it now in any case. */ if (cur) - cur->bc_ino.allocated = 0; + cur->bc_bmap.allocated = 0; xfs_bmap_check_leaf_extents(cur, ip, whichfork); done: @@ -3898,14 +3977,18 @@ xfs_bmapi_read( ASSERT(*nmap >= 1); ASSERT(!(flags & ~(XFS_BMAPI_ATTRFORK | XFS_BMAPI_ENTIRE))); - ASSERT(xfs_isilocked(ip, XFS_ILOCK_SHARED|XFS_ILOCK_EXCL)); + xfs_assert_ilocked(ip, XFS_ILOCK_SHARED | XFS_ILOCK_EXCL); - if (WARN_ON_ONCE(!ifp)) + if (WARN_ON_ONCE(!ifp)) { + xfs_bmap_mark_sick(ip, whichfork); return -EFSCORRUPTED; + } if (XFS_IS_CORRUPT(mp, !xfs_ifork_has_extents(ifp)) || - XFS_TEST_ERROR(false, mp, XFS_ERRTAG_BMAPIFORMAT)) + XFS_TEST_ERROR(false, mp, XFS_ERRTAG_BMAPIFORMAT)) { + xfs_bmap_mark_sick(ip, whichfork); return -EFSCORRUPTED; + } if (xfs_is_shutdown(mp)) return -EIO; @@ -4160,9 +4243,8 @@ xfs_bmapi_allocate( */ bma->nallocs++; - if (bma->cur) - bma->cur->bc_ino.flags = - bma->wasdel ? XFS_BTCUR_BMBT_WASDEL : 0; + if (bma->cur && bma->wasdel) + bma->cur->bc_flags |= XFS_BTREE_BMBT_WASDEL; bma->got.br_startoff = bma->offset; bma->got.br_startblock = bma->blkno; @@ -4369,7 +4451,7 @@ xfs_bmapi_write( ASSERT(tp != NULL); ASSERT(len > 0); ASSERT(ifp->if_format != XFS_DINODE_FMT_LOCAL); - ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL)); + xfs_assert_ilocked(ip, XFS_ILOCK_EXCL); ASSERT(!(flags & XFS_BMAPI_REMAP)); /* zeroing is for currently only for data extents, not metadata */ @@ -4386,6 +4468,7 @@ xfs_bmapi_write( if (XFS_IS_CORRUPT(mp, !xfs_ifork_has_extents(ifp)) || XFS_TEST_ERROR(false, mp, XFS_ERRTAG_BMAPIFORMAT)) { + xfs_bmap_mark_sick(ip, whichfork); return -EFSCORRUPTED; } @@ -4613,9 +4696,11 @@ xfs_bmapi_convert_delalloc( error = -ENOSPC; if (WARN_ON_ONCE(bma.blkno == NULLFSBLOCK)) goto out_finish; - error = -EFSCORRUPTED; - if (WARN_ON_ONCE(!xfs_valid_startblock(ip, bma.got.br_startblock))) + if (WARN_ON_ONCE(!xfs_valid_startblock(ip, bma.got.br_startblock))) { + xfs_bmap_mark_sick(ip, whichfork); + error = -EFSCORRUPTED; goto out_finish; + } XFS_STATS_ADD(mp, xs_xstrat_bytes, XFS_FSB_TO_B(mp, bma.length)); XFS_STATS_INC(mp, xs_xstrat_quick); @@ -4666,7 +4751,7 @@ xfs_bmapi_remap( ifp = xfs_ifork_ptr(ip, whichfork); ASSERT(len > 0); ASSERT(len <= (xfs_filblks_t)XFS_MAX_BMBT_EXTLEN); - ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL)); + xfs_assert_ilocked(ip, XFS_ILOCK_EXCL); ASSERT(!(flags & ~(XFS_BMAPI_ATTRFORK | XFS_BMAPI_PREALLOC | XFS_BMAPI_NORMAP))); ASSERT((flags & (XFS_BMAPI_ATTRFORK | XFS_BMAPI_PREALLOC)) != @@ -4674,6 +4759,7 @@ xfs_bmapi_remap( if (XFS_IS_CORRUPT(mp, !xfs_ifork_has_extents(ifp)) || XFS_TEST_ERROR(false, mp, XFS_ERRTAG_BMAPIFORMAT)) { + xfs_bmap_mark_sick(ip, whichfork); return -EFSCORRUPTED; } @@ -4693,10 +4779,8 @@ xfs_bmapi_remap( ip->i_nblocks += len; xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE); - if (ifp->if_format == XFS_DINODE_FMT_BTREE) { + if (ifp->if_format == XFS_DINODE_FMT_BTREE) cur = xfs_bmbt_init_cursor(mp, tp, ip, whichfork); - cur->bc_ino.flags = 0; - } got.br_startoff = bno; got.br_startblock = startblock; @@ -4831,7 +4915,7 @@ xfs_bmap_del_extent_delay( XFS_STATS_INC(mp, xs_del_exlist); - isrt = (whichfork == XFS_DATA_FORK) && XFS_IS_REALTIME_INODE(ip); + isrt = xfs_ifork_is_realtime(ip, whichfork); del_endoff = del->br_startoff + del->br_blockcount; got_endoff = got->br_startoff + got->br_blockcount; da_old = startblockval(got->br_startblock); @@ -5067,7 +5151,7 @@ xfs_bmap_del_extent_real( return -ENOSPC; *logflagsp = XFS_ILOG_CORE; - if (whichfork == XFS_DATA_FORK && XFS_IS_REALTIME_INODE(ip)) { + if (xfs_ifork_is_realtime(ip, whichfork)) { if (!(bflags & XFS_BMAPI_REMAP)) { error = xfs_rtfree_blocks(tp, del->br_startblock, del->br_blockcount); @@ -5088,8 +5172,10 @@ xfs_bmap_del_extent_real( error = xfs_bmbt_lookup_eq(cur, &got, &i); if (error) return error; - if (XFS_IS_CORRUPT(mp, i != 1)) + if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cur); return -EFSCORRUPTED; + } } if (got.br_startoff == del->br_startoff) @@ -5113,8 +5199,10 @@ xfs_bmap_del_extent_real( } if ((error = xfs_btree_delete(cur, &i))) return error; - if (XFS_IS_CORRUPT(mp, i != 1)) + if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cur); return -EFSCORRUPTED; + } break; case BMAP_LEFT_FILLING: /* @@ -5186,8 +5274,10 @@ xfs_bmap_del_extent_real( error = xfs_bmbt_lookup_eq(cur, &got, &i); if (error) return error; - if (XFS_IS_CORRUPT(mp, i != 1)) + if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cur); return -EFSCORRUPTED; + } /* * Update the btree record back * to the original value. @@ -5203,8 +5293,10 @@ xfs_bmap_del_extent_real( *logflagsp = 0; return -ENOSPC; } - if (XFS_IS_CORRUPT(mp, i != 1)) + if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cur); return -EFSCORRUPTED; + } } else *logflagsp |= xfs_ilog_fext(whichfork); @@ -5286,12 +5378,14 @@ __xfs_bunmapi( whichfork = xfs_bmapi_whichfork(flags); ASSERT(whichfork != XFS_COW_FORK); ifp = xfs_ifork_ptr(ip, whichfork); - if (XFS_IS_CORRUPT(mp, !xfs_ifork_has_extents(ifp))) + if (XFS_IS_CORRUPT(mp, !xfs_ifork_has_extents(ifp))) { + xfs_bmap_mark_sick(ip, whichfork); return -EFSCORRUPTED; + } if (xfs_is_shutdown(mp)) return -EIO; - ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL)); + xfs_assert_ilocked(ip, XFS_ILOCK_EXCL); ASSERT(len > 0); ASSERT(nexts >= 0); @@ -5304,7 +5398,7 @@ __xfs_bunmapi( return 0; } XFS_STATS_INC(mp, xs_blk_unmap); - isrt = (whichfork == XFS_DATA_FORK) && XFS_IS_REALTIME_INODE(ip); + isrt = xfs_ifork_is_realtime(ip, whichfork); end = start + len; if (!xfs_iext_lookup_extent_before(ip, ifp, &end, &icur, &got)) { @@ -5317,7 +5411,6 @@ __xfs_bunmapi( if (ifp->if_format == XFS_DINODE_FMT_BTREE) { ASSERT(ifp->if_format == XFS_DINODE_FMT_BTREE); cur = xfs_bmbt_init_cursor(mp, tp, ip, whichfork); - cur->bc_ino.flags = 0; } else cur = NULL; @@ -5367,7 +5460,7 @@ __xfs_bunmapi( if (del.br_startoff + del.br_blockcount > end + 1) del.br_blockcount = end + 1 - del.br_startoff; - if (!isrt) + if (!isrt || (flags & XFS_BMAPI_REMAP)) goto delete; mod = xfs_rtb_to_rtxoff(mp, @@ -5385,7 +5478,7 @@ __xfs_bunmapi( * This piece is unwritten, or we're not * using unwritten extents. Skip over it. */ - ASSERT(end >= mod); + ASSERT((flags & XFS_BMAPI_REMAP) || end >= mod); end -= mod > del.br_blockcount ? del.br_blockcount : mod; if (end < got.br_startoff && @@ -5555,7 +5648,7 @@ error0: xfs_trans_log_inode(tp, ip, logflags); if (cur) { if (!error) - cur->bc_ino.allocated = 0; + cur->bc_bmap.allocated = 0; xfs_btree_del_cursor(cur, error); } return error; @@ -5635,8 +5728,7 @@ xfs_bmse_merge( blockcount = left->br_blockcount + got->br_blockcount; - ASSERT(xfs_isilocked(ip, XFS_IOLOCK_EXCL)); - ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL)); + xfs_assert_ilocked(ip, XFS_IOLOCK_EXCL | XFS_ILOCK_EXCL); ASSERT(xfs_bmse_can_merge(left, got, shift)); new = *left; @@ -5657,21 +5749,27 @@ xfs_bmse_merge( error = xfs_bmbt_lookup_eq(cur, got, &i); if (error) return error; - if (XFS_IS_CORRUPT(mp, i != 1)) + if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cur); return -EFSCORRUPTED; + } error = xfs_btree_delete(cur, &i); if (error) return error; - if (XFS_IS_CORRUPT(mp, i != 1)) + if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cur); return -EFSCORRUPTED; + } /* lookup and update size of the previous extent */ error = xfs_bmbt_lookup_eq(cur, left, &i); if (error) return error; - if (XFS_IS_CORRUPT(mp, i != 1)) + if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cur); return -EFSCORRUPTED; + } error = xfs_bmbt_update(cur, &new); if (error) @@ -5719,8 +5817,10 @@ xfs_bmap_shift_update_extent( error = xfs_bmbt_lookup_eq(cur, &prev, &i); if (error) return error; - if (XFS_IS_CORRUPT(mp, i != 1)) + if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cur); return -EFSCORRUPTED; + } error = xfs_bmbt_update(cur, got); if (error) @@ -5758,28 +5858,28 @@ xfs_bmap_collapse_extents( if (XFS_IS_CORRUPT(mp, !xfs_ifork_has_extents(ifp)) || XFS_TEST_ERROR(false, mp, XFS_ERRTAG_BMAPIFORMAT)) { + xfs_bmap_mark_sick(ip, whichfork); return -EFSCORRUPTED; } if (xfs_is_shutdown(mp)) return -EIO; - ASSERT(xfs_isilocked(ip, XFS_IOLOCK_EXCL | XFS_ILOCK_EXCL)); + xfs_assert_ilocked(ip, XFS_IOLOCK_EXCL | XFS_ILOCK_EXCL); error = xfs_iread_extents(tp, ip, whichfork); if (error) return error; - if (ifp->if_format == XFS_DINODE_FMT_BTREE) { + if (ifp->if_format == XFS_DINODE_FMT_BTREE) cur = xfs_bmbt_init_cursor(mp, tp, ip, whichfork); - cur->bc_ino.flags = 0; - } if (!xfs_iext_lookup_extent(ip, ifp, *next_fsb, &icur, &got)) { *done = true; goto del_cursor; } if (XFS_IS_CORRUPT(mp, isnullstartblock(got.br_startblock))) { + xfs_bmap_mark_sick(ip, whichfork); error = -EFSCORRUPTED; goto del_cursor; } @@ -5837,7 +5937,7 @@ xfs_bmap_can_insert_extents( int is_empty; int error = 0; - ASSERT(xfs_isilocked(ip, XFS_IOLOCK_EXCL)); + xfs_assert_ilocked(ip, XFS_IOLOCK_EXCL); if (xfs_is_shutdown(ip->i_mount)) return -EIO; @@ -5873,22 +5973,21 @@ xfs_bmap_insert_extents( if (XFS_IS_CORRUPT(mp, !xfs_ifork_has_extents(ifp)) || XFS_TEST_ERROR(false, mp, XFS_ERRTAG_BMAPIFORMAT)) { + xfs_bmap_mark_sick(ip, whichfork); return -EFSCORRUPTED; } if (xfs_is_shutdown(mp)) return -EIO; - ASSERT(xfs_isilocked(ip, XFS_IOLOCK_EXCL | XFS_ILOCK_EXCL)); + xfs_assert_ilocked(ip, XFS_IOLOCK_EXCL | XFS_ILOCK_EXCL); error = xfs_iread_extents(tp, ip, whichfork); if (error) return error; - if (ifp->if_format == XFS_DINODE_FMT_BTREE) { + if (ifp->if_format == XFS_DINODE_FMT_BTREE) cur = xfs_bmbt_init_cursor(mp, tp, ip, whichfork); - cur->bc_ino.flags = 0; - } if (*next_fsb == NULLFSBLOCK) { xfs_iext_last(ifp, &icur); @@ -5904,11 +6003,13 @@ xfs_bmap_insert_extents( } } if (XFS_IS_CORRUPT(mp, isnullstartblock(got.br_startblock))) { + xfs_bmap_mark_sick(ip, whichfork); error = -EFSCORRUPTED; goto del_cursor; } if (XFS_IS_CORRUPT(mp, stop_fsb > got.br_startoff)) { + xfs_bmap_mark_sick(ip, whichfork); error = -EFSCORRUPTED; goto del_cursor; } @@ -5976,6 +6077,7 @@ xfs_bmap_split_extent( if (XFS_IS_CORRUPT(mp, !xfs_ifork_has_extents(ifp)) || XFS_TEST_ERROR(false, mp, XFS_ERRTAG_BMAPIFORMAT)) { + xfs_bmap_mark_sick(ip, whichfork); return -EFSCORRUPTED; } @@ -6002,11 +6104,11 @@ xfs_bmap_split_extent( if (ifp->if_format == XFS_DINODE_FMT_BTREE) { cur = xfs_bmbt_init_cursor(mp, tp, ip, whichfork); - cur->bc_ino.flags = 0; error = xfs_bmbt_lookup_eq(cur, &got, &i); if (error) goto del_cursor; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto del_cursor; } @@ -6034,6 +6136,7 @@ xfs_bmap_split_extent( if (error) goto del_cursor; if (XFS_IS_CORRUPT(mp, i != 0)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto del_cursor; } @@ -6041,6 +6144,7 @@ xfs_bmap_split_extent( if (error) goto del_cursor; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto del_cursor; } @@ -6060,7 +6164,7 @@ xfs_bmap_split_extent( del_cursor: if (cur) { - cur->bc_ino.allocated = 0; + cur->bc_bmap.allocated = 0; xfs_btree_del_cursor(cur, error); } @@ -6069,17 +6173,8 @@ del_cursor: return error; } -/* Deferred mapping is only for real extents in the data fork. */ -static bool -xfs_bmap_is_update_needed( - struct xfs_bmbt_irec *bmap) -{ - return bmap->br_startblock != HOLESTARTBLOCK && - bmap->br_startblock != DELAYSTARTBLOCK; -} - /* Record a bmap intent. */ -static int +static inline void __xfs_bmap_add( struct xfs_trans *tp, enum xfs_bmap_intent_type type, @@ -6089,25 +6184,19 @@ __xfs_bmap_add( { struct xfs_bmap_intent *bi; - trace_xfs_bmap_defer(tp->t_mountp, - XFS_FSB_TO_AGNO(tp->t_mountp, bmap->br_startblock), - type, - XFS_FSB_TO_AGBNO(tp->t_mountp, bmap->br_startblock), - ip->i_ino, whichfork, - bmap->br_startoff, - bmap->br_blockcount, - bmap->br_state); + if ((whichfork != XFS_DATA_FORK && whichfork != XFS_ATTR_FORK) || + bmap->br_startblock == HOLESTARTBLOCK || + bmap->br_startblock == DELAYSTARTBLOCK) + return; - bi = kmem_cache_alloc(xfs_bmap_intent_cache, GFP_NOFS | __GFP_NOFAIL); + bi = kmem_cache_alloc(xfs_bmap_intent_cache, GFP_KERNEL | __GFP_NOFAIL); INIT_LIST_HEAD(&bi->bi_list); bi->bi_type = type; bi->bi_owner = ip; bi->bi_whichfork = whichfork; bi->bi_bmap = *bmap; - xfs_bmap_update_get_group(tp->t_mountp, bi); - xfs_defer_add(tp, &bi->bi_list, &xfs_bmap_update_defer_type); - return 0; + xfs_bmap_defer_add(tp, bi); } /* Map an extent into a file. */ @@ -6115,12 +6204,10 @@ void xfs_bmap_map_extent( struct xfs_trans *tp, struct xfs_inode *ip, + int whichfork, struct xfs_bmbt_irec *PREV) { - if (!xfs_bmap_is_update_needed(PREV)) - return; - - __xfs_bmap_add(tp, XFS_BMAP_MAP, ip, XFS_DATA_FORK, PREV); + __xfs_bmap_add(tp, XFS_BMAP_MAP, ip, whichfork, PREV); } /* Unmap an extent out of a file. */ @@ -6128,12 +6215,10 @@ void xfs_bmap_unmap_extent( struct xfs_trans *tp, struct xfs_inode *ip, + int whichfork, struct xfs_bmbt_irec *PREV) { - if (!xfs_bmap_is_update_needed(PREV)) - return; - - __xfs_bmap_add(tp, XFS_BMAP_UNMAP, ip, XFS_DATA_FORK, PREV); + __xfs_bmap_add(tp, XFS_BMAP_UNMAP, ip, whichfork, PREV); } /* @@ -6147,36 +6232,35 @@ xfs_bmap_finish_one( { struct xfs_bmbt_irec *bmap = &bi->bi_bmap; int error = 0; + int flags = 0; - ASSERT(tp->t_highest_agno == NULLAGNUMBER); + if (bi->bi_whichfork == XFS_ATTR_FORK) + flags |= XFS_BMAPI_ATTRFORK; - trace_xfs_bmap_deferred(tp->t_mountp, - XFS_FSB_TO_AGNO(tp->t_mountp, bmap->br_startblock), - bi->bi_type, - XFS_FSB_TO_AGBNO(tp->t_mountp, bmap->br_startblock), - bi->bi_owner->i_ino, bi->bi_whichfork, - bmap->br_startoff, bmap->br_blockcount, - bmap->br_state); + ASSERT(tp->t_highest_agno == NULLAGNUMBER); - if (WARN_ON_ONCE(bi->bi_whichfork != XFS_DATA_FORK)) - return -EFSCORRUPTED; + trace_xfs_bmap_deferred(bi); - if (XFS_TEST_ERROR(false, tp->t_mountp, - XFS_ERRTAG_BMAP_FINISH_ONE)) + if (XFS_TEST_ERROR(false, tp->t_mountp, XFS_ERRTAG_BMAP_FINISH_ONE)) return -EIO; switch (bi->bi_type) { case XFS_BMAP_MAP: + if (bi->bi_bmap.br_state == XFS_EXT_UNWRITTEN) + flags |= XFS_BMAPI_PREALLOC; error = xfs_bmapi_remap(tp, bi->bi_owner, bmap->br_startoff, - bmap->br_blockcount, bmap->br_startblock, 0); + bmap->br_blockcount, bmap->br_startblock, + flags); bmap->br_blockcount = 0; break; case XFS_BMAP_UNMAP: error = __xfs_bunmapi(tp, bi->bi_owner, bmap->br_startoff, - &bmap->br_blockcount, XFS_BMAPI_REMAP, 1); + &bmap->br_blockcount, flags | XFS_BMAPI_REMAP, + 1); break; default: ASSERT(0); + xfs_bmap_mark_sick(bi->bi_owner, bi->bi_whichfork); error = -EFSCORRUPTED; } @@ -6257,7 +6341,7 @@ xfs_bunmapi_range( xfs_filblks_t unmap_len = endoff - startoff + 1; int error = 0; - ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL)); + xfs_assert_ilocked(ip, XFS_ILOCK_EXCL); while (unmap_len > 0) { ASSERT((*tpp)->t_highest_agno == NULLAGNUMBER); @@ -6274,3 +6358,46 @@ xfs_bunmapi_range( out: return error; } + +struct xfs_bmap_query_range { + xfs_bmap_query_range_fn fn; + void *priv; +}; + +/* Format btree record and pass to our callback. */ +STATIC int +xfs_bmap_query_range_helper( + struct xfs_btree_cur *cur, + const union xfs_btree_rec *rec, + void *priv) +{ + struct xfs_bmap_query_range *query = priv; + struct xfs_bmbt_irec irec; + xfs_failaddr_t fa; + + xfs_bmbt_disk_get_all(&rec->bmbt, &irec); + fa = xfs_bmap_validate_extent(cur->bc_ino.ip, cur->bc_ino.whichfork, + &irec); + if (fa) { + xfs_btree_mark_sick(cur); + return xfs_bmap_complain_bad_rec(cur->bc_ino.ip, + cur->bc_ino.whichfork, fa, &irec); + } + + return query->fn(cur, &irec, query->priv); +} + +/* Find all bmaps. */ +int +xfs_bmap_query_all( + struct xfs_btree_cur *cur, + xfs_bmap_query_range_fn fn, + void *priv) +{ + struct xfs_bmap_query_range query = { + .priv = priv, + .fn = fn, + }; + + return xfs_btree_query_all(cur, xfs_bmap_query_range_helper, &query); +} diff --git a/fs/xfs/libxfs/xfs_bmap.h b/fs/xfs/libxfs/xfs_bmap.h index f6b73f1bad5f..f7662595309d 100644 --- a/fs/xfs/libxfs/xfs_bmap.h +++ b/fs/xfs/libxfs/xfs_bmap.h @@ -232,6 +232,10 @@ enum xfs_bmap_intent_type { XFS_BMAP_UNMAP, }; +#define XFS_BMAP_INTENT_STRINGS \ + { XFS_BMAP_MAP, "map" }, \ + { XFS_BMAP_UNMAP, "unmap" } + struct xfs_bmap_intent { struct list_head bi_list; enum xfs_bmap_intent_type bi_type; @@ -241,14 +245,11 @@ struct xfs_bmap_intent { struct xfs_bmbt_irec bi_bmap; }; -void xfs_bmap_update_get_group(struct xfs_mount *mp, - struct xfs_bmap_intent *bi); - int xfs_bmap_finish_one(struct xfs_trans *tp, struct xfs_bmap_intent *bi); void xfs_bmap_map_extent(struct xfs_trans *tp, struct xfs_inode *ip, - struct xfs_bmbt_irec *imap); + int whichfork, struct xfs_bmbt_irec *imap); void xfs_bmap_unmap_extent(struct xfs_trans *tp, struct xfs_inode *ip, - struct xfs_bmbt_irec *imap); + int whichfork, struct xfs_bmbt_irec *imap); static inline uint32_t xfs_bmap_fork_to_state(int whichfork) { @@ -280,4 +281,12 @@ extern struct kmem_cache *xfs_bmap_intent_cache; int __init xfs_bmap_intent_init_cache(void); void xfs_bmap_intent_destroy_cache(void); +typedef int (*xfs_bmap_query_range_fn)( + struct xfs_btree_cur *cur, + struct xfs_bmbt_irec *rec, + void *priv); + +int xfs_bmap_query_all(struct xfs_btree_cur *cur, xfs_bmap_query_range_fn fn, + void *priv); + #endif /* __XFS_BMAP_H__ */ diff --git a/fs/xfs/libxfs/xfs_bmap_btree.c b/fs/xfs/libxfs/xfs_bmap_btree.c index 71f2d50f7823..f5d84dcb58da 100644 --- a/fs/xfs/libxfs/xfs_bmap_btree.c +++ b/fs/xfs/libxfs/xfs_bmap_btree.c @@ -26,6 +26,22 @@ static struct kmem_cache *xfs_bmbt_cur_cache; +void +xfs_bmbt_init_block( + struct xfs_inode *ip, + struct xfs_btree_block *buf, + struct xfs_buf *bp, + __u16 level, + __u16 numrecs) +{ + if (bp) + xfs_btree_init_buf(ip->i_mount, bp, &xfs_bmbt_ops, level, + numrecs, ip->i_ino); + else + xfs_btree_init_block(ip->i_mount, buf, &xfs_bmbt_ops, level, + numrecs, ip->i_ino); +} + /* * Convert on-disk form of btree root to in-memory form. */ @@ -44,9 +60,7 @@ xfs_bmdr_to_bmbt( xfs_bmbt_key_t *tkp; __be64 *tpp; - xfs_btree_init_block_int(mp, rblock, XFS_BUF_DADDR_NULL, - XFS_BTNUM_BMAP, 0, 0, ip->i_ino, - XFS_BTREE_LONG_PTRS); + xfs_bmbt_init_block(ip, rblock, NULL, 0, 0); rblock->bb_level = dblock->bb_level; ASSERT(be16_to_cpu(rblock->bb_level) > 0); rblock->bb_numrecs = dblock->bb_numrecs; @@ -171,13 +185,8 @@ xfs_bmbt_dup_cursor( new = xfs_bmbt_init_cursor(cur->bc_mp, cur->bc_tp, cur->bc_ino.ip, cur->bc_ino.whichfork); - - /* - * Copy the firstblock, dfops, and flags values, - * since init cursor doesn't get them. - */ - new->bc_ino.flags = cur->bc_ino.flags; - + new->bc_flags |= (cur->bc_flags & + (XFS_BTREE_BMBT_INVALID_OWNER | XFS_BTREE_BMBT_WASDEL)); return new; } @@ -189,10 +198,10 @@ xfs_bmbt_update_cursor( ASSERT((dst->bc_tp->t_highest_agno != NULLAGNUMBER) || (dst->bc_ino.ip->i_diflags & XFS_DIFLAG_REALTIME)); - dst->bc_ino.allocated += src->bc_ino.allocated; + dst->bc_bmap.allocated += src->bc_bmap.allocated; dst->bc_tp->t_highest_agno = src->bc_tp->t_highest_agno; - src->bc_ino.allocated = 0; + src->bc_bmap.allocated = 0; } STATIC int @@ -211,7 +220,7 @@ xfs_bmbt_alloc_block( xfs_rmap_ino_bmbt_owner(&args.oinfo, cur->bc_ino.ip->i_ino, cur->bc_ino.whichfork); args.minlen = args.maxlen = args.prod = 1; - args.wasdel = cur->bc_ino.flags & XFS_BTCUR_BMBT_WASDEL; + args.wasdel = cur->bc_flags & XFS_BTREE_BMBT_WASDEL; if (!args.wasdel && args.tp->t_blk_res == 0) return -ENOSPC; @@ -247,7 +256,7 @@ xfs_bmbt_alloc_block( } ASSERT(args.len == 1); - cur->bc_ino.allocated++; + cur->bc_bmap.allocated++; cur->bc_ino.ip->i_nblocks++; xfs_trans_log_inode(args.tp, cur->bc_ino.ip, XFS_ILOG_CORE); xfs_trans_mod_dquot_byino(args.tp, cur->bc_ino.ip, @@ -360,14 +369,6 @@ xfs_bmbt_init_rec_from_cur( xfs_bmbt_disk_set_all(&rec->bmbt, &cur->bc_rec.b); } -STATIC void -xfs_bmbt_init_ptr_from_cur( - struct xfs_btree_cur *cur, - union xfs_btree_ptr *ptr) -{ - ptr->l = 0; -} - STATIC int64_t xfs_bmbt_key_diff( struct xfs_btree_cur *cur, @@ -419,7 +420,7 @@ xfs_bmbt_verify( * XXX: need a better way of verifying the owner here. Right now * just make sure there has been one set. */ - fa = xfs_btree_lblock_v5hdr_verify(bp, XFS_RMAP_OWN_UNKNOWN); + fa = xfs_btree_fsblock_v5hdr_verify(bp, XFS_RMAP_OWN_UNKNOWN); if (fa) return fa; } @@ -435,7 +436,7 @@ xfs_bmbt_verify( if (level > max(mp->m_bm_maxlevels[0], mp->m_bm_maxlevels[1])) return __this_address; - return xfs_btree_lblock_verify(bp, mp->m_bmap_dmxr[level != 0]); + return xfs_btree_fsblock_verify(bp, mp->m_bmap_dmxr[level != 0]); } static void @@ -444,7 +445,7 @@ xfs_bmbt_read_verify( { xfs_failaddr_t fa; - if (!xfs_btree_lblock_verify_crc(bp)) + if (!xfs_btree_fsblock_verify_crc(bp)) xfs_verifier_error(bp, -EFSBADCRC, __this_address); else { fa = xfs_bmbt_verify(bp); @@ -468,7 +469,7 @@ xfs_bmbt_write_verify( xfs_verifier_error(bp, -EFSCORRUPTED, fa); return; } - xfs_btree_lblock_calc_crc(bp); + xfs_btree_fsblock_calc_crc(bp); } const struct xfs_buf_ops xfs_bmbt_buf_ops = { @@ -515,9 +516,16 @@ xfs_bmbt_keys_contiguous( be64_to_cpu(key2->bmbt.br_startoff)); } -static const struct xfs_btree_ops xfs_bmbt_ops = { +const struct xfs_btree_ops xfs_bmbt_ops = { + .name = "bmap", + .type = XFS_BTREE_TYPE_INODE, + .rec_len = sizeof(xfs_bmbt_rec_t), .key_len = sizeof(xfs_bmbt_key_t), + .ptr_len = XFS_BTREE_LONG_PTR_LEN, + + .lru_refs = XFS_BMAP_BTREE_REF, + .statoff = XFS_STATS_CALC_INDEX(xs_bmbt_2), .dup_cursor = xfs_bmbt_dup_cursor, .update_cursor = xfs_bmbt_update_cursor, @@ -529,7 +537,6 @@ static const struct xfs_btree_ops xfs_bmbt_ops = { .init_key_from_rec = xfs_bmbt_init_key_from_rec, .init_high_key_from_rec = xfs_bmbt_init_high_key_from_rec, .init_rec_from_cur = xfs_bmbt_init_rec_from_cur, - .init_ptr_from_cur = xfs_bmbt_init_ptr_from_cur, .key_diff = xfs_bmbt_key_diff, .diff_two_keys = xfs_bmbt_diff_two_keys, .buf_ops = &xfs_bmbt_buf_ops, @@ -538,35 +545,10 @@ static const struct xfs_btree_ops xfs_bmbt_ops = { .keys_contiguous = xfs_bmbt_keys_contiguous, }; -static struct xfs_btree_cur * -xfs_bmbt_init_common( - struct xfs_mount *mp, - struct xfs_trans *tp, - struct xfs_inode *ip, - int whichfork) -{ - struct xfs_btree_cur *cur; - - ASSERT(whichfork != XFS_COW_FORK); - - cur = xfs_btree_alloc_cursor(mp, tp, XFS_BTNUM_BMAP, - mp->m_bm_maxlevels[whichfork], xfs_bmbt_cur_cache); - cur->bc_statoff = XFS_STATS_CALC_INDEX(xs_bmbt_2); - - cur->bc_ops = &xfs_bmbt_ops; - cur->bc_flags = XFS_BTREE_LONG_PTRS | XFS_BTREE_ROOT_IN_INODE; - if (xfs_has_crc(mp)) - cur->bc_flags |= XFS_BTREE_CRC_BLOCKS; - - cur->bc_ino.ip = ip; - cur->bc_ino.allocated = 0; - cur->bc_ino.flags = 0; - - return cur; -} - /* - * Allocate a new bmap btree cursor. + * Create a new bmap btree cursor. + * + * For staging cursors -1 in passed in whichfork. */ struct xfs_btree_cur * xfs_bmbt_init_cursor( @@ -575,15 +557,34 @@ xfs_bmbt_init_cursor( struct xfs_inode *ip, int whichfork) { - struct xfs_ifork *ifp = xfs_ifork_ptr(ip, whichfork); struct xfs_btree_cur *cur; + unsigned int maxlevels; - cur = xfs_bmbt_init_common(mp, tp, ip, whichfork); + ASSERT(whichfork != XFS_COW_FORK); - cur->bc_nlevels = be16_to_cpu(ifp->if_broot->bb_level) + 1; - cur->bc_ino.forksize = xfs_inode_fork_size(ip, whichfork); + /* + * The Data fork always has larger maxlevel, so use that for staging + * cursors. + */ + switch (whichfork) { + case XFS_STAGING_FORK: + maxlevels = mp->m_bm_maxlevels[XFS_DATA_FORK]; + break; + default: + maxlevels = mp->m_bm_maxlevels[whichfork]; + break; + } + cur = xfs_btree_alloc_cursor(mp, tp, &xfs_bmbt_ops, maxlevels, + xfs_bmbt_cur_cache); + cur->bc_ino.ip = ip; cur->bc_ino.whichfork = whichfork; + cur->bc_bmap.allocated = 0; + if (whichfork != XFS_STAGING_FORK) { + struct xfs_ifork *ifp = xfs_ifork_ptr(ip, whichfork); + cur->bc_nlevels = be16_to_cpu(ifp->if_broot->bb_level) + 1; + cur->bc_ino.forksize = xfs_inode_fork_size(ip, whichfork); + } return cur; } @@ -599,33 +600,6 @@ xfs_bmbt_block_maxrecs( } /* - * Allocate a new bmap btree cursor for reloading an inode block mapping data - * structure. Note that callers can use the staged cursor to reload extents - * format inode forks if they rebuild the iext tree and commit the staged - * cursor immediately. - */ -struct xfs_btree_cur * -xfs_bmbt_stage_cursor( - struct xfs_mount *mp, - struct xfs_inode *ip, - struct xbtree_ifakeroot *ifake) -{ - struct xfs_btree_cur *cur; - struct xfs_btree_ops *ops; - - /* data fork always has larger maxheight */ - cur = xfs_bmbt_init_common(mp, NULL, ip, XFS_DATA_FORK); - cur->bc_nlevels = ifake->if_levels; - cur->bc_ino.forksize = ifake->if_fork_size; - - /* Don't let anyone think we're attached to the real fork yet. */ - cur->bc_ino.whichfork = -1; - xfs_btree_stage_ifakeroot(cur, ifake, &ops); - ops->update_cursor = NULL; - return cur; -} - -/* * Swap in the new inode fork root. Once we pass this point the newly rebuilt * mappings are in place and we have to kill off any old btree blocks. */ @@ -665,7 +639,7 @@ xfs_bmbt_commit_staged_btree( break; } xfs_trans_log_inode(tp, cur->bc_ino.ip, flags); - xfs_btree_commit_ifakeroot(cur, tp, whichfork, &xfs_bmbt_ops); + xfs_btree_commit_ifakeroot(cur, tp, whichfork); } /* @@ -751,7 +725,7 @@ xfs_bmbt_change_owner( ASSERT(xfs_ifork_ptr(ip, whichfork)->if_format == XFS_DINODE_FMT_BTREE); cur = xfs_bmbt_init_cursor(ip->i_mount, tp, ip, whichfork); - cur->bc_ino.flags |= XFS_BTCUR_BMBT_INVALID_OWNER; + cur->bc_flags |= XFS_BTREE_BMBT_INVALID_OWNER; error = xfs_btree_change_owner(cur, new_owner, buffer_list); xfs_btree_del_cursor(cur, error); diff --git a/fs/xfs/libxfs/xfs_bmap_btree.h b/fs/xfs/libxfs/xfs_bmap_btree.h index 151b8491f60e..de1b73f1225c 100644 --- a/fs/xfs/libxfs/xfs_bmap_btree.h +++ b/fs/xfs/libxfs/xfs_bmap_btree.h @@ -107,8 +107,6 @@ extern int xfs_bmbt_change_owner(struct xfs_trans *tp, struct xfs_inode *ip, extern struct xfs_btree_cur *xfs_bmbt_init_cursor(struct xfs_mount *, struct xfs_trans *, struct xfs_inode *, int); -struct xfs_btree_cur *xfs_bmbt_stage_cursor(struct xfs_mount *mp, - struct xfs_inode *ip, struct xbtree_ifakeroot *ifake); void xfs_bmbt_commit_staged_btree(struct xfs_btree_cur *cur, struct xfs_trans *tp, int whichfork); @@ -120,4 +118,7 @@ unsigned int xfs_bmbt_maxlevels_ondisk(void); int __init xfs_bmbt_init_cur_cache(void); void xfs_bmbt_destroy_cur_cache(void); +void xfs_bmbt_init_block(struct xfs_inode *ip, struct xfs_btree_block *buf, + struct xfs_buf *bp, __u16 level, __u16 numrecs); + #endif /* __XFS_BMAP_BTREE_H__ */ diff --git a/fs/xfs/libxfs/xfs_btree.c b/fs/xfs/libxfs/xfs_btree.c index ea8d3659df20..d29547572a68 100644 --- a/fs/xfs/libxfs/xfs_btree.c +++ b/fs/xfs/libxfs/xfs_btree.c @@ -27,28 +27,24 @@ #include "xfs_bmap_btree.h" #include "xfs_rmap_btree.h" #include "xfs_refcount_btree.h" +#include "xfs_health.h" +#include "xfs_buf_mem.h" +#include "xfs_btree_mem.h" /* * Btree magic numbers. */ -static const uint32_t xfs_magics[2][XFS_BTNUM_MAX] = { - { XFS_ABTB_MAGIC, XFS_ABTC_MAGIC, 0, XFS_BMAP_MAGIC, XFS_IBT_MAGIC, - XFS_FIBT_MAGIC, 0 }, - { XFS_ABTB_CRC_MAGIC, XFS_ABTC_CRC_MAGIC, XFS_RMAP_CRC_MAGIC, - XFS_BMAP_CRC_MAGIC, XFS_IBT_CRC_MAGIC, XFS_FIBT_CRC_MAGIC, - XFS_REFC_CRC_MAGIC } -}; - uint32_t xfs_btree_magic( - int crc, - xfs_btnum_t btnum) + struct xfs_mount *mp, + const struct xfs_btree_ops *ops) { - uint32_t magic = xfs_magics[crc][btnum]; + int idx = xfs_has_crc(mp) ? 1 : 0; + __be32 magic = ops->buf_ops->magic[idx]; /* Ensure we asked for crc for crc-only magics. */ ASSERT(magic != 0); - return magic; + return be32_to_cpu(magic); } /* @@ -63,10 +59,8 @@ xfs_btree_magic( * bytes. */ static inline xfs_failaddr_t -xfs_btree_check_lblock_siblings( +xfs_btree_check_fsblock_siblings( struct xfs_mount *mp, - struct xfs_btree_cur *cur, - int level, xfs_fsblock_t fsb, __be64 dsibling) { @@ -78,22 +72,33 @@ xfs_btree_check_lblock_siblings( sibling = be64_to_cpu(dsibling); if (sibling == fsb) return __this_address; - if (level >= 0) { - if (!xfs_btree_check_lptr(cur, sibling, level + 1)) - return __this_address; - } else { - if (!xfs_verify_fsbno(mp, sibling)) - return __this_address; - } + if (!xfs_verify_fsbno(mp, sibling)) + return __this_address; + return NULL; +} +static inline xfs_failaddr_t +xfs_btree_check_memblock_siblings( + struct xfs_buftarg *btp, + xfbno_t bno, + __be64 dsibling) +{ + xfbno_t sibling; + + if (dsibling == cpu_to_be64(NULLFSBLOCK)) + return NULL; + + sibling = be64_to_cpu(dsibling); + if (sibling == bno) + return __this_address; + if (!xmbuf_verify_daddr(btp, xfbno_to_daddr(sibling))) + return __this_address; return NULL; } static inline xfs_failaddr_t -xfs_btree_check_sblock_siblings( +xfs_btree_check_agblock_siblings( struct xfs_perag *pag, - struct xfs_btree_cur *cur, - int level, xfs_agblock_t agbno, __be32 dsibling) { @@ -105,34 +110,21 @@ xfs_btree_check_sblock_siblings( sibling = be32_to_cpu(dsibling); if (sibling == agbno) return __this_address; - if (level >= 0) { - if (!xfs_btree_check_sptr(cur, sibling, level + 1)) - return __this_address; - } else { - if (!xfs_verify_agbno(pag, sibling)) - return __this_address; - } + if (!xfs_verify_agbno(pag, sibling)) + return __this_address; return NULL; } -/* - * Check a long btree block header. Return the address of the failing check, - * or NULL if everything is ok. - */ -xfs_failaddr_t -__xfs_btree_check_lblock( +static xfs_failaddr_t +__xfs_btree_check_lblock_hdr( struct xfs_btree_cur *cur, struct xfs_btree_block *block, int level, struct xfs_buf *bp) { struct xfs_mount *mp = cur->bc_mp; - xfs_btnum_t btnum = cur->bc_btnum; - int crc = xfs_has_crc(mp); - xfs_failaddr_t fa; - xfs_fsblock_t fsb = NULLFSBLOCK; - if (crc) { + if (xfs_has_crc(mp)) { if (!uuid_equal(&block->bb_u.l.bb_uuid, &mp->m_sb.sb_meta_uuid)) return __this_address; if (block->bb_u.l.bb_blkno != @@ -142,7 +134,7 @@ __xfs_btree_check_lblock( return __this_address; } - if (be32_to_cpu(block->bb_magic) != xfs_btree_magic(crc, btnum)) + if (be32_to_cpu(block->bb_magic) != xfs_btree_magic(mp, cur->bc_ops)) return __this_address; if (be16_to_cpu(block->bb_level) != level) return __this_address; @@ -150,44 +142,83 @@ __xfs_btree_check_lblock( cur->bc_ops->get_maxrecs(cur, level)) return __this_address; - if (bp) - fsb = XFS_DADDR_TO_FSB(mp, xfs_buf_daddr(bp)); + return NULL; +} + +/* + * Check a long btree block header. Return the address of the failing check, + * or NULL if everything is ok. + */ +static xfs_failaddr_t +__xfs_btree_check_fsblock( + struct xfs_btree_cur *cur, + struct xfs_btree_block *block, + int level, + struct xfs_buf *bp) +{ + struct xfs_mount *mp = cur->bc_mp; + xfs_failaddr_t fa; + xfs_fsblock_t fsb; + + fa = __xfs_btree_check_lblock_hdr(cur, block, level, bp); + if (fa) + return fa; - fa = xfs_btree_check_lblock_siblings(mp, cur, level, fsb, + /* + * For inode-rooted btrees, the root block sits in the inode fork. In + * that case bp is NULL, and the block must not have any siblings. + */ + if (!bp) { + if (block->bb_u.l.bb_leftsib != cpu_to_be64(NULLFSBLOCK)) + return __this_address; + if (block->bb_u.l.bb_rightsib != cpu_to_be64(NULLFSBLOCK)) + return __this_address; + return NULL; + } + + fsb = XFS_DADDR_TO_FSB(mp, xfs_buf_daddr(bp)); + fa = xfs_btree_check_fsblock_siblings(mp, fsb, block->bb_u.l.bb_leftsib); if (!fa) - fa = xfs_btree_check_lblock_siblings(mp, cur, level, fsb, + fa = xfs_btree_check_fsblock_siblings(mp, fsb, block->bb_u.l.bb_rightsib); return fa; } -/* Check a long btree block header. */ -static int -xfs_btree_check_lblock( +/* + * Check an in-memory btree block header. Return the address of the failing + * check, or NULL if everything is ok. + */ +static xfs_failaddr_t +__xfs_btree_check_memblock( struct xfs_btree_cur *cur, struct xfs_btree_block *block, int level, struct xfs_buf *bp) { - struct xfs_mount *mp = cur->bc_mp; + struct xfs_buftarg *btp = cur->bc_mem.xfbtree->target; xfs_failaddr_t fa; + xfbno_t bno; - fa = __xfs_btree_check_lblock(cur, block, level, bp); - if (XFS_IS_CORRUPT(mp, fa != NULL) || - XFS_TEST_ERROR(false, mp, XFS_ERRTAG_BTREE_CHECK_LBLOCK)) { - if (bp) - trace_xfs_btree_corrupt(bp, _RET_IP_); - return -EFSCORRUPTED; - } - return 0; + fa = __xfs_btree_check_lblock_hdr(cur, block, level, bp); + if (fa) + return fa; + + bno = xfs_daddr_to_xfbno(xfs_buf_daddr(bp)); + fa = xfs_btree_check_memblock_siblings(btp, bno, + block->bb_u.l.bb_leftsib); + if (!fa) + fa = xfs_btree_check_memblock_siblings(btp, bno, + block->bb_u.l.bb_rightsib); + return fa; } /* * Check a short btree block header. Return the address of the failing check, * or NULL if everything is ok. */ -xfs_failaddr_t -__xfs_btree_check_sblock( +static xfs_failaddr_t +__xfs_btree_check_agblock( struct xfs_btree_cur *cur, struct xfs_btree_block *block, int level, @@ -195,20 +226,17 @@ __xfs_btree_check_sblock( { struct xfs_mount *mp = cur->bc_mp; struct xfs_perag *pag = cur->bc_ag.pag; - xfs_btnum_t btnum = cur->bc_btnum; - int crc = xfs_has_crc(mp); xfs_failaddr_t fa; - xfs_agblock_t agbno = NULLAGBLOCK; + xfs_agblock_t agbno; - if (crc) { + if (xfs_has_crc(mp)) { if (!uuid_equal(&block->bb_u.s.bb_uuid, &mp->m_sb.sb_meta_uuid)) return __this_address; - if (block->bb_u.s.bb_blkno != - cpu_to_be64(bp ? xfs_buf_daddr(bp) : XFS_BUF_DADDR_NULL)) + if (block->bb_u.s.bb_blkno != cpu_to_be64(xfs_buf_daddr(bp))) return __this_address; } - if (be32_to_cpu(block->bb_magic) != xfs_btree_magic(crc, btnum)) + if (be32_to_cpu(block->bb_magic) != xfs_btree_magic(mp, cur->bc_ops)) return __this_address; if (be16_to_cpu(block->bb_level) != level) return __this_address; @@ -216,36 +244,45 @@ __xfs_btree_check_sblock( cur->bc_ops->get_maxrecs(cur, level)) return __this_address; - if (bp) - agbno = xfs_daddr_to_agbno(mp, xfs_buf_daddr(bp)); - - fa = xfs_btree_check_sblock_siblings(pag, cur, level, agbno, + agbno = xfs_daddr_to_agbno(mp, xfs_buf_daddr(bp)); + fa = xfs_btree_check_agblock_siblings(pag, agbno, block->bb_u.s.bb_leftsib); if (!fa) - fa = xfs_btree_check_sblock_siblings(pag, cur, level, agbno, + fa = xfs_btree_check_agblock_siblings(pag, agbno, block->bb_u.s.bb_rightsib); return fa; } -/* Check a short btree block header. */ -STATIC int -xfs_btree_check_sblock( +/* + * Internal btree block check. + * + * Return NULL if the block is ok or the address of the failed check otherwise. + */ +xfs_failaddr_t +__xfs_btree_check_block( struct xfs_btree_cur *cur, struct xfs_btree_block *block, int level, struct xfs_buf *bp) { - struct xfs_mount *mp = cur->bc_mp; - xfs_failaddr_t fa; - - fa = __xfs_btree_check_sblock(cur, block, level, bp); - if (XFS_IS_CORRUPT(mp, fa != NULL) || - XFS_TEST_ERROR(false, mp, XFS_ERRTAG_BTREE_CHECK_SBLOCK)) { - if (bp) - trace_xfs_btree_corrupt(bp, _RET_IP_); - return -EFSCORRUPTED; + switch (cur->bc_ops->type) { + case XFS_BTREE_TYPE_MEM: + return __xfs_btree_check_memblock(cur, block, level, bp); + case XFS_BTREE_TYPE_AG: + return __xfs_btree_check_agblock(cur, block, level, bp); + case XFS_BTREE_TYPE_INODE: + return __xfs_btree_check_fsblock(cur, block, level, bp); + default: + ASSERT(0); + return __this_address; } - return 0; +} + +static inline unsigned int xfs_btree_block_errtag(struct xfs_btree_cur *cur) +{ + if (cur->bc_ops->ptr_len == XFS_BTREE_SHORT_PTR_LEN) + return XFS_ERRTAG_BTREE_CHECK_SBLOCK; + return XFS_ERRTAG_BTREE_CHECK_LBLOCK; } /* @@ -258,34 +295,49 @@ xfs_btree_check_block( int level, /* level of the btree block */ struct xfs_buf *bp) /* buffer containing block, if any */ { - if (cur->bc_flags & XFS_BTREE_LONG_PTRS) - return xfs_btree_check_lblock(cur, block, level, bp); - else - return xfs_btree_check_sblock(cur, block, level, bp); -} + struct xfs_mount *mp = cur->bc_mp; + xfs_failaddr_t fa; -/* Check that this long pointer is valid and points within the fs. */ -bool -xfs_btree_check_lptr( - struct xfs_btree_cur *cur, - xfs_fsblock_t fsbno, - int level) -{ - if (level <= 0) - return false; - return xfs_verify_fsbno(cur->bc_mp, fsbno); + fa = __xfs_btree_check_block(cur, block, level, bp); + if (XFS_IS_CORRUPT(mp, fa != NULL) || + XFS_TEST_ERROR(false, mp, xfs_btree_block_errtag(cur))) { + if (bp) + trace_xfs_btree_corrupt(bp, _RET_IP_); + xfs_btree_mark_sick(cur); + return -EFSCORRUPTED; + } + return 0; } -/* Check that this short pointer is valid and points within the AG. */ -bool -xfs_btree_check_sptr( - struct xfs_btree_cur *cur, - xfs_agblock_t agbno, - int level) +int +__xfs_btree_check_ptr( + struct xfs_btree_cur *cur, + const union xfs_btree_ptr *ptr, + int index, + int level) { if (level <= 0) - return false; - return xfs_verify_agbno(cur->bc_ag.pag, agbno); + return -EFSCORRUPTED; + + switch (cur->bc_ops->type) { + case XFS_BTREE_TYPE_MEM: + if (!xfbtree_verify_bno(cur->bc_mem.xfbtree, + be64_to_cpu((&ptr->l)[index]))) + return -EFSCORRUPTED; + break; + case XFS_BTREE_TYPE_INODE: + if (!xfs_verify_fsbno(cur->bc_mp, + be64_to_cpu((&ptr->l)[index]))) + return -EFSCORRUPTED; + break; + case XFS_BTREE_TYPE_AG: + if (!xfs_verify_agbno(cur->bc_ag.pag, + be32_to_cpu((&ptr->s)[index]))) + return -EFSCORRUPTED; + break; + } + + return 0; } /* @@ -299,26 +351,35 @@ xfs_btree_check_ptr( int index, int level) { - if (cur->bc_flags & XFS_BTREE_LONG_PTRS) { - if (xfs_btree_check_lptr(cur, be64_to_cpu((&ptr->l)[index]), - level)) - return 0; - xfs_err(cur->bc_mp, -"Inode %llu fork %d: Corrupt btree %d pointer at level %d index %d.", + int error; + + error = __xfs_btree_check_ptr(cur, ptr, index, level); + if (error) { + switch (cur->bc_ops->type) { + case XFS_BTREE_TYPE_MEM: + xfs_err(cur->bc_mp, +"In-memory: Corrupt %sbt flags 0x%x pointer at level %d index %d fa %pS.", + cur->bc_ops->name, cur->bc_flags, level, index, + __this_address); + break; + case XFS_BTREE_TYPE_INODE: + xfs_err(cur->bc_mp, +"Inode %llu fork %d: Corrupt %sbt pointer at level %d index %d.", cur->bc_ino.ip->i_ino, - cur->bc_ino.whichfork, cur->bc_btnum, + cur->bc_ino.whichfork, cur->bc_ops->name, level, index); - } else { - if (xfs_btree_check_sptr(cur, be32_to_cpu((&ptr->s)[index]), - level)) - return 0; - xfs_err(cur->bc_mp, -"AG %u: Corrupt btree %d pointer at level %d index %d.", - cur->bc_ag.pag->pag_agno, cur->bc_btnum, + break; + case XFS_BTREE_TYPE_AG: + xfs_err(cur->bc_mp, +"AG %u: Corrupt %sbt pointer at level %d index %d.", + cur->bc_ag.pag->pag_agno, cur->bc_ops->name, level, index); + break; + } + xfs_btree_mark_sick(cur); } - return -EFSCORRUPTED; + return error; } #ifdef DEBUG @@ -336,7 +397,7 @@ xfs_btree_check_ptr( * it to disk. */ void -xfs_btree_lblock_calc_crc( +xfs_btree_fsblock_calc_crc( struct xfs_buf *bp) { struct xfs_btree_block *block = XFS_BUF_TO_BLOCK(bp); @@ -350,7 +411,7 @@ xfs_btree_lblock_calc_crc( } bool -xfs_btree_lblock_verify_crc( +xfs_btree_fsblock_verify_crc( struct xfs_buf *bp) { struct xfs_btree_block *block = XFS_BUF_TO_BLOCK(bp); @@ -374,7 +435,7 @@ xfs_btree_lblock_verify_crc( * it to disk. */ void -xfs_btree_sblock_calc_crc( +xfs_btree_agblock_calc_crc( struct xfs_buf *bp) { struct xfs_btree_block *block = XFS_BUF_TO_BLOCK(bp); @@ -388,7 +449,7 @@ xfs_btree_sblock_calc_crc( } bool -xfs_btree_sblock_verify_crc( +xfs_btree_agblock_verify_crc( struct xfs_buf *bp) { struct xfs_btree_block *block = XFS_BUF_TO_BLOCK(bp); @@ -410,6 +471,17 @@ xfs_btree_free_block( { int error; + trace_xfs_btree_free_block(cur, bp); + + /* + * Don't allow block freeing for a staging cursor, because staging + * cursors do not support regular btree modifications. + */ + if (unlikely(cur->bc_flags & XFS_BTREE_STAGING)) { + ASSERT(0); + return -EFSCORRUPTED; + } + error = cur->bc_ops->free_block(cur, bp); if (!error) { xfs_trans_binval(cur->bc_tp, bp); @@ -448,33 +520,70 @@ xfs_btree_del_cursor( * zero, then we should be shut down or on our way to shutdown due to * cancelling a dirty transaction on error. */ - ASSERT(cur->bc_btnum != XFS_BTNUM_BMAP || cur->bc_ino.allocated == 0 || + ASSERT(!xfs_btree_is_bmap(cur->bc_ops) || cur->bc_bmap.allocated == 0 || xfs_is_shutdown(cur->bc_mp) || error != 0); - if (unlikely(cur->bc_flags & XFS_BTREE_STAGING)) - kmem_free(cur->bc_ops); - if (!(cur->bc_flags & XFS_BTREE_LONG_PTRS) && cur->bc_ag.pag) - xfs_perag_put(cur->bc_ag.pag); + + switch (cur->bc_ops->type) { + case XFS_BTREE_TYPE_AG: + if (cur->bc_ag.pag) + xfs_perag_put(cur->bc_ag.pag); + break; + case XFS_BTREE_TYPE_INODE: + /* nothing to do */ + break; + case XFS_BTREE_TYPE_MEM: + if (cur->bc_mem.pag) + xfs_perag_put(cur->bc_mem.pag); + break; + } + kmem_cache_free(cur->bc_cache, cur); } +/* Return the buffer target for this btree's buffer. */ +static inline struct xfs_buftarg * +xfs_btree_buftarg( + struct xfs_btree_cur *cur) +{ + if (cur->bc_ops->type == XFS_BTREE_TYPE_MEM) + return cur->bc_mem.xfbtree->target; + return cur->bc_mp->m_ddev_targp; +} + +/* Return the block size (in units of 512b sectors) for this btree. */ +static inline unsigned int +xfs_btree_bbsize( + struct xfs_btree_cur *cur) +{ + if (cur->bc_ops->type == XFS_BTREE_TYPE_MEM) + return XFBNO_BBSIZE; + return cur->bc_mp->m_bsize; +} + /* * Duplicate the btree cursor. * Allocate a new one, copy the record, re-get the buffers. */ -int /* error */ +int /* error */ xfs_btree_dup_cursor( - struct xfs_btree_cur *cur, /* input cursor */ - struct xfs_btree_cur **ncur) /* output cursor */ + struct xfs_btree_cur *cur, /* input cursor */ + struct xfs_btree_cur **ncur) /* output cursor */ { - struct xfs_buf *bp; /* btree block's buffer pointer */ - int error; /* error return value */ - int i; /* level number of btree block */ - xfs_mount_t *mp; /* mount structure for filesystem */ - struct xfs_btree_cur *new; /* new cursor value */ - xfs_trans_t *tp; /* transaction pointer, can be NULL */ + struct xfs_mount *mp = cur->bc_mp; + struct xfs_trans *tp = cur->bc_tp; + struct xfs_buf *bp; + struct xfs_btree_cur *new; + int error; + int i; - tp = cur->bc_tp; - mp = cur->bc_mp; + /* + * Don't allow staging cursors to be duplicated because they're supposed + * to be kept private to a single thread. + */ + if (unlikely(cur->bc_flags & XFS_BTREE_STAGING)) { + ASSERT(0); + return -EFSCORRUPTED; + } /* * Allocate a new cursor like the old one. @@ -494,10 +603,13 @@ xfs_btree_dup_cursor( new->bc_levels[i].ra = cur->bc_levels[i].ra; bp = cur->bc_levels[i].bp; if (bp) { - error = xfs_trans_read_buf(mp, tp, mp->m_ddev_targp, - xfs_buf_daddr(bp), mp->m_bsize, - 0, &bp, - cur->bc_ops->buf_ops); + error = xfs_trans_read_buf(mp, tp, + xfs_btree_buftarg(cur), + xfs_buf_daddr(bp), + xfs_btree_bbsize(cur), 0, &bp, + cur->bc_ops->buf_ops); + if (xfs_metadata_is_sick(error)) + xfs_btree_mark_sick(new); if (error) { xfs_btree_del_cursor(new, error); *ncur = NULL; @@ -539,7 +651,7 @@ xfs_btree_dup_cursor( * record, key or pointer (xfs_btree_*_addr). Note that all addressing * inside the btree block is done using indices starting at one, not zero! * - * If XFS_BTREE_OVERLAPPING is set, then this btree supports keys containing + * If XFS_BTGEO_OVERLAPPING is set, then this btree supports keys containing * overlapping intervals. In such a tree, records are still sorted lowest to * highest and indexed by the smallest key value that refers to the record. * However, nodes are different: each pointer has two associated keys -- one @@ -589,26 +701,17 @@ xfs_btree_dup_cursor( */ static inline size_t xfs_btree_block_len(struct xfs_btree_cur *cur) { - if (cur->bc_flags & XFS_BTREE_LONG_PTRS) { - if (cur->bc_flags & XFS_BTREE_CRC_BLOCKS) + if (cur->bc_ops->ptr_len == XFS_BTREE_LONG_PTR_LEN) { + if (xfs_has_crc(cur->bc_mp)) return XFS_BTREE_LBLOCK_CRC_LEN; return XFS_BTREE_LBLOCK_LEN; } - if (cur->bc_flags & XFS_BTREE_CRC_BLOCKS) + if (xfs_has_crc(cur->bc_mp)) return XFS_BTREE_SBLOCK_CRC_LEN; return XFS_BTREE_SBLOCK_LEN; } /* - * Return size of btree block pointers for this btree instance. - */ -static inline size_t xfs_btree_ptr_len(struct xfs_btree_cur *cur) -{ - return (cur->bc_flags & XFS_BTREE_LONG_PTRS) ? - sizeof(__be64) : sizeof(__be32); -} - -/* * Calculate offset of the n-th record in a btree block. */ STATIC size_t @@ -655,7 +758,7 @@ xfs_btree_ptr_offset( { return xfs_btree_block_len(cur) + cur->bc_ops->get_maxrecs(cur, level) * cur->bc_ops->key_len + - (n - 1) * xfs_btree_ptr_len(cur); + (n - 1) * cur->bc_ops->ptr_len; } /* @@ -718,7 +821,7 @@ struct xfs_ifork * xfs_btree_ifork_ptr( struct xfs_btree_cur *cur) { - ASSERT(cur->bc_flags & XFS_BTREE_ROOT_IN_INODE); + ASSERT(cur->bc_ops->type == XFS_BTREE_TYPE_INODE); if (cur->bc_flags & XFS_BTREE_STAGING) return cur->bc_ino.ifake->if_fork; @@ -750,8 +853,7 @@ xfs_btree_get_block( int level, /* level in btree */ struct xfs_buf **bpp) /* buffer containing the block */ { - if ((cur->bc_flags & XFS_BTREE_ROOT_IN_INODE) && - (level == cur->bc_nlevels - 1)) { + if (xfs_btree_at_iroot(cur, level)) { *bpp = NULL; return xfs_btree_get_iroot(cur); } @@ -856,95 +958,52 @@ xfs_btree_offsets( } } -/* - * Get a buffer for the block, return it read in. - * Long-form addressing. - */ -int -xfs_btree_read_bufl( - struct xfs_mount *mp, /* file system mount point */ - struct xfs_trans *tp, /* transaction pointer */ - xfs_fsblock_t fsbno, /* file system block number */ - struct xfs_buf **bpp, /* buffer for fsbno */ - int refval, /* ref count value for buffer */ - const struct xfs_buf_ops *ops) -{ - struct xfs_buf *bp; /* return value */ - xfs_daddr_t d; /* real disk block address */ - int error; - - if (!xfs_verify_fsbno(mp, fsbno)) - return -EFSCORRUPTED; - d = XFS_FSB_TO_DADDR(mp, fsbno); - error = xfs_trans_read_buf(mp, tp, mp->m_ddev_targp, d, - mp->m_bsize, 0, &bp, ops); - if (error) - return error; - if (bp) - xfs_buf_set_ref(bp, refval); - *bpp = bp; - return 0; -} - -/* - * Read-ahead the block, don't wait for it, don't return a buffer. - * Long-form addressing. - */ -/* ARGSUSED */ -void -xfs_btree_reada_bufl( - struct xfs_mount *mp, /* file system mount point */ - xfs_fsblock_t fsbno, /* file system block number */ - xfs_extlen_t count, /* count of filesystem blocks */ - const struct xfs_buf_ops *ops) +STATIC int +xfs_btree_readahead_fsblock( + struct xfs_btree_cur *cur, + int lr, + struct xfs_btree_block *block) { - xfs_daddr_t d; + struct xfs_mount *mp = cur->bc_mp; + xfs_fsblock_t left = be64_to_cpu(block->bb_u.l.bb_leftsib); + xfs_fsblock_t right = be64_to_cpu(block->bb_u.l.bb_rightsib); + int rval = 0; - ASSERT(fsbno != NULLFSBLOCK); - d = XFS_FSB_TO_DADDR(mp, fsbno); - xfs_buf_readahead(mp->m_ddev_targp, d, mp->m_bsize * count, ops); -} + if ((lr & XFS_BTCUR_LEFTRA) && left != NULLFSBLOCK) { + xfs_buf_readahead(mp->m_ddev_targp, XFS_FSB_TO_DADDR(mp, left), + mp->m_bsize, cur->bc_ops->buf_ops); + rval++; + } -/* - * Read-ahead the block, don't wait for it, don't return a buffer. - * Short-form addressing. - */ -/* ARGSUSED */ -void -xfs_btree_reada_bufs( - struct xfs_mount *mp, /* file system mount point */ - xfs_agnumber_t agno, /* allocation group number */ - xfs_agblock_t agbno, /* allocation group block number */ - xfs_extlen_t count, /* count of filesystem blocks */ - const struct xfs_buf_ops *ops) -{ - xfs_daddr_t d; + if ((lr & XFS_BTCUR_RIGHTRA) && right != NULLFSBLOCK) { + xfs_buf_readahead(mp->m_ddev_targp, XFS_FSB_TO_DADDR(mp, right), + mp->m_bsize, cur->bc_ops->buf_ops); + rval++; + } - ASSERT(agno != NULLAGNUMBER); - ASSERT(agbno != NULLAGBLOCK); - d = XFS_AGB_TO_DADDR(mp, agno, agbno); - xfs_buf_readahead(mp->m_ddev_targp, d, mp->m_bsize * count, ops); + return rval; } STATIC int -xfs_btree_readahead_lblock( +xfs_btree_readahead_memblock( struct xfs_btree_cur *cur, int lr, struct xfs_btree_block *block) { + struct xfs_buftarg *btp = cur->bc_mem.xfbtree->target; + xfbno_t left = be64_to_cpu(block->bb_u.l.bb_leftsib); + xfbno_t right = be64_to_cpu(block->bb_u.l.bb_rightsib); int rval = 0; - xfs_fsblock_t left = be64_to_cpu(block->bb_u.l.bb_leftsib); - xfs_fsblock_t right = be64_to_cpu(block->bb_u.l.bb_rightsib); if ((lr & XFS_BTCUR_LEFTRA) && left != NULLFSBLOCK) { - xfs_btree_reada_bufl(cur->bc_mp, left, 1, - cur->bc_ops->buf_ops); + xfs_buf_readahead(btp, xfbno_to_daddr(left), XFBNO_BBSIZE, + cur->bc_ops->buf_ops); rval++; } if ((lr & XFS_BTCUR_RIGHTRA) && right != NULLFSBLOCK) { - xfs_btree_reada_bufl(cur->bc_mp, right, 1, - cur->bc_ops->buf_ops); + xfs_buf_readahead(btp, xfbno_to_daddr(right), XFBNO_BBSIZE, + cur->bc_ops->buf_ops); rval++; } @@ -952,25 +1011,28 @@ xfs_btree_readahead_lblock( } STATIC int -xfs_btree_readahead_sblock( +xfs_btree_readahead_agblock( struct xfs_btree_cur *cur, int lr, - struct xfs_btree_block *block) + struct xfs_btree_block *block) { - int rval = 0; + struct xfs_mount *mp = cur->bc_mp; + xfs_agnumber_t agno = cur->bc_ag.pag->pag_agno; xfs_agblock_t left = be32_to_cpu(block->bb_u.s.bb_leftsib); xfs_agblock_t right = be32_to_cpu(block->bb_u.s.bb_rightsib); - + int rval = 0; if ((lr & XFS_BTCUR_LEFTRA) && left != NULLAGBLOCK) { - xfs_btree_reada_bufs(cur->bc_mp, cur->bc_ag.pag->pag_agno, - left, 1, cur->bc_ops->buf_ops); + xfs_buf_readahead(mp->m_ddev_targp, + XFS_AGB_TO_DADDR(mp, agno, left), + mp->m_bsize, cur->bc_ops->buf_ops); rval++; } if ((lr & XFS_BTCUR_RIGHTRA) && right != NULLAGBLOCK) { - xfs_btree_reada_bufs(cur->bc_mp, cur->bc_ag.pag->pag_agno, - right, 1, cur->bc_ops->buf_ops); + xfs_buf_readahead(mp->m_ddev_targp, + XFS_AGB_TO_DADDR(mp, agno, right), + mp->m_bsize, cur->bc_ops->buf_ops); rval++; } @@ -993,8 +1055,7 @@ xfs_btree_readahead( * No readahead needed if we are at the root level and the * btree root is stored in the inode. */ - if ((cur->bc_flags & XFS_BTREE_ROOT_IN_INODE) && - (lev == cur->bc_nlevels - 1)) + if (xfs_btree_at_iroot(cur, lev)) return 0; if ((cur->bc_levels[lev].ra | lr) == cur->bc_levels[lev].ra) @@ -1003,9 +1064,17 @@ xfs_btree_readahead( cur->bc_levels[lev].ra |= lr; block = XFS_BUF_TO_BLOCK(cur->bc_levels[lev].bp); - if (cur->bc_flags & XFS_BTREE_LONG_PTRS) - return xfs_btree_readahead_lblock(cur, lr, block); - return xfs_btree_readahead_sblock(cur, lr, block); + switch (cur->bc_ops->type) { + case XFS_BTREE_TYPE_AG: + return xfs_btree_readahead_agblock(cur, lr, block); + case XFS_BTREE_TYPE_INODE: + return xfs_btree_readahead_fsblock(cur, lr, block); + case XFS_BTREE_TYPE_MEM: + return xfs_btree_readahead_memblock(cur, lr, block); + default: + ASSERT(0); + return 0; + } } STATIC int @@ -1014,23 +1083,24 @@ xfs_btree_ptr_to_daddr( const union xfs_btree_ptr *ptr, xfs_daddr_t *daddr) { - xfs_fsblock_t fsbno; - xfs_agblock_t agbno; int error; error = xfs_btree_check_ptr(cur, ptr, 0, 1); if (error) return error; - if (cur->bc_flags & XFS_BTREE_LONG_PTRS) { - fsbno = be64_to_cpu(ptr->l); - *daddr = XFS_FSB_TO_DADDR(cur->bc_mp, fsbno); - } else { - agbno = be32_to_cpu(ptr->s); + switch (cur->bc_ops->type) { + case XFS_BTREE_TYPE_AG: *daddr = XFS_AGB_TO_DADDR(cur->bc_mp, cur->bc_ag.pag->pag_agno, - agbno); + be32_to_cpu(ptr->s)); + break; + case XFS_BTREE_TYPE_INODE: + *daddr = XFS_FSB_TO_DADDR(cur->bc_mp, be64_to_cpu(ptr->l)); + break; + case XFS_BTREE_TYPE_MEM: + *daddr = xfbno_to_daddr(be64_to_cpu(ptr->l)); + break; } - return 0; } @@ -1050,8 +1120,9 @@ xfs_btree_readahead_ptr( if (xfs_btree_ptr_to_daddr(cur, ptr, &daddr)) return; - xfs_buf_readahead(cur->bc_mp->m_ddev_targp, daddr, - cur->bc_mp->m_bsize * count, cur->bc_ops->buf_ops); + xfs_buf_readahead(xfs_btree_buftarg(cur), daddr, + xfs_btree_bbsize(cur) * count, + cur->bc_ops->buf_ops); } /* @@ -1072,7 +1143,7 @@ xfs_btree_setbuf( cur->bc_levels[lev].ra = 0; b = XFS_BUF_TO_BLOCK(bp); - if (cur->bc_flags & XFS_BTREE_LONG_PTRS) { + if (cur->bc_ops->ptr_len == XFS_BTREE_LONG_PTR_LEN) { if (b->bb_u.l.bb_leftsib == cpu_to_be64(NULLFSBLOCK)) cur->bc_levels[lev].ra |= XFS_BTCUR_LEFTRA; if (b->bb_u.l.bb_rightsib == cpu_to_be64(NULLFSBLOCK)) @@ -1090,7 +1161,7 @@ xfs_btree_ptr_is_null( struct xfs_btree_cur *cur, const union xfs_btree_ptr *ptr) { - if (cur->bc_flags & XFS_BTREE_LONG_PTRS) + if (cur->bc_ops->ptr_len == XFS_BTREE_LONG_PTR_LEN) return ptr->l == cpu_to_be64(NULLFSBLOCK); else return ptr->s == cpu_to_be32(NULLAGBLOCK); @@ -1101,12 +1172,23 @@ xfs_btree_set_ptr_null( struct xfs_btree_cur *cur, union xfs_btree_ptr *ptr) { - if (cur->bc_flags & XFS_BTREE_LONG_PTRS) + if (cur->bc_ops->ptr_len == XFS_BTREE_LONG_PTR_LEN) ptr->l = cpu_to_be64(NULLFSBLOCK); else ptr->s = cpu_to_be32(NULLAGBLOCK); } +static inline bool +xfs_btree_ptrs_equal( + struct xfs_btree_cur *cur, + union xfs_btree_ptr *ptr1, + union xfs_btree_ptr *ptr2) +{ + if (cur->bc_ops->ptr_len == XFS_BTREE_LONG_PTR_LEN) + return ptr1->l == ptr2->l; + return ptr1->s == ptr2->s; +} + /* * Get/set/init sibling pointers */ @@ -1119,7 +1201,7 @@ xfs_btree_get_sibling( { ASSERT(lr == XFS_BB_LEFTSIB || lr == XFS_BB_RIGHTSIB); - if (cur->bc_flags & XFS_BTREE_LONG_PTRS) { + if (cur->bc_ops->ptr_len == XFS_BTREE_LONG_PTR_LEN) { if (lr == XFS_BB_RIGHTSIB) ptr->l = block->bb_u.l.bb_rightsib; else @@ -1141,7 +1223,7 @@ xfs_btree_set_sibling( { ASSERT(lr == XFS_BB_LEFTSIB || lr == XFS_BB_RIGHTSIB); - if (cur->bc_flags & XFS_BTREE_LONG_PTRS) { + if (cur->bc_ops->ptr_len == XFS_BTREE_LONG_PTR_LEN) { if (lr == XFS_BB_RIGHTSIB) block->bb_u.l.bb_rightsib = ptr->l; else @@ -1154,25 +1236,24 @@ xfs_btree_set_sibling( } } -void -xfs_btree_init_block_int( +static void +__xfs_btree_init_block( struct xfs_mount *mp, struct xfs_btree_block *buf, + const struct xfs_btree_ops *ops, xfs_daddr_t blkno, - xfs_btnum_t btnum, __u16 level, __u16 numrecs, - __u64 owner, - unsigned int flags) + __u64 owner) { - int crc = xfs_has_crc(mp); - __u32 magic = xfs_btree_magic(crc, btnum); + bool crc = xfs_has_crc(mp); + __u32 magic = xfs_btree_magic(mp, ops); buf->bb_magic = cpu_to_be32(magic); buf->bb_level = cpu_to_be16(level); buf->bb_numrecs = cpu_to_be16(numrecs); - if (flags & XFS_BTREE_LONG_PTRS) { + if (ops->ptr_len == XFS_BTREE_LONG_PTR_LEN) { buf->bb_u.l.bb_leftsib = cpu_to_be64(NULLFSBLOCK); buf->bb_u.l.bb_rightsib = cpu_to_be64(NULLFSBLOCK); if (crc) { @@ -1183,14 +1264,12 @@ xfs_btree_init_block_int( buf->bb_u.l.bb_lsn = 0; } } else { - /* owner is a 32 bit value on short blocks */ - __u32 __owner = (__u32)owner; - buf->bb_u.s.bb_leftsib = cpu_to_be32(NULLAGBLOCK); buf->bb_u.s.bb_rightsib = cpu_to_be32(NULLAGBLOCK); if (crc) { buf->bb_u.s.bb_blkno = cpu_to_be64(blkno); - buf->bb_u.s.bb_owner = cpu_to_be32(__owner); + /* owner is a 32 bit value on short blocks */ + buf->bb_u.s.bb_owner = cpu_to_be32((__u32)owner); uuid_copy(&buf->bb_u.s.bb_uuid, &mp->m_sb.sb_meta_uuid); buf->bb_u.s.bb_lsn = 0; } @@ -1199,15 +1278,46 @@ xfs_btree_init_block_int( void xfs_btree_init_block( - struct xfs_mount *mp, - struct xfs_buf *bp, - xfs_btnum_t btnum, - __u16 level, - __u16 numrecs, - __u64 owner) + struct xfs_mount *mp, + struct xfs_btree_block *block, + const struct xfs_btree_ops *ops, + __u16 level, + __u16 numrecs, + __u64 owner) +{ + __xfs_btree_init_block(mp, block, ops, XFS_BUF_DADDR_NULL, level, + numrecs, owner); +} + +void +xfs_btree_init_buf( + struct xfs_mount *mp, + struct xfs_buf *bp, + const struct xfs_btree_ops *ops, + __u16 level, + __u16 numrecs, + __u64 owner) +{ + __xfs_btree_init_block(mp, XFS_BUF_TO_BLOCK(bp), ops, + xfs_buf_daddr(bp), level, numrecs, owner); + bp->b_ops = ops->buf_ops; +} + +static inline __u64 +xfs_btree_owner( + struct xfs_btree_cur *cur) { - xfs_btree_init_block_int(mp, XFS_BUF_TO_BLOCK(bp), xfs_buf_daddr(bp), - btnum, level, numrecs, owner, 0); + switch (cur->bc_ops->type) { + case XFS_BTREE_TYPE_MEM: + return cur->bc_mem.xfbtree->owner; + case XFS_BTREE_TYPE_INODE: + return cur->bc_ino.ip->i_ino; + case XFS_BTREE_TYPE_AG: + return cur->bc_ag.pag->pag_agno; + default: + ASSERT(0); + return 0; + } } void @@ -1217,22 +1327,8 @@ xfs_btree_init_block_cur( int level, int numrecs) { - __u64 owner; - - /* - * we can pull the owner from the cursor right now as the different - * owners align directly with the pointer size of the btree. This may - * change in future, but is safe for current users of the generic btree - * code. - */ - if (cur->bc_flags & XFS_BTREE_LONG_PTRS) - owner = cur->bc_ino.ip->i_ino; - else - owner = cur->bc_ag.pag->pag_agno; - - xfs_btree_init_block_int(cur->bc_mp, XFS_BUF_TO_BLOCK(bp), - xfs_buf_daddr(bp), cur->bc_btnum, level, - numrecs, owner, cur->bc_flags); + xfs_btree_init_buf(cur->bc_mp, bp, cur->bc_ops, level, numrecs, + xfs_btree_owner(cur)); } /* @@ -1250,7 +1346,7 @@ xfs_btree_is_lastrec( if (level > 0) return 0; - if (!(cur->bc_flags & XFS_BTREE_LASTREC_UPDATE)) + if (!(cur->bc_ops->geom_flags & XFS_BTGEO_LASTREC_UPDATE)) return 0; xfs_btree_get_sibling(cur, block, &ptr, XFS_BB_RIGHTSIB); @@ -1265,41 +1361,27 @@ xfs_btree_buf_to_ptr( struct xfs_buf *bp, union xfs_btree_ptr *ptr) { - if (cur->bc_flags & XFS_BTREE_LONG_PTRS) - ptr->l = cpu_to_be64(XFS_DADDR_TO_FSB(cur->bc_mp, - xfs_buf_daddr(bp))); - else { + switch (cur->bc_ops->type) { + case XFS_BTREE_TYPE_AG: ptr->s = cpu_to_be32(xfs_daddr_to_agbno(cur->bc_mp, xfs_buf_daddr(bp))); + break; + case XFS_BTREE_TYPE_INODE: + ptr->l = cpu_to_be64(XFS_DADDR_TO_FSB(cur->bc_mp, + xfs_buf_daddr(bp))); + break; + case XFS_BTREE_TYPE_MEM: + ptr->l = cpu_to_be64(xfs_daddr_to_xfbno(xfs_buf_daddr(bp))); + break; } } -STATIC void +static inline void xfs_btree_set_refs( struct xfs_btree_cur *cur, struct xfs_buf *bp) { - switch (cur->bc_btnum) { - case XFS_BTNUM_BNO: - case XFS_BTNUM_CNT: - xfs_buf_set_ref(bp, XFS_ALLOC_BTREE_REF); - break; - case XFS_BTNUM_INO: - case XFS_BTNUM_FINO: - xfs_buf_set_ref(bp, XFS_INO_BTREE_REF); - break; - case XFS_BTNUM_BMAP: - xfs_buf_set_ref(bp, XFS_BMAP_BTREE_REF); - break; - case XFS_BTNUM_RMAP: - xfs_buf_set_ref(bp, XFS_RMAP_BTREE_REF); - break; - case XFS_BTNUM_REFC: - xfs_buf_set_ref(bp, XFS_REFC_BTREE_REF); - break; - default: - ASSERT(0); - } + xfs_buf_set_ref(bp, cur->bc_ops->lru_refs); } int @@ -1309,15 +1391,14 @@ xfs_btree_get_buf_block( struct xfs_btree_block **block, struct xfs_buf **bpp) { - struct xfs_mount *mp = cur->bc_mp; - xfs_daddr_t d; - int error; + xfs_daddr_t d; + int error; error = xfs_btree_ptr_to_daddr(cur, ptr, &d); if (error) return error; - error = xfs_trans_get_buf(cur->bc_tp, mp->m_ddev_targp, d, mp->m_bsize, - 0, bpp); + error = xfs_trans_get_buf(cur->bc_tp, xfs_btree_buftarg(cur), d, + xfs_btree_bbsize(cur), 0, bpp); if (error) return error; @@ -1348,9 +1429,11 @@ xfs_btree_read_buf_block( error = xfs_btree_ptr_to_daddr(cur, ptr, &d); if (error) return error; - error = xfs_trans_read_buf(mp, cur->bc_tp, mp->m_ddev_targp, d, - mp->m_bsize, flags, bpp, - cur->bc_ops->buf_ops); + error = xfs_trans_read_buf(mp, cur->bc_tp, xfs_btree_buftarg(cur), d, + xfs_btree_bbsize(cur), flags, bpp, + cur->bc_ops->buf_ops); + if (xfs_metadata_is_sick(error)) + xfs_btree_mark_sick(cur); if (error) return error; @@ -1398,7 +1481,7 @@ xfs_btree_copy_ptrs( int numptrs) { ASSERT(numptrs >= 0); - memcpy(dst_ptr, src_ptr, numptrs * xfs_btree_ptr_len(cur)); + memcpy(dst_ptr, src_ptr, numptrs * cur->bc_ops->ptr_len); } /* @@ -1454,8 +1537,8 @@ xfs_btree_shift_ptrs( ASSERT(numptrs >= 0); ASSERT(dir == 1 || dir == -1); - dst_ptr = (char *)ptr + (dir * xfs_btree_ptr_len(cur)); - memmove(dst_ptr, ptr, numptrs * xfs_btree_ptr_len(cur)); + dst_ptr = (char *)ptr + (dir * cur->bc_ops->ptr_len); + memmove(dst_ptr, ptr, numptrs * cur->bc_ops->ptr_len); } /* @@ -1566,7 +1649,7 @@ xfs_btree_log_block( if (bp) { int nbits; - if (cur->bc_flags & XFS_BTREE_CRC_BLOCKS) { + if (xfs_has_crc(cur->bc_mp)) { /* * We don't log the CRC when updating a btree * block but instead recreate it during log @@ -1581,7 +1664,7 @@ xfs_btree_log_block( nbits = XFS_BB_NUM_BITS; } xfs_btree_offsets(fields, - (cur->bc_flags & XFS_BTREE_LONG_PTRS) ? + (cur->bc_ops->ptr_len == XFS_BTREE_LONG_PTR_LEN) ? loffsets : soffsets, nbits, &first, &last); xfs_trans_buf_set_type(cur->bc_tp, bp, XFS_BLFT_BTREE_BUF); @@ -1658,9 +1741,10 @@ xfs_btree_increment( * confused or have the tree root in an inode. */ if (lev == cur->bc_nlevels) { - if (cur->bc_flags & XFS_BTREE_ROOT_IN_INODE) + if (cur->bc_ops->type == XFS_BTREE_TYPE_INODE) goto out0; ASSERT(0); + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto error0; } @@ -1751,9 +1835,10 @@ xfs_btree_decrement( * or the root of the tree is in an inode. */ if (lev == cur->bc_nlevels) { - if (cur->bc_flags & XFS_BTREE_ROOT_IN_INODE) + if (cur->bc_ops->type == XFS_BTREE_TYPE_INODE) goto out0; ASSERT(0); + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto error0; } @@ -1786,6 +1871,33 @@ error0: return error; } +/* + * Check the btree block owner now that we have the context to know who the + * real owner is. + */ +static inline xfs_failaddr_t +xfs_btree_check_block_owner( + struct xfs_btree_cur *cur, + struct xfs_btree_block *block) +{ + __u64 owner; + + if (!xfs_has_crc(cur->bc_mp) || + (cur->bc_flags & XFS_BTREE_BMBT_INVALID_OWNER)) + return NULL; + + owner = xfs_btree_owner(cur); + if (cur->bc_ops->ptr_len == XFS_BTREE_LONG_PTR_LEN) { + if (be64_to_cpu(block->bb_u.l.bb_owner) != owner) + return __this_address; + } else { + if (be32_to_cpu(block->bb_u.s.bb_owner) != owner) + return __this_address; + } + + return NULL; +} + int xfs_btree_lookup_get_block( struct xfs_btree_cur *cur, /* btree cursor */ @@ -1798,8 +1910,7 @@ xfs_btree_lookup_get_block( int error = 0; /* special case the root block if in an inode */ - if ((cur->bc_flags & XFS_BTREE_ROOT_IN_INODE) && - (level == cur->bc_nlevels - 1)) { + if (xfs_btree_at_iroot(cur, level)) { *blkp = xfs_btree_get_iroot(cur); return 0; } @@ -1824,11 +1935,7 @@ xfs_btree_lookup_get_block( return error; /* Check the inode owner since the verifiers don't. */ - if (xfs_has_crc(cur->bc_mp) && - !(cur->bc_ino.flags & XFS_BTCUR_BMBT_INVALID_OWNER) && - (cur->bc_flags & XFS_BTREE_LONG_PTRS) && - be64_to_cpu((*blkp)->bb_u.l.bb_owner) != - cur->bc_ino.ip->i_ino) + if (xfs_btree_check_block_owner(cur, *blkp) != NULL) goto out_bad; /* Did we get the level we were looking for? */ @@ -1846,6 +1953,7 @@ out_bad: *blkp = NULL; xfs_buf_mark_corrupt(bp); xfs_trans_brelse(cur->bc_tp, bp); + xfs_btree_mark_sick(cur); return -EFSCORRUPTED; } @@ -1872,6 +1980,27 @@ xfs_lookup_get_search_key( } /* + * Initialize a pointer to the root block. + */ +void +xfs_btree_init_ptr_from_cur( + struct xfs_btree_cur *cur, + union xfs_btree_ptr *ptr) +{ + if (cur->bc_ops->type == XFS_BTREE_TYPE_INODE) { + /* + * Inode-rooted btrees call xfs_btree_get_iroot to find the root + * in xfs_btree_lookup_get_block and don't need a pointer here. + */ + ptr->l = 0; + } else if (cur->bc_flags & XFS_BTREE_STAGING) { + ptr->s = cpu_to_be32(cur->bc_ag.afake->af_root); + } else { + cur->bc_ops->init_ptr_from_cur(cur, ptr); + } +} + +/* * Lookup the record. The cursor is made to point to it, based on dir. * stat is set to 0 if can't find any such record, 1 for success. */ @@ -1892,14 +2021,16 @@ xfs_btree_lookup( XFS_BTREE_STATS_INC(cur, lookup); /* No such thing as a zero-level tree. */ - if (XFS_IS_CORRUPT(cur->bc_mp, cur->bc_nlevels == 0)) + if (XFS_IS_CORRUPT(cur->bc_mp, cur->bc_nlevels == 0)) { + xfs_btree_mark_sick(cur); return -EFSCORRUPTED; + } block = NULL; keyno = 0; /* initialise start pointer from cursor */ - cur->bc_ops->init_ptr_from_cur(cur, &ptr); + xfs_btree_init_ptr_from_cur(cur, &ptr); pp = &ptr; /* @@ -1936,6 +2067,7 @@ xfs_btree_lookup( XFS_ERRLEVEL_LOW, cur->bc_mp, block, sizeof(*block)); + xfs_btree_mark_sick(cur); return -EFSCORRUPTED; } @@ -2012,8 +2144,10 @@ xfs_btree_lookup( error = xfs_btree_increment(cur, 0, &i); if (error) goto error0; - if (XFS_IS_CORRUPT(cur->bc_mp, i != 1)) + if (XFS_IS_CORRUPT(cur->bc_mp, i != 1)) { + xfs_btree_mark_sick(cur); return -EFSCORRUPTED; + } *stat = 1; return 0; } @@ -2040,7 +2174,7 @@ xfs_btree_high_key_from_key( struct xfs_btree_cur *cur, union xfs_btree_key *key) { - ASSERT(cur->bc_flags & XFS_BTREE_OVERLAPPING); + ASSERT(cur->bc_ops->geom_flags & XFS_BTGEO_OVERLAPPING); return (union xfs_btree_key *)((char *)key + (cur->bc_ops->key_len / 2)); } @@ -2061,7 +2195,7 @@ xfs_btree_get_leaf_keys( rec = xfs_btree_rec_addr(cur, 1, block); cur->bc_ops->init_key_from_rec(key, rec); - if (cur->bc_flags & XFS_BTREE_OVERLAPPING) { + if (cur->bc_ops->geom_flags & XFS_BTGEO_OVERLAPPING) { cur->bc_ops->init_high_key_from_rec(&max_hkey, rec); for (n = 2; n <= xfs_btree_get_numrecs(block); n++) { @@ -2088,7 +2222,7 @@ xfs_btree_get_node_keys( union xfs_btree_key *high; int n; - if (cur->bc_flags & XFS_BTREE_OVERLAPPING) { + if (cur->bc_ops->geom_flags & XFS_BTGEO_OVERLAPPING) { memcpy(key, xfs_btree_key_addr(cur, 1, block), cur->bc_ops->key_len / 2); @@ -2132,7 +2266,7 @@ xfs_btree_needs_key_update( struct xfs_btree_cur *cur, int ptr) { - return (cur->bc_flags & XFS_BTREE_OVERLAPPING) || ptr == 1; + return (cur->bc_ops->geom_flags & XFS_BTGEO_OVERLAPPING) || ptr == 1; } /* @@ -2156,7 +2290,7 @@ __xfs_btree_updkeys( struct xfs_buf *bp; int ptr; - ASSERT(cur->bc_flags & XFS_BTREE_OVERLAPPING); + ASSERT(cur->bc_ops->geom_flags & XFS_BTGEO_OVERLAPPING); /* Exit if there aren't any parent levels to update. */ if (level + 1 >= cur->bc_nlevels) @@ -2225,7 +2359,7 @@ xfs_btree_update_keys( ASSERT(level >= 0); block = xfs_btree_get_block(cur, level, &bp); - if (cur->bc_flags & XFS_BTREE_OVERLAPPING) + if (cur->bc_ops->geom_flags & XFS_BTGEO_OVERLAPPING) return __xfs_btree_updkeys(cur, level, block, bp, false); /* @@ -2332,8 +2466,7 @@ xfs_btree_lshift( int error; /* error return value */ int i; - if ((cur->bc_flags & XFS_BTREE_ROOT_IN_INODE) && - level == cur->bc_nlevels - 1) + if (xfs_btree_at_iroot(cur, level)) goto out0; /* Set up variables for this block as "right". */ @@ -2460,12 +2593,13 @@ xfs_btree_lshift( * Using a temporary cursor, update the parent key values of the * block on the left. */ - if (cur->bc_flags & XFS_BTREE_OVERLAPPING) { + if (cur->bc_ops->geom_flags & XFS_BTGEO_OVERLAPPING) { error = xfs_btree_dup_cursor(cur, &tcur); if (error) goto error0; i = xfs_btree_firstrec(tcur, level); if (XFS_IS_CORRUPT(tcur->bc_mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto error0; } @@ -2527,8 +2661,7 @@ xfs_btree_rshift( int error; /* error return value */ int i; /* loop counter */ - if ((cur->bc_flags & XFS_BTREE_ROOT_IN_INODE) && - (level == cur->bc_nlevels - 1)) + if (xfs_btree_at_iroot(cur, level)) goto out0; /* Set up variables for this block as "left". */ @@ -2636,6 +2769,7 @@ xfs_btree_rshift( goto error0; i = xfs_btree_lastrec(tcur, level); if (XFS_IS_CORRUPT(tcur->bc_mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto error0; } @@ -2645,7 +2779,7 @@ xfs_btree_rshift( goto error1; /* Update the parent high keys of the left block, if needed. */ - if (cur->bc_flags & XFS_BTREE_OVERLAPPING) { + if (cur->bc_ops->geom_flags & XFS_BTGEO_OVERLAPPING) { error = xfs_btree_update_keys(cur, level); if (error) goto error1; @@ -2673,6 +2807,32 @@ error1: return error; } +static inline int +xfs_btree_alloc_block( + struct xfs_btree_cur *cur, + const union xfs_btree_ptr *hint_block, + union xfs_btree_ptr *new_block, + int *stat) +{ + int error; + + /* + * Don't allow block allocation for a staging cursor, because staging + * cursors do not support regular btree modifications. + * + * Bulk loading uses a separate callback to obtain new blocks from a + * preallocated list, which prevents ENOSPC failures during loading. + */ + if (unlikely(cur->bc_flags & XFS_BTREE_STAGING)) { + ASSERT(0); + return -EFSCORRUPTED; + } + + error = cur->bc_ops->alloc_block(cur, hint_block, new_block, stat); + trace_xfs_btree_alloc_block(cur, new_block, *stat, error); + return error; +} + /* * Split cur/level block in half. * Return new block number and the key to its first @@ -2716,7 +2876,7 @@ __xfs_btree_split( xfs_btree_buf_to_ptr(cur, lbp, &lptr); /* Allocate the new block. If we can't do it, we're toast. Give up. */ - error = cur->bc_ops->alloc_block(cur, &lptr, &rptr, stat); + error = xfs_btree_alloc_block(cur, &lptr, &rptr, stat); if (error) goto error0; if (*stat == 0) @@ -2823,7 +2983,7 @@ __xfs_btree_split( } /* Update the parent high keys of the left block, if needed. */ - if (cur->bc_flags & XFS_BTREE_OVERLAPPING) { + if (cur->bc_ops->geom_flags & XFS_BTGEO_OVERLAPPING) { error = xfs_btree_update_keys(cur, level); if (error) goto error0; @@ -2941,7 +3101,7 @@ xfs_btree_split( struct xfs_btree_split_args args; DECLARE_COMPLETION_ONSTACK(done); - if (cur->bc_btnum != XFS_BTNUM_BMAP || + if (!xfs_btree_is_bmap(cur->bc_ops) || cur->bc_tp->t_highest_agno == NULLAGNUMBER) return __xfs_btree_split(cur, level, ptrp, key, curp, stat); @@ -2963,7 +3123,6 @@ xfs_btree_split( #define xfs_btree_split __xfs_btree_split #endif /* __KERNEL__ */ - /* * Copy the old inode root contents into a real block and make the * broot point to it. @@ -2988,7 +3147,7 @@ xfs_btree_new_iroot( XFS_BTREE_STATS_INC(cur, newroot); - ASSERT(cur->bc_flags & XFS_BTREE_ROOT_IN_INODE); + ASSERT(cur->bc_ops->type == XFS_BTREE_TYPE_INODE); level = cur->bc_nlevels - 1; @@ -2996,7 +3155,7 @@ xfs_btree_new_iroot( pp = xfs_btree_ptr_addr(cur, 1, block); /* Allocate the new block. If we can't do it, we're toast. Give up. */ - error = cur->bc_ops->alloc_block(cur, pp, &nptr, stat); + error = xfs_btree_alloc_block(cur, pp, &nptr, stat); if (error) goto error0; if (*stat == 0) @@ -3014,9 +3173,9 @@ xfs_btree_new_iroot( * In that case have to also ensure the blkno remains correct */ memcpy(cblock, block, xfs_btree_block_len(cur)); - if (cur->bc_flags & XFS_BTREE_CRC_BLOCKS) { + if (xfs_has_crc(cur->bc_mp)) { __be64 bno = cpu_to_be64(xfs_buf_daddr(cbp)); - if (cur->bc_flags & XFS_BTREE_LONG_PTRS) + if (cur->bc_ops->ptr_len == XFS_BTREE_LONG_PTR_LEN) cblock->bb_u.l.bb_blkno = bno; else cblock->bb_u.s.bb_blkno = bno; @@ -3069,6 +3228,21 @@ error0: return error; } +static void +xfs_btree_set_root( + struct xfs_btree_cur *cur, + const union xfs_btree_ptr *ptr, + int inc) +{ + if (cur->bc_flags & XFS_BTREE_STAGING) { + /* Update the btree root information for a per-AG fake root. */ + cur->bc_ag.afake->af_root = be32_to_cpu(ptr->s); + cur->bc_ag.afake->af_levels += inc; + } else { + cur->bc_ops->set_root(cur, ptr, inc); + } +} + /* * Allocate a new root block, fill it in. */ @@ -3093,10 +3267,10 @@ xfs_btree_new_root( XFS_BTREE_STATS_INC(cur, newroot); /* initialise our start point from the cursor */ - cur->bc_ops->init_ptr_from_cur(cur, &rptr); + xfs_btree_init_ptr_from_cur(cur, &rptr); /* Allocate the new block. If we can't do it, we're toast. Give up. */ - error = cur->bc_ops->alloc_block(cur, &rptr, &lptr, stat); + error = xfs_btree_alloc_block(cur, &rptr, &lptr, stat); if (error) goto error0; if (*stat == 0) @@ -3109,7 +3283,7 @@ xfs_btree_new_root( goto error0; /* Set the root in the holding structure increasing the level by 1. */ - cur->bc_ops->set_root(cur, &lptr, 1); + xfs_btree_set_root(cur, &lptr, 1); /* * At the previous root level there are now two blocks: the old root, @@ -3213,8 +3387,7 @@ xfs_btree_make_block_unfull( { int error = 0; - if ((cur->bc_flags & XFS_BTREE_ROOT_IN_INODE) && - level == cur->bc_nlevels - 1) { + if (xfs_btree_at_iroot(cur, level)) { struct xfs_inode *ip = cur->bc_ino.ip; if (numrecs < cur->bc_ops->get_dmaxrecs(cur, level)) { @@ -3299,8 +3472,8 @@ xfs_btree_insrec( * If we have an external root pointer, and we've made it to the * root level, allocate a new root block and we're done. */ - if (!(cur->bc_flags & XFS_BTREE_ROOT_IN_INODE) && - (level >= cur->bc_nlevels)) { + if (cur->bc_ops->type != XFS_BTREE_TYPE_INODE && + level >= cur->bc_nlevels) { error = xfs_btree_new_root(cur, stat); xfs_btree_set_ptr_null(cur, ptrp); @@ -3524,6 +3697,7 @@ xfs_btree_insert( } if (XFS_IS_CORRUPT(cur->bc_mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto error0; } @@ -3537,7 +3711,8 @@ xfs_btree_insert( if (pcur != cur && (ncur || xfs_btree_ptr_is_null(cur, &nptr))) { /* Save the state from the cursor before we trash it */ - if (cur->bc_ops->update_cursor) + if (cur->bc_ops->update_cursor && + !(cur->bc_flags & XFS_BTREE_STAGING)) cur->bc_ops->update_cursor(pcur, cur); cur->bc_nlevels = pcur->bc_nlevels; xfs_btree_del_cursor(pcur, XFS_BTREE_NOERROR); @@ -3586,7 +3761,7 @@ xfs_btree_kill_iroot( #endif int i; - ASSERT(cur->bc_flags & XFS_BTREE_ROOT_IN_INODE); + ASSERT(cur->bc_ops->type == XFS_BTREE_TYPE_INODE); ASSERT(cur->bc_nlevels > 1); /* @@ -3680,7 +3855,7 @@ xfs_btree_kill_root( * Update the root pointer, decreasing the level by 1 and then * free the old root. */ - cur->bc_ops->set_root(cur, newroot, -1); + xfs_btree_set_root(cur, newroot, -1); error = xfs_btree_free_block(cur, bp); if (error) @@ -3822,27 +3997,25 @@ xfs_btree_delrec( * Try to get rid of the next level down. If we can't then there's * nothing left to do. */ - if (level == cur->bc_nlevels - 1) { - if (cur->bc_flags & XFS_BTREE_ROOT_IN_INODE) { - xfs_iroot_realloc(cur->bc_ino.ip, -1, - cur->bc_ino.whichfork); + if (xfs_btree_at_iroot(cur, level)) { + xfs_iroot_realloc(cur->bc_ino.ip, -1, cur->bc_ino.whichfork); - error = xfs_btree_kill_iroot(cur); - if (error) - goto error0; + error = xfs_btree_kill_iroot(cur); + if (error) + goto error0; - error = xfs_btree_dec_cursor(cur, level, stat); - if (error) - goto error0; - *stat = 1; - return 0; - } + error = xfs_btree_dec_cursor(cur, level, stat); + if (error) + goto error0; + *stat = 1; + return 0; + } - /* - * If this is the root level, and there's only one entry left, - * and it's NOT the leaf level, then we can get rid of this - * level. - */ + /* + * If this is the root level, and there's only one entry left, and it's + * NOT the leaf level, then we can get rid of this level. + */ + if (level == cur->bc_nlevels - 1) { if (numrecs == 1 && level > 0) { union xfs_btree_ptr *pp; /* @@ -3891,7 +4064,7 @@ xfs_btree_delrec( xfs_btree_get_sibling(cur, block, &rptr, XFS_BB_RIGHTSIB); xfs_btree_get_sibling(cur, block, &lptr, XFS_BB_LEFTSIB); - if (cur->bc_flags & XFS_BTREE_ROOT_IN_INODE) { + if (cur->bc_ops->type == XFS_BTREE_TYPE_INODE) { /* * One child of root, need to get a chance to copy its contents * into the root and delete it. Can't go up to next level, @@ -3931,6 +4104,7 @@ xfs_btree_delrec( */ i = xfs_btree_lastrec(tcur, level); if (XFS_IS_CORRUPT(cur->bc_mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto error0; } @@ -3939,12 +4113,14 @@ xfs_btree_delrec( if (error) goto error0; if (XFS_IS_CORRUPT(cur->bc_mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto error0; } i = xfs_btree_lastrec(tcur, level); if (XFS_IS_CORRUPT(cur->bc_mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto error0; } @@ -3992,6 +4168,7 @@ xfs_btree_delrec( if (!xfs_btree_ptr_is_null(cur, &lptr)) { i = xfs_btree_firstrec(tcur, level); if (XFS_IS_CORRUPT(cur->bc_mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto error0; } @@ -4000,6 +4177,7 @@ xfs_btree_delrec( if (error) goto error0; if (XFS_IS_CORRUPT(cur->bc_mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto error0; } @@ -4017,6 +4195,7 @@ xfs_btree_delrec( */ i = xfs_btree_firstrec(tcur, level); if (XFS_IS_CORRUPT(cur->bc_mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto error0; } @@ -4026,6 +4205,7 @@ xfs_btree_delrec( goto error0; i = xfs_btree_firstrec(tcur, level); if (XFS_IS_CORRUPT(cur->bc_mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto error0; } @@ -4201,8 +4381,8 @@ xfs_btree_delrec( * If we joined with the right neighbor and there's a level above * us, increment the cursor at that level. */ - else if ((cur->bc_flags & XFS_BTREE_ROOT_IN_INODE) || - (level + 1 < cur->bc_nlevels)) { + else if (cur->bc_ops->type == XFS_BTREE_TYPE_INODE || + level + 1 < cur->bc_nlevels) { error = xfs_btree_increment(cur, level + 1, &i); if (error) goto error0; @@ -4270,7 +4450,7 @@ xfs_btree_delete( * If we combined blocks as part of deleting the record, delrec won't * have updated the parent high keys so we have to do that here. */ - if (joined && (cur->bc_flags & XFS_BTREE_OVERLAPPING)) { + if (joined && (cur->bc_ops->geom_flags & XFS_BTGEO_OVERLAPPING)) { error = xfs_btree_updkeys_force(cur, 0); if (error) goto error0; @@ -4344,7 +4524,7 @@ xfs_btree_visit_block( { struct xfs_btree_block *block; struct xfs_buf *bp; - union xfs_btree_ptr rptr; + union xfs_btree_ptr rptr, bufptr; int error; /* do right sibling readahead */ @@ -4367,15 +4547,12 @@ xfs_btree_visit_block( * return the same block without checking if the right sibling points * back to us and creates a cyclic reference in the btree. */ - if (cur->bc_flags & XFS_BTREE_LONG_PTRS) { - if (be64_to_cpu(rptr.l) == XFS_DADDR_TO_FSB(cur->bc_mp, - xfs_buf_daddr(bp))) - return -EFSCORRUPTED; - } else { - if (be32_to_cpu(rptr.s) == xfs_daddr_to_agbno(cur->bc_mp, - xfs_buf_daddr(bp))) - return -EFSCORRUPTED; + xfs_btree_buf_to_ptr(cur, bp, &bufptr); + if (xfs_btree_ptrs_equal(cur, &rptr, &bufptr)) { + xfs_btree_mark_sick(cur); + return -EFSCORRUPTED; } + return xfs_btree_lookup_get_block(cur, level, &rptr, &block); } @@ -4393,7 +4570,7 @@ xfs_btree_visit_blocks( struct xfs_btree_block *block = NULL; int error = 0; - cur->bc_ops->init_ptr_from_cur(cur, &lptr); + xfs_btree_init_ptr_from_cur(cur, &lptr); /* for each level */ for (level = cur->bc_nlevels - 1; level >= 0; level--) { @@ -4471,7 +4648,7 @@ xfs_btree_block_change_owner( /* modify the owner */ block = xfs_btree_get_block(cur, level, &bp); - if (cur->bc_flags & XFS_BTREE_LONG_PTRS) { + if (cur->bc_ops->ptr_len == XFS_BTREE_LONG_PTR_LEN) { if (block->bb_u.l.bb_owner == cpu_to_be64(bbcoi->new_owner)) return 0; block->bb_u.l.bb_owner = cpu_to_be64(bbcoi->new_owner); @@ -4489,7 +4666,7 @@ xfs_btree_block_change_owner( * though, so everything is consistent in memory. */ if (!bp) { - ASSERT(cur->bc_flags & XFS_BTREE_ROOT_IN_INODE); + ASSERT(cur->bc_ops->type == XFS_BTREE_TYPE_INODE); ASSERT(level == cur->bc_nlevels - 1); return 0; } @@ -4523,7 +4700,7 @@ xfs_btree_change_owner( /* Verify the v5 fields of a long-format btree block. */ xfs_failaddr_t -xfs_btree_lblock_v5hdr_verify( +xfs_btree_fsblock_v5hdr_verify( struct xfs_buf *bp, uint64_t owner) { @@ -4544,7 +4721,7 @@ xfs_btree_lblock_v5hdr_verify( /* Verify a long-format btree block. */ xfs_failaddr_t -xfs_btree_lblock_verify( +xfs_btree_fsblock_verify( struct xfs_buf *bp, unsigned int max_recs) { @@ -4553,28 +4730,60 @@ xfs_btree_lblock_verify( xfs_fsblock_t fsb; xfs_failaddr_t fa; + ASSERT(!xfs_buftarg_is_mem(bp->b_target)); + /* numrecs verification */ if (be16_to_cpu(block->bb_numrecs) > max_recs) return __this_address; /* sibling pointer verification */ fsb = XFS_DADDR_TO_FSB(mp, xfs_buf_daddr(bp)); - fa = xfs_btree_check_lblock_siblings(mp, NULL, -1, fsb, + fa = xfs_btree_check_fsblock_siblings(mp, fsb, block->bb_u.l.bb_leftsib); if (!fa) - fa = xfs_btree_check_lblock_siblings(mp, NULL, -1, fsb, + fa = xfs_btree_check_fsblock_siblings(mp, fsb, block->bb_u.l.bb_rightsib); return fa; } +/* Verify an in-memory btree block. */ +xfs_failaddr_t +xfs_btree_memblock_verify( + struct xfs_buf *bp, + unsigned int max_recs) +{ + struct xfs_btree_block *block = XFS_BUF_TO_BLOCK(bp); + struct xfs_buftarg *btp = bp->b_target; + xfs_failaddr_t fa; + xfbno_t bno; + + ASSERT(xfs_buftarg_is_mem(bp->b_target)); + + /* numrecs verification */ + if (be16_to_cpu(block->bb_numrecs) > max_recs) + return __this_address; + + /* sibling pointer verification */ + bno = xfs_daddr_to_xfbno(xfs_buf_daddr(bp)); + fa = xfs_btree_check_memblock_siblings(btp, bno, + block->bb_u.l.bb_leftsib); + if (fa) + return fa; + fa = xfs_btree_check_memblock_siblings(btp, bno, + block->bb_u.l.bb_rightsib); + if (fa) + return fa; + + return NULL; +} /** - * xfs_btree_sblock_v5hdr_verify() -- verify the v5 fields of a short-format + * xfs_btree_agblock_v5hdr_verify() -- verify the v5 fields of a short-format * btree block * * @bp: buffer containing the btree block */ xfs_failaddr_t -xfs_btree_sblock_v5hdr_verify( +xfs_btree_agblock_v5hdr_verify( struct xfs_buf *bp) { struct xfs_mount *mp = bp->b_mount; @@ -4593,13 +4802,13 @@ xfs_btree_sblock_v5hdr_verify( } /** - * xfs_btree_sblock_verify() -- verify a short-format btree block + * xfs_btree_agblock_verify() -- verify a short-format btree block * * @bp: buffer containing the btree block * @max_recs: maximum records allowed in this btree node */ xfs_failaddr_t -xfs_btree_sblock_verify( +xfs_btree_agblock_verify( struct xfs_buf *bp, unsigned int max_recs) { @@ -4608,16 +4817,18 @@ xfs_btree_sblock_verify( xfs_agblock_t agbno; xfs_failaddr_t fa; + ASSERT(!xfs_buftarg_is_mem(bp->b_target)); + /* numrecs verification */ if (be16_to_cpu(block->bb_numrecs) > max_recs) return __this_address; /* sibling pointer verification */ agbno = xfs_daddr_to_agbno(mp, xfs_buf_daddr(bp)); - fa = xfs_btree_check_sblock_siblings(bp->b_pag, NULL, -1, agbno, + fa = xfs_btree_check_agblock_siblings(bp->b_pag, agbno, block->bb_u.s.bb_leftsib); if (!fa) - fa = xfs_btree_check_sblock_siblings(bp->b_pag, NULL, -1, agbno, + fa = xfs_btree_check_agblock_siblings(bp->b_pag, agbno, block->bb_u.s.bb_rightsib); return fa; } @@ -4815,7 +5026,7 @@ xfs_btree_overlapped_query_range( /* Load the root of the btree. */ level = cur->bc_nlevels - 1; - cur->bc_ops->init_ptr_from_cur(cur, &ptr); + xfs_btree_init_ptr_from_cur(cur, &ptr); error = xfs_btree_lookup_get_block(cur, level, &ptr, &block); if (error) return error; @@ -4966,7 +5177,7 @@ xfs_btree_query_range( if (!xfs_btree_keycmp_le(cur, &low_key, &high_key)) return -EINVAL; - if (!(cur->bc_flags & XFS_BTREE_OVERLAPPING)) + if (!(cur->bc_ops->geom_flags & XFS_BTGEO_OVERLAPPING)) return xfs_btree_simple_query_range(cur, &low_key, &high_key, fn, priv); return xfs_btree_overlapped_query_range(cur, &low_key, &high_key, @@ -5020,7 +5231,7 @@ xfs_btree_diff_two_ptrs( const union xfs_btree_ptr *a, const union xfs_btree_ptr *b) { - if (cur->bc_flags & XFS_BTREE_LONG_PTRS) + if (cur->bc_ops->ptr_len == XFS_BTREE_LONG_PTR_LEN) return (int64_t)be64_to_cpu(a->l) - be64_to_cpu(b->l); return (int64_t)be32_to_cpu(a->s) - be32_to_cpu(b->s); } @@ -5074,7 +5285,7 @@ xfs_btree_has_records_helper( key_contig = cur->bc_ops->keys_contiguous(cur, &info->high_key, &rec_key, info->key_mask); if (key_contig == XBTREE_KEY_OVERLAP && - !(cur->bc_flags & XFS_BTREE_OVERLAPPING)) + !(cur->bc_ops->geom_flags & XFS_BTGEO_OVERLAPPING)) return -EFSCORRUPTED; if (key_contig == XBTREE_KEY_GAP) return -ECANCELED; @@ -5168,7 +5379,7 @@ xfs_btree_has_more_records( return true; /* There are more record blocks. */ - if (cur->bc_flags & XFS_BTREE_LONG_PTRS) + if (cur->bc_ops->ptr_len == XFS_BTREE_LONG_PTR_LEN) return block->bb_u.l.bb_rightsib != cpu_to_be64(NULLFSBLOCK); else return block->bb_u.s.bb_rightsib != cpu_to_be32(NULLAGBLOCK); @@ -5233,6 +5444,7 @@ xfs_btree_goto_left_edge( return error; if (stat != 0) { ASSERT(0); + xfs_btree_mark_sick(cur); return -EFSCORRUPTED; } diff --git a/fs/xfs/libxfs/xfs_btree.h b/fs/xfs/libxfs/xfs_btree.h index d906324e25c8..f93374278aa1 100644 --- a/fs/xfs/libxfs/xfs_btree.h +++ b/fs/xfs/libxfs/xfs_btree.h @@ -55,15 +55,8 @@ union xfs_btree_rec { #define XFS_LOOKUP_LE ((xfs_lookup_t)XFS_LOOKUP_LEi) #define XFS_LOOKUP_GE ((xfs_lookup_t)XFS_LOOKUP_GEi) -#define XFS_BTNUM_BNO ((xfs_btnum_t)XFS_BTNUM_BNOi) -#define XFS_BTNUM_CNT ((xfs_btnum_t)XFS_BTNUM_CNTi) -#define XFS_BTNUM_BMAP ((xfs_btnum_t)XFS_BTNUM_BMAPi) -#define XFS_BTNUM_INO ((xfs_btnum_t)XFS_BTNUM_INOi) -#define XFS_BTNUM_FINO ((xfs_btnum_t)XFS_BTNUM_FINOi) -#define XFS_BTNUM_RMAP ((xfs_btnum_t)XFS_BTNUM_RMAPi) -#define XFS_BTNUM_REFC ((xfs_btnum_t)XFS_BTNUM_REFCi) - -uint32_t xfs_btree_magic(int crc, xfs_btnum_t btnum); +struct xfs_btree_ops; +uint32_t xfs_btree_magic(struct xfs_mount *mp, const struct xfs_btree_ops *ops); /* * For logging record fields. @@ -86,9 +79,11 @@ uint32_t xfs_btree_magic(int crc, xfs_btnum_t btnum); * Generic stats interface */ #define XFS_BTREE_STATS_INC(cur, stat) \ - XFS_STATS_INC_OFF((cur)->bc_mp, (cur)->bc_statoff + __XBTS_ ## stat) + XFS_STATS_INC_OFF((cur)->bc_mp, \ + (cur)->bc_ops->statoff + __XBTS_ ## stat) #define XFS_BTREE_STATS_ADD(cur, stat, val) \ - XFS_STATS_ADD_OFF((cur)->bc_mp, (cur)->bc_statoff + __XBTS_ ## stat, val) + XFS_STATS_ADD_OFF((cur)->bc_mp, \ + (cur)->bc_ops->statoff + __XBTS_ ## stat, val) enum xbtree_key_contig { XBTREE_KEY_GAP = 0, @@ -111,10 +106,37 @@ static inline enum xbtree_key_contig xbtree_key_contig(uint64_t x, uint64_t y) return XBTREE_KEY_OVERLAP; } +#define XFS_BTREE_LONG_PTR_LEN (sizeof(__be64)) +#define XFS_BTREE_SHORT_PTR_LEN (sizeof(__be32)) + +enum xfs_btree_type { + XFS_BTREE_TYPE_AG, + XFS_BTREE_TYPE_INODE, + XFS_BTREE_TYPE_MEM, +}; + struct xfs_btree_ops { - /* size of the key and record structures */ - size_t key_len; - size_t rec_len; + const char *name; + + /* Type of btree - AG-rooted or inode-rooted */ + enum xfs_btree_type type; + + /* XFS_BTGEO_* flags that determine the geometry of the btree */ + unsigned int geom_flags; + + /* size of the key, pointer, and record structures */ + size_t key_len; + size_t ptr_len; + size_t rec_len; + + /* LRU refcount to set on each btree buffer created */ + unsigned int lru_refs; + + /* offset of btree stats array */ + unsigned int statoff; + + /* sick mask for health reporting (only for XFS_BTREE_TYPE_AG) */ + unsigned int sick_mask; /* cursor operations */ struct xfs_btree_cur *(*dup_cursor)(struct xfs_btree_cur *); @@ -199,6 +221,10 @@ struct xfs_btree_ops { const union xfs_btree_key *mask); }; +/* btree geometry flags */ +#define XFS_BTGEO_LASTREC_UPDATE (1U << 0) /* track last rec externally */ +#define XFS_BTGEO_OVERLAPPING (1U << 1) /* overlapping intervals */ + /* * Reasons for the update_lastrec method to be called. */ @@ -215,39 +241,6 @@ union xfs_btree_irec { struct xfs_refcount_irec rc; }; -/* Per-AG btree information. */ -struct xfs_btree_cur_ag { - struct xfs_perag *pag; - union { - struct xfs_buf *agbp; - struct xbtree_afakeroot *afake; /* for staging cursor */ - }; - union { - struct { - unsigned int nr_ops; /* # record updates */ - unsigned int shape_changes; /* # of extent splits */ - } refc; - struct { - bool active; /* allocation cursor state */ - } abt; - }; -}; - -/* Btree-in-inode cursor information */ -struct xfs_btree_cur_ino { - struct xfs_inode *ip; - struct xbtree_ifakeroot *ifake; /* for staging cursor */ - int allocated; - short forksize; - char whichfork; - char flags; -/* We are converting a delalloc reservation */ -#define XFS_BTCUR_BMBT_WASDEL (1 << 0) - -/* For extent swap, ignore owner check in verifier */ -#define XFS_BTCUR_BMBT_INVALID_OWNER (1 << 1) -}; - struct xfs_btree_level { /* buffer pointer */ struct xfs_buf *bp; @@ -272,21 +265,38 @@ struct xfs_btree_cur const struct xfs_btree_ops *bc_ops; struct kmem_cache *bc_cache; /* cursor cache */ unsigned int bc_flags; /* btree features - below */ - xfs_btnum_t bc_btnum; /* identifies which btree type */ union xfs_btree_irec bc_rec; /* current insert/search record value */ uint8_t bc_nlevels; /* number of levels in the tree */ uint8_t bc_maxlevels; /* maximum levels for this btree type */ - int bc_statoff; /* offset of btree stats array */ - /* - * Short btree pointers need an agno to be able to turn the pointers - * into physical addresses for IO, so the btree cursor switches between - * bc_ino and bc_ag based on whether XFS_BTREE_LONG_PTRS is set for the - * cursor. - */ + /* per-type information */ union { - struct xfs_btree_cur_ag bc_ag; - struct xfs_btree_cur_ino bc_ino; + struct { + struct xfs_inode *ip; + short forksize; + char whichfork; + struct xbtree_ifakeroot *ifake; /* for staging cursor */ + } bc_ino; + struct { + struct xfs_perag *pag; + struct xfs_buf *agbp; + struct xbtree_afakeroot *afake; /* for staging cursor */ + } bc_ag; + struct { + struct xfbtree *xfbtree; + struct xfs_perag *pag; + } bc_mem; + }; + + /* per-format private data */ + union { + struct { + int allocated; + } bc_bmap; /* bmapbt */ + struct { + unsigned int nr_ops; /* # record updates */ + unsigned int shape_changes; /* # of extent splits */ + } bc_refc; /* refcountbt */ }; /* Must be at the end of the struct! */ @@ -304,18 +314,22 @@ xfs_btree_cur_sizeof(unsigned int nlevels) return struct_size_t(struct xfs_btree_cur, bc_levels, nlevels); } -/* cursor flags */ -#define XFS_BTREE_LONG_PTRS (1<<0) /* pointers are 64bits long */ -#define XFS_BTREE_ROOT_IN_INODE (1<<1) /* root may be variable size */ -#define XFS_BTREE_LASTREC_UPDATE (1<<2) /* track last rec externally */ -#define XFS_BTREE_CRC_BLOCKS (1<<3) /* uses extended btree blocks */ -#define XFS_BTREE_OVERLAPPING (1<<4) /* overlapping intervals */ +/* cursor state flags */ /* * The root of this btree is a fakeroot structure so that we can stage a btree * rebuild without leaving it accessible via primary metadata. The ops struct * is dynamically allocated and must be freed when the cursor is deleted. */ -#define XFS_BTREE_STAGING (1<<5) +#define XFS_BTREE_STAGING (1U << 0) + +/* We are converting a delalloc reservation (only for bmbt btrees) */ +#define XFS_BTREE_BMBT_WASDEL (1U << 1) + +/* For extent swap, ignore owner check in verifier (only for bmbt btrees) */ +#define XFS_BTREE_BMBT_INVALID_OWNER (1U << 2) + +/* Cursor is active (only for allocbt btrees) */ +#define XFS_BTREE_ALLOCBT_ACTIVE (1U << 3) #define XFS_BTREE_NOERROR 0 #define XFS_BTREE_ERROR 1 @@ -325,14 +339,10 @@ xfs_btree_cur_sizeof(unsigned int nlevels) */ #define XFS_BUF_TO_BLOCK(bp) ((struct xfs_btree_block *)((bp)->b_addr)) -/* - * Internal long and short btree block checks. They return NULL if the - * block is ok or the address of the failed check otherwise. - */ -xfs_failaddr_t __xfs_btree_check_lblock(struct xfs_btree_cur *cur, - struct xfs_btree_block *block, int level, struct xfs_buf *bp); -xfs_failaddr_t __xfs_btree_check_sblock(struct xfs_btree_cur *cur, +xfs_failaddr_t __xfs_btree_check_block(struct xfs_btree_cur *cur, struct xfs_btree_block *block, int level, struct xfs_buf *bp); +int __xfs_btree_check_ptr(struct xfs_btree_cur *cur, + const union xfs_btree_ptr *ptr, int index, int level); /* * Check that block header is ok. @@ -345,24 +355,6 @@ xfs_btree_check_block( struct xfs_buf *bp); /* buffer containing block, if any */ /* - * Check that (long) pointer is ok. - */ -bool /* error (0 or EFSCORRUPTED) */ -xfs_btree_check_lptr( - struct xfs_btree_cur *cur, /* btree cursor */ - xfs_fsblock_t fsbno, /* btree block disk address */ - int level); /* btree block level */ - -/* - * Check that (short) pointer is ok. - */ -bool /* error (0 or EFSCORRUPTED) */ -xfs_btree_check_sptr( - struct xfs_btree_cur *cur, /* btree cursor */ - xfs_agblock_t agbno, /* btree block disk address */ - int level); /* btree block level */ - -/* * Delete the btree cursor. */ void @@ -392,63 +384,14 @@ xfs_btree_offsets( int *last); /* output: last byte offset */ /* - * Get a buffer for the block, return it read in. - * Long-form addressing. - */ -int /* error */ -xfs_btree_read_bufl( - struct xfs_mount *mp, /* file system mount point */ - struct xfs_trans *tp, /* transaction pointer */ - xfs_fsblock_t fsbno, /* file system block number */ - struct xfs_buf **bpp, /* buffer for fsbno */ - int refval, /* ref count value for buffer */ - const struct xfs_buf_ops *ops); - -/* - * Read-ahead the block, don't wait for it, don't return a buffer. - * Long-form addressing. - */ -void /* error */ -xfs_btree_reada_bufl( - struct xfs_mount *mp, /* file system mount point */ - xfs_fsblock_t fsbno, /* file system block number */ - xfs_extlen_t count, /* count of filesystem blocks */ - const struct xfs_buf_ops *ops); - -/* - * Read-ahead the block, don't wait for it, don't return a buffer. - * Short-form addressing. - */ -void /* error */ -xfs_btree_reada_bufs( - struct xfs_mount *mp, /* file system mount point */ - xfs_agnumber_t agno, /* allocation group number */ - xfs_agblock_t agbno, /* allocation group block number */ - xfs_extlen_t count, /* count of filesystem blocks */ - const struct xfs_buf_ops *ops); - -/* * Initialise a new btree block header */ -void -xfs_btree_init_block( - struct xfs_mount *mp, - struct xfs_buf *bp, - xfs_btnum_t btnum, - __u16 level, - __u16 numrecs, - __u64 owner); - -void -xfs_btree_init_block_int( - struct xfs_mount *mp, - struct xfs_btree_block *buf, - xfs_daddr_t blkno, - xfs_btnum_t btnum, - __u16 level, - __u16 numrecs, - __u64 owner, - unsigned int flags); +void xfs_btree_init_buf(struct xfs_mount *mp, struct xfs_buf *bp, + const struct xfs_btree_ops *ops, __u16 level, __u16 numrecs, + __u64 owner); +void xfs_btree_init_block(struct xfs_mount *mp, + struct xfs_btree_block *buf, const struct xfs_btree_ops *ops, + __u16 level, __u16 numrecs, __u64 owner); /* * Common btree core entry points. @@ -467,10 +410,10 @@ int xfs_btree_change_owner(struct xfs_btree_cur *cur, uint64_t new_owner, /* * btree block CRC helpers */ -void xfs_btree_lblock_calc_crc(struct xfs_buf *); -bool xfs_btree_lblock_verify_crc(struct xfs_buf *); -void xfs_btree_sblock_calc_crc(struct xfs_buf *); -bool xfs_btree_sblock_verify_crc(struct xfs_buf *); +void xfs_btree_fsblock_calc_crc(struct xfs_buf *); +bool xfs_btree_fsblock_verify_crc(struct xfs_buf *); +void xfs_btree_agblock_calc_crc(struct xfs_buf *); +bool xfs_btree_agblock_verify_crc(struct xfs_buf *); /* * Internal btree helpers also used by xfs_bmap.c. @@ -510,12 +453,14 @@ static inline int xfs_btree_get_level(const struct xfs_btree_block *block) #define XFS_FILBLKS_MIN(a,b) min_t(xfs_filblks_t, (a), (b)) #define XFS_FILBLKS_MAX(a,b) max_t(xfs_filblks_t, (a), (b)) -xfs_failaddr_t xfs_btree_sblock_v5hdr_verify(struct xfs_buf *bp); -xfs_failaddr_t xfs_btree_sblock_verify(struct xfs_buf *bp, +xfs_failaddr_t xfs_btree_agblock_v5hdr_verify(struct xfs_buf *bp); +xfs_failaddr_t xfs_btree_agblock_verify(struct xfs_buf *bp, unsigned int max_recs); -xfs_failaddr_t xfs_btree_lblock_v5hdr_verify(struct xfs_buf *bp, +xfs_failaddr_t xfs_btree_fsblock_v5hdr_verify(struct xfs_buf *bp, uint64_t owner); -xfs_failaddr_t xfs_btree_lblock_verify(struct xfs_buf *bp, +xfs_failaddr_t xfs_btree_fsblock_verify(struct xfs_buf *bp, + unsigned int max_recs); +xfs_failaddr_t xfs_btree_memblock_verify(struct xfs_buf *bp, unsigned int max_recs); unsigned int xfs_btree_compute_maxlevels(const unsigned int *limits, @@ -690,7 +635,7 @@ xfs_btree_islastblock( block = xfs_btree_get_block(cur, level, &bp); - if (cur->bc_flags & XFS_BTREE_LONG_PTRS) + if (cur->bc_ops->ptr_len == XFS_BTREE_LONG_PTR_LEN) return block->bb_u.l.bb_rightsib == cpu_to_be64(NULLFSBLOCK); return block->bb_u.s.bb_rightsib == cpu_to_be32(NULLAGBLOCK); } @@ -714,21 +659,28 @@ void xfs_btree_copy_ptrs(struct xfs_btree_cur *cur, void xfs_btree_copy_keys(struct xfs_btree_cur *cur, union xfs_btree_key *dst_key, const union xfs_btree_key *src_key, int numkeys); +void xfs_btree_init_ptr_from_cur(struct xfs_btree_cur *cur, + union xfs_btree_ptr *ptr); static inline struct xfs_btree_cur * xfs_btree_alloc_cursor( struct xfs_mount *mp, struct xfs_trans *tp, - xfs_btnum_t btnum, + const struct xfs_btree_ops *ops, uint8_t maxlevels, struct kmem_cache *cache) { struct xfs_btree_cur *cur; - cur = kmem_cache_zalloc(cache, GFP_NOFS | __GFP_NOFAIL); + ASSERT(ops->ptr_len == XFS_BTREE_LONG_PTR_LEN || + ops->ptr_len == XFS_BTREE_SHORT_PTR_LEN); + + /* BMBT allocations can come through from non-transactional context. */ + cur = kmem_cache_zalloc(cache, + GFP_KERNEL | __GFP_NOLOCKDEP | __GFP_NOFAIL); + cur->bc_ops = ops; cur->bc_tp = tp; cur->bc_mp = mp; - cur->bc_btnum = btnum; cur->bc_maxlevels = maxlevels; cur->bc_cache = cache; @@ -740,4 +692,14 @@ void xfs_btree_destroy_cur_caches(void); int xfs_btree_goto_left_edge(struct xfs_btree_cur *cur); +/* Does this level of the cursor point to the inode root (and not a block)? */ +static inline bool +xfs_btree_at_iroot( + const struct xfs_btree_cur *cur, + int level) +{ + return cur->bc_ops->type == XFS_BTREE_TYPE_INODE && + level == cur->bc_nlevels - 1; +} + #endif /* __XFS_BTREE_H__ */ diff --git a/fs/xfs/libxfs/xfs_btree_mem.c b/fs/xfs/libxfs/xfs_btree_mem.c new file mode 100644 index 000000000000..036061fe32cc --- /dev/null +++ b/fs/xfs/libxfs/xfs_btree_mem.c @@ -0,0 +1,347 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * Copyright (c) 2021-2024 Oracle. All Rights Reserved. + * Author: Darrick J. Wong <djwong@kernel.org> + */ +#include "xfs.h" +#include "xfs_fs.h" +#include "xfs_shared.h" +#include "xfs_format.h" +#include "xfs_log_format.h" +#include "xfs_trans_resv.h" +#include "xfs_mount.h" +#include "xfs_trans.h" +#include "xfs_btree.h" +#include "xfs_error.h" +#include "xfs_buf_mem.h" +#include "xfs_btree_mem.h" +#include "xfs_ag.h" +#include "xfs_buf_item.h" +#include "xfs_trace.h" + +/* Set the root of an in-memory btree. */ +void +xfbtree_set_root( + struct xfs_btree_cur *cur, + const union xfs_btree_ptr *ptr, + int inc) +{ + ASSERT(cur->bc_ops->type == XFS_BTREE_TYPE_MEM); + + cur->bc_mem.xfbtree->root = *ptr; + cur->bc_mem.xfbtree->nlevels += inc; +} + +/* Initialize a pointer from the in-memory btree header. */ +void +xfbtree_init_ptr_from_cur( + struct xfs_btree_cur *cur, + union xfs_btree_ptr *ptr) +{ + ASSERT(cur->bc_ops->type == XFS_BTREE_TYPE_MEM); + + *ptr = cur->bc_mem.xfbtree->root; +} + +/* Duplicate an in-memory btree cursor. */ +struct xfs_btree_cur * +xfbtree_dup_cursor( + struct xfs_btree_cur *cur) +{ + struct xfs_btree_cur *ncur; + + ASSERT(cur->bc_ops->type == XFS_BTREE_TYPE_MEM); + + ncur = xfs_btree_alloc_cursor(cur->bc_mp, cur->bc_tp, cur->bc_ops, + cur->bc_maxlevels, cur->bc_cache); + ncur->bc_flags = cur->bc_flags; + ncur->bc_nlevels = cur->bc_nlevels; + ncur->bc_mem.xfbtree = cur->bc_mem.xfbtree; + + if (cur->bc_mem.pag) + ncur->bc_mem.pag = xfs_perag_hold(cur->bc_mem.pag); + + return ncur; +} + +/* Close the btree xfile and release all resources. */ +void +xfbtree_destroy( + struct xfbtree *xfbt) +{ + xfs_buftarg_drain(xfbt->target); +} + +/* Compute the number of bytes available for records. */ +static inline unsigned int +xfbtree_rec_bytes( + struct xfs_mount *mp, + const struct xfs_btree_ops *ops) +{ + return XMBUF_BLOCKSIZE - XFS_BTREE_LBLOCK_CRC_LEN; +} + +/* Initialize an empty leaf block as the btree root. */ +STATIC int +xfbtree_init_leaf_block( + struct xfs_mount *mp, + struct xfbtree *xfbt, + const struct xfs_btree_ops *ops) +{ + struct xfs_buf *bp; + xfbno_t bno = xfbt->highest_bno++; + int error; + + error = xfs_buf_get(xfbt->target, xfbno_to_daddr(bno), XFBNO_BBSIZE, + &bp); + if (error) + return error; + + trace_xfbtree_create_root_buf(xfbt, bp); + + bp->b_ops = ops->buf_ops; + xfs_btree_init_buf(mp, bp, ops, 0, 0, xfbt->owner); + xfs_buf_relse(bp); + + xfbt->root.l = cpu_to_be64(bno); + return 0; +} + +/* + * Create an in-memory btree root that can be used with the given xmbuf. + * Callers must set xfbt->owner. + */ +int +xfbtree_init( + struct xfs_mount *mp, + struct xfbtree *xfbt, + struct xfs_buftarg *btp, + const struct xfs_btree_ops *ops) +{ + unsigned int blocklen = xfbtree_rec_bytes(mp, ops); + unsigned int keyptr_len; + int error; + + /* Requires a long-format CRC-format btree */ + if (!xfs_has_crc(mp)) { + ASSERT(xfs_has_crc(mp)); + return -EINVAL; + } + if (ops->ptr_len != XFS_BTREE_LONG_PTR_LEN) { + ASSERT(ops->ptr_len == XFS_BTREE_LONG_PTR_LEN); + return -EINVAL; + } + + memset(xfbt, 0, sizeof(*xfbt)); + xfbt->target = btp; + + /* Set up min/maxrecs for this btree. */ + keyptr_len = ops->key_len + sizeof(__be64); + xfbt->maxrecs[0] = blocklen / ops->rec_len; + xfbt->maxrecs[1] = blocklen / keyptr_len; + xfbt->minrecs[0] = xfbt->maxrecs[0] / 2; + xfbt->minrecs[1] = xfbt->maxrecs[1] / 2; + xfbt->highest_bno = 0; + xfbt->nlevels = 1; + + /* Initialize the empty btree. */ + error = xfbtree_init_leaf_block(mp, xfbt, ops); + if (error) + goto err_freesp; + + trace_xfbtree_init(mp, xfbt, ops); + + return 0; + +err_freesp: + xfs_buftarg_drain(xfbt->target); + return error; +} + +/* Allocate a block to our in-memory btree. */ +int +xfbtree_alloc_block( + struct xfs_btree_cur *cur, + const union xfs_btree_ptr *start, + union xfs_btree_ptr *new, + int *stat) +{ + struct xfbtree *xfbt = cur->bc_mem.xfbtree; + xfbno_t bno = xfbt->highest_bno++; + + ASSERT(cur->bc_ops->type == XFS_BTREE_TYPE_MEM); + + trace_xfbtree_alloc_block(xfbt, cur, bno); + + /* Fail if the block address exceeds the maximum for the buftarg. */ + if (!xfbtree_verify_bno(xfbt, bno)) { + ASSERT(xfbtree_verify_bno(xfbt, bno)); + *stat = 0; + return 0; + } + + new->l = cpu_to_be64(bno); + *stat = 1; + return 0; +} + +/* Free a block from our in-memory btree. */ +int +xfbtree_free_block( + struct xfs_btree_cur *cur, + struct xfs_buf *bp) +{ + struct xfbtree *xfbt = cur->bc_mem.xfbtree; + xfs_daddr_t daddr = xfs_buf_daddr(bp); + xfbno_t bno = xfs_daddr_to_xfbno(daddr); + + ASSERT(cur->bc_ops->type == XFS_BTREE_TYPE_MEM); + + trace_xfbtree_free_block(xfbt, cur, bno); + + if (bno + 1 == xfbt->highest_bno) + xfbt->highest_bno--; + + return 0; +} + +/* Return the minimum number of records for a btree block. */ +int +xfbtree_get_minrecs( + struct xfs_btree_cur *cur, + int level) +{ + struct xfbtree *xfbt = cur->bc_mem.xfbtree; + + return xfbt->minrecs[level != 0]; +} + +/* Return the maximum number of records for a btree block. */ +int +xfbtree_get_maxrecs( + struct xfs_btree_cur *cur, + int level) +{ + struct xfbtree *xfbt = cur->bc_mem.xfbtree; + + return xfbt->maxrecs[level != 0]; +} + +/* If this log item is a buffer item that came from the xfbtree, return it. */ +static inline struct xfs_buf * +xfbtree_buf_match( + struct xfbtree *xfbt, + const struct xfs_log_item *lip) +{ + const struct xfs_buf_log_item *bli; + struct xfs_buf *bp; + + if (lip->li_type != XFS_LI_BUF) + return NULL; + + bli = container_of(lip, struct xfs_buf_log_item, bli_item); + bp = bli->bli_buf; + if (bp->b_target != xfbt->target) + return NULL; + + return bp; +} + +/* + * Commit changes to the incore btree immediately by writing all dirty xfbtree + * buffers to the backing xfile. This detaches all xfbtree buffers from the + * transaction, even on failure. The buffer locks are dropped between the + * delwri queue and submit, so the caller must synchronize btree access. + * + * Normally we'd let the buffers commit with the transaction and get written to + * the xfile via the log, but online repair stages ephemeral btrees in memory + * and uses the btree_staging functions to write new btrees to disk atomically. + * The in-memory btree (and its backing store) are discarded at the end of the + * repair phase, which means that xfbtree buffers cannot commit with the rest + * of a transaction. + * + * In other words, online repair only needs the transaction to collect buffer + * pointers and to avoid buffer deadlocks, not to guarantee consistency of + * updates. + */ +int +xfbtree_trans_commit( + struct xfbtree *xfbt, + struct xfs_trans *tp) +{ + struct xfs_log_item *lip, *n; + bool tp_dirty = false; + int error = 0; + + /* + * For each xfbtree buffer attached to the transaction, write the dirty + * buffers to the xfile and release them. + */ + list_for_each_entry_safe(lip, n, &tp->t_items, li_trans) { + struct xfs_buf *bp = xfbtree_buf_match(xfbt, lip); + + if (!bp) { + if (test_bit(XFS_LI_DIRTY, &lip->li_flags)) + tp_dirty |= true; + continue; + } + + trace_xfbtree_trans_commit_buf(xfbt, bp); + + xmbuf_trans_bdetach(tp, bp); + + /* + * If the buffer fails verification, note the failure but + * continue walking the transaction items so that we remove all + * ephemeral btree buffers. + */ + if (!error) + error = xmbuf_finalize(bp); + + xfs_buf_relse(bp); + } + + /* + * Reset the transaction's dirty flag to reflect the dirty state of the + * log items that are still attached. + */ + tp->t_flags = (tp->t_flags & ~XFS_TRANS_DIRTY) | + (tp_dirty ? XFS_TRANS_DIRTY : 0); + + return error; +} + +/* + * Cancel changes to the incore btree by detaching all the xfbtree buffers. + * Changes are not undone, so callers must not access the btree ever again. + */ +void +xfbtree_trans_cancel( + struct xfbtree *xfbt, + struct xfs_trans *tp) +{ + struct xfs_log_item *lip, *n; + bool tp_dirty = false; + + list_for_each_entry_safe(lip, n, &tp->t_items, li_trans) { + struct xfs_buf *bp = xfbtree_buf_match(xfbt, lip); + + if (!bp) { + if (test_bit(XFS_LI_DIRTY, &lip->li_flags)) + tp_dirty |= true; + continue; + } + + trace_xfbtree_trans_cancel_buf(xfbt, bp); + + xmbuf_trans_bdetach(tp, bp); + xfs_buf_relse(bp); + } + + /* + * Reset the transaction's dirty flag to reflect the dirty state of the + * log items that are still attached. + */ + tp->t_flags = (tp->t_flags & ~XFS_TRANS_DIRTY) | + (tp_dirty ? XFS_TRANS_DIRTY : 0); +} diff --git a/fs/xfs/libxfs/xfs_btree_mem.h b/fs/xfs/libxfs/xfs_btree_mem.h new file mode 100644 index 000000000000..1c3825786ec8 --- /dev/null +++ b/fs/xfs/libxfs/xfs_btree_mem.h @@ -0,0 +1,75 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * Copyright (c) 2021-2024 Oracle. All Rights Reserved. + * Author: Darrick J. Wong <djwong@kernel.org> + */ +#ifndef __XFS_BTREE_MEM_H__ +#define __XFS_BTREE_MEM_H__ + +typedef uint64_t xfbno_t; + +#define XFBNO_BLOCKSIZE (XMBUF_BLOCKSIZE) +#define XFBNO_BBSHIFT (XMBUF_BLOCKSHIFT - BBSHIFT) +#define XFBNO_BBSIZE (XFBNO_BLOCKSIZE >> BBSHIFT) + +static inline xfs_daddr_t xfbno_to_daddr(xfbno_t blkno) +{ + return blkno << XFBNO_BBSHIFT; +} + +static inline xfbno_t xfs_daddr_to_xfbno(xfs_daddr_t daddr) +{ + return daddr >> XFBNO_BBSHIFT; +} + +struct xfbtree { + /* buffer cache target for this in-memory btree */ + struct xfs_buftarg *target; + + /* Highest block number that has been written to. */ + xfbno_t highest_bno; + + /* Owner of this btree. */ + unsigned long long owner; + + /* Btree header */ + union xfs_btree_ptr root; + unsigned int nlevels; + + /* Minimum and maximum records per block. */ + unsigned int maxrecs[2]; + unsigned int minrecs[2]; +}; + +#ifdef CONFIG_XFS_BTREE_IN_MEM +static inline bool xfbtree_verify_bno(struct xfbtree *xfbt, xfbno_t bno) +{ + return xmbuf_verify_daddr(xfbt->target, xfbno_to_daddr(bno)); +} + +void xfbtree_set_root(struct xfs_btree_cur *cur, + const union xfs_btree_ptr *ptr, int inc); +void xfbtree_init_ptr_from_cur(struct xfs_btree_cur *cur, + union xfs_btree_ptr *ptr); +struct xfs_btree_cur *xfbtree_dup_cursor(struct xfs_btree_cur *cur); + +int xfbtree_get_minrecs(struct xfs_btree_cur *cur, int level); +int xfbtree_get_maxrecs(struct xfs_btree_cur *cur, int level); + +int xfbtree_alloc_block(struct xfs_btree_cur *cur, + const union xfs_btree_ptr *start, union xfs_btree_ptr *ptr, + int *stat); +int xfbtree_free_block(struct xfs_btree_cur *cur, struct xfs_buf *bp); + +/* Callers must set xfbt->target and xfbt->owner before calling this */ +int xfbtree_init(struct xfs_mount *mp, struct xfbtree *xfbt, + struct xfs_buftarg *btp, const struct xfs_btree_ops *ops); +void xfbtree_destroy(struct xfbtree *xfbt); + +int xfbtree_trans_commit(struct xfbtree *xfbt, struct xfs_trans *tp); +void xfbtree_trans_cancel(struct xfbtree *xfbt, struct xfs_trans *tp); +#else +# define xfbtree_verify_bno(...) (false) +#endif /* CONFIG_XFS_BTREE_IN_MEM */ + +#endif /* __XFS_BTREE_MEM_H__ */ diff --git a/fs/xfs/libxfs/xfs_btree_staging.c b/fs/xfs/libxfs/xfs_btree_staging.c index e276eba87cb1..694929703152 100644 --- a/fs/xfs/libxfs/xfs_btree_staging.c +++ b/fs/xfs/libxfs/xfs_btree_staging.c @@ -39,63 +39,6 @@ */ /* - * Don't allow staging cursors to be duplicated because they're supposed to be - * kept private to a single thread. - */ -STATIC struct xfs_btree_cur * -xfs_btree_fakeroot_dup_cursor( - struct xfs_btree_cur *cur) -{ - ASSERT(0); - return NULL; -} - -/* - * Don't allow block allocation for a staging cursor, because staging cursors - * do not support regular btree modifications. - * - * Bulk loading uses a separate callback to obtain new blocks from a - * preallocated list, which prevents ENOSPC failures during loading. - */ -STATIC int -xfs_btree_fakeroot_alloc_block( - struct xfs_btree_cur *cur, - const union xfs_btree_ptr *start_bno, - union xfs_btree_ptr *new_bno, - int *stat) -{ - ASSERT(0); - return -EFSCORRUPTED; -} - -/* - * Don't allow block freeing for a staging cursor, because staging cursors - * do not support regular btree modifications. - */ -STATIC int -xfs_btree_fakeroot_free_block( - struct xfs_btree_cur *cur, - struct xfs_buf *bp) -{ - ASSERT(0); - return -EFSCORRUPTED; -} - -/* Initialize a pointer to the root block from the fakeroot. */ -STATIC void -xfs_btree_fakeroot_init_ptr_from_cur( - struct xfs_btree_cur *cur, - union xfs_btree_ptr *ptr) -{ - struct xbtree_afakeroot *afake; - - ASSERT(cur->bc_flags & XFS_BTREE_STAGING); - - afake = cur->bc_ag.afake; - ptr->s = cpu_to_be32(afake->af_root); -} - -/* * Bulk Loading for AG Btrees * ========================== * @@ -109,47 +52,20 @@ xfs_btree_fakeroot_init_ptr_from_cur( * cursor into a regular btree cursor. */ -/* Update the btree root information for a per-AG fake root. */ -STATIC void -xfs_btree_afakeroot_set_root( - struct xfs_btree_cur *cur, - const union xfs_btree_ptr *ptr, - int inc) -{ - struct xbtree_afakeroot *afake = cur->bc_ag.afake; - - ASSERT(cur->bc_flags & XFS_BTREE_STAGING); - afake->af_root = be32_to_cpu(ptr->s); - afake->af_levels += inc; -} - /* * Initialize a AG-rooted btree cursor with the given AG btree fake root. - * The btree cursor's bc_ops will be overridden as needed to make the staging - * functionality work. */ void xfs_btree_stage_afakeroot( struct xfs_btree_cur *cur, struct xbtree_afakeroot *afake) { - struct xfs_btree_ops *nops; - ASSERT(!(cur->bc_flags & XFS_BTREE_STAGING)); - ASSERT(!(cur->bc_flags & XFS_BTREE_ROOT_IN_INODE)); + ASSERT(cur->bc_ops->type != XFS_BTREE_TYPE_INODE); ASSERT(cur->bc_tp == NULL); - nops = kmem_alloc(sizeof(struct xfs_btree_ops), KM_NOFS); - memcpy(nops, cur->bc_ops, sizeof(struct xfs_btree_ops)); - nops->alloc_block = xfs_btree_fakeroot_alloc_block; - nops->free_block = xfs_btree_fakeroot_free_block; - nops->init_ptr_from_cur = xfs_btree_fakeroot_init_ptr_from_cur; - nops->set_root = xfs_btree_afakeroot_set_root; - nops->dup_cursor = xfs_btree_fakeroot_dup_cursor; - cur->bc_ag.afake = afake; cur->bc_nlevels = afake->af_levels; - cur->bc_ops = nops; cur->bc_flags |= XFS_BTREE_STAGING; } @@ -163,17 +79,15 @@ void xfs_btree_commit_afakeroot( struct xfs_btree_cur *cur, struct xfs_trans *tp, - struct xfs_buf *agbp, - const struct xfs_btree_ops *ops) + struct xfs_buf *agbp) { ASSERT(cur->bc_flags & XFS_BTREE_STAGING); ASSERT(cur->bc_tp == NULL); trace_xfs_btree_commit_afakeroot(cur); - kmem_free((void *)cur->bc_ops); + cur->bc_ag.afake = NULL; cur->bc_ag.agbp = agbp; - cur->bc_ops = ops; cur->bc_flags &= ~XFS_BTREE_STAGING; cur->bc_tp = tp; } @@ -211,29 +125,16 @@ xfs_btree_commit_afakeroot( void xfs_btree_stage_ifakeroot( struct xfs_btree_cur *cur, - struct xbtree_ifakeroot *ifake, - struct xfs_btree_ops **new_ops) + struct xbtree_ifakeroot *ifake) { - struct xfs_btree_ops *nops; - ASSERT(!(cur->bc_flags & XFS_BTREE_STAGING)); - ASSERT(cur->bc_flags & XFS_BTREE_ROOT_IN_INODE); + ASSERT(cur->bc_ops->type == XFS_BTREE_TYPE_INODE); ASSERT(cur->bc_tp == NULL); - nops = kmem_alloc(sizeof(struct xfs_btree_ops), KM_NOFS); - memcpy(nops, cur->bc_ops, sizeof(struct xfs_btree_ops)); - nops->alloc_block = xfs_btree_fakeroot_alloc_block; - nops->free_block = xfs_btree_fakeroot_free_block; - nops->init_ptr_from_cur = xfs_btree_fakeroot_init_ptr_from_cur; - nops->dup_cursor = xfs_btree_fakeroot_dup_cursor; - cur->bc_ino.ifake = ifake; cur->bc_nlevels = ifake->if_levels; - cur->bc_ops = nops; + cur->bc_ino.forksize = ifake->if_fork_size; cur->bc_flags |= XFS_BTREE_STAGING; - - if (new_ops) - *new_ops = nops; } /* @@ -246,18 +147,15 @@ void xfs_btree_commit_ifakeroot( struct xfs_btree_cur *cur, struct xfs_trans *tp, - int whichfork, - const struct xfs_btree_ops *ops) + int whichfork) { ASSERT(cur->bc_flags & XFS_BTREE_STAGING); ASSERT(cur->bc_tp == NULL); trace_xfs_btree_commit_ifakeroot(cur); - kmem_free((void *)cur->bc_ops); cur->bc_ino.ifake = NULL; cur->bc_ino.whichfork = whichfork; - cur->bc_ops = ops; cur->bc_flags &= ~XFS_BTREE_STAGING; cur->bc_tp = tp; } @@ -397,8 +295,7 @@ xfs_btree_bload_prep_block( struct xfs_btree_block *new_block; int ret; - if ((cur->bc_flags & XFS_BTREE_ROOT_IN_INODE) && - level == cur->bc_nlevels - 1) { + if (xfs_btree_at_iroot(cur, level)) { struct xfs_ifork *ifp = xfs_btree_ifork_ptr(cur); size_t new_size; @@ -406,14 +303,12 @@ xfs_btree_bload_prep_block( /* Allocate a new incore btree root block. */ new_size = bbl->iroot_size(cur, level, nr_this_block, priv); - ifp->if_broot = kmem_zalloc(new_size, 0); + ifp->if_broot = kzalloc(new_size, GFP_KERNEL | __GFP_NOFAIL); ifp->if_broot_bytes = (int)new_size; /* Initialize it and send it out. */ - xfs_btree_init_block_int(cur->bc_mp, ifp->if_broot, - XFS_BUF_DADDR_NULL, cur->bc_btnum, level, - nr_this_block, cur->bc_ino.ip->i_ino, - cur->bc_flags); + xfs_btree_init_block(cur->bc_mp, ifp->if_broot, cur->bc_ops, + level, nr_this_block, cur->bc_ino.ip->i_ino); *bpp = NULL; *blockp = ifp->if_broot; @@ -704,7 +599,7 @@ xfs_btree_bload_compute_geometry( xfs_btree_bload_level_geometry(cur, bbl, level, nr_this_level, &avg_per_block, &level_blocks, &dontcare64); - if (cur->bc_flags & XFS_BTREE_ROOT_IN_INODE) { + if (cur->bc_ops->type == XFS_BTREE_TYPE_INODE) { /* * If all the items we want to store at this level * would fit in the inode root block, then we have our @@ -763,7 +658,7 @@ xfs_btree_bload_compute_geometry( return -EOVERFLOW; bbl->btree_height = cur->bc_nlevels; - if (cur->bc_flags & XFS_BTREE_ROOT_IN_INODE) + if (cur->bc_ops->type == XFS_BTREE_TYPE_INODE) bbl->nr_blocks = nr_blocks - 1; else bbl->nr_blocks = nr_blocks; @@ -890,7 +785,7 @@ xfs_btree_bload( } /* Initialize the new root. */ - if (cur->bc_flags & XFS_BTREE_ROOT_IN_INODE) { + if (cur->bc_ops->type == XFS_BTREE_TYPE_INODE) { ASSERT(xfs_btree_ptr_is_null(cur, &ptr)); cur->bc_ino.ifake->if_levels = cur->bc_nlevels; cur->bc_ino.ifake->if_blocks = total_blocks - 1; diff --git a/fs/xfs/libxfs/xfs_btree_staging.h b/fs/xfs/libxfs/xfs_btree_staging.h index 055ea43b1e18..0c9c2ffb127a 100644 --- a/fs/xfs/libxfs/xfs_btree_staging.h +++ b/fs/xfs/libxfs/xfs_btree_staging.h @@ -22,7 +22,7 @@ struct xbtree_afakeroot { void xfs_btree_stage_afakeroot(struct xfs_btree_cur *cur, struct xbtree_afakeroot *afake); void xfs_btree_commit_afakeroot(struct xfs_btree_cur *cur, struct xfs_trans *tp, - struct xfs_buf *agbp, const struct xfs_btree_ops *ops); + struct xfs_buf *agbp); /* Fake root for an inode-rooted btree. */ struct xbtree_ifakeroot { @@ -41,10 +41,9 @@ struct xbtree_ifakeroot { /* Cursor interactions with fake roots for inode-rooted btrees. */ void xfs_btree_stage_ifakeroot(struct xfs_btree_cur *cur, - struct xbtree_ifakeroot *ifake, - struct xfs_btree_ops **new_ops); + struct xbtree_ifakeroot *ifake); void xfs_btree_commit_ifakeroot(struct xfs_btree_cur *cur, struct xfs_trans *tp, - int whichfork, const struct xfs_btree_ops *ops); + int whichfork); /* Bulk loading of staged btrees. */ typedef int (*xfs_btree_bload_get_records_fn)(struct xfs_btree_cur *cur, @@ -76,8 +75,7 @@ struct xfs_btree_bload { /* * This function should return the size of the in-core btree root - * block. It is only necessary for XFS_BTREE_ROOT_IN_INODE btree - * types. + * block. It is only necessary for XFS_BTREE_TYPE_INODE btrees. */ xfs_btree_bload_iroot_size_fn iroot_size; diff --git a/fs/xfs/libxfs/xfs_da_btree.c b/fs/xfs/libxfs/xfs_da_btree.c index 5457188bb4de..718d071bb21a 100644 --- a/fs/xfs/libxfs/xfs_da_btree.c +++ b/fs/xfs/libxfs/xfs_da_btree.c @@ -23,6 +23,7 @@ #include "xfs_buf_item.h" #include "xfs_log.h" #include "xfs_errortag.h" +#include "xfs_health.h" /* * xfs_da_btree.c @@ -85,7 +86,8 @@ xfs_da_state_alloc( { struct xfs_da_state *state; - state = kmem_cache_zalloc(xfs_da_state_cache, GFP_NOFS | __GFP_NOFAIL); + state = kmem_cache_zalloc(xfs_da_state_cache, + GFP_KERNEL | __GFP_NOLOCKDEP | __GFP_NOFAIL); state->args = args; state->mp = args->dp->i_mount; return state; @@ -352,6 +354,8 @@ const struct xfs_buf_ops xfs_da3_node_buf_ops = { static int xfs_da3_node_set_type( struct xfs_trans *tp, + struct xfs_inode *dp, + int whichfork, struct xfs_buf *bp) { struct xfs_da_blkinfo *info = bp->b_addr; @@ -373,6 +377,7 @@ xfs_da3_node_set_type( XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, tp->t_mountp, info, sizeof(*info)); xfs_trans_brelse(tp, bp); + xfs_dirattr_mark_sick(dp, whichfork); return -EFSCORRUPTED; } } @@ -391,7 +396,7 @@ xfs_da3_node_read( &xfs_da3_node_buf_ops); if (error || !*bpp || !tp) return error; - return xfs_da3_node_set_type(tp, *bpp); + return xfs_da3_node_set_type(tp, dp, whichfork, *bpp); } int @@ -408,6 +413,8 @@ xfs_da3_node_read_mapped( error = xfs_trans_read_buf(mp, tp, mp->m_ddev_targp, mappedbno, XFS_FSB_TO_BB(mp, xfs_dabuf_nfsb(mp, whichfork)), 0, bpp, &xfs_da3_node_buf_ops); + if (xfs_metadata_is_sick(error)) + xfs_dirattr_mark_sick(dp, whichfork); if (error || !*bpp) return error; @@ -418,7 +425,7 @@ xfs_da3_node_read_mapped( if (!tp) return 0; - return xfs_da3_node_set_type(tp, *bpp); + return xfs_da3_node_set_type(tp, dp, whichfork, *bpp); } /* @@ -631,6 +638,7 @@ xfs_da3_split( if (node->hdr.info.forw) { if (be32_to_cpu(node->hdr.info.forw) != addblk->blkno) { xfs_buf_mark_corrupt(oldblk->bp); + xfs_da_mark_sick(state->args); error = -EFSCORRUPTED; goto out; } @@ -644,6 +652,7 @@ xfs_da3_split( if (node->hdr.info.back) { if (be32_to_cpu(node->hdr.info.back) != addblk->blkno) { xfs_buf_mark_corrupt(oldblk->bp); + xfs_da_mark_sick(state->args); error = -EFSCORRUPTED; goto out; } @@ -1635,6 +1644,7 @@ xfs_da3_node_lookup_int( if (magic != XFS_DA_NODE_MAGIC && magic != XFS_DA3_NODE_MAGIC) { xfs_buf_mark_corrupt(blk->bp); + xfs_da_mark_sick(args); return -EFSCORRUPTED; } @@ -1650,6 +1660,7 @@ xfs_da3_node_lookup_int( /* Tree taller than we can handle; bail out! */ if (nodehdr.level >= XFS_DA_NODE_MAXDEPTH) { xfs_buf_mark_corrupt(blk->bp); + xfs_da_mark_sick(args); return -EFSCORRUPTED; } @@ -1658,6 +1669,7 @@ xfs_da3_node_lookup_int( expected_level = nodehdr.level - 1; else if (expected_level != nodehdr.level) { xfs_buf_mark_corrupt(blk->bp); + xfs_da_mark_sick(args); return -EFSCORRUPTED; } else expected_level--; @@ -1709,12 +1721,16 @@ xfs_da3_node_lookup_int( } /* We can't point back to the root. */ - if (XFS_IS_CORRUPT(dp->i_mount, blkno == args->geo->leafblk)) + if (XFS_IS_CORRUPT(dp->i_mount, blkno == args->geo->leafblk)) { + xfs_da_mark_sick(args); return -EFSCORRUPTED; + } } - if (XFS_IS_CORRUPT(dp->i_mount, expected_level != 0)) + if (XFS_IS_CORRUPT(dp->i_mount, expected_level != 0)) { + xfs_da_mark_sick(args); return -EFSCORRUPTED; + } /* * A leaf block that ends in the hashval that we are interested in @@ -1732,6 +1748,7 @@ xfs_da3_node_lookup_int( args->blkno = blk->blkno; } else { ASSERT(0); + xfs_da_mark_sick(args); return -EFSCORRUPTED; } if (((retval == -ENOENT) || (retval == -ENOATTR)) && @@ -2182,7 +2199,8 @@ xfs_da_grow_inode_int( * If we didn't get it and the block might work if fragmented, * try without the CONTIG flag. Loop until we get it all. */ - mapp = kmem_alloc(sizeof(*mapp) * count, 0); + mapp = kmalloc(sizeof(*mapp) * count, + GFP_KERNEL | __GFP_NOFAIL); for (b = *bno, mapi = 0; b < *bno + count; ) { c = (int)(*bno + count - b); nmap = min(XFS_BMAP_MAX_NMAP, c); @@ -2219,7 +2237,7 @@ xfs_da_grow_inode_int( out_free_map: if (mapp != &map) - kmem_free(mapp); + kfree(mapp); return error; } @@ -2297,8 +2315,10 @@ xfs_da3_swap_lastblock( error = xfs_bmap_last_before(tp, dp, &lastoff, w); if (error) return error; - if (XFS_IS_CORRUPT(mp, lastoff == 0)) + if (XFS_IS_CORRUPT(mp, lastoff == 0)) { + xfs_da_mark_sick(args); return -EFSCORRUPTED; + } /* * Read the last block in the btree space. */ @@ -2348,6 +2368,7 @@ xfs_da3_swap_lastblock( if (XFS_IS_CORRUPT(mp, be32_to_cpu(sib_info->forw) != last_blkno || sib_info->magic != dead_info->magic)) { + xfs_da_mark_sick(args); error = -EFSCORRUPTED; goto done; } @@ -2368,6 +2389,7 @@ xfs_da3_swap_lastblock( if (XFS_IS_CORRUPT(mp, be32_to_cpu(sib_info->back) != last_blkno || sib_info->magic != dead_info->magic)) { + xfs_da_mark_sick(args); error = -EFSCORRUPTED; goto done; } @@ -2390,6 +2412,7 @@ xfs_da3_swap_lastblock( xfs_da3_node_hdr_from_disk(dp->i_mount, &par_hdr, par_node); if (XFS_IS_CORRUPT(mp, level >= 0 && level != par_hdr.level + 1)) { + xfs_da_mark_sick(args); error = -EFSCORRUPTED; goto done; } @@ -2401,6 +2424,7 @@ xfs_da3_swap_lastblock( entno++) continue; if (XFS_IS_CORRUPT(mp, entno == par_hdr.count)) { + xfs_da_mark_sick(args); error = -EFSCORRUPTED; goto done; } @@ -2426,6 +2450,7 @@ xfs_da3_swap_lastblock( xfs_trans_brelse(tp, par_buf); par_buf = NULL; if (XFS_IS_CORRUPT(mp, par_blkno == 0)) { + xfs_da_mark_sick(args); error = -EFSCORRUPTED; goto done; } @@ -2435,6 +2460,7 @@ xfs_da3_swap_lastblock( par_node = par_buf->b_addr; xfs_da3_node_hdr_from_disk(dp->i_mount, &par_hdr, par_node); if (XFS_IS_CORRUPT(mp, par_hdr.level != level)) { + xfs_da_mark_sick(args); error = -EFSCORRUPTED; goto done; } @@ -2518,7 +2544,8 @@ xfs_dabuf_map( int error = 0, nirecs, i; if (nfsb > 1) - irecs = kmem_zalloc(sizeof(irec) * nfsb, KM_NOFS); + irecs = kzalloc(sizeof(irec) * nfsb, + GFP_KERNEL | __GFP_NOLOCKDEP | __GFP_NOFAIL); nirecs = nfsb; error = xfs_bmapi_read(dp, bno, nfsb, irecs, &nirecs, @@ -2531,7 +2558,8 @@ xfs_dabuf_map( * larger one that needs to be free by the caller. */ if (nirecs > 1) { - map = kmem_zalloc(nirecs * sizeof(struct xfs_buf_map), KM_NOFS); + map = kzalloc(nirecs * sizeof(struct xfs_buf_map), + GFP_KERNEL | __GFP_NOLOCKDEP | __GFP_NOFAIL); if (!map) { error = -ENOMEM; goto out_free_irecs; @@ -2557,12 +2585,13 @@ xfs_dabuf_map( *nmaps = nirecs; out_free_irecs: if (irecs != &irec) - kmem_free(irecs); + kfree(irecs); return error; invalid_mapping: /* Caller ok with no mapping. */ if (XFS_IS_CORRUPT(mp, !(flags & XFS_DABUF_MAP_HOLE_OK))) { + xfs_dirattr_mark_sick(dp, whichfork); error = -EFSCORRUPTED; if (xfs_error_level >= XFS_ERRLEVEL_LOW) { xfs_alert(mp, "%s: bno %u inode %llu", @@ -2613,7 +2642,7 @@ xfs_da_get_buf( out_free: if (mapp != &map) - kmem_free(mapp); + kfree(mapp); return error; } @@ -2644,6 +2673,8 @@ xfs_da_read_buf( error = xfs_trans_read_buf_map(mp, tp, mp->m_ddev_targp, mapp, nmap, 0, &bp, ops); + if (xfs_metadata_is_sick(error)) + xfs_dirattr_mark_sick(dp, whichfork); if (error) goto out_free; @@ -2654,7 +2685,7 @@ xfs_da_read_buf( *bpp = bp; out_free: if (mapp != &map) - kmem_free(mapp); + kfree(mapp); return error; } @@ -2685,7 +2716,7 @@ xfs_da_reada_buf( out_free: if (mapp != &map) - kmem_free(mapp); + kfree(mapp); return error; } diff --git a/fs/xfs/libxfs/xfs_da_format.h b/fs/xfs/libxfs/xfs_da_format.h index 24f9d1461f9a..060e5c96b70f 100644 --- a/fs/xfs/libxfs/xfs_da_format.h +++ b/fs/xfs/libxfs/xfs_da_format.h @@ -159,6 +159,17 @@ struct xfs_da3_intnode { #define XFS_DIR3_FT_MAX 9 +#define XFS_DIR3_FTYPE_STR \ + { XFS_DIR3_FT_UNKNOWN, "unknown" }, \ + { XFS_DIR3_FT_REG_FILE, "file" }, \ + { XFS_DIR3_FT_DIR, "directory" }, \ + { XFS_DIR3_FT_CHRDEV, "char" }, \ + { XFS_DIR3_FT_BLKDEV, "block" }, \ + { XFS_DIR3_FT_FIFO, "fifo" }, \ + { XFS_DIR3_FT_SOCK, "sock" }, \ + { XFS_DIR3_FT_SYMLINK, "symlink" }, \ + { XFS_DIR3_FT_WHT, "whiteout" } + /* * Byte offset in data block and shortform entry. */ diff --git a/fs/xfs/libxfs/xfs_defer.c b/fs/xfs/libxfs/xfs_defer.c index 66a17910d021..c13276095cc0 100644 --- a/fs/xfs/libxfs/xfs_defer.c +++ b/fs/xfs/libxfs/xfs_defer.c @@ -819,16 +819,16 @@ xfs_defer_can_append( /* Create a new pending item at the end of the transaction list. */ static inline struct xfs_defer_pending * xfs_defer_alloc( - struct xfs_trans *tp, + struct list_head *dfops, const struct xfs_defer_op_type *ops) { struct xfs_defer_pending *dfp; dfp = kmem_cache_zalloc(xfs_defer_pending_cache, - GFP_NOFS | __GFP_NOFAIL); + GFP_KERNEL | __GFP_NOFAIL); dfp->dfp_ops = ops; INIT_LIST_HEAD(&dfp->dfp_work); - list_add_tail(&dfp->dfp_list, &tp->t_dfops); + list_add_tail(&dfp->dfp_list, dfops); return dfp; } @@ -846,7 +846,7 @@ xfs_defer_add( dfp = xfs_defer_find_last(tp, ops); if (!dfp || !xfs_defer_can_append(dfp, ops)) - dfp = xfs_defer_alloc(tp, ops); + dfp = xfs_defer_alloc(&tp->t_dfops, ops); xfs_defer_add_item(dfp, li); trace_xfs_defer_add_item(tp->t_mountp, dfp, li); @@ -870,7 +870,7 @@ xfs_defer_add_barrier( if (dfp) return; - xfs_defer_alloc(tp, &xfs_barrier_defer_type); + xfs_defer_alloc(&tp->t_dfops, &xfs_barrier_defer_type); trace_xfs_defer_add_item(tp->t_mountp, dfp, NULL); } @@ -885,14 +885,9 @@ xfs_defer_start_recovery( struct list_head *r_dfops, const struct xfs_defer_op_type *ops) { - struct xfs_defer_pending *dfp; + struct xfs_defer_pending *dfp = xfs_defer_alloc(r_dfops, ops); - dfp = kmem_cache_zalloc(xfs_defer_pending_cache, - GFP_NOFS | __GFP_NOFAIL); - dfp->dfp_ops = ops; dfp->dfp_intent = lip; - INIT_LIST_HEAD(&dfp->dfp_work); - list_add_tail(&dfp->dfp_list, r_dfops); } /* @@ -979,7 +974,7 @@ xfs_defer_ops_capture( return ERR_PTR(error); /* Create an object to capture the defer ops. */ - dfc = kmem_zalloc(sizeof(*dfc), KM_NOFS); + dfc = kzalloc(sizeof(*dfc), GFP_KERNEL | __GFP_NOFAIL); INIT_LIST_HEAD(&dfc->dfc_list); INIT_LIST_HEAD(&dfc->dfc_dfops); @@ -1011,7 +1006,7 @@ xfs_defer_ops_capture( * transaction. */ for (i = 0; i < dfc->dfc_held.dr_inos; i++) { - ASSERT(xfs_isilocked(dfc->dfc_held.dr_ip[i], XFS_ILOCK_EXCL)); + xfs_assert_ilocked(dfc->dfc_held.dr_ip[i], XFS_ILOCK_EXCL); ihold(VFS_I(dfc->dfc_held.dr_ip[i])); } @@ -1038,7 +1033,7 @@ xfs_defer_ops_capture_abort( for (i = 0; i < dfc->dfc_held.dr_inos; i++) xfs_irele(dfc->dfc_held.dr_ip[i]); - kmem_free(dfc); + kfree(dfc); } /* @@ -1114,7 +1109,7 @@ xfs_defer_ops_continue( list_splice_init(&dfc->dfc_dfops, &tp->t_dfops); tp->t_flags |= dfc->dfc_tpflags; - kmem_free(dfc); + kfree(dfc); } /* Release the resources captured and continued during recovery. */ diff --git a/fs/xfs/libxfs/xfs_dir2.c b/fs/xfs/libxfs/xfs_dir2.c index a76673281514..4821519efad4 100644 --- a/fs/xfs/libxfs/xfs_dir2.c +++ b/fs/xfs/libxfs/xfs_dir2.c @@ -18,6 +18,7 @@ #include "xfs_errortag.h" #include "xfs_error.h" #include "xfs_trace.h" +#include "xfs_health.h" const struct xfs_name xfs_name_dotdot = { .name = (const unsigned char *)"..", @@ -25,6 +26,12 @@ const struct xfs_name xfs_name_dotdot = { .type = XFS_DIR3_FT_DIR, }; +const struct xfs_name xfs_name_dot = { + .name = (const unsigned char *)".", + .len = 1, + .type = XFS_DIR3_FT_DIR, +}; + /* * Convert inode mode to directory entry filetype */ @@ -104,13 +111,13 @@ xfs_da_mount( ASSERT(mp->m_sb.sb_versionnum & XFS_SB_VERSION_DIRV2BIT); ASSERT(xfs_dir2_dirblock_bytes(&mp->m_sb) <= XFS_MAX_BLOCKSIZE); - mp->m_dir_geo = kmem_zalloc(sizeof(struct xfs_da_geometry), - KM_MAYFAIL); - mp->m_attr_geo = kmem_zalloc(sizeof(struct xfs_da_geometry), - KM_MAYFAIL); + mp->m_dir_geo = kzalloc(sizeof(struct xfs_da_geometry), + GFP_KERNEL | __GFP_RETRY_MAYFAIL); + mp->m_attr_geo = kzalloc(sizeof(struct xfs_da_geometry), + GFP_KERNEL | __GFP_RETRY_MAYFAIL); if (!mp->m_dir_geo || !mp->m_attr_geo) { - kmem_free(mp->m_dir_geo); - kmem_free(mp->m_attr_geo); + kfree(mp->m_dir_geo); + kfree(mp->m_attr_geo); return -ENOMEM; } @@ -178,8 +185,8 @@ void xfs_da_unmount( struct xfs_mount *mp) { - kmem_free(mp->m_dir_geo); - kmem_free(mp->m_attr_geo); + kfree(mp->m_dir_geo); + kfree(mp->m_attr_geo); } /* @@ -236,7 +243,7 @@ xfs_dir_init( if (error) return error; - args = kmem_zalloc(sizeof(*args), KM_NOFS); + args = kzalloc(sizeof(*args), GFP_KERNEL | __GFP_NOFAIL); if (!args) return -ENOMEM; @@ -244,7 +251,7 @@ xfs_dir_init( args->dp = dp; args->trans = tp; error = xfs_dir2_sf_create(args, pdp->i_ino); - kmem_free(args); + kfree(args); return error; } @@ -273,7 +280,7 @@ xfs_dir_createname( XFS_STATS_INC(dp->i_mount, xs_dir_create); } - args = kmem_zalloc(sizeof(*args), KM_NOFS); + args = kzalloc(sizeof(*args), GFP_KERNEL | __GFP_NOFAIL); if (!args) return -ENOMEM; @@ -313,7 +320,7 @@ xfs_dir_createname( rval = xfs_dir2_node_addname(args); out_free: - kmem_free(args); + kfree(args); return rval; } @@ -333,7 +340,8 @@ xfs_dir_cilookup_result( !(args->op_flags & XFS_DA_OP_CILOOKUP)) return -EEXIST; - args->value = kmem_alloc(len, KM_NOFS | KM_MAYFAIL); + args->value = kmalloc(len, + GFP_KERNEL | __GFP_NOLOCKDEP | __GFP_RETRY_MAYFAIL); if (!args->value) return -ENOMEM; @@ -364,15 +372,8 @@ xfs_dir_lookup( ASSERT(S_ISDIR(VFS_I(dp)->i_mode)); XFS_STATS_INC(dp->i_mount, xs_dir_lookup); - /* - * We need to use KM_NOFS here so that lockdep will not throw false - * positive deadlock warnings on a non-transactional lookup path. It is - * safe to recurse into inode recalim in that case, but lockdep can't - * easily be taught about it. Hence KM_NOFS avoids having to add more - * lockdep Doing this avoids having to add a bunch of lockdep class - * annotations into the reclaim path for the ilock. - */ - args = kmem_zalloc(sizeof(*args), KM_NOFS); + args = kzalloc(sizeof(*args), + GFP_KERNEL | __GFP_NOLOCKDEP | __GFP_NOFAIL); args->geo = dp->i_mount->m_dir_geo; args->name = name->name; args->namelen = name->len; @@ -419,7 +420,7 @@ out_check_rval: } out_free: xfs_iunlock(dp, lock_mode); - kmem_free(args); + kfree(args); return rval; } @@ -441,7 +442,7 @@ xfs_dir_removename( ASSERT(S_ISDIR(VFS_I(dp)->i_mode)); XFS_STATS_INC(dp->i_mount, xs_dir_remove); - args = kmem_zalloc(sizeof(*args), KM_NOFS); + args = kzalloc(sizeof(*args), GFP_KERNEL | __GFP_NOFAIL); if (!args) return -ENOMEM; @@ -477,7 +478,7 @@ xfs_dir_removename( else rval = xfs_dir2_node_removename(args); out_free: - kmem_free(args); + kfree(args); return rval; } @@ -502,7 +503,7 @@ xfs_dir_replace( if (rval) return rval; - args = kmem_zalloc(sizeof(*args), KM_NOFS); + args = kzalloc(sizeof(*args), GFP_KERNEL | __GFP_NOFAIL); if (!args) return -ENOMEM; @@ -538,7 +539,7 @@ xfs_dir_replace( else rval = xfs_dir2_node_replace(args); out_free: - kmem_free(args); + kfree(args); return rval; } @@ -626,8 +627,10 @@ xfs_dir2_isblock( return 0; *isblock = true; - if (XFS_IS_CORRUPT(mp, args->dp->i_disk_size != args->geo->blksize)) + if (XFS_IS_CORRUPT(mp, args->dp->i_disk_size != args->geo->blksize)) { + xfs_da_mark_sick(args); return -EFSCORRUPTED; + } return 0; } diff --git a/fs/xfs/libxfs/xfs_dir2.h b/fs/xfs/libxfs/xfs_dir2.h index 19af22a16c41..8497d041f316 100644 --- a/fs/xfs/libxfs/xfs_dir2.h +++ b/fs/xfs/libxfs/xfs_dir2.h @@ -22,6 +22,19 @@ struct xfs_dir3_icfree_hdr; struct xfs_dir3_icleaf_hdr; extern const struct xfs_name xfs_name_dotdot; +extern const struct xfs_name xfs_name_dot; + +static inline bool +xfs_dir2_samename( + const struct xfs_name *n1, + const struct xfs_name *n2) +{ + if (n1 == n2) + return true; + if (n1->len != n2->len) + return false; + return !memcmp(n1->name, n2->name, n1->len); +} /* * Convert inode mode to directory entry filetype diff --git a/fs/xfs/libxfs/xfs_dir2_block.c b/fs/xfs/libxfs/xfs_dir2_block.c index 3c256d4cc40b..a2da007adb46 100644 --- a/fs/xfs/libxfs/xfs_dir2_block.c +++ b/fs/xfs/libxfs/xfs_dir2_block.c @@ -20,6 +20,7 @@ #include "xfs_error.h" #include "xfs_trace.h" #include "xfs_log.h" +#include "xfs_health.h" /* * Local function prototypes. @@ -152,6 +153,7 @@ xfs_dir3_block_read( __xfs_buf_mark_corrupt(*bpp, fa); xfs_trans_brelse(tp, *bpp); *bpp = NULL; + xfs_dirattr_mark_sick(dp, XFS_DATA_FORK); return -EFSCORRUPTED; } @@ -1108,7 +1110,7 @@ xfs_dir2_sf_to_block( * Copy the directory into a temporary buffer. * Then pitch the incore inode data so we can make extents. */ - sfp = kmem_alloc(ifp->if_bytes, 0); + sfp = kmalloc(ifp->if_bytes, GFP_KERNEL | __GFP_NOFAIL); memcpy(sfp, oldsfp, ifp->if_bytes); xfs_idata_realloc(dp, -ifp->if_bytes, XFS_DATA_FORK); @@ -1253,7 +1255,7 @@ xfs_dir2_sf_to_block( sfep = xfs_dir2_sf_nextentry(mp, sfp, sfep); } /* Done with the temporary buffer */ - kmem_free(sfp); + kfree(sfp); /* * Sort the leaf entries by hash value. */ @@ -1268,6 +1270,6 @@ xfs_dir2_sf_to_block( xfs_dir3_data_check(dp, bp); return 0; out_free: - kmem_free(sfp); + kfree(sfp); return error; } diff --git a/fs/xfs/libxfs/xfs_dir2_data.c b/fs/xfs/libxfs/xfs_dir2_data.c index dbcf58979a59..7a6d965bea71 100644 --- a/fs/xfs/libxfs/xfs_dir2_data.c +++ b/fs/xfs/libxfs/xfs_dir2_data.c @@ -18,6 +18,7 @@ #include "xfs_trans.h" #include "xfs_buf_item.h" #include "xfs_log.h" +#include "xfs_health.h" static xfs_failaddr_t xfs_dir2_data_freefind_verify( struct xfs_dir2_data_hdr *hdr, struct xfs_dir2_data_free *bf, @@ -433,6 +434,7 @@ xfs_dir3_data_read( __xfs_buf_mark_corrupt(*bpp, fa); xfs_trans_brelse(tp, *bpp); *bpp = NULL; + xfs_dirattr_mark_sick(dp, XFS_DATA_FORK); return -EFSCORRUPTED; } @@ -1198,6 +1200,7 @@ xfs_dir2_data_use_free( corrupt: xfs_corruption_error(__func__, XFS_ERRLEVEL_LOW, args->dp->i_mount, hdr, sizeof(*hdr), __FILE__, __LINE__, fa); + xfs_da_mark_sick(args); return -EFSCORRUPTED; } diff --git a/fs/xfs/libxfs/xfs_dir2_leaf.c b/fs/xfs/libxfs/xfs_dir2_leaf.c index cb9e950a911d..08dda5ce9d91 100644 --- a/fs/xfs/libxfs/xfs_dir2_leaf.c +++ b/fs/xfs/libxfs/xfs_dir2_leaf.c @@ -19,6 +19,7 @@ #include "xfs_trace.h" #include "xfs_trans.h" #include "xfs_buf_item.h" +#include "xfs_health.h" /* * Local function declarations. @@ -1393,8 +1394,10 @@ xfs_dir2_leaf_removename( bestsp = xfs_dir2_leaf_bests_p(ltp); if (be16_to_cpu(bestsp[db]) != oldbest) { xfs_buf_mark_corrupt(lbp); + xfs_da_mark_sick(args); return -EFSCORRUPTED; } + /* * Mark the former data entry unused. */ diff --git a/fs/xfs/libxfs/xfs_dir2_node.c b/fs/xfs/libxfs/xfs_dir2_node.c index 7a03aeb9f4c9..be0b8834028c 100644 --- a/fs/xfs/libxfs/xfs_dir2_node.c +++ b/fs/xfs/libxfs/xfs_dir2_node.c @@ -20,6 +20,7 @@ #include "xfs_trans.h" #include "xfs_buf_item.h" #include "xfs_log.h" +#include "xfs_health.h" /* * Function declarations. @@ -231,6 +232,7 @@ __xfs_dir3_free_read( __xfs_buf_mark_corrupt(*bpp, fa); xfs_trans_brelse(tp, *bpp); *bpp = NULL; + xfs_dirattr_mark_sick(dp, XFS_DATA_FORK); return -EFSCORRUPTED; } @@ -443,6 +445,7 @@ xfs_dir2_leaf_to_node( if (be32_to_cpu(ltp->bestcount) > (uint)dp->i_disk_size / args->geo->blksize) { xfs_buf_mark_corrupt(lbp); + xfs_da_mark_sick(args); return -EFSCORRUPTED; } @@ -517,6 +520,7 @@ xfs_dir2_leafn_add( */ if (index < 0) { xfs_buf_mark_corrupt(bp); + xfs_da_mark_sick(args); return -EFSCORRUPTED; } @@ -736,6 +740,7 @@ xfs_dir2_leafn_lookup_for_addname( cpu_to_be16(NULLDATAOFF))) { if (curfdb != newfdb) xfs_trans_brelse(tp, curbp); + xfs_da_mark_sick(args); return -EFSCORRUPTED; } curfdb = newfdb; @@ -804,6 +809,7 @@ xfs_dir2_leafn_lookup_for_entry( xfs_dir3_leaf_check(dp, bp); if (leafhdr.count <= 0) { xfs_buf_mark_corrupt(bp); + xfs_da_mark_sick(args); return -EFSCORRUPTED; } @@ -1739,6 +1745,7 @@ xfs_dir2_node_add_datablk( } else { xfs_alert(mp, " ... fblk is NULL"); } + xfs_da_mark_sick(args); return -EFSCORRUPTED; } diff --git a/fs/xfs/libxfs/xfs_dir2_sf.c b/fs/xfs/libxfs/xfs_dir2_sf.c index e1f83fc7b6ad..17a20384c8b7 100644 --- a/fs/xfs/libxfs/xfs_dir2_sf.c +++ b/fs/xfs/libxfs/xfs_dir2_sf.c @@ -276,7 +276,7 @@ xfs_dir2_block_to_sf( * format the data into. Once we have formatted the data, we can free * the block and copy the formatted data into the inode literal area. */ - sfp = kmem_alloc(mp->m_sb.sb_inodesize, 0); + sfp = kmalloc(mp->m_sb.sb_inodesize, GFP_KERNEL | __GFP_NOFAIL); memcpy(sfp, sfhp, xfs_dir2_sf_hdr_size(sfhp->i8count)); /* @@ -350,7 +350,7 @@ xfs_dir2_block_to_sf( xfs_dir2_sf_check(args); out: xfs_trans_log_inode(args->trans, dp, logflags); - kmem_free(sfp); + kfree(sfp); return error; } @@ -524,7 +524,7 @@ xfs_dir2_sf_addname_hard( * Copy the old directory to the stack buffer. */ old_isize = (int)dp->i_disk_size; - buf = kmem_alloc(old_isize, 0); + buf = kmalloc(old_isize, GFP_KERNEL | __GFP_NOFAIL); oldsfp = (xfs_dir2_sf_hdr_t *)buf; memcpy(oldsfp, dp->i_df.if_data, old_isize); /* @@ -576,7 +576,7 @@ xfs_dir2_sf_addname_hard( sfep = xfs_dir2_sf_nextentry(mp, sfp, sfep); memcpy(sfep, oldsfep, old_isize - nbytes); } - kmem_free(buf); + kfree(buf); dp->i_disk_size = new_isize; xfs_dir2_sf_check(args); } @@ -1151,7 +1151,7 @@ xfs_dir2_sf_toino4( * Don't want xfs_idata_realloc copying the data here. */ oldsize = dp->i_df.if_bytes; - buf = kmem_alloc(oldsize, 0); + buf = kmalloc(oldsize, GFP_KERNEL | __GFP_NOFAIL); ASSERT(oldsfp->i8count == 1); memcpy(buf, oldsfp, oldsize); /* @@ -1190,7 +1190,7 @@ xfs_dir2_sf_toino4( /* * Clean up the inode. */ - kmem_free(buf); + kfree(buf); dp->i_disk_size = newsize; xfs_trans_log_inode(args->trans, dp, XFS_ILOG_CORE | XFS_ILOG_DDATA); } @@ -1223,7 +1223,7 @@ xfs_dir2_sf_toino8( * Don't want xfs_idata_realloc copying the data here. */ oldsize = dp->i_df.if_bytes; - buf = kmem_alloc(oldsize, 0); + buf = kmalloc(oldsize, GFP_KERNEL | __GFP_NOFAIL); ASSERT(oldsfp->i8count == 0); memcpy(buf, oldsfp, oldsize); /* @@ -1262,7 +1262,7 @@ xfs_dir2_sf_toino8( /* * Clean up the inode. */ - kmem_free(buf); + kfree(buf); dp->i_disk_size = newsize; xfs_trans_log_inode(args->trans, dp, XFS_ILOG_CORE | XFS_ILOG_DDATA); } diff --git a/fs/xfs/libxfs/xfs_format.h b/fs/xfs/libxfs/xfs_format.h index 382ab1e71c0b..2b2f9050fbfb 100644 --- a/fs/xfs/libxfs/xfs_format.h +++ b/fs/xfs/libxfs/xfs_format.h @@ -477,15 +477,9 @@ xfs_is_quota_inode(struct xfs_sb *sbp, xfs_ino_t ino) #define XFS_AGI_GOOD_VERSION(v) ((v) == XFS_AGI_VERSION) /* - * Btree number 0 is bno, 1 is cnt, 2 is rmap. This value gives the size of the - * arrays below. - */ -#define XFS_BTNUM_AGF ((int)XFS_BTNUM_RMAPi + 1) - -/* - * The second word of agf_levels in the first a.g. overlaps the EFS - * superblock's magic number. Since the magic numbers valid for EFS - * are > 64k, our value cannot be confused for an EFS superblock's. + * agf_cnt_level in the first AGF overlaps the EFS superblock's magic number. + * Since the magic numbers valid for EFS are > 64k, our value cannot be confused + * for an EFS superblock. */ typedef struct xfs_agf { @@ -499,8 +493,13 @@ typedef struct xfs_agf { /* * Freespace and rmap information */ - __be32 agf_roots[XFS_BTNUM_AGF]; /* root blocks */ - __be32 agf_levels[XFS_BTNUM_AGF]; /* btree levels */ + __be32 agf_bno_root; /* bnobt root block */ + __be32 agf_cnt_root; /* cntbt root block */ + __be32 agf_rmap_root; /* rmapbt root block */ + + __be32 agf_bno_level; /* bnobt btree levels */ + __be32 agf_cnt_level; /* cntbt btree levels */ + __be32 agf_rmap_level; /* rmapbt btree levels */ __be32 agf_flfirst; /* first freelist block's index */ __be32 agf_fllast; /* last freelist block's index */ diff --git a/fs/xfs/libxfs/xfs_fs.h b/fs/xfs/libxfs/xfs_fs.h index 6360073865db..ca1b17d01437 100644 --- a/fs/xfs/libxfs/xfs_fs.h +++ b/fs/xfs/libxfs/xfs_fs.h @@ -195,6 +195,8 @@ struct xfs_fsop_geom { #define XFS_FSOP_GEOM_SICK_PQUOTA (1 << 3) /* project quota */ #define XFS_FSOP_GEOM_SICK_RT_BITMAP (1 << 4) /* realtime bitmap */ #define XFS_FSOP_GEOM_SICK_RT_SUMMARY (1 << 5) /* realtime summary */ +#define XFS_FSOP_GEOM_SICK_QUOTACHECK (1 << 6) /* quota counts */ +#define XFS_FSOP_GEOM_SICK_NLINKS (1 << 7) /* inode link counts */ /* Output for XFS_FS_COUNTS */ typedef struct xfs_fsop_counts { @@ -292,6 +294,7 @@ struct xfs_ag_geometry { #define XFS_AG_GEOM_SICK_FINOBT (1 << 7) /* free inode index */ #define XFS_AG_GEOM_SICK_RMAPBT (1 << 8) /* reverse mappings */ #define XFS_AG_GEOM_SICK_REFCNTBT (1 << 9) /* reference counts */ +#define XFS_AG_GEOM_SICK_INODES (1 << 10) /* bad inodes were seen */ /* * Structures for XFS_IOC_FSGROWFSDATA, XFS_IOC_FSGROWFSLOG & XFS_IOC_FSGROWFSRT @@ -709,9 +712,12 @@ struct xfs_scrub_metadata { #define XFS_SCRUB_TYPE_GQUOTA 22 /* group quotas */ #define XFS_SCRUB_TYPE_PQUOTA 23 /* project quotas */ #define XFS_SCRUB_TYPE_FSCOUNTERS 24 /* fs summary counters */ +#define XFS_SCRUB_TYPE_QUOTACHECK 25 /* quota counters */ +#define XFS_SCRUB_TYPE_NLINKS 26 /* inode link counts */ +#define XFS_SCRUB_TYPE_HEALTHY 27 /* everything checked out ok */ /* Number of scrub subcommands. */ -#define XFS_SCRUB_TYPE_NR 25 +#define XFS_SCRUB_TYPE_NR 28 /* i: Repair this metadata. */ #define XFS_SCRUB_IFLAG_REPAIR (1u << 0) diff --git a/fs/xfs/libxfs/xfs_health.h b/fs/xfs/libxfs/xfs_health.h index 6296993ff8f3..3c64b5f9bd68 100644 --- a/fs/xfs/libxfs/xfs_health.h +++ b/fs/xfs/libxfs/xfs_health.h @@ -26,21 +26,40 @@ * and the "sick" field tells us if that piece was found to need repairs. * Therefore we can conclude that for a given sick flag value: * - * - checked && sick => metadata needs repair - * - checked && !sick => metadata is ok - * - !checked => has not been examined since mount + * - checked && sick => metadata needs repair + * - checked && !sick => metadata is ok + * - !checked && sick => errors have been observed during normal operation, + * but the metadata has not been checked thoroughly + * - !checked && !sick => has not been examined since mount + * + * Evidence of health problems can be sorted into three basic categories: + * + * a) Primary evidence, which signals that something is defective within the + * general grouping of metadata. + * + * b) Secondary evidence, which are side effects of primary problem but are + * not themselves problems. These can be forgotten when the primary + * health problems are addressed. + * + * c) Indirect evidence, which points to something being wrong in another + * group, but we had to release resources and this is all that's left of + * that state. */ struct xfs_mount; struct xfs_perag; struct xfs_inode; struct xfs_fsop_geom; +struct xfs_btree_cur; +struct xfs_da_args; /* Observable health issues for metadata spanning the entire filesystem. */ #define XFS_SICK_FS_COUNTERS (1 << 0) /* summary counters */ #define XFS_SICK_FS_UQUOTA (1 << 1) /* user quota */ #define XFS_SICK_FS_GQUOTA (1 << 2) /* group quota */ #define XFS_SICK_FS_PQUOTA (1 << 3) /* project quota */ +#define XFS_SICK_FS_QUOTACHECK (1 << 4) /* quota counts */ +#define XFS_SICK_FS_NLINKS (1 << 5) /* inode link counts */ /* Observable health issues for realtime volume metadata. */ #define XFS_SICK_RT_BITMAP (1 << 0) /* realtime bitmap */ @@ -57,6 +76,7 @@ struct xfs_fsop_geom; #define XFS_SICK_AG_FINOBT (1 << 7) /* free inode index */ #define XFS_SICK_AG_RMAPBT (1 << 8) /* reverse mappings */ #define XFS_SICK_AG_REFCNTBT (1 << 9) /* reference counts */ +#define XFS_SICK_AG_INODES (1 << 10) /* inactivated bad inodes */ /* Observable health issues for inode metadata. */ #define XFS_SICK_INO_CORE (1 << 0) /* inode core */ @@ -73,11 +93,16 @@ struct xfs_fsop_geom; #define XFS_SICK_INO_DIR_ZAPPED (1 << 10) /* directory erased */ #define XFS_SICK_INO_SYMLINK_ZAPPED (1 << 11) /* symlink erased */ +/* Don't propagate sick status to ag health summary during inactivation */ +#define XFS_SICK_INO_FORGET (1 << 12) + /* Primary evidence of health problems in a given group. */ #define XFS_SICK_FS_PRIMARY (XFS_SICK_FS_COUNTERS | \ XFS_SICK_FS_UQUOTA | \ XFS_SICK_FS_GQUOTA | \ - XFS_SICK_FS_PQUOTA) + XFS_SICK_FS_PQUOTA | \ + XFS_SICK_FS_QUOTACHECK | \ + XFS_SICK_FS_NLINKS) #define XFS_SICK_RT_PRIMARY (XFS_SICK_RT_BITMAP | \ XFS_SICK_RT_SUMMARY) @@ -107,29 +132,86 @@ struct xfs_fsop_geom; XFS_SICK_INO_DIR_ZAPPED | \ XFS_SICK_INO_SYMLINK_ZAPPED) -/* These functions must be provided by the xfs implementation. */ +/* Secondary state related to (but not primary evidence of) health problems. */ +#define XFS_SICK_FS_SECONDARY (0) +#define XFS_SICK_RT_SECONDARY (0) +#define XFS_SICK_AG_SECONDARY (0) +#define XFS_SICK_INO_SECONDARY (XFS_SICK_INO_FORGET) + +/* Evidence of health problems elsewhere. */ +#define XFS_SICK_FS_INDIRECT (0) +#define XFS_SICK_RT_INDIRECT (0) +#define XFS_SICK_AG_INDIRECT (XFS_SICK_AG_INODES) +#define XFS_SICK_INO_INDIRECT (0) + +/* All health masks. */ +#define XFS_SICK_FS_ALL (XFS_SICK_FS_PRIMARY | \ + XFS_SICK_FS_SECONDARY | \ + XFS_SICK_FS_INDIRECT) + +#define XFS_SICK_RT_ALL (XFS_SICK_RT_PRIMARY | \ + XFS_SICK_RT_SECONDARY | \ + XFS_SICK_RT_INDIRECT) + +#define XFS_SICK_AG_ALL (XFS_SICK_AG_PRIMARY | \ + XFS_SICK_AG_SECONDARY | \ + XFS_SICK_AG_INDIRECT) + +#define XFS_SICK_INO_ALL (XFS_SICK_INO_PRIMARY | \ + XFS_SICK_INO_SECONDARY | \ + XFS_SICK_INO_INDIRECT | \ + XFS_SICK_INO_ZAPPED) + +/* + * These functions must be provided by the xfs implementation. Function + * behavior with respect to the first argument should be as follows: + * + * xfs_*_mark_sick: Set the sick flags and do not set checked flags. + * Runtime code should call this upon encountering + * a corruption. + * + * xfs_*_mark_corrupt: Set the sick and checked flags simultaneously. + * Fsck tools should call this when corruption is + * found. + * + * xfs_*_mark_healthy: Clear the sick flags and set the checked flags. + * Fsck tools should call this after correcting errors. + * + * xfs_*_measure_sickness: Return the sick and check status in the provided + * out parameters. + */ void xfs_fs_mark_sick(struct xfs_mount *mp, unsigned int mask); +void xfs_fs_mark_corrupt(struct xfs_mount *mp, unsigned int mask); void xfs_fs_mark_healthy(struct xfs_mount *mp, unsigned int mask); void xfs_fs_measure_sickness(struct xfs_mount *mp, unsigned int *sick, unsigned int *checked); void xfs_rt_mark_sick(struct xfs_mount *mp, unsigned int mask); +void xfs_rt_mark_corrupt(struct xfs_mount *mp, unsigned int mask); void xfs_rt_mark_healthy(struct xfs_mount *mp, unsigned int mask); void xfs_rt_measure_sickness(struct xfs_mount *mp, unsigned int *sick, unsigned int *checked); +void xfs_agno_mark_sick(struct xfs_mount *mp, xfs_agnumber_t agno, + unsigned int mask); void xfs_ag_mark_sick(struct xfs_perag *pag, unsigned int mask); +void xfs_ag_mark_corrupt(struct xfs_perag *pag, unsigned int mask); void xfs_ag_mark_healthy(struct xfs_perag *pag, unsigned int mask); void xfs_ag_measure_sickness(struct xfs_perag *pag, unsigned int *sick, unsigned int *checked); void xfs_inode_mark_sick(struct xfs_inode *ip, unsigned int mask); +void xfs_inode_mark_corrupt(struct xfs_inode *ip, unsigned int mask); void xfs_inode_mark_healthy(struct xfs_inode *ip, unsigned int mask); void xfs_inode_measure_sickness(struct xfs_inode *ip, unsigned int *sick, unsigned int *checked); void xfs_health_unmount(struct xfs_mount *mp); +void xfs_bmap_mark_sick(struct xfs_inode *ip, int whichfork); +void xfs_btree_mark_sick(struct xfs_btree_cur *cur); +void xfs_dirattr_mark_sick(struct xfs_inode *ip, int whichfork); +void xfs_da_mark_sick(struct xfs_da_args *args); /* Now some helpers. */ @@ -197,4 +279,7 @@ void xfs_fsop_geom_health(struct xfs_mount *mp, struct xfs_fsop_geom *geo); void xfs_ag_geom_health(struct xfs_perag *pag, struct xfs_ag_geometry *ageo); void xfs_bulkstat_health(struct xfs_inode *ip, struct xfs_bulkstat *bs); +#define xfs_metadata_is_sick(error) \ + (unlikely((error) == -EFSCORRUPTED || (error) == -EFSBADCRC)) + #endif /* __XFS_HEALTH_H__ */ diff --git a/fs/xfs/libxfs/xfs_ialloc.c b/fs/xfs/libxfs/xfs_ialloc.c index 2361a22035b0..e5ac3e5430c4 100644 --- a/fs/xfs/libxfs/xfs_ialloc.c +++ b/fs/xfs/libxfs/xfs_ialloc.c @@ -27,6 +27,7 @@ #include "xfs_log.h" #include "xfs_rmap.h" #include "xfs_ag.h" +#include "xfs_health.h" /* * Lookup a record by ino in the btree given by cur. @@ -140,13 +141,13 @@ xfs_inobt_complain_bad_rec( struct xfs_mount *mp = cur->bc_mp; xfs_warn(mp, - "%s Inode BTree record corruption in AG %d detected at %pS!", - cur->bc_btnum == XFS_BTNUM_INO ? "Used" : "Free", - cur->bc_ag.pag->pag_agno, fa); + "%sbt record corruption in AG %d detected at %pS!", + cur->bc_ops->name, cur->bc_ag.pag->pag_agno, fa); xfs_warn(mp, "start inode 0x%x, count 0x%x, free 0x%x freemask 0x%llx, holemask 0x%x", irec->ir_startino, irec->ir_count, irec->ir_freecount, irec->ir_free, irec->ir_holemask); + xfs_btree_mark_sick(cur); return -EFSCORRUPTED; } @@ -205,14 +206,17 @@ xfs_inobt_insert( struct xfs_buf *agbp, xfs_agino_t newino, xfs_agino_t newlen, - xfs_btnum_t btnum) + bool is_finobt) { struct xfs_btree_cur *cur; xfs_agino_t thisino; int i; int error; - cur = xfs_inobt_init_cursor(pag, tp, agbp, btnum); + if (is_finobt) + cur = xfs_finobt_init_cursor(pag, tp, agbp); + else + cur = xfs_inobt_init_cursor(pag, tp, agbp); for (thisino = newino; thisino < newino + newlen; @@ -528,16 +532,14 @@ __xfs_inobt_rec_merge( } /* - * Insert a new sparse inode chunk into the associated inode btree. The inode - * record for the sparse chunk is pre-aligned to a startino that should match - * any pre-existing sparse inode record in the tree. This allows sparse chunks - * to fill over time. + * Insert a new sparse inode chunk into the associated inode allocation btree. + * The inode record for the sparse chunk is pre-aligned to a startino that + * should match any pre-existing sparse inode record in the tree. This allows + * sparse chunks to fill over time. * - * This function supports two modes of handling preexisting records depending on - * the merge flag. If merge is true, the provided record is merged with the + * If no preexisting record exists, the provided record is inserted. + * If there is a preexisting record, the provided record is merged with the * existing record and updated in place. The merged record is returned in nrec. - * If merge is false, an existing record is replaced with the provided record. - * If no preexisting record exists, the provided record is always inserted. * * It is considered corruption if a merge is requested and not possible. Given * the sparse inode alignment constraints, this should never happen. @@ -547,9 +549,7 @@ xfs_inobt_insert_sprec( struct xfs_perag *pag, struct xfs_trans *tp, struct xfs_buf *agbp, - int btnum, - struct xfs_inobt_rec_incore *nrec, /* in/out: new/merged rec. */ - bool merge) /* merge or replace */ + struct xfs_inobt_rec_incore *nrec) /* in/out: new/merged rec. */ { struct xfs_mount *mp = pag->pag_mount; struct xfs_btree_cur *cur; @@ -557,7 +557,7 @@ xfs_inobt_insert_sprec( int i; struct xfs_inobt_rec_incore rec; - cur = xfs_inobt_init_cursor(pag, tp, agbp, btnum); + cur = xfs_inobt_init_cursor(pag, tp, agbp); /* the new record is pre-aligned so we know where to look */ error = xfs_inobt_lookup(cur, nrec->ir_startino, XFS_LOOKUP_EQ, &i); @@ -571,6 +571,7 @@ xfs_inobt_insert_sprec( if (error) goto error; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto error; } @@ -579,45 +580,45 @@ xfs_inobt_insert_sprec( } /* - * A record exists at this startino. Merge or replace the record - * depending on what we've been asked to do. + * A record exists at this startino. Merge the records. */ - if (merge) { - error = xfs_inobt_get_rec(cur, &rec, &i); - if (error) - goto error; - if (XFS_IS_CORRUPT(mp, i != 1)) { - error = -EFSCORRUPTED; - goto error; - } - if (XFS_IS_CORRUPT(mp, rec.ir_startino != nrec->ir_startino)) { - error = -EFSCORRUPTED; - goto error; - } + error = xfs_inobt_get_rec(cur, &rec, &i); + if (error) + goto error; + if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cur); + error = -EFSCORRUPTED; + goto error; + } + if (XFS_IS_CORRUPT(mp, rec.ir_startino != nrec->ir_startino)) { + xfs_btree_mark_sick(cur); + error = -EFSCORRUPTED; + goto error; + } - /* - * This should never fail. If we have coexisting records that - * cannot merge, something is seriously wrong. - */ - if (XFS_IS_CORRUPT(mp, !__xfs_inobt_can_merge(nrec, &rec))) { - error = -EFSCORRUPTED; - goto error; - } + /* + * This should never fail. If we have coexisting records that + * cannot merge, something is seriously wrong. + */ + if (XFS_IS_CORRUPT(mp, !__xfs_inobt_can_merge(nrec, &rec))) { + xfs_btree_mark_sick(cur); + error = -EFSCORRUPTED; + goto error; + } - trace_xfs_irec_merge_pre(mp, pag->pag_agno, rec.ir_startino, - rec.ir_holemask, nrec->ir_startino, - nrec->ir_holemask); + trace_xfs_irec_merge_pre(mp, pag->pag_agno, rec.ir_startino, + rec.ir_holemask, nrec->ir_startino, + nrec->ir_holemask); - /* merge to nrec to output the updated record */ - __xfs_inobt_rec_merge(nrec, &rec); + /* merge to nrec to output the updated record */ + __xfs_inobt_rec_merge(nrec, &rec); - trace_xfs_irec_merge_post(mp, pag->pag_agno, nrec->ir_startino, - nrec->ir_holemask); + trace_xfs_irec_merge_post(mp, pag->pag_agno, nrec->ir_startino, + nrec->ir_holemask); - error = xfs_inobt_rec_check_count(mp, nrec); - if (error) - goto error; - } + error = xfs_inobt_rec_check_count(mp, nrec); + if (error) + goto error; error = xfs_inobt_update(cur, nrec); if (error) @@ -632,6 +633,59 @@ error: } /* + * Insert a new sparse inode chunk into the free inode btree. The inode + * record for the sparse chunk is pre-aligned to a startino that should match + * any pre-existing sparse inode record in the tree. This allows sparse chunks + * to fill over time. + * + * The new record is always inserted, overwriting a pre-existing record if + * there is one. + */ +STATIC int +xfs_finobt_insert_sprec( + struct xfs_perag *pag, + struct xfs_trans *tp, + struct xfs_buf *agbp, + struct xfs_inobt_rec_incore *nrec) /* in/out: new rec. */ +{ + struct xfs_mount *mp = pag->pag_mount; + struct xfs_btree_cur *cur; + int error; + int i; + + cur = xfs_finobt_init_cursor(pag, tp, agbp); + + /* the new record is pre-aligned so we know where to look */ + error = xfs_inobt_lookup(cur, nrec->ir_startino, XFS_LOOKUP_EQ, &i); + if (error) + goto error; + /* if nothing there, insert a new record and return */ + if (i == 0) { + error = xfs_inobt_insert_rec(cur, nrec->ir_holemask, + nrec->ir_count, nrec->ir_freecount, + nrec->ir_free, &i); + if (error) + goto error; + if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cur); + error = -EFSCORRUPTED; + goto error; + } + } else { + error = xfs_inobt_update(cur, nrec); + if (error) + goto error; + } + + xfs_btree_del_cursor(cur, XFS_BTREE_NOERROR); + return 0; +error: + xfs_btree_del_cursor(cur, XFS_BTREE_ERROR); + return error; +} + + +/* * Allocate new inodes in the allocation group specified by agbp. Returns 0 if * inodes were allocated in this AG; -EAGAIN if there was no space in this AG so * the caller knows it can try another AG, a hard -ENOSPC when over the maximum @@ -857,8 +911,7 @@ sparse_alloc: * if necessary. If a merge does occur, rec is updated to the * merged record. */ - error = xfs_inobt_insert_sprec(pag, tp, agbp, - XFS_BTNUM_INO, &rec, true); + error = xfs_inobt_insert_sprec(pag, tp, agbp, &rec); if (error == -EFSCORRUPTED) { xfs_alert(args.mp, "invalid sparse inode record: ino 0x%llx holemask 0x%x count %u", @@ -882,21 +935,19 @@ sparse_alloc: * existing record with this one. */ if (xfs_has_finobt(args.mp)) { - error = xfs_inobt_insert_sprec(pag, tp, agbp, - XFS_BTNUM_FINO, &rec, false); + error = xfs_finobt_insert_sprec(pag, tp, agbp, &rec); if (error) return error; } } else { /* full chunk - insert new records to both btrees */ - error = xfs_inobt_insert(pag, tp, agbp, newino, newlen, - XFS_BTNUM_INO); + error = xfs_inobt_insert(pag, tp, agbp, newino, newlen, false); if (error) return error; if (xfs_has_finobt(args.mp)) { error = xfs_inobt_insert(pag, tp, agbp, newino, - newlen, XFS_BTNUM_FINO); + newlen, true); if (error) return error; } @@ -949,8 +1000,10 @@ xfs_ialloc_next_rec( error = xfs_inobt_get_rec(cur, rec, &i); if (error) return error; - if (XFS_IS_CORRUPT(cur->bc_mp, i != 1)) + if (XFS_IS_CORRUPT(cur->bc_mp, i != 1)) { + xfs_btree_mark_sick(cur); return -EFSCORRUPTED; + } } return 0; @@ -974,8 +1027,10 @@ xfs_ialloc_get_rec( error = xfs_inobt_get_rec(cur, rec, &i); if (error) return error; - if (XFS_IS_CORRUPT(cur->bc_mp, i != 1)) + if (XFS_IS_CORRUPT(cur->bc_mp, i != 1)) { + xfs_btree_mark_sick(cur); return -EFSCORRUPTED; + } } return 0; @@ -1030,7 +1085,7 @@ xfs_dialloc_ag_inobt( ASSERT(pag->pagi_freecount > 0); restart_pagno: - cur = xfs_inobt_init_cursor(pag, tp, agbp, XFS_BTNUM_INO); + cur = xfs_inobt_init_cursor(pag, tp, agbp); /* * If pagino is 0 (this is the root inode allocation) use newino. * This must work because we've just allocated some. @@ -1053,6 +1108,7 @@ xfs_dialloc_ag_inobt( if (error) goto error0; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto error0; } @@ -1061,6 +1117,7 @@ xfs_dialloc_ag_inobt( if (error) goto error0; if (XFS_IS_CORRUPT(mp, j != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto error0; } @@ -1219,6 +1276,7 @@ xfs_dialloc_ag_inobt( if (error) goto error0; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto error0; } @@ -1228,6 +1286,7 @@ xfs_dialloc_ag_inobt( if (error) goto error0; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto error0; } @@ -1237,6 +1296,7 @@ xfs_dialloc_ag_inobt( if (error) goto error0; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto error0; } @@ -1297,8 +1357,10 @@ xfs_dialloc_ag_finobt_near( error = xfs_inobt_get_rec(lcur, rec, &i); if (error) return error; - if (XFS_IS_CORRUPT(lcur->bc_mp, i != 1)) + if (XFS_IS_CORRUPT(lcur->bc_mp, i != 1)) { + xfs_btree_mark_sick(lcur); return -EFSCORRUPTED; + } /* * See if we've landed in the parent inode record. The finobt @@ -1322,12 +1384,14 @@ xfs_dialloc_ag_finobt_near( if (error) goto error_rcur; if (XFS_IS_CORRUPT(lcur->bc_mp, j != 1)) { + xfs_btree_mark_sick(lcur); error = -EFSCORRUPTED; goto error_rcur; } } if (XFS_IS_CORRUPT(lcur->bc_mp, i != 1 && j != 1)) { + xfs_btree_mark_sick(lcur); error = -EFSCORRUPTED; goto error_rcur; } @@ -1383,8 +1447,10 @@ xfs_dialloc_ag_finobt_newino( error = xfs_inobt_get_rec(cur, rec, &i); if (error) return error; - if (XFS_IS_CORRUPT(cur->bc_mp, i != 1)) + if (XFS_IS_CORRUPT(cur->bc_mp, i != 1)) { + xfs_btree_mark_sick(cur); return -EFSCORRUPTED; + } return 0; } } @@ -1395,14 +1461,18 @@ xfs_dialloc_ag_finobt_newino( error = xfs_inobt_lookup(cur, 0, XFS_LOOKUP_GE, &i); if (error) return error; - if (XFS_IS_CORRUPT(cur->bc_mp, i != 1)) + if (XFS_IS_CORRUPT(cur->bc_mp, i != 1)) { + xfs_btree_mark_sick(cur); return -EFSCORRUPTED; + } error = xfs_inobt_get_rec(cur, rec, &i); if (error) return error; - if (XFS_IS_CORRUPT(cur->bc_mp, i != 1)) + if (XFS_IS_CORRUPT(cur->bc_mp, i != 1)) { + xfs_btree_mark_sick(cur); return -EFSCORRUPTED; + } return 0; } @@ -1424,14 +1494,18 @@ xfs_dialloc_ag_update_inobt( error = xfs_inobt_lookup(cur, frec->ir_startino, XFS_LOOKUP_EQ, &i); if (error) return error; - if (XFS_IS_CORRUPT(cur->bc_mp, i != 1)) + if (XFS_IS_CORRUPT(cur->bc_mp, i != 1)) { + xfs_btree_mark_sick(cur); return -EFSCORRUPTED; + } error = xfs_inobt_get_rec(cur, &rec, &i); if (error) return error; - if (XFS_IS_CORRUPT(cur->bc_mp, i != 1)) + if (XFS_IS_CORRUPT(cur->bc_mp, i != 1)) { + xfs_btree_mark_sick(cur); return -EFSCORRUPTED; + } ASSERT((XFS_AGINO_TO_OFFSET(cur->bc_mp, rec.ir_startino) % XFS_INODES_PER_CHUNK) == 0); @@ -1440,8 +1514,10 @@ xfs_dialloc_ag_update_inobt( if (XFS_IS_CORRUPT(cur->bc_mp, rec.ir_free != frec->ir_free || - rec.ir_freecount != frec->ir_freecount)) + rec.ir_freecount != frec->ir_freecount)) { + xfs_btree_mark_sick(cur); return -EFSCORRUPTED; + } return xfs_inobt_update(cur, &rec); } @@ -1483,7 +1559,7 @@ xfs_dialloc_ag( if (!pagino) pagino = be32_to_cpu(agi->agi_newino); - cur = xfs_inobt_init_cursor(pag, tp, agbp, XFS_BTNUM_FINO); + cur = xfs_finobt_init_cursor(pag, tp, agbp); error = xfs_check_agi_freecount(cur); if (error) @@ -1526,7 +1602,7 @@ xfs_dialloc_ag( * the original freecount. If all is well, make the equivalent update to * the inobt using the finobt record and offset information. */ - icur = xfs_inobt_init_cursor(pag, tp, agbp, XFS_BTNUM_INO); + icur = xfs_inobt_init_cursor(pag, tp, agbp); error = xfs_check_agi_freecount(icur); if (error) @@ -1943,7 +2019,7 @@ xfs_difree_inobt( /* * Initialize the cursor. */ - cur = xfs_inobt_init_cursor(pag, tp, agbp, XFS_BTNUM_INO); + cur = xfs_inobt_init_cursor(pag, tp, agbp); error = xfs_check_agi_freecount(cur); if (error) @@ -1958,6 +2034,7 @@ xfs_difree_inobt( goto error0; } if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto error0; } @@ -1968,6 +2045,7 @@ xfs_difree_inobt( goto error0; } if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto error0; } @@ -2068,7 +2146,7 @@ xfs_difree_finobt( int error; int i; - cur = xfs_inobt_init_cursor(pag, tp, agbp, XFS_BTNUM_FINO); + cur = xfs_finobt_init_cursor(pag, tp, agbp); error = xfs_inobt_lookup(cur, ibtrec->ir_startino, XFS_LOOKUP_EQ, &i); if (error) @@ -2080,6 +2158,7 @@ xfs_difree_finobt( * something is out of sync. */ if (XFS_IS_CORRUPT(mp, ibtrec->ir_freecount != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto error; } @@ -2106,6 +2185,7 @@ xfs_difree_finobt( if (error) goto error; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto error; } @@ -2116,6 +2196,7 @@ xfs_difree_finobt( if (XFS_IS_CORRUPT(mp, rec.ir_free != ibtrec->ir_free || rec.ir_freecount != ibtrec->ir_freecount)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto error; } @@ -2265,7 +2346,7 @@ xfs_imap_lookup( * we have a record, we need to ensure it contains the inode number * we are looking up. */ - cur = xfs_inobt_init_cursor(pag, tp, agbp, XFS_BTNUM_INO); + cur = xfs_inobt_init_cursor(pag, tp, agbp); error = xfs_inobt_lookup(cur, agino, XFS_LOOKUP_LE, &i); if (!error) { if (i) @@ -2604,6 +2685,8 @@ xfs_read_agi( error = xfs_trans_read_buf(mp, tp, mp->m_ddev_targp, XFS_AG_DADDR(mp, pag->pag_agno, XFS_AGI_DADDR(mp)), XFS_FSS_TO_BB(mp, 1), 0, agibpp, &xfs_agi_buf_ops); + if (xfs_metadata_is_sick(error)) + xfs_ag_mark_sick(pag, XFS_SICK_AG_AGI); if (error) return error; if (tp) @@ -2765,7 +2848,7 @@ xfs_ialloc_count_inodes( struct xfs_ialloc_count_inodes ci = {0}; int error; - ASSERT(cur->bc_btnum == XFS_BTNUM_INO); + ASSERT(xfs_btree_is_ino(cur->bc_ops)); error = xfs_btree_query_all(cur, xfs_ialloc_count_inodes_rec, &ci); if (error) return error; @@ -2982,7 +3065,7 @@ xfs_ialloc_check_shrink( if (!xfs_has_sparseinodes(pag->pag_mount)) return 0; - cur = xfs_inobt_init_cursor(pag, tp, agibp, XFS_BTNUM_INO); + cur = xfs_inobt_init_cursor(pag, tp, agibp); /* Look up the inobt record that would correspond to the new EOFS. */ agino = XFS_AGB_TO_AGINO(pag->pag_mount, new_length); @@ -2995,6 +3078,7 @@ xfs_ialloc_check_shrink( goto out; if (!has) { + xfs_ag_mark_sick(pag, XFS_SICK_AG_INOBT); error = -EFSCORRUPTED; goto out; } diff --git a/fs/xfs/libxfs/xfs_ialloc_btree.c b/fs/xfs/libxfs/xfs_ialloc_btree.c index 42a5e1f227a0..cc661fca6ff5 100644 --- a/fs/xfs/libxfs/xfs_ialloc_btree.c +++ b/fs/xfs/libxfs/xfs_ialloc_btree.c @@ -17,6 +17,7 @@ #include "xfs_ialloc_btree.h" #include "xfs_alloc.h" #include "xfs_error.h" +#include "xfs_health.h" #include "xfs_trace.h" #include "xfs_trans.h" #include "xfs_rmap.h" @@ -37,7 +38,15 @@ xfs_inobt_dup_cursor( struct xfs_btree_cur *cur) { return xfs_inobt_init_cursor(cur->bc_ag.pag, cur->bc_tp, - cur->bc_ag.agbp, cur->bc_btnum); + cur->bc_ag.agbp); +} + +STATIC struct xfs_btree_cur * +xfs_finobt_dup_cursor( + struct xfs_btree_cur *cur) +{ + return xfs_finobt_init_cursor(cur->bc_ag.pag, cur->bc_tp, + cur->bc_ag.agbp); } STATIC void @@ -81,9 +90,9 @@ xfs_inobt_mod_blockcount( if (!xfs_has_inobtcounts(cur->bc_mp)) return; - if (cur->bc_btnum == XFS_BTNUM_FINO) + if (xfs_btree_is_fino(cur->bc_ops)) be32_add_cpu(&agi->agi_fblocks, howmuch); - else if (cur->bc_btnum == XFS_BTNUM_INO) + else be32_add_cpu(&agi->agi_iblocks, howmuch); xfs_ialloc_log_agi(cur->bc_tp, agbp, XFS_AGI_IBLOCKS); } @@ -300,7 +309,7 @@ xfs_inobt_verify( * xfs_perag_initialised_agi(pag)) if we ever do. */ if (xfs_has_crc(mp)) { - fa = xfs_btree_sblock_v5hdr_verify(bp); + fa = xfs_btree_agblock_v5hdr_verify(bp); if (fa) return fa; } @@ -310,7 +319,7 @@ xfs_inobt_verify( if (level >= M_IGEO(mp)->inobt_maxlevels) return __this_address; - return xfs_btree_sblock_verify(bp, + return xfs_btree_agblock_verify(bp, M_IGEO(mp)->inobt_mxr[level != 0]); } @@ -320,7 +329,7 @@ xfs_inobt_read_verify( { xfs_failaddr_t fa; - if (!xfs_btree_sblock_verify_crc(bp)) + if (!xfs_btree_agblock_verify_crc(bp)) xfs_verifier_error(bp, -EFSBADCRC, __this_address); else { fa = xfs_inobt_verify(bp); @@ -344,7 +353,7 @@ xfs_inobt_write_verify( xfs_verifier_error(bp, -EFSCORRUPTED, fa); return; } - xfs_btree_sblock_calc_crc(bp); + xfs_btree_agblock_calc_crc(bp); } @@ -398,9 +407,17 @@ xfs_inobt_keys_contiguous( be32_to_cpu(key2->inobt.ir_startino)); } -static const struct xfs_btree_ops xfs_inobt_ops = { +const struct xfs_btree_ops xfs_inobt_ops = { + .name = "ino", + .type = XFS_BTREE_TYPE_AG, + .rec_len = sizeof(xfs_inobt_rec_t), .key_len = sizeof(xfs_inobt_key_t), + .ptr_len = XFS_BTREE_SHORT_PTR_LEN, + + .lru_refs = XFS_INO_BTREE_REF, + .statoff = XFS_STATS_CALC_INDEX(xs_ibt_2), + .sick_mask = XFS_SICK_AG_INOBT, .dup_cursor = xfs_inobt_dup_cursor, .set_root = xfs_inobt_set_root, @@ -420,11 +437,19 @@ static const struct xfs_btree_ops xfs_inobt_ops = { .keys_contiguous = xfs_inobt_keys_contiguous, }; -static const struct xfs_btree_ops xfs_finobt_ops = { +const struct xfs_btree_ops xfs_finobt_ops = { + .name = "fino", + .type = XFS_BTREE_TYPE_AG, + .rec_len = sizeof(xfs_inobt_rec_t), .key_len = sizeof(xfs_inobt_key_t), + .ptr_len = XFS_BTREE_SHORT_PTR_LEN, - .dup_cursor = xfs_inobt_dup_cursor, + .lru_refs = XFS_INO_BTREE_REF, + .statoff = XFS_STATS_CALC_INDEX(xs_fibt_2), + .sick_mask = XFS_SICK_AG_FINOBT, + + .dup_cursor = xfs_finobt_dup_cursor, .set_root = xfs_finobt_set_root, .alloc_block = xfs_finobt_alloc_block, .free_block = xfs_finobt_free_block, @@ -443,65 +468,54 @@ static const struct xfs_btree_ops xfs_finobt_ops = { }; /* - * Initialize a new inode btree cursor. + * Create an inode btree cursor. + * + * For staging cursors tp and agbp are NULL. */ -static struct xfs_btree_cur * -xfs_inobt_init_common( +struct xfs_btree_cur * +xfs_inobt_init_cursor( struct xfs_perag *pag, - struct xfs_trans *tp, /* transaction pointer */ - xfs_btnum_t btnum) /* ialloc or free ino btree */ + struct xfs_trans *tp, + struct xfs_buf *agbp) { struct xfs_mount *mp = pag->pag_mount; struct xfs_btree_cur *cur; - cur = xfs_btree_alloc_cursor(mp, tp, btnum, + cur = xfs_btree_alloc_cursor(mp, tp, &xfs_inobt_ops, M_IGEO(mp)->inobt_maxlevels, xfs_inobt_cur_cache); - if (btnum == XFS_BTNUM_INO) { - cur->bc_statoff = XFS_STATS_CALC_INDEX(xs_ibt_2); - cur->bc_ops = &xfs_inobt_ops; - } else { - cur->bc_statoff = XFS_STATS_CALC_INDEX(xs_fibt_2); - cur->bc_ops = &xfs_finobt_ops; - } - - if (xfs_has_crc(mp)) - cur->bc_flags |= XFS_BTREE_CRC_BLOCKS; - cur->bc_ag.pag = xfs_perag_hold(pag); + cur->bc_ag.agbp = agbp; + if (agbp) { + struct xfs_agi *agi = agbp->b_addr; + + cur->bc_nlevels = be32_to_cpu(agi->agi_level); + } return cur; } -/* Create an inode btree cursor. */ +/* + * Create a free inode btree cursor. + * + * For staging cursors tp and agbp are NULL. + */ struct xfs_btree_cur * -xfs_inobt_init_cursor( +xfs_finobt_init_cursor( struct xfs_perag *pag, struct xfs_trans *tp, - struct xfs_buf *agbp, - xfs_btnum_t btnum) + struct xfs_buf *agbp) { + struct xfs_mount *mp = pag->pag_mount; struct xfs_btree_cur *cur; - struct xfs_agi *agi = agbp->b_addr; - cur = xfs_inobt_init_common(pag, tp, btnum); - if (btnum == XFS_BTNUM_INO) - cur->bc_nlevels = be32_to_cpu(agi->agi_level); - else - cur->bc_nlevels = be32_to_cpu(agi->agi_free_level); + cur = xfs_btree_alloc_cursor(mp, tp, &xfs_finobt_ops, + M_IGEO(mp)->inobt_maxlevels, xfs_inobt_cur_cache); + cur->bc_ag.pag = xfs_perag_hold(pag); cur->bc_ag.agbp = agbp; - return cur; -} + if (agbp) { + struct xfs_agi *agi = agbp->b_addr; -/* Create an inode btree cursor with a fake root for staging. */ -struct xfs_btree_cur * -xfs_inobt_stage_cursor( - struct xfs_perag *pag, - struct xbtree_afakeroot *afake, - xfs_btnum_t btnum) -{ - struct xfs_btree_cur *cur; - - cur = xfs_inobt_init_common(pag, NULL, btnum); - xfs_btree_stage_afakeroot(cur, afake); + cur->bc_nlevels = be32_to_cpu(agi->agi_free_level); + } return cur; } @@ -521,7 +535,7 @@ xfs_inobt_commit_staged_btree( ASSERT(cur->bc_flags & XFS_BTREE_STAGING); - if (cur->bc_btnum == XFS_BTNUM_INO) { + if (xfs_btree_is_ino(cur->bc_ops)) { fields = XFS_AGI_ROOT | XFS_AGI_LEVEL; agi->agi_root = cpu_to_be32(afake->af_root); agi->agi_level = cpu_to_be32(afake->af_levels); @@ -530,7 +544,7 @@ xfs_inobt_commit_staged_btree( fields |= XFS_AGI_IBLOCKS; } xfs_ialloc_log_agi(tp, agbp, fields); - xfs_btree_commit_afakeroot(cur, tp, agbp, &xfs_inobt_ops); + xfs_btree_commit_afakeroot(cur, tp, agbp); } else { fields = XFS_AGI_FREE_ROOT | XFS_AGI_FREE_LEVEL; agi->agi_free_root = cpu_to_be32(afake->af_root); @@ -540,7 +554,7 @@ xfs_inobt_commit_staged_btree( fields |= XFS_AGI_IBLOCKS; } xfs_ialloc_log_agi(tp, agbp, fields); - xfs_btree_commit_afakeroot(cur, tp, agbp, &xfs_finobt_ops); + xfs_btree_commit_afakeroot(cur, tp, agbp); } } @@ -721,45 +735,21 @@ xfs_inobt_max_size( XFS_INODES_PER_CHUNK); } -/* Read AGI and create inobt cursor. */ -int -xfs_inobt_cur( - struct xfs_perag *pag, - struct xfs_trans *tp, - xfs_btnum_t which, - struct xfs_btree_cur **curpp, - struct xfs_buf **agi_bpp) -{ - struct xfs_btree_cur *cur; - int error; - - ASSERT(*agi_bpp == NULL); - ASSERT(*curpp == NULL); - - error = xfs_ialloc_read_agi(pag, tp, agi_bpp); - if (error) - return error; - - cur = xfs_inobt_init_cursor(pag, tp, *agi_bpp, which); - *curpp = cur; - return 0; -} - static int -xfs_inobt_count_blocks( +xfs_finobt_count_blocks( struct xfs_perag *pag, struct xfs_trans *tp, - xfs_btnum_t btnum, xfs_extlen_t *tree_blocks) { struct xfs_buf *agbp = NULL; - struct xfs_btree_cur *cur = NULL; + struct xfs_btree_cur *cur; int error; - error = xfs_inobt_cur(pag, tp, btnum, &cur, &agbp); + error = xfs_ialloc_read_agi(pag, tp, &agbp); if (error) return error; + cur = xfs_inobt_init_cursor(pag, tp, agbp); error = xfs_btree_count_blocks(cur, tree_blocks); xfs_btree_del_cursor(cur, error); xfs_trans_brelse(tp, agbp); @@ -807,8 +797,7 @@ xfs_finobt_calc_reserves( if (xfs_has_inobtcounts(pag->pag_mount)) error = xfs_finobt_read_blocks(pag, tp, &tree_len); else - error = xfs_inobt_count_blocks(pag, tp, XFS_BTNUM_FINO, - &tree_len); + error = xfs_finobt_count_blocks(pag, tp, &tree_len); if (error) return error; diff --git a/fs/xfs/libxfs/xfs_ialloc_btree.h b/fs/xfs/libxfs/xfs_ialloc_btree.h index 3262c3fe5ebe..6472ec1ecbb4 100644 --- a/fs/xfs/libxfs/xfs_ialloc_btree.h +++ b/fs/xfs/libxfs/xfs_ialloc_btree.h @@ -46,10 +46,10 @@ struct xfs_perag; (maxrecs) * sizeof(xfs_inobt_key_t) + \ ((index) - 1) * sizeof(xfs_inobt_ptr_t))) -extern struct xfs_btree_cur *xfs_inobt_init_cursor(struct xfs_perag *pag, - struct xfs_trans *tp, struct xfs_buf *agbp, xfs_btnum_t btnum); -struct xfs_btree_cur *xfs_inobt_stage_cursor(struct xfs_perag *pag, - struct xbtree_afakeroot *afake, xfs_btnum_t btnum); +struct xfs_btree_cur *xfs_inobt_init_cursor(struct xfs_perag *pag, + struct xfs_trans *tp, struct xfs_buf *agbp); +struct xfs_btree_cur *xfs_finobt_init_cursor(struct xfs_perag *pag, + struct xfs_trans *tp, struct xfs_buf *agbp); extern int xfs_inobt_maxrecs(struct xfs_mount *, int, int); /* ir_holemask to inode allocation bitmap conversion */ @@ -66,9 +66,6 @@ int xfs_finobt_calc_reserves(struct xfs_perag *perag, struct xfs_trans *tp, xfs_extlen_t *ask, xfs_extlen_t *used); extern xfs_extlen_t xfs_iallocbt_calc_size(struct xfs_mount *mp, unsigned long long len); -int xfs_inobt_cur(struct xfs_perag *pag, struct xfs_trans *tp, - xfs_btnum_t btnum, struct xfs_btree_cur **curpp, - struct xfs_buf **agi_bpp); void xfs_inobt_commit_staged_btree(struct xfs_btree_cur *cur, struct xfs_trans *tp, struct xfs_buf *agbp); diff --git a/fs/xfs/libxfs/xfs_iext_tree.c b/fs/xfs/libxfs/xfs_iext_tree.c index f4e6b200cdf8..8796f2b3e534 100644 --- a/fs/xfs/libxfs/xfs_iext_tree.c +++ b/fs/xfs/libxfs/xfs_iext_tree.c @@ -394,11 +394,18 @@ xfs_iext_leaf_key( return leaf->recs[n].lo & XFS_IEXT_STARTOFF_MASK; } +static inline void * +xfs_iext_alloc_node( + int size) +{ + return kzalloc(size, GFP_KERNEL | __GFP_NOLOCKDEP | __GFP_NOFAIL); +} + static void xfs_iext_grow( struct xfs_ifork *ifp) { - struct xfs_iext_node *node = kmem_zalloc(NODE_SIZE, KM_NOFS); + struct xfs_iext_node *node = xfs_iext_alloc_node(NODE_SIZE); int i; if (ifp->if_height == 1) { @@ -454,7 +461,7 @@ xfs_iext_split_node( int *nr_entries) { struct xfs_iext_node *node = *nodep; - struct xfs_iext_node *new = kmem_zalloc(NODE_SIZE, KM_NOFS); + struct xfs_iext_node *new = xfs_iext_alloc_node(NODE_SIZE); const int nr_move = KEYS_PER_NODE / 2; int nr_keep = nr_move + (KEYS_PER_NODE & 1); int i = 0; @@ -542,7 +549,7 @@ xfs_iext_split_leaf( int *nr_entries) { struct xfs_iext_leaf *leaf = cur->leaf; - struct xfs_iext_leaf *new = kmem_zalloc(NODE_SIZE, KM_NOFS); + struct xfs_iext_leaf *new = xfs_iext_alloc_node(NODE_SIZE); const int nr_move = RECS_PER_LEAF / 2; int nr_keep = nr_move + (RECS_PER_LEAF & 1); int i; @@ -583,7 +590,7 @@ xfs_iext_alloc_root( { ASSERT(ifp->if_bytes == 0); - ifp->if_data = kmem_zalloc(sizeof(struct xfs_iext_rec), KM_NOFS); + ifp->if_data = xfs_iext_alloc_node(sizeof(struct xfs_iext_rec)); ifp->if_height = 1; /* now that we have a node step into it */ @@ -603,7 +610,8 @@ xfs_iext_realloc_root( if (new_size / sizeof(struct xfs_iext_rec) == RECS_PER_LEAF) new_size = NODE_SIZE; - new = krealloc(ifp->if_data, new_size, GFP_NOFS | __GFP_NOFAIL); + new = krealloc(ifp->if_data, new_size, + GFP_KERNEL | __GFP_NOLOCKDEP | __GFP_NOFAIL); memset(new + ifp->if_bytes, 0, new_size - ifp->if_bytes); ifp->if_data = new; cur->leaf = new; @@ -743,7 +751,7 @@ xfs_iext_remove_node( again: ASSERT(node->ptrs[pos]); ASSERT(node->ptrs[pos] == victim); - kmem_free(victim); + kfree(victim); nr_entries = xfs_iext_node_nr_entries(node, pos) - 1; offset = node->keys[0]; @@ -789,7 +797,7 @@ again: ASSERT(node == ifp->if_data); ifp->if_data = node->ptrs[0]; ifp->if_height--; - kmem_free(node); + kfree(node); } } @@ -863,7 +871,7 @@ xfs_iext_free_last_leaf( struct xfs_ifork *ifp) { ifp->if_height--; - kmem_free(ifp->if_data); + kfree(ifp->if_data); ifp->if_data = NULL; } @@ -1044,7 +1052,7 @@ xfs_iext_destroy_node( } } - kmem_free(node); + kfree(node); } void diff --git a/fs/xfs/libxfs/xfs_inode_buf.c b/fs/xfs/libxfs/xfs_inode_buf.c index 137a65bda95d..d0dcce462bf4 100644 --- a/fs/xfs/libxfs/xfs_inode_buf.c +++ b/fs/xfs/libxfs/xfs_inode_buf.c @@ -18,6 +18,7 @@ #include "xfs_trans.h" #include "xfs_ialloc.h" #include "xfs_dir2.h" +#include "xfs_health.h" #include <linux/iversion.h> @@ -132,9 +133,14 @@ xfs_imap_to_bp( struct xfs_imap *imap, struct xfs_buf **bpp) { - return xfs_trans_read_buf(mp, tp, mp->m_ddev_targp, imap->im_blkno, - imap->im_len, XBF_UNMAPPED, bpp, - &xfs_inode_buf_ops); + int error; + + error = xfs_trans_read_buf(mp, tp, mp->m_ddev_targp, imap->im_blkno, + imap->im_len, XBF_UNMAPPED, bpp, &xfs_inode_buf_ops); + if (xfs_metadata_is_sick(error)) + xfs_agno_mark_sick(mp, xfs_daddr_to_agno(mp, imap->im_blkno), + XFS_SICK_AG_INODES); + return error; } static inline struct timespec64 xfs_inode_decode_bigtime(uint64_t ts) diff --git a/fs/xfs/libxfs/xfs_inode_fork.c b/fs/xfs/libxfs/xfs_inode_fork.c index f4569e18a8d0..7d660a973909 100644 --- a/fs/xfs/libxfs/xfs_inode_fork.c +++ b/fs/xfs/libxfs/xfs_inode_fork.c @@ -25,6 +25,8 @@ #include "xfs_attr_leaf.h" #include "xfs_types.h" #include "xfs_errortag.h" +#include "xfs_health.h" +#include "xfs_symlink_remote.h" struct kmem_cache *xfs_ifork_cache; @@ -50,7 +52,8 @@ xfs_init_local_fork( mem_size++; if (size) { - char *new_data = kmem_alloc(mem_size, KM_NOFS); + char *new_data = kmalloc(mem_size, + GFP_KERNEL | __GFP_NOLOCKDEP | __GFP_NOFAIL); memcpy(new_data, data, size); if (zero_terminate) @@ -77,7 +80,7 @@ xfs_iformat_local( /* * If the size is unreasonable, then something * is wrong and we just bail out rather than crash in - * kmem_alloc() or memcpy() below. + * kmalloc() or memcpy() below. */ if (unlikely(size > XFS_DFORK_SIZE(dip, ip->i_mount, whichfork))) { xfs_warn(ip->i_mount, @@ -87,6 +90,7 @@ xfs_iformat_local( xfs_inode_verifier_error(ip, -EFSCORRUPTED, "xfs_iformat_local", dip, sizeof(*dip), __this_address); + xfs_inode_mark_sick(ip, XFS_SICK_INO_CORE); return -EFSCORRUPTED; } @@ -116,7 +120,7 @@ xfs_iformat_extents( /* * If the number of extents is unreasonable, then something is wrong and - * we just bail out rather than crash in kmem_alloc() or memcpy() below. + * we just bail out rather than crash in kmalloc() or memcpy() below. */ if (unlikely(size < 0 || size > XFS_DFORK_SIZE(dip, mp, whichfork))) { xfs_warn(ip->i_mount, "corrupt inode %llu ((a)extents = %llu).", @@ -124,6 +128,7 @@ xfs_iformat_extents( xfs_inode_verifier_error(ip, -EFSCORRUPTED, "xfs_iformat_extents(1)", dip, sizeof(*dip), __this_address); + xfs_inode_mark_sick(ip, XFS_SICK_INO_CORE); return -EFSCORRUPTED; } @@ -143,6 +148,7 @@ xfs_iformat_extents( xfs_inode_verifier_error(ip, -EFSCORRUPTED, "xfs_iformat_extents(2)", dp, sizeof(*dp), fa); + xfs_inode_mark_sick(ip, XFS_SICK_INO_CORE); return xfs_bmap_complain_bad_rec(ip, whichfork, fa, &new); } @@ -201,11 +207,13 @@ xfs_iformat_btree( xfs_inode_verifier_error(ip, -EFSCORRUPTED, "xfs_iformat_btree", dfp, size, __this_address); + xfs_inode_mark_sick(ip, XFS_SICK_INO_CORE); return -EFSCORRUPTED; } ifp->if_broot_bytes = size; - ifp->if_broot = kmem_alloc(size, KM_NOFS); + ifp->if_broot = kmalloc(size, + GFP_KERNEL | __GFP_NOLOCKDEP | __GFP_NOFAIL); ASSERT(ifp->if_broot != NULL); /* * Copy and convert from the on-disk structure @@ -265,12 +273,14 @@ xfs_iformat_data_fork( default: xfs_inode_verifier_error(ip, -EFSCORRUPTED, __func__, dip, sizeof(*dip), __this_address); + xfs_inode_mark_sick(ip, XFS_SICK_INO_CORE); return -EFSCORRUPTED; } break; default: xfs_inode_verifier_error(ip, -EFSCORRUPTED, __func__, dip, sizeof(*dip), __this_address); + xfs_inode_mark_sick(ip, XFS_SICK_INO_CORE); return -EFSCORRUPTED; } } @@ -342,6 +352,7 @@ xfs_iformat_attr_fork( default: xfs_inode_verifier_error(ip, error, __func__, dip, sizeof(*dip), __this_address); + xfs_inode_mark_sick(ip, XFS_SICK_INO_CORE); error = -EFSCORRUPTED; break; } @@ -399,7 +410,8 @@ xfs_iroot_realloc( */ if (ifp->if_broot_bytes == 0) { new_size = XFS_BMAP_BROOT_SPACE_CALC(mp, rec_diff); - ifp->if_broot = kmem_alloc(new_size, KM_NOFS); + ifp->if_broot = kmalloc(new_size, + GFP_KERNEL | __GFP_NOFAIL); ifp->if_broot_bytes = (int)new_size; return; } @@ -414,7 +426,7 @@ xfs_iroot_realloc( new_max = cur_max + rec_diff; new_size = XFS_BMAP_BROOT_SPACE_CALC(mp, new_max); ifp->if_broot = krealloc(ifp->if_broot, new_size, - GFP_NOFS | __GFP_NOFAIL); + GFP_KERNEL | __GFP_NOFAIL); op = (char *)XFS_BMAP_BROOT_PTR_ADDR(mp, ifp->if_broot, 1, ifp->if_broot_bytes); np = (char *)XFS_BMAP_BROOT_PTR_ADDR(mp, ifp->if_broot, 1, @@ -440,7 +452,7 @@ xfs_iroot_realloc( else new_size = 0; if (new_size > 0) { - new_broot = kmem_alloc(new_size, KM_NOFS); + new_broot = kmalloc(new_size, GFP_KERNEL | __GFP_NOFAIL); /* * First copy over the btree block header. */ @@ -470,7 +482,7 @@ xfs_iroot_realloc( (int)new_size); memcpy(np, op, new_max * (uint)sizeof(xfs_fsblock_t)); } - kmem_free(ifp->if_broot); + kfree(ifp->if_broot); ifp->if_broot = new_broot; ifp->if_broot_bytes = (int)new_size; if (ifp->if_broot) @@ -488,7 +500,7 @@ xfs_iroot_realloc( * * If the amount of space needed has decreased below the size of the * inline buffer, then switch to using the inline buffer. Otherwise, - * use kmem_realloc() or kmem_alloc() to adjust the size of the buffer + * use krealloc() or kmalloc() to adjust the size of the buffer * to what is needed. * * ip -- the inode whose if_data area is changing @@ -509,7 +521,7 @@ xfs_idata_realloc( if (byte_diff) { ifp->if_data = krealloc(ifp->if_data, new_size, - GFP_NOFS | __GFP_NOFAIL); + GFP_KERNEL | __GFP_NOFAIL); if (new_size == 0) ifp->if_data = NULL; ifp->if_bytes = new_size; @@ -524,13 +536,13 @@ xfs_idestroy_fork( struct xfs_ifork *ifp) { if (ifp->if_broot != NULL) { - kmem_free(ifp->if_broot); + kfree(ifp->if_broot); ifp->if_broot = NULL; } switch (ifp->if_format) { case XFS_DINODE_FMT_LOCAL: - kmem_free(ifp->if_data); + kfree(ifp->if_data); ifp->if_data = NULL; break; case XFS_DINODE_FMT_EXTENTS: @@ -562,7 +574,7 @@ xfs_iextents_copy( struct xfs_bmbt_irec rec; int64_t copied = 0; - ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL | XFS_ILOCK_SHARED)); + xfs_assert_ilocked(ip, XFS_ILOCK_EXCL | XFS_ILOCK_SHARED); ASSERT(ifp->if_bytes > 0); for_each_xfs_iext(ifp, &icur, &rec) { @@ -689,7 +701,7 @@ xfs_ifork_init_cow( return; ip->i_cowfp = kmem_cache_zalloc(xfs_ifork_cache, - GFP_NOFS | __GFP_NOFAIL); + GFP_KERNEL | __GFP_NOLOCKDEP | __GFP_NOFAIL); ip->i_cowfp->if_format = XFS_DINODE_FMT_EXTENTS; } @@ -802,3 +814,12 @@ xfs_iext_count_upgrade( return 0; } + +/* Decide if a file mapping is on the realtime device or not. */ +bool +xfs_ifork_is_realtime( + struct xfs_inode *ip, + int whichfork) +{ + return XFS_IS_REALTIME_INODE(ip) && whichfork != XFS_ATTR_FORK; +} diff --git a/fs/xfs/libxfs/xfs_inode_fork.h b/fs/xfs/libxfs/xfs_inode_fork.h index 96303249d28a..bd53eb951b65 100644 --- a/fs/xfs/libxfs/xfs_inode_fork.h +++ b/fs/xfs/libxfs/xfs_inode_fork.h @@ -260,6 +260,7 @@ int xfs_iext_count_may_overflow(struct xfs_inode *ip, int whichfork, int nr_to_add); int xfs_iext_count_upgrade(struct xfs_trans *tp, struct xfs_inode *ip, uint nr_to_add); +bool xfs_ifork_is_realtime(struct xfs_inode *ip, int whichfork); /* returns true if the fork has extents but they are not read in yet. */ static inline bool xfs_need_iread_extents(const struct xfs_ifork *ifp) diff --git a/fs/xfs/libxfs/xfs_log_format.h b/fs/xfs/libxfs/xfs_log_format.h index 269573c82808..16872972e1e9 100644 --- a/fs/xfs/libxfs/xfs_log_format.h +++ b/fs/xfs/libxfs/xfs_log_format.h @@ -838,10 +838,12 @@ struct xfs_cud_log_format { #define XFS_BMAP_EXTENT_ATTR_FORK (1U << 31) #define XFS_BMAP_EXTENT_UNWRITTEN (1U << 30) +#define XFS_BMAP_EXTENT_REALTIME (1U << 29) #define XFS_BMAP_EXTENT_FLAGS (XFS_BMAP_EXTENT_TYPE_MASK | \ XFS_BMAP_EXTENT_ATTR_FORK | \ - XFS_BMAP_EXTENT_UNWRITTEN) + XFS_BMAP_EXTENT_UNWRITTEN | \ + XFS_BMAP_EXTENT_REALTIME) /* * This is the structure used to lay out an bui log item in the diff --git a/fs/xfs/libxfs/xfs_refcount.c b/fs/xfs/libxfs/xfs_refcount.c index 6709a7f8bad5..511c912d515c 100644 --- a/fs/xfs/libxfs/xfs_refcount.c +++ b/fs/xfs/libxfs/xfs_refcount.c @@ -23,6 +23,7 @@ #include "xfs_refcount.h" #include "xfs_rmap.h" #include "xfs_ag.h" +#include "xfs_health.h" struct kmem_cache *xfs_refcount_intent_cache; @@ -156,6 +157,7 @@ xfs_refcount_complain_bad_rec( xfs_warn(mp, "Start block 0x%x, block count 0x%x, references 0x%x", irec->rc_startblock, irec->rc_blockcount, irec->rc_refcount); + xfs_btree_mark_sick(cur); return -EFSCORRUPTED; } @@ -238,6 +240,7 @@ xfs_refcount_insert( if (error) goto out_error; if (XFS_IS_CORRUPT(cur->bc_mp, *i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto out_error; } @@ -268,12 +271,14 @@ xfs_refcount_delete( if (error) goto out_error; if (XFS_IS_CORRUPT(cur->bc_mp, found_rec != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto out_error; } trace_xfs_refcount_delete(cur->bc_mp, cur->bc_ag.pag->pag_agno, &irec); error = xfs_btree_delete(cur, i); if (XFS_IS_CORRUPT(cur->bc_mp, *i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto out_error; } @@ -398,6 +403,7 @@ xfs_refcount_split_extent( if (error) goto out_error; if (XFS_IS_CORRUPT(cur->bc_mp, found_rec != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto out_error; } @@ -425,6 +431,7 @@ xfs_refcount_split_extent( if (error) goto out_error; if (XFS_IS_CORRUPT(cur->bc_mp, found_rec != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto out_error; } @@ -470,6 +477,7 @@ xfs_refcount_merge_center_extents( if (error) goto out_error; if (XFS_IS_CORRUPT(cur->bc_mp, found_rec != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto out_error; } @@ -478,6 +486,7 @@ xfs_refcount_merge_center_extents( if (error) goto out_error; if (XFS_IS_CORRUPT(cur->bc_mp, found_rec != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto out_error; } @@ -487,6 +496,7 @@ xfs_refcount_merge_center_extents( if (error) goto out_error; if (XFS_IS_CORRUPT(cur->bc_mp, found_rec != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto out_error; } @@ -498,6 +508,7 @@ xfs_refcount_merge_center_extents( if (error) goto out_error; if (XFS_IS_CORRUPT(cur->bc_mp, found_rec != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto out_error; } @@ -542,6 +553,7 @@ xfs_refcount_merge_left_extent( if (error) goto out_error; if (XFS_IS_CORRUPT(cur->bc_mp, found_rec != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto out_error; } @@ -550,6 +562,7 @@ xfs_refcount_merge_left_extent( if (error) goto out_error; if (XFS_IS_CORRUPT(cur->bc_mp, found_rec != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto out_error; } @@ -561,6 +574,7 @@ xfs_refcount_merge_left_extent( if (error) goto out_error; if (XFS_IS_CORRUPT(cur->bc_mp, found_rec != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto out_error; } @@ -608,6 +622,7 @@ xfs_refcount_merge_right_extent( if (error) goto out_error; if (XFS_IS_CORRUPT(cur->bc_mp, found_rec != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto out_error; } @@ -616,6 +631,7 @@ xfs_refcount_merge_right_extent( if (error) goto out_error; if (XFS_IS_CORRUPT(cur->bc_mp, found_rec != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto out_error; } @@ -627,6 +643,7 @@ xfs_refcount_merge_right_extent( if (error) goto out_error; if (XFS_IS_CORRUPT(cur->bc_mp, found_rec != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto out_error; } @@ -674,6 +691,7 @@ xfs_refcount_find_left_extents( if (error) goto out_error; if (XFS_IS_CORRUPT(cur->bc_mp, found_rec != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto out_error; } @@ -693,6 +711,7 @@ xfs_refcount_find_left_extents( if (error) goto out_error; if (XFS_IS_CORRUPT(cur->bc_mp, found_rec != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto out_error; } @@ -767,6 +786,7 @@ xfs_refcount_find_right_extents( if (error) goto out_error; if (XFS_IS_CORRUPT(cur->bc_mp, found_rec != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto out_error; } @@ -786,6 +806,7 @@ xfs_refcount_find_right_extents( if (error) goto out_error; if (XFS_IS_CORRUPT(cur->bc_mp, found_rec != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto out_error; } @@ -1056,7 +1077,7 @@ xfs_refcount_still_have_space( * to handle each of the shape changes to the refcount btree. */ overhead = xfs_allocfree_block_count(cur->bc_mp, - cur->bc_ag.refc.shape_changes); + cur->bc_refc.shape_changes); overhead += cur->bc_mp->m_refc_maxlevels; overhead *= cur->bc_mp->m_sb.sb_blocksize; @@ -1064,17 +1085,17 @@ xfs_refcount_still_have_space( * Only allow 2 refcount extent updates per transaction if the * refcount continue update "error" has been injected. */ - if (cur->bc_ag.refc.nr_ops > 2 && + if (cur->bc_refc.nr_ops > 2 && XFS_TEST_ERROR(false, cur->bc_mp, XFS_ERRTAG_REFCOUNT_CONTINUE_UPDATE)) return false; - if (cur->bc_ag.refc.nr_ops == 0) + if (cur->bc_refc.nr_ops == 0) return true; else if (overhead > cur->bc_tp->t_log_res) return false; - return cur->bc_tp->t_log_res - overhead > - cur->bc_ag.refc.nr_ops * XFS_REFCOUNT_ITEM_OVERHEAD; + return cur->bc_tp->t_log_res - overhead > + cur->bc_refc.nr_ops * XFS_REFCOUNT_ITEM_OVERHEAD; } /* @@ -1134,7 +1155,7 @@ xfs_refcount_adjust_extents( * Either cover the hole (increment) or * delete the range (decrement). */ - cur->bc_ag.refc.nr_ops++; + cur->bc_refc.nr_ops++; if (tmp.rc_refcount) { error = xfs_refcount_insert(cur, &tmp, &found_tmp); @@ -1142,6 +1163,7 @@ xfs_refcount_adjust_extents( goto out_error; if (XFS_IS_CORRUPT(cur->bc_mp, found_tmp != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto out_error; } @@ -1180,6 +1202,7 @@ xfs_refcount_adjust_extents( */ if (XFS_IS_CORRUPT(cur->bc_mp, ext.rc_blockcount == 0) || XFS_IS_CORRUPT(cur->bc_mp, ext.rc_blockcount > *aglen)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto out_error; } @@ -1193,7 +1216,7 @@ xfs_refcount_adjust_extents( ext.rc_refcount += adj; trace_xfs_refcount_modify_extent(cur->bc_mp, cur->bc_ag.pag->pag_agno, &ext); - cur->bc_ag.refc.nr_ops++; + cur->bc_refc.nr_ops++; if (ext.rc_refcount > 1) { error = xfs_refcount_update(cur, &ext); if (error) @@ -1203,6 +1226,7 @@ xfs_refcount_adjust_extents( if (error) goto out_error; if (XFS_IS_CORRUPT(cur->bc_mp, found_rec != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto out_error; } @@ -1281,7 +1305,7 @@ xfs_refcount_adjust( if (shape_changed) shape_changes++; if (shape_changes) - cur->bc_ag.refc.shape_changes++; + cur->bc_refc.shape_changes++; /* Now that we've taken care of the ends, adjust the middle extents */ error = xfs_refcount_adjust_extents(cur, agbno, aglen, adj); @@ -1327,8 +1351,10 @@ xfs_refcount_continue_op( struct xfs_perag *pag = cur->bc_ag.pag; if (XFS_IS_CORRUPT(mp, !xfs_verify_agbext(pag, new_agbno, - ri->ri_blockcount))) + ri->ri_blockcount))) { + xfs_btree_mark_sick(cur); return -EFSCORRUPTED; + } ri->ri_startblock = XFS_AGB_TO_FSB(mp, pag->pag_agno, new_agbno); @@ -1374,8 +1400,8 @@ xfs_refcount_finish_one( */ rcur = *pcur; if (rcur != NULL && rcur->bc_ag.pag != ri->ri_pag) { - nr_ops = rcur->bc_ag.refc.nr_ops; - shape_changes = rcur->bc_ag.refc.shape_changes; + nr_ops = rcur->bc_refc.nr_ops; + shape_changes = rcur->bc_refc.shape_changes; xfs_refcount_finish_one_cleanup(tp, rcur, 0); rcur = NULL; *pcur = NULL; @@ -1387,8 +1413,8 @@ xfs_refcount_finish_one( return error; rcur = xfs_refcountbt_init_cursor(mp, tp, agbp, ri->ri_pag); - rcur->bc_ag.refc.nr_ops = nr_ops; - rcur->bc_ag.refc.shape_changes = shape_changes; + rcur->bc_refc.nr_ops = nr_ops; + rcur->bc_refc.shape_changes = shape_changes; } *pcur = rcur; @@ -1449,7 +1475,7 @@ __xfs_refcount_add( blockcount); ri = kmem_cache_alloc(xfs_refcount_intent_cache, - GFP_NOFS | __GFP_NOFAIL); + GFP_KERNEL | __GFP_NOFAIL); INIT_LIST_HEAD(&ri->ri_list); ri->ri_type = type; ri->ri_startblock = startblock; @@ -1535,6 +1561,7 @@ xfs_refcount_find_shared( if (error) goto out_error; if (XFS_IS_CORRUPT(cur->bc_mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto out_error; } @@ -1552,6 +1579,7 @@ xfs_refcount_find_shared( if (error) goto out_error; if (XFS_IS_CORRUPT(cur->bc_mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto out_error; } @@ -1585,6 +1613,7 @@ xfs_refcount_find_shared( if (error) goto out_error; if (XFS_IS_CORRUPT(cur->bc_mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto out_error; } @@ -1682,6 +1711,7 @@ xfs_refcount_adjust_cow_extents( goto out_error; if (XFS_IS_CORRUPT(cur->bc_mp, found_rec && ext.rc_domain != XFS_REFC_DOMAIN_COW)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto out_error; } @@ -1697,6 +1727,7 @@ xfs_refcount_adjust_cow_extents( /* Adding a CoW reservation, there should be nothing here. */ if (XFS_IS_CORRUPT(cur->bc_mp, agbno + aglen > ext.rc_startblock)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto out_error; } @@ -1714,6 +1745,7 @@ xfs_refcount_adjust_cow_extents( if (error) goto out_error; if (XFS_IS_CORRUPT(cur->bc_mp, found_tmp != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto out_error; } @@ -1721,14 +1753,17 @@ xfs_refcount_adjust_cow_extents( case XFS_REFCOUNT_ADJUST_COW_FREE: /* Removing a CoW reservation, there should be one extent. */ if (XFS_IS_CORRUPT(cur->bc_mp, ext.rc_startblock != agbno)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto out_error; } if (XFS_IS_CORRUPT(cur->bc_mp, ext.rc_blockcount != aglen)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto out_error; } if (XFS_IS_CORRUPT(cur->bc_mp, ext.rc_refcount != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto out_error; } @@ -1740,6 +1775,7 @@ xfs_refcount_adjust_cow_extents( if (error) goto out_error; if (XFS_IS_CORRUPT(cur->bc_mp, found_rec != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto out_error; } @@ -1889,8 +1925,10 @@ xfs_refcount_recover_extent( struct xfs_refcount_recovery *rr; if (XFS_IS_CORRUPT(cur->bc_mp, - be32_to_cpu(rec->refc.rc_refcount) != 1)) + be32_to_cpu(rec->refc.rc_refcount) != 1)) { + xfs_btree_mark_sick(cur); return -EFSCORRUPTED; + } rr = kmalloc(sizeof(struct xfs_refcount_recovery), GFP_KERNEL | __GFP_NOFAIL); @@ -1900,6 +1938,7 @@ xfs_refcount_recover_extent( if (xfs_refcount_check_irec(cur->bc_ag.pag, &rr->rr_rrec) != NULL || XFS_IS_CORRUPT(cur->bc_mp, rr->rr_rrec.rc_domain != XFS_REFC_DOMAIN_COW)) { + xfs_btree_mark_sick(cur); kfree(rr); return -EFSCORRUPTED; } diff --git a/fs/xfs/libxfs/xfs_refcount_btree.c b/fs/xfs/libxfs/xfs_refcount_btree.c index 0d80bd99147c..ca59f6c89f3e 100644 --- a/fs/xfs/libxfs/xfs_refcount_btree.c +++ b/fs/xfs/libxfs/xfs_refcount_btree.c @@ -16,6 +16,7 @@ #include "xfs_refcount.h" #include "xfs_alloc.h" #include "xfs_error.h" +#include "xfs_health.h" #include "xfs_trace.h" #include "xfs_trans.h" #include "xfs_bit.h" @@ -77,8 +78,6 @@ xfs_refcountbt_alloc_block( xfs_refc_block(args.mp))); if (error) goto out_error; - trace_xfs_refcountbt_alloc_block(cur->bc_mp, cur->bc_ag.pag->pag_agno, - args.agbno, 1); if (args.fsbno == NULLFSBLOCK) { *stat = 0; return 0; @@ -107,8 +106,6 @@ xfs_refcountbt_free_block( struct xfs_agf *agf = agbp->b_addr; xfs_fsblock_t fsbno = XFS_DADDR_TO_FSB(mp, xfs_buf_daddr(bp)); - trace_xfs_refcountbt_free_block(cur->bc_mp, cur->bc_ag.pag->pag_agno, - XFS_FSB_TO_AGBNO(cur->bc_mp, fsbno), 1); be32_add_cpu(&agf->agf_refcount_blocks, -1); xfs_alloc_log_agf(cur->bc_tp, agbp, XFS_AGF_REFCOUNT_BLOCKS); return xfs_free_extent_later(cur->bc_tp, fsbno, 1, @@ -220,7 +217,7 @@ xfs_refcountbt_verify( if (!xfs_has_reflink(mp)) return __this_address; - fa = xfs_btree_sblock_v5hdr_verify(bp); + fa = xfs_btree_agblock_v5hdr_verify(bp); if (fa) return fa; @@ -242,7 +239,7 @@ xfs_refcountbt_verify( } else if (level >= mp->m_refc_maxlevels) return __this_address; - return xfs_btree_sblock_verify(bp, mp->m_refc_mxr[level != 0]); + return xfs_btree_agblock_verify(bp, mp->m_refc_mxr[level != 0]); } STATIC void @@ -251,7 +248,7 @@ xfs_refcountbt_read_verify( { xfs_failaddr_t fa; - if (!xfs_btree_sblock_verify_crc(bp)) + if (!xfs_btree_agblock_verify_crc(bp)) xfs_verifier_error(bp, -EFSBADCRC, __this_address); else { fa = xfs_refcountbt_verify(bp); @@ -275,7 +272,7 @@ xfs_refcountbt_write_verify( xfs_verifier_error(bp, -EFSCORRUPTED, fa); return; } - xfs_btree_sblock_calc_crc(bp); + xfs_btree_agblock_calc_crc(bp); } @@ -321,9 +318,17 @@ xfs_refcountbt_keys_contiguous( be32_to_cpu(key2->refc.rc_startblock)); } -static const struct xfs_btree_ops xfs_refcountbt_ops = { +const struct xfs_btree_ops xfs_refcountbt_ops = { + .name = "refcount", + .type = XFS_BTREE_TYPE_AG, + .rec_len = sizeof(struct xfs_refcount_rec), .key_len = sizeof(struct xfs_refcount_key), + .ptr_len = XFS_BTREE_SHORT_PTR_LEN, + + .lru_refs = XFS_REFC_BTREE_REF, + .statoff = XFS_STATS_CALC_INDEX(xs_refcbt_2), + .sick_mask = XFS_SICK_AG_REFCNTBT, .dup_cursor = xfs_refcountbt_dup_cursor, .set_root = xfs_refcountbt_set_root, @@ -344,59 +349,32 @@ static const struct xfs_btree_ops xfs_refcountbt_ops = { }; /* - * Initialize a new refcount btree cursor. + * Create a new refcount btree cursor. + * + * For staging cursors tp and agbp are NULL. */ -static struct xfs_btree_cur * -xfs_refcountbt_init_common( +struct xfs_btree_cur * +xfs_refcountbt_init_cursor( struct xfs_mount *mp, struct xfs_trans *tp, + struct xfs_buf *agbp, struct xfs_perag *pag) { struct xfs_btree_cur *cur; ASSERT(pag->pag_agno < mp->m_sb.sb_agcount); - cur = xfs_btree_alloc_cursor(mp, tp, XFS_BTNUM_REFC, + cur = xfs_btree_alloc_cursor(mp, tp, &xfs_refcountbt_ops, mp->m_refc_maxlevels, xfs_refcountbt_cur_cache); - cur->bc_statoff = XFS_STATS_CALC_INDEX(xs_refcbt_2); - - cur->bc_flags |= XFS_BTREE_CRC_BLOCKS; - cur->bc_ag.pag = xfs_perag_hold(pag); - cur->bc_ag.refc.nr_ops = 0; - cur->bc_ag.refc.shape_changes = 0; - cur->bc_ops = &xfs_refcountbt_ops; - return cur; -} - -/* Create a btree cursor. */ -struct xfs_btree_cur * -xfs_refcountbt_init_cursor( - struct xfs_mount *mp, - struct xfs_trans *tp, - struct xfs_buf *agbp, - struct xfs_perag *pag) -{ - struct xfs_agf *agf = agbp->b_addr; - struct xfs_btree_cur *cur; - - cur = xfs_refcountbt_init_common(mp, tp, pag); - cur->bc_nlevels = be32_to_cpu(agf->agf_refcount_level); + cur->bc_refc.nr_ops = 0; + cur->bc_refc.shape_changes = 0; cur->bc_ag.agbp = agbp; - return cur; -} + if (agbp) { + struct xfs_agf *agf = agbp->b_addr; -/* Create a btree cursor with a fake root for staging. */ -struct xfs_btree_cur * -xfs_refcountbt_stage_cursor( - struct xfs_mount *mp, - struct xbtree_afakeroot *afake, - struct xfs_perag *pag) -{ - struct xfs_btree_cur *cur; - - cur = xfs_refcountbt_init_common(mp, NULL, pag); - xfs_btree_stage_afakeroot(cur, afake); + cur->bc_nlevels = be32_to_cpu(agf->agf_refcount_level); + } return cur; } @@ -421,7 +399,7 @@ xfs_refcountbt_commit_staged_btree( xfs_alloc_log_agf(tp, agbp, XFS_AGF_REFCOUNT_BLOCKS | XFS_AGF_REFCOUNT_ROOT | XFS_AGF_REFCOUNT_LEVEL); - xfs_btree_commit_afakeroot(cur, tp, agbp, &xfs_refcountbt_ops); + xfs_btree_commit_afakeroot(cur, tp, agbp); } /* Calculate number of records in a refcount btree block. */ diff --git a/fs/xfs/libxfs/xfs_refcount_btree.h b/fs/xfs/libxfs/xfs_refcount_btree.h index d66b37259bed..1e0ab25f6c68 100644 --- a/fs/xfs/libxfs/xfs_refcount_btree.h +++ b/fs/xfs/libxfs/xfs_refcount_btree.h @@ -48,8 +48,6 @@ struct xbtree_afakeroot; extern struct xfs_btree_cur *xfs_refcountbt_init_cursor(struct xfs_mount *mp, struct xfs_trans *tp, struct xfs_buf *agbp, struct xfs_perag *pag); -struct xfs_btree_cur *xfs_refcountbt_stage_cursor(struct xfs_mount *mp, - struct xbtree_afakeroot *afake, struct xfs_perag *pag); extern int xfs_refcountbt_maxrecs(int blocklen, bool leaf); extern void xfs_refcountbt_compute_maxlevels(struct xfs_mount *mp); diff --git a/fs/xfs/libxfs/xfs_rmap.c b/fs/xfs/libxfs/xfs_rmap.c index 76bf7f48cb5a..ef16f6f9cef6 100644 --- a/fs/xfs/libxfs/xfs_rmap.c +++ b/fs/xfs/libxfs/xfs_rmap.c @@ -23,6 +23,7 @@ #include "xfs_error.h" #include "xfs_inode.h" #include "xfs_ag.h" +#include "xfs_health.h" struct kmem_cache *xfs_rmap_intent_cache; @@ -56,8 +57,10 @@ xfs_rmap_lookup_le( error = xfs_rmap_get_rec(cur, irec, &get_stat); if (error) return error; - if (!get_stat) + if (!get_stat) { + xfs_btree_mark_sick(cur); return -EFSCORRUPTED; + } return 0; } @@ -132,6 +135,7 @@ xfs_rmap_insert( if (error) goto done; if (XFS_IS_CORRUPT(rcur->bc_mp, i != 0)) { + xfs_btree_mark_sick(rcur); error = -EFSCORRUPTED; goto done; } @@ -145,6 +149,7 @@ xfs_rmap_insert( if (error) goto done; if (XFS_IS_CORRUPT(rcur->bc_mp, i != 1)) { + xfs_btree_mark_sick(rcur); error = -EFSCORRUPTED; goto done; } @@ -174,6 +179,7 @@ xfs_rmap_delete( if (error) goto done; if (XFS_IS_CORRUPT(rcur->bc_mp, i != 1)) { + xfs_btree_mark_sick(rcur); error = -EFSCORRUPTED; goto done; } @@ -182,6 +188,7 @@ xfs_rmap_delete( if (error) goto done; if (XFS_IS_CORRUPT(rcur->bc_mp, i != 1)) { + xfs_btree_mark_sick(rcur); error = -EFSCORRUPTED; goto done; } @@ -208,10 +215,10 @@ xfs_rmap_btrec_to_irec( /* Simple checks for rmap records. */ xfs_failaddr_t xfs_rmap_check_irec( - struct xfs_btree_cur *cur, + struct xfs_perag *pag, const struct xfs_rmap_irec *irec) { - struct xfs_mount *mp = cur->bc_mp; + struct xfs_mount *mp = pag->pag_mount; bool is_inode; bool is_unwritten; bool is_bmbt; @@ -226,8 +233,8 @@ xfs_rmap_check_irec( return __this_address; } else { /* check for valid extent range, including overflow */ - if (!xfs_verify_agbext(cur->bc_ag.pag, irec->rm_startblock, - irec->rm_blockcount)) + if (!xfs_verify_agbext(pag, irec->rm_startblock, + irec->rm_blockcount)) return __this_address; } @@ -262,6 +269,16 @@ xfs_rmap_check_irec( return NULL; } +static inline xfs_failaddr_t +xfs_rmap_check_btrec( + struct xfs_btree_cur *cur, + const struct xfs_rmap_irec *irec) +{ + if (xfs_btree_is_mem_rmap(cur->bc_ops)) + return xfs_rmap_check_irec(cur->bc_mem.pag, irec); + return xfs_rmap_check_irec(cur->bc_ag.pag, irec); +} + static inline int xfs_rmap_complain_bad_rec( struct xfs_btree_cur *cur, @@ -270,13 +287,18 @@ xfs_rmap_complain_bad_rec( { struct xfs_mount *mp = cur->bc_mp; - xfs_warn(mp, - "Reverse Mapping BTree record corruption in AG %d detected at %pS!", - cur->bc_ag.pag->pag_agno, fa); + if (xfs_btree_is_mem_rmap(cur->bc_ops)) + xfs_warn(mp, + "In-Memory Reverse Mapping BTree record corruption detected at %pS!", fa); + else + xfs_warn(mp, + "Reverse Mapping BTree record corruption in AG %d detected at %pS!", + cur->bc_ag.pag->pag_agno, fa); xfs_warn(mp, "Owner 0x%llx, flags 0x%x, start block 0x%x block count 0x%x", irec->rm_owner, irec->rm_flags, irec->rm_startblock, irec->rm_blockcount); + xfs_btree_mark_sick(cur); return -EFSCORRUPTED; } @@ -299,7 +321,7 @@ xfs_rmap_get_rec( fa = xfs_rmap_btrec_to_irec(rec, irec); if (!fa) - fa = xfs_rmap_check_irec(cur, irec); + fa = xfs_rmap_check_btrec(cur, irec); if (fa) return xfs_rmap_complain_bad_rec(cur, fa, irec); @@ -512,7 +534,7 @@ xfs_rmap_lookup_le_range( */ static int xfs_rmap_free_check_owner( - struct xfs_mount *mp, + struct xfs_btree_cur *cur, uint64_t ltoff, struct xfs_rmap_irec *rec, xfs_filblks_t len, @@ -520,6 +542,7 @@ xfs_rmap_free_check_owner( uint64_t offset, unsigned int flags) { + struct xfs_mount *mp = cur->bc_mp; int error = 0; if (owner == XFS_RMAP_OWN_UNKNOWN) @@ -529,12 +552,14 @@ xfs_rmap_free_check_owner( if (XFS_IS_CORRUPT(mp, (flags & XFS_RMAP_UNWRITTEN) != (rec->rm_flags & XFS_RMAP_UNWRITTEN))) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto out; } /* Make sure the owner matches what we expect to find in the tree. */ if (XFS_IS_CORRUPT(mp, owner != rec->rm_owner)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto out; } @@ -546,16 +571,19 @@ xfs_rmap_free_check_owner( if (flags & XFS_RMAP_BMBT_BLOCK) { if (XFS_IS_CORRUPT(mp, !(rec->rm_flags & XFS_RMAP_BMBT_BLOCK))) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto out; } } else { if (XFS_IS_CORRUPT(mp, rec->rm_offset > offset)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto out; } if (XFS_IS_CORRUPT(mp, offset + len > ltoff + rec->rm_blockcount)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto out; } @@ -618,6 +646,7 @@ xfs_rmap_unmap( if (error) goto out_error; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto out_error; } @@ -639,6 +668,7 @@ xfs_rmap_unmap( if (XFS_IS_CORRUPT(mp, bno < ltrec.rm_startblock + ltrec.rm_blockcount)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto out_error; } @@ -665,6 +695,7 @@ xfs_rmap_unmap( if (error) goto out_error; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto out_error; } @@ -677,12 +708,13 @@ xfs_rmap_unmap( ltrec.rm_startblock > bno || ltrec.rm_startblock + ltrec.rm_blockcount < bno + len)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto out_error; } /* Check owner information. */ - error = xfs_rmap_free_check_owner(mp, ltoff, <rec, len, owner, + error = xfs_rmap_free_check_owner(cur, ltoff, <rec, len, owner, offset, flags); if (error) goto out_error; @@ -697,6 +729,7 @@ xfs_rmap_unmap( if (error) goto out_error; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto out_error; } @@ -788,6 +821,86 @@ out_error: return error; } +#ifdef CONFIG_XFS_LIVE_HOOKS +/* + * Use a static key here to reduce the overhead of rmapbt live updates. If + * the compiler supports jump labels, the static branch will be replaced by a + * nop sled when there are no hook users. Online fsck is currently the only + * caller, so this is a reasonable tradeoff. + * + * Note: Patching the kernel code requires taking the cpu hotplug lock. Other + * parts of the kernel allocate memory with that lock held, which means that + * XFS callers cannot hold any locks that might be used by memory reclaim or + * writeback when calling the static_branch_{inc,dec} functions. + */ +DEFINE_STATIC_XFS_HOOK_SWITCH(xfs_rmap_hooks_switch); + +void +xfs_rmap_hook_disable(void) +{ + xfs_hooks_switch_off(&xfs_rmap_hooks_switch); +} + +void +xfs_rmap_hook_enable(void) +{ + xfs_hooks_switch_on(&xfs_rmap_hooks_switch); +} + +/* Call downstream hooks for a reverse mapping update. */ +static inline void +xfs_rmap_update_hook( + struct xfs_trans *tp, + struct xfs_perag *pag, + enum xfs_rmap_intent_type op, + xfs_agblock_t startblock, + xfs_extlen_t blockcount, + bool unwritten, + const struct xfs_owner_info *oinfo) +{ + if (xfs_hooks_switched_on(&xfs_rmap_hooks_switch)) { + struct xfs_rmap_update_params p = { + .startblock = startblock, + .blockcount = blockcount, + .unwritten = unwritten, + .oinfo = *oinfo, /* struct copy */ + }; + + if (pag) + xfs_hooks_call(&pag->pag_rmap_update_hooks, op, &p); + } +} + +/* Call the specified function during a reverse mapping update. */ +int +xfs_rmap_hook_add( + struct xfs_perag *pag, + struct xfs_rmap_hook *hook) +{ + return xfs_hooks_add(&pag->pag_rmap_update_hooks, &hook->rmap_hook); +} + +/* Stop calling the specified function during a reverse mapping update. */ +void +xfs_rmap_hook_del( + struct xfs_perag *pag, + struct xfs_rmap_hook *hook) +{ + xfs_hooks_del(&pag->pag_rmap_update_hooks, &hook->rmap_hook); +} + +/* Configure rmap update hook functions. */ +void +xfs_rmap_hook_setup( + struct xfs_rmap_hook *hook, + notifier_fn_t mod_fn) +{ + xfs_hook_setup(&hook->rmap_hook, mod_fn); +} +#else +# define xfs_rmap_update_hook(t, p, o, s, b, u, oi) do { } while (0) +#endif /* CONFIG_XFS_LIVE_HOOKS */ + /* * Remove a reference to an extent in the rmap btree. */ @@ -808,7 +921,7 @@ xfs_rmap_free( return 0; cur = xfs_rmapbt_init_cursor(mp, tp, agbp, pag); - + xfs_rmap_update_hook(tp, pag, XFS_RMAP_UNMAP, bno, len, false, oinfo); error = xfs_rmap_unmap(cur, bno, len, false, oinfo); xfs_btree_del_cursor(cur, error); @@ -900,6 +1013,7 @@ xfs_rmap_map( if (XFS_IS_CORRUPT(mp, have_lt != 0 && ltrec.rm_startblock + ltrec.rm_blockcount > bno)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto out_error; } @@ -917,10 +1031,12 @@ xfs_rmap_map( if (error) goto out_error; if (XFS_IS_CORRUPT(mp, have_gt != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto out_error; } if (XFS_IS_CORRUPT(mp, bno + len > gtrec.rm_startblock)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto out_error; } @@ -974,6 +1090,7 @@ xfs_rmap_map( if (error) goto out_error; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto out_error; } @@ -1021,6 +1138,7 @@ xfs_rmap_map( if (error) goto out_error; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto out_error; } @@ -1055,6 +1173,7 @@ xfs_rmap_alloc( return 0; cur = xfs_rmapbt_init_cursor(mp, tp, agbp, pag); + xfs_rmap_update_hook(tp, pag, XFS_RMAP_MAP, bno, len, false, oinfo); error = xfs_rmap_map(cur, bno, len, false, oinfo); xfs_btree_del_cursor(cur, error); @@ -1116,6 +1235,7 @@ xfs_rmap_convert( if (error) goto done; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto done; } @@ -1153,12 +1273,14 @@ xfs_rmap_convert( if (error) goto done; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto done; } if (XFS_IS_CORRUPT(mp, LEFT.rm_startblock + LEFT.rm_blockcount > bno)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto done; } @@ -1181,6 +1303,7 @@ xfs_rmap_convert( if (error) goto done; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto done; } @@ -1193,10 +1316,12 @@ xfs_rmap_convert( if (error) goto done; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto done; } if (XFS_IS_CORRUPT(mp, bno + len > RIGHT.rm_startblock)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto done; } @@ -1227,6 +1352,7 @@ xfs_rmap_convert( if (error) goto done; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto done; } @@ -1246,6 +1372,7 @@ xfs_rmap_convert( if (error) goto done; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto done; } @@ -1257,6 +1384,7 @@ xfs_rmap_convert( if (error) goto done; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto done; } @@ -1264,6 +1392,7 @@ xfs_rmap_convert( if (error) goto done; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto done; } @@ -1275,6 +1404,7 @@ xfs_rmap_convert( if (error) goto done; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto done; } @@ -1282,6 +1412,7 @@ xfs_rmap_convert( if (error) goto done; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto done; } @@ -1305,6 +1436,7 @@ xfs_rmap_convert( if (error) goto done; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto done; } @@ -1312,6 +1444,7 @@ xfs_rmap_convert( if (error) goto done; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto done; } @@ -1331,6 +1464,7 @@ xfs_rmap_convert( if (error) goto done; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto done; } @@ -1342,6 +1476,7 @@ xfs_rmap_convert( if (error) goto done; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto done; } @@ -1349,6 +1484,7 @@ xfs_rmap_convert( if (error) goto done; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto done; } @@ -1419,6 +1555,7 @@ xfs_rmap_convert( if (error) goto done; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto done; } @@ -1461,6 +1598,7 @@ xfs_rmap_convert( if (error) goto done; if (XFS_IS_CORRUPT(mp, i != 0)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto done; } @@ -1476,6 +1614,7 @@ xfs_rmap_convert( if (error) goto done; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto done; } @@ -1509,6 +1648,7 @@ xfs_rmap_convert( if (error) goto done; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto done; } @@ -1522,6 +1662,7 @@ xfs_rmap_convert( if (error) goto done; if (XFS_IS_CORRUPT(mp, i != 0)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto done; } @@ -1534,6 +1675,7 @@ xfs_rmap_convert( if (error) goto done; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto done; } @@ -1606,6 +1748,7 @@ xfs_rmap_convert_shared( if (error) goto done; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto done; } @@ -1634,6 +1777,7 @@ xfs_rmap_convert_shared( if (XFS_IS_CORRUPT(mp, LEFT.rm_startblock + LEFT.rm_blockcount > bno)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto done; } @@ -1652,10 +1796,12 @@ xfs_rmap_convert_shared( if (error) goto done; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto done; } if (XFS_IS_CORRUPT(mp, bno + len > RIGHT.rm_startblock)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto done; } @@ -1706,6 +1852,7 @@ xfs_rmap_convert_shared( if (error) goto done; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto done; } @@ -1732,6 +1879,7 @@ xfs_rmap_convert_shared( if (error) goto done; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto done; } @@ -1758,6 +1906,7 @@ xfs_rmap_convert_shared( if (error) goto done; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto done; } @@ -1781,6 +1930,7 @@ xfs_rmap_convert_shared( if (error) goto done; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto done; } @@ -1816,6 +1966,7 @@ xfs_rmap_convert_shared( if (error) goto done; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto done; } @@ -1861,6 +2012,7 @@ xfs_rmap_convert_shared( if (error) goto done; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto done; } @@ -1896,6 +2048,7 @@ xfs_rmap_convert_shared( if (error) goto done; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto done; } @@ -1934,6 +2087,7 @@ xfs_rmap_convert_shared( if (error) goto done; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto done; } @@ -2023,6 +2177,7 @@ xfs_rmap_unmap_shared( if (error) goto out_error; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto out_error; } @@ -2033,12 +2188,14 @@ xfs_rmap_unmap_shared( ltrec.rm_startblock > bno || ltrec.rm_startblock + ltrec.rm_blockcount < bno + len)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto out_error; } /* Make sure the owner matches what we expect to find in the tree. */ if (XFS_IS_CORRUPT(mp, owner != ltrec.rm_owner)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto out_error; } @@ -2047,16 +2204,19 @@ xfs_rmap_unmap_shared( if (XFS_IS_CORRUPT(mp, (flags & XFS_RMAP_UNWRITTEN) != (ltrec.rm_flags & XFS_RMAP_UNWRITTEN))) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto out_error; } /* Check the offset. */ if (XFS_IS_CORRUPT(mp, ltrec.rm_offset > offset)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto out_error; } if (XFS_IS_CORRUPT(mp, offset > ltoff + ltrec.rm_blockcount)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto out_error; } @@ -2113,6 +2273,7 @@ xfs_rmap_unmap_shared( if (error) goto out_error; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto out_error; } @@ -2142,6 +2303,7 @@ xfs_rmap_unmap_shared( if (error) goto out_error; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto out_error; } @@ -2221,6 +2383,7 @@ xfs_rmap_map_shared( if (error) goto out_error; if (XFS_IS_CORRUPT(mp, have_gt != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto out_error; } @@ -2273,6 +2436,7 @@ xfs_rmap_map_shared( if (error) goto out_error; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto out_error; } @@ -2335,15 +2499,12 @@ xfs_rmap_map_raw( { struct xfs_owner_info oinfo; - oinfo.oi_owner = rmap->rm_owner; - oinfo.oi_offset = rmap->rm_offset; - oinfo.oi_flags = 0; - if (rmap->rm_flags & XFS_RMAP_ATTR_FORK) - oinfo.oi_flags |= XFS_OWNER_INFO_ATTR_FORK; - if (rmap->rm_flags & XFS_RMAP_BMBT_BLOCK) - oinfo.oi_flags |= XFS_OWNER_INFO_BMBT_BLOCK; + xfs_owner_info_pack(&oinfo, rmap->rm_owner, rmap->rm_offset, + rmap->rm_flags); - if (rmap->rm_flags || XFS_RMAP_NON_INODE_OWNER(rmap->rm_owner)) + if ((rmap->rm_flags & (XFS_RMAP_ATTR_FORK | XFS_RMAP_BMBT_BLOCK | + XFS_RMAP_UNWRITTEN)) || + XFS_RMAP_NON_INODE_OWNER(rmap->rm_owner)) return xfs_rmap_map(cur, rmap->rm_startblock, rmap->rm_blockcount, rmap->rm_flags & XFS_RMAP_UNWRITTEN, @@ -2373,7 +2534,7 @@ xfs_rmap_query_range_helper( fa = xfs_rmap_btrec_to_irec(rec, &irec); if (!fa) - fa = xfs_rmap_check_irec(cur, &irec); + fa = xfs_rmap_check_btrec(cur, &irec); if (fa) return xfs_rmap_complain_bad_rec(cur, fa, &irec); @@ -2428,6 +2589,38 @@ xfs_rmap_finish_one_cleanup( xfs_trans_brelse(tp, agbp); } +/* Commit an rmap operation into the ondisk tree. */ +int +__xfs_rmap_finish_intent( + struct xfs_btree_cur *rcur, + enum xfs_rmap_intent_type op, + xfs_agblock_t bno, + xfs_extlen_t len, + const struct xfs_owner_info *oinfo, + bool unwritten) +{ + switch (op) { + case XFS_RMAP_ALLOC: + case XFS_RMAP_MAP: + return xfs_rmap_map(rcur, bno, len, unwritten, oinfo); + case XFS_RMAP_MAP_SHARED: + return xfs_rmap_map_shared(rcur, bno, len, unwritten, oinfo); + case XFS_RMAP_FREE: + case XFS_RMAP_UNMAP: + return xfs_rmap_unmap(rcur, bno, len, unwritten, oinfo); + case XFS_RMAP_UNMAP_SHARED: + return xfs_rmap_unmap_shared(rcur, bno, len, unwritten, oinfo); + case XFS_RMAP_CONVERT: + return xfs_rmap_convert(rcur, bno, len, !unwritten, oinfo); + case XFS_RMAP_CONVERT_SHARED: + return xfs_rmap_convert_shared(rcur, bno, len, !unwritten, + oinfo); + default: + ASSERT(0); + return -EFSCORRUPTED; + } +} + /* * Process one of the deferred rmap operations. We pass back the * btree cursor to maintain our lock on the rmapbt between calls. @@ -2476,10 +2669,14 @@ xfs_rmap_finish_one( * allocate blocks. */ error = xfs_free_extent_fix_freelist(tp, ri->ri_pag, &agbp); - if (error) + if (error) { + xfs_ag_mark_sick(ri->ri_pag, XFS_SICK_AG_AGFL); return error; - if (XFS_IS_CORRUPT(tp->t_mountp, !agbp)) + } + if (XFS_IS_CORRUPT(tp->t_mountp, !agbp)) { + xfs_ag_mark_sick(ri->ri_pag, XFS_SICK_AG_AGFL); return -EFSCORRUPTED; + } rcur = xfs_rmapbt_init_cursor(mp, tp, agbp, ri->ri_pag); } @@ -2490,39 +2687,14 @@ xfs_rmap_finish_one( unwritten = ri->ri_bmap.br_state == XFS_EXT_UNWRITTEN; bno = XFS_FSB_TO_AGBNO(rcur->bc_mp, ri->ri_bmap.br_startblock); - switch (ri->ri_type) { - case XFS_RMAP_ALLOC: - case XFS_RMAP_MAP: - error = xfs_rmap_map(rcur, bno, ri->ri_bmap.br_blockcount, - unwritten, &oinfo); - break; - case XFS_RMAP_MAP_SHARED: - error = xfs_rmap_map_shared(rcur, bno, - ri->ri_bmap.br_blockcount, unwritten, &oinfo); - break; - case XFS_RMAP_FREE: - case XFS_RMAP_UNMAP: - error = xfs_rmap_unmap(rcur, bno, ri->ri_bmap.br_blockcount, - unwritten, &oinfo); - break; - case XFS_RMAP_UNMAP_SHARED: - error = xfs_rmap_unmap_shared(rcur, bno, - ri->ri_bmap.br_blockcount, unwritten, &oinfo); - break; - case XFS_RMAP_CONVERT: - error = xfs_rmap_convert(rcur, bno, ri->ri_bmap.br_blockcount, - !unwritten, &oinfo); - break; - case XFS_RMAP_CONVERT_SHARED: - error = xfs_rmap_convert_shared(rcur, bno, - ri->ri_bmap.br_blockcount, !unwritten, &oinfo); - break; - default: - ASSERT(0); - error = -EFSCORRUPTED; - } + error = __xfs_rmap_finish_intent(rcur, ri->ri_type, bno, + ri->ri_bmap.br_blockcount, &oinfo, unwritten); + if (error) + return error; - return error; + xfs_rmap_update_hook(tp, ri->ri_pag, ri->ri_type, bno, + ri->ri_bmap.br_blockcount, unwritten, &oinfo); + return 0; } /* @@ -2559,7 +2731,7 @@ __xfs_rmap_add( bmap->br_blockcount, bmap->br_state); - ri = kmem_cache_alloc(xfs_rmap_intent_cache, GFP_NOFS | __GFP_NOFAIL); + ri = kmem_cache_alloc(xfs_rmap_intent_cache, GFP_KERNEL | __GFP_NOFAIL); INIT_LIST_HEAD(&ri->ri_list); ri->ri_type = type; ri->ri_owner = owner; diff --git a/fs/xfs/libxfs/xfs_rmap.h b/fs/xfs/libxfs/xfs_rmap.h index 3c98d9d50afb..9d01fe689497 100644 --- a/fs/xfs/libxfs/xfs_rmap.h +++ b/fs/xfs/libxfs/xfs_rmap.h @@ -186,6 +186,10 @@ void xfs_rmap_finish_one_cleanup(struct xfs_trans *tp, struct xfs_btree_cur *rcur, int error); int xfs_rmap_finish_one(struct xfs_trans *tp, struct xfs_rmap_intent *ri, struct xfs_btree_cur **pcur); +int __xfs_rmap_finish_intent(struct xfs_btree_cur *rcur, + enum xfs_rmap_intent_type op, xfs_agblock_t bno, + xfs_extlen_t len, const struct xfs_owner_info *oinfo, + bool unwritten); int xfs_rmap_lookup_le_range(struct xfs_btree_cur *cur, xfs_agblock_t bno, uint64_t owner, uint64_t offset, unsigned int flags, @@ -195,7 +199,7 @@ int xfs_rmap_compare(const struct xfs_rmap_irec *a, union xfs_btree_rec; xfs_failaddr_t xfs_rmap_btrec_to_irec(const union xfs_btree_rec *rec, struct xfs_rmap_irec *irec); -xfs_failaddr_t xfs_rmap_check_irec(struct xfs_btree_cur *cur, +xfs_failaddr_t xfs_rmap_check_irec(struct xfs_perag *pag, const struct xfs_rmap_irec *irec); int xfs_rmap_has_records(struct xfs_btree_cur *cur, xfs_agblock_t bno, @@ -235,4 +239,29 @@ extern struct kmem_cache *xfs_rmap_intent_cache; int __init xfs_rmap_intent_init_cache(void); void xfs_rmap_intent_destroy_cache(void); +/* + * Parameters for tracking reverse mapping changes. The hook function arg + * parameter is enum xfs_rmap_intent_type, and the rest is below. + */ +struct xfs_rmap_update_params { + xfs_agblock_t startblock; + xfs_extlen_t blockcount; + struct xfs_owner_info oinfo; + bool unwritten; +}; + +#ifdef CONFIG_XFS_LIVE_HOOKS + +struct xfs_rmap_hook { + struct xfs_hook rmap_hook; +}; + +void xfs_rmap_hook_disable(void); +void xfs_rmap_hook_enable(void); + +int xfs_rmap_hook_add(struct xfs_perag *pag, struct xfs_rmap_hook *hook); +void xfs_rmap_hook_del(struct xfs_perag *pag, struct xfs_rmap_hook *hook); +void xfs_rmap_hook_setup(struct xfs_rmap_hook *hook, notifier_fn_t mod_fn); +#endif + #endif /* __XFS_RMAP_H__ */ diff --git a/fs/xfs/libxfs/xfs_rmap_btree.c b/fs/xfs/libxfs/xfs_rmap_btree.c index 6c81b20e97d2..9e759efa81cc 100644 --- a/fs/xfs/libxfs/xfs_rmap_btree.c +++ b/fs/xfs/libxfs/xfs_rmap_btree.c @@ -16,11 +16,14 @@ #include "xfs_btree_staging.h" #include "xfs_rmap.h" #include "xfs_rmap_btree.h" +#include "xfs_health.h" #include "xfs_trace.h" #include "xfs_error.h" #include "xfs_extent_busy.h" #include "xfs_ag.h" #include "xfs_ag_resv.h" +#include "xfs_buf_mem.h" +#include "xfs_btree_mem.h" static struct kmem_cache *xfs_rmapbt_cur_cache; @@ -65,13 +68,12 @@ xfs_rmapbt_set_root( { struct xfs_buf *agbp = cur->bc_ag.agbp; struct xfs_agf *agf = agbp->b_addr; - int btnum = cur->bc_btnum; ASSERT(ptr->s != 0); - agf->agf_roots[btnum] = ptr->s; - be32_add_cpu(&agf->agf_levels[btnum], inc); - cur->bc_ag.pag->pagf_levels[btnum] += inc; + agf->agf_rmap_root = ptr->s; + be32_add_cpu(&agf->agf_rmap_level, inc); + cur->bc_ag.pag->pagf_rmap_level += inc; xfs_alloc_log_agf(cur->bc_tp, agbp, XFS_AGF_ROOTS | XFS_AGF_LEVELS); } @@ -94,8 +96,6 @@ xfs_rmapbt_alloc_block( &bno, 1); if (error) return error; - - trace_xfs_rmapbt_alloc_block(cur->bc_mp, pag->pag_agno, bno, 1); if (bno == NULLAGBLOCK) { *stat = 0; return 0; @@ -125,8 +125,6 @@ xfs_rmapbt_free_block( int error; bno = xfs_daddr_to_agbno(cur->bc_mp, xfs_buf_daddr(bp)); - trace_xfs_rmapbt_free_block(cur->bc_mp, pag->pag_agno, - bno, 1); be32_add_cpu(&agf->agf_rmap_blocks, -1); xfs_alloc_log_agf(cur->bc_tp, agbp, XFS_AGF_RMAP_BLOCKS); error = xfs_alloc_put_freelist(pag, cur->bc_tp, agbp, NULL, bno, 1); @@ -226,7 +224,7 @@ xfs_rmapbt_init_ptr_from_cur( ASSERT(cur->bc_ag.pag->pag_agno == be32_to_cpu(agf->agf_seqno)); - ptr->s = agf->agf_roots[cur->bc_btnum]; + ptr->s = agf->agf_rmap_root; } /* @@ -340,18 +338,29 @@ xfs_rmapbt_verify( if (!xfs_has_rmapbt(mp)) return __this_address; - fa = xfs_btree_sblock_v5hdr_verify(bp); + fa = xfs_btree_agblock_v5hdr_verify(bp); if (fa) return fa; level = be16_to_cpu(block->bb_level); if (pag && xfs_perag_initialised_agf(pag)) { - if (level >= pag->pagf_levels[XFS_BTNUM_RMAPi]) + unsigned int maxlevel = pag->pagf_rmap_level; + +#ifdef CONFIG_XFS_ONLINE_REPAIR + /* + * Online repair could be rewriting the free space btrees, so + * we'll validate against the larger of either tree while this + * is going on. + */ + maxlevel = max_t(unsigned int, maxlevel, + pag->pagf_repair_rmap_level); +#endif + if (level >= maxlevel) return __this_address; } else if (level >= mp->m_rmap_maxlevels) return __this_address; - return xfs_btree_sblock_verify(bp, mp->m_rmap_mxr[level != 0]); + return xfs_btree_agblock_verify(bp, mp->m_rmap_mxr[level != 0]); } static void @@ -360,7 +369,7 @@ xfs_rmapbt_read_verify( { xfs_failaddr_t fa; - if (!xfs_btree_sblock_verify_crc(bp)) + if (!xfs_btree_agblock_verify_crc(bp)) xfs_verifier_error(bp, -EFSBADCRC, __this_address); else { fa = xfs_rmapbt_verify(bp); @@ -384,7 +393,7 @@ xfs_rmapbt_write_verify( xfs_verifier_error(bp, -EFSCORRUPTED, fa); return; } - xfs_btree_sblock_calc_crc(bp); + xfs_btree_agblock_calc_crc(bp); } @@ -476,9 +485,19 @@ xfs_rmapbt_keys_contiguous( be32_to_cpu(key2->rmap.rm_startblock)); } -static const struct xfs_btree_ops xfs_rmapbt_ops = { +const struct xfs_btree_ops xfs_rmapbt_ops = { + .name = "rmap", + .type = XFS_BTREE_TYPE_AG, + .geom_flags = XFS_BTGEO_OVERLAPPING, + .rec_len = sizeof(struct xfs_rmap_rec), + /* Overlapping btree; 2 keys per pointer. */ .key_len = 2 * sizeof(struct xfs_rmap_key), + .ptr_len = XFS_BTREE_SHORT_PTR_LEN, + + .lru_refs = XFS_RMAP_BTREE_REF, + .statoff = XFS_STATS_CALC_INDEX(xs_rmap_2), + .sick_mask = XFS_SICK_AG_RMAPBT, .dup_cursor = xfs_rmapbt_dup_cursor, .set_root = xfs_rmapbt_set_root, @@ -498,55 +517,176 @@ static const struct xfs_btree_ops xfs_rmapbt_ops = { .keys_contiguous = xfs_rmapbt_keys_contiguous, }; -static struct xfs_btree_cur * -xfs_rmapbt_init_common( +/* + * Create a new reverse mapping btree cursor. + * + * For staging cursors tp and agbp are NULL. + */ +struct xfs_btree_cur * +xfs_rmapbt_init_cursor( struct xfs_mount *mp, struct xfs_trans *tp, + struct xfs_buf *agbp, struct xfs_perag *pag) { struct xfs_btree_cur *cur; - /* Overlapping btree; 2 keys per pointer. */ - cur = xfs_btree_alloc_cursor(mp, tp, XFS_BTNUM_RMAP, + cur = xfs_btree_alloc_cursor(mp, tp, &xfs_rmapbt_ops, mp->m_rmap_maxlevels, xfs_rmapbt_cur_cache); - cur->bc_flags = XFS_BTREE_CRC_BLOCKS | XFS_BTREE_OVERLAPPING; - cur->bc_statoff = XFS_STATS_CALC_INDEX(xs_rmap_2); - cur->bc_ops = &xfs_rmapbt_ops; - cur->bc_ag.pag = xfs_perag_hold(pag); + cur->bc_ag.agbp = agbp; + if (agbp) { + struct xfs_agf *agf = agbp->b_addr; + + cur->bc_nlevels = be32_to_cpu(agf->agf_rmap_level); + } return cur; } -/* Create a new reverse mapping btree cursor. */ +#ifdef CONFIG_XFS_BTREE_IN_MEM +static inline unsigned int +xfs_rmapbt_mem_block_maxrecs( + unsigned int blocklen, + bool leaf) +{ + if (leaf) + return blocklen / sizeof(struct xfs_rmap_rec); + return blocklen / + (2 * sizeof(struct xfs_rmap_key) + sizeof(__be64)); +} + +/* + * Validate an in-memory rmap btree block. Callers are allowed to generate an + * in-memory btree even if the ondisk feature is not enabled. + */ +static xfs_failaddr_t +xfs_rmapbt_mem_verify( + struct xfs_buf *bp) +{ + struct xfs_btree_block *block = XFS_BUF_TO_BLOCK(bp); + xfs_failaddr_t fa; + unsigned int level; + unsigned int maxrecs; + + if (!xfs_verify_magic(bp, block->bb_magic)) + return __this_address; + + fa = xfs_btree_fsblock_v5hdr_verify(bp, XFS_RMAP_OWN_UNKNOWN); + if (fa) + return fa; + + level = be16_to_cpu(block->bb_level); + if (level >= xfs_rmapbt_maxlevels_ondisk()) + return __this_address; + + maxrecs = xfs_rmapbt_mem_block_maxrecs( + XFBNO_BLOCKSIZE - XFS_BTREE_LBLOCK_CRC_LEN, level == 0); + return xfs_btree_memblock_verify(bp, maxrecs); +} + +static void +xfs_rmapbt_mem_rw_verify( + struct xfs_buf *bp) +{ + xfs_failaddr_t fa = xfs_rmapbt_mem_verify(bp); + + if (fa) + xfs_verifier_error(bp, -EFSCORRUPTED, fa); +} + +/* skip crc checks on in-memory btrees to save time */ +static const struct xfs_buf_ops xfs_rmapbt_mem_buf_ops = { + .name = "xfs_rmapbt_mem", + .magic = { 0, cpu_to_be32(XFS_RMAP_CRC_MAGIC) }, + .verify_read = xfs_rmapbt_mem_rw_verify, + .verify_write = xfs_rmapbt_mem_rw_verify, + .verify_struct = xfs_rmapbt_mem_verify, +}; + +const struct xfs_btree_ops xfs_rmapbt_mem_ops = { + .name = "mem_rmap", + .type = XFS_BTREE_TYPE_MEM, + .geom_flags = XFS_BTGEO_OVERLAPPING, + + .rec_len = sizeof(struct xfs_rmap_rec), + /* Overlapping btree; 2 keys per pointer. */ + .key_len = 2 * sizeof(struct xfs_rmap_key), + .ptr_len = XFS_BTREE_LONG_PTR_LEN, + + .lru_refs = XFS_RMAP_BTREE_REF, + .statoff = XFS_STATS_CALC_INDEX(xs_rmap_mem_2), + + .dup_cursor = xfbtree_dup_cursor, + .set_root = xfbtree_set_root, + .alloc_block = xfbtree_alloc_block, + .free_block = xfbtree_free_block, + .get_minrecs = xfbtree_get_minrecs, + .get_maxrecs = xfbtree_get_maxrecs, + .init_key_from_rec = xfs_rmapbt_init_key_from_rec, + .init_high_key_from_rec = xfs_rmapbt_init_high_key_from_rec, + .init_rec_from_cur = xfs_rmapbt_init_rec_from_cur, + .init_ptr_from_cur = xfbtree_init_ptr_from_cur, + .key_diff = xfs_rmapbt_key_diff, + .buf_ops = &xfs_rmapbt_mem_buf_ops, + .diff_two_keys = xfs_rmapbt_diff_two_keys, + .keys_inorder = xfs_rmapbt_keys_inorder, + .recs_inorder = xfs_rmapbt_recs_inorder, + .keys_contiguous = xfs_rmapbt_keys_contiguous, +}; + +/* Create a cursor for an in-memory btree. */ struct xfs_btree_cur * -xfs_rmapbt_init_cursor( - struct xfs_mount *mp, +xfs_rmapbt_mem_cursor( + struct xfs_perag *pag, struct xfs_trans *tp, - struct xfs_buf *agbp, - struct xfs_perag *pag) + struct xfbtree *xfbt) { - struct xfs_agf *agf = agbp->b_addr; struct xfs_btree_cur *cur; + struct xfs_mount *mp = pag->pag_mount; - cur = xfs_rmapbt_init_common(mp, tp, pag); - cur->bc_nlevels = be32_to_cpu(agf->agf_levels[XFS_BTNUM_RMAP]); - cur->bc_ag.agbp = agbp; + cur = xfs_btree_alloc_cursor(mp, tp, &xfs_rmapbt_mem_ops, + xfs_rmapbt_maxlevels_ondisk(), xfs_rmapbt_cur_cache); + cur->bc_mem.xfbtree = xfbt; + cur->bc_nlevels = xfbt->nlevels; + + cur->bc_mem.pag = xfs_perag_hold(pag); return cur; } -/* Create a new reverse mapping btree cursor with a fake root for staging. */ -struct xfs_btree_cur * -xfs_rmapbt_stage_cursor( +/* Create an in-memory rmap btree. */ +int +xfs_rmapbt_mem_init( struct xfs_mount *mp, - struct xbtree_afakeroot *afake, - struct xfs_perag *pag) + struct xfbtree *xfbt, + struct xfs_buftarg *btp, + xfs_agnumber_t agno) { - struct xfs_btree_cur *cur; + xfbt->owner = agno; + return xfbtree_init(mp, xfbt, btp, &xfs_rmapbt_mem_ops); +} - cur = xfs_rmapbt_init_common(mp, NULL, pag); - xfs_btree_stage_afakeroot(cur, afake); - return cur; +/* Compute the max possible height for reverse mapping btrees in memory. */ +static unsigned int +xfs_rmapbt_mem_maxlevels(void) +{ + unsigned int minrecs[2]; + unsigned int blocklen; + + blocklen = XFBNO_BLOCKSIZE - XFS_BTREE_LBLOCK_CRC_LEN; + + minrecs[0] = xfs_rmapbt_mem_block_maxrecs(blocklen, true) / 2; + minrecs[1] = xfs_rmapbt_mem_block_maxrecs(blocklen, false) / 2; + + /* + * How tall can an in-memory rmap btree become if we filled the entire + * AG with rmap records? + */ + return xfs_btree_compute_maxlevels(minrecs, + XFS_MAX_AG_BYTES / sizeof(struct xfs_rmap_rec)); } +#else +# define xfs_rmapbt_mem_maxlevels() (0) +#endif /* CONFIG_XFS_BTREE_IN_MEM */ /* * Install a new reverse mapping btree root. Caller is responsible for @@ -563,12 +703,12 @@ xfs_rmapbt_commit_staged_btree( ASSERT(cur->bc_flags & XFS_BTREE_STAGING); - agf->agf_roots[cur->bc_btnum] = cpu_to_be32(afake->af_root); - agf->agf_levels[cur->bc_btnum] = cpu_to_be32(afake->af_levels); + agf->agf_rmap_root = cpu_to_be32(afake->af_root); + agf->agf_rmap_level = cpu_to_be32(afake->af_levels); agf->agf_rmap_blocks = cpu_to_be32(afake->af_blocks); xfs_alloc_log_agf(tp, agbp, XFS_AGF_ROOTS | XFS_AGF_LEVELS | XFS_AGF_RMAP_BLOCKS); - xfs_btree_commit_afakeroot(cur, tp, agbp, &xfs_rmapbt_ops); + xfs_btree_commit_afakeroot(cur, tp, agbp); } /* Calculate number of records in a reverse mapping btree block. */ @@ -618,7 +758,8 @@ xfs_rmapbt_maxlevels_ondisk(void) * like if it consumes almost all the blocks in the AG due to maximal * sharing factor. */ - return xfs_btree_space_to_height(minrecs, XFS_MAX_CRC_AG_BLOCKS); + return max(xfs_btree_space_to_height(minrecs, XFS_MAX_CRC_AG_BLOCKS), + xfs_rmapbt_mem_maxlevels()); } /* Compute the maximum height of an rmap btree. */ diff --git a/fs/xfs/libxfs/xfs_rmap_btree.h b/fs/xfs/libxfs/xfs_rmap_btree.h index 3244715dd111..eb90d89e8086 100644 --- a/fs/xfs/libxfs/xfs_rmap_btree.h +++ b/fs/xfs/libxfs/xfs_rmap_btree.h @@ -10,6 +10,7 @@ struct xfs_buf; struct xfs_btree_cur; struct xfs_mount; struct xbtree_afakeroot; +struct xfbtree; /* rmaps only exist on crc enabled filesystems */ #define XFS_RMAP_BLOCK_LEN XFS_BTREE_SBLOCK_CRC_LEN @@ -44,8 +45,6 @@ struct xbtree_afakeroot; struct xfs_btree_cur *xfs_rmapbt_init_cursor(struct xfs_mount *mp, struct xfs_trans *tp, struct xfs_buf *bp, struct xfs_perag *pag); -struct xfs_btree_cur *xfs_rmapbt_stage_cursor(struct xfs_mount *mp, - struct xbtree_afakeroot *afake, struct xfs_perag *pag); void xfs_rmapbt_commit_staged_btree(struct xfs_btree_cur *cur, struct xfs_trans *tp, struct xfs_buf *agbp); int xfs_rmapbt_maxrecs(int blocklen, int leaf); @@ -64,4 +63,9 @@ unsigned int xfs_rmapbt_maxlevels_ondisk(void); int __init xfs_rmapbt_init_cur_cache(void); void xfs_rmapbt_destroy_cur_cache(void); +struct xfs_btree_cur *xfs_rmapbt_mem_cursor(struct xfs_perag *pag, + struct xfs_trans *tp, struct xfbtree *xfbtree); +int xfs_rmapbt_mem_init(struct xfs_mount *mp, struct xfbtree *xfbtree, + struct xfs_buftarg *btp, xfs_agnumber_t agno); + #endif /* __XFS_RMAP_BTREE_H__ */ diff --git a/fs/xfs/libxfs/xfs_rtbitmap.c b/fs/xfs/libxfs/xfs_rtbitmap.c index e31663cb7b43..f246d6dbf4ec 100644 --- a/fs/xfs/libxfs/xfs_rtbitmap.c +++ b/fs/xfs/libxfs/xfs_rtbitmap.c @@ -17,6 +17,7 @@ #include "xfs_rtalloc.h" #include "xfs_error.h" #include "xfs_rtbitmap.h" +#include "xfs_health.h" /* * Realtime allocator bitmap functions shared with userspace. @@ -115,13 +116,19 @@ xfs_rtbuf_get( if (error) return error; - if (XFS_IS_CORRUPT(mp, nmap == 0 || !xfs_bmap_is_written_extent(&map))) + if (XFS_IS_CORRUPT(mp, nmap == 0 || !xfs_bmap_is_written_extent(&map))) { + xfs_rt_mark_sick(mp, issum ? XFS_SICK_RT_SUMMARY : + XFS_SICK_RT_BITMAP); return -EFSCORRUPTED; + } ASSERT(map.br_startblock != NULLFSBLOCK); error = xfs_trans_read_buf(mp, args->tp, mp->m_ddev_targp, XFS_FSB_TO_DADDR(mp, map.br_startblock), mp->m_bsize, 0, &bp, &xfs_rtbuf_ops); + if (xfs_metadata_is_sick(error)) + xfs_rt_mark_sick(mp, issum ? XFS_SICK_RT_SUMMARY : + XFS_SICK_RT_BITMAP); if (error) return error; @@ -934,7 +941,7 @@ xfs_rtfree_extent( struct timespec64 atime; ASSERT(mp->m_rbmip->i_itemp != NULL); - ASSERT(xfs_isilocked(mp->m_rbmip, XFS_ILOCK_EXCL)); + xfs_assert_ilocked(mp->m_rbmip, XFS_ILOCK_EXCL); error = xfs_rtcheck_alloc_range(&args, start, len); if (error) diff --git a/fs/xfs/libxfs/xfs_sb.c b/fs/xfs/libxfs/xfs_sb.c index 5bb6e2bd6dee..d991eec05436 100644 --- a/fs/xfs/libxfs/xfs_sb.c +++ b/fs/xfs/libxfs/xfs_sb.c @@ -1290,6 +1290,8 @@ xfs_sb_read_secondary( error = xfs_trans_read_buf(mp, tp, mp->m_ddev_targp, XFS_AG_DADDR(mp, agno, XFS_SB_BLOCK(mp)), XFS_FSS_TO_BB(mp, 1), 0, &bp, &xfs_sb_buf_ops); + if (xfs_metadata_is_sick(error)) + xfs_agno_mark_sick(mp, agno, XFS_SICK_AG_SB); if (error) return error; xfs_buf_set_ref(bp, XFS_SSB_REF); diff --git a/fs/xfs/libxfs/xfs_shared.h b/fs/xfs/libxfs/xfs_shared.h index 4220d3584c1b..dfd61fa8332e 100644 --- a/fs/xfs/libxfs/xfs_shared.h +++ b/fs/xfs/libxfs/xfs_shared.h @@ -43,6 +43,60 @@ extern const struct xfs_buf_ops xfs_sb_buf_ops; extern const struct xfs_buf_ops xfs_sb_quiet_buf_ops; extern const struct xfs_buf_ops xfs_symlink_buf_ops; +/* btree ops */ +extern const struct xfs_btree_ops xfs_bnobt_ops; +extern const struct xfs_btree_ops xfs_cntbt_ops; +extern const struct xfs_btree_ops xfs_inobt_ops; +extern const struct xfs_btree_ops xfs_finobt_ops; +extern const struct xfs_btree_ops xfs_bmbt_ops; +extern const struct xfs_btree_ops xfs_refcountbt_ops; +extern const struct xfs_btree_ops xfs_rmapbt_ops; +extern const struct xfs_btree_ops xfs_rmapbt_mem_ops; + +static inline bool xfs_btree_is_bno(const struct xfs_btree_ops *ops) +{ + return ops == &xfs_bnobt_ops; +} + +static inline bool xfs_btree_is_cnt(const struct xfs_btree_ops *ops) +{ + return ops == &xfs_cntbt_ops; +} + +static inline bool xfs_btree_is_bmap(const struct xfs_btree_ops *ops) +{ + return ops == &xfs_bmbt_ops; +} + +static inline bool xfs_btree_is_ino(const struct xfs_btree_ops *ops) +{ + return ops == &xfs_inobt_ops; +} + +static inline bool xfs_btree_is_fino(const struct xfs_btree_ops *ops) +{ + return ops == &xfs_finobt_ops; +} + +static inline bool xfs_btree_is_refcount(const struct xfs_btree_ops *ops) +{ + return ops == &xfs_refcountbt_ops; +} + +static inline bool xfs_btree_is_rmap(const struct xfs_btree_ops *ops) +{ + return ops == &xfs_rmapbt_ops; +} + +#ifdef CONFIG_XFS_BTREE_IN_MEM +static inline bool xfs_btree_is_mem_rmap(const struct xfs_btree_ops *ops) +{ + return ops == &xfs_rmapbt_mem_ops; +} +#else +# define xfs_btree_is_mem_rmap(...) (false) +#endif + /* log size calculation functions */ int xfs_log_calc_unit_res(struct xfs_mount *mp, int unit_bytes); int xfs_log_calc_minimum_size(struct xfs_mount *); @@ -128,19 +182,6 @@ void xfs_log_get_max_trans_res(struct xfs_mount *mp, #define XFS_ICHGTIME_CHG 0x2 /* inode field change timestamp */ #define XFS_ICHGTIME_CREATE 0x4 /* inode create timestamp */ - -/* - * Symlink decoding/encoding functions - */ -int xfs_symlink_blocks(struct xfs_mount *mp, int pathlen); -int xfs_symlink_hdr_set(struct xfs_mount *mp, xfs_ino_t ino, uint32_t offset, - uint32_t size, struct xfs_buf *bp); -bool xfs_symlink_hdr_ok(xfs_ino_t ino, uint32_t offset, - uint32_t size, struct xfs_buf *bp); -void xfs_symlink_local_to_remote(struct xfs_trans *tp, struct xfs_buf *bp, - struct xfs_inode *ip, struct xfs_ifork *ifp); -xfs_failaddr_t xfs_symlink_shortform_verify(void *sfp, int64_t size); - /* Computed inode geometry for the filesystem. */ struct xfs_ino_geometry { /* Maximum inode count in this filesystem. */ diff --git a/fs/xfs/libxfs/xfs_symlink_remote.c b/fs/xfs/libxfs/xfs_symlink_remote.c index 160aa20aa441..ffb1317a9212 100644 --- a/fs/xfs/libxfs/xfs_symlink_remote.c +++ b/fs/xfs/libxfs/xfs_symlink_remote.c @@ -16,7 +16,10 @@ #include "xfs_trans.h" #include "xfs_buf_item.h" #include "xfs_log.h" - +#include "xfs_symlink_remote.h" +#include "xfs_bit.h" +#include "xfs_bmap.h" +#include "xfs_health.h" /* * Each contiguous block has a header, so it is not just a simple pathlen @@ -227,3 +230,153 @@ xfs_symlink_shortform_verify( return __this_address; return NULL; } + +/* Read a remote symlink target into the buffer. */ +int +xfs_symlink_remote_read( + struct xfs_inode *ip, + char *link) +{ + struct xfs_mount *mp = ip->i_mount; + struct xfs_bmbt_irec mval[XFS_SYMLINK_MAPS]; + struct xfs_buf *bp; + xfs_daddr_t d; + char *cur_chunk; + int pathlen = ip->i_disk_size; + int nmaps = XFS_SYMLINK_MAPS; + int byte_cnt; + int n; + int error = 0; + int fsblocks = 0; + int offset; + + xfs_assert_ilocked(ip, XFS_ILOCK_SHARED | XFS_ILOCK_EXCL); + + fsblocks = xfs_symlink_blocks(mp, pathlen); + error = xfs_bmapi_read(ip, 0, fsblocks, mval, &nmaps, 0); + if (error) + goto out; + + offset = 0; + for (n = 0; n < nmaps; n++) { + d = XFS_FSB_TO_DADDR(mp, mval[n].br_startblock); + byte_cnt = XFS_FSB_TO_B(mp, mval[n].br_blockcount); + + error = xfs_buf_read(mp->m_ddev_targp, d, BTOBB(byte_cnt), 0, + &bp, &xfs_symlink_buf_ops); + if (xfs_metadata_is_sick(error)) + xfs_inode_mark_sick(ip, XFS_SICK_INO_SYMLINK); + if (error) + return error; + byte_cnt = XFS_SYMLINK_BUF_SPACE(mp, byte_cnt); + if (pathlen < byte_cnt) + byte_cnt = pathlen; + + cur_chunk = bp->b_addr; + if (xfs_has_crc(mp)) { + if (!xfs_symlink_hdr_ok(ip->i_ino, offset, + byte_cnt, bp)) { + xfs_inode_mark_sick(ip, XFS_SICK_INO_SYMLINK); + error = -EFSCORRUPTED; + xfs_alert(mp, +"symlink header does not match required off/len/owner (0x%x/0x%x,0x%llx)", + offset, byte_cnt, ip->i_ino); + xfs_buf_relse(bp); + goto out; + + } + + cur_chunk += sizeof(struct xfs_dsymlink_hdr); + } + + memcpy(link + offset, cur_chunk, byte_cnt); + + pathlen -= byte_cnt; + offset += byte_cnt; + + xfs_buf_relse(bp); + } + ASSERT(pathlen == 0); + + link[ip->i_disk_size] = '\0'; + error = 0; + + out: + return error; +} + +/* Write the symlink target into the inode. */ +int +xfs_symlink_write_target( + struct xfs_trans *tp, + struct xfs_inode *ip, + const char *target_path, + int pathlen, + xfs_fsblock_t fs_blocks, + uint resblks) +{ + struct xfs_bmbt_irec mval[XFS_SYMLINK_MAPS]; + struct xfs_mount *mp = tp->t_mountp; + const char *cur_chunk; + struct xfs_buf *bp; + xfs_daddr_t d; + int byte_cnt; + int nmaps; + int offset = 0; + int n; + int error; + + /* + * If the symlink will fit into the inode, write it inline. + */ + if (pathlen <= xfs_inode_data_fork_size(ip)) { + xfs_init_local_fork(ip, XFS_DATA_FORK, target_path, pathlen); + + ip->i_disk_size = pathlen; + ip->i_df.if_format = XFS_DINODE_FMT_LOCAL; + xfs_trans_log_inode(tp, ip, XFS_ILOG_DDATA | XFS_ILOG_CORE); + return 0; + } + + nmaps = XFS_SYMLINK_MAPS; + error = xfs_bmapi_write(tp, ip, 0, fs_blocks, XFS_BMAPI_METADATA, + resblks, mval, &nmaps); + if (error) + return error; + + ip->i_disk_size = pathlen; + xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE); + + cur_chunk = target_path; + offset = 0; + for (n = 0; n < nmaps; n++) { + char *buf; + + d = XFS_FSB_TO_DADDR(mp, mval[n].br_startblock); + byte_cnt = XFS_FSB_TO_B(mp, mval[n].br_blockcount); + error = xfs_trans_get_buf(tp, mp->m_ddev_targp, d, + BTOBB(byte_cnt), 0, &bp); + if (error) + return error; + bp->b_ops = &xfs_symlink_buf_ops; + + byte_cnt = XFS_SYMLINK_BUF_SPACE(mp, byte_cnt); + byte_cnt = min(byte_cnt, pathlen); + + buf = bp->b_addr; + buf += xfs_symlink_hdr_set(mp, ip->i_ino, offset, byte_cnt, + bp); + + memcpy(buf, cur_chunk, byte_cnt); + + cur_chunk += byte_cnt; + pathlen -= byte_cnt; + offset += byte_cnt; + + xfs_trans_buf_set_type(tp, bp, XFS_BLFT_SYMLINK_BUF); + xfs_trans_log_buf(tp, bp, 0, (buf + byte_cnt - 1) - + (char *)bp->b_addr); + } + ASSERT(pathlen == 0); + return 0; +} diff --git a/fs/xfs/libxfs/xfs_symlink_remote.h b/fs/xfs/libxfs/xfs_symlink_remote.h new file mode 100644 index 000000000000..a63bd38ae4fa --- /dev/null +++ b/fs/xfs/libxfs/xfs_symlink_remote.h @@ -0,0 +1,26 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright (c) 2000-2005 Silicon Graphics, Inc. + * Copyright (c) 2013 Red Hat, Inc. + * All Rights Reserved. + */ +#ifndef __XFS_SYMLINK_REMOTE_H +#define __XFS_SYMLINK_REMOTE_H + +/* + * Symlink decoding/encoding functions + */ +int xfs_symlink_blocks(struct xfs_mount *mp, int pathlen); +int xfs_symlink_hdr_set(struct xfs_mount *mp, xfs_ino_t ino, uint32_t offset, + uint32_t size, struct xfs_buf *bp); +bool xfs_symlink_hdr_ok(xfs_ino_t ino, uint32_t offset, + uint32_t size, struct xfs_buf *bp); +void xfs_symlink_local_to_remote(struct xfs_trans *tp, struct xfs_buf *bp, + struct xfs_inode *ip, struct xfs_ifork *ifp); +xfs_failaddr_t xfs_symlink_shortform_verify(void *sfp, int64_t size); +int xfs_symlink_remote_read(struct xfs_inode *ip, char *link); +int xfs_symlink_write_target(struct xfs_trans *tp, struct xfs_inode *ip, + const char *target_path, int pathlen, xfs_fsblock_t fs_blocks, + uint resblks); + +#endif /* __XFS_SYMLINK_REMOTE_H */ diff --git a/fs/xfs/libxfs/xfs_trans_inode.c b/fs/xfs/libxfs/xfs_trans_inode.c index 70e97ea6eee7..69fc5b981352 100644 --- a/fs/xfs/libxfs/xfs_trans_inode.c +++ b/fs/xfs/libxfs/xfs_trans_inode.c @@ -31,7 +31,7 @@ xfs_trans_ijoin( { struct xfs_inode_log_item *iip; - ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL)); + xfs_assert_ilocked(ip, XFS_ILOCK_EXCL); if (ip->i_itemp == NULL) xfs_inode_item_init(ip, ip->i_mount); iip = ip->i_itemp; @@ -60,7 +60,7 @@ xfs_trans_ichgtime( struct timespec64 tv; ASSERT(tp); - ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL)); + xfs_assert_ilocked(ip, XFS_ILOCK_EXCL); tv = current_time(inode); @@ -90,7 +90,7 @@ xfs_trans_log_inode( struct inode *inode = VFS_I(ip); ASSERT(iip); - ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL)); + xfs_assert_ilocked(ip, XFS_ILOCK_EXCL); ASSERT(!xfs_iflags_test(ip, XFS_ISTALE)); tp->t_flags |= XFS_TRANS_DIRTY; diff --git a/fs/xfs/libxfs/xfs_types.h b/fs/xfs/libxfs/xfs_types.h index 62e02d5380ad..76eb9e328835 100644 --- a/fs/xfs/libxfs/xfs_types.h +++ b/fs/xfs/libxfs/xfs_types.h @@ -80,11 +80,13 @@ typedef void * xfs_failaddr_t; /* * Inode fork identifiers. */ -#define XFS_DATA_FORK 0 -#define XFS_ATTR_FORK 1 -#define XFS_COW_FORK 2 +#define XFS_STAGING_FORK (-1) /* fake fork for staging a btree */ +#define XFS_DATA_FORK (0) +#define XFS_ATTR_FORK (1) +#define XFS_COW_FORK (2) #define XFS_WHICHFORK_STRINGS \ + { XFS_STAGING_FORK, "staging" }, \ { XFS_DATA_FORK, "data" }, \ { XFS_ATTR_FORK, "attr" }, \ { XFS_COW_FORK, "cow" } @@ -114,24 +116,6 @@ typedef enum { { XFS_LOOKUP_LEi, "le" }, \ { XFS_LOOKUP_GEi, "ge" } -/* - * This enum is used in string mapping in xfs_trace.h and scrub/trace.h; - * please keep the TRACE_DEFINE_ENUMs for it up to date. - */ -typedef enum { - XFS_BTNUM_BNOi, XFS_BTNUM_CNTi, XFS_BTNUM_RMAPi, XFS_BTNUM_BMAPi, - XFS_BTNUM_INOi, XFS_BTNUM_FINOi, XFS_BTNUM_REFCi, XFS_BTNUM_MAX -} xfs_btnum_t; - -#define XFS_BTNUM_STRINGS \ - { XFS_BTNUM_BNOi, "bnobt" }, \ - { XFS_BTNUM_CNTi, "cntbt" }, \ - { XFS_BTNUM_RMAPi, "rmapbt" }, \ - { XFS_BTNUM_BMAPi, "bmbt" }, \ - { XFS_BTNUM_INOi, "inobt" }, \ - { XFS_BTNUM_FINOi, "finobt" }, \ - { XFS_BTNUM_REFCi, "refcbt" } - struct xfs_name { const unsigned char *name; int len; diff --git a/fs/xfs/mrlock.h b/fs/xfs/mrlock.h deleted file mode 100644 index 79155eec341b..000000000000 --- a/fs/xfs/mrlock.h +++ /dev/null @@ -1,78 +0,0 @@ -// SPDX-License-Identifier: GPL-2.0 -/* - * Copyright (c) 2000-2006 Silicon Graphics, Inc. - * All Rights Reserved. - */ -#ifndef __XFS_SUPPORT_MRLOCK_H__ -#define __XFS_SUPPORT_MRLOCK_H__ - -#include <linux/rwsem.h> - -typedef struct { - struct rw_semaphore mr_lock; -#if defined(DEBUG) || defined(XFS_WARN) - int mr_writer; -#endif -} mrlock_t; - -#if defined(DEBUG) || defined(XFS_WARN) -#define mrinit(mrp, name) \ - do { (mrp)->mr_writer = 0; init_rwsem(&(mrp)->mr_lock); } while (0) -#else -#define mrinit(mrp, name) \ - do { init_rwsem(&(mrp)->mr_lock); } while (0) -#endif - -#define mrlock_init(mrp, t,n,s) mrinit(mrp, n) -#define mrfree(mrp) do { } while (0) - -static inline void mraccess_nested(mrlock_t *mrp, int subclass) -{ - down_read_nested(&mrp->mr_lock, subclass); -} - -static inline void mrupdate_nested(mrlock_t *mrp, int subclass) -{ - down_write_nested(&mrp->mr_lock, subclass); -#if defined(DEBUG) || defined(XFS_WARN) - mrp->mr_writer = 1; -#endif -} - -static inline int mrtryaccess(mrlock_t *mrp) -{ - return down_read_trylock(&mrp->mr_lock); -} - -static inline int mrtryupdate(mrlock_t *mrp) -{ - if (!down_write_trylock(&mrp->mr_lock)) - return 0; -#if defined(DEBUG) || defined(XFS_WARN) - mrp->mr_writer = 1; -#endif - return 1; -} - -static inline void mrunlock_excl(mrlock_t *mrp) -{ -#if defined(DEBUG) || defined(XFS_WARN) - mrp->mr_writer = 0; -#endif - up_write(&mrp->mr_lock); -} - -static inline void mrunlock_shared(mrlock_t *mrp) -{ - up_read(&mrp->mr_lock); -} - -static inline void mrdemote(mrlock_t *mrp) -{ -#if defined(DEBUG) || defined(XFS_WARN) - mrp->mr_writer = 0; -#endif - downgrade_write(&mrp->mr_lock); -} - -#endif /* __XFS_SUPPORT_MRLOCK_H__ */ diff --git a/fs/xfs/scrub/agb_bitmap.h b/fs/xfs/scrub/agb_bitmap.h index ed08f76ff4f3..e488e1f4f63d 100644 --- a/fs/xfs/scrub/agb_bitmap.h +++ b/fs/xfs/scrub/agb_bitmap.h @@ -65,4 +65,9 @@ int xagb_bitmap_set_btblocks(struct xagb_bitmap *bitmap, int xagb_bitmap_set_btcur_path(struct xagb_bitmap *bitmap, struct xfs_btree_cur *cur); +static inline uint32_t xagb_bitmap_count_set_regions(struct xagb_bitmap *b) +{ + return xbitmap32_count_set_regions(&b->agbitmap); +} + #endif /* __XFS_SCRUB_AGB_BITMAP_H__ */ diff --git a/fs/xfs/scrub/agheader.c b/fs/xfs/scrub/agheader.c index 6c6e5eba42c8..e954f07679dd 100644 --- a/fs/xfs/scrub/agheader.c +++ b/fs/xfs/scrub/agheader.c @@ -556,28 +556,28 @@ xchk_agf( xchk_block_set_corrupt(sc, sc->sa.agf_bp); /* Check the AGF btree roots and levels */ - agbno = be32_to_cpu(agf->agf_roots[XFS_BTNUM_BNO]); + agbno = be32_to_cpu(agf->agf_bno_root); if (!xfs_verify_agbno(pag, agbno)) xchk_block_set_corrupt(sc, sc->sa.agf_bp); - agbno = be32_to_cpu(agf->agf_roots[XFS_BTNUM_CNT]); + agbno = be32_to_cpu(agf->agf_cnt_root); if (!xfs_verify_agbno(pag, agbno)) xchk_block_set_corrupt(sc, sc->sa.agf_bp); - level = be32_to_cpu(agf->agf_levels[XFS_BTNUM_BNO]); + level = be32_to_cpu(agf->agf_bno_level); if (level <= 0 || level > mp->m_alloc_maxlevels) xchk_block_set_corrupt(sc, sc->sa.agf_bp); - level = be32_to_cpu(agf->agf_levels[XFS_BTNUM_CNT]); + level = be32_to_cpu(agf->agf_cnt_level); if (level <= 0 || level > mp->m_alloc_maxlevels) xchk_block_set_corrupt(sc, sc->sa.agf_bp); if (xfs_has_rmapbt(mp)) { - agbno = be32_to_cpu(agf->agf_roots[XFS_BTNUM_RMAP]); + agbno = be32_to_cpu(agf->agf_rmap_root); if (!xfs_verify_agbno(pag, agbno)) xchk_block_set_corrupt(sc, sc->sa.agf_bp); - level = be32_to_cpu(agf->agf_levels[XFS_BTNUM_RMAP]); + level = be32_to_cpu(agf->agf_rmap_level); if (level <= 0 || level > mp->m_rmap_maxlevels) xchk_block_set_corrupt(sc, sc->sa.agf_bp); } diff --git a/fs/xfs/scrub/agheader_repair.c b/fs/xfs/scrub/agheader_repair.c index 26bd1ff68f1b..427054b65b23 100644 --- a/fs/xfs/scrub/agheader_repair.c +++ b/fs/xfs/scrub/agheader_repair.c @@ -174,8 +174,7 @@ xrep_agf_find_btrees( * We relied on the rmapbt to reconstruct the AGF. If we get a * different root then something's seriously wrong. */ - if (fab[XREP_AGF_RMAPBT].root != - be32_to_cpu(old_agf->agf_roots[XFS_BTNUM_RMAPi])) + if (fab[XREP_AGF_RMAPBT].root != be32_to_cpu(old_agf->agf_rmap_root)) return -EFSCORRUPTED; /* We must find the refcountbt root if that feature is enabled. */ @@ -224,20 +223,14 @@ xrep_agf_set_roots( struct xfs_agf *agf, struct xrep_find_ag_btree *fab) { - agf->agf_roots[XFS_BTNUM_BNOi] = - cpu_to_be32(fab[XREP_AGF_BNOBT].root); - agf->agf_levels[XFS_BTNUM_BNOi] = - cpu_to_be32(fab[XREP_AGF_BNOBT].height); + agf->agf_bno_root = cpu_to_be32(fab[XREP_AGF_BNOBT].root); + agf->agf_bno_level = cpu_to_be32(fab[XREP_AGF_BNOBT].height); - agf->agf_roots[XFS_BTNUM_CNTi] = - cpu_to_be32(fab[XREP_AGF_CNTBT].root); - agf->agf_levels[XFS_BTNUM_CNTi] = - cpu_to_be32(fab[XREP_AGF_CNTBT].height); + agf->agf_cnt_root = cpu_to_be32(fab[XREP_AGF_CNTBT].root); + agf->agf_cnt_level = cpu_to_be32(fab[XREP_AGF_CNTBT].height); - agf->agf_roots[XFS_BTNUM_RMAPi] = - cpu_to_be32(fab[XREP_AGF_RMAPBT].root); - agf->agf_levels[XFS_BTNUM_RMAPi] = - cpu_to_be32(fab[XREP_AGF_RMAPBT].height); + agf->agf_rmap_root = cpu_to_be32(fab[XREP_AGF_RMAPBT].root); + agf->agf_rmap_level = cpu_to_be32(fab[XREP_AGF_RMAPBT].height); if (xfs_has_reflink(sc->mp)) { agf->agf_refcount_root = @@ -262,8 +255,7 @@ xrep_agf_calc_from_btrees( int error; /* Update the AGF counters from the bnobt. */ - cur = xfs_allocbt_init_cursor(mp, sc->tp, agf_bp, - sc->sa.pag, XFS_BTNUM_BNO); + cur = xfs_bnobt_init_cursor(mp, sc->tp, agf_bp, sc->sa.pag); error = xfs_alloc_query_all(cur, xrep_agf_walk_allocbt, &raa); if (error) goto err; @@ -276,8 +268,7 @@ xrep_agf_calc_from_btrees( agf->agf_longest = cpu_to_be32(raa.longest); /* Update the AGF counters from the cntbt. */ - cur = xfs_allocbt_init_cursor(mp, sc->tp, agf_bp, - sc->sa.pag, XFS_BTNUM_CNT); + cur = xfs_cntbt_init_cursor(mp, sc->tp, agf_bp, sc->sa.pag); error = xfs_btree_count_blocks(cur, &blocks); if (error) goto err; @@ -333,12 +324,9 @@ xrep_agf_commit_new( pag->pagf_btreeblks = be32_to_cpu(agf->agf_btreeblks); pag->pagf_freeblks = be32_to_cpu(agf->agf_freeblks); pag->pagf_longest = be32_to_cpu(agf->agf_longest); - pag->pagf_levels[XFS_BTNUM_BNOi] = - be32_to_cpu(agf->agf_levels[XFS_BTNUM_BNOi]); - pag->pagf_levels[XFS_BTNUM_CNTi] = - be32_to_cpu(agf->agf_levels[XFS_BTNUM_CNTi]); - pag->pagf_levels[XFS_BTNUM_RMAPi] = - be32_to_cpu(agf->agf_levels[XFS_BTNUM_RMAPi]); + pag->pagf_bno_level = be32_to_cpu(agf->agf_bno_level); + pag->pagf_cnt_level = be32_to_cpu(agf->agf_cnt_level); + pag->pagf_rmap_level = be32_to_cpu(agf->agf_rmap_level); pag->pagf_refcount_level = be32_to_cpu(agf->agf_refcount_level); set_bit(XFS_AGSTATE_AGF_INIT, &pag->pag_opstate); @@ -559,16 +547,14 @@ xrep_agfl_collect_blocks( goto out_bmp; /* Find all blocks currently being used by the bnobt. */ - cur = xfs_allocbt_init_cursor(mp, sc->tp, agf_bp, - sc->sa.pag, XFS_BTNUM_BNO); + cur = xfs_bnobt_init_cursor(mp, sc->tp, agf_bp, sc->sa.pag); error = xagb_bitmap_set_btblocks(&ra.agmetablocks, cur); xfs_btree_del_cursor(cur, error); if (error) goto out_bmp; /* Find all blocks currently being used by the cntbt. */ - cur = xfs_allocbt_init_cursor(mp, sc->tp, agf_bp, - sc->sa.pag, XFS_BTNUM_CNT); + cur = xfs_cntbt_init_cursor(mp, sc->tp, agf_bp, sc->sa.pag); error = xagb_bitmap_set_btblocks(&ra.agmetablocks, cur); xfs_btree_del_cursor(cur, error); if (error) @@ -908,7 +894,7 @@ xrep_agi_calc_from_btrees( xfs_agino_t freecount; int error; - cur = xfs_inobt_init_cursor(sc->sa.pag, sc->tp, agi_bp, XFS_BTNUM_INO); + cur = xfs_inobt_init_cursor(sc->sa.pag, sc->tp, agi_bp); error = xfs_ialloc_count_inodes(cur, &count, &freecount); if (error) goto err; @@ -928,8 +914,7 @@ xrep_agi_calc_from_btrees( if (xfs_has_finobt(mp) && xfs_has_inobtcounts(mp)) { xfs_agblock_t blocks; - cur = xfs_inobt_init_cursor(sc->sa.pag, sc->tp, agi_bp, - XFS_BTNUM_FINO); + cur = xfs_finobt_init_cursor(sc->sa.pag, sc->tp, agi_bp); error = xfs_btree_count_blocks(cur, &blocks); if (error) goto err; diff --git a/fs/xfs/scrub/alloc_repair.c b/fs/xfs/scrub/alloc_repair.c index 45edda096869..d421b253923e 100644 --- a/fs/xfs/scrub/alloc_repair.c +++ b/fs/xfs/scrub/alloc_repair.c @@ -687,8 +687,8 @@ xrep_abt_reset_counters( * height values before re-initializing the perag info from the updated * AGF to capture all the new values. */ - pag->pagf_repair_levels[XFS_BTNUM_BNOi] = pag->pagf_levels[XFS_BTNUM_BNOi]; - pag->pagf_repair_levels[XFS_BTNUM_CNTi] = pag->pagf_levels[XFS_BTNUM_CNTi]; + pag->pagf_repair_bno_level = pag->pagf_bno_level; + pag->pagf_repair_cnt_level = pag->pagf_cnt_level; /* Reinitialize with the values we just logged. */ return xrep_reinit_pagf(sc); @@ -735,10 +735,11 @@ xrep_abt_build_new_trees( ra->new_cntbt.bload.claim_block = xrep_abt_claim_block; /* Allocate cursors for the staged btrees. */ - bno_cur = xfs_allocbt_stage_cursor(sc->mp, &ra->new_bnobt.afake, - pag, XFS_BTNUM_BNO); - cnt_cur = xfs_allocbt_stage_cursor(sc->mp, &ra->new_cntbt.afake, - pag, XFS_BTNUM_CNT); + bno_cur = xfs_bnobt_init_cursor(sc->mp, NULL, NULL, pag); + xfs_btree_stage_afakeroot(bno_cur, &ra->new_bnobt.afake); + + cnt_cur = xfs_cntbt_init_cursor(sc->mp, NULL, NULL, pag); + xfs_btree_stage_afakeroot(cnt_cur, &ra->new_cntbt.afake); /* Last chance to abort before we start committing fixes. */ if (xchk_should_terminate(sc, &error)) @@ -765,10 +766,8 @@ xrep_abt_build_new_trees( * height so that we don't trip the verifiers when writing the new * btree blocks to disk. */ - pag->pagf_repair_levels[XFS_BTNUM_BNOi] = - ra->new_bnobt.bload.btree_height; - pag->pagf_repair_levels[XFS_BTNUM_CNTi] = - ra->new_cntbt.bload.btree_height; + pag->pagf_repair_bno_level = ra->new_bnobt.bload.btree_height; + pag->pagf_repair_cnt_level = ra->new_cntbt.bload.btree_height; /* Load the free space by length tree. */ ra->array_cur = XFARRAY_CURSOR_INIT; @@ -807,8 +806,8 @@ xrep_abt_build_new_trees( return xrep_roll_ag_trans(sc); err_levels: - pag->pagf_repair_levels[XFS_BTNUM_BNOi] = 0; - pag->pagf_repair_levels[XFS_BTNUM_CNTi] = 0; + pag->pagf_repair_bno_level = 0; + pag->pagf_repair_cnt_level = 0; err_cur: xfs_btree_del_cursor(cnt_cur, error); xfs_btree_del_cursor(bno_cur, error); @@ -838,8 +837,8 @@ xrep_abt_remove_old_trees( * Now that we've zapped all the old allocbt blocks we can turn off * the alternate height mechanism. */ - pag->pagf_repair_levels[XFS_BTNUM_BNOi] = 0; - pag->pagf_repair_levels[XFS_BTNUM_CNTi] = 0; + pag->pagf_repair_bno_level = 0; + pag->pagf_repair_cnt_level = 0; return 0; } diff --git a/fs/xfs/scrub/bitmap.c b/fs/xfs/scrub/bitmap.c index 1449bb5262d9..0cb8d43912a8 100644 --- a/fs/xfs/scrub/bitmap.c +++ b/fs/xfs/scrub/bitmap.c @@ -566,3 +566,17 @@ xbitmap32_test( *len = bn->bn_start - start; return false; } + +/* Count the number of set regions in this bitmap. */ +uint32_t +xbitmap32_count_set_regions( + struct xbitmap32 *bitmap) +{ + struct xbitmap32_node *bn; + uint32_t nr = 0; + + for_each_xbitmap32_extent(bn, bitmap) + nr++; + + return nr; +} diff --git a/fs/xfs/scrub/bitmap.h b/fs/xfs/scrub/bitmap.h index 2df8911606d6..710c1ac5e323 100644 --- a/fs/xfs/scrub/bitmap.h +++ b/fs/xfs/scrub/bitmap.h @@ -62,4 +62,6 @@ int xbitmap32_walk(struct xbitmap32 *bitmap, xbitmap32_walk_fn fn, bool xbitmap32_empty(struct xbitmap32 *bitmap); bool xbitmap32_test(struct xbitmap32 *bitmap, uint32_t start, uint32_t *len); +uint32_t xbitmap32_count_set_regions(struct xbitmap32 *bitmap); + #endif /* __XFS_SCRUB_BITMAP_H__ */ diff --git a/fs/xfs/scrub/bmap.c b/fs/xfs/scrub/bmap.c index b169cddde6da..24a15bf784f1 100644 --- a/fs/xfs/scrub/bmap.c +++ b/fs/xfs/scrub/bmap.c @@ -924,7 +924,7 @@ xchk_bmap( if (!ifp) return -ENOENT; - info.is_rt = whichfork == XFS_DATA_FORK && XFS_IS_REALTIME_INODE(ip); + info.is_rt = xfs_ifork_is_realtime(ip, whichfork); info.whichfork = whichfork; info.is_shared = whichfork == XFS_DATA_FORK && xfs_is_reflink_inode(ip); info.sc = sc; diff --git a/fs/xfs/scrub/bmap_repair.c b/fs/xfs/scrub/bmap_repair.c index a4bb89fdd510..1e656fab5e41 100644 --- a/fs/xfs/scrub/bmap_repair.c +++ b/fs/xfs/scrub/bmap_repair.c @@ -639,7 +639,13 @@ xrep_bmap_build_new_fork( rb->new_bmapbt.bload.get_records = xrep_bmap_get_records; rb->new_bmapbt.bload.claim_block = xrep_bmap_claim_block; rb->new_bmapbt.bload.iroot_size = xrep_bmap_iroot_size; - bmap_cur = xfs_bmbt_stage_cursor(sc->mp, sc->ip, ifake); + + /* + * Allocate a new bmap btree cursor for reloading an inode block mapping + * data structure. + */ + bmap_cur = xfs_bmbt_init_cursor(sc->mp, NULL, sc->ip, XFS_STAGING_FORK); + xfs_btree_stage_ifakeroot(bmap_cur, ifake); /* * Figure out the size and format of the new fork, then fill it with diff --git a/fs/xfs/scrub/btree.c b/fs/xfs/scrub/btree.c index 1935b9ce1885..fe678a0438bc 100644 --- a/fs/xfs/scrub/btree.c +++ b/fs/xfs/scrub/btree.c @@ -47,7 +47,7 @@ __xchk_btree_process_error( *error = 0; fallthrough; default: - if (cur->bc_flags & XFS_BTREE_ROOT_IN_INODE) + if (cur->bc_ops->type == XFS_BTREE_TYPE_INODE) trace_xchk_ifork_btree_op_error(sc, cur, level, *error, ret_ip); else @@ -91,7 +91,7 @@ __xchk_btree_set_corrupt( { sc->sm->sm_flags |= errflag; - if (cur->bc_flags & XFS_BTREE_ROOT_IN_INODE) + if (cur->bc_ops->type == XFS_BTREE_TYPE_INODE) trace_xchk_ifork_btree_error(sc, cur, level, ret_ip); else @@ -168,7 +168,7 @@ xchk_btree_rec( if (xfs_btree_keycmp_lt(cur, &key, keyp)) xchk_btree_set_corrupt(bs->sc, cur, 1); - if (!(cur->bc_flags & XFS_BTREE_OVERLAPPING)) + if (!(cur->bc_ops->geom_flags & XFS_BTGEO_OVERLAPPING)) return; /* Is high_key(rec) no larger than the parent high key? */ @@ -215,7 +215,7 @@ xchk_btree_key( if (xfs_btree_keycmp_lt(cur, key, keyp)) xchk_btree_set_corrupt(bs->sc, cur, level); - if (!(cur->bc_flags & XFS_BTREE_OVERLAPPING)) + if (!(cur->bc_ops->geom_flags & XFS_BTGEO_OVERLAPPING)) return; /* Is this block's high key no larger than the parent high key? */ @@ -236,22 +236,18 @@ xchk_btree_ptr_ok( int level, union xfs_btree_ptr *ptr) { - bool res; - /* A btree rooted in an inode has no block pointer to the root. */ - if ((bs->cur->bc_flags & XFS_BTREE_ROOT_IN_INODE) && + if (bs->cur->bc_ops->type == XFS_BTREE_TYPE_INODE && level == bs->cur->bc_nlevels) return true; /* Otherwise, check the pointers. */ - if (bs->cur->bc_flags & XFS_BTREE_LONG_PTRS) - res = xfs_btree_check_lptr(bs->cur, be64_to_cpu(ptr->l), level); - else - res = xfs_btree_check_sptr(bs->cur, be32_to_cpu(ptr->s), level); - if (!res) + if (__xfs_btree_check_ptr(bs->cur, ptr, 0, level)) { xchk_btree_set_corrupt(bs->sc, bs->cur, level); + return false; + } - return res; + return true; } /* Check that a btree block's sibling matches what we expect it. */ @@ -374,18 +370,21 @@ xchk_btree_check_block_owner( { xfs_agnumber_t agno; xfs_agblock_t agbno; - xfs_btnum_t btnum; bool init_sa; int error = 0; if (!bs->cur) return 0; - btnum = bs->cur->bc_btnum; agno = xfs_daddr_to_agno(bs->cur->bc_mp, daddr); agbno = xfs_daddr_to_agbno(bs->cur->bc_mp, daddr); - init_sa = bs->cur->bc_flags & XFS_BTREE_LONG_PTRS; + /* + * If the btree being examined is not itself a per-AG btree, initialize + * sc->sa so that we can check for the presence of an ownership record + * in the rmap btree for the AG containing the block. + */ + init_sa = bs->cur->bc_ops->type != XFS_BTREE_TYPE_AG; if (init_sa) { error = xchk_ag_init_existing(bs->sc, agno, &bs->sc->sa); if (!xchk_btree_xref_process_error(bs->sc, bs->cur, @@ -399,11 +398,11 @@ xchk_btree_check_block_owner( * have to nullify it (to shut down further block owner checks) if * self-xref encounters problems. */ - if (!bs->sc->sa.bno_cur && btnum == XFS_BTNUM_BNO) + if (!bs->sc->sa.bno_cur && xfs_btree_is_bno(bs->cur->bc_ops)) bs->cur = NULL; xchk_xref_is_only_owned_by(bs->sc, agbno, 1, bs->oinfo); - if (!bs->sc->sa.rmap_cur && btnum == XFS_BTNUM_RMAP) + if (!bs->sc->sa.rmap_cur && xfs_btree_is_rmap(bs->cur->bc_ops)) bs->cur = NULL; out_free: @@ -429,7 +428,7 @@ xchk_btree_check_owner( * up. */ if (bp == NULL) { - if (!(cur->bc_flags & XFS_BTREE_ROOT_IN_INODE)) + if (cur->bc_ops->type != XFS_BTREE_TYPE_INODE) xchk_btree_set_corrupt(bs->sc, bs->cur, level); return 0; } @@ -442,7 +441,7 @@ xchk_btree_check_owner( * duplicate cursors. Therefore, save the buffer daddr for * later scanning. */ - if (cur->bc_btnum == XFS_BTNUM_BNO || cur->bc_btnum == XFS_BTNUM_RMAP) { + if (xfs_btree_is_bno(cur->bc_ops) || xfs_btree_is_rmap(cur->bc_ops)) { struct check_owner *co; co = kmalloc(sizeof(struct check_owner), XCHK_GFP_FLAGS); @@ -475,7 +474,7 @@ xchk_btree_check_iroot_minrecs( * existing filesystems, so instead we disable the check for data fork * bmap btrees when there's an attr fork. */ - if (bs->cur->bc_btnum == XFS_BTNUM_BMAP && + if (xfs_btree_is_bmap(bs->cur->bc_ops) && bs->cur->bc_ino.whichfork == XFS_DATA_FORK && xfs_inode_has_attr_fork(bs->sc->ip)) return false; @@ -508,7 +507,7 @@ xchk_btree_check_minrecs( * child block might be less than the standard minrecs, but that's ok * provided that there's only one direct child of the root. */ - if ((cur->bc_flags & XFS_BTREE_ROOT_IN_INODE) && + if (cur->bc_ops->type == XFS_BTREE_TYPE_INODE && level == cur->bc_nlevels - 2) { struct xfs_btree_block *root_block; struct xfs_buf *root_bp; @@ -562,7 +561,7 @@ xchk_btree_block_check_keys( return; } - if (!(cur->bc_flags & XFS_BTREE_OVERLAPPING)) + if (!(cur->bc_ops->geom_flags & XFS_BTGEO_OVERLAPPING)) return; /* Make sure the high key of this block matches the parent. */ @@ -585,7 +584,6 @@ xchk_btree_get_block( struct xfs_btree_block **pblock, struct xfs_buf **pbp) { - xfs_failaddr_t failed_at; int error; *pblock = NULL; @@ -597,13 +595,7 @@ xchk_btree_get_block( return error; xfs_btree_get_block(bs->cur, level, pbp); - if (bs->cur->bc_flags & XFS_BTREE_LONG_PTRS) - failed_at = __xfs_btree_check_lblock(bs->cur, *pblock, - level, *pbp); - else - failed_at = __xfs_btree_check_sblock(bs->cur, *pblock, - level, *pbp); - if (failed_at) { + if (__xfs_btree_check_block(bs->cur, *pblock, level, *pbp)) { xchk_btree_set_corrupt(bs->sc, bs->cur, level); return 0; } @@ -664,7 +656,7 @@ xchk_btree_block_keys( if (xfs_btree_keycmp_ne(cur, &block_keys, parent_keys)) xchk_btree_set_corrupt(bs->sc, cur, 1); - if (!(cur->bc_flags & XFS_BTREE_OVERLAPPING)) + if (!(cur->bc_ops->geom_flags & XFS_BTGEO_OVERLAPPING)) return; /* Get high keys */ @@ -728,7 +720,7 @@ xchk_btree( * error codes for us. */ level = cur->bc_nlevels - 1; - cur->bc_ops->init_ptr_from_cur(cur, &ptr); + xfs_btree_init_ptr_from_cur(cur, &ptr); if (!xchk_btree_ptr_ok(bs, cur->bc_nlevels, &ptr)) goto out; error = xchk_btree_get_block(bs, level, &ptr, &block, &bp); diff --git a/fs/xfs/scrub/common.c b/fs/xfs/scrub/common.c index 81f2b96bb5a7..abff79a77c72 100644 --- a/fs/xfs/scrub/common.c +++ b/fs/xfs/scrub/common.c @@ -29,6 +29,8 @@ #include "xfs_attr.h" #include "xfs_reflink.h" #include "xfs_ag.h" +#include "xfs_error.h" +#include "xfs_quota.h" #include "scrub/scrub.h" #include "scrub/common.h" #include "scrub/trace.h" @@ -82,6 +84,15 @@ __xchk_process_error( sc->ip ? sc->ip : XFS_I(file_inode(sc->file)), sc->sm, *error); break; + case -ECANCELED: + /* + * ECANCELED here means that the caller set one of the scrub + * outcome flags (corrupt, xfail, xcorrupt) and wants to exit + * quickly. Set error to zero and do not continue. + */ + trace_xchk_op_error(sc, agno, bno, *error, ret_ip); + *error = 0; + break; case -EFSBADCRC: case -EFSCORRUPTED: /* Note the badness but don't abort. */ @@ -89,8 +100,7 @@ __xchk_process_error( *error = 0; fallthrough; default: - trace_xchk_op_error(sc, agno, bno, *error, - ret_ip); + trace_xchk_op_error(sc, agno, bno, *error, ret_ip); break; } return false; @@ -136,6 +146,16 @@ __xchk_fblock_process_error( /* Used to restart an op with deadlock avoidance. */ trace_xchk_deadlock_retry(sc->ip, sc->sm, *error); break; + case -ECANCELED: + /* + * ECANCELED here means that the caller set one of the scrub + * outcome flags (corrupt, xfail, xcorrupt) and wants to exit + * quickly. Set error to zero and do not continue. + */ + trace_xchk_file_op_error(sc, whichfork, offset, *error, + ret_ip); + *error = 0; + break; case -EFSBADCRC: case -EFSCORRUPTED: /* Note the badness but don't abort. */ @@ -227,6 +247,19 @@ xchk_block_set_corrupt( trace_xchk_block_error(sc, xfs_buf_daddr(bp), __return_address); } +#ifdef CONFIG_XFS_QUOTA +/* Record a corrupt quota counter. */ +void +xchk_qcheck_set_corrupt( + struct xfs_scrub *sc, + unsigned int dqtype, + xfs_dqid_t id) +{ + sc->sm->sm_flags |= XFS_SCRUB_OFLAG_CORRUPT; + trace_xchk_qcheck_error(sc, dqtype, id, __return_address); +} +#endif + /* Record a corruption while cross-referencing. */ void xchk_block_xref_set_corrupt( @@ -427,7 +460,7 @@ xchk_perag_read_headers( * Grab the AG headers for the attached perag structure and wait for pending * intents to drain. */ -static int +int xchk_perag_drain_and_lock( struct xfs_scrub *sc) { @@ -555,46 +588,50 @@ xchk_ag_btcur_init( { struct xfs_mount *mp = sc->mp; - if (sa->agf_bp && - xchk_ag_btree_healthy_enough(sc, sa->pag, XFS_BTNUM_BNO)) { + if (sa->agf_bp) { /* Set up a bnobt cursor for cross-referencing. */ - sa->bno_cur = xfs_allocbt_init_cursor(mp, sc->tp, sa->agf_bp, - sa->pag, XFS_BTNUM_BNO); - } + sa->bno_cur = xfs_bnobt_init_cursor(mp, sc->tp, sa->agf_bp, + sa->pag); + xchk_ag_btree_del_cursor_if_sick(sc, &sa->bno_cur, + XFS_SCRUB_TYPE_BNOBT); - if (sa->agf_bp && - xchk_ag_btree_healthy_enough(sc, sa->pag, XFS_BTNUM_CNT)) { /* Set up a cntbt cursor for cross-referencing. */ - sa->cnt_cur = xfs_allocbt_init_cursor(mp, sc->tp, sa->agf_bp, - sa->pag, XFS_BTNUM_CNT); - } - - /* Set up a inobt cursor for cross-referencing. */ - if (sa->agi_bp && - xchk_ag_btree_healthy_enough(sc, sa->pag, XFS_BTNUM_INO)) { - sa->ino_cur = xfs_inobt_init_cursor(sa->pag, sc->tp, sa->agi_bp, - XFS_BTNUM_INO); - } - - /* Set up a finobt cursor for cross-referencing. */ - if (sa->agi_bp && xfs_has_finobt(mp) && - xchk_ag_btree_healthy_enough(sc, sa->pag, XFS_BTNUM_FINO)) { - sa->fino_cur = xfs_inobt_init_cursor(sa->pag, sc->tp, sa->agi_bp, - XFS_BTNUM_FINO); - } - - /* Set up a rmapbt cursor for cross-referencing. */ - if (sa->agf_bp && xfs_has_rmapbt(mp) && - xchk_ag_btree_healthy_enough(sc, sa->pag, XFS_BTNUM_RMAP)) { - sa->rmap_cur = xfs_rmapbt_init_cursor(mp, sc->tp, sa->agf_bp, + sa->cnt_cur = xfs_cntbt_init_cursor(mp, sc->tp, sa->agf_bp, sa->pag); + xchk_ag_btree_del_cursor_if_sick(sc, &sa->cnt_cur, + XFS_SCRUB_TYPE_CNTBT); + + /* Set up a rmapbt cursor for cross-referencing. */ + if (xfs_has_rmapbt(mp)) { + sa->rmap_cur = xfs_rmapbt_init_cursor(mp, sc->tp, + sa->agf_bp, sa->pag); + xchk_ag_btree_del_cursor_if_sick(sc, &sa->rmap_cur, + XFS_SCRUB_TYPE_RMAPBT); + } + + /* Set up a refcountbt cursor for cross-referencing. */ + if (xfs_has_reflink(mp)) { + sa->refc_cur = xfs_refcountbt_init_cursor(mp, sc->tp, + sa->agf_bp, sa->pag); + xchk_ag_btree_del_cursor_if_sick(sc, &sa->refc_cur, + XFS_SCRUB_TYPE_REFCNTBT); + } } - /* Set up a refcountbt cursor for cross-referencing. */ - if (sa->agf_bp && xfs_has_reflink(mp) && - xchk_ag_btree_healthy_enough(sc, sa->pag, XFS_BTNUM_REFC)) { - sa->refc_cur = xfs_refcountbt_init_cursor(mp, sc->tp, - sa->agf_bp, sa->pag); + if (sa->agi_bp) { + /* Set up a inobt cursor for cross-referencing. */ + sa->ino_cur = xfs_inobt_init_cursor(sa->pag, sc->tp, + sa->agi_bp); + xchk_ag_btree_del_cursor_if_sick(sc, &sa->ino_cur, + XFS_SCRUB_TYPE_INOBT); + + /* Set up a finobt cursor for cross-referencing. */ + if (xfs_has_finobt(mp)) { + sa->fino_cur = xfs_finobt_init_cursor(sa->pag, sc->tp, + sa->agi_bp); + xchk_ag_btree_del_cursor_if_sick(sc, &sa->fino_cur, + XFS_SCRUB_TYPE_FINOBT); + } } } @@ -653,6 +690,13 @@ xchk_trans_cancel( sc->tp = NULL; } +int +xchk_trans_alloc_empty( + struct xfs_scrub *sc) +{ + return xfs_trans_alloc_empty(sc->mp, &sc->tp); +} + /* * Grab an empty transaction so that we can re-grab locked buffers if * one of our btrees turns out to be cyclic. @@ -672,7 +716,7 @@ xchk_trans_alloc( return xfs_trans_alloc(sc->mp, &M_RES(sc->mp)->tr_itruncate, resblks, 0, 0, &sc->tp); - return xfs_trans_alloc_empty(sc->mp, &sc->tp); + return xchk_trans_alloc_empty(sc); } /* Set us up with a transaction and an empty context. */ @@ -1259,6 +1303,15 @@ xchk_fsgates_enable( if (scrub_fsgates & XCHK_FSGATES_DRAIN) xfs_drain_wait_enable(); + if (scrub_fsgates & XCHK_FSGATES_QUOTA) + xfs_dqtrx_hook_enable(); + + if (scrub_fsgates & XCHK_FSGATES_DIRENTS) + xfs_dir_hook_enable(); + + if (scrub_fsgates & XCHK_FSGATES_RMAP) + xfs_rmap_hook_enable(); + sc->flags |= scrub_fsgates; } diff --git a/fs/xfs/scrub/common.h b/fs/xfs/scrub/common.h index da09580b454a..89f7bbec887e 100644 --- a/fs/xfs/scrub/common.h +++ b/fs/xfs/scrub/common.h @@ -32,6 +32,7 @@ xchk_should_terminate( } int xchk_trans_alloc(struct xfs_scrub *sc, uint resblks); +int xchk_trans_alloc_empty(struct xfs_scrub *sc); void xchk_trans_cancel(struct xfs_scrub *sc); bool xchk_process_error(struct xfs_scrub *sc, xfs_agnumber_t agno, @@ -54,6 +55,10 @@ void xchk_block_set_corrupt(struct xfs_scrub *sc, void xchk_ino_set_corrupt(struct xfs_scrub *sc, xfs_ino_t ino); void xchk_fblock_set_corrupt(struct xfs_scrub *sc, int whichfork, xfs_fileoff_t offset); +#ifdef CONFIG_XFS_QUOTA +void xchk_qcheck_set_corrupt(struct xfs_scrub *sc, unsigned int dqtype, + xfs_dqid_t id); +#endif void xchk_block_xref_set_corrupt(struct xfs_scrub *sc, struct xfs_buf *bp); @@ -105,6 +110,7 @@ xchk_setup_rtsummary(struct xfs_scrub *sc) #ifdef CONFIG_XFS_QUOTA int xchk_ino_dqattach(struct xfs_scrub *sc); int xchk_setup_quota(struct xfs_scrub *sc); +int xchk_setup_quotacheck(struct xfs_scrub *sc); #else static inline int xchk_ino_dqattach(struct xfs_scrub *sc) @@ -116,12 +122,19 @@ xchk_setup_quota(struct xfs_scrub *sc) { return -ENOENT; } +static inline int +xchk_setup_quotacheck(struct xfs_scrub *sc) +{ + return -ENOENT; +} #endif int xchk_setup_fscounters(struct xfs_scrub *sc); +int xchk_setup_nlinks(struct xfs_scrub *sc); void xchk_ag_free(struct xfs_scrub *sc, struct xchk_ag *sa); int xchk_ag_init(struct xfs_scrub *sc, xfs_agnumber_t agno, struct xchk_ag *sa); +int xchk_perag_drain_and_lock(struct xfs_scrub *sc); /* * Grab all AG resources, treating the inability to grab the perag structure as diff --git a/fs/xfs/scrub/cow_repair.c b/fs/xfs/scrub/cow_repair.c index 1e82c727af8e..4de3f0f40f48 100644 --- a/fs/xfs/scrub/cow_repair.c +++ b/fs/xfs/scrub/cow_repair.c @@ -609,6 +609,6 @@ xrep_bmap_cow( out_bitmap: xfsb_bitmap_destroy(&xc->old_cowfork_fsblocks); xoff_bitmap_destroy(&xc->bad_fileoffs); - kmem_free(xc); + kfree(xc); return error; } diff --git a/fs/xfs/scrub/dir.c b/fs/xfs/scrub/dir.c index d86ab51af928..076a310b8eb0 100644 --- a/fs/xfs/scrub/dir.c +++ b/fs/xfs/scrub/dir.c @@ -93,11 +93,11 @@ xchk_dir_actor( return -ECANCELED; } - if (!strncmp(".", name->name, name->len)) { + if (xfs_dir2_samename(name, &xfs_name_dot)) { /* If this is "." then check that the inum matches the dir. */ if (ino != dp->i_ino) xchk_fblock_set_corrupt(sc, XFS_DATA_FORK, offset); - } else if (!strncmp("..", name->name, name->len)) { + } else if (xfs_dir2_samename(name, &xfs_name_dotdot)) { /* * If this is ".." in the root inode, check that the inum * matches this dir. diff --git a/fs/xfs/scrub/fscounters.c b/fs/xfs/scrub/fscounters.c index 5799e9a94f1f..d310737c8823 100644 --- a/fs/xfs/scrub/fscounters.c +++ b/fs/xfs/scrub/fscounters.c @@ -22,6 +22,7 @@ #include "scrub/scrub.h" #include "scrub/common.h" #include "scrub/trace.h" +#include "scrub/fscounters.h" /* * FS Summary Counters @@ -48,17 +49,6 @@ * our tolerance for mismatch between expected and actual counter values. */ -struct xchk_fscounters { - struct xfs_scrub *sc; - uint64_t icount; - uint64_t ifree; - uint64_t fdblocks; - uint64_t frextents; - unsigned long long icount_min; - unsigned long long icount_max; - bool frozen; -}; - /* * Since the expected value computation is lockless but only browses incore * values, the percpu counters should be fairly close to each other. However, @@ -235,14 +225,19 @@ xchk_setup_fscounters( * Pause all writer activity in the filesystem while we're scrubbing to * reduce the likelihood of background perturbations to the counters * throwing off our calculations. + * + * If we're repairing, we need to prevent any other thread from + * changing the global fs summary counters while we're repairing them. + * This requires the fs to be frozen, which will disable background + * reclaim and purge all inactive inodes. */ - if (sc->flags & XCHK_TRY_HARDER) { + if ((sc->flags & XCHK_TRY_HARDER) || xchk_could_repair(sc)) { error = xchk_fscounters_freeze(sc); if (error) return error; } - return xfs_trans_alloc_empty(sc->mp, &sc->tp); + return xchk_trans_alloc_empty(sc); } /* @@ -254,7 +249,9 @@ xchk_setup_fscounters( * set the INCOMPLETE flag even when a negative errno is returned. This care * must be taken with certain errno values (i.e. EFSBADCRC, EFSCORRUPTED, * ECANCELED) that are absorbed into a scrub state flag update by - * xchk_*_process_error. + * xchk_*_process_error. Scrub and repair share the same incore data + * structures, so the INCOMPLETE flag is critical to prevent a repair based on + * insufficient information. */ /* Count free space btree blocks manually for pre-lazysbcount filesystems. */ @@ -482,6 +479,10 @@ xchk_fscount_within_range( if (curr_value == expected) return true; + /* We require exact matches when repair is running. */ + if (sc->sm->sm_flags & XFS_SCRUB_IFLAG_REPAIR) + return false; + min_value = min(old_value, curr_value); max_value = max(old_value, curr_value); diff --git a/fs/xfs/scrub/fscounters.h b/fs/xfs/scrub/fscounters.h new file mode 100644 index 000000000000..461a13d25f4b --- /dev/null +++ b/fs/xfs/scrub/fscounters.h @@ -0,0 +1,20 @@ +/* SPDX-License-Identifier: GPL-2.0-or-later */ +/* + * Copyright (c) 2021-2024 Oracle. All Rights Reserved. + * Author: Darrick J. Wong <djwong@kernel.org> + */ +#ifndef __XFS_SCRUB_FSCOUNTERS_H__ +#define __XFS_SCRUB_FSCOUNTERS_H__ + +struct xchk_fscounters { + struct xfs_scrub *sc; + uint64_t icount; + uint64_t ifree; + uint64_t fdblocks; + uint64_t frextents; + unsigned long long icount_min; + unsigned long long icount_max; + bool frozen; +}; + +#endif /* __XFS_SCRUB_FSCOUNTERS_H__ */ diff --git a/fs/xfs/scrub/fscounters_repair.c b/fs/xfs/scrub/fscounters_repair.c new file mode 100644 index 000000000000..94cdb852bee4 --- /dev/null +++ b/fs/xfs/scrub/fscounters_repair.c @@ -0,0 +1,72 @@ +// SPDX-License-Identifier: GPL-2.0-or-later +/* + * Copyright (c) 2018-2024 Oracle. All Rights Reserved. + * Author: Darrick J. Wong <djwong@kernel.org> + */ +#include "xfs.h" +#include "xfs_fs.h" +#include "xfs_shared.h" +#include "xfs_format.h" +#include "xfs_trans_resv.h" +#include "xfs_mount.h" +#include "xfs_defer.h" +#include "xfs_btree.h" +#include "xfs_bit.h" +#include "xfs_log_format.h" +#include "xfs_trans.h" +#include "xfs_sb.h" +#include "xfs_inode.h" +#include "xfs_alloc.h" +#include "xfs_ialloc.h" +#include "xfs_rmap.h" +#include "xfs_health.h" +#include "scrub/xfs_scrub.h" +#include "scrub/scrub.h" +#include "scrub/common.h" +#include "scrub/trace.h" +#include "scrub/repair.h" +#include "scrub/fscounters.h" + +/* + * FS Summary Counters + * =================== + * + * We correct errors in the filesystem summary counters by setting them to the + * values computed during the obligatory scrub phase. However, we must be + * careful not to allow any other thread to change the counters while we're + * computing and setting new values. To achieve this, we freeze the + * filesystem for the whole operation if the REPAIR flag is set. The checking + * function is stricter when we've frozen the fs. + */ + +/* + * Reset the superblock counters. Caller is responsible for freezing the + * filesystem during the calculation and reset phases. + */ +int +xrep_fscounters( + struct xfs_scrub *sc) +{ + struct xfs_mount *mp = sc->mp; + struct xchk_fscounters *fsc = sc->buf; + + /* + * Reinitialize the in-core counters from what we computed. We froze + * the filesystem, so there shouldn't be anyone else trying to modify + * these counters. + */ + if (!fsc->frozen) { + ASSERT(fsc->frozen); + return -EFSCORRUPTED; + } + + trace_xrep_reset_counters(mp, fsc); + + percpu_counter_set(&mp->m_icount, fsc->icount); + percpu_counter_set(&mp->m_ifree, fsc->ifree); + percpu_counter_set(&mp->m_fdblocks, fsc->fdblocks); + percpu_counter_set(&mp->m_frextents, fsc->frextents); + mp->m_sb.sb_frextents = fsc->frextents; + + return 0; +} diff --git a/fs/xfs/scrub/health.c b/fs/xfs/scrub/health.c index 531006910ca9..9020a6bef7f1 100644 --- a/fs/xfs/scrub/health.c +++ b/fs/xfs/scrub/health.c @@ -14,6 +14,7 @@ #include "xfs_health.h" #include "scrub/scrub.h" #include "scrub/health.h" +#include "scrub/common.h" /* * Scrub and In-Core Filesystem Health Assessments @@ -105,6 +106,8 @@ static const struct xchk_health_map type_to_health_flag[XFS_SCRUB_TYPE_NR] = { [XFS_SCRUB_TYPE_GQUOTA] = { XHG_FS, XFS_SICK_FS_GQUOTA }, [XFS_SCRUB_TYPE_PQUOTA] = { XHG_FS, XFS_SICK_FS_PQUOTA }, [XFS_SCRUB_TYPE_FSCOUNTERS] = { XHG_FS, XFS_SICK_FS_COUNTERS }, + [XFS_SCRUB_TYPE_QUOTACHECK] = { XHG_FS, XFS_SICK_FS_QUOTACHECK }, + [XFS_SCRUB_TYPE_NLINKS] = { XHG_FS, XFS_SICK_FS_NLINKS }, }; /* Return the health status mask for this scrub type. */ @@ -148,6 +151,24 @@ xchk_file_looks_zapped( } /* + * Scrub gave the filesystem a clean bill of health, so clear all the indirect + * markers of past problems (at least for the fs and ags) so that we can be + * healthy again. + */ +STATIC void +xchk_mark_all_healthy( + struct xfs_mount *mp) +{ + struct xfs_perag *pag; + xfs_agnumber_t agno; + + xfs_fs_mark_healthy(mp, XFS_SICK_FS_INDIRECT); + xfs_rt_mark_healthy(mp, XFS_SICK_RT_INDIRECT); + for_each_perag(mp, agno, pag) + xfs_ag_mark_healthy(pag, XFS_SICK_AG_INDIRECT); +} + +/* * Update filesystem health assessments based on what we found and did. * * If the scrubber finds errors, we mark sick whatever's mentioned in @@ -164,6 +185,18 @@ xchk_update_health( struct xfs_perag *pag; bool bad; + /* + * The HEALTHY scrub type is a request from userspace to clear all the + * indirect flags after a clean scan of the entire filesystem. As such + * there's no sick flag defined for it, so we branch here ahead of the + * mask check. + */ + if (sc->sm->sm_type == XFS_SCRUB_TYPE_HEALTHY && + !(sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT)) { + xchk_mark_all_healthy(sc->mp); + return; + } + if (!sc->sick_mask) return; @@ -173,7 +206,7 @@ xchk_update_health( case XHG_AG: pag = xfs_perag_get(sc->mp, sc->sm->sm_agno); if (bad) - xfs_ag_mark_sick(pag, sc->sick_mask); + xfs_ag_mark_corrupt(pag, sc->sick_mask); else xfs_ag_mark_healthy(pag, sc->sick_mask); xfs_perag_put(pag); @@ -181,20 +214,30 @@ xchk_update_health( case XHG_INO: if (!sc->ip) return; - if (bad) - xfs_inode_mark_sick(sc->ip, sc->sick_mask); - else + if (bad) { + unsigned int mask = sc->sick_mask; + + /* + * If we're coming in for repairs then we don't want + * sickness flags to propagate to the incore health + * status if the inode gets inactivated before we can + * fix it. + */ + if (sc->sm->sm_flags & XFS_SCRUB_IFLAG_REPAIR) + mask |= XFS_SICK_INO_FORGET; + xfs_inode_mark_corrupt(sc->ip, mask); + } else xfs_inode_mark_healthy(sc->ip, sc->sick_mask); break; case XHG_FS: if (bad) - xfs_fs_mark_sick(sc->mp, sc->sick_mask); + xfs_fs_mark_corrupt(sc->mp, sc->sick_mask); else xfs_fs_mark_healthy(sc->mp, sc->sick_mask); break; case XHG_RT: if (bad) - xfs_rt_mark_sick(sc->mp, sc->sick_mask); + xfs_rt_mark_corrupt(sc->mp, sc->sick_mask); else xfs_rt_mark_healthy(sc->mp, sc->sick_mask); break; @@ -205,13 +248,13 @@ xchk_update_health( } /* Is the given per-AG btree healthy enough for scanning? */ -bool -xchk_ag_btree_healthy_enough( +void +xchk_ag_btree_del_cursor_if_sick( struct xfs_scrub *sc, - struct xfs_perag *pag, - xfs_btnum_t btnum) + struct xfs_btree_cur **curp, + unsigned int sm_type) { - unsigned int mask = 0; + unsigned int mask = (*curp)->bc_ops->sick_mask; /* * We always want the cursor if it's the same type as whatever we're @@ -220,41 +263,8 @@ xchk_ag_btree_healthy_enough( * Otherwise, we're only interested in the btree for cross-referencing. * If we know the btree is bad then don't bother, just set XFAIL. */ - switch (btnum) { - case XFS_BTNUM_BNO: - if (sc->sm->sm_type == XFS_SCRUB_TYPE_BNOBT) - return true; - mask = XFS_SICK_AG_BNOBT; - break; - case XFS_BTNUM_CNT: - if (sc->sm->sm_type == XFS_SCRUB_TYPE_CNTBT) - return true; - mask = XFS_SICK_AG_CNTBT; - break; - case XFS_BTNUM_INO: - if (sc->sm->sm_type == XFS_SCRUB_TYPE_INOBT) - return true; - mask = XFS_SICK_AG_INOBT; - break; - case XFS_BTNUM_FINO: - if (sc->sm->sm_type == XFS_SCRUB_TYPE_FINOBT) - return true; - mask = XFS_SICK_AG_FINOBT; - break; - case XFS_BTNUM_RMAP: - if (sc->sm->sm_type == XFS_SCRUB_TYPE_RMAPBT) - return true; - mask = XFS_SICK_AG_RMAPBT; - break; - case XFS_BTNUM_REFC: - if (sc->sm->sm_type == XFS_SCRUB_TYPE_REFCNTBT) - return true; - mask = XFS_SICK_AG_REFCNTBT; - break; - default: - ASSERT(0); - return true; - } + if (sc->sm->sm_type == sm_type) + return; /* * If we just repaired some AG metadata, sc->sick_mask will reflect all @@ -266,10 +276,42 @@ xchk_ag_btree_healthy_enough( type_to_health_flag[sc->sm->sm_type].group == XHG_AG) mask &= ~sc->sick_mask; - if (xfs_ag_has_sickness(pag, mask)) { + if (xfs_ag_has_sickness((*curp)->bc_ag.pag, mask)) { sc->sm->sm_flags |= XFS_SCRUB_OFLAG_XFAIL; - return false; + xfs_btree_del_cursor(*curp, XFS_BTREE_NOERROR); + *curp = NULL; + } +} + +/* + * Quick scan to double-check that there isn't any evidence of lingering + * primary health problems. If we're still clear, then the health update will + * take care of clearing the indirect evidence. + */ +int +xchk_health_record( + struct xfs_scrub *sc) +{ + struct xfs_mount *mp = sc->mp; + struct xfs_perag *pag; + xfs_agnumber_t agno; + + unsigned int sick; + unsigned int checked; + + xfs_fs_measure_sickness(mp, &sick, &checked); + if (sick & XFS_SICK_FS_PRIMARY) + xchk_set_corrupt(sc); + + xfs_rt_measure_sickness(mp, &sick, &checked); + if (sick & XFS_SICK_RT_PRIMARY) + xchk_set_corrupt(sc); + + for_each_perag(mp, agno, pag) { + xfs_ag_measure_sickness(pag, &sick, &checked); + if (sick & XFS_SICK_AG_PRIMARY) + xchk_set_corrupt(sc); } - return true; + return 0; } diff --git a/fs/xfs/scrub/health.h b/fs/xfs/scrub/health.h index a731b2467399..63fc426eb5ae 100644 --- a/fs/xfs/scrub/health.h +++ b/fs/xfs/scrub/health.h @@ -8,9 +8,10 @@ unsigned int xchk_health_mask_for_scrub_type(__u32 scrub_type); void xchk_update_health(struct xfs_scrub *sc); -bool xchk_ag_btree_healthy_enough(struct xfs_scrub *sc, struct xfs_perag *pag, - xfs_btnum_t btnum); +void xchk_ag_btree_del_cursor_if_sick(struct xfs_scrub *sc, + struct xfs_btree_cur **curp, unsigned int sm_type); void xchk_mark_healthy_if_clean(struct xfs_scrub *sc, unsigned int mask); bool xchk_file_looks_zapped(struct xfs_scrub *sc, unsigned int mask); +int xchk_health_record(struct xfs_scrub *sc); #endif /* __XFS_SCRUB_HEALTH_H__ */ diff --git a/fs/xfs/scrub/ialloc.c b/fs/xfs/scrub/ialloc.c index a720fc62262a..750d7b0cd25a 100644 --- a/fs/xfs/scrub/ialloc.c +++ b/fs/xfs/scrub/ialloc.c @@ -76,7 +76,7 @@ xchk_inobt_xref_finobt( int has_record; int error; - ASSERT(cur->bc_btnum == XFS_BTNUM_FINO); + ASSERT(xfs_btree_is_fino(cur->bc_ops)); error = xfs_inobt_lookup(cur, agino, XFS_LOOKUP_LE, &has_record); if (error) @@ -179,7 +179,7 @@ xchk_finobt_xref_inobt( int has_record; int error; - ASSERT(cur->bc_btnum == XFS_BTNUM_INO); + ASSERT(xfs_btree_is_ino(cur->bc_ops)); error = xfs_inobt_lookup(cur, agino, XFS_LOOKUP_LE, &has_record); if (error) @@ -514,7 +514,7 @@ xchk_iallocbt_rec_alignment( * Otherwise, we expect that the finobt record is aligned to the * cluster alignment as told by the superblock. */ - if (bs->cur->bc_btnum == XFS_BTNUM_FINO) { + if (xfs_btree_is_fino(bs->cur->bc_ops)) { unsigned int imask; imask = min_t(unsigned int, XFS_INODES_PER_CHUNK, @@ -649,8 +649,7 @@ out: */ STATIC void xchk_iallocbt_xref_rmap_btreeblks( - struct xfs_scrub *sc, - int which) + struct xfs_scrub *sc) { xfs_filblks_t blocks; xfs_extlen_t inobt_blocks = 0; @@ -688,7 +687,6 @@ xchk_iallocbt_xref_rmap_btreeblks( STATIC void xchk_iallocbt_xref_rmap_inodes( struct xfs_scrub *sc, - int which, unsigned long long inodes) { xfs_filblks_t blocks; @@ -719,17 +717,14 @@ xchk_iallocbt( .next_startino = NULLAGINO, .next_cluster_ino = NULLAGINO, }; - xfs_btnum_t which; int error; switch (sc->sm->sm_type) { case XFS_SCRUB_TYPE_INOBT: cur = sc->sa.ino_cur; - which = XFS_BTNUM_INO; break; case XFS_SCRUB_TYPE_FINOBT: cur = sc->sa.fino_cur; - which = XFS_BTNUM_FINO; break; default: ASSERT(0); @@ -741,7 +736,7 @@ xchk_iallocbt( if (error) return error; - xchk_iallocbt_xref_rmap_btreeblks(sc, which); + xchk_iallocbt_xref_rmap_btreeblks(sc); /* * If we're scrubbing the inode btree, inode_blocks is the number of @@ -750,9 +745,8 @@ xchk_iallocbt( * knows about. We can't do this for the finobt since it only points * to inode chunks with free inodes. */ - if (which == XFS_BTNUM_INO) - xchk_iallocbt_xref_rmap_inodes(sc, which, iabt.inodes); - + if (sc->sm->sm_type == XFS_SCRUB_TYPE_INOBT) + xchk_iallocbt_xref_rmap_inodes(sc, iabt.inodes); return error; } diff --git a/fs/xfs/scrub/ialloc_repair.c b/fs/xfs/scrub/ialloc_repair.c index b3f7182dd2f5..a00ec7ae1792 100644 --- a/fs/xfs/scrub/ialloc_repair.c +++ b/fs/xfs/scrub/ialloc_repair.c @@ -369,7 +369,7 @@ xrep_ibt_check_inode_ext( * On a sparse inode fs, this cluster could be part of a sparse chunk. * Sparse clusters must be aligned to sparse chunk alignment. */ - if (xfs_has_sparseinodes(mp) && + if (xfs_has_sparseinodes(mp) && mp->m_sb.sb_spino_align && (!IS_ALIGNED(agbno, mp->m_sb.sb_spino_align) || !IS_ALIGNED(agbno + len, mp->m_sb.sb_spino_align))) return -EFSCORRUPTED; @@ -663,8 +663,8 @@ xrep_ibt_build_new_trees( ri->new_inobt.bload.claim_block = xrep_ibt_claim_block; ri->new_inobt.bload.get_records = xrep_ibt_get_records; - ino_cur = xfs_inobt_stage_cursor(sc->sa.pag, &ri->new_inobt.afake, - XFS_BTNUM_INO); + ino_cur = xfs_inobt_init_cursor(sc->sa.pag, NULL, NULL); + xfs_btree_stage_afakeroot(ino_cur, &ri->new_inobt.afake); error = xfs_btree_bload_compute_geometry(ino_cur, &ri->new_inobt.bload, xfarray_length(ri->inode_records)); if (error) @@ -684,8 +684,8 @@ xrep_ibt_build_new_trees( ri->new_finobt.bload.claim_block = xrep_fibt_claim_block; ri->new_finobt.bload.get_records = xrep_fibt_get_records; - fino_cur = xfs_inobt_stage_cursor(sc->sa.pag, - &ri->new_finobt.afake, XFS_BTNUM_FINO); + fino_cur = xfs_finobt_init_cursor(sc->sa.pag, NULL, NULL); + xfs_btree_stage_afakeroot(fino_cur, &ri->new_finobt.afake); error = xfs_btree_bload_compute_geometry(fino_cur, &ri->new_finobt.bload, ri->finobt_recs); if (error) diff --git a/fs/xfs/scrub/inode_repair.c b/fs/xfs/scrub/inode_repair.c index 0ca62d59f84a..eab380e95ef4 100644 --- a/fs/xfs/scrub/inode_repair.c +++ b/fs/xfs/scrub/inode_repair.c @@ -37,12 +37,15 @@ #include "xfs_attr_leaf.h" #include "xfs_log_priv.h" #include "xfs_health.h" +#include "xfs_symlink_remote.h" #include "scrub/xfs_scrub.h" #include "scrub/scrub.h" #include "scrub/common.h" #include "scrub/btree.h" #include "scrub/trace.h" #include "scrub/repair.h" +#include "scrub/iscan.h" +#include "scrub/readdir.h" /* * Inode Record Repair @@ -126,6 +129,10 @@ struct xrep_inode { /* Must we remove all access from this file? */ bool zap_acls; + + /* Inode scanner to see if we can find the ftype from dirents */ + struct xchk_iscan ftype_iscan; + uint8_t alleged_ftype; }; /* @@ -227,26 +234,233 @@ xrep_dinode_header( dip->di_gen = cpu_to_be32(sc->sm->sm_gen); } -/* Turn di_mode into /something/ recognizable. */ -STATIC void +/* + * If this directory entry points to the scrub target inode, then the directory + * we're scanning is the parent of the scrub target inode. + */ +STATIC int +xrep_dinode_findmode_dirent( + struct xfs_scrub *sc, + struct xfs_inode *dp, + xfs_dir2_dataptr_t dapos, + const struct xfs_name *name, + xfs_ino_t ino, + void *priv) +{ + struct xrep_inode *ri = priv; + int error = 0; + + if (xchk_should_terminate(ri->sc, &error)) + return error; + + if (ino != sc->sm->sm_ino) + return 0; + + /* Ignore garbage directory entry names. */ + if (name->len == 0 || !xfs_dir2_namecheck(name->name, name->len)) + return -EFSCORRUPTED; + + /* Don't pick up dot or dotdot entries; we only want child dirents. */ + if (xfs_dir2_samename(name, &xfs_name_dotdot) || + xfs_dir2_samename(name, &xfs_name_dot)) + return 0; + + /* + * Uhoh, more than one parent for this inode and they don't agree on + * the file type? + */ + if (ri->alleged_ftype != XFS_DIR3_FT_UNKNOWN && + ri->alleged_ftype != name->type) { + trace_xrep_dinode_findmode_dirent_inval(ri->sc, dp, name->type, + ri->alleged_ftype); + return -EFSCORRUPTED; + } + + /* We found a potential parent; remember the ftype. */ + trace_xrep_dinode_findmode_dirent(ri->sc, dp, name->type); + ri->alleged_ftype = name->type; + return 0; +} + +/* + * If this is a directory, walk the dirents looking for any that point to the + * scrub target inode. + */ +STATIC int +xrep_dinode_findmode_walk_directory( + struct xrep_inode *ri, + struct xfs_inode *dp) +{ + struct xfs_scrub *sc = ri->sc; + unsigned int lock_mode; + int error = 0; + + /* + * Scan the directory to see if there it contains an entry pointing to + * the directory that we are repairing. + */ + lock_mode = xfs_ilock_data_map_shared(dp); + + /* + * If this directory is known to be sick, we cannot scan it reliably + * and must abort. + */ + if (xfs_inode_has_sickness(dp, XFS_SICK_INO_CORE | + XFS_SICK_INO_BMBTD | + XFS_SICK_INO_DIR)) { + error = -EFSCORRUPTED; + goto out_unlock; + } + + /* + * We cannot complete our parent pointer scan if a directory looks as + * though it has been zapped by the inode record repair code. + */ + if (xchk_dir_looks_zapped(dp)) { + error = -EBUSY; + goto out_unlock; + } + + error = xchk_dir_walk(sc, dp, xrep_dinode_findmode_dirent, ri); + if (error) + goto out_unlock; + +out_unlock: + xfs_iunlock(dp, lock_mode); + return error; +} + +/* + * Try to find the mode of the inode being repaired by looking for directories + * that point down to this file. + */ +STATIC int +xrep_dinode_find_mode( + struct xrep_inode *ri, + uint16_t *mode) +{ + struct xfs_scrub *sc = ri->sc; + struct xfs_inode *dp; + int error; + + /* No ftype means we have no other metadata to consult. */ + if (!xfs_has_ftype(sc->mp)) { + *mode = S_IFREG; + return 0; + } + + /* + * Scan all directories for parents that might point down to this + * inode. Skip the inode being repaired during the scan since it + * cannot be its own parent. Note that we still hold the AGI locked + * so there's a real possibility that _iscan_iter can return EBUSY. + */ + xchk_iscan_start(sc, 5000, 100, &ri->ftype_iscan); + ri->ftype_iscan.skip_ino = sc->sm->sm_ino; + ri->alleged_ftype = XFS_DIR3_FT_UNKNOWN; + while ((error = xchk_iscan_iter(&ri->ftype_iscan, &dp)) == 1) { + if (S_ISDIR(VFS_I(dp)->i_mode)) + error = xrep_dinode_findmode_walk_directory(ri, dp); + xchk_iscan_mark_visited(&ri->ftype_iscan, dp); + xchk_irele(sc, dp); + if (error < 0) + break; + if (xchk_should_terminate(sc, &error)) + break; + } + xchk_iscan_iter_finish(&ri->ftype_iscan); + xchk_iscan_teardown(&ri->ftype_iscan); + + if (error == -EBUSY) { + if (ri->alleged_ftype != XFS_DIR3_FT_UNKNOWN) { + /* + * If we got an EBUSY after finding at least one + * dirent, that means the scan found an inode on the + * inactivation list and could not open it. Accept the + * alleged ftype and install a new mode below. + */ + error = 0; + } else if (!(sc->flags & XCHK_TRY_HARDER)) { + /* + * Otherwise, retry the operation one time to see if + * the reason for the delay is an inode from the same + * cluster buffer waiting on the inactivation list. + */ + error = -EDEADLOCK; + } + } + if (error) + return error; + + /* + * Convert the discovered ftype into the file mode. If all else fails, + * return S_IFREG. + */ + switch (ri->alleged_ftype) { + case XFS_DIR3_FT_DIR: + *mode = S_IFDIR; + break; + case XFS_DIR3_FT_WHT: + case XFS_DIR3_FT_CHRDEV: + *mode = S_IFCHR; + break; + case XFS_DIR3_FT_BLKDEV: + *mode = S_IFBLK; + break; + case XFS_DIR3_FT_FIFO: + *mode = S_IFIFO; + break; + case XFS_DIR3_FT_SOCK: + *mode = S_IFSOCK; + break; + case XFS_DIR3_FT_SYMLINK: + *mode = S_IFLNK; + break; + default: + *mode = S_IFREG; + break; + } + return 0; +} + +/* Turn di_mode into /something/ recognizable. Returns true if we succeed. */ +STATIC int xrep_dinode_mode( struct xrep_inode *ri, struct xfs_dinode *dip) { struct xfs_scrub *sc = ri->sc; uint16_t mode = be16_to_cpu(dip->di_mode); + int error; trace_xrep_dinode_mode(sc, dip); if (mode == 0 || xfs_mode_to_ftype(mode) != XFS_DIR3_FT_UNKNOWN) - return; + return 0; + + /* Try to fix the mode. If we cannot, then leave everything alone. */ + error = xrep_dinode_find_mode(ri, &mode); + switch (error) { + case -EINTR: + case -EBUSY: + case -EDEADLOCK: + /* temporary failure or fatal signal */ + return error; + case 0: + /* found mode */ + break; + default: + /* some other error, assume S_IFREG */ + mode = S_IFREG; + break; + } /* bad mode, so we set it to a file that only root can read */ - mode = S_IFREG; dip->di_mode = cpu_to_be16(mode); dip->di_uid = 0; dip->di_gid = 0; ri->zap_acls = true; + return 0; } /* Fix any conflicting flags that the verifiers complain about. */ @@ -1107,12 +1321,15 @@ xrep_dinode_core( /* Fix everything the verifier will complain about. */ dip = xfs_buf_offset(bp, ri->imap.im_boffset); xrep_dinode_header(sc, dip); - xrep_dinode_mode(ri, dip); + iget_error = xrep_dinode_mode(ri, dip); + if (iget_error) + goto write; xrep_dinode_flags(sc, dip, ri->rt_extents > 0); xrep_dinode_size(ri, dip); xrep_dinode_extsize_hints(sc, dip); xrep_dinode_zap_forks(ri, dip); +write: /* Write out the inode. */ trace_xrep_dinode_fixed(sc, dip); xfs_dinode_calc_crc(sc->mp, dip); @@ -1128,7 +1345,8 @@ xrep_dinode_core( * accessing the inode. If iget fails, we still need to commit the * changes. */ - iget_error = xchk_iget(sc, ino, &sc->ip); + if (!iget_error) + iget_error = xchk_iget(sc, ino, &sc->ip); if (!iget_error) xchk_ilock(sc, XFS_IOLOCK_EXCL); @@ -1496,6 +1714,13 @@ xrep_inode( ASSERT(ri != NULL); error = xrep_dinode_problems(ri); + if (error == -EBUSY) { + /* + * Directory scan to recover inode mode encountered a + * busy inode, so we did not continue repairing things. + */ + return 0; + } if (error) return error; diff --git a/fs/xfs/scrub/iscan.c b/fs/xfs/scrub/iscan.c new file mode 100644 index 000000000000..ec3478bc505e --- /dev/null +++ b/fs/xfs/scrub/iscan.c @@ -0,0 +1,767 @@ +// SPDX-License-Identifier: GPL-2.0-or-later +/* + * Copyright (c) 2021-2024 Oracle. All Rights Reserved. + * Author: Darrick J. Wong <djwong@kernel.org> + */ +#include "xfs.h" +#include "xfs_fs.h" +#include "xfs_shared.h" +#include "xfs_format.h" +#include "xfs_trans_resv.h" +#include "xfs_mount.h" +#include "xfs_log_format.h" +#include "xfs_trans.h" +#include "xfs_inode.h" +#include "xfs_btree.h" +#include "xfs_ialloc.h" +#include "xfs_ialloc_btree.h" +#include "xfs_ag.h" +#include "xfs_error.h" +#include "xfs_bit.h" +#include "xfs_icache.h" +#include "scrub/scrub.h" +#include "scrub/iscan.h" +#include "scrub/common.h" +#include "scrub/trace.h" + +/* + * Live File Scan + * ============== + * + * Live file scans walk every inode in a live filesystem. This is more or + * less like a regular iwalk, except that when we're advancing the scan cursor, + * we must ensure that inodes cannot be added or deleted anywhere between the + * old cursor value and the new cursor value. If we're advancing the cursor + * by one inode, the caller must hold that inode; if we're finding the next + * inode to scan, we must grab the AGI and hold it until we've updated the + * scan cursor. + * + * Callers are expected to use this code to scan all files in the filesystem to + * construct a new metadata index of some kind. The scan races against other + * live updates, which means there must be a provision to update the new index + * when updates are made to inodes that already been scanned. The iscan lock + * can be used in live update hook code to stop the scan and protect this data + * structure. + * + * To keep the new index up to date with other metadata updates being made to + * the live filesystem, it is assumed that the caller will add hooks as needed + * to be notified when a metadata update occurs. The inode scanner must tell + * the hook code when an inode has been visited with xchk_iscan_mark_visit. + * Hook functions can use xchk_iscan_want_live_update to decide if the + * scanner's observations must be updated. + */ + +/* + * If the inobt record @rec covers @iscan->skip_ino, mark the inode free so + * that the scan ignores that inode. + */ +STATIC void +xchk_iscan_mask_skipino( + struct xchk_iscan *iscan, + struct xfs_perag *pag, + struct xfs_inobt_rec_incore *rec, + xfs_agino_t lastrecino) +{ + struct xfs_scrub *sc = iscan->sc; + struct xfs_mount *mp = sc->mp; + xfs_agnumber_t skip_agno = XFS_INO_TO_AGNO(mp, iscan->skip_ino); + xfs_agnumber_t skip_agino = XFS_INO_TO_AGINO(mp, iscan->skip_ino); + + if (pag->pag_agno != skip_agno) + return; + if (skip_agino < rec->ir_startino) + return; + if (skip_agino > lastrecino) + return; + + rec->ir_free |= xfs_inobt_maskn(skip_agino - rec->ir_startino, 1); +} + +/* + * Set *cursor to the next allocated inode after whatever it's set to now. + * If there are no more inodes in this AG, cursor is set to NULLAGINO. + */ +STATIC int +xchk_iscan_find_next( + struct xchk_iscan *iscan, + struct xfs_buf *agi_bp, + struct xfs_perag *pag, + xfs_inofree_t *allocmaskp, + xfs_agino_t *cursor, + uint8_t *nr_inodesp) +{ + struct xfs_scrub *sc = iscan->sc; + struct xfs_inobt_rec_incore rec; + struct xfs_btree_cur *cur; + struct xfs_mount *mp = sc->mp; + struct xfs_trans *tp = sc->tp; + xfs_agnumber_t agno = pag->pag_agno; + xfs_agino_t lastino = NULLAGINO; + xfs_agino_t first, last; + xfs_agino_t agino = *cursor; + int has_rec; + int error; + + /* If the cursor is beyond the end of this AG, move to the next one. */ + xfs_agino_range(mp, agno, &first, &last); + if (agino > last) { + *cursor = NULLAGINO; + return 0; + } + + /* + * Look up the inode chunk for the current cursor position. If there + * is no chunk here, we want the next one. + */ + cur = xfs_inobt_init_cursor(pag, tp, agi_bp); + error = xfs_inobt_lookup(cur, agino, XFS_LOOKUP_LE, &has_rec); + if (!error && !has_rec) + error = xfs_btree_increment(cur, 0, &has_rec); + for (; !error; error = xfs_btree_increment(cur, 0, &has_rec)) { + xfs_inofree_t allocmask; + + /* + * If we've run out of inobt records in this AG, move the + * cursor on to the next AG and exit. The caller can try + * again with the next AG. + */ + if (!has_rec) { + *cursor = NULLAGINO; + break; + } + + error = xfs_inobt_get_rec(cur, &rec, &has_rec); + if (error) + break; + if (!has_rec) { + error = -EFSCORRUPTED; + break; + } + + /* Make sure that we always move forward. */ + if (lastino != NULLAGINO && + XFS_IS_CORRUPT(mp, lastino >= rec.ir_startino)) { + error = -EFSCORRUPTED; + break; + } + lastino = rec.ir_startino + XFS_INODES_PER_CHUNK - 1; + + /* + * If this record only covers inodes that come before the + * cursor, advance to the next record. + */ + if (rec.ir_startino + XFS_INODES_PER_CHUNK <= agino) + continue; + + if (iscan->skip_ino) + xchk_iscan_mask_skipino(iscan, pag, &rec, lastino); + + /* + * If the incoming lookup put us in the middle of an inobt + * record, mark it and the previous inodes "free" so that the + * search for allocated inodes will start at the cursor. + * We don't care about ir_freecount here. + */ + if (agino >= rec.ir_startino) + rec.ir_free |= xfs_inobt_maskn(0, + agino + 1 - rec.ir_startino); + + /* + * If there are allocated inodes in this chunk, find them + * and update the scan cursor. + */ + allocmask = ~rec.ir_free; + if (hweight64(allocmask) > 0) { + int next = xfs_lowbit64(allocmask); + + ASSERT(next >= 0); + *cursor = rec.ir_startino + next; + *allocmaskp = allocmask >> next; + *nr_inodesp = XFS_INODES_PER_CHUNK - next; + break; + } + } + + xfs_btree_del_cursor(cur, error); + return error; +} + +/* + * Advance both the scan and the visited cursors. + * + * The inumber address space for a given filesystem is sparse, which means that + * the scan cursor can jump a long ways in a single iter() call. There are no + * inodes in these sparse areas, so we must move the visited cursor forward at + * the same time so that the scan user can receive live updates for inodes that + * may get created once we release the AGI buffer. + */ +static inline void +xchk_iscan_move_cursor( + struct xchk_iscan *iscan, + xfs_agnumber_t agno, + xfs_agino_t agino) +{ + struct xfs_scrub *sc = iscan->sc; + struct xfs_mount *mp = sc->mp; + xfs_ino_t cursor, visited; + + BUILD_BUG_ON(XFS_MAXINUMBER == NULLFSINO); + + /* + * Special-case ino == 0 here so that we never set visited_ino to + * NULLFSINO when wrapping around EOFS, for that will let through all + * live updates. + */ + cursor = XFS_AGINO_TO_INO(mp, agno, agino); + if (cursor == 0) + visited = XFS_MAXINUMBER; + else + visited = cursor - 1; + + mutex_lock(&iscan->lock); + iscan->cursor_ino = cursor; + iscan->__visited_ino = visited; + trace_xchk_iscan_move_cursor(iscan); + mutex_unlock(&iscan->lock); +} + +/* + * Prepare to return agno/agino to the iscan caller by moving the lastino + * cursor to the previous inode. Do this while we still hold the AGI so that + * no other threads can create or delete inodes in this AG. + */ +static inline void +xchk_iscan_finish( + struct xchk_iscan *iscan) +{ + mutex_lock(&iscan->lock); + iscan->cursor_ino = NULLFSINO; + + /* All live updates will be applied from now on */ + iscan->__visited_ino = NULLFSINO; + + mutex_unlock(&iscan->lock); +} + +/* + * Advance ino to the next inode that the inobt thinks is allocated, being + * careful to jump to the next AG if we've reached the right end of this AG's + * inode btree. Advancing ino effectively means that we've pushed the inode + * scan forward, so set the iscan cursor to (ino - 1) so that our live update + * predicates will track inode allocations in that part of the inode number + * key space once we release the AGI buffer. + * + * Returns 1 if there's a new inode to examine, 0 if we've run out of inodes, + * -ECANCELED if the live scan aborted, or the usual negative errno. + */ +STATIC int +xchk_iscan_advance( + struct xchk_iscan *iscan, + struct xfs_perag **pagp, + struct xfs_buf **agi_bpp, + xfs_inofree_t *allocmaskp, + uint8_t *nr_inodesp) +{ + struct xfs_scrub *sc = iscan->sc; + struct xfs_mount *mp = sc->mp; + struct xfs_buf *agi_bp; + struct xfs_perag *pag; + xfs_agnumber_t agno; + xfs_agino_t agino; + int ret; + + ASSERT(iscan->cursor_ino >= iscan->__visited_ino); + + do { + if (xchk_iscan_aborted(iscan)) + return -ECANCELED; + + agno = XFS_INO_TO_AGNO(mp, iscan->cursor_ino); + pag = xfs_perag_get(mp, agno); + if (!pag) + return -ECANCELED; + + ret = xfs_ialloc_read_agi(pag, sc->tp, &agi_bp); + if (ret) + goto out_pag; + + agino = XFS_INO_TO_AGINO(mp, iscan->cursor_ino); + ret = xchk_iscan_find_next(iscan, agi_bp, pag, allocmaskp, + &agino, nr_inodesp); + if (ret) + goto out_buf; + + if (agino != NULLAGINO) { + /* + * Found the next inode in this AG, so return it along + * with the AGI buffer and the perag structure to + * ensure it cannot go away. + */ + xchk_iscan_move_cursor(iscan, agno, agino); + *agi_bpp = agi_bp; + *pagp = pag; + return 1; + } + + /* + * Did not find any more inodes in this AG, move on to the next + * AG. + */ + agno = (agno + 1) % mp->m_sb.sb_agcount; + xchk_iscan_move_cursor(iscan, agno, 0); + xfs_trans_brelse(sc->tp, agi_bp); + xfs_perag_put(pag); + + trace_xchk_iscan_advance_ag(iscan); + } while (iscan->cursor_ino != iscan->scan_start_ino); + + xchk_iscan_finish(iscan); + return 0; + +out_buf: + xfs_trans_brelse(sc->tp, agi_bp); +out_pag: + xfs_perag_put(pag); + return ret; +} + +/* + * Grabbing the inode failed, so we need to back up the scan and ask the caller + * to try to _advance the scan again. Returns -EBUSY if we've run out of retry + * opportunities, -ECANCELED if the process has a fatal signal pending, or + * -EAGAIN if we should try again. + */ +STATIC int +xchk_iscan_iget_retry( + struct xchk_iscan *iscan, + bool wait) +{ + ASSERT(iscan->cursor_ino == iscan->__visited_ino + 1); + + if (!iscan->iget_timeout || + time_is_before_jiffies(iscan->__iget_deadline)) + return -EBUSY; + + if (wait) { + unsigned long relax; + + /* + * Sleep for a period of time to let the rest of the system + * catch up. If we return early, someone sent a kill signal to + * the calling process. + */ + relax = msecs_to_jiffies(iscan->iget_retry_delay); + trace_xchk_iscan_iget_retry_wait(iscan); + + if (schedule_timeout_killable(relax) || + xchk_iscan_aborted(iscan)) + return -ECANCELED; + } + + iscan->cursor_ino--; + return -EAGAIN; +} + +/* + * Grab an inode as part of an inode scan. While scanning this inode, the + * caller must ensure that no other threads can modify the inode until a call + * to xchk_iscan_visit succeeds. + * + * Returns the number of incore inodes grabbed; -EAGAIN if the caller should + * call again xchk_iscan_advance; -EBUSY if we couldn't grab an inode; + * -ECANCELED if there's a fatal signal pending; or some other negative errno. + */ +STATIC int +xchk_iscan_iget( + struct xchk_iscan *iscan, + struct xfs_perag *pag, + struct xfs_buf *agi_bp, + xfs_inofree_t allocmask, + uint8_t nr_inodes) +{ + struct xfs_scrub *sc = iscan->sc; + struct xfs_mount *mp = sc->mp; + xfs_ino_t ino = iscan->cursor_ino; + unsigned int idx = 0; + unsigned int i; + int error; + + ASSERT(iscan->__inodes[0] == NULL); + + /* Fill the first slot in the inode array. */ + error = xfs_iget(sc->mp, sc->tp, ino, XFS_IGET_NORETRY, 0, + &iscan->__inodes[idx]); + + trace_xchk_iscan_iget(iscan, error); + + if (error == -ENOENT || error == -EAGAIN) { + xfs_trans_brelse(sc->tp, agi_bp); + xfs_perag_put(pag); + + /* + * It's possible that this inode has lost all of its links but + * hasn't yet been inactivated. If we don't have a transaction + * or it's not writable, flush the inodegc workers and wait. + */ + xfs_inodegc_flush(mp); + return xchk_iscan_iget_retry(iscan, true); + } + + if (error == -EINVAL) { + xfs_trans_brelse(sc->tp, agi_bp); + xfs_perag_put(pag); + + /* + * We thought the inode was allocated, but the inode btree + * lookup failed, which means that it was freed since the last + * time we advanced the cursor. Back up and try again. This + * should never happen since still hold the AGI buffer from the + * inobt check, but we need to be careful about infinite loops. + */ + return xchk_iscan_iget_retry(iscan, false); + } + + if (error) { + xfs_trans_brelse(sc->tp, agi_bp); + xfs_perag_put(pag); + return error; + } + idx++; + ino++; + allocmask >>= 1; + + /* + * Now that we've filled the first slot in __inodes, try to fill the + * rest of the batch with consecutively ordered inodes. to reduce the + * number of _iter calls. Make a bitmap of unallocated inodes from the + * zeroes in the inuse bitmap; these inodes will not be scanned, but + * the _want_live_update predicate will pass through all live updates. + * + * If we can't iget an allocated inode, stop and return what we have. + */ + mutex_lock(&iscan->lock); + iscan->__batch_ino = ino - 1; + iscan->__skipped_inomask = 0; + mutex_unlock(&iscan->lock); + + for (i = 1; i < nr_inodes; i++, ino++, allocmask >>= 1) { + if (!(allocmask & 1)) { + ASSERT(!(iscan->__skipped_inomask & (1ULL << i))); + + mutex_lock(&iscan->lock); + iscan->cursor_ino = ino; + iscan->__skipped_inomask |= (1ULL << i); + mutex_unlock(&iscan->lock); + continue; + } + + ASSERT(iscan->__inodes[idx] == NULL); + + error = xfs_iget(sc->mp, sc->tp, ino, XFS_IGET_NORETRY, 0, + &iscan->__inodes[idx]); + if (error) + break; + + mutex_lock(&iscan->lock); + iscan->cursor_ino = ino; + mutex_unlock(&iscan->lock); + idx++; + } + + trace_xchk_iscan_iget_batch(sc->mp, iscan, nr_inodes, idx); + xfs_trans_brelse(sc->tp, agi_bp); + xfs_perag_put(pag); + return idx; +} + +/* + * Advance the visit cursor to reflect skipped inodes beyond whatever we + * scanned. + */ +STATIC void +xchk_iscan_finish_batch( + struct xchk_iscan *iscan) +{ + xfs_ino_t highest_skipped; + + mutex_lock(&iscan->lock); + + if (iscan->__batch_ino != NULLFSINO) { + highest_skipped = iscan->__batch_ino + + xfs_highbit64(iscan->__skipped_inomask); + iscan->__visited_ino = max(iscan->__visited_ino, + highest_skipped); + + trace_xchk_iscan_skip(iscan); + } + + iscan->__batch_ino = NULLFSINO; + iscan->__skipped_inomask = 0; + + mutex_unlock(&iscan->lock); +} + +/* + * Advance the inode scan cursor to the next allocated inode and return up to + * 64 consecutive allocated inodes starting with the cursor position. + */ +STATIC int +xchk_iscan_iter_batch( + struct xchk_iscan *iscan) +{ + struct xfs_scrub *sc = iscan->sc; + int ret; + + xchk_iscan_finish_batch(iscan); + + if (iscan->iget_timeout) + iscan->__iget_deadline = jiffies + + msecs_to_jiffies(iscan->iget_timeout); + + do { + struct xfs_buf *agi_bp = NULL; + struct xfs_perag *pag = NULL; + xfs_inofree_t allocmask = 0; + uint8_t nr_inodes = 0; + + ret = xchk_iscan_advance(iscan, &pag, &agi_bp, &allocmask, + &nr_inodes); + if (ret != 1) + return ret; + + if (xchk_iscan_aborted(iscan)) { + xfs_trans_brelse(sc->tp, agi_bp); + xfs_perag_put(pag); + ret = -ECANCELED; + break; + } + + ret = xchk_iscan_iget(iscan, pag, agi_bp, allocmask, nr_inodes); + } while (ret == -EAGAIN); + + return ret; +} + +/* + * Advance the inode scan cursor to the next allocated inode and return the + * incore inode structure associated with it. + * + * Returns 1 if there's a new inode to examine, 0 if we've run out of inodes, + * -ECANCELED if the live scan aborted, -EBUSY if the incore inode could not be + * grabbed, or the usual negative errno. + * + * If the function returns -EBUSY and the caller can handle skipping an inode, + * it may call this function again to continue the scan with the next allocated + * inode. + */ +int +xchk_iscan_iter( + struct xchk_iscan *iscan, + struct xfs_inode **ipp) +{ + unsigned int i; + int error; + + /* Find a cached inode, or go get another batch. */ + for (i = 0; i < XFS_INODES_PER_CHUNK; i++) { + if (iscan->__inodes[i]) + goto foundit; + } + + error = xchk_iscan_iter_batch(iscan); + if (error <= 0) + return error; + + ASSERT(iscan->__inodes[0] != NULL); + i = 0; + +foundit: + /* Give the caller our reference. */ + *ipp = iscan->__inodes[i]; + iscan->__inodes[i] = NULL; + return 1; +} + +/* Clean up an xfs_iscan_iter call by dropping any inodes that we still hold. */ +void +xchk_iscan_iter_finish( + struct xchk_iscan *iscan) +{ + struct xfs_scrub *sc = iscan->sc; + unsigned int i; + + for (i = 0; i < XFS_INODES_PER_CHUNK; i++) { + if (iscan->__inodes[i]) { + xchk_irele(sc, iscan->__inodes[i]); + iscan->__inodes[i] = NULL; + } + } +} + +/* Mark this inode scan finished and release resources. */ +void +xchk_iscan_teardown( + struct xchk_iscan *iscan) +{ + xchk_iscan_iter_finish(iscan); + xchk_iscan_finish(iscan); + mutex_destroy(&iscan->lock); +} + +/* Pick an AG from which to start a scan. */ +static inline xfs_ino_t +xchk_iscan_rotor( + struct xfs_mount *mp) +{ + static atomic_t agi_rotor; + unsigned int r = atomic_inc_return(&agi_rotor) - 1; + + /* + * Rotoring *backwards* through the AGs, so we add one here before + * subtracting from the agcount to arrive at an AG number. + */ + r = (r % mp->m_sb.sb_agcount) + 1; + + return XFS_AGINO_TO_INO(mp, mp->m_sb.sb_agcount - r, 0); +} + +/* + * Set ourselves up to start an inode scan. If the @iget_timeout and + * @iget_retry_delay parameters are set, the scan will try to iget each inode + * for @iget_timeout milliseconds. If an iget call indicates that the inode is + * waiting to be inactivated, the CPU will relax for @iget_retry_delay + * milliseconds after pushing the inactivation workers. + */ +void +xchk_iscan_start( + struct xfs_scrub *sc, + unsigned int iget_timeout, + unsigned int iget_retry_delay, + struct xchk_iscan *iscan) +{ + xfs_ino_t start_ino; + + start_ino = xchk_iscan_rotor(sc->mp); + + iscan->__batch_ino = NULLFSINO; + iscan->__skipped_inomask = 0; + + iscan->sc = sc; + clear_bit(XCHK_ISCAN_OPSTATE_ABORTED, &iscan->__opstate); + iscan->iget_timeout = iget_timeout; + iscan->iget_retry_delay = iget_retry_delay; + iscan->__visited_ino = start_ino; + iscan->cursor_ino = start_ino; + iscan->scan_start_ino = start_ino; + mutex_init(&iscan->lock); + memset(iscan->__inodes, 0, sizeof(iscan->__inodes)); + + trace_xchk_iscan_start(iscan, start_ino); +} + +/* + * Mark this inode as having been visited. Callers must hold a sufficiently + * exclusive lock on the inode to prevent concurrent modifications. + */ +void +xchk_iscan_mark_visited( + struct xchk_iscan *iscan, + struct xfs_inode *ip) +{ + mutex_lock(&iscan->lock); + iscan->__visited_ino = ip->i_ino; + trace_xchk_iscan_visit(iscan); + mutex_unlock(&iscan->lock); +} + +/* + * Did we skip this inode because it wasn't allocated when we loaded the batch? + * If so, it is newly allocated and will not be scanned. All live updates to + * this inode must be passed to the caller to maintain scan correctness. + */ +static inline bool +xchk_iscan_skipped( + const struct xchk_iscan *iscan, + xfs_ino_t ino) +{ + if (iscan->__batch_ino == NULLFSINO) + return false; + if (ino < iscan->__batch_ino) + return false; + if (ino >= iscan->__batch_ino + XFS_INODES_PER_CHUNK) + return false; + + return iscan->__skipped_inomask & (1ULL << (ino - iscan->__batch_ino)); +} + +/* + * Do we need a live update for this inode? This is true if the scanner thread + * has visited this inode and the scan hasn't been aborted due to errors. + * Callers must hold a sufficiently exclusive lock on the inode to prevent + * scanners from reading any inode metadata. + */ +bool +xchk_iscan_want_live_update( + struct xchk_iscan *iscan, + xfs_ino_t ino) +{ + bool ret = false; + + if (xchk_iscan_aborted(iscan)) + return false; + + mutex_lock(&iscan->lock); + + trace_xchk_iscan_want_live_update(iscan, ino); + + /* Scan is finished, caller should receive all updates. */ + if (iscan->__visited_ino == NULLFSINO) { + ret = true; + goto unlock; + } + + /* + * No inodes have been visited yet, so the visited cursor points at the + * start of the scan range. The caller should not receive any updates. + */ + if (iscan->scan_start_ino == iscan->__visited_ino) { + ret = false; + goto unlock; + } + + /* + * This inode was not allocated at the time of the iscan batch. + * The caller should receive all updates. + */ + if (xchk_iscan_skipped(iscan, ino)) { + ret = true; + goto unlock; + } + + /* + * The visited cursor hasn't yet wrapped around the end of the FS. If + * @ino is inside the starred range, the caller should receive updates: + * + * 0 ------------ S ************ V ------------ EOFS + */ + if (iscan->scan_start_ino <= iscan->__visited_ino) { + if (ino >= iscan->scan_start_ino && + ino <= iscan->__visited_ino) + ret = true; + + goto unlock; + } + + /* + * The visited cursor wrapped around the end of the FS. If @ino is + * inside the starred range, the caller should receive updates: + * + * 0 ************ V ------------ S ************ EOFS + */ + if (ino >= iscan->scan_start_ino || ino <= iscan->__visited_ino) + ret = true; + +unlock: + mutex_unlock(&iscan->lock); + return ret; +} diff --git a/fs/xfs/scrub/iscan.h b/fs/xfs/scrub/iscan.h new file mode 100644 index 000000000000..71f657552dfa --- /dev/null +++ b/fs/xfs/scrub/iscan.h @@ -0,0 +1,84 @@ +/* SPDX-License-Identifier: GPL-2.0-or-later */ +/* + * Copyright (c) 2021-2024 Oracle. All Rights Reserved. + * Author: Darrick J. Wong <djwong@kernel.org> + */ +#ifndef __XFS_SCRUB_ISCAN_H__ +#define __XFS_SCRUB_ISCAN_H__ + +struct xchk_iscan { + struct xfs_scrub *sc; + + /* Lock to protect the scan cursor. */ + struct mutex lock; + + /* + * This is the first inode in the inumber address space that we + * examined. When the scan wraps around back to here, the scan is + * finished. + */ + xfs_ino_t scan_start_ino; + + /* This is the inode that will be examined next. */ + xfs_ino_t cursor_ino; + + /* If nonzero and non-NULL, skip this inode when scanning. */ + xfs_ino_t skip_ino; + + /* + * This is the last inode that we've successfully scanned, either + * because the caller scanned it, or we moved the cursor past an empty + * part of the inode address space. Scan callers should only use the + * xchk_iscan_visit function to modify this. + */ + xfs_ino_t __visited_ino; + + /* Operational state of the livescan. */ + unsigned long __opstate; + + /* Give up on iterating @cursor_ino if we can't iget it by this time. */ + unsigned long __iget_deadline; + + /* Amount of time (in ms) that we will try to iget an inode. */ + unsigned int iget_timeout; + + /* Wait this many ms to retry an iget. */ + unsigned int iget_retry_delay; + + /* + * The scan grabs batches of inodes and stashes them here before + * handing them out with _iter. Unallocated inodes are set in the + * mask so that all updates to that inode are selected for live + * update propagation. + */ + xfs_ino_t __batch_ino; + xfs_inofree_t __skipped_inomask; + struct xfs_inode *__inodes[XFS_INODES_PER_CHUNK]; +}; + +/* Set if the scan has been aborted due to some event in the fs. */ +#define XCHK_ISCAN_OPSTATE_ABORTED (1) + +static inline bool +xchk_iscan_aborted(const struct xchk_iscan *iscan) +{ + return test_bit(XCHK_ISCAN_OPSTATE_ABORTED, &iscan->__opstate); +} + +static inline void +xchk_iscan_abort(struct xchk_iscan *iscan) +{ + set_bit(XCHK_ISCAN_OPSTATE_ABORTED, &iscan->__opstate); +} + +void xchk_iscan_start(struct xfs_scrub *sc, unsigned int iget_timeout, + unsigned int iget_retry_delay, struct xchk_iscan *iscan); +void xchk_iscan_teardown(struct xchk_iscan *iscan); + +int xchk_iscan_iter(struct xchk_iscan *iscan, struct xfs_inode **ipp); +void xchk_iscan_iter_finish(struct xchk_iscan *iscan); + +void xchk_iscan_mark_visited(struct xchk_iscan *iscan, struct xfs_inode *ip); +bool xchk_iscan_want_live_update(struct xchk_iscan *iscan, xfs_ino_t ino); + +#endif /* __XFS_SCRUB_ISCAN_H__ */ diff --git a/fs/xfs/scrub/newbt.c b/fs/xfs/scrub/newbt.c index bb6d980b4fcd..4a0271123d94 100644 --- a/fs/xfs/scrub/newbt.c +++ b/fs/xfs/scrub/newbt.c @@ -239,7 +239,11 @@ xrep_newbt_alloc_ag_blocks( xrep_newbt_validate_ag_alloc_hint(xnr); - error = xfs_alloc_vextent_near_bno(&args, xnr->alloc_hint); + if (xnr->alloc_vextent) + error = xnr->alloc_vextent(sc, &args, xnr->alloc_hint); + else + error = xfs_alloc_vextent_near_bno(&args, + xnr->alloc_hint); if (error) return error; if (args.fsbno == NULLFSBLOCK) @@ -309,7 +313,11 @@ xrep_newbt_alloc_file_blocks( xrep_newbt_validate_file_alloc_hint(xnr); - error = xfs_alloc_vextent_start_ag(&args, xnr->alloc_hint); + if (xnr->alloc_vextent) + error = xnr->alloc_vextent(sc, &args, xnr->alloc_hint); + else + error = xfs_alloc_vextent_start_ag(&args, + xnr->alloc_hint); if (error) return error; if (args.fsbno == NULLFSBLOCK) @@ -535,7 +543,7 @@ xrep_newbt_claim_block( trace_xrep_newbt_claim_block(mp, resv->pag->pag_agno, agbno, 1, xnr->oinfo.oi_owner); - if (cur->bc_flags & XFS_BTREE_LONG_PTRS) + if (cur->bc_ops->ptr_len == XFS_BTREE_LONG_PTR_LEN) ptr->l = cpu_to_be64(XFS_AGB_TO_FSB(mp, resv->pag->pag_agno, agbno)); else diff --git a/fs/xfs/scrub/newbt.h b/fs/xfs/scrub/newbt.h index 89f8e3970b1f..3d804d31af24 100644 --- a/fs/xfs/scrub/newbt.h +++ b/fs/xfs/scrub/newbt.h @@ -6,6 +6,8 @@ #ifndef __XFS_SCRUB_NEWBT_H__ #define __XFS_SCRUB_NEWBT_H__ +struct xfs_alloc_arg; + struct xrep_newbt_resv { /* Link to list of extents that we've reserved. */ struct list_head list; @@ -28,6 +30,11 @@ struct xrep_newbt_resv { struct xrep_newbt { struct xfs_scrub *sc; + /* Custom allocation function, or NULL for xfs_alloc_vextent */ + int (*alloc_vextent)(struct xfs_scrub *sc, + struct xfs_alloc_arg *args, + xfs_fsblock_t alloc_hint); + /* List of extents that we've reserved. */ struct list_head resv_list; diff --git a/fs/xfs/scrub/nlinks.c b/fs/xfs/scrub/nlinks.c new file mode 100644 index 000000000000..8a7d9557897c --- /dev/null +++ b/fs/xfs/scrub/nlinks.c @@ -0,0 +1,930 @@ +// SPDX-License-Identifier: GPL-2.0-or-later +/* + * Copyright (c) 2021-2024 Oracle. All Rights Reserved. + * Author: Darrick J. Wong <djwong@kernel.org> + */ +#include "xfs.h" +#include "xfs_fs.h" +#include "xfs_shared.h" +#include "xfs_format.h" +#include "xfs_trans_resv.h" +#include "xfs_mount.h" +#include "xfs_log_format.h" +#include "xfs_trans.h" +#include "xfs_inode.h" +#include "xfs_icache.h" +#include "xfs_iwalk.h" +#include "xfs_ialloc.h" +#include "xfs_dir2.h" +#include "xfs_dir2_priv.h" +#include "xfs_ag.h" +#include "scrub/scrub.h" +#include "scrub/common.h" +#include "scrub/repair.h" +#include "scrub/xfile.h" +#include "scrub/xfarray.h" +#include "scrub/iscan.h" +#include "scrub/nlinks.h" +#include "scrub/trace.h" +#include "scrub/readdir.h" + +/* + * Live Inode Link Count Checking + * ============================== + * + * Inode link counts are "summary" metadata, in the sense that they are + * computed as the number of directory entries referencing each file on the + * filesystem. Therefore, we compute the correct link counts by creating a + * shadow link count structure and walking every inode. + */ + +/* Set us up to scrub inode link counts. */ +int +xchk_setup_nlinks( + struct xfs_scrub *sc) +{ + xchk_fsgates_enable(sc, XCHK_FSGATES_DIRENTS); + + sc->buf = kzalloc(sizeof(struct xchk_nlink_ctrs), XCHK_GFP_FLAGS); + if (!sc->buf) + return -ENOMEM; + + return xchk_setup_fs(sc); +} + +/* + * Part 1: Collecting file link counts. For each file, we create a shadow link + * counting structure, then walk the entire directory tree, incrementing parent + * and child link counts for each directory entry seen. + * + * To avoid false corruption reports in part 2, any failure in this part must + * set the INCOMPLETE flag even when a negative errno is returned. This care + * must be taken with certain errno values (i.e. EFSBADCRC, EFSCORRUPTED, + * ECANCELED) that are absorbed into a scrub state flag update by + * xchk_*_process_error. Scrub and repair share the same incore data + * structures, so the INCOMPLETE flag is critical to prevent a repair based on + * insufficient information. + * + * Because we are scanning a live filesystem, it's possible that another thread + * will try to update the link counts for an inode that we've already scanned. + * This will cause our counts to be incorrect. Therefore, we hook all + * directory entry updates because that is when link count updates occur. By + * shadowing transaction updates in this manner, live nlink check can ensure by + * locking the inode and the shadow structure that its own copies are not out + * of date. Because the hook code runs in a different process context from the + * scrub code and the scrub state flags are not accessed atomically, failures + * in the hook code must abort the iscan and the scrubber must notice the + * aborted scan and set the incomplete flag. + * + * Note that we use jump labels and srcu notifier hooks to minimize the + * overhead when live nlinks is /not/ running. Locking order for nlink + * observations is inode ILOCK -> iscan_lock/xchk_nlink_ctrs lock. + */ + +/* + * Add a delta to an nlink counter, clamping the value to U32_MAX. Because + * XFS_MAXLINK < U32_MAX, the checking code will produce the correct results + * even if we lose some precision. + */ +static inline void +careful_add( + xfs_nlink_t *nlinkp, + int delta) +{ + uint64_t new_value = (uint64_t)(*nlinkp) + delta; + + BUILD_BUG_ON(XFS_MAXLINK > U32_MAX); + *nlinkp = min_t(uint64_t, new_value, U32_MAX); +} + +/* Update incore link count information. Caller must hold the nlinks lock. */ +STATIC int +xchk_nlinks_update_incore( + struct xchk_nlink_ctrs *xnc, + xfs_ino_t ino, + int parents_delta, + int backrefs_delta, + int children_delta) +{ + struct xchk_nlink nl; + int error; + + if (!xnc->nlinks) + return 0; + + error = xfarray_load_sparse(xnc->nlinks, ino, &nl); + if (error) + return error; + + trace_xchk_nlinks_update_incore(xnc->sc->mp, ino, &nl, parents_delta, + backrefs_delta, children_delta); + + careful_add(&nl.parents, parents_delta); + careful_add(&nl.backrefs, backrefs_delta); + careful_add(&nl.children, children_delta); + + nl.flags |= XCHK_NLINK_WRITTEN; + error = xfarray_store(xnc->nlinks, ino, &nl); + if (error == -EFBIG) { + /* + * EFBIG means we tried to store data at too high a byte offset + * in the sparse array. IOWs, we cannot complete the check and + * must notify userspace that the check was incomplete. + */ + error = -ECANCELED; + } + return error; +} + +/* + * Apply a link count change from the regular filesystem into our shadow link + * count structure based on a directory update in progress. + */ +STATIC int +xchk_nlinks_live_update( + struct notifier_block *nb, + unsigned long action, + void *data) +{ + struct xfs_dir_update_params *p = data; + struct xchk_nlink_ctrs *xnc; + int error; + + xnc = container_of(nb, struct xchk_nlink_ctrs, dhook.dirent_hook.nb); + + trace_xchk_nlinks_live_update(xnc->sc->mp, p->dp, action, p->ip->i_ino, + p->delta, p->name->name, p->name->len); + + /* + * If we've already scanned @dp, update the number of parents that link + * to @ip. If @ip is a subdirectory, update the number of child links + * going out of @dp. + */ + if (xchk_iscan_want_live_update(&xnc->collect_iscan, p->dp->i_ino)) { + mutex_lock(&xnc->lock); + error = xchk_nlinks_update_incore(xnc, p->ip->i_ino, p->delta, + 0, 0); + if (!error && S_ISDIR(VFS_IC(p->ip)->i_mode)) + error = xchk_nlinks_update_incore(xnc, p->dp->i_ino, 0, + 0, p->delta); + mutex_unlock(&xnc->lock); + if (error) + goto out_abort; + } + + /* + * If @ip is a subdirectory and we've already scanned it, update the + * number of backrefs pointing to @dp. + */ + if (S_ISDIR(VFS_IC(p->ip)->i_mode) && + xchk_iscan_want_live_update(&xnc->collect_iscan, p->ip->i_ino)) { + mutex_lock(&xnc->lock); + error = xchk_nlinks_update_incore(xnc, p->dp->i_ino, 0, + p->delta, 0); + mutex_unlock(&xnc->lock); + if (error) + goto out_abort; + } + + return NOTIFY_DONE; + +out_abort: + xchk_iscan_abort(&xnc->collect_iscan); + return NOTIFY_DONE; +} + +/* Bump the observed link count for the inode referenced by this entry. */ +STATIC int +xchk_nlinks_collect_dirent( + struct xfs_scrub *sc, + struct xfs_inode *dp, + xfs_dir2_dataptr_t dapos, + const struct xfs_name *name, + xfs_ino_t ino, + void *priv) +{ + struct xchk_nlink_ctrs *xnc = priv; + bool dot = false, dotdot = false; + int error; + + /* Does this name make sense? */ + if (name->len == 0 || !xfs_dir2_namecheck(name->name, name->len)) { + error = -ECANCELED; + goto out_abort; + } + + if (name->len == 1 && name->name[0] == '.') + dot = true; + else if (name->len == 2 && name->name[0] == '.' && + name->name[1] == '.') + dotdot = true; + + /* Don't accept a '.' entry that points somewhere else. */ + if (dot && ino != dp->i_ino) { + error = -ECANCELED; + goto out_abort; + } + + /* Don't accept an invalid inode number. */ + if (!xfs_verify_dir_ino(sc->mp, ino)) { + error = -ECANCELED; + goto out_abort; + } + + /* Update the shadow link counts if we haven't already failed. */ + + if (xchk_iscan_aborted(&xnc->collect_iscan)) { + error = -ECANCELED; + goto out_incomplete; + } + + trace_xchk_nlinks_collect_dirent(sc->mp, dp, ino, name); + + mutex_lock(&xnc->lock); + + /* + * If this is a dotdot entry, it is a back link from dp to ino. How + * we handle this depends on whether or not dp is the root directory. + * + * The root directory is its own parent, so we pretend the dotdot entry + * establishes the "parent" of the root directory. Increment the + * number of parents of the root directory. + * + * Otherwise, increment the number of backrefs pointing back to ino. + */ + if (dotdot) { + if (dp == sc->mp->m_rootip) + error = xchk_nlinks_update_incore(xnc, ino, 1, 0, 0); + else + error = xchk_nlinks_update_incore(xnc, ino, 0, 1, 0); + if (error) + goto out_unlock; + } + + /* + * If this dirent is a forward link from dp to ino, increment the + * number of parents linking into ino. + */ + if (!dot && !dotdot) { + error = xchk_nlinks_update_incore(xnc, ino, 1, 0, 0); + if (error) + goto out_unlock; + } + + /* + * If this dirent is a forward link to a subdirectory, increment the + * number of child links of dp. + */ + if (!dot && !dotdot && name->type == XFS_DIR3_FT_DIR) { + error = xchk_nlinks_update_incore(xnc, dp->i_ino, 0, 0, 1); + if (error) + goto out_unlock; + } + + mutex_unlock(&xnc->lock); + return 0; + +out_unlock: + mutex_unlock(&xnc->lock); +out_abort: + xchk_iscan_abort(&xnc->collect_iscan); +out_incomplete: + xchk_set_incomplete(sc); + return error; +} + +/* Walk a directory to bump the observed link counts of the children. */ +STATIC int +xchk_nlinks_collect_dir( + struct xchk_nlink_ctrs *xnc, + struct xfs_inode *dp) +{ + struct xfs_scrub *sc = xnc->sc; + unsigned int lock_mode; + int error = 0; + + /* Prevent anyone from changing this directory while we walk it. */ + xfs_ilock(dp, XFS_IOLOCK_SHARED); + lock_mode = xfs_ilock_data_map_shared(dp); + + /* + * The dotdot entry of an unlinked directory still points to the last + * parent, but the parent no longer links to this directory. Skip the + * directory to avoid overcounting. + */ + if (VFS_I(dp)->i_nlink == 0) + goto out_unlock; + + /* + * We cannot count file links if the directory looks as though it has + * been zapped by the inode record repair code. + */ + if (xchk_dir_looks_zapped(dp)) { + error = -EBUSY; + goto out_abort; + } + + error = xchk_dir_walk(sc, dp, xchk_nlinks_collect_dirent, xnc); + if (error == -ECANCELED) { + error = 0; + goto out_unlock; + } + if (error) + goto out_abort; + + xchk_iscan_mark_visited(&xnc->collect_iscan, dp); + goto out_unlock; + +out_abort: + xchk_set_incomplete(sc); + xchk_iscan_abort(&xnc->collect_iscan); +out_unlock: + xfs_iunlock(dp, lock_mode); + xfs_iunlock(dp, XFS_IOLOCK_SHARED); + return error; +} + +/* If this looks like a valid pointer, count it. */ +static inline int +xchk_nlinks_collect_metafile( + struct xchk_nlink_ctrs *xnc, + xfs_ino_t ino) +{ + if (!xfs_verify_ino(xnc->sc->mp, ino)) + return 0; + + trace_xchk_nlinks_collect_metafile(xnc->sc->mp, ino); + return xchk_nlinks_update_incore(xnc, ino, 1, 0, 0); +} + +/* Bump the link counts of metadata files rooted in the superblock. */ +STATIC int +xchk_nlinks_collect_metafiles( + struct xchk_nlink_ctrs *xnc) +{ + struct xfs_mount *mp = xnc->sc->mp; + int error = -ECANCELED; + + + if (xchk_iscan_aborted(&xnc->collect_iscan)) + goto out_incomplete; + + mutex_lock(&xnc->lock); + error = xchk_nlinks_collect_metafile(xnc, mp->m_sb.sb_rbmino); + if (error) + goto out_abort; + + error = xchk_nlinks_collect_metafile(xnc, mp->m_sb.sb_rsumino); + if (error) + goto out_abort; + + error = xchk_nlinks_collect_metafile(xnc, mp->m_sb.sb_uquotino); + if (error) + goto out_abort; + + error = xchk_nlinks_collect_metafile(xnc, mp->m_sb.sb_gquotino); + if (error) + goto out_abort; + + error = xchk_nlinks_collect_metafile(xnc, mp->m_sb.sb_pquotino); + if (error) + goto out_abort; + mutex_unlock(&xnc->lock); + + return 0; + +out_abort: + mutex_unlock(&xnc->lock); + xchk_iscan_abort(&xnc->collect_iscan); +out_incomplete: + xchk_set_incomplete(xnc->sc); + return error; +} + +/* Advance the collection scan cursor for this non-directory file. */ +static inline int +xchk_nlinks_collect_file( + struct xchk_nlink_ctrs *xnc, + struct xfs_inode *ip) +{ + xfs_ilock(ip, XFS_IOLOCK_SHARED); + xchk_iscan_mark_visited(&xnc->collect_iscan, ip); + xfs_iunlock(ip, XFS_IOLOCK_SHARED); + return 0; +} + +/* Walk all directories and count inode links. */ +STATIC int +xchk_nlinks_collect( + struct xchk_nlink_ctrs *xnc) +{ + struct xfs_scrub *sc = xnc->sc; + struct xfs_inode *ip; + int error; + + /* Count the rt and quota files that are rooted in the superblock. */ + error = xchk_nlinks_collect_metafiles(xnc); + if (error) + return error; + + /* + * Set up for a potentially lengthy filesystem scan by reducing our + * transaction resource usage for the duration. Specifically: + * + * Cancel the transaction to release the log grant space while we scan + * the filesystem. + * + * Create a new empty transaction to eliminate the possibility of the + * inode scan deadlocking on cyclical metadata. + * + * We pass the empty transaction to the file scanning function to avoid + * repeatedly cycling empty transactions. This can be done even though + * we take the IOLOCK to quiesce the file because empty transactions + * do not take sb_internal. + */ + xchk_trans_cancel(sc); + error = xchk_trans_alloc_empty(sc); + if (error) + return error; + + while ((error = xchk_iscan_iter(&xnc->collect_iscan, &ip)) == 1) { + if (S_ISDIR(VFS_I(ip)->i_mode)) + error = xchk_nlinks_collect_dir(xnc, ip); + else + error = xchk_nlinks_collect_file(xnc, ip); + xchk_irele(sc, ip); + if (error) + break; + + if (xchk_should_terminate(sc, &error)) + break; + } + xchk_iscan_iter_finish(&xnc->collect_iscan); + if (error) { + xchk_set_incomplete(sc); + /* + * If we couldn't grab an inode that was busy with a state + * change, change the error code so that we exit to userspace + * as quickly as possible. + */ + if (error == -EBUSY) + return -ECANCELED; + return error; + } + + /* + * Switch out for a real transaction in preparation for building a new + * tree. + */ + xchk_trans_cancel(sc); + return xchk_setup_fs(sc); +} + +/* + * Part 2: Comparing file link counters. Walk each inode and compare the link + * counts against our shadow information; and then walk each shadow link count + * structure (that wasn't covered in the first part), comparing it against the + * file. + */ + +/* Read the observed link count for comparison with the actual inode. */ +STATIC int +xchk_nlinks_comparison_read( + struct xchk_nlink_ctrs *xnc, + xfs_ino_t ino, + struct xchk_nlink *obs) +{ + struct xchk_nlink nl; + int error; + + error = xfarray_load_sparse(xnc->nlinks, ino, &nl); + if (error) + return error; + + nl.flags |= (XCHK_NLINK_COMPARE_SCANNED | XCHK_NLINK_WRITTEN); + + error = xfarray_store(xnc->nlinks, ino, &nl); + if (error == -EFBIG) { + /* + * EFBIG means we tried to store data at too high a byte offset + * in the sparse array. IOWs, we cannot complete the check and + * must notify userspace that the check was incomplete. This + * shouldn't really happen outside of the collection phase. + */ + xchk_set_incomplete(xnc->sc); + return -ECANCELED; + } + if (error) + return error; + + /* Copy the counters, but do not expose the internal state. */ + obs->parents = nl.parents; + obs->backrefs = nl.backrefs; + obs->children = nl.children; + obs->flags = 0; + return 0; +} + +/* Check our link count against an inode. */ +STATIC int +xchk_nlinks_compare_inode( + struct xchk_nlink_ctrs *xnc, + struct xfs_inode *ip) +{ + struct xchk_nlink obs; + struct xfs_scrub *sc = xnc->sc; + uint64_t total_links; + unsigned int actual_nlink; + int error; + + xfs_ilock(ip, XFS_ILOCK_SHARED); + mutex_lock(&xnc->lock); + + if (xchk_iscan_aborted(&xnc->collect_iscan)) { + xchk_set_incomplete(xnc->sc); + error = -ECANCELED; + goto out_scanlock; + } + + error = xchk_nlinks_comparison_read(xnc, ip->i_ino, &obs); + if (error) + goto out_scanlock; + + /* + * If we don't have ftype to get an accurate count of the subdirectory + * entries in this directory, take advantage of the fact that on a + * consistent ftype=0 filesystem, the number of subdirectory + * backreferences (dotdot entries) pointing towards this directory + * should be equal to the number of subdirectory entries in the + * directory. + */ + if (!xfs_has_ftype(sc->mp) && S_ISDIR(VFS_I(ip)->i_mode)) + obs.children = obs.backrefs; + + total_links = xchk_nlink_total(ip, &obs); + actual_nlink = VFS_I(ip)->i_nlink; + + trace_xchk_nlinks_compare_inode(sc->mp, ip, &obs); + + /* + * If we found so many parents that we'd overflow i_nlink, we must flag + * this as a corruption. The VFS won't let users increase the link + * count, but it will let them decrease it. + */ + if (total_links > XFS_MAXLINK) { + xchk_ino_set_corrupt(sc, ip->i_ino); + goto out_corrupt; + } + + /* Link counts should match. */ + if (total_links != actual_nlink) { + xchk_ino_set_corrupt(sc, ip->i_ino); + goto out_corrupt; + } + + if (S_ISDIR(VFS_I(ip)->i_mode) && actual_nlink > 0) { + /* + * The collection phase ignores directories with zero link + * count, so we ignore them here too. + * + * The number of subdirectory backreferences (dotdot entries) + * pointing towards this directory should be equal to the + * number of subdirectory entries in the directory. + */ + if (obs.children != obs.backrefs) + xchk_ino_xref_set_corrupt(sc, ip->i_ino); + } else { + /* + * Non-directories and unlinked directories should not have + * back references. + */ + if (obs.backrefs != 0) { + xchk_ino_set_corrupt(sc, ip->i_ino); + goto out_corrupt; + } + + /* + * Non-directories and unlinked directories should not have + * children. + */ + if (obs.children != 0) { + xchk_ino_set_corrupt(sc, ip->i_ino); + goto out_corrupt; + } + } + + if (ip == sc->mp->m_rootip) { + /* + * For the root of a directory tree, both the '.' and '..' + * entries should point to the root directory. The dotdot + * entry is counted as a parent of the root /and/ a backref of + * the root directory. + */ + if (obs.parents != 1) { + xchk_ino_set_corrupt(sc, ip->i_ino); + goto out_corrupt; + } + } else if (actual_nlink > 0) { + /* + * Linked files that are not the root directory should have at + * least one parent. + */ + if (obs.parents == 0) { + xchk_ino_set_corrupt(sc, ip->i_ino); + goto out_corrupt; + } + } + +out_corrupt: + if (sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT) + error = -ECANCELED; +out_scanlock: + mutex_unlock(&xnc->lock); + xfs_iunlock(ip, XFS_ILOCK_SHARED); + return error; +} + +/* + * Check our link count against an inode that wasn't checked previously. This + * is intended to catch directories with dangling links, though we could be + * racing with inode allocation in other threads. + */ +STATIC int +xchk_nlinks_compare_inum( + struct xchk_nlink_ctrs *xnc, + xfs_ino_t ino) +{ + struct xchk_nlink obs; + struct xfs_mount *mp = xnc->sc->mp; + struct xfs_trans *tp = xnc->sc->tp; + struct xfs_buf *agi_bp; + struct xfs_inode *ip; + int error; + + /* + * The first iget failed, so try again with the variant that returns + * either an incore inode or the AGI buffer. If the function returns + * EINVAL/ENOENT, it should have passed us the AGI buffer so that we + * can guarantee that the inode won't be allocated while we check for + * a zero link count in the observed link count data. + */ + error = xchk_iget_agi(xnc->sc, ino, &agi_bp, &ip); + if (!error) { + /* Actually got an inode, so use the inode compare. */ + error = xchk_nlinks_compare_inode(xnc, ip); + xchk_irele(xnc->sc, ip); + return error; + } + if (error == -ENOENT || error == -EINVAL) { + /* No inode was found. Check for zero link count below. */ + error = 0; + } + if (error) + goto out_agi; + + /* Ensure that we have protected against inode allocation/freeing. */ + if (agi_bp == NULL) { + ASSERT(agi_bp != NULL); + xchk_set_incomplete(xnc->sc); + return -ECANCELED; + } + + if (xchk_iscan_aborted(&xnc->collect_iscan)) { + xchk_set_incomplete(xnc->sc); + error = -ECANCELED; + goto out_agi; + } + + mutex_lock(&xnc->lock); + error = xchk_nlinks_comparison_read(xnc, ino, &obs); + if (error) + goto out_scanlock; + + trace_xchk_nlinks_check_zero(mp, ino, &obs); + + /* + * If we can't grab the inode, the link count had better be zero. We + * still hold the AGI to prevent inode allocation/freeing. + */ + if (xchk_nlink_total(NULL, &obs) != 0) { + xchk_ino_set_corrupt(xnc->sc, ino); + error = -ECANCELED; + } + +out_scanlock: + mutex_unlock(&xnc->lock); +out_agi: + if (agi_bp) + xfs_trans_brelse(tp, agi_bp); + return error; +} + +/* + * Try to visit every inode in the filesystem to compare the link count. Move + * on if we can't grab an inode, since we'll revisit unchecked nlink records in + * the second part. + */ +static int +xchk_nlinks_compare_iter( + struct xchk_nlink_ctrs *xnc, + struct xfs_inode **ipp) +{ + int error; + + do { + error = xchk_iscan_iter(&xnc->compare_iscan, ipp); + } while (error == -EBUSY); + + return error; +} + +/* Compare the link counts we observed against the live information. */ +STATIC int +xchk_nlinks_compare( + struct xchk_nlink_ctrs *xnc) +{ + struct xchk_nlink nl; + struct xfs_scrub *sc = xnc->sc; + struct xfs_inode *ip; + xfarray_idx_t cur = XFARRAY_CURSOR_INIT; + int error; + + if (sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT) + return 0; + + /* + * Create a new empty transaction so that we can advance the iscan + * cursor without deadlocking if the inobt has a cycle and push on the + * inactivation workqueue. + */ + xchk_trans_cancel(sc); + error = xchk_trans_alloc_empty(sc); + if (error) + return error; + + /* + * Use the inobt to walk all allocated inodes to compare the link + * counts. Inodes skipped by _compare_iter will be tried again in the + * next phase of the scan. + */ + xchk_iscan_start(sc, 0, 0, &xnc->compare_iscan); + while ((error = xchk_nlinks_compare_iter(xnc, &ip)) == 1) { + error = xchk_nlinks_compare_inode(xnc, ip); + xchk_iscan_mark_visited(&xnc->compare_iscan, ip); + xchk_irele(sc, ip); + if (error) + break; + + if (xchk_should_terminate(sc, &error)) + break; + } + xchk_iscan_iter_finish(&xnc->compare_iscan); + xchk_iscan_teardown(&xnc->compare_iscan); + if (error) + return error; + + if (sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT) + return 0; + + /* + * Walk all the non-null nlink observations that weren't checked in the + * previous step. + */ + mutex_lock(&xnc->lock); + while ((error = xfarray_iter(xnc->nlinks, &cur, &nl)) == 1) { + xfs_ino_t ino = cur - 1; + + if (nl.flags & XCHK_NLINK_COMPARE_SCANNED) + continue; + + mutex_unlock(&xnc->lock); + + error = xchk_nlinks_compare_inum(xnc, ino); + if (error) + return error; + + if (xchk_should_terminate(xnc->sc, &error)) + return error; + + mutex_lock(&xnc->lock); + } + mutex_unlock(&xnc->lock); + + return error; +} + +/* Tear down everything associated with a nlinks check. */ +static void +xchk_nlinks_teardown_scan( + void *priv) +{ + struct xchk_nlink_ctrs *xnc = priv; + + /* Discourage any hook functions that might be running. */ + xchk_iscan_abort(&xnc->collect_iscan); + + xfs_dir_hook_del(xnc->sc->mp, &xnc->dhook); + + xfarray_destroy(xnc->nlinks); + xnc->nlinks = NULL; + + xchk_iscan_teardown(&xnc->collect_iscan); + mutex_destroy(&xnc->lock); + xnc->sc = NULL; +} + +/* + * Scan all inodes in the entire filesystem to generate link count data. If + * the scan is successful, the counts will be left alive for a repair. If any + * error occurs, we'll tear everything down. + */ +STATIC int +xchk_nlinks_setup_scan( + struct xfs_scrub *sc, + struct xchk_nlink_ctrs *xnc) +{ + struct xfs_mount *mp = sc->mp; + char *descr; + unsigned long long max_inos; + xfs_agnumber_t last_agno = mp->m_sb.sb_agcount - 1; + xfs_agino_t first_agino, last_agino; + int error; + + ASSERT(xnc->sc == NULL); + xnc->sc = sc; + + mutex_init(&xnc->lock); + + /* Retry iget every tenth of a second for up to 30 seconds. */ + xchk_iscan_start(sc, 30000, 100, &xnc->collect_iscan); + + /* + * Set up enough space to store an nlink record for the highest + * possible inode number in this system. + */ + xfs_agino_range(mp, last_agno, &first_agino, &last_agino); + max_inos = XFS_AGINO_TO_INO(mp, last_agno, last_agino) + 1; + descr = xchk_xfile_descr(sc, "file link counts"); + error = xfarray_create(descr, min(XFS_MAXINUMBER + 1, max_inos), + sizeof(struct xchk_nlink), &xnc->nlinks); + kfree(descr); + if (error) + goto out_teardown; + + /* + * Hook into the directory entry code so that we can capture updates to + * file link counts. The hook only triggers for inodes that were + * already scanned, and the scanner thread takes each inode's ILOCK, + * which means that any in-progress inode updates will finish before we + * can scan the inode. + */ + ASSERT(sc->flags & XCHK_FSGATES_DIRENTS); + xfs_dir_hook_setup(&xnc->dhook, xchk_nlinks_live_update); + error = xfs_dir_hook_add(mp, &xnc->dhook); + if (error) + goto out_teardown; + + /* Use deferred cleanup to pass the inode link count data to repair. */ + sc->buf_cleanup = xchk_nlinks_teardown_scan; + return 0; + +out_teardown: + xchk_nlinks_teardown_scan(xnc); + return error; +} + +/* Scrub the link count of all inodes on the filesystem. */ +int +xchk_nlinks( + struct xfs_scrub *sc) +{ + struct xchk_nlink_ctrs *xnc = sc->buf; + int error = 0; + + /* Set ourselves up to check link counts on the live filesystem. */ + error = xchk_nlinks_setup_scan(sc, xnc); + if (error) + return error; + + /* Walk all inodes, picking up link count information. */ + error = xchk_nlinks_collect(xnc); + if (!xchk_xref_process_error(sc, 0, 0, &error)) + return error; + + /* Fail fast if we're not playing with a full dataset. */ + if (xchk_iscan_aborted(&xnc->collect_iscan)) + xchk_set_incomplete(sc); + if (sc->sm->sm_flags & XFS_SCRUB_OFLAG_INCOMPLETE) + return 0; + + /* Compare link counts. */ + error = xchk_nlinks_compare(xnc); + if (!xchk_xref_process_error(sc, 0, 0, &error)) + return error; + + /* Check one last time for an incomplete dataset. */ + if (xchk_iscan_aborted(&xnc->collect_iscan)) + xchk_set_incomplete(sc); + + return 0; +} diff --git a/fs/xfs/scrub/nlinks.h b/fs/xfs/scrub/nlinks.h new file mode 100644 index 000000000000..a950f3daf204 --- /dev/null +++ b/fs/xfs/scrub/nlinks.h @@ -0,0 +1,102 @@ +/* SPDX-License-Identifier: GPL-2.0-or-later */ +/* + * Copyright (c) 2021-2024 Oracle. All Rights Reserved. + * Author: Darrick J. Wong <djwong@kernel.org> + */ +#ifndef __XFS_SCRUB_NLINKS_H__ +#define __XFS_SCRUB_NLINKS_H__ + +/* Live link count control structure. */ +struct xchk_nlink_ctrs { + struct xfs_scrub *sc; + + /* Shadow link count data and its mutex. */ + struct xfarray *nlinks; + struct mutex lock; + + /* + * The collection step uses a separate iscan context from the compare + * step because the collection iscan coordinates live updates to the + * observation data while this scanner is running. The compare iscan + * is secondary and can be reinitialized as needed. + */ + struct xchk_iscan collect_iscan; + struct xchk_iscan compare_iscan; + + /* + * Hook into directory updates so that we can receive live updates + * from other writer threads. + */ + struct xfs_dir_hook dhook; +}; + +/* + * In-core link counts for a given inode in the filesystem. + * + * For an empty rootdir, the directory entries and the field to which they are + * accounted are as follows: + * + * Root directory: + * + * . points to self (root.child) + * .. points to self (root.parent) + * f1 points to a child file (f1.parent) + * d1 points to a child dir (d1.parent, root.child) + * + * Subdirectory d1: + * + * . points to self (d1.child) + * .. points to root dir (root.backref) + * f2 points to child file (f2.parent) + * f3 points to root.f1 (f1.parent) + * + * root.nlink == 3 (root.dot, root.dotdot, root.d1) + * d1.nlink == 2 (root.d1, d1.dot) + * f1.nlink == 2 (root.f1, d1.f3) + * f2.nlink == 1 (d1.f2) + */ +struct xchk_nlink { + /* Count of forward links from parent directories to this file. */ + xfs_nlink_t parents; + + /* + * Count of back links to this parent directory from child + * subdirectories. + */ + xfs_nlink_t backrefs; + + /* + * Count of forward links from this directory to all child files and + * the number of dot entries. Should be zero for non-directories. + */ + xfs_nlink_t children; + + /* Record state flags */ + unsigned int flags; +}; + +/* + * This incore link count has been written at least once. We never want to + * store an xchk_nlink that looks uninitialized. + */ +#define XCHK_NLINK_WRITTEN (1U << 0) + +/* Already checked this link count record. */ +#define XCHK_NLINK_COMPARE_SCANNED (1U << 1) + +/* Already made a repair with this link count record. */ +#define XREP_NLINK_DIRTY (1U << 2) + +/* Compute total link count, using large enough variables to detect overflow. */ +static inline uint64_t +xchk_nlink_total(struct xfs_inode *ip, const struct xchk_nlink *live) +{ + uint64_t ret = live->parents; + + /* Add one link count for the dot entry of any linked directory. */ + if (ip && S_ISDIR(VFS_I(ip)->i_mode) && VFS_I(ip)->i_nlink) + ret++; + return ret + live->children; +} + +#endif /* __XFS_SCRUB_NLINKS_H__ */ diff --git a/fs/xfs/scrub/nlinks_repair.c b/fs/xfs/scrub/nlinks_repair.c new file mode 100644 index 000000000000..b87618322f55 --- /dev/null +++ b/fs/xfs/scrub/nlinks_repair.c @@ -0,0 +1,223 @@ +// SPDX-License-Identifier: GPL-2.0-or-later +/* + * Copyright (c) 2021-2024 Oracle. All Rights Reserved. + * Author: Darrick J. Wong <djwong@kernel.org> + */ +#include "xfs.h" +#include "xfs_fs.h" +#include "xfs_shared.h" +#include "xfs_format.h" +#include "xfs_trans_resv.h" +#include "xfs_mount.h" +#include "xfs_log_format.h" +#include "xfs_trans.h" +#include "xfs_inode.h" +#include "xfs_icache.h" +#include "xfs_bmap_util.h" +#include "xfs_iwalk.h" +#include "xfs_ialloc.h" +#include "xfs_sb.h" +#include "scrub/scrub.h" +#include "scrub/common.h" +#include "scrub/repair.h" +#include "scrub/xfile.h" +#include "scrub/xfarray.h" +#include "scrub/iscan.h" +#include "scrub/nlinks.h" +#include "scrub/trace.h" + +/* + * Live Inode Link Count Repair + * ============================ + * + * Use the live inode link count information that we collected to replace the + * nlink values of the incore inodes. A scrub->repair cycle should have left + * the live data and hooks active, so this is safe so long as we make sure the + * inode is locked. + */ + +/* + * Correct the link count of the given inode. Because we have to grab locks + * and resources in a certain order, it's possible that this will be a no-op. + */ +STATIC int +xrep_nlinks_repair_inode( + struct xchk_nlink_ctrs *xnc) +{ + struct xchk_nlink obs; + struct xfs_scrub *sc = xnc->sc; + struct xfs_mount *mp = sc->mp; + struct xfs_inode *ip = sc->ip; + uint64_t total_links; + uint64_t actual_nlink; + bool dirty = false; + int error; + + xchk_ilock(sc, XFS_IOLOCK_EXCL); + + error = xfs_trans_alloc(mp, &M_RES(mp)->tr_link, 0, 0, 0, &sc->tp); + if (error) + return error; + + xchk_ilock(sc, XFS_ILOCK_EXCL); + xfs_trans_ijoin(sc->tp, ip, 0); + + mutex_lock(&xnc->lock); + + if (xchk_iscan_aborted(&xnc->collect_iscan)) { + error = -ECANCELED; + goto out_scanlock; + } + + error = xfarray_load_sparse(xnc->nlinks, ip->i_ino, &obs); + if (error) + goto out_scanlock; + + /* + * We're done accessing the shared scan data, so we can drop the lock. + * We still hold @ip's ILOCK, so its link count cannot change. + */ + mutex_unlock(&xnc->lock); + + total_links = xchk_nlink_total(ip, &obs); + actual_nlink = VFS_I(ip)->i_nlink; + + /* + * Non-directories cannot have directories pointing up to them. + * + * We previously set error to zero, but set it again because one static + * checker author fears that programmers will fail to maintain this + * invariant and built their tool to flag this as a security risk. A + * different tool author made their bot complain about the redundant + * store. This is a never-ending and stupid battle; both tools missed + * *actual bugs* elsewhere; and I no longer care. + */ + if (!S_ISDIR(VFS_I(ip)->i_mode) && obs.children != 0) { + trace_xrep_nlinks_unfixable_inode(mp, ip, &obs); + error = 0; + goto out_trans; + } + + /* + * We did not find any links to this inode. If the inode agrees, we + * have nothing further to do. If not, the inode has a nonzero link + * count and we don't have anywhere to graft the child onto. Dropping + * a live inode's link count to zero can cause unexpected shutdowns in + * inactivation, so leave it alone. + */ + if (total_links == 0) { + if (actual_nlink != 0) + trace_xrep_nlinks_unfixable_inode(mp, ip, &obs); + goto out_trans; + } + + /* Commit the new link count if it changed. */ + if (total_links != actual_nlink) { + if (total_links > XFS_MAXLINK) { + trace_xrep_nlinks_unfixable_inode(mp, ip, &obs); + goto out_trans; + } + + trace_xrep_nlinks_update_inode(mp, ip, &obs); + + set_nlink(VFS_I(ip), total_links); + dirty = true; + } + + if (!dirty) { + error = 0; + goto out_trans; + } + + xfs_trans_log_inode(sc->tp, ip, XFS_ILOG_CORE); + + error = xrep_trans_commit(sc); + xchk_iunlock(sc, XFS_ILOCK_EXCL | XFS_IOLOCK_EXCL); + return error; + +out_scanlock: + mutex_unlock(&xnc->lock); +out_trans: + xchk_trans_cancel(sc); + xchk_iunlock(sc, XFS_ILOCK_EXCL | XFS_IOLOCK_EXCL); + return error; +} + +/* + * Try to visit every inode in the filesystem for repairs. Move on if we can't + * grab an inode, since we're still making forward progress. + */ +static int +xrep_nlinks_iter( + struct xchk_nlink_ctrs *xnc, + struct xfs_inode **ipp) +{ + int error; + + do { + error = xchk_iscan_iter(&xnc->compare_iscan, ipp); + } while (error == -EBUSY); + + return error; +} + +/* Commit the new inode link counters. */ +int +xrep_nlinks( + struct xfs_scrub *sc) +{ + struct xchk_nlink_ctrs *xnc = sc->buf; + int error; + + /* + * We need ftype for an accurate count of the number of child + * subdirectory links. Child subdirectories with a back link (dotdot + * entry) but no forward link are unfixable, so we cannot repair the + * link count of the parent directory based on the back link count + * alone. Filesystems without ftype support are rare (old V4) so we + * just skip out here. + */ + if (!xfs_has_ftype(sc->mp)) + return -EOPNOTSUPP; + + /* + * Use the inobt to walk all allocated inodes to compare and fix the + * link counts. Retry iget every tenth of a second for up to 30 + * seconds -- even if repair misses a few inodes, we still try to fix + * as many of them as we can. + */ + xchk_iscan_start(sc, 30000, 100, &xnc->compare_iscan); + ASSERT(sc->ip == NULL); + + while ((error = xrep_nlinks_iter(xnc, &sc->ip)) == 1) { + /* + * Commit the scrub transaction so that we can create repair + * transactions with the correct reservations. + */ + xchk_trans_cancel(sc); + + error = xrep_nlinks_repair_inode(xnc); + xchk_iscan_mark_visited(&xnc->compare_iscan, sc->ip); + xchk_irele(sc, sc->ip); + sc->ip = NULL; + if (error) + break; + + if (xchk_should_terminate(sc, &error)) + break; + + /* + * Create a new empty transaction so that we can advance the + * iscan cursor without deadlocking if the inobt has a cycle. + * We can only push the inactivation workqueues with an empty + * transaction. + */ + error = xchk_trans_alloc_empty(sc); + if (error) + break; + } + xchk_iscan_iter_finish(&xnc->compare_iscan); + xchk_iscan_teardown(&xnc->compare_iscan); + + return error; +} diff --git a/fs/xfs/scrub/quotacheck.c b/fs/xfs/scrub/quotacheck.c new file mode 100644 index 000000000000..c77eb2de8df7 --- /dev/null +++ b/fs/xfs/scrub/quotacheck.c @@ -0,0 +1,867 @@ +// SPDX-License-Identifier: GPL-2.0-or-later +/* + * Copyright (c) 2020-2024 Oracle. All Rights Reserved. + * Author: Darrick J. Wong <djwong@kernel.org> + */ +#include "xfs.h" +#include "xfs_fs.h" +#include "xfs_shared.h" +#include "xfs_format.h" +#include "xfs_trans_resv.h" +#include "xfs_mount.h" +#include "xfs_log_format.h" +#include "xfs_trans.h" +#include "xfs_inode.h" +#include "xfs_quota.h" +#include "xfs_qm.h" +#include "xfs_icache.h" +#include "xfs_bmap_util.h" +#include "xfs_ialloc.h" +#include "xfs_ag.h" +#include "scrub/scrub.h" +#include "scrub/common.h" +#include "scrub/repair.h" +#include "scrub/xfile.h" +#include "scrub/xfarray.h" +#include "scrub/iscan.h" +#include "scrub/quota.h" +#include "scrub/quotacheck.h" +#include "scrub/trace.h" + +/* + * Live Quotacheck + * =============== + * + * Quota counters are "summary" metadata, in the sense that they are computed + * as the summation of the block usage counts for every file on the filesystem. + * Therefore, we compute the correct icount, bcount, and rtbcount values by + * creating a shadow quota counter structure and walking every inode. + */ + +/* Track the quota deltas for a dquot in a transaction. */ +struct xqcheck_dqtrx { + xfs_dqtype_t q_type; + xfs_dqid_t q_id; + + int64_t icount_delta; + + int64_t bcount_delta; + int64_t delbcnt_delta; + + int64_t rtbcount_delta; + int64_t delrtb_delta; +}; + +#define XQCHECK_MAX_NR_DQTRXS (XFS_QM_TRANS_DQTYPES * XFS_QM_TRANS_MAXDQS) + +/* + * Track the quota deltas for all dquots attached to a transaction if the + * quota deltas are being applied to an inode that we already scanned. + */ +struct xqcheck_dqacct { + struct rhash_head hash; + uintptr_t tx_id; + struct xqcheck_dqtrx dqtrx[XQCHECK_MAX_NR_DQTRXS]; + unsigned int refcount; +}; + +/* Free a shadow dquot accounting structure. */ +static void +xqcheck_dqacct_free( + void *ptr, + void *arg) +{ + struct xqcheck_dqacct *dqa = ptr; + + kfree(dqa); +} + +/* Set us up to scrub quota counters. */ +int +xchk_setup_quotacheck( + struct xfs_scrub *sc) +{ + if (!XFS_IS_QUOTA_ON(sc->mp)) + return -ENOENT; + + xchk_fsgates_enable(sc, XCHK_FSGATES_QUOTA); + + sc->buf = kzalloc(sizeof(struct xqcheck), XCHK_GFP_FLAGS); + if (!sc->buf) + return -ENOMEM; + + return xchk_setup_fs(sc); +} + +/* + * Part 1: Collecting dquot resource usage counts. For each xfs_dquot attached + * to each inode, we create a shadow dquot, and compute the inode count and add + * the data/rt block usage from what we see. + * + * To avoid false corruption reports in part 2, any failure in this part must + * set the INCOMPLETE flag even when a negative errno is returned. This care + * must be taken with certain errno values (i.e. EFSBADCRC, EFSCORRUPTED, + * ECANCELED) that are absorbed into a scrub state flag update by + * xchk_*_process_error. Scrub and repair share the same incore data + * structures, so the INCOMPLETE flag is critical to prevent a repair based on + * insufficient information. + * + * Because we are scanning a live filesystem, it's possible that another thread + * will try to update the quota counters for an inode that we've already + * scanned. This will cause our counts to be incorrect. Therefore, we hook + * the live transaction code in two places: (1) when the callers update the + * per-transaction dqtrx structure to log quota counter updates; and (2) when + * transaction commit actually logs those updates to the incore dquot. By + * shadowing transaction updates in this manner, live quotacheck can ensure + * by locking the dquot and the shadow structure that its own copies are not + * out of date. Because the hook code runs in a different process context from + * the scrub code and the scrub state flags are not accessed atomically, + * failures in the hook code must abort the iscan and the scrubber must notice + * the aborted scan and set the incomplete flag. + * + * Note that we use srcu notifier hooks to minimize the overhead when live + * quotacheck is /not/ running. + */ + +/* Update an incore dquot counter information from a live update. */ +static int +xqcheck_update_incore_counts( + struct xqcheck *xqc, + struct xfarray *counts, + xfs_dqid_t id, + int64_t inodes, + int64_t nblks, + int64_t rtblks) +{ + struct xqcheck_dquot xcdq; + int error; + + error = xfarray_load_sparse(counts, id, &xcdq); + if (error) + return error; + + xcdq.flags |= XQCHECK_DQUOT_WRITTEN; + xcdq.icount += inodes; + xcdq.bcount += nblks; + xcdq.rtbcount += rtblks; + + error = xfarray_store(counts, id, &xcdq); + if (error == -EFBIG) { + /* + * EFBIG means we tried to store data at too high a byte offset + * in the sparse array. IOWs, we cannot complete the check and + * must notify userspace that the check was incomplete. + */ + error = -ECANCELED; + } + return error; +} + +/* Decide if this is the shadow dquot accounting structure for a transaction. */ +static int +xqcheck_dqacct_obj_cmpfn( + struct rhashtable_compare_arg *arg, + const void *obj) +{ + const uintptr_t *tx_idp = arg->key; + const struct xqcheck_dqacct *dqa = obj; + + if (dqa->tx_id != *tx_idp) + return 1; + return 0; +} + +static const struct rhashtable_params xqcheck_dqacct_hash_params = { + .min_size = 32, + .key_len = sizeof(uintptr_t), + .key_offset = offsetof(struct xqcheck_dqacct, tx_id), + .head_offset = offsetof(struct xqcheck_dqacct, hash), + .automatic_shrinking = true, + .obj_cmpfn = xqcheck_dqacct_obj_cmpfn, +}; + +/* Find a shadow dqtrx slot for the given dquot. */ +STATIC struct xqcheck_dqtrx * +xqcheck_get_dqtrx( + struct xqcheck_dqacct *dqa, + xfs_dqtype_t q_type, + xfs_dqid_t q_id) +{ + int i; + + for (i = 0; i < XQCHECK_MAX_NR_DQTRXS; i++) { + if (dqa->dqtrx[i].q_type == 0 || + (dqa->dqtrx[i].q_type == q_type && + dqa->dqtrx[i].q_id == q_id)) + return &dqa->dqtrx[i]; + } + + return NULL; +} + +/* + * Create and fill out a quota delta tracking structure to shadow the updates + * going on in the regular quota code. + */ +static int +xqcheck_mod_live_ino_dqtrx( + struct notifier_block *nb, + unsigned long action, + void *data) +{ + struct xfs_mod_ino_dqtrx_params *p = data; + struct xqcheck *xqc; + struct xqcheck_dqacct *dqa; + struct xqcheck_dqtrx *dqtrx; + int error; + + xqc = container_of(nb, struct xqcheck, qhook.mod_hook.nb); + + /* Skip quota reservation fields. */ + switch (action) { + case XFS_TRANS_DQ_BCOUNT: + case XFS_TRANS_DQ_DELBCOUNT: + case XFS_TRANS_DQ_ICOUNT: + case XFS_TRANS_DQ_RTBCOUNT: + case XFS_TRANS_DQ_DELRTBCOUNT: + break; + default: + return NOTIFY_DONE; + } + + /* Ignore dqtrx updates for quota types we don't care about. */ + switch (p->q_type) { + case XFS_DQTYPE_USER: + if (!xqc->ucounts) + return NOTIFY_DONE; + break; + case XFS_DQTYPE_GROUP: + if (!xqc->gcounts) + return NOTIFY_DONE; + break; + case XFS_DQTYPE_PROJ: + if (!xqc->pcounts) + return NOTIFY_DONE; + break; + default: + return NOTIFY_DONE; + } + + /* Skip inodes that haven't been scanned yet. */ + if (!xchk_iscan_want_live_update(&xqc->iscan, p->ino)) + return NOTIFY_DONE; + + /* Make a shadow quota accounting tracker for this transaction. */ + mutex_lock(&xqc->lock); + dqa = rhashtable_lookup_fast(&xqc->shadow_dquot_acct, &p->tx_id, + xqcheck_dqacct_hash_params); + if (!dqa) { + dqa = kzalloc(sizeof(struct xqcheck_dqacct), XCHK_GFP_FLAGS); + if (!dqa) + goto out_abort; + + dqa->tx_id = p->tx_id; + error = rhashtable_insert_fast(&xqc->shadow_dquot_acct, + &dqa->hash, xqcheck_dqacct_hash_params); + if (error) + goto out_abort; + } + + /* Find the shadow dqtrx (or an empty slot) here. */ + dqtrx = xqcheck_get_dqtrx(dqa, p->q_type, p->q_id); + if (!dqtrx) + goto out_abort; + if (dqtrx->q_type == 0) { + dqtrx->q_type = p->q_type; + dqtrx->q_id = p->q_id; + dqa->refcount++; + } + + /* Update counter */ + switch (action) { + case XFS_TRANS_DQ_BCOUNT: + dqtrx->bcount_delta += p->delta; + break; + case XFS_TRANS_DQ_DELBCOUNT: + dqtrx->delbcnt_delta += p->delta; + break; + case XFS_TRANS_DQ_ICOUNT: + dqtrx->icount_delta += p->delta; + break; + case XFS_TRANS_DQ_RTBCOUNT: + dqtrx->rtbcount_delta += p->delta; + break; + case XFS_TRANS_DQ_DELRTBCOUNT: + dqtrx->delrtb_delta += p->delta; + break; + } + + mutex_unlock(&xqc->lock); + return NOTIFY_DONE; + +out_abort: + xchk_iscan_abort(&xqc->iscan); + mutex_unlock(&xqc->lock); + return NOTIFY_DONE; +} + +/* + * Apply the transaction quota deltas to our shadow quota accounting info when + * the regular quota code are doing the same. + */ +static int +xqcheck_apply_live_dqtrx( + struct notifier_block *nb, + unsigned long action, + void *data) +{ + struct xfs_apply_dqtrx_params *p = data; + struct xqcheck *xqc; + struct xqcheck_dqacct *dqa; + struct xqcheck_dqtrx *dqtrx; + struct xfarray *counts; + int error; + + xqc = container_of(nb, struct xqcheck, qhook.apply_hook.nb); + + /* Map the dquot type to an incore counter object. */ + switch (p->q_type) { + case XFS_DQTYPE_USER: + counts = xqc->ucounts; + break; + case XFS_DQTYPE_GROUP: + counts = xqc->gcounts; + break; + case XFS_DQTYPE_PROJ: + counts = xqc->pcounts; + break; + default: + return NOTIFY_DONE; + } + + if (xchk_iscan_aborted(&xqc->iscan) || counts == NULL) + return NOTIFY_DONE; + + /* + * Find the shadow dqtrx for this transaction and dquot, if any deltas + * need to be applied here. If not, we're finished early. + */ + mutex_lock(&xqc->lock); + dqa = rhashtable_lookup_fast(&xqc->shadow_dquot_acct, &p->tx_id, + xqcheck_dqacct_hash_params); + if (!dqa) + goto out_unlock; + dqtrx = xqcheck_get_dqtrx(dqa, p->q_type, p->q_id); + if (!dqtrx || dqtrx->q_type == 0) + goto out_unlock; + + /* Update our shadow dquot if we're committing. */ + if (action == XFS_APPLY_DQTRX_COMMIT) { + error = xqcheck_update_incore_counts(xqc, counts, p->q_id, + dqtrx->icount_delta, + dqtrx->bcount_delta + dqtrx->delbcnt_delta, + dqtrx->rtbcount_delta + dqtrx->delrtb_delta); + if (error) + goto out_abort; + } + + /* Free the shadow accounting structure if that was the last user. */ + dqa->refcount--; + if (dqa->refcount == 0) { + error = rhashtable_remove_fast(&xqc->shadow_dquot_acct, + &dqa->hash, xqcheck_dqacct_hash_params); + if (error) + goto out_abort; + xqcheck_dqacct_free(dqa, NULL); + } + + mutex_unlock(&xqc->lock); + return NOTIFY_DONE; + +out_abort: + xchk_iscan_abort(&xqc->iscan); +out_unlock: + mutex_unlock(&xqc->lock); + return NOTIFY_DONE; +} + +/* Record this inode's quota usage in our shadow quota counter data. */ +STATIC int +xqcheck_collect_inode( + struct xqcheck *xqc, + struct xfs_inode *ip) +{ + struct xfs_trans *tp = xqc->sc->tp; + xfs_filblks_t nblks, rtblks; + uint ilock_flags = 0; + xfs_dqid_t id; + bool isreg = S_ISREG(VFS_I(ip)->i_mode); + int error = 0; + + if (xfs_is_quota_inode(&tp->t_mountp->m_sb, ip->i_ino)) { + /* + * Quota files are never counted towards quota, so we do not + * need to take the lock. + */ + xchk_iscan_mark_visited(&xqc->iscan, ip); + return 0; + } + + /* Figure out the data / rt device block counts. */ + xfs_ilock(ip, XFS_IOLOCK_SHARED); + if (isreg) + xfs_ilock(ip, XFS_MMAPLOCK_SHARED); + if (XFS_IS_REALTIME_INODE(ip)) { + /* + * Read in the data fork for rt files so that _count_blocks + * can count the number of blocks allocated from the rt volume. + * Inodes do not track that separately. + */ + ilock_flags = xfs_ilock_data_map_shared(ip); + error = xfs_iread_extents(tp, ip, XFS_DATA_FORK); + if (error) + goto out_abort; + } else { + ilock_flags = XFS_ILOCK_SHARED; + xfs_ilock(ip, XFS_ILOCK_SHARED); + } + xfs_inode_count_blocks(tp, ip, &nblks, &rtblks); + + if (xchk_iscan_aborted(&xqc->iscan)) { + error = -ECANCELED; + goto out_incomplete; + } + + /* Update the shadow dquot counters. */ + mutex_lock(&xqc->lock); + if (xqc->ucounts) { + id = xfs_qm_id_for_quotatype(ip, XFS_DQTYPE_USER); + error = xqcheck_update_incore_counts(xqc, xqc->ucounts, id, 1, + nblks, rtblks); + if (error) + goto out_mutex; + } + + if (xqc->gcounts) { + id = xfs_qm_id_for_quotatype(ip, XFS_DQTYPE_GROUP); + error = xqcheck_update_incore_counts(xqc, xqc->gcounts, id, 1, + nblks, rtblks); + if (error) + goto out_mutex; + } + + if (xqc->pcounts) { + id = xfs_qm_id_for_quotatype(ip, XFS_DQTYPE_PROJ); + error = xqcheck_update_incore_counts(xqc, xqc->pcounts, id, 1, + nblks, rtblks); + if (error) + goto out_mutex; + } + mutex_unlock(&xqc->lock); + + xchk_iscan_mark_visited(&xqc->iscan, ip); + goto out_ilock; + +out_mutex: + mutex_unlock(&xqc->lock); +out_abort: + xchk_iscan_abort(&xqc->iscan); +out_incomplete: + xchk_set_incomplete(xqc->sc); +out_ilock: + xfs_iunlock(ip, ilock_flags); + if (isreg) + xfs_iunlock(ip, XFS_MMAPLOCK_SHARED); + xfs_iunlock(ip, XFS_IOLOCK_SHARED); + return error; +} + +/* Walk all the allocated inodes and run a quota scan on them. */ +STATIC int +xqcheck_collect_counts( + struct xqcheck *xqc) +{ + struct xfs_scrub *sc = xqc->sc; + struct xfs_inode *ip; + int error; + + /* + * Set up for a potentially lengthy filesystem scan by reducing our + * transaction resource usage for the duration. Specifically: + * + * Cancel the transaction to release the log grant space while we scan + * the filesystem. + * + * Create a new empty transaction to eliminate the possibility of the + * inode scan deadlocking on cyclical metadata. + * + * We pass the empty transaction to the file scanning function to avoid + * repeatedly cycling empty transactions. This can be done without + * risk of deadlock between sb_internal and the IOLOCK (we take the + * IOLOCK to quiesce the file before scanning) because empty + * transactions do not take sb_internal. + */ + xchk_trans_cancel(sc); + error = xchk_trans_alloc_empty(sc); + if (error) + return error; + + while ((error = xchk_iscan_iter(&xqc->iscan, &ip)) == 1) { + error = xqcheck_collect_inode(xqc, ip); + xchk_irele(sc, ip); + if (error) + break; + + if (xchk_should_terminate(sc, &error)) + break; + } + xchk_iscan_iter_finish(&xqc->iscan); + if (error) { + xchk_set_incomplete(sc); + /* + * If we couldn't grab an inode that was busy with a state + * change, change the error code so that we exit to userspace + * as quickly as possible. + */ + if (error == -EBUSY) + return -ECANCELED; + return error; + } + + /* + * Switch out for a real transaction in preparation for building a new + * tree. + */ + xchk_trans_cancel(sc); + return xchk_setup_fs(sc); +} + +/* + * Part 2: Comparing dquot resource counters. Walk each xfs_dquot, comparing + * the resource usage counters against our shadow dquots; and then walk each + * shadow dquot (that wasn't covered in the first part), comparing it against + * the xfs_dquot. + */ + +/* + * Check the dquot data against what we observed. Caller must hold the dquot + * lock. + */ +STATIC int +xqcheck_compare_dquot( + struct xqcheck *xqc, + xfs_dqtype_t dqtype, + struct xfs_dquot *dq) +{ + struct xqcheck_dquot xcdq; + struct xfarray *counts = xqcheck_counters_for(xqc, dqtype); + int error; + + if (xchk_iscan_aborted(&xqc->iscan)) { + xchk_set_incomplete(xqc->sc); + return -ECANCELED; + } + + mutex_lock(&xqc->lock); + error = xfarray_load_sparse(counts, dq->q_id, &xcdq); + if (error) + goto out_unlock; + + if (xcdq.icount != dq->q_ino.count) + xchk_qcheck_set_corrupt(xqc->sc, dqtype, dq->q_id); + + if (xcdq.bcount != dq->q_blk.count) + xchk_qcheck_set_corrupt(xqc->sc, dqtype, dq->q_id); + + if (xcdq.rtbcount != dq->q_rtb.count) + xchk_qcheck_set_corrupt(xqc->sc, dqtype, dq->q_id); + + xcdq.flags |= (XQCHECK_DQUOT_COMPARE_SCANNED | XQCHECK_DQUOT_WRITTEN); + error = xfarray_store(counts, dq->q_id, &xcdq); + if (error == -EFBIG) { + /* + * EFBIG means we tried to store data at too high a byte offset + * in the sparse array. IOWs, we cannot complete the check and + * must notify userspace that the check was incomplete. This + * should never happen outside of the collection phase. + */ + xchk_set_incomplete(xqc->sc); + error = -ECANCELED; + } + mutex_unlock(&xqc->lock); + if (error) + return error; + + if (xqc->sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT) + return -ECANCELED; + + return 0; + +out_unlock: + mutex_unlock(&xqc->lock); + return error; +} + +/* + * Walk all the observed dquots, and make sure there's a matching incore + * dquot and that its counts match ours. + */ +STATIC int +xqcheck_walk_observations( + struct xqcheck *xqc, + xfs_dqtype_t dqtype) +{ + struct xqcheck_dquot xcdq; + struct xfs_dquot *dq; + struct xfarray *counts = xqcheck_counters_for(xqc, dqtype); + xfarray_idx_t cur = XFARRAY_CURSOR_INIT; + int error; + + mutex_lock(&xqc->lock); + while ((error = xfarray_iter(counts, &cur, &xcdq)) == 1) { + xfs_dqid_t id = cur - 1; + + if (xcdq.flags & XQCHECK_DQUOT_COMPARE_SCANNED) + continue; + + mutex_unlock(&xqc->lock); + + error = xfs_qm_dqget(xqc->sc->mp, id, dqtype, false, &dq); + if (error == -ENOENT) { + xchk_qcheck_set_corrupt(xqc->sc, dqtype, id); + return 0; + } + if (error) + return error; + + error = xqcheck_compare_dquot(xqc, dqtype, dq); + xfs_qm_dqput(dq); + if (error) + return error; + + if (xchk_should_terminate(xqc->sc, &error)) + return error; + + mutex_lock(&xqc->lock); + } + mutex_unlock(&xqc->lock); + + return error; +} + +/* Compare the quota counters we observed against the live dquots. */ +STATIC int +xqcheck_compare_dqtype( + struct xqcheck *xqc, + xfs_dqtype_t dqtype) +{ + struct xchk_dqiter cursor = { }; + struct xfs_scrub *sc = xqc->sc; + struct xfs_dquot *dq; + int error; + + if (sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT) + return 0; + + /* If the quota CHKD flag is cleared, we need to repair this quota. */ + if (!(xfs_quota_chkd_flag(dqtype) & sc->mp->m_qflags)) { + xchk_qcheck_set_corrupt(xqc->sc, dqtype, 0); + return 0; + } + + /* Compare what we observed against the actual dquots. */ + xchk_dqiter_init(&cursor, sc, dqtype); + while ((error = xchk_dquot_iter(&cursor, &dq)) == 1) { + error = xqcheck_compare_dquot(xqc, dqtype, dq); + xfs_qm_dqput(dq); + if (error) + break; + } + if (error) + return error; + + /* Walk all the observed dquots and compare to the incore ones. */ + return xqcheck_walk_observations(xqc, dqtype); +} + +/* Tear down everything associated with a quotacheck. */ +static void +xqcheck_teardown_scan( + void *priv) +{ + struct xqcheck *xqc = priv; + struct xfs_quotainfo *qi = xqc->sc->mp->m_quotainfo; + + /* Discourage any hook functions that might be running. */ + xchk_iscan_abort(&xqc->iscan); + + /* + * As noted above, the apply hook is responsible for cleaning up the + * shadow dquot accounting data when a transaction completes. The mod + * hook must be removed before the apply hook so that we don't + * mistakenly leave an active shadow account for the mod hook to get + * its hands on. No hooks should be running after these functions + * return. + */ + xfs_dqtrx_hook_del(qi, &xqc->qhook); + + if (xqc->shadow_dquot_acct.key_len) { + rhashtable_free_and_destroy(&xqc->shadow_dquot_acct, + xqcheck_dqacct_free, NULL); + xqc->shadow_dquot_acct.key_len = 0; + } + + if (xqc->pcounts) { + xfarray_destroy(xqc->pcounts); + xqc->pcounts = NULL; + } + + if (xqc->gcounts) { + xfarray_destroy(xqc->gcounts); + xqc->gcounts = NULL; + } + + if (xqc->ucounts) { + xfarray_destroy(xqc->ucounts); + xqc->ucounts = NULL; + } + + xchk_iscan_teardown(&xqc->iscan); + mutex_destroy(&xqc->lock); + xqc->sc = NULL; +} + +/* + * Scan all inodes in the entire filesystem to generate quota counter data. + * If the scan is successful, the quota data will be left alive for a repair. + * If any error occurs, we'll tear everything down. + */ +STATIC int +xqcheck_setup_scan( + struct xfs_scrub *sc, + struct xqcheck *xqc) +{ + char *descr; + struct xfs_quotainfo *qi = sc->mp->m_quotainfo; + unsigned long long max_dquots = XFS_DQ_ID_MAX + 1ULL; + int error; + + ASSERT(xqc->sc == NULL); + xqc->sc = sc; + + mutex_init(&xqc->lock); + + /* Retry iget every tenth of a second for up to 30 seconds. */ + xchk_iscan_start(sc, 30000, 100, &xqc->iscan); + + error = -ENOMEM; + if (xfs_this_quota_on(sc->mp, XFS_DQTYPE_USER)) { + descr = xchk_xfile_descr(sc, "user dquot records"); + error = xfarray_create(descr, max_dquots, + sizeof(struct xqcheck_dquot), &xqc->ucounts); + kfree(descr); + if (error) + goto out_teardown; + } + + if (xfs_this_quota_on(sc->mp, XFS_DQTYPE_GROUP)) { + descr = xchk_xfile_descr(sc, "group dquot records"); + error = xfarray_create(descr, max_dquots, + sizeof(struct xqcheck_dquot), &xqc->gcounts); + kfree(descr); + if (error) + goto out_teardown; + } + + if (xfs_this_quota_on(sc->mp, XFS_DQTYPE_PROJ)) { + descr = xchk_xfile_descr(sc, "project dquot records"); + error = xfarray_create(descr, max_dquots, + sizeof(struct xqcheck_dquot), &xqc->pcounts); + kfree(descr); + if (error) + goto out_teardown; + } + + /* + * Set up hash table to map transactions to our internal shadow dqtrx + * structures. + */ + error = rhashtable_init(&xqc->shadow_dquot_acct, + &xqcheck_dqacct_hash_params); + if (error) + goto out_teardown; + + /* + * Hook into the quota code. The hook only triggers for inodes that + * were already scanned, and the scanner thread takes each inode's + * ILOCK, which means that any in-progress inode updates will finish + * before we can scan the inode. + * + * The apply hook (which removes the shadow dquot accounting struct) + * must be installed before the mod hook so that we never fail to catch + * the end of a quota update sequence and leave stale shadow data. + */ + ASSERT(sc->flags & XCHK_FSGATES_QUOTA); + xfs_dqtrx_hook_setup(&xqc->qhook, xqcheck_mod_live_ino_dqtrx, + xqcheck_apply_live_dqtrx); + + error = xfs_dqtrx_hook_add(qi, &xqc->qhook); + if (error) + goto out_teardown; + + /* Use deferred cleanup to pass the quota count data to repair. */ + sc->buf_cleanup = xqcheck_teardown_scan; + return 0; + +out_teardown: + xqcheck_teardown_scan(xqc); + return error; +} + +/* Scrub all counters for a given quota type. */ +int +xchk_quotacheck( + struct xfs_scrub *sc) +{ + struct xqcheck *xqc = sc->buf; + int error = 0; + + /* Check quota counters on the live filesystem. */ + error = xqcheck_setup_scan(sc, xqc); + if (error) + return error; + + /* Walk all inodes, picking up quota information. */ + error = xqcheck_collect_counts(xqc); + if (!xchk_xref_process_error(sc, 0, 0, &error)) + return error; + + /* Fail fast if we're not playing with a full dataset. */ + if (xchk_iscan_aborted(&xqc->iscan)) + xchk_set_incomplete(sc); + if (sc->sm->sm_flags & XFS_SCRUB_OFLAG_INCOMPLETE) + return 0; + + /* Compare quota counters. */ + if (xqc->ucounts) { + error = xqcheck_compare_dqtype(xqc, XFS_DQTYPE_USER); + if (!xchk_xref_process_error(sc, 0, 0, &error)) + return error; + } + if (xqc->gcounts) { + error = xqcheck_compare_dqtype(xqc, XFS_DQTYPE_GROUP); + if (!xchk_xref_process_error(sc, 0, 0, &error)) + return error; + } + if (xqc->pcounts) { + error = xqcheck_compare_dqtype(xqc, XFS_DQTYPE_PROJ); + if (!xchk_xref_process_error(sc, 0, 0, &error)) + return error; + } + + /* Check one last time for an incomplete dataset. */ + if (xchk_iscan_aborted(&xqc->iscan)) + xchk_set_incomplete(sc); + + return 0; +} diff --git a/fs/xfs/scrub/quotacheck.h b/fs/xfs/scrub/quotacheck.h new file mode 100644 index 000000000000..4ea5f249c978 --- /dev/null +++ b/fs/xfs/scrub/quotacheck.h @@ -0,0 +1,76 @@ +/* SPDX-License-Identifier: GPL-2.0-or-later */ +/* + * Copyright (c) 2020-2024 Oracle. All Rights Reserved. + * Author: Darrick J. Wong <djwong@kernel.org> + */ +#ifndef __XFS_SCRUB_QUOTACHECK_H__ +#define __XFS_SCRUB_QUOTACHECK_H__ + +/* Quota counters for live quotacheck. */ +struct xqcheck_dquot { + /* block usage count */ + int64_t bcount; + + /* inode usage count */ + int64_t icount; + + /* realtime block usage count */ + int64_t rtbcount; + + /* Record state */ + unsigned int flags; +}; + +/* + * This incore dquot record has been written at least once. We never want to + * store an xqcheck_dquot that looks uninitialized. + */ +#define XQCHECK_DQUOT_WRITTEN (1U << 0) + +/* Already checked this dquot. */ +#define XQCHECK_DQUOT_COMPARE_SCANNED (1U << 1) + +/* Already repaired this dquot. */ +#define XQCHECK_DQUOT_REPAIR_SCANNED (1U << 2) + +/* Live quotacheck control structure. */ +struct xqcheck { + struct xfs_scrub *sc; + + /* Shadow dquot counter data. */ + struct xfarray *ucounts; + struct xfarray *gcounts; + struct xfarray *pcounts; + + /* Lock protecting quotacheck count observations */ + struct mutex lock; + + struct xchk_iscan iscan; + + /* Hooks into the quota code. */ + struct xfs_dqtrx_hook qhook; + + /* Shadow quota delta tracking structure. */ + struct rhashtable shadow_dquot_acct; +}; + +/* Return the incore counter array for a given quota type. */ +static inline struct xfarray * +xqcheck_counters_for( + struct xqcheck *xqc, + xfs_dqtype_t dqtype) +{ + switch (dqtype) { + case XFS_DQTYPE_USER: + return xqc->ucounts; + case XFS_DQTYPE_GROUP: + return xqc->gcounts; + case XFS_DQTYPE_PROJ: + return xqc->pcounts; + } + + ASSERT(0); + return NULL; +} + +#endif /* __XFS_SCRUB_QUOTACHECK_H__ */ diff --git a/fs/xfs/scrub/quotacheck_repair.c b/fs/xfs/scrub/quotacheck_repair.c new file mode 100644 index 000000000000..dd8554c755b5 --- /dev/null +++ b/fs/xfs/scrub/quotacheck_repair.c @@ -0,0 +1,261 @@ +// SPDX-License-Identifier: GPL-2.0-or-later +/* + * Copyright (c) 2020-2024 Oracle. All Rights Reserved. + * Author: Darrick J. Wong <djwong@kernel.org> + */ +#include "xfs.h" +#include "xfs_fs.h" +#include "xfs_shared.h" +#include "xfs_format.h" +#include "xfs_trans_resv.h" +#include "xfs_mount.h" +#include "xfs_log_format.h" +#include "xfs_trans.h" +#include "xfs_inode.h" +#include "xfs_quota.h" +#include "xfs_qm.h" +#include "xfs_icache.h" +#include "xfs_bmap_util.h" +#include "xfs_iwalk.h" +#include "xfs_ialloc.h" +#include "xfs_sb.h" +#include "scrub/scrub.h" +#include "scrub/common.h" +#include "scrub/repair.h" +#include "scrub/xfile.h" +#include "scrub/xfarray.h" +#include "scrub/iscan.h" +#include "scrub/quota.h" +#include "scrub/quotacheck.h" +#include "scrub/trace.h" + +/* + * Live Quotacheck Repair + * ====================== + * + * Use the live quota counter information that we collected to replace the + * counter values in the incore dquots. A scrub->repair cycle should have left + * the live data and hooks active, so this is safe so long as we make sure the + * dquot is locked. + */ + +/* Commit new counters to a dquot. */ +static int +xqcheck_commit_dquot( + struct xqcheck *xqc, + xfs_dqtype_t dqtype, + struct xfs_dquot *dq) +{ + struct xqcheck_dquot xcdq; + struct xfarray *counts = xqcheck_counters_for(xqc, dqtype); + int64_t delta; + bool dirty = false; + int error = 0; + + /* Unlock the dquot just long enough to allocate a transaction. */ + xfs_dqunlock(dq); + error = xchk_trans_alloc(xqc->sc, 0); + xfs_dqlock(dq); + if (error) + return error; + + xfs_trans_dqjoin(xqc->sc->tp, dq); + + if (xchk_iscan_aborted(&xqc->iscan)) { + error = -ECANCELED; + goto out_cancel; + } + + mutex_lock(&xqc->lock); + error = xfarray_load_sparse(counts, dq->q_id, &xcdq); + if (error) + goto out_unlock; + + /* Adjust counters as needed. */ + delta = (int64_t)xcdq.icount - dq->q_ino.count; + if (delta) { + dq->q_ino.reserved += delta; + dq->q_ino.count += delta; + dirty = true; + } + + delta = (int64_t)xcdq.bcount - dq->q_blk.count; + if (delta) { + dq->q_blk.reserved += delta; + dq->q_blk.count += delta; + dirty = true; + } + + delta = (int64_t)xcdq.rtbcount - dq->q_rtb.count; + if (delta) { + dq->q_rtb.reserved += delta; + dq->q_rtb.count += delta; + dirty = true; + } + + xcdq.flags |= (XQCHECK_DQUOT_REPAIR_SCANNED | XQCHECK_DQUOT_WRITTEN); + error = xfarray_store(counts, dq->q_id, &xcdq); + if (error == -EFBIG) { + /* + * EFBIG means we tried to store data at too high a byte offset + * in the sparse array. IOWs, we cannot complete the repair + * and must cancel the whole operation. This should never + * happen, but we need to catch it anyway. + */ + error = -ECANCELED; + } + mutex_unlock(&xqc->lock); + if (error || !dirty) + goto out_cancel; + + trace_xrep_quotacheck_dquot(xqc->sc->mp, dq->q_type, dq->q_id); + + /* Commit the dirty dquot to disk. */ + dq->q_flags |= XFS_DQFLAG_DIRTY; + if (dq->q_id) + xfs_qm_adjust_dqtimers(dq); + xfs_trans_log_dquot(xqc->sc->tp, dq); + + /* + * Transaction commit unlocks the dquot, so we must re-lock it so that + * the caller can put the reference (which apparently requires a locked + * dquot). + */ + error = xrep_trans_commit(xqc->sc); + xfs_dqlock(dq); + return error; + +out_unlock: + mutex_unlock(&xqc->lock); +out_cancel: + xchk_trans_cancel(xqc->sc); + + /* Re-lock the dquot so the caller can put the reference. */ + xfs_dqlock(dq); + return error; +} + +/* Commit new quota counters for a particular quota type. */ +STATIC int +xqcheck_commit_dqtype( + struct xqcheck *xqc, + unsigned int dqtype) +{ + struct xchk_dqiter cursor = { }; + struct xqcheck_dquot xcdq; + struct xfs_scrub *sc = xqc->sc; + struct xfs_mount *mp = sc->mp; + struct xfarray *counts = xqcheck_counters_for(xqc, dqtype); + struct xfs_dquot *dq; + xfarray_idx_t cur = XFARRAY_CURSOR_INIT; + int error; + + /* + * Update the counters of every dquot that the quota file knows about. + */ + xchk_dqiter_init(&cursor, sc, dqtype); + while ((error = xchk_dquot_iter(&cursor, &dq)) == 1) { + error = xqcheck_commit_dquot(xqc, dqtype, dq); + xfs_qm_dqput(dq); + if (error) + break; + } + if (error) + return error; + + /* + * Make a second pass to deal with the dquots that we know about but + * the quota file previously did not know about. + */ + mutex_lock(&xqc->lock); + while ((error = xfarray_iter(counts, &cur, &xcdq)) == 1) { + xfs_dqid_t id = cur - 1; + + if (xcdq.flags & XQCHECK_DQUOT_REPAIR_SCANNED) + continue; + + mutex_unlock(&xqc->lock); + + /* + * Grab the dquot, allowing for dquot block allocation in a + * separate transaction. We committed the scrub transaction + * in a previous step, so we will not be creating nested + * transactions here. + */ + error = xfs_qm_dqget(mp, id, dqtype, true, &dq); + if (error) + return error; + + error = xqcheck_commit_dquot(xqc, dqtype, dq); + xfs_qm_dqput(dq); + if (error) + return error; + + mutex_lock(&xqc->lock); + } + mutex_unlock(&xqc->lock); + + return error; +} + +/* Figure out quota CHKD flags for the running quota types. */ +static inline unsigned int +xqcheck_chkd_flags( + struct xfs_mount *mp) +{ + unsigned int ret = 0; + + if (XFS_IS_UQUOTA_ON(mp)) + ret |= XFS_UQUOTA_CHKD; + if (XFS_IS_GQUOTA_ON(mp)) + ret |= XFS_GQUOTA_CHKD; + if (XFS_IS_PQUOTA_ON(mp)) + ret |= XFS_PQUOTA_CHKD; + return ret; +} + +/* Commit the new dquot counters. */ +int +xrep_quotacheck( + struct xfs_scrub *sc) +{ + struct xqcheck *xqc = sc->buf; + unsigned int qflags = xqcheck_chkd_flags(sc->mp); + int error; + + /* + * Clear the CHKD flag for the running quota types and commit the scrub + * transaction so that we can allocate new quota block mappings if we + * have to. If we crash after this point, the sb still has the CHKD + * flags cleared, so mount quotacheck will fix all of this up. + */ + xrep_update_qflags(sc, qflags, 0); + error = xrep_trans_commit(sc); + if (error) + return error; + + /* Commit the new counters to the dquots. */ + if (xqc->ucounts) { + error = xqcheck_commit_dqtype(xqc, XFS_DQTYPE_USER); + if (error) + return error; + } + if (xqc->gcounts) { + error = xqcheck_commit_dqtype(xqc, XFS_DQTYPE_GROUP); + if (error) + return error; + } + if (xqc->pcounts) { + error = xqcheck_commit_dqtype(xqc, XFS_DQTYPE_PROJ); + if (error) + return error; + } + + /* Set the CHKD flags now that we've fixed quota counts. */ + error = xchk_trans_alloc(sc, 0); + if (error) + return error; + + xrep_update_qflags(sc, 0, qflags); + return xrep_trans_commit(sc); +} diff --git a/fs/xfs/scrub/rcbag.c b/fs/xfs/scrub/rcbag.c new file mode 100644 index 000000000000..e1e52bc20713 --- /dev/null +++ b/fs/xfs/scrub/rcbag.c @@ -0,0 +1,307 @@ +// SPDX-License-Identifier: GPL-2.0-or-later +/* + * Copyright (c) 2022-2024 Oracle. All Rights Reserved. + * Author: Darrick J. Wong <djwong@kernel.org> + */ +#include "xfs.h" +#include "xfs_fs.h" +#include "xfs_shared.h" +#include "xfs_format.h" +#include "xfs_log_format.h" +#include "xfs_trans.h" +#include "xfs_trans_resv.h" +#include "xfs_mount.h" +#include "xfs_defer.h" +#include "xfs_btree.h" +#include "xfs_buf_mem.h" +#include "xfs_btree_mem.h" +#include "xfs_error.h" +#include "scrub/scrub.h" +#include "scrub/rcbag_btree.h" +#include "scrub/rcbag.h" +#include "scrub/trace.h" + +struct rcbag { + struct xfs_mount *mp; + struct xfbtree xfbtree; + uint64_t nr_items; +}; + +int +rcbag_init( + struct xfs_mount *mp, + struct xfs_buftarg *btp, + struct rcbag **bagp) +{ + struct rcbag *bag; + int error; + + bag = kzalloc(sizeof(struct rcbag), XCHK_GFP_FLAGS); + if (!bag) + return -ENOMEM; + + bag->nr_items = 0; + bag->mp = mp; + + error = rcbagbt_mem_init(mp, &bag->xfbtree, btp); + if (error) + goto out_bag; + + *bagp = bag; + return 0; + +out_bag: + kfree(bag); + return error; +} + +void +rcbag_free( + struct rcbag **bagp) +{ + struct rcbag *bag = *bagp; + + xfbtree_destroy(&bag->xfbtree); + kfree(bag); + *bagp = NULL; +} + +/* Track an rmap in the refcount bag. */ +int +rcbag_add( + struct rcbag *bag, + struct xfs_trans *tp, + const struct xfs_rmap_irec *rmap) +{ + struct rcbag_rec bagrec; + struct xfs_mount *mp = bag->mp; + struct xfs_btree_cur *cur; + int has; + int error; + + cur = rcbagbt_mem_cursor(mp, tp, &bag->xfbtree); + error = rcbagbt_lookup_eq(cur, rmap, &has); + if (error) + goto out_cur; + + if (has) { + error = rcbagbt_get_rec(cur, &bagrec, &has); + if (error) + goto out_cur; + if (!has) { + error = -EFSCORRUPTED; + goto out_cur; + } + + bagrec.rbg_refcount++; + error = rcbagbt_update(cur, &bagrec); + if (error) + goto out_cur; + } else { + bagrec.rbg_startblock = rmap->rm_startblock; + bagrec.rbg_blockcount = rmap->rm_blockcount; + bagrec.rbg_refcount = 1; + + error = rcbagbt_insert(cur, &bagrec, &has); + if (error) + goto out_cur; + if (!has) { + error = -EFSCORRUPTED; + goto out_cur; + } + } + + xfs_btree_del_cursor(cur, 0); + + error = xfbtree_trans_commit(&bag->xfbtree, tp); + if (error) + return error; + + bag->nr_items++; + return 0; + +out_cur: + xfs_btree_del_cursor(cur, error); + xfbtree_trans_cancel(&bag->xfbtree, tp); + return error; +} + +/* Return the number of records in the bag. */ +uint64_t +rcbag_count( + const struct rcbag *rcbag) +{ + return rcbag->nr_items; +} + +static inline uint32_t rcbag_rec_next_bno(const struct rcbag_rec *r) +{ + return r->rbg_startblock + r->rbg_blockcount; +} + +/* + * Find the next block where the refcount changes, given the next rmap we + * looked at and the ones we're already tracking. + */ +int +rcbag_next_edge( + struct rcbag *bag, + struct xfs_trans *tp, + const struct xfs_rmap_irec *next_rmap, + bool next_valid, + uint32_t *next_bnop) +{ + struct rcbag_rec bagrec; + struct xfs_mount *mp = bag->mp; + struct xfs_btree_cur *cur; + uint32_t next_bno = NULLAGBLOCK; + int has; + int error; + + if (next_valid) + next_bno = next_rmap->rm_startblock; + + cur = rcbagbt_mem_cursor(mp, tp, &bag->xfbtree); + error = xfs_btree_goto_left_edge(cur); + if (error) + goto out_cur; + + while (true) { + error = xfs_btree_increment(cur, 0, &has); + if (error) + goto out_cur; + if (!has) + break; + + error = rcbagbt_get_rec(cur, &bagrec, &has); + if (error) + goto out_cur; + if (!has) { + error = -EFSCORRUPTED; + goto out_cur; + } + + next_bno = min(next_bno, rcbag_rec_next_bno(&bagrec)); + } + + /* + * We should have found /something/ because either next_rrm is the next + * interesting rmap to look at after emitting this refcount extent, or + * there are other rmaps in rmap_bag contributing to the current + * sharing count. But if something is seriously wrong, bail out. + */ + if (next_bno == NULLAGBLOCK) { + error = -EFSCORRUPTED; + goto out_cur; + } + + xfs_btree_del_cursor(cur, 0); + + *next_bnop = next_bno; + return 0; + +out_cur: + xfs_btree_del_cursor(cur, error); + return error; +} + +/* Pop all refcount bag records that end at next_bno */ +int +rcbag_remove_ending_at( + struct rcbag *bag, + struct xfs_trans *tp, + uint32_t next_bno) +{ + struct rcbag_rec bagrec; + struct xfs_mount *mp = bag->mp; + struct xfs_btree_cur *cur; + int has; + int error; + + /* go to the right edge of the tree */ + cur = rcbagbt_mem_cursor(mp, tp, &bag->xfbtree); + memset(&cur->bc_rec, 0xFF, sizeof(cur->bc_rec)); + error = xfs_btree_lookup(cur, XFS_LOOKUP_GE, &has); + if (error) + goto out_cur; + + while (true) { + error = xfs_btree_decrement(cur, 0, &has); + if (error) + goto out_cur; + if (!has) + break; + + error = rcbagbt_get_rec(cur, &bagrec, &has); + if (error) + goto out_cur; + if (!has) { + error = -EFSCORRUPTED; + goto out_cur; + } + + if (rcbag_rec_next_bno(&bagrec) != next_bno) + continue; + + error = xfs_btree_delete(cur, &has); + if (error) + goto out_cur; + if (!has) { + error = -EFSCORRUPTED; + goto out_cur; + } + + bag->nr_items -= bagrec.rbg_refcount; + } + + xfs_btree_del_cursor(cur, 0); + return xfbtree_trans_commit(&bag->xfbtree, tp); +out_cur: + xfs_btree_del_cursor(cur, error); + xfbtree_trans_cancel(&bag->xfbtree, tp); + return error; +} + +/* Dump the rcbag. */ +void +rcbag_dump( + struct rcbag *bag, + struct xfs_trans *tp) +{ + struct rcbag_rec bagrec; + struct xfs_mount *mp = bag->mp; + struct xfs_btree_cur *cur; + unsigned long long nr = 0; + int has; + int error; + + cur = rcbagbt_mem_cursor(mp, tp, &bag->xfbtree); + error = xfs_btree_goto_left_edge(cur); + if (error) + goto out_cur; + + while (true) { + error = xfs_btree_increment(cur, 0, &has); + if (error) + goto out_cur; + if (!has) + break; + + error = rcbagbt_get_rec(cur, &bagrec, &has); + if (error) + goto out_cur; + if (!has) { + error = -EFSCORRUPTED; + goto out_cur; + } + + xfs_err(bag->mp, "[%llu]: bno 0x%x fsbcount 0x%x refcount 0x%llx\n", + nr++, + (unsigned int)bagrec.rbg_startblock, + (unsigned int)bagrec.rbg_blockcount, + (unsigned long long)bagrec.rbg_refcount); + } + +out_cur: + xfs_btree_del_cursor(cur, error); +} diff --git a/fs/xfs/scrub/rcbag.h b/fs/xfs/scrub/rcbag.h new file mode 100644 index 000000000000..e29ef788ba72 --- /dev/null +++ b/fs/xfs/scrub/rcbag.h @@ -0,0 +1,28 @@ +// SPDX-License-Identifier: GPL-2.0-or-later +/* + * Copyright (c) 2022-2024 Oracle. All Rights Reserved. + * Author: Darrick J. Wong <djwong@kernel.org> + */ +#ifndef __XFS_SCRUB_RCBAG_H__ +#define __XFS_SCRUB_RCBAG_H__ + +struct xfs_mount; +struct rcbag; +struct xfs_buftarg; + +int rcbag_init(struct xfs_mount *mp, struct xfs_buftarg *btp, + struct rcbag **bagp); +void rcbag_free(struct rcbag **bagp); +int rcbag_add(struct rcbag *bag, struct xfs_trans *tp, + const struct xfs_rmap_irec *rmap); +uint64_t rcbag_count(const struct rcbag *bag); + +int rcbag_next_edge(struct rcbag *bag, struct xfs_trans *tp, + const struct xfs_rmap_irec *next_rmap, bool next_valid, + uint32_t *next_bnop); +int rcbag_remove_ending_at(struct rcbag *bag, struct xfs_trans *tp, + uint32_t next_bno); + +void rcbag_dump(struct rcbag *bag, struct xfs_trans *tp); + +#endif /* __XFS_SCRUB_RCBAG_H__ */ diff --git a/fs/xfs/scrub/rcbag_btree.c b/fs/xfs/scrub/rcbag_btree.c new file mode 100644 index 000000000000..709356dc6256 --- /dev/null +++ b/fs/xfs/scrub/rcbag_btree.c @@ -0,0 +1,370 @@ +// SPDX-License-Identifier: GPL-2.0-or-later +/* + * Copyright (c) 2022-2024 Oracle. All Rights Reserved. + * Author: Darrick J. Wong <djwong@kernel.org> + */ +#include "xfs.h" +#include "xfs_fs.h" +#include "xfs_shared.h" +#include "xfs_format.h" +#include "xfs_trans_resv.h" +#include "xfs_mount.h" +#include "xfs_defer.h" +#include "xfs_btree.h" +#include "xfs_buf_mem.h" +#include "xfs_btree_mem.h" +#include "xfs_error.h" +#include "scrub/rcbag_btree.h" +#include "scrub/trace.h" + +static struct kmem_cache *rcbagbt_cur_cache; + +STATIC void +rcbagbt_init_key_from_rec( + union xfs_btree_key *key, + const union xfs_btree_rec *rec) +{ + struct rcbag_key *bag_key = (struct rcbag_key *)key; + const struct rcbag_rec *bag_rec = (const struct rcbag_rec *)rec; + + BUILD_BUG_ON(sizeof(struct rcbag_key) > sizeof(union xfs_btree_key)); + BUILD_BUG_ON(sizeof(struct rcbag_rec) > sizeof(union xfs_btree_rec)); + + bag_key->rbg_startblock = bag_rec->rbg_startblock; + bag_key->rbg_blockcount = bag_rec->rbg_blockcount; +} + +STATIC void +rcbagbt_init_rec_from_cur( + struct xfs_btree_cur *cur, + union xfs_btree_rec *rec) +{ + struct rcbag_rec *bag_rec = (struct rcbag_rec *)rec; + struct rcbag_rec *bag_irec = (struct rcbag_rec *)&cur->bc_rec; + + bag_rec->rbg_startblock = bag_irec->rbg_startblock; + bag_rec->rbg_blockcount = bag_irec->rbg_blockcount; + bag_rec->rbg_refcount = bag_irec->rbg_refcount; +} + +STATIC int64_t +rcbagbt_key_diff( + struct xfs_btree_cur *cur, + const union xfs_btree_key *key) +{ + struct rcbag_rec *rec = (struct rcbag_rec *)&cur->bc_rec; + const struct rcbag_key *kp = (const struct rcbag_key *)key; + + if (kp->rbg_startblock > rec->rbg_startblock) + return 1; + if (kp->rbg_startblock < rec->rbg_startblock) + return -1; + + if (kp->rbg_blockcount > rec->rbg_blockcount) + return 1; + if (kp->rbg_blockcount < rec->rbg_blockcount) + return -1; + + return 0; +} + +STATIC int64_t +rcbagbt_diff_two_keys( + struct xfs_btree_cur *cur, + const union xfs_btree_key *k1, + const union xfs_btree_key *k2, + const union xfs_btree_key *mask) +{ + const struct rcbag_key *kp1 = (const struct rcbag_key *)k1; + const struct rcbag_key *kp2 = (const struct rcbag_key *)k2; + + ASSERT(mask == NULL); + + if (kp1->rbg_startblock > kp2->rbg_startblock) + return 1; + if (kp1->rbg_startblock < kp2->rbg_startblock) + return -1; + + if (kp1->rbg_blockcount > kp2->rbg_blockcount) + return 1; + if (kp1->rbg_blockcount < kp2->rbg_blockcount) + return -1; + + return 0; +} + +STATIC int +rcbagbt_keys_inorder( + struct xfs_btree_cur *cur, + const union xfs_btree_key *k1, + const union xfs_btree_key *k2) +{ + const struct rcbag_key *kp1 = (const struct rcbag_key *)k1; + const struct rcbag_key *kp2 = (const struct rcbag_key *)k2; + + if (kp1->rbg_startblock > kp2->rbg_startblock) + return 0; + if (kp1->rbg_startblock < kp2->rbg_startblock) + return 1; + + if (kp1->rbg_blockcount > kp2->rbg_blockcount) + return 0; + if (kp1->rbg_blockcount < kp2->rbg_blockcount) + return 1; + + return 0; +} + +STATIC int +rcbagbt_recs_inorder( + struct xfs_btree_cur *cur, + const union xfs_btree_rec *r1, + const union xfs_btree_rec *r2) +{ + const struct rcbag_rec *rp1 = (const struct rcbag_rec *)r1; + const struct rcbag_rec *rp2 = (const struct rcbag_rec *)r2; + + if (rp1->rbg_startblock > rp2->rbg_startblock) + return 0; + if (rp1->rbg_startblock < rp2->rbg_startblock) + return 1; + + if (rp1->rbg_blockcount > rp2->rbg_blockcount) + return 0; + if (rp1->rbg_blockcount < rp2->rbg_blockcount) + return 1; + + return 0; +} + +static xfs_failaddr_t +rcbagbt_verify( + struct xfs_buf *bp) +{ + struct xfs_mount *mp = bp->b_mount; + struct xfs_btree_block *block = XFS_BUF_TO_BLOCK(bp); + xfs_failaddr_t fa; + unsigned int level; + unsigned int maxrecs; + + if (!xfs_verify_magic(bp, block->bb_magic)) + return __this_address; + + fa = xfs_btree_fsblock_v5hdr_verify(bp, XFS_RMAP_OWN_UNKNOWN); + if (fa) + return fa; + + level = be16_to_cpu(block->bb_level); + if (level >= rcbagbt_maxlevels_possible()) + return __this_address; + + maxrecs = rcbagbt_maxrecs(mp, XFBNO_BLOCKSIZE, level == 0); + return xfs_btree_memblock_verify(bp, maxrecs); +} + +static void +rcbagbt_rw_verify( + struct xfs_buf *bp) +{ + xfs_failaddr_t fa = rcbagbt_verify(bp); + + if (fa) + xfs_verifier_error(bp, -EFSCORRUPTED, fa); +} + +/* skip crc checks on in-memory btrees to save time */ +static const struct xfs_buf_ops rcbagbt_mem_buf_ops = { + .name = "rcbagbt_mem", + .magic = { 0, cpu_to_be32(RCBAG_MAGIC) }, + .verify_read = rcbagbt_rw_verify, + .verify_write = rcbagbt_rw_verify, + .verify_struct = rcbagbt_verify, +}; + +static const struct xfs_btree_ops rcbagbt_mem_ops = { + .name = "rcbag", + .type = XFS_BTREE_TYPE_MEM, + + .rec_len = sizeof(struct rcbag_rec), + .key_len = sizeof(struct rcbag_key), + .ptr_len = XFS_BTREE_LONG_PTR_LEN, + + .lru_refs = 1, + .statoff = XFS_STATS_CALC_INDEX(xs_rcbag_2), + + .dup_cursor = xfbtree_dup_cursor, + .set_root = xfbtree_set_root, + .alloc_block = xfbtree_alloc_block, + .free_block = xfbtree_free_block, + .get_minrecs = xfbtree_get_minrecs, + .get_maxrecs = xfbtree_get_maxrecs, + .init_key_from_rec = rcbagbt_init_key_from_rec, + .init_rec_from_cur = rcbagbt_init_rec_from_cur, + .init_ptr_from_cur = xfbtree_init_ptr_from_cur, + .key_diff = rcbagbt_key_diff, + .buf_ops = &rcbagbt_mem_buf_ops, + .diff_two_keys = rcbagbt_diff_two_keys, + .keys_inorder = rcbagbt_keys_inorder, + .recs_inorder = rcbagbt_recs_inorder, +}; + +/* Create a cursor for an in-memory btree. */ +struct xfs_btree_cur * +rcbagbt_mem_cursor( + struct xfs_mount *mp, + struct xfs_trans *tp, + struct xfbtree *xfbtree) +{ + struct xfs_btree_cur *cur; + + cur = xfs_btree_alloc_cursor(mp, tp, &rcbagbt_mem_ops, + rcbagbt_maxlevels_possible(), rcbagbt_cur_cache); + + cur->bc_mem.xfbtree = xfbtree; + cur->bc_nlevels = xfbtree->nlevels; + return cur; +} + +/* Create an in-memory refcount bag btree. */ +int +rcbagbt_mem_init( + struct xfs_mount *mp, + struct xfbtree *xfbt, + struct xfs_buftarg *btp) +{ + xfbt->owner = 0; + return xfbtree_init(mp, xfbt, btp, &rcbagbt_mem_ops); +} + +/* Calculate number of records in a refcount bag btree block. */ +static inline unsigned int +rcbagbt_block_maxrecs( + unsigned int blocklen, + bool leaf) +{ + if (leaf) + return blocklen / sizeof(struct rcbag_rec); + return blocklen / + (sizeof(struct rcbag_key) + sizeof(rcbag_ptr_t)); +} + +/* + * Calculate number of records in an refcount bag btree block. + */ +unsigned int +rcbagbt_maxrecs( + struct xfs_mount *mp, + unsigned int blocklen, + bool leaf) +{ + blocklen -= RCBAG_BLOCK_LEN; + return rcbagbt_block_maxrecs(blocklen, leaf); +} + +/* Compute the max possible height for refcount bag btrees. */ +unsigned int +rcbagbt_maxlevels_possible(void) +{ + unsigned int minrecs[2]; + unsigned int blocklen; + + blocklen = XFBNO_BLOCKSIZE - XFS_BTREE_LBLOCK_CRC_LEN; + + minrecs[0] = rcbagbt_block_maxrecs(blocklen, true) / 2; + minrecs[1] = rcbagbt_block_maxrecs(blocklen, false) / 2; + + return xfs_btree_space_to_height(minrecs, ULLONG_MAX); +} + +/* Calculate the refcount bag btree size for some records. */ +unsigned long long +rcbagbt_calc_size( + unsigned long long nr_records) +{ + unsigned int minrecs[2]; + unsigned int blocklen; + + blocklen = XFBNO_BLOCKSIZE - XFS_BTREE_LBLOCK_CRC_LEN; + + minrecs[0] = rcbagbt_block_maxrecs(blocklen, true) / 2; + minrecs[1] = rcbagbt_block_maxrecs(blocklen, false) / 2; + + return xfs_btree_calc_size(minrecs, nr_records); +} + +int __init +rcbagbt_init_cur_cache(void) +{ + rcbagbt_cur_cache = kmem_cache_create("xfs_rcbagbt_cur", + xfs_btree_cur_sizeof(rcbagbt_maxlevels_possible()), + 0, 0, NULL); + + if (!rcbagbt_cur_cache) + return -ENOMEM; + return 0; +} + +void +rcbagbt_destroy_cur_cache(void) +{ + kmem_cache_destroy(rcbagbt_cur_cache); + rcbagbt_cur_cache = NULL; +} + +/* Look up the refcount bag record corresponding to this reverse mapping. */ +int +rcbagbt_lookup_eq( + struct xfs_btree_cur *cur, + const struct xfs_rmap_irec *rmap, + int *success) +{ + struct rcbag_rec *rec = (struct rcbag_rec *)&cur->bc_rec; + + rec->rbg_startblock = rmap->rm_startblock; + rec->rbg_blockcount = rmap->rm_blockcount; + + return xfs_btree_lookup(cur, XFS_LOOKUP_EQ, success); +} + +/* Get the data from the pointed-to record. */ +int +rcbagbt_get_rec( + struct xfs_btree_cur *cur, + struct rcbag_rec *rec, + int *has) +{ + union xfs_btree_rec *btrec; + int error; + + error = xfs_btree_get_rec(cur, &btrec, has); + if (error || !(*has)) + return error; + + memcpy(rec, btrec, sizeof(struct rcbag_rec)); + return 0; +} + +/* Update the record referred to by cur to the value given. */ +int +rcbagbt_update( + struct xfs_btree_cur *cur, + const struct rcbag_rec *rec) +{ + union xfs_btree_rec btrec; + + memcpy(&btrec, rec, sizeof(struct rcbag_rec)); + return xfs_btree_update(cur, &btrec); +} + +/* Update the record referred to by cur to the value given. */ +int +rcbagbt_insert( + struct xfs_btree_cur *cur, + const struct rcbag_rec *rec, + int *success) +{ + struct rcbag_rec *btrec = (struct rcbag_rec *)&cur->bc_rec; + + memcpy(btrec, rec, sizeof(struct rcbag_rec)); + return xfs_btree_insert(cur, success); +} diff --git a/fs/xfs/scrub/rcbag_btree.h b/fs/xfs/scrub/rcbag_btree.h new file mode 100644 index 000000000000..03cadb032552 --- /dev/null +++ b/fs/xfs/scrub/rcbag_btree.h @@ -0,0 +1,81 @@ +// SPDX-License-Identifier: GPL-2.0-or-later +/* + * Copyright (c) 2022-2024 Oracle. All Rights Reserved. + * Author: Darrick J. Wong <djwong@kernel.org> + */ +#ifndef __XFS_SCRUB_RCBAG_BTREE_H__ +#define __XFS_SCRUB_RCBAG_BTREE_H__ + +#ifdef CONFIG_XFS_BTREE_IN_MEM + +struct xfs_buf; +struct xfs_btree_cur; +struct xfs_mount; + +#define RCBAG_MAGIC 0x74826671 /* 'JRBG' */ + +struct rcbag_key { + uint32_t rbg_startblock; + uint32_t rbg_blockcount; +}; + +struct rcbag_rec { + uint32_t rbg_startblock; + uint32_t rbg_blockcount; + uint64_t rbg_refcount; +}; + +typedef __be64 rcbag_ptr_t; + +/* reflinks only exist on crc enabled filesystems */ +#define RCBAG_BLOCK_LEN XFS_BTREE_LBLOCK_CRC_LEN + +/* + * Record, key, and pointer address macros for btree blocks. + * + * (note that some of these may appear unused, but they are used in userspace) + */ +#define RCBAG_REC_ADDR(block, index) \ + ((struct rcbag_rec *) \ + ((char *)(block) + RCBAG_BLOCK_LEN + \ + (((index) - 1) * sizeof(struct rcbag_rec)))) + +#define RCBAG_KEY_ADDR(block, index) \ + ((struct rcbag_key *) \ + ((char *)(block) + RCBAG_BLOCK_LEN + \ + ((index) - 1) * sizeof(struct rcbag_key))) + +#define RCBAG_PTR_ADDR(block, index, maxrecs) \ + ((rcbag_ptr_t *) \ + ((char *)(block) + RCBAG_BLOCK_LEN + \ + (maxrecs) * sizeof(struct rcbag_key) + \ + ((index) - 1) * sizeof(rcbag_ptr_t))) + +unsigned int rcbagbt_maxrecs(struct xfs_mount *mp, unsigned int blocklen, + bool leaf); + +unsigned long long rcbagbt_calc_size(unsigned long long nr_records); + +unsigned int rcbagbt_maxlevels_possible(void); + +int __init rcbagbt_init_cur_cache(void); +void rcbagbt_destroy_cur_cache(void); + +struct xfs_btree_cur *rcbagbt_mem_cursor(struct xfs_mount *mp, + struct xfs_trans *tp, struct xfbtree *xfbtree); +int rcbagbt_mem_init(struct xfs_mount *mp, struct xfbtree *xfbtree, + struct xfs_buftarg *btp); + +int rcbagbt_lookup_eq(struct xfs_btree_cur *cur, + const struct xfs_rmap_irec *rmap, int *success); +int rcbagbt_get_rec(struct xfs_btree_cur *cur, struct rcbag_rec *rec, int *has); +int rcbagbt_update(struct xfs_btree_cur *cur, const struct rcbag_rec *rec); +int rcbagbt_insert(struct xfs_btree_cur *cur, const struct rcbag_rec *rec, + int *success); + +#else +# define rcbagbt_init_cur_cache() 0 +# define rcbagbt_destroy_cur_cache() ((void)0) +#endif /* CONFIG_XFS_BTREE_IN_MEM */ + +#endif /* __XFS_SCRUB_RCBAG_BTREE_H__ */ diff --git a/fs/xfs/scrub/readdir.c b/fs/xfs/scrub/readdir.c index 16462332c897..dfdcb96b6c16 100644 --- a/fs/xfs/scrub/readdir.c +++ b/fs/xfs/scrub/readdir.c @@ -281,7 +281,7 @@ xchk_dir_walk( return -EIO; ASSERT(S_ISDIR(VFS_I(dp)->i_mode)); - ASSERT(xfs_isilocked(dp, XFS_ILOCK_SHARED | XFS_ILOCK_EXCL)); + xfs_assert_ilocked(dp, XFS_ILOCK_SHARED | XFS_ILOCK_EXCL); if (dp->i_df.if_format == XFS_DINODE_FMT_LOCAL) return xchk_dir_walk_sf(sc, dp, dirent_fn, priv); @@ -332,7 +332,7 @@ xchk_dir_lookup( return -EIO; ASSERT(S_ISDIR(VFS_I(dp)->i_mode)); - ASSERT(xfs_isilocked(dp, XFS_ILOCK_SHARED | XFS_ILOCK_EXCL)); + xfs_assert_ilocked(dp, XFS_ILOCK_SHARED | XFS_ILOCK_EXCL); if (dp->i_df.if_format == XFS_DINODE_FMT_LOCAL) { error = xfs_dir2_sf_lookup(&args); diff --git a/fs/xfs/scrub/reap.c b/fs/xfs/scrub/reap.c index f99eca799809..0252a3b5b65a 100644 --- a/fs/xfs/scrub/reap.c +++ b/fs/xfs/scrub/reap.c @@ -114,7 +114,7 @@ xreap_put_freelist( int error; /* Make sure there's space on the freelist. */ - error = xrep_fix_freelist(sc, true); + error = xrep_fix_freelist(sc, 0); if (error) return error; diff --git a/fs/xfs/scrub/refcount.c b/fs/xfs/scrub/refcount.c index bf22f245bbfa..d0c7d4a29c0f 100644 --- a/fs/xfs/scrub/refcount.c +++ b/fs/xfs/scrub/refcount.c @@ -7,8 +7,10 @@ #include "xfs_fs.h" #include "xfs_shared.h" #include "xfs_format.h" +#include "xfs_log_format.h" #include "xfs_trans_resv.h" #include "xfs_mount.h" +#include "xfs_trans.h" #include "xfs_ag.h" #include "xfs_btree.h" #include "xfs_rmap.h" @@ -17,6 +19,7 @@ #include "scrub/common.h" #include "scrub/btree.h" #include "scrub/trace.h" +#include "scrub/repair.h" /* * Set us up to scrub reference count btrees. @@ -27,6 +30,15 @@ xchk_setup_ag_refcountbt( { if (xchk_need_intent_drain(sc)) xchk_fsgates_enable(sc, XCHK_FSGATES_DRAIN); + + if (xchk_could_repair(sc)) { + int error; + + error = xrep_setup_ag_refcountbt(sc); + if (error) + return error; + } + return xchk_setup_ag_btree(sc, false); } diff --git a/fs/xfs/scrub/refcount_repair.c b/fs/xfs/scrub/refcount_repair.c index f38fccc42a20..a00d7ce7ae5b 100644 --- a/fs/xfs/scrub/refcount_repair.c +++ b/fs/xfs/scrub/refcount_repair.c @@ -25,6 +25,7 @@ #include "xfs_refcount_btree.h" #include "xfs_error.h" #include "xfs_ag.h" +#include "xfs_health.h" #include "scrub/xfs_scrub.h" #include "scrub/scrub.h" #include "scrub/common.h" @@ -37,6 +38,7 @@ #include "scrub/xfarray.h" #include "scrub/newbt.h" #include "scrub/reap.h" +#include "scrub/rcbag.h" /* * Rebuilding the Reference Count Btree @@ -97,12 +99,6 @@ * insert all the records. */ -/* The only parts of the rmap that we care about for computing refcounts. */ -struct xrep_refc_rmap { - xfs_agblock_t startblock; - xfs_extlen_t blockcount; -} __packed; - struct xrep_refc { /* refcount extents */ struct xfarray *refcount_records; @@ -122,6 +118,20 @@ struct xrep_refc { xfs_extlen_t btblocks; }; +/* Set us up to repair refcount btrees. */ +int +xrep_setup_ag_refcountbt( + struct xfs_scrub *sc) +{ + char *descr; + int error; + + descr = xchk_xfile_ag_descr(sc, "rmap record bag"); + error = xrep_setup_xfbtree(sc, descr); + kfree(descr); + return error; +} + /* Check for any obvious conflicts with this shared/CoW staging extent. */ STATIC int xrep_refc_check_ext( @@ -223,10 +233,9 @@ xrep_refc_rmap_shareable( STATIC int xrep_refc_walk_rmaps( struct xrep_refc *rr, - struct xrep_refc_rmap *rrm, + struct xfs_rmap_irec *rmap, bool *have_rec) { - struct xfs_rmap_irec rmap; struct xfs_btree_cur *cur = rr->sc->sa.rmap_cur; struct xfs_mount *mp = cur->bc_mp; int have_gt; @@ -250,29 +259,30 @@ xrep_refc_walk_rmaps( if (!have_gt) return 0; - error = xfs_rmap_get_rec(cur, &rmap, &have_gt); + error = xfs_rmap_get_rec(cur, rmap, &have_gt); if (error) return error; - if (XFS_IS_CORRUPT(mp, !have_gt)) + if (XFS_IS_CORRUPT(mp, !have_gt)) { + xfs_btree_mark_sick(cur); return -EFSCORRUPTED; + } - if (rmap.rm_owner == XFS_RMAP_OWN_COW) { - error = xrep_refc_stash_cow(rr, rmap.rm_startblock, - rmap.rm_blockcount); + if (rmap->rm_owner == XFS_RMAP_OWN_COW) { + error = xrep_refc_stash_cow(rr, rmap->rm_startblock, + rmap->rm_blockcount); if (error) return error; - } else if (rmap.rm_owner == XFS_RMAP_OWN_REFC) { + } else if (rmap->rm_owner == XFS_RMAP_OWN_REFC) { /* refcountbt block, dump it when we're done. */ - rr->btblocks += rmap.rm_blockcount; + rr->btblocks += rmap->rm_blockcount; error = xagb_bitmap_set(&rr->old_refcountbt_blocks, - rmap.rm_startblock, rmap.rm_blockcount); + rmap->rm_startblock, + rmap->rm_blockcount); if (error) return error; } - } while (!xrep_refc_rmap_shareable(mp, &rmap)); + } while (!xrep_refc_rmap_shareable(mp, rmap)); - rrm->startblock = rmap.rm_startblock; - rrm->blockcount = rmap.rm_blockcount; *have_rec = true; return 0; } @@ -354,45 +364,6 @@ xrep_refc_sort_records( return error; } -#define RRM_NEXT(r) ((r).startblock + (r).blockcount) -/* - * Find the next block where the refcount changes, given the next rmap we - * looked at and the ones we're already tracking. - */ -static inline int -xrep_refc_next_edge( - struct xfarray *rmap_bag, - struct xrep_refc_rmap *next_rrm, - bool next_valid, - xfs_agblock_t *nbnop) -{ - struct xrep_refc_rmap rrm; - xfarray_idx_t array_cur = XFARRAY_CURSOR_INIT; - xfs_agblock_t nbno = NULLAGBLOCK; - int error; - - if (next_valid) - nbno = next_rrm->startblock; - - while ((error = xfarray_iter(rmap_bag, &array_cur, &rrm)) == 1) - nbno = min_t(xfs_agblock_t, nbno, RRM_NEXT(rrm)); - - if (error) - return error; - - /* - * We should have found /something/ because either next_rrm is the next - * interesting rmap to look at after emitting this refcount extent, or - * there are other rmaps in rmap_bag contributing to the current - * sharing count. But if something is seriously wrong, bail out. - */ - if (nbno == NULLAGBLOCK) - return -EFSCORRUPTED; - - *nbnop = nbno; - return 0; -} - /* * Walk forward through the rmap btree to collect all rmaps starting at * @bno in @rmap_bag. These represent the file(s) that share ownership of @@ -402,22 +373,21 @@ xrep_refc_next_edge( static int xrep_refc_push_rmaps_at( struct xrep_refc *rr, - struct xfarray *rmap_bag, + struct rcbag *rcstack, xfs_agblock_t bno, - struct xrep_refc_rmap *rrm, - bool *have, - uint64_t *stack_sz) + struct xfs_rmap_irec *rmap, + bool *have) { struct xfs_scrub *sc = rr->sc; int have_gt; int error; - while (*have && rrm->startblock == bno) { - error = xfarray_store_anywhere(rmap_bag, rrm); + while (*have && rmap->rm_startblock == bno) { + error = rcbag_add(rcstack, rr->sc->tp, rmap); if (error) return error; - (*stack_sz)++; - error = xrep_refc_walk_rmaps(rr, rrm, have); + + error = xrep_refc_walk_rmaps(rr, rmap, have); if (error) return error; } @@ -425,8 +395,10 @@ xrep_refc_push_rmaps_at( error = xfs_btree_decrement(sc->sa.rmap_cur, 0, &have_gt); if (error) return error; - if (XFS_IS_CORRUPT(sc->mp, !have_gt)) + if (XFS_IS_CORRUPT(sc->mp, !have_gt)) { + xfs_btree_mark_sick(sc->sa.rmap_cur); return -EFSCORRUPTED; + } return 0; } @@ -436,12 +408,9 @@ STATIC int xrep_refc_find_refcounts( struct xrep_refc *rr) { - struct xrep_refc_rmap rrm; struct xfs_scrub *sc = rr->sc; - struct xfarray *rmap_bag; - char *descr; - uint64_t old_stack_sz; - uint64_t stack_sz = 0; + struct rcbag *rcstack; + uint64_t old_stack_height; xfs_agblock_t sbno; xfs_agblock_t cbno; xfs_agblock_t nbno; @@ -451,14 +420,11 @@ xrep_refc_find_refcounts( xrep_ag_btcur_init(sc, &sc->sa); /* - * Set up a sparse array to store all the rmap records that we're - * tracking to generate a reference count record. If this exceeds + * Set up a bag to store all the rmap records that we're tracking to + * generate a reference count record. If the size of the bag exceeds * MAXREFCOUNT, we clamp rc_refcount. */ - descr = xchk_xfile_ag_descr(sc, "rmap record bag"); - error = xfarray_create(descr, 0, sizeof(struct xrep_refc_rmap), - &rmap_bag); - kfree(descr); + error = rcbag_init(sc->mp, sc->xmbtp, &rcstack); if (error) goto out_cur; @@ -469,62 +435,54 @@ xrep_refc_find_refcounts( /* Process reverse mappings into refcount data. */ while (xfs_btree_has_more_records(sc->sa.rmap_cur)) { + struct xfs_rmap_irec rmap; + /* Push all rmaps with pblk == sbno onto the stack */ - error = xrep_refc_walk_rmaps(rr, &rrm, &have); + error = xrep_refc_walk_rmaps(rr, &rmap, &have); if (error) goto out_bag; if (!have) break; - sbno = cbno = rrm.startblock; - error = xrep_refc_push_rmaps_at(rr, rmap_bag, sbno, - &rrm, &have, &stack_sz); + sbno = cbno = rmap.rm_startblock; + error = xrep_refc_push_rmaps_at(rr, rcstack, sbno, &rmap, + &have); if (error) goto out_bag; /* Set nbno to the bno of the next refcount change */ - error = xrep_refc_next_edge(rmap_bag, &rrm, have, &nbno); + error = rcbag_next_edge(rcstack, sc->tp, &rmap, have, &nbno); if (error) goto out_bag; ASSERT(nbno > sbno); - old_stack_sz = stack_sz; + old_stack_height = rcbag_count(rcstack); /* While stack isn't empty... */ - while (stack_sz) { - xfarray_idx_t array_cur = XFARRAY_CURSOR_INIT; - + while (rcbag_count(rcstack) > 0) { /* Pop all rmaps that end at nbno */ - while ((error = xfarray_iter(rmap_bag, &array_cur, - &rrm)) == 1) { - if (RRM_NEXT(rrm) != nbno) - continue; - error = xfarray_unset(rmap_bag, array_cur - 1); - if (error) - goto out_bag; - stack_sz--; - } + error = rcbag_remove_ending_at(rcstack, sc->tp, nbno); if (error) goto out_bag; /* Push array items that start at nbno */ - error = xrep_refc_walk_rmaps(rr, &rrm, &have); + error = xrep_refc_walk_rmaps(rr, &rmap, &have); if (error) goto out_bag; if (have) { - error = xrep_refc_push_rmaps_at(rr, rmap_bag, - nbno, &rrm, &have, &stack_sz); + error = xrep_refc_push_rmaps_at(rr, rcstack, + nbno, &rmap, &have); if (error) goto out_bag; } /* Emit refcount if necessary */ ASSERT(nbno > cbno); - if (stack_sz != old_stack_sz) { - if (old_stack_sz > 1) { + if (rcbag_count(rcstack) != old_stack_height) { + if (old_stack_height > 1) { error = xrep_refc_stash(rr, XFS_REFC_DOMAIN_SHARED, cbno, nbno - cbno, - old_stack_sz); + old_stack_height); if (error) goto out_bag; } @@ -532,13 +490,13 @@ xrep_refc_find_refcounts( } /* Stack empty, go find the next rmap */ - if (stack_sz == 0) + if (rcbag_count(rcstack) == 0) break; - old_stack_sz = stack_sz; + old_stack_height = rcbag_count(rcstack); sbno = nbno; /* Set nbno to the bno of the next refcount change */ - error = xrep_refc_next_edge(rmap_bag, &rrm, have, + error = rcbag_next_edge(rcstack, sc->tp, &rmap, have, &nbno); if (error) goto out_bag; @@ -547,14 +505,13 @@ xrep_refc_find_refcounts( } } - ASSERT(stack_sz == 0); + ASSERT(rcbag_count(rcstack) == 0); out_bag: - xfarray_destroy(rmap_bag); + rcbag_free(&rcstack); out_cur: xchk_ag_btcur_free(&sc->sa); return error; } -#undef RRM_NEXT /* Retrieve refcountbt data for bulk load. */ STATIC int @@ -653,8 +610,8 @@ xrep_refc_build_new_tree( rr->new_btree.bload.claim_block = xrep_refc_claim_block; /* Compute how many blocks we'll need. */ - refc_cur = xfs_refcountbt_stage_cursor(sc->mp, &rr->new_btree.afake, - pag); + refc_cur = xfs_refcountbt_init_cursor(sc->mp, NULL, NULL, pag); + xfs_btree_stage_afakeroot(refc_cur, &rr->new_btree.afake); error = xfs_btree_bload_compute_geometry(refc_cur, &rr->new_btree.bload, xfarray_length(rr->refcount_records)); diff --git a/fs/xfs/scrub/repair.c b/fs/xfs/scrub/repair.c index 745d5b8f405a..f43dce771cdd 100644 --- a/fs/xfs/scrub/repair.c +++ b/fs/xfs/scrub/repair.c @@ -30,12 +30,15 @@ #include "xfs_errortag.h" #include "xfs_error.h" #include "xfs_reflink.h" +#include "xfs_health.h" +#include "xfs_buf_mem.h" #include "scrub/scrub.h" #include "scrub/common.h" #include "scrub/trace.h" #include "scrub/repair.h" #include "scrub/bitmap.h" #include "scrub/stats.h" +#include "scrub/xfile.h" /* * Attempt to repair some metadata, if the metadata is corrupt and userspace @@ -400,7 +403,7 @@ xrep_calc_ag_resblks( int xrep_fix_freelist( struct xfs_scrub *sc, - bool can_shrink) + int alloc_flags) { struct xfs_alloc_arg args = {0}; @@ -410,8 +413,7 @@ xrep_fix_freelist( args.alignment = 1; args.pag = sc->sa.pag; - return xfs_alloc_fix_freelist(&args, - can_shrink ? 0 : XFS_ALLOC_FLAG_NOSHRINK); + return xfs_alloc_fix_freelist(&args, alloc_flags); } /* @@ -687,6 +689,44 @@ xrep_find_ag_btree_roots( } #ifdef CONFIG_XFS_QUOTA +/* Update some quota flags in the superblock. */ +void +xrep_update_qflags( + struct xfs_scrub *sc, + unsigned int clear_flags, + unsigned int set_flags) +{ + struct xfs_mount *mp = sc->mp; + struct xfs_buf *bp; + + mutex_lock(&mp->m_quotainfo->qi_quotaofflock); + if ((mp->m_qflags & clear_flags) == 0 && + (mp->m_qflags & set_flags) == set_flags) + goto no_update; + + mp->m_qflags &= ~clear_flags; + mp->m_qflags |= set_flags; + + spin_lock(&mp->m_sb_lock); + mp->m_sb.sb_qflags &= ~clear_flags; + mp->m_sb.sb_qflags |= set_flags; + spin_unlock(&mp->m_sb_lock); + + /* + * Update the quota flags in the ondisk superblock without touching + * the summary counters. We have not quiesced inode chunk allocation, + * so we cannot coordinate with updates to the icount and ifree percpu + * counters. + */ + bp = xfs_trans_getsb(sc->tp); + xfs_sb_to_disk(bp->b_addr, &mp->m_sb); + xfs_trans_buf_set_type(sc->tp, bp, XFS_BLFT_SB_BUF); + xfs_trans_log_buf(sc->tp, bp, 0, sizeof(struct xfs_dsb) - 1); + +no_update: + mutex_unlock(&sc->mp->m_quotainfo->qi_quotaofflock); +} + /* Force a quotacheck the next time we mount. */ void xrep_force_quotacheck( @@ -699,13 +739,7 @@ xrep_force_quotacheck( if (!(flag & sc->mp->m_qflags)) return; - mutex_lock(&sc->mp->m_quotainfo->qi_quotaofflock); - sc->mp->m_qflags &= ~flag; - spin_lock(&sc->mp->m_sb_lock); - sc->mp->m_sb.sb_qflags &= ~flag; - spin_unlock(&sc->mp->m_sb_lock); - xfs_log_sb(sc->tp); - mutex_unlock(&sc->mp->m_quotainfo->qi_quotaofflock); + xrep_update_qflags(sc, flag, 0); } /* @@ -799,20 +833,20 @@ xrep_ag_btcur_init( /* Set up a bnobt cursor for cross-referencing. */ if (sc->sm->sm_type != XFS_SCRUB_TYPE_BNOBT && sc->sm->sm_type != XFS_SCRUB_TYPE_CNTBT) { - sa->bno_cur = xfs_allocbt_init_cursor(mp, sc->tp, sa->agf_bp, - sc->sa.pag, XFS_BTNUM_BNO); - sa->cnt_cur = xfs_allocbt_init_cursor(mp, sc->tp, sa->agf_bp, - sc->sa.pag, XFS_BTNUM_CNT); + sa->bno_cur = xfs_bnobt_init_cursor(mp, sc->tp, sa->agf_bp, + sc->sa.pag); + sa->cnt_cur = xfs_cntbt_init_cursor(mp, sc->tp, sa->agf_bp, + sc->sa.pag); } /* Set up a inobt cursor for cross-referencing. */ if (sc->sm->sm_type != XFS_SCRUB_TYPE_INOBT && sc->sm->sm_type != XFS_SCRUB_TYPE_FINOBT) { sa->ino_cur = xfs_inobt_init_cursor(sc->sa.pag, sc->tp, - sa->agi_bp, XFS_BTNUM_INO); + sa->agi_bp); if (xfs_has_finobt(mp)) - sa->fino_cur = xfs_inobt_init_cursor(sc->sa.pag, - sc->tp, sa->agi_bp, XFS_BTNUM_FINO); + sa->fino_cur = xfs_finobt_init_cursor(sc->sa.pag, + sc->tp, sa->agi_bp); } /* Set up a rmapbt cursor for cross-referencing. */ @@ -1115,3 +1149,55 @@ xrep_metadata_inode_forks( return 0; } + +/* + * Set up an in-memory buffer cache so that we can use the xfbtree. Allocating + * a shmem file might take loks, so we cannot be in transaction context. Park + * our resources in the scrub context and let the teardown function take care + * of them at the right time. + */ +int +xrep_setup_xfbtree( + struct xfs_scrub *sc, + const char *descr) +{ + ASSERT(sc->tp == NULL); + + return xmbuf_alloc(sc->mp, descr, &sc->xmbtp); +} + +/* + * Create a dummy transaction for use in a live update hook function. This + * function MUST NOT be called from regular repair code because the current + * process' transaction is saved via the cookie. + */ +int +xrep_trans_alloc_hook_dummy( + struct xfs_mount *mp, + void **cookiep, + struct xfs_trans **tpp) +{ + int error; + + *cookiep = current->journal_info; + current->journal_info = NULL; + + error = xfs_trans_alloc_empty(mp, tpp); + if (!error) + return 0; + + current->journal_info = *cookiep; + *cookiep = NULL; + return error; +} + +/* Cancel a dummy transaction used by a live update hook function. */ +void +xrep_trans_cancel_hook_dummy( + void **cookiep, + struct xfs_trans *tp) +{ + xfs_trans_cancel(tp); + current->journal_info = *cookiep; + *cookiep = NULL; +} diff --git a/fs/xfs/scrub/repair.h b/fs/xfs/scrub/repair.h index 17114327e6fa..ce082d941459 100644 --- a/fs/xfs/scrub/repair.h +++ b/fs/xfs/scrub/repair.h @@ -51,7 +51,7 @@ struct xbitmap; struct xagb_bitmap; struct xfsb_bitmap; -int xrep_fix_freelist(struct xfs_scrub *sc, bool can_shrink); +int xrep_fix_freelist(struct xfs_scrub *sc, int alloc_flags); struct xrep_find_ag_btree { /* in: rmap owner of the btree we're looking for */ @@ -72,6 +72,8 @@ int xrep_find_ag_btree_roots(struct xfs_scrub *sc, struct xfs_buf *agf_bp, struct xrep_find_ag_btree *btree_info, struct xfs_buf *agfl_bp); #ifdef CONFIG_XFS_QUOTA +void xrep_update_qflags(struct xfs_scrub *sc, unsigned int clear_flags, + unsigned int set_flags); void xrep_force_quotacheck(struct xfs_scrub *sc, xfs_dqtype_t type); int xrep_ino_dqattach(struct xfs_scrub *sc); #else @@ -79,11 +81,15 @@ int xrep_ino_dqattach(struct xfs_scrub *sc); # define xrep_ino_dqattach(sc) (0) #endif /* CONFIG_XFS_QUOTA */ +int xrep_setup_xfbtree(struct xfs_scrub *sc, const char *descr); + int xrep_ino_ensure_extent_count(struct xfs_scrub *sc, int whichfork, xfs_extnum_t nextents); int xrep_reset_perag_resv(struct xfs_scrub *sc); int xrep_bmap(struct xfs_scrub *sc, int whichfork, bool allow_unwritten); int xrep_metadata_inode_forks(struct xfs_scrub *sc); +int xrep_setup_ag_rmapbt(struct xfs_scrub *sc); +int xrep_setup_ag_refcountbt(struct xfs_scrub *sc); /* Repair setup functions */ int xrep_setup_ag_allocbt(struct xfs_scrub *sc); @@ -109,11 +115,14 @@ int xrep_agfl(struct xfs_scrub *sc); int xrep_agi(struct xfs_scrub *sc); int xrep_allocbt(struct xfs_scrub *sc); int xrep_iallocbt(struct xfs_scrub *sc); +int xrep_rmapbt(struct xfs_scrub *sc); int xrep_refcountbt(struct xfs_scrub *sc); int xrep_inode(struct xfs_scrub *sc); int xrep_bmap_data(struct xfs_scrub *sc); int xrep_bmap_attr(struct xfs_scrub *sc); int xrep_bmap_cow(struct xfs_scrub *sc); +int xrep_nlinks(struct xfs_scrub *sc); +int xrep_fscounters(struct xfs_scrub *sc); #ifdef CONFIG_XFS_RT int xrep_rtbitmap(struct xfs_scrub *sc); @@ -123,13 +132,19 @@ int xrep_rtbitmap(struct xfs_scrub *sc); #ifdef CONFIG_XFS_QUOTA int xrep_quota(struct xfs_scrub *sc); +int xrep_quotacheck(struct xfs_scrub *sc); #else # define xrep_quota xrep_notsupported +# define xrep_quotacheck xrep_notsupported #endif /* CONFIG_XFS_QUOTA */ int xrep_reinit_pagf(struct xfs_scrub *sc); int xrep_reinit_pagi(struct xfs_scrub *sc); +int xrep_trans_alloc_hook_dummy(struct xfs_mount *mp, void **cookiep, + struct xfs_trans **tpp); +void xrep_trans_cancel_hook_dummy(void **cookiep, struct xfs_trans *tp); + #else #define xrep_ino_dqattach(sc) (0) @@ -171,6 +186,8 @@ xrep_setup_nothing( return 0; } #define xrep_setup_ag_allocbt xrep_setup_nothing +#define xrep_setup_ag_rmapbt xrep_setup_nothing +#define xrep_setup_ag_refcountbt xrep_setup_nothing #define xrep_setup_inode(sc, imap) ((void)0) @@ -184,6 +201,7 @@ xrep_setup_nothing( #define xrep_agi xrep_notsupported #define xrep_allocbt xrep_notsupported #define xrep_iallocbt xrep_notsupported +#define xrep_rmapbt xrep_notsupported #define xrep_refcountbt xrep_notsupported #define xrep_inode xrep_notsupported #define xrep_bmap_data xrep_notsupported @@ -191,6 +209,9 @@ xrep_setup_nothing( #define xrep_bmap_cow xrep_notsupported #define xrep_rtbitmap xrep_notsupported #define xrep_quota xrep_notsupported +#define xrep_quotacheck xrep_notsupported +#define xrep_nlinks xrep_notsupported +#define xrep_fscounters xrep_notsupported #endif /* CONFIG_XFS_ONLINE_REPAIR */ diff --git a/fs/xfs/scrub/rmap.c b/fs/xfs/scrub/rmap.c index c99d1714f283..ba5bbc3fb754 100644 --- a/fs/xfs/scrub/rmap.c +++ b/fs/xfs/scrub/rmap.c @@ -25,6 +25,7 @@ #include "scrub/btree.h" #include "scrub/bitmap.h" #include "scrub/agb_bitmap.h" +#include "scrub/repair.h" /* * Set us up to scrub reverse mapping btrees. @@ -36,6 +37,14 @@ xchk_setup_ag_rmapbt( if (xchk_need_intent_drain(sc)) xchk_fsgates_enable(sc, XCHK_FSGATES_DRAIN); + if (xchk_could_repair(sc)) { + int error; + + error = xrep_setup_ag_rmapbt(sc); + if (error) + return error; + } + return xchk_setup_ag_btree(sc, false); } @@ -349,7 +358,7 @@ xchk_rmapbt_rec( struct xfs_rmap_irec irec; if (xfs_rmap_btrec_to_irec(rec, &irec) != NULL || - xfs_rmap_check_irec(bs->cur, &irec) != NULL) { + xfs_rmap_check_irec(bs->cur->bc_ag.pag, &irec) != NULL) { xchk_btree_set_corrupt(bs->sc, bs->cur, 0); return 0; } @@ -412,8 +421,8 @@ xchk_rmapbt_walk_ag_metadata( /* OWN_AG: bnobt, cntbt, rmapbt, and AGFL */ cur = sc->sa.bno_cur; if (!cur) - cur = xfs_allocbt_init_cursor(sc->mp, sc->tp, sc->sa.agf_bp, - sc->sa.pag, XFS_BTNUM_BNO); + cur = xfs_bnobt_init_cursor(sc->mp, sc->tp, sc->sa.agf_bp, + sc->sa.pag); error = xagb_bitmap_set_btblocks(&cr->ag_owned, cur); if (cur != sc->sa.bno_cur) xfs_btree_del_cursor(cur, error); @@ -422,8 +431,8 @@ xchk_rmapbt_walk_ag_metadata( cur = sc->sa.cnt_cur; if (!cur) - cur = xfs_allocbt_init_cursor(sc->mp, sc->tp, sc->sa.agf_bp, - sc->sa.pag, XFS_BTNUM_CNT); + cur = xfs_cntbt_init_cursor(sc->mp, sc->tp, sc->sa.agf_bp, + sc->sa.pag); error = xagb_bitmap_set_btblocks(&cr->ag_owned, cur); if (cur != sc->sa.cnt_cur) xfs_btree_del_cursor(cur, error); @@ -447,8 +456,7 @@ xchk_rmapbt_walk_ag_metadata( /* OWN_INOBT: inobt, finobt */ cur = sc->sa.ino_cur; if (!cur) - cur = xfs_inobt_init_cursor(sc->sa.pag, sc->tp, sc->sa.agi_bp, - XFS_BTNUM_INO); + cur = xfs_inobt_init_cursor(sc->sa.pag, sc->tp, sc->sa.agi_bp); error = xagb_bitmap_set_btblocks(&cr->inobt_owned, cur); if (cur != sc->sa.ino_cur) xfs_btree_del_cursor(cur, error); @@ -458,8 +466,8 @@ xchk_rmapbt_walk_ag_metadata( if (xfs_has_finobt(sc->mp)) { cur = sc->sa.fino_cur; if (!cur) - cur = xfs_inobt_init_cursor(sc->sa.pag, sc->tp, - sc->sa.agi_bp, XFS_BTNUM_FINO); + cur = xfs_finobt_init_cursor(sc->sa.pag, sc->tp, + sc->sa.agi_bp); error = xagb_bitmap_set_btblocks(&cr->inobt_owned, cur); if (cur != sc->sa.fino_cur) xfs_btree_del_cursor(cur, error); diff --git a/fs/xfs/scrub/rmap_repair.c b/fs/xfs/scrub/rmap_repair.c new file mode 100644 index 000000000000..e8e07b683eab --- /dev/null +++ b/fs/xfs/scrub/rmap_repair.c @@ -0,0 +1,1697 @@ +// SPDX-License-Identifier: GPL-2.0-or-later +/* + * Copyright (c) 2018-2024 Oracle. All Rights Reserved. + * Author: Darrick J. Wong <djwong@kernel.org> + */ +#include "xfs.h" +#include "xfs_fs.h" +#include "xfs_shared.h" +#include "xfs_format.h" +#include "xfs_trans_resv.h" +#include "xfs_mount.h" +#include "xfs_defer.h" +#include "xfs_btree.h" +#include "xfs_btree_staging.h" +#include "xfs_buf_mem.h" +#include "xfs_btree_mem.h" +#include "xfs_bit.h" +#include "xfs_log_format.h" +#include "xfs_trans.h" +#include "xfs_sb.h" +#include "xfs_alloc.h" +#include "xfs_alloc_btree.h" +#include "xfs_ialloc.h" +#include "xfs_ialloc_btree.h" +#include "xfs_rmap.h" +#include "xfs_rmap_btree.h" +#include "xfs_inode.h" +#include "xfs_icache.h" +#include "xfs_bmap.h" +#include "xfs_bmap_btree.h" +#include "xfs_refcount.h" +#include "xfs_refcount_btree.h" +#include "xfs_ag.h" +#include "scrub/xfs_scrub.h" +#include "scrub/scrub.h" +#include "scrub/common.h" +#include "scrub/btree.h" +#include "scrub/trace.h" +#include "scrub/repair.h" +#include "scrub/bitmap.h" +#include "scrub/agb_bitmap.h" +#include "scrub/xfile.h" +#include "scrub/xfarray.h" +#include "scrub/iscan.h" +#include "scrub/newbt.h" +#include "scrub/reap.h" + +/* + * Reverse Mapping Btree Repair + * ============================ + * + * This is the most involved of all the AG space btree rebuilds. Everywhere + * else in XFS we lock inodes and then AG data structures, but generating the + * list of rmap records requires that we be able to scan both block mapping + * btrees of every inode in the filesystem to see if it owns any extents in + * this AG. We can't tolerate any inode updates while we do this, so we + * freeze the filesystem to lock everyone else out, and grant ourselves + * special privileges to run transactions with regular background reclamation + * turned off. + * + * We also have to be very careful not to allow inode reclaim to start a + * transaction because all transactions (other than our own) will block. + * Deferred inode inactivation helps us out there. + * + * I) Reverse mappings for all non-space metadata and file data are collected + * according to the following algorithm: + * + * 1. For each fork of each inode: + * 1.1. Create a bitmap BMBIT to track bmbt blocks if necessary. + * 1.2. If the incore extent map isn't loaded, walk the bmbt to accumulate + * bmaps into rmap records (see 1.1.4). Set bits in BMBIT for each btree + * block. + * 1.3. If the incore extent map is loaded but the fork is in btree format, + * just visit the bmbt blocks to set the corresponding BMBIT areas. + * 1.4. From the incore extent map, accumulate each bmap that falls into our + * target AG. Remember, multiple bmap records can map to a single rmap + * record, so we cannot simply emit rmap records 1:1. + * 1.5. Emit rmap records for each extent in BMBIT and free it. + * 2. Create bitmaps INOBIT and ICHUNKBIT. + * 3. For each record in the inobt, set the corresponding areas in ICHUNKBIT, + * and set bits in INOBIT for each btree block. If the inobt has no records + * at all, we must be careful to record its root in INOBIT. + * 4. For each block in the finobt, set the corresponding INOBIT area. + * 5. Emit rmap records for each extent in INOBIT and ICHUNKBIT and free them. + * 6. Create bitmaps REFCBIT and COWBIT. + * 7. For each CoW staging extent in the refcountbt, set the corresponding + * areas in COWBIT. + * 8. For each block in the refcountbt, set the corresponding REFCBIT area. + * 9. Emit rmap records for each extent in REFCBIT and COWBIT and free them. + * A. Emit rmap for the AG headers. + * B. Emit rmap for the log, if there is one. + * + * II) The rmapbt shape and space metadata rmaps are computed as follows: + * + * 1. Count the rmaps collected in the previous step. (= NR) + * 2. Estimate the number of rmapbt blocks needed to store NR records. (= RMB) + * 3. Reserve RMB blocks through the newbt using the allocator in normap mode. + * 4. Create bitmap AGBIT. + * 5. For each reservation in the newbt, set the corresponding areas in AGBIT. + * 6. For each block in the AGFL, bnobt, and cntbt, set the bits in AGBIT. + * 7. Count the extents in AGBIT. (= AGNR) + * 8. Estimate the number of rmapbt blocks needed for NR + AGNR rmaps. (= RMB') + * 9. If RMB' >= RMB, reserve RMB' - RMB more newbt blocks, set RMB = RMB', + * and clear AGBIT. Go to step 5. + * A. Emit rmaps for each extent in AGBIT. + * + * III) The rmapbt is constructed and set in place as follows: + * + * 1. Sort the rmap records. + * 2. Bulk load the rmaps. + * + * IV) Reap the old btree blocks. + * + * 1. Create a bitmap OLDRMBIT. + * 2. For each gap in the new rmapbt, set the corresponding areas of OLDRMBIT. + * 3. For each extent in the bnobt, clear the corresponding parts of OLDRMBIT. + * 4. Reap the extents corresponding to the set areas in OLDRMBIT. These are + * the parts of the AG that the rmap didn't find during its scan of the + * primary metadata and aren't known to be in the free space, which implies + * that they were the old rmapbt blocks. + * 5. Commit. + * + * We use the 'xrep_rmap' prefix for all the rmap functions. + */ + +/* Context for collecting rmaps */ +struct xrep_rmap { + /* new rmapbt information */ + struct xrep_newbt new_btree; + + /* lock for the xfbtree and xfile */ + struct mutex lock; + + /* rmap records generated from primary metadata */ + struct xfbtree rmap_btree; + + struct xfs_scrub *sc; + + /* in-memory btree cursor for the xfs_btree_bload iteration */ + struct xfs_btree_cur *mcur; + + /* Hooks into rmap update code. */ + struct xfs_rmap_hook rhook; + + /* inode scan cursor */ + struct xchk_iscan iscan; + + /* Number of non-freespace records found. */ + unsigned long long nr_records; + + /* bnobt/cntbt contribution to btreeblks */ + xfs_agblock_t freesp_btblocks; + + /* old agf_rmap_blocks counter */ + unsigned int old_rmapbt_fsbcount; +}; + +/* Set us up to repair reverse mapping btrees. */ +int +xrep_setup_ag_rmapbt( + struct xfs_scrub *sc) +{ + struct xrep_rmap *rr; + char *descr; + int error; + + xchk_fsgates_enable(sc, XCHK_FSGATES_RMAP); + + descr = xchk_xfile_ag_descr(sc, "reverse mapping records"); + error = xrep_setup_xfbtree(sc, descr); + kfree(descr); + if (error) + return error; + + rr = kzalloc(sizeof(struct xrep_rmap), XCHK_GFP_FLAGS); + if (!rr) + return -ENOMEM; + + rr->sc = sc; + sc->buf = rr; + return 0; +} + +/* Make sure there's nothing funny about this mapping. */ +STATIC int +xrep_rmap_check_mapping( + struct xfs_scrub *sc, + const struct xfs_rmap_irec *rec) +{ + enum xbtree_recpacking outcome; + int error; + + if (xfs_rmap_check_irec(sc->sa.pag, rec) != NULL) + return -EFSCORRUPTED; + + /* Make sure this isn't free space. */ + error = xfs_alloc_has_records(sc->sa.bno_cur, rec->rm_startblock, + rec->rm_blockcount, &outcome); + if (error) + return error; + if (outcome != XBTREE_RECPACKING_EMPTY) + return -EFSCORRUPTED; + + return 0; +} + +/* Store a reverse-mapping record. */ +static inline int +xrep_rmap_stash( + struct xrep_rmap *rr, + xfs_agblock_t startblock, + xfs_extlen_t blockcount, + uint64_t owner, + uint64_t offset, + unsigned int flags) +{ + struct xfs_rmap_irec rmap = { + .rm_startblock = startblock, + .rm_blockcount = blockcount, + .rm_owner = owner, + .rm_offset = offset, + .rm_flags = flags, + }; + struct xfs_scrub *sc = rr->sc; + struct xfs_btree_cur *mcur; + int error = 0; + + if (xchk_should_terminate(sc, &error)) + return error; + + if (xchk_iscan_aborted(&rr->iscan)) + return -EFSCORRUPTED; + + trace_xrep_rmap_found(sc->mp, sc->sa.pag->pag_agno, &rmap); + + mutex_lock(&rr->lock); + mcur = xfs_rmapbt_mem_cursor(sc->sa.pag, sc->tp, &rr->rmap_btree); + error = xfs_rmap_map_raw(mcur, &rmap); + xfs_btree_del_cursor(mcur, error); + if (error) + goto out_cancel; + + error = xfbtree_trans_commit(&rr->rmap_btree, sc->tp); + if (error) + goto out_abort; + + mutex_unlock(&rr->lock); + return 0; + +out_cancel: + xfbtree_trans_cancel(&rr->rmap_btree, sc->tp); +out_abort: + xchk_iscan_abort(&rr->iscan); + mutex_unlock(&rr->lock); + return error; +} + +struct xrep_rmap_stash_run { + struct xrep_rmap *rr; + uint64_t owner; + unsigned int rmap_flags; +}; + +static int +xrep_rmap_stash_run( + uint32_t start, + uint32_t len, + void *priv) +{ + struct xrep_rmap_stash_run *rsr = priv; + struct xrep_rmap *rr = rsr->rr; + + return xrep_rmap_stash(rr, start, len, rsr->owner, 0, rsr->rmap_flags); +} + +/* + * Emit rmaps for every extent of bits set in the bitmap. Caller must ensure + * that the ranges are in units of FS blocks. + */ +STATIC int +xrep_rmap_stash_bitmap( + struct xrep_rmap *rr, + struct xagb_bitmap *bitmap, + const struct xfs_owner_info *oinfo) +{ + struct xrep_rmap_stash_run rsr = { + .rr = rr, + .owner = oinfo->oi_owner, + .rmap_flags = 0, + }; + + if (oinfo->oi_flags & XFS_OWNER_INFO_ATTR_FORK) + rsr.rmap_flags |= XFS_RMAP_ATTR_FORK; + if (oinfo->oi_flags & XFS_OWNER_INFO_BMBT_BLOCK) + rsr.rmap_flags |= XFS_RMAP_BMBT_BLOCK; + + return xagb_bitmap_walk(bitmap, xrep_rmap_stash_run, &rsr); +} + +/* Section (I): Finding all file and bmbt extents. */ + +/* Context for accumulating rmaps for an inode fork. */ +struct xrep_rmap_ifork { + /* + * Accumulate rmap data here to turn multiple adjacent bmaps into a + * single rmap. + */ + struct xfs_rmap_irec accum; + + /* Bitmap of bmbt blocks in this AG. */ + struct xagb_bitmap bmbt_blocks; + + struct xrep_rmap *rr; + + /* Which inode fork? */ + int whichfork; +}; + +/* Stash an rmap that we accumulated while walking an inode fork. */ +STATIC int +xrep_rmap_stash_accumulated( + struct xrep_rmap_ifork *rf) +{ + if (rf->accum.rm_blockcount == 0) + return 0; + + return xrep_rmap_stash(rf->rr, rf->accum.rm_startblock, + rf->accum.rm_blockcount, rf->accum.rm_owner, + rf->accum.rm_offset, rf->accum.rm_flags); +} + +/* Accumulate a bmbt record. */ +STATIC int +xrep_rmap_visit_bmbt( + struct xfs_btree_cur *cur, + struct xfs_bmbt_irec *rec, + void *priv) +{ + struct xrep_rmap_ifork *rf = priv; + struct xfs_mount *mp = rf->rr->sc->mp; + struct xfs_rmap_irec *accum = &rf->accum; + xfs_agblock_t agbno; + unsigned int rmap_flags = 0; + int error; + + if (XFS_FSB_TO_AGNO(mp, rec->br_startblock) != + rf->rr->sc->sa.pag->pag_agno) + return 0; + + agbno = XFS_FSB_TO_AGBNO(mp, rec->br_startblock); + if (rf->whichfork == XFS_ATTR_FORK) + rmap_flags |= XFS_RMAP_ATTR_FORK; + if (rec->br_state == XFS_EXT_UNWRITTEN) + rmap_flags |= XFS_RMAP_UNWRITTEN; + + /* If this bmap is adjacent to the previous one, just add it. */ + if (accum->rm_blockcount > 0 && + rec->br_startoff == accum->rm_offset + accum->rm_blockcount && + agbno == accum->rm_startblock + accum->rm_blockcount && + rmap_flags == accum->rm_flags) { + accum->rm_blockcount += rec->br_blockcount; + return 0; + } + + /* Otherwise stash the old rmap and start accumulating a new one. */ + error = xrep_rmap_stash_accumulated(rf); + if (error) + return error; + + accum->rm_startblock = agbno; + accum->rm_blockcount = rec->br_blockcount; + accum->rm_offset = rec->br_startoff; + accum->rm_flags = rmap_flags; + return 0; +} + +/* Add a btree block to the bitmap. */ +STATIC int +xrep_rmap_visit_iroot_btree_block( + struct xfs_btree_cur *cur, + int level, + void *priv) +{ + struct xrep_rmap_ifork *rf = priv; + struct xfs_buf *bp; + xfs_fsblock_t fsbno; + xfs_agblock_t agbno; + + xfs_btree_get_block(cur, level, &bp); + if (!bp) + return 0; + + fsbno = XFS_DADDR_TO_FSB(cur->bc_mp, xfs_buf_daddr(bp)); + if (XFS_FSB_TO_AGNO(cur->bc_mp, fsbno) != rf->rr->sc->sa.pag->pag_agno) + return 0; + + agbno = XFS_FSB_TO_AGBNO(cur->bc_mp, fsbno); + return xagb_bitmap_set(&rf->bmbt_blocks, agbno, 1); +} + +/* + * Iterate a metadata btree rooted in an inode to collect rmap records for + * anything in this fork that matches the AG. + */ +STATIC int +xrep_rmap_scan_iroot_btree( + struct xrep_rmap_ifork *rf, + struct xfs_btree_cur *cur) +{ + struct xfs_owner_info oinfo; + struct xrep_rmap *rr = rf->rr; + int error; + + xagb_bitmap_init(&rf->bmbt_blocks); + + /* Record all the blocks in the btree itself. */ + error = xfs_btree_visit_blocks(cur, xrep_rmap_visit_iroot_btree_block, + XFS_BTREE_VISIT_ALL, rf); + if (error) + goto out; + + /* Emit rmaps for the btree blocks. */ + xfs_rmap_ino_bmbt_owner(&oinfo, rf->accum.rm_owner, rf->whichfork); + error = xrep_rmap_stash_bitmap(rr, &rf->bmbt_blocks, &oinfo); + if (error) + goto out; + + /* Stash any remaining accumulated rmaps. */ + error = xrep_rmap_stash_accumulated(rf); +out: + xagb_bitmap_destroy(&rf->bmbt_blocks); + return error; +} + +static inline bool +is_rt_data_fork( + struct xfs_inode *ip, + int whichfork) +{ + return XFS_IS_REALTIME_INODE(ip) && whichfork == XFS_DATA_FORK; +} + +/* + * Iterate the block mapping btree to collect rmap records for anything in this + * fork that matches the AG. Sets @mappings_done to true if we've scanned the + * block mappings in this fork. + */ +STATIC int +xrep_rmap_scan_bmbt( + struct xrep_rmap_ifork *rf, + struct xfs_inode *ip, + bool *mappings_done) +{ + struct xrep_rmap *rr = rf->rr; + struct xfs_btree_cur *cur; + struct xfs_ifork *ifp; + int error; + + *mappings_done = false; + ifp = xfs_ifork_ptr(ip, rf->whichfork); + cur = xfs_bmbt_init_cursor(rr->sc->mp, rr->sc->tp, ip, rf->whichfork); + + if (!xfs_ifork_is_realtime(ip, rf->whichfork) && + xfs_need_iread_extents(ifp)) { + /* + * If the incore extent cache isn't loaded, scan the bmbt for + * mapping records. This avoids loading the incore extent + * tree, which will increase memory pressure at a time when + * we're trying to run as quickly as we possibly can. Ignore + * realtime extents. + */ + error = xfs_bmap_query_all(cur, xrep_rmap_visit_bmbt, rf); + if (error) + goto out_cur; + + *mappings_done = true; + } + + /* Scan for the bmbt blocks, which always live on the data device. */ + error = xrep_rmap_scan_iroot_btree(rf, cur); +out_cur: + xfs_btree_del_cursor(cur, error); + return error; +} + +/* + * Iterate the in-core extent cache to collect rmap records for anything in + * this fork that matches the AG. + */ +STATIC int +xrep_rmap_scan_iext( + struct xrep_rmap_ifork *rf, + struct xfs_ifork *ifp) +{ + struct xfs_bmbt_irec rec; + struct xfs_iext_cursor icur; + int error; + + for_each_xfs_iext(ifp, &icur, &rec) { + if (isnullstartblock(rec.br_startblock)) + continue; + error = xrep_rmap_visit_bmbt(NULL, &rec, rf); + if (error) + return error; + } + + return xrep_rmap_stash_accumulated(rf); +} + +/* Find all the extents from a given AG in an inode fork. */ +STATIC int +xrep_rmap_scan_ifork( + struct xrep_rmap *rr, + struct xfs_inode *ip, + int whichfork) +{ + struct xrep_rmap_ifork rf = { + .accum = { .rm_owner = ip->i_ino, }, + .rr = rr, + .whichfork = whichfork, + }; + struct xfs_ifork *ifp = xfs_ifork_ptr(ip, whichfork); + int error = 0; + + if (!ifp) + return 0; + + if (ifp->if_format == XFS_DINODE_FMT_BTREE) { + bool mappings_done; + + /* + * Scan the bmap btree for data device mappings. This includes + * the btree blocks themselves, even if this is a realtime + * file. + */ + error = xrep_rmap_scan_bmbt(&rf, ip, &mappings_done); + if (error || mappings_done) + return error; + } else if (ifp->if_format != XFS_DINODE_FMT_EXTENTS) { + return 0; + } + + /* Scan incore extent cache if this isn't a realtime file. */ + if (xfs_ifork_is_realtime(ip, whichfork)) + return 0; + + return xrep_rmap_scan_iext(&rf, ifp); +} + +/* + * Take ILOCK on a file that we want to scan. + * + * Select ILOCK_EXCL if the file has an unloaded data bmbt or has an unloaded + * attr bmbt. Otherwise, take ILOCK_SHARED. + */ +static inline unsigned int +xrep_rmap_scan_ilock( + struct xfs_inode *ip) +{ + uint lock_mode = XFS_ILOCK_SHARED; + + if (xfs_need_iread_extents(&ip->i_df)) { + lock_mode = XFS_ILOCK_EXCL; + goto lock; + } + + if (xfs_inode_has_attr_fork(ip) && xfs_need_iread_extents(&ip->i_af)) + lock_mode = XFS_ILOCK_EXCL; + +lock: + xfs_ilock(ip, lock_mode); + return lock_mode; +} + +/* Record reverse mappings for a file. */ +STATIC int +xrep_rmap_scan_inode( + struct xrep_rmap *rr, + struct xfs_inode *ip) +{ + unsigned int lock_mode = 0; + int error; + + /* + * Directory updates (create/link/unlink/rename) drop the directory's + * ILOCK before finishing any rmapbt updates associated with directory + * shape changes. For this scan to coordinate correctly with the live + * update hook, we must take the only lock (i_rwsem) that is held all + * the way to dir op completion. This will get fixed by the parent + * pointer patchset. + */ + if (S_ISDIR(VFS_I(ip)->i_mode)) { + lock_mode = XFS_IOLOCK_SHARED; + xfs_ilock(ip, lock_mode); + } + lock_mode |= xrep_rmap_scan_ilock(ip); + + /* Check the data fork. */ + error = xrep_rmap_scan_ifork(rr, ip, XFS_DATA_FORK); + if (error) + goto out_unlock; + + /* Check the attr fork. */ + error = xrep_rmap_scan_ifork(rr, ip, XFS_ATTR_FORK); + if (error) + goto out_unlock; + + /* COW fork extents are "owned" by the refcount btree. */ + + xchk_iscan_mark_visited(&rr->iscan, ip); +out_unlock: + xfs_iunlock(ip, lock_mode); + return error; +} + +/* Section (I): Find all AG metadata extents except for free space metadata. */ + +struct xrep_rmap_inodes { + struct xrep_rmap *rr; + struct xagb_bitmap inobt_blocks; /* INOBIT */ + struct xagb_bitmap ichunk_blocks; /* ICHUNKBIT */ +}; + +/* Record inode btree rmaps. */ +STATIC int +xrep_rmap_walk_inobt( + struct xfs_btree_cur *cur, + const union xfs_btree_rec *rec, + void *priv) +{ + struct xfs_inobt_rec_incore irec; + struct xrep_rmap_inodes *ri = priv; + struct xfs_mount *mp = cur->bc_mp; + xfs_agblock_t agbno; + xfs_extlen_t aglen; + xfs_agino_t agino; + xfs_agino_t iperhole; + unsigned int i; + int error; + + /* Record the inobt blocks. */ + error = xagb_bitmap_set_btcur_path(&ri->inobt_blocks, cur); + if (error) + return error; + + xfs_inobt_btrec_to_irec(mp, rec, &irec); + if (xfs_inobt_check_irec(cur->bc_ag.pag, &irec) != NULL) + return -EFSCORRUPTED; + + agino = irec.ir_startino; + + /* Record a non-sparse inode chunk. */ + if (!xfs_inobt_issparse(irec.ir_holemask)) { + agbno = XFS_AGINO_TO_AGBNO(mp, agino); + aglen = max_t(xfs_extlen_t, 1, + XFS_INODES_PER_CHUNK / mp->m_sb.sb_inopblock); + + return xagb_bitmap_set(&ri->ichunk_blocks, agbno, aglen); + } + + /* Iterate each chunk. */ + iperhole = max_t(xfs_agino_t, mp->m_sb.sb_inopblock, + XFS_INODES_PER_HOLEMASK_BIT); + aglen = iperhole / mp->m_sb.sb_inopblock; + for (i = 0, agino = irec.ir_startino; + i < XFS_INOBT_HOLEMASK_BITS; + i += iperhole / XFS_INODES_PER_HOLEMASK_BIT, agino += iperhole) { + /* Skip holes. */ + if (irec.ir_holemask & (1 << i)) + continue; + + /* Record the inode chunk otherwise. */ + agbno = XFS_AGINO_TO_AGBNO(mp, agino); + error = xagb_bitmap_set(&ri->ichunk_blocks, agbno, aglen); + if (error) + return error; + } + + return 0; +} + +/* Collect rmaps for the blocks containing inode btrees and the inode chunks. */ +STATIC int +xrep_rmap_find_inode_rmaps( + struct xrep_rmap *rr) +{ + struct xrep_rmap_inodes ri = { + .rr = rr, + }; + struct xfs_scrub *sc = rr->sc; + int error; + + xagb_bitmap_init(&ri.inobt_blocks); + xagb_bitmap_init(&ri.ichunk_blocks); + + /* + * Iterate every record in the inobt so we can capture all the inode + * chunks and the blocks in the inobt itself. + */ + error = xfs_btree_query_all(sc->sa.ino_cur, xrep_rmap_walk_inobt, &ri); + if (error) + goto out_bitmap; + + /* + * Note that if there are zero records in the inobt then query_all does + * nothing and we have to account the empty inobt root manually. + */ + if (xagb_bitmap_empty(&ri.ichunk_blocks)) { + struct xfs_agi *agi = sc->sa.agi_bp->b_addr; + + error = xagb_bitmap_set(&ri.inobt_blocks, + be32_to_cpu(agi->agi_root), 1); + if (error) + goto out_bitmap; + } + + /* Scan the finobt too. */ + if (xfs_has_finobt(sc->mp)) { + error = xagb_bitmap_set_btblocks(&ri.inobt_blocks, + sc->sa.fino_cur); + if (error) + goto out_bitmap; + } + + /* Generate rmaps for everything. */ + error = xrep_rmap_stash_bitmap(rr, &ri.inobt_blocks, + &XFS_RMAP_OINFO_INOBT); + if (error) + goto out_bitmap; + error = xrep_rmap_stash_bitmap(rr, &ri.ichunk_blocks, + &XFS_RMAP_OINFO_INODES); + +out_bitmap: + xagb_bitmap_destroy(&ri.inobt_blocks); + xagb_bitmap_destroy(&ri.ichunk_blocks); + return error; +} + +/* Record a CoW staging extent. */ +STATIC int +xrep_rmap_walk_cowblocks( + struct xfs_btree_cur *cur, + const struct xfs_refcount_irec *irec, + void *priv) +{ + struct xagb_bitmap *bitmap = priv; + + if (!xfs_refcount_check_domain(irec) || + irec->rc_domain != XFS_REFC_DOMAIN_COW) + return -EFSCORRUPTED; + + return xagb_bitmap_set(bitmap, irec->rc_startblock, irec->rc_blockcount); +} + +/* + * Collect rmaps for the blocks containing the refcount btree, and all CoW + * staging extents. + */ +STATIC int +xrep_rmap_find_refcount_rmaps( + struct xrep_rmap *rr) +{ + struct xagb_bitmap refcountbt_blocks; /* REFCBIT */ + struct xagb_bitmap cow_blocks; /* COWBIT */ + struct xfs_refcount_irec low = { + .rc_startblock = 0, + .rc_domain = XFS_REFC_DOMAIN_COW, + }; + struct xfs_refcount_irec high = { + .rc_startblock = -1U, + .rc_domain = XFS_REFC_DOMAIN_COW, + }; + struct xfs_scrub *sc = rr->sc; + int error; + + if (!xfs_has_reflink(sc->mp)) + return 0; + + xagb_bitmap_init(&refcountbt_blocks); + xagb_bitmap_init(&cow_blocks); + + /* refcountbt */ + error = xagb_bitmap_set_btblocks(&refcountbt_blocks, sc->sa.refc_cur); + if (error) + goto out_bitmap; + + /* Collect rmaps for CoW staging extents. */ + error = xfs_refcount_query_range(sc->sa.refc_cur, &low, &high, + xrep_rmap_walk_cowblocks, &cow_blocks); + if (error) + goto out_bitmap; + + /* Generate rmaps for everything. */ + error = xrep_rmap_stash_bitmap(rr, &cow_blocks, &XFS_RMAP_OINFO_COW); + if (error) + goto out_bitmap; + error = xrep_rmap_stash_bitmap(rr, &refcountbt_blocks, + &XFS_RMAP_OINFO_REFC); + +out_bitmap: + xagb_bitmap_destroy(&cow_blocks); + xagb_bitmap_destroy(&refcountbt_blocks); + return error; +} + +/* Generate rmaps for the AG headers (AGI/AGF/AGFL) */ +STATIC int +xrep_rmap_find_agheader_rmaps( + struct xrep_rmap *rr) +{ + struct xfs_scrub *sc = rr->sc; + + /* Create a record for the AG sb->agfl. */ + return xrep_rmap_stash(rr, XFS_SB_BLOCK(sc->mp), + XFS_AGFL_BLOCK(sc->mp) - XFS_SB_BLOCK(sc->mp) + 1, + XFS_RMAP_OWN_FS, 0, 0); +} + +/* Generate rmaps for the log, if it's in this AG. */ +STATIC int +xrep_rmap_find_log_rmaps( + struct xrep_rmap *rr) +{ + struct xfs_scrub *sc = rr->sc; + + if (!xfs_ag_contains_log(sc->mp, sc->sa.pag->pag_agno)) + return 0; + + return xrep_rmap_stash(rr, + XFS_FSB_TO_AGBNO(sc->mp, sc->mp->m_sb.sb_logstart), + sc->mp->m_sb.sb_logblocks, XFS_RMAP_OWN_LOG, 0, 0); +} + +/* Check and count all the records that we gathered. */ +STATIC int +xrep_rmap_check_record( + struct xfs_btree_cur *cur, + const struct xfs_rmap_irec *rec, + void *priv) +{ + struct xrep_rmap *rr = priv; + int error; + + error = xrep_rmap_check_mapping(rr->sc, rec); + if (error) + return error; + + rr->nr_records++; + return 0; +} + +/* + * Generate all the reverse-mappings for this AG, a list of the old rmapbt + * blocks, and the new btreeblks count. Figure out if we have enough free + * space to reconstruct the inode btrees. The caller must clean up the lists + * if anything goes wrong. This implements section (I) above. + */ +STATIC int +xrep_rmap_find_rmaps( + struct xrep_rmap *rr) +{ + struct xfs_scrub *sc = rr->sc; + struct xchk_ag *sa = &sc->sa; + struct xfs_inode *ip; + struct xfs_btree_cur *mcur; + int error; + + /* Find all the per-AG metadata. */ + xrep_ag_btcur_init(sc, &sc->sa); + + error = xrep_rmap_find_inode_rmaps(rr); + if (error) + goto end_agscan; + + error = xrep_rmap_find_refcount_rmaps(rr); + if (error) + goto end_agscan; + + error = xrep_rmap_find_agheader_rmaps(rr); + if (error) + goto end_agscan; + + error = xrep_rmap_find_log_rmaps(rr); +end_agscan: + xchk_ag_btcur_free(&sc->sa); + if (error) + return error; + + /* + * Set up for a potentially lengthy filesystem scan by reducing our + * transaction resource usage for the duration. Specifically: + * + * Unlock the AG header buffers and cancel the transaction to release + * the log grant space while we scan the filesystem. + * + * Create a new empty transaction to eliminate the possibility of the + * inode scan deadlocking on cyclical metadata. + * + * We pass the empty transaction to the file scanning function to avoid + * repeatedly cycling empty transactions. This can be done even though + * we take the IOLOCK to quiesce the file because empty transactions + * do not take sb_internal. + */ + sa->agf_bp = NULL; + sa->agi_bp = NULL; + xchk_trans_cancel(sc); + error = xchk_trans_alloc_empty(sc); + if (error) + return error; + + /* Iterate all AGs for inodes rmaps. */ + while ((error = xchk_iscan_iter(&rr->iscan, &ip)) == 1) { + error = xrep_rmap_scan_inode(rr, ip); + xchk_irele(sc, ip); + if (error) + break; + + if (xchk_should_terminate(sc, &error)) + break; + } + xchk_iscan_iter_finish(&rr->iscan); + if (error) + return error; + + /* + * Switch out for a real transaction and lock the AG headers in + * preparation for building a new tree. + */ + xchk_trans_cancel(sc); + error = xchk_setup_fs(sc); + if (error) + return error; + error = xchk_perag_drain_and_lock(sc); + if (error) + return error; + + /* + * If a hook failed to update the in-memory btree, we lack the data to + * continue the repair. + */ + if (xchk_iscan_aborted(&rr->iscan)) + return -EFSCORRUPTED; + + /* + * Now that we have everything locked again, we need to count the + * number of rmap records stashed in the btree. This should reflect + * all actively-owned space in the filesystem. At the same time, check + * all our records before we start building a new btree, which requires + * a bnobt cursor. + */ + mcur = xfs_rmapbt_mem_cursor(rr->sc->sa.pag, NULL, &rr->rmap_btree); + sc->sa.bno_cur = xfs_bnobt_init_cursor(sc->mp, sc->tp, sc->sa.agf_bp, + sc->sa.pag); + + rr->nr_records = 0; + error = xfs_rmap_query_all(mcur, xrep_rmap_check_record, rr); + + xfs_btree_del_cursor(sc->sa.bno_cur, error); + sc->sa.bno_cur = NULL; + xfs_btree_del_cursor(mcur, error); + + return error; +} + +/* Section (II): Reserving space for new rmapbt and setting free space bitmap */ + +struct xrep_rmap_agfl { + struct xagb_bitmap *bitmap; + xfs_agnumber_t agno; +}; + +/* Add an AGFL block to the rmap list. */ +STATIC int +xrep_rmap_walk_agfl( + struct xfs_mount *mp, + xfs_agblock_t agbno, + void *priv) +{ + struct xrep_rmap_agfl *ra = priv; + + return xagb_bitmap_set(ra->bitmap, agbno, 1); +} + +/* + * Run one round of reserving space for the new rmapbt and recomputing the + * number of blocks needed to store the previously observed rmapbt records and + * the ones we'll create for the free space metadata. When we don't need more + * blocks, return a bitmap of OWN_AG extents in @freesp_blocks and set @done to + * true. + */ +STATIC int +xrep_rmap_try_reserve( + struct xrep_rmap *rr, + struct xfs_btree_cur *rmap_cur, + struct xagb_bitmap *freesp_blocks, + uint64_t *blocks_reserved, + bool *done) +{ + struct xrep_rmap_agfl ra = { + .bitmap = freesp_blocks, + .agno = rr->sc->sa.pag->pag_agno, + }; + struct xfs_scrub *sc = rr->sc; + struct xrep_newbt_resv *resv, *n; + struct xfs_agf *agf = sc->sa.agf_bp->b_addr; + struct xfs_buf *agfl_bp; + uint64_t nr_blocks; /* RMB */ + uint64_t freesp_records; + int error; + + /* + * We're going to recompute new_btree.bload.nr_blocks at the end of + * this function to reflect however many btree blocks we need to store + * all the rmap records (including the ones that reflect the changes we + * made to support the new rmapbt blocks), so we save the old value + * here so we can decide if we've reserved enough blocks. + */ + nr_blocks = rr->new_btree.bload.nr_blocks; + + /* + * Make sure we've reserved enough space for the new btree. This can + * change the shape of the free space btrees, which can cause secondary + * interactions with the rmap records because all three space btrees + * have the same rmap owner. We'll account for all that below. + */ + error = xrep_newbt_alloc_blocks(&rr->new_btree, + nr_blocks - *blocks_reserved); + if (error) + return error; + + *blocks_reserved = rr->new_btree.bload.nr_blocks; + + /* Clear everything in the bitmap. */ + xagb_bitmap_destroy(freesp_blocks); + + /* Set all the bnobt blocks in the bitmap. */ + sc->sa.bno_cur = xfs_bnobt_init_cursor(sc->mp, sc->tp, sc->sa.agf_bp, + sc->sa.pag); + error = xagb_bitmap_set_btblocks(freesp_blocks, sc->sa.bno_cur); + xfs_btree_del_cursor(sc->sa.bno_cur, error); + sc->sa.bno_cur = NULL; + if (error) + return error; + + /* Set all the cntbt blocks in the bitmap. */ + sc->sa.cnt_cur = xfs_cntbt_init_cursor(sc->mp, sc->tp, sc->sa.agf_bp, + sc->sa.pag); + error = xagb_bitmap_set_btblocks(freesp_blocks, sc->sa.cnt_cur); + xfs_btree_del_cursor(sc->sa.cnt_cur, error); + sc->sa.cnt_cur = NULL; + if (error) + return error; + + /* Record our new btreeblks value. */ + rr->freesp_btblocks = xagb_bitmap_hweight(freesp_blocks) - 2; + + /* Set all the new rmapbt blocks in the bitmap. */ + list_for_each_entry_safe(resv, n, &rr->new_btree.resv_list, list) { + error = xagb_bitmap_set(freesp_blocks, resv->agbno, resv->len); + if (error) + return error; + } + + /* Set all the AGFL blocks in the bitmap. */ + error = xfs_alloc_read_agfl(sc->sa.pag, sc->tp, &agfl_bp); + if (error) + return error; + + error = xfs_agfl_walk(sc->mp, agf, agfl_bp, xrep_rmap_walk_agfl, &ra); + if (error) + return error; + + /* Count the extents in the bitmap. */ + freesp_records = xagb_bitmap_count_set_regions(freesp_blocks); + + /* Compute how many blocks we'll need for all the rmaps. */ + error = xfs_btree_bload_compute_geometry(rmap_cur, + &rr->new_btree.bload, rr->nr_records + freesp_records); + if (error) + return error; + + /* We're done when we don't need more blocks. */ + *done = nr_blocks >= rr->new_btree.bload.nr_blocks; + return 0; +} + +/* + * Iteratively reserve space for rmap btree while recording OWN_AG rmaps for + * the free space metadata. This implements section (II) above. + */ +STATIC int +xrep_rmap_reserve_space( + struct xrep_rmap *rr, + struct xfs_btree_cur *rmap_cur) +{ + struct xagb_bitmap freesp_blocks; /* AGBIT */ + uint64_t blocks_reserved = 0; + bool done = false; + int error; + + /* Compute how many blocks we'll need for the rmaps collected so far. */ + error = xfs_btree_bload_compute_geometry(rmap_cur, + &rr->new_btree.bload, rr->nr_records); + if (error) + return error; + + /* Last chance to abort before we start committing fixes. */ + if (xchk_should_terminate(rr->sc, &error)) + return error; + + xagb_bitmap_init(&freesp_blocks); + + /* + * Iteratively reserve space for the new rmapbt and recompute the + * number of blocks needed to store the previously observed rmapbt + * records and the ones we'll create for the free space metadata. + * Finish when we don't need more blocks. + */ + do { + error = xrep_rmap_try_reserve(rr, rmap_cur, &freesp_blocks, + &blocks_reserved, &done); + if (error) + goto out_bitmap; + } while (!done); + + /* Emit rmaps for everything in the free space bitmap. */ + xrep_ag_btcur_init(rr->sc, &rr->sc->sa); + error = xrep_rmap_stash_bitmap(rr, &freesp_blocks, &XFS_RMAP_OINFO_AG); + xchk_ag_btcur_free(&rr->sc->sa); + +out_bitmap: + xagb_bitmap_destroy(&freesp_blocks); + return error; +} + +/* Section (III): Building the new rmap btree. */ + +/* Update the AGF counters. */ +STATIC int +xrep_rmap_reset_counters( + struct xrep_rmap *rr) +{ + struct xfs_scrub *sc = rr->sc; + struct xfs_perag *pag = sc->sa.pag; + struct xfs_agf *agf = sc->sa.agf_bp->b_addr; + xfs_agblock_t rmap_btblocks; + + /* + * The AGF header contains extra information related to the reverse + * mapping btree, so we must update those fields here. + */ + rmap_btblocks = rr->new_btree.afake.af_blocks - 1; + agf->agf_btreeblks = cpu_to_be32(rr->freesp_btblocks + rmap_btblocks); + xfs_alloc_log_agf(sc->tp, sc->sa.agf_bp, XFS_AGF_BTREEBLKS); + + /* + * After we commit the new btree to disk, it is possible that the + * process to reap the old btree blocks will race with the AIL trying + * to checkpoint the old btree blocks into the filesystem. If the new + * tree is shorter than the old one, the rmapbt write verifier will + * fail and the AIL will shut down the filesystem. + * + * To avoid this, save the old incore btree height values as the alt + * height values before re-initializing the perag info from the updated + * AGF to capture all the new values. + */ + pag->pagf_repair_rmap_level = pag->pagf_rmap_level; + + /* Reinitialize with the values we just logged. */ + return xrep_reinit_pagf(sc); +} + +/* Retrieve rmapbt data for bulk load. */ +STATIC int +xrep_rmap_get_records( + struct xfs_btree_cur *cur, + unsigned int idx, + struct xfs_btree_block *block, + unsigned int nr_wanted, + void *priv) +{ + struct xrep_rmap *rr = priv; + union xfs_btree_rec *block_rec; + unsigned int loaded; + int error; + + for (loaded = 0; loaded < nr_wanted; loaded++, idx++) { + int stat = 0; + + error = xfs_btree_increment(rr->mcur, 0, &stat); + if (error) + return error; + if (!stat) + return -EFSCORRUPTED; + + error = xfs_rmap_get_rec(rr->mcur, &cur->bc_rec.r, &stat); + if (error) + return error; + if (!stat) + return -EFSCORRUPTED; + + block_rec = xfs_btree_rec_addr(cur, idx, block); + cur->bc_ops->init_rec_from_cur(cur, block_rec); + } + + return loaded; +} + +/* Feed one of the new btree blocks to the bulk loader. */ +STATIC int +xrep_rmap_claim_block( + struct xfs_btree_cur *cur, + union xfs_btree_ptr *ptr, + void *priv) +{ + struct xrep_rmap *rr = priv; + + return xrep_newbt_claim_block(cur, &rr->new_btree, ptr); +} + +/* Custom allocation function for new rmap btrees. */ +STATIC int +xrep_rmap_alloc_vextent( + struct xfs_scrub *sc, + struct xfs_alloc_arg *args, + xfs_fsblock_t alloc_hint) +{ + int error; + + /* + * We don't want an rmap update on the allocation, since we iteratively + * compute the OWN_AG records /after/ allocating blocks for the records + * that we already know we need to store. Therefore, fix the freelist + * with the NORMAP flag set so that we don't also try to create an rmap + * for new AGFL blocks. + */ + error = xrep_fix_freelist(sc, XFS_ALLOC_FLAG_NORMAP); + if (error) + return error; + + /* + * If xrep_fix_freelist fixed the freelist by moving blocks from the + * free space btrees or by removing blocks from the AGFL and queueing + * an EFI to free the block, the transaction will be dirty. This + * second case is of interest to us. + * + * Later on, we will need to compare gaps in the new recordset against + * the block usage of all OWN_AG owners in order to free the old + * btree's blocks, which means that we can't have EFIs for former AGFL + * blocks attached to the repair transaction when we commit the new + * btree. + * + * xrep_newbt_alloc_blocks guarantees this for us by calling + * xrep_defer_finish to commit anything that fix_freelist may have + * added to the transaction. + */ + return xfs_alloc_vextent_near_bno(args, alloc_hint); +} + + +/* Count the records in this btree. */ +STATIC int +xrep_rmap_count_records( + struct xfs_btree_cur *cur, + unsigned long long *nr) +{ + int running = 1; + int error; + + *nr = 0; + + error = xfs_btree_goto_left_edge(cur); + if (error) + return error; + + while (running && !(error = xfs_btree_increment(cur, 0, &running))) { + if (running) + (*nr)++; + } + + return error; +} +/* + * Use the collected rmap information to stage a new rmap btree. If this is + * successful we'll return with the new btree root information logged to the + * repair transaction but not yet committed. This implements section (III) + * above. + */ +STATIC int +xrep_rmap_build_new_tree( + struct xrep_rmap *rr) +{ + struct xfs_scrub *sc = rr->sc; + struct xfs_perag *pag = sc->sa.pag; + struct xfs_agf *agf = sc->sa.agf_bp->b_addr; + struct xfs_btree_cur *rmap_cur; + xfs_fsblock_t fsbno; + int error; + + /* + * Preserve the old rmapbt block count so that we can adjust the + * per-AG rmapbt reservation after we commit the new btree root and + * want to dispose of the old btree blocks. + */ + rr->old_rmapbt_fsbcount = be32_to_cpu(agf->agf_rmap_blocks); + + /* + * Prepare to construct the new btree by reserving disk space for the + * new btree and setting up all the accounting information we'll need + * to root the new btree while it's under construction and before we + * attach it to the AG header. The new blocks are accounted to the + * rmapbt per-AG reservation, which we will adjust further after + * committing the new btree. + */ + fsbno = XFS_AGB_TO_FSB(sc->mp, pag->pag_agno, XFS_RMAP_BLOCK(sc->mp)); + xrep_newbt_init_ag(&rr->new_btree, sc, &XFS_RMAP_OINFO_SKIP_UPDATE, + fsbno, XFS_AG_RESV_RMAPBT); + rr->new_btree.bload.get_records = xrep_rmap_get_records; + rr->new_btree.bload.claim_block = xrep_rmap_claim_block; + rr->new_btree.alloc_vextent = xrep_rmap_alloc_vextent; + rmap_cur = xfs_rmapbt_init_cursor(sc->mp, NULL, NULL, pag); + xfs_btree_stage_afakeroot(rmap_cur, &rr->new_btree.afake); + + /* + * Initialize @rr->new_btree, reserve space for the new rmapbt, + * and compute OWN_AG rmaps. + */ + error = xrep_rmap_reserve_space(rr, rmap_cur); + if (error) + goto err_cur; + + /* + * Count the rmapbt records again, because the space reservation + * for the rmapbt itself probably added more records to the btree. + */ + rr->mcur = xfs_rmapbt_mem_cursor(rr->sc->sa.pag, NULL, + &rr->rmap_btree); + + error = xrep_rmap_count_records(rr->mcur, &rr->nr_records); + if (error) + goto err_mcur; + + /* + * Due to btree slack factors, it's possible for a new btree to be one + * level taller than the old btree. Update the incore btree height so + * that we don't trip the verifiers when writing the new btree blocks + * to disk. + */ + pag->pagf_repair_rmap_level = rr->new_btree.bload.btree_height; + + /* + * Move the cursor to the left edge of the tree so that the first + * increment in ->get_records positions us at the first record. + */ + error = xfs_btree_goto_left_edge(rr->mcur); + if (error) + goto err_level; + + /* Add all observed rmap records. */ + error = xfs_btree_bload(rmap_cur, &rr->new_btree.bload, rr); + if (error) + goto err_level; + + /* + * Install the new btree in the AG header. After this point the old + * btree is no longer accessible and the new tree is live. + */ + xfs_rmapbt_commit_staged_btree(rmap_cur, sc->tp, sc->sa.agf_bp); + xfs_btree_del_cursor(rmap_cur, 0); + xfs_btree_del_cursor(rr->mcur, 0); + rr->mcur = NULL; + + /* + * Now that we've written the new btree to disk, we don't need to keep + * updating the in-memory btree. Abort the scan to stop live updates. + */ + xchk_iscan_abort(&rr->iscan); + + /* + * The newly committed rmap recordset includes mappings for the blocks + * that we reserved to build the new btree. If there is excess space + * reservation to be freed, the corresponding rmap records must also be + * removed. + */ + rr->new_btree.oinfo = XFS_RMAP_OINFO_AG; + + /* Reset the AGF counters now that we've changed the btree shape. */ + error = xrep_rmap_reset_counters(rr); + if (error) + goto err_newbt; + + /* Dispose of any unused blocks and the accounting information. */ + error = xrep_newbt_commit(&rr->new_btree); + if (error) + return error; + + return xrep_roll_ag_trans(sc); + +err_level: + pag->pagf_repair_rmap_level = 0; +err_mcur: + xfs_btree_del_cursor(rr->mcur, error); +err_cur: + xfs_btree_del_cursor(rmap_cur, error); +err_newbt: + xrep_newbt_cancel(&rr->new_btree); + return error; +} + +/* Section (IV): Reaping the old btree. */ + +struct xrep_rmap_find_gaps { + struct xagb_bitmap rmap_gaps; + xfs_agblock_t next_agbno; +}; + +/* Subtract each free extent in the bnobt from the rmap gaps. */ +STATIC int +xrep_rmap_find_freesp( + struct xfs_btree_cur *cur, + const struct xfs_alloc_rec_incore *rec, + void *priv) +{ + struct xrep_rmap_find_gaps *rfg = priv; + + return xagb_bitmap_clear(&rfg->rmap_gaps, rec->ar_startblock, + rec->ar_blockcount); +} + +/* Record the free space we find, as part of cleaning out the btree. */ +STATIC int +xrep_rmap_find_gaps( + struct xfs_btree_cur *cur, + const struct xfs_rmap_irec *rec, + void *priv) +{ + struct xrep_rmap_find_gaps *rfg = priv; + int error; + + if (rec->rm_startblock > rfg->next_agbno) { + error = xagb_bitmap_set(&rfg->rmap_gaps, rfg->next_agbno, + rec->rm_startblock - rfg->next_agbno); + if (error) + return error; + } + + rfg->next_agbno = max_t(xfs_agblock_t, rfg->next_agbno, + rec->rm_startblock + rec->rm_blockcount); + return 0; +} + +/* + * Reap the old rmapbt blocks. Now that the rmapbt is fully rebuilt, we make + * a list of gaps in the rmap records and a list of the extents mentioned in + * the bnobt. Any block that's in the new rmapbt gap list but not mentioned + * in the bnobt is a block from the old rmapbt and can be removed. + */ +STATIC int +xrep_rmap_remove_old_tree( + struct xrep_rmap *rr) +{ + struct xrep_rmap_find_gaps rfg = { + .next_agbno = 0, + }; + struct xfs_scrub *sc = rr->sc; + struct xfs_agf *agf = sc->sa.agf_bp->b_addr; + struct xfs_perag *pag = sc->sa.pag; + struct xfs_btree_cur *mcur; + xfs_agblock_t agend; + int error; + + xagb_bitmap_init(&rfg.rmap_gaps); + + /* Compute free space from the new rmapbt. */ + mcur = xfs_rmapbt_mem_cursor(rr->sc->sa.pag, NULL, &rr->rmap_btree); + + error = xfs_rmap_query_all(mcur, xrep_rmap_find_gaps, &rfg); + xfs_btree_del_cursor(mcur, error); + if (error) + goto out_bitmap; + + /* Insert a record for space between the last rmap and EOAG. */ + agend = be32_to_cpu(agf->agf_length); + if (rfg.next_agbno < agend) { + error = xagb_bitmap_set(&rfg.rmap_gaps, rfg.next_agbno, + agend - rfg.next_agbno); + if (error) + goto out_bitmap; + } + + /* Compute free space from the existing bnobt. */ + sc->sa.bno_cur = xfs_bnobt_init_cursor(sc->mp, sc->tp, sc->sa.agf_bp, + sc->sa.pag); + error = xfs_alloc_query_all(sc->sa.bno_cur, xrep_rmap_find_freesp, + &rfg); + xfs_btree_del_cursor(sc->sa.bno_cur, error); + sc->sa.bno_cur = NULL; + if (error) + goto out_bitmap; + + /* + * Free the "free" blocks that the new rmapbt knows about but the bnobt + * doesn't--these are the old rmapbt blocks. Credit the old rmapbt + * block usage count back to the per-AG rmapbt reservation (and not + * fdblocks, since the rmap btree lives in free space) to keep the + * reservation and free space accounting correct. + */ + error = xrep_reap_agblocks(sc, &rfg.rmap_gaps, + &XFS_RMAP_OINFO_ANY_OWNER, XFS_AG_RESV_RMAPBT); + if (error) + goto out_bitmap; + + /* + * Now that we've zapped all the old rmapbt blocks we can turn off + * the alternate height mechanism and reset the per-AG space + * reservation. + */ + pag->pagf_repair_rmap_level = 0; + sc->flags |= XREP_RESET_PERAG_RESV; +out_bitmap: + xagb_bitmap_destroy(&rfg.rmap_gaps); + return error; +} + +static inline bool +xrep_rmapbt_want_live_update( + struct xchk_iscan *iscan, + const struct xfs_owner_info *oi) +{ + if (xchk_iscan_aborted(iscan)) + return false; + + /* + * Before unlocking the AG header to perform the inode scan, we + * recorded reverse mappings for all AG metadata except for the OWN_AG + * metadata. IOWs, the in-memory btree knows about the AG headers, the + * two inode btrees, the CoW staging extents, and the refcount btrees. + * For these types of metadata, we need to record the live updates in + * the in-memory rmap btree. + * + * However, we do not scan the free space btrees or the AGFL until we + * have re-locked the AGF and are ready to reserve space for the new + * rmap btree, so we do not want live updates for OWN_AG metadata. + */ + if (XFS_RMAP_NON_INODE_OWNER(oi->oi_owner)) + return oi->oi_owner != XFS_RMAP_OWN_AG; + + /* Ignore updates to files that the scanner hasn't visited yet. */ + return xchk_iscan_want_live_update(iscan, oi->oi_owner); +} + +/* + * Apply a rmapbt update from the regular filesystem into our shadow btree. + * We're running from the thread that owns the AGF buffer and is generating + * the update, so we must be careful about which parts of the struct xrep_rmap + * that we change. + */ +static int +xrep_rmapbt_live_update( + struct notifier_block *nb, + unsigned long action, + void *data) +{ + struct xfs_rmap_update_params *p = data; + struct xrep_rmap *rr; + struct xfs_mount *mp; + struct xfs_btree_cur *mcur; + struct xfs_trans *tp; + void *txcookie; + int error; + + rr = container_of(nb, struct xrep_rmap, rhook.rmap_hook.nb); + mp = rr->sc->mp; + + if (!xrep_rmapbt_want_live_update(&rr->iscan, &p->oinfo)) + goto out_unlock; + + trace_xrep_rmap_live_update(mp, rr->sc->sa.pag->pag_agno, action, p); + + error = xrep_trans_alloc_hook_dummy(mp, &txcookie, &tp); + if (error) + goto out_abort; + + mutex_lock(&rr->lock); + mcur = xfs_rmapbt_mem_cursor(rr->sc->sa.pag, tp, &rr->rmap_btree); + error = __xfs_rmap_finish_intent(mcur, action, p->startblock, + p->blockcount, &p->oinfo, p->unwritten); + xfs_btree_del_cursor(mcur, error); + if (error) + goto out_cancel; + + error = xfbtree_trans_commit(&rr->rmap_btree, tp); + if (error) + goto out_cancel; + + xrep_trans_cancel_hook_dummy(&txcookie, tp); + mutex_unlock(&rr->lock); + return NOTIFY_DONE; + +out_cancel: + xfbtree_trans_cancel(&rr->rmap_btree, tp); + xrep_trans_cancel_hook_dummy(&txcookie, tp); +out_abort: + mutex_unlock(&rr->lock); + xchk_iscan_abort(&rr->iscan); +out_unlock: + return NOTIFY_DONE; +} + +/* Set up the filesystem scan components. */ +STATIC int +xrep_rmap_setup_scan( + struct xrep_rmap *rr) +{ + struct xfs_scrub *sc = rr->sc; + int error; + + mutex_init(&rr->lock); + + /* Set up in-memory rmap btree */ + error = xfs_rmapbt_mem_init(sc->mp, &rr->rmap_btree, sc->xmbtp, + sc->sa.pag->pag_agno); + if (error) + goto out_mutex; + + /* Retry iget every tenth of a second for up to 30 seconds. */ + xchk_iscan_start(sc, 30000, 100, &rr->iscan); + + /* + * Hook into live rmap operations so that we can update our in-memory + * btree to reflect live changes on the filesystem. Since we drop the + * AGF buffer to scan all the inodes, we need this piece to avoid + * installing a stale btree. + */ + ASSERT(sc->flags & XCHK_FSGATES_RMAP); + xfs_rmap_hook_setup(&rr->rhook, xrep_rmapbt_live_update); + error = xfs_rmap_hook_add(sc->sa.pag, &rr->rhook); + if (error) + goto out_iscan; + return 0; + +out_iscan: + xchk_iscan_teardown(&rr->iscan); + xfbtree_destroy(&rr->rmap_btree); +out_mutex: + mutex_destroy(&rr->lock); + return error; +} + +/* Tear down scan components. */ +STATIC void +xrep_rmap_teardown( + struct xrep_rmap *rr) +{ + struct xfs_scrub *sc = rr->sc; + + xchk_iscan_abort(&rr->iscan); + xfs_rmap_hook_del(sc->sa.pag, &rr->rhook); + xchk_iscan_teardown(&rr->iscan); + xfbtree_destroy(&rr->rmap_btree); + mutex_destroy(&rr->lock); +} + +/* Repair the rmap btree for some AG. */ +int +xrep_rmapbt( + struct xfs_scrub *sc) +{ + struct xrep_rmap *rr = sc->buf; + int error; + + error = xrep_rmap_setup_scan(rr); + if (error) + return error; + + /* + * Collect rmaps for everything in this AG that isn't space metadata. + * These rmaps won't change even as we try to allocate blocks. + */ + error = xrep_rmap_find_rmaps(rr); + if (error) + goto out_records; + + /* Rebuild the rmap information. */ + error = xrep_rmap_build_new_tree(rr); + if (error) + goto out_records; + + /* Kill the old tree. */ + error = xrep_rmap_remove_old_tree(rr); + if (error) + goto out_records; + +out_records: + xrep_rmap_teardown(rr); + return error; +} diff --git a/fs/xfs/scrub/rtsummary.c b/fs/xfs/scrub/rtsummary.c index b1ff4f33324a..5055092bd9e8 100644 --- a/fs/xfs/scrub/rtsummary.c +++ b/fs/xfs/scrub/rtsummary.c @@ -119,7 +119,7 @@ xfsum_load( xfs_rtsumoff_t sumoff, union xfs_suminfo_raw *rawinfo) { - return xfile_obj_load(sc->xfile, rawinfo, + return xfile_load(sc->xfile, rawinfo, sizeof(union xfs_suminfo_raw), sumoff << XFS_WORDLOG); } @@ -130,7 +130,7 @@ xfsum_store( xfs_rtsumoff_t sumoff, const union xfs_suminfo_raw rawinfo) { - return xfile_obj_store(sc->xfile, &rawinfo, + return xfile_store(sc->xfile, &rawinfo, sizeof(union xfs_suminfo_raw), sumoff << XFS_WORDLOG); } @@ -142,7 +142,7 @@ xfsum_copyout( union xfs_suminfo_raw *rawinfo, unsigned int nr_words) { - return xfile_obj_load(sc->xfile, rawinfo, nr_words << XFS_WORDLOG, + return xfile_load(sc->xfile, rawinfo, nr_words << XFS_WORDLOG, sumoff << XFS_WORDLOG); } diff --git a/fs/xfs/scrub/scrub.c b/fs/xfs/scrub/scrub.c index caf324c2b991..20fac9723c08 100644 --- a/fs/xfs/scrub/scrub.c +++ b/fs/xfs/scrub/scrub.c @@ -15,6 +15,8 @@ #include "xfs_quota.h" #include "xfs_qm.h" #include "xfs_scrub.h" +#include "xfs_buf_mem.h" +#include "xfs_rmap.h" #include "scrub/scrub.h" #include "scrub/common.h" #include "scrub/trace.h" @@ -157,6 +159,15 @@ xchk_fsgates_disable( if (sc->flags & XCHK_FSGATES_DRAIN) xfs_drain_wait_disable(); + if (sc->flags & XCHK_FSGATES_QUOTA) + xfs_dqtrx_hook_disable(); + + if (sc->flags & XCHK_FSGATES_DIRENTS) + xfs_dir_hook_disable(); + + if (sc->flags & XCHK_FSGATES_RMAP) + xfs_rmap_hook_disable(); + sc->flags &= ~XCHK_FSGATES_ALL; } @@ -184,6 +195,10 @@ xchk_teardown( sc->flags &= ~XCHK_HAVE_FREEZE_PROT; mnt_drop_write_file(sc->file); } + if (sc->xmbtp) { + xmbuf_free(sc->xmbtp); + sc->xmbtp = NULL; + } if (sc->xfile) { xfile_destroy(sc->xfile); sc->xfile = NULL; @@ -267,7 +282,7 @@ static const struct xchk_meta_ops meta_scrub_ops[] = { .setup = xchk_setup_ag_rmapbt, .scrub = xchk_rmapbt, .has = xfs_has_rmapbt, - .repair = xrep_notsupported, + .repair = xrep_rmapbt, }, [XFS_SCRUB_TYPE_REFCNTBT] = { /* refcountbt */ .type = ST_PERAG, @@ -358,7 +373,25 @@ static const struct xchk_meta_ops meta_scrub_ops[] = { .type = ST_FS, .setup = xchk_setup_fscounters, .scrub = xchk_fscounters, - .repair = xrep_notsupported, + .repair = xrep_fscounters, + }, + [XFS_SCRUB_TYPE_QUOTACHECK] = { /* quota counters */ + .type = ST_FS, + .setup = xchk_setup_quotacheck, + .scrub = xchk_quotacheck, + .repair = xrep_quotacheck, + }, + [XFS_SCRUB_TYPE_NLINKS] = { /* inode link counts */ + .type = ST_FS, + .setup = xchk_setup_nlinks, + .scrub = xchk_nlinks, + .repair = xrep_nlinks, + }, + [XFS_SCRUB_TYPE_HEALTHY] = { /* fs healthy; clean all reminders */ + .type = ST_FS, + .setup = xchk_setup_fs, + .scrub = xchk_health_record, + .repair = xrep_notsupported, }, }; diff --git a/fs/xfs/scrub/scrub.h b/fs/xfs/scrub/scrub.h index 7fc50654c4fe..9ad65b604fe1 100644 --- a/fs/xfs/scrub/scrub.h +++ b/fs/xfs/scrub/scrub.h @@ -99,6 +99,9 @@ struct xfs_scrub { /* xfile used by the scrubbers; freed at teardown. */ struct xfile *xfile; + /* buffer target for in-memory btrees; also freed at teardown. */ + struct xfs_buftarg *xmbtp; + /* Lock flags for @ip. */ uint ilock_flags; @@ -121,6 +124,9 @@ struct xfs_scrub { #define XCHK_HAVE_FREEZE_PROT (1U << 1) /* do we have freeze protection? */ #define XCHK_FSGATES_DRAIN (1U << 2) /* defer ops draining enabled */ #define XCHK_NEED_DRAIN (1U << 3) /* scrub needs to drain defer ops */ +#define XCHK_FSGATES_QUOTA (1U << 4) /* quota live update enabled */ +#define XCHK_FSGATES_DIRENTS (1U << 5) /* directory live update enabled */ +#define XCHK_FSGATES_RMAP (1U << 6) /* rmapbt live update enabled */ #define XREP_RESET_PERAG_RESV (1U << 30) /* must reset AG space reservation */ #define XREP_ALREADY_FIXED (1U << 31) /* checking our repair work */ @@ -130,7 +136,10 @@ struct xfs_scrub { * features are gated off via dynamic code patching, which is why the state * must be enabled during scrub setup and can only be torn down afterwards. */ -#define XCHK_FSGATES_ALL (XCHK_FSGATES_DRAIN) +#define XCHK_FSGATES_ALL (XCHK_FSGATES_DRAIN | \ + XCHK_FSGATES_QUOTA | \ + XCHK_FSGATES_DIRENTS | \ + XCHK_FSGATES_RMAP) /* Metadata scrubbers */ int xchk_tester(struct xfs_scrub *sc); @@ -167,14 +176,21 @@ xchk_rtsummary(struct xfs_scrub *sc) #endif #ifdef CONFIG_XFS_QUOTA int xchk_quota(struct xfs_scrub *sc); +int xchk_quotacheck(struct xfs_scrub *sc); #else static inline int xchk_quota(struct xfs_scrub *sc) { return -ENOENT; } +static inline int +xchk_quotacheck(struct xfs_scrub *sc) +{ + return -ENOENT; +} #endif int xchk_fscounters(struct xfs_scrub *sc); +int xchk_nlinks(struct xfs_scrub *sc); /* cross-referencing helpers */ void xchk_xref_is_used_space(struct xfs_scrub *sc, xfs_agblock_t agbno, diff --git a/fs/xfs/scrub/stats.c b/fs/xfs/scrub/stats.c index cd91db4a5548..42cafbed94ac 100644 --- a/fs/xfs/scrub/stats.c +++ b/fs/xfs/scrub/stats.c @@ -77,6 +77,8 @@ static const char *name_map[XFS_SCRUB_TYPE_NR] = { [XFS_SCRUB_TYPE_GQUOTA] = "grpquota", [XFS_SCRUB_TYPE_PQUOTA] = "prjquota", [XFS_SCRUB_TYPE_FSCOUNTERS] = "fscounters", + [XFS_SCRUB_TYPE_QUOTACHECK] = "quotacheck", + [XFS_SCRUB_TYPE_NLINKS] = "nlinks", }; /* Format the scrub stats into a text buffer, similar to pcp style. */ @@ -329,9 +331,9 @@ xchk_stats_register( if (!cs->cs_debugfs) return; - debugfs_create_file("stats", 0644, cs->cs_debugfs, cs, + debugfs_create_file("stats", 0444, cs->cs_debugfs, cs, &scrub_stats_fops); - debugfs_create_file("clear_stats", 0400, cs->cs_debugfs, cs, + debugfs_create_file("clear_stats", 0200, cs->cs_debugfs, cs, &clear_scrub_stats_fops); } diff --git a/fs/xfs/scrub/symlink.c b/fs/xfs/scrub/symlink.c index ddff86713df3..d77d8a9598f6 100644 --- a/fs/xfs/scrub/symlink.c +++ b/fs/xfs/scrub/symlink.c @@ -13,6 +13,7 @@ #include "xfs_inode.h" #include "xfs_symlink.h" #include "xfs_health.h" +#include "xfs_symlink_remote.h" #include "scrub/scrub.h" #include "scrub/common.h" #include "scrub/health.h" @@ -67,7 +68,7 @@ xchk_symlink( } /* Remote symlink; must read the contents. */ - error = xfs_readlink_bmap_ilocked(sc->ip, sc->buf); + error = xfs_symlink_remote_read(sc->ip, sc->buf); if (!xchk_fblock_process_error(sc, XFS_DATA_FORK, 0, &error)) return error; if (strnlen(sc->buf, XFS_SYMLINK_MAXLEN) < len) diff --git a/fs/xfs/scrub/trace.c b/fs/xfs/scrub/trace.c index d0e24ffaf754..3dd281d6d185 100644 --- a/fs/xfs/scrub/trace.c +++ b/fs/xfs/scrub/trace.c @@ -16,10 +16,16 @@ #include "xfs_rtbitmap.h" #include "xfs_quota.h" #include "xfs_quota_defs.h" +#include "xfs_da_format.h" +#include "xfs_dir2.h" +#include "xfs_rmap.h" #include "scrub/scrub.h" #include "scrub/xfile.h" #include "scrub/xfarray.h" #include "scrub/quota.h" +#include "scrub/iscan.h" +#include "scrub/nlinks.h" +#include "scrub/fscounters.h" /* Figure out which block the btree cursor was pointing to. */ static inline xfs_fsblock_t @@ -32,7 +38,7 @@ xchk_btree_cur_fsbno( xfs_buf_daddr(cur->bc_levels[level].bp)); if (level == cur->bc_nlevels - 1 && - (cur->bc_flags & XFS_BTREE_ROOT_IN_INODE)) + cur->bc_ops->type == XFS_BTREE_TYPE_INODE) return XFS_INO_TO_FSB(cur->bc_mp, cur->bc_ino.ip->i_ino); return NULLFSBLOCK; diff --git a/fs/xfs/scrub/trace.h b/fs/xfs/scrub/trace.h index 6bbb4e8639dc..5b294be52c55 100644 --- a/fs/xfs/scrub/trace.h +++ b/fs/xfs/scrub/trace.h @@ -15,11 +15,17 @@ #include <linux/tracepoint.h> #include "xfs_bit.h" +#include "xfs_quota_defs.h" +struct xfs_scrub; struct xfile; struct xfarray; struct xfarray_sortinfo; struct xchk_dqiter; +struct xchk_iscan; +struct xchk_nlink; +struct xchk_fscounters; +struct xfs_rmap_update_params; /* * ftrace's __print_symbolic requires that all enum values be wrapped in the @@ -27,14 +33,6 @@ struct xchk_dqiter; * ring buffer. Somehow this was only worth mentioning in the ftrace sample * code. */ -TRACE_DEFINE_ENUM(XFS_BTNUM_BNOi); -TRACE_DEFINE_ENUM(XFS_BTNUM_CNTi); -TRACE_DEFINE_ENUM(XFS_BTNUM_BMAPi); -TRACE_DEFINE_ENUM(XFS_BTNUM_INOi); -TRACE_DEFINE_ENUM(XFS_BTNUM_FINOi); -TRACE_DEFINE_ENUM(XFS_BTNUM_RMAPi); -TRACE_DEFINE_ENUM(XFS_BTNUM_REFCi); - TRACE_DEFINE_ENUM(XFS_REFC_DOMAIN_SHARED); TRACE_DEFINE_ENUM(XFS_REFC_DOMAIN_COW); @@ -63,6 +61,9 @@ TRACE_DEFINE_ENUM(XFS_SCRUB_TYPE_UQUOTA); TRACE_DEFINE_ENUM(XFS_SCRUB_TYPE_GQUOTA); TRACE_DEFINE_ENUM(XFS_SCRUB_TYPE_PQUOTA); TRACE_DEFINE_ENUM(XFS_SCRUB_TYPE_FSCOUNTERS); +TRACE_DEFINE_ENUM(XFS_SCRUB_TYPE_QUOTACHECK); +TRACE_DEFINE_ENUM(XFS_SCRUB_TYPE_NLINKS); +TRACE_DEFINE_ENUM(XFS_SCRUB_TYPE_HEALTHY); #define XFS_SCRUB_TYPE_STRINGS \ { XFS_SCRUB_TYPE_PROBE, "probe" }, \ @@ -89,7 +90,10 @@ TRACE_DEFINE_ENUM(XFS_SCRUB_TYPE_FSCOUNTERS); { XFS_SCRUB_TYPE_UQUOTA, "usrquota" }, \ { XFS_SCRUB_TYPE_GQUOTA, "grpquota" }, \ { XFS_SCRUB_TYPE_PQUOTA, "prjquota" }, \ - { XFS_SCRUB_TYPE_FSCOUNTERS, "fscounters" } + { XFS_SCRUB_TYPE_FSCOUNTERS, "fscounters" }, \ + { XFS_SCRUB_TYPE_QUOTACHECK, "quotacheck" }, \ + { XFS_SCRUB_TYPE_NLINKS, "nlinks" }, \ + { XFS_SCRUB_TYPE_HEALTHY, "healthy" } #define XFS_SCRUB_FLAG_STRINGS \ { XFS_SCRUB_IFLAG_REPAIR, "repair" }, \ @@ -107,9 +111,21 @@ TRACE_DEFINE_ENUM(XFS_SCRUB_TYPE_FSCOUNTERS); { XCHK_HAVE_FREEZE_PROT, "nofreeze" }, \ { XCHK_FSGATES_DRAIN, "fsgates_drain" }, \ { XCHK_NEED_DRAIN, "need_drain" }, \ + { XCHK_FSGATES_QUOTA, "fsgates_quota" }, \ + { XCHK_FSGATES_DIRENTS, "fsgates_dirents" }, \ + { XCHK_FSGATES_RMAP, "fsgates_rmap" }, \ { XREP_RESET_PERAG_RESV, "reset_perag_resv" }, \ { XREP_ALREADY_FIXED, "already_fixed" } +TRACE_DEFINE_ENUM(XFS_RMAP_MAP); +TRACE_DEFINE_ENUM(XFS_RMAP_MAP_SHARED); +TRACE_DEFINE_ENUM(XFS_RMAP_UNMAP); +TRACE_DEFINE_ENUM(XFS_RMAP_UNMAP_SHARED); +TRACE_DEFINE_ENUM(XFS_RMAP_CONVERT); +TRACE_DEFINE_ENUM(XFS_RMAP_CONVERT_SHARED); +TRACE_DEFINE_ENUM(XFS_RMAP_ALLOC); +TRACE_DEFINE_ENUM(XFS_RMAP_FREE); + DECLARE_EVENT_CLASS(xchk_class, TP_PROTO(struct xfs_inode *ip, struct xfs_scrub_metadata *sm, int error), @@ -395,6 +411,29 @@ DEFINE_SCRUB_DQITER_EVENT(xchk_dquot_iter_revalidate_bmap); DEFINE_SCRUB_DQITER_EVENT(xchk_dquot_iter_advance_bmap); DEFINE_SCRUB_DQITER_EVENT(xchk_dquot_iter_advance_incore); DEFINE_SCRUB_DQITER_EVENT(xchk_dquot_iter); + +TRACE_EVENT(xchk_qcheck_error, + TP_PROTO(struct xfs_scrub *sc, xfs_dqtype_t dqtype, xfs_dqid_t id, + void *ret_ip), + TP_ARGS(sc, dqtype, id, ret_ip), + TP_STRUCT__entry( + __field(dev_t, dev) + __field(xfs_dqtype_t, dqtype) + __field(xfs_dqid_t, id) + __field(void *, ret_ip) + ), + TP_fast_assign( + __entry->dev = sc->mp->m_super->s_dev; + __entry->dqtype = dqtype; + __entry->id = id; + __entry->ret_ip = ret_ip; + ), + TP_printk("dev %d:%d dquot type %s id 0x%x ret_ip %pS", + MAJOR(__entry->dev), MINOR(__entry->dev), + __print_symbolic(__entry->dqtype, XFS_DQTYPE_STRINGS), + __entry->id, + __entry->ret_ip) +); #endif /* CONFIG_XFS_QUOTA */ TRACE_EVENT(xchk_incomplete, @@ -423,7 +462,7 @@ TRACE_EVENT(xchk_btree_op_error, TP_STRUCT__entry( __field(dev_t, dev) __field(unsigned int, type) - __field(xfs_btnum_t, btnum) + __string(name, cur->bc_ops->name) __field(int, level) __field(xfs_agnumber_t, agno) __field(xfs_agblock_t, bno) @@ -436,7 +475,7 @@ TRACE_EVENT(xchk_btree_op_error, __entry->dev = sc->mp->m_super->s_dev; __entry->type = sc->sm->sm_type; - __entry->btnum = cur->bc_btnum; + __assign_str(name, cur->bc_ops->name); __entry->level = level; __entry->agno = XFS_FSB_TO_AGNO(cur->bc_mp, fsbno); __entry->bno = XFS_FSB_TO_AGBNO(cur->bc_mp, fsbno); @@ -444,10 +483,10 @@ TRACE_EVENT(xchk_btree_op_error, __entry->error = error; __entry->ret_ip = ret_ip; ), - TP_printk("dev %d:%d type %s btree %s level %d ptr %d agno 0x%x agbno 0x%x error %d ret_ip %pS", + TP_printk("dev %d:%d type %s %sbt level %d ptr %d agno 0x%x agbno 0x%x error %d ret_ip %pS", MAJOR(__entry->dev), MINOR(__entry->dev), __print_symbolic(__entry->type, XFS_SCRUB_TYPE_STRINGS), - __print_symbolic(__entry->btnum, XFS_BTNUM_STRINGS), + __get_str(name), __entry->level, __entry->ptr, __entry->agno, @@ -465,7 +504,7 @@ TRACE_EVENT(xchk_ifork_btree_op_error, __field(xfs_ino_t, ino) __field(int, whichfork) __field(unsigned int, type) - __field(xfs_btnum_t, btnum) + __string(name, cur->bc_ops->name) __field(int, level) __field(int, ptr) __field(xfs_agnumber_t, agno) @@ -479,7 +518,7 @@ TRACE_EVENT(xchk_ifork_btree_op_error, __entry->ino = sc->ip->i_ino; __entry->whichfork = cur->bc_ino.whichfork; __entry->type = sc->sm->sm_type; - __entry->btnum = cur->bc_btnum; + __assign_str(name, cur->bc_ops->name); __entry->level = level; __entry->ptr = cur->bc_levels[level].ptr; __entry->agno = XFS_FSB_TO_AGNO(cur->bc_mp, fsbno); @@ -487,12 +526,12 @@ TRACE_EVENT(xchk_ifork_btree_op_error, __entry->error = error; __entry->ret_ip = ret_ip; ), - TP_printk("dev %d:%d ino 0x%llx fork %s type %s btree %s level %d ptr %d agno 0x%x agbno 0x%x error %d ret_ip %pS", + TP_printk("dev %d:%d ino 0x%llx fork %s type %s %sbt level %d ptr %d agno 0x%x agbno 0x%x error %d ret_ip %pS", MAJOR(__entry->dev), MINOR(__entry->dev), __entry->ino, __print_symbolic(__entry->whichfork, XFS_WHICHFORK_STRINGS), __print_symbolic(__entry->type, XFS_SCRUB_TYPE_STRINGS), - __print_symbolic(__entry->btnum, XFS_BTNUM_STRINGS), + __get_str(name), __entry->level, __entry->ptr, __entry->agno, @@ -508,7 +547,7 @@ TRACE_EVENT(xchk_btree_error, TP_STRUCT__entry( __field(dev_t, dev) __field(unsigned int, type) - __field(xfs_btnum_t, btnum) + __string(name, cur->bc_ops->name) __field(int, level) __field(xfs_agnumber_t, agno) __field(xfs_agblock_t, bno) @@ -519,17 +558,17 @@ TRACE_EVENT(xchk_btree_error, xfs_fsblock_t fsbno = xchk_btree_cur_fsbno(cur, level); __entry->dev = sc->mp->m_super->s_dev; __entry->type = sc->sm->sm_type; - __entry->btnum = cur->bc_btnum; + __assign_str(name, cur->bc_ops->name); __entry->level = level; __entry->agno = XFS_FSB_TO_AGNO(cur->bc_mp, fsbno); __entry->bno = XFS_FSB_TO_AGBNO(cur->bc_mp, fsbno); __entry->ptr = cur->bc_levels[level].ptr; __entry->ret_ip = ret_ip; ), - TP_printk("dev %d:%d type %s btree %s level %d ptr %d agno 0x%x agbno 0x%x ret_ip %pS", + TP_printk("dev %d:%d type %s %sbt level %d ptr %d agno 0x%x agbno 0x%x ret_ip %pS", MAJOR(__entry->dev), MINOR(__entry->dev), __print_symbolic(__entry->type, XFS_SCRUB_TYPE_STRINGS), - __print_symbolic(__entry->btnum, XFS_BTNUM_STRINGS), + __get_str(name), __entry->level, __entry->ptr, __entry->agno, @@ -546,7 +585,7 @@ TRACE_EVENT(xchk_ifork_btree_error, __field(xfs_ino_t, ino) __field(int, whichfork) __field(unsigned int, type) - __field(xfs_btnum_t, btnum) + __string(name, cur->bc_ops->name) __field(int, level) __field(xfs_agnumber_t, agno) __field(xfs_agblock_t, bno) @@ -559,19 +598,19 @@ TRACE_EVENT(xchk_ifork_btree_error, __entry->ino = sc->ip->i_ino; __entry->whichfork = cur->bc_ino.whichfork; __entry->type = sc->sm->sm_type; - __entry->btnum = cur->bc_btnum; + __assign_str(name, cur->bc_ops->name); __entry->level = level; __entry->agno = XFS_FSB_TO_AGNO(cur->bc_mp, fsbno); __entry->bno = XFS_FSB_TO_AGBNO(cur->bc_mp, fsbno); __entry->ptr = cur->bc_levels[level].ptr; __entry->ret_ip = ret_ip; ), - TP_printk("dev %d:%d ino 0x%llx fork %s type %s btree %s level %d ptr %d agno 0x%x agbno 0x%x ret_ip %pS", + TP_printk("dev %d:%d ino 0x%llx fork %s type %s %sbt level %d ptr %d agno 0x%x agbno 0x%x ret_ip %pS", MAJOR(__entry->dev), MINOR(__entry->dev), __entry->ino, __print_symbolic(__entry->whichfork, XFS_WHICHFORK_STRINGS), __print_symbolic(__entry->type, XFS_SCRUB_TYPE_STRINGS), - __print_symbolic(__entry->btnum, XFS_BTNUM_STRINGS), + __get_str(name), __entry->level, __entry->ptr, __entry->agno, @@ -586,7 +625,7 @@ DECLARE_EVENT_CLASS(xchk_sbtree_class, TP_STRUCT__entry( __field(dev_t, dev) __field(int, type) - __field(xfs_btnum_t, btnum) + __string(name, cur->bc_ops->name) __field(xfs_agnumber_t, agno) __field(xfs_agblock_t, bno) __field(int, level) @@ -598,17 +637,17 @@ DECLARE_EVENT_CLASS(xchk_sbtree_class, __entry->dev = sc->mp->m_super->s_dev; __entry->type = sc->sm->sm_type; - __entry->btnum = cur->bc_btnum; + __assign_str(name, cur->bc_ops->name); __entry->agno = XFS_FSB_TO_AGNO(cur->bc_mp, fsbno); __entry->bno = XFS_FSB_TO_AGBNO(cur->bc_mp, fsbno); __entry->level = level; __entry->nlevels = cur->bc_nlevels; __entry->ptr = cur->bc_levels[level].ptr; ), - TP_printk("dev %d:%d type %s btree %s agno 0x%x agbno 0x%x level %d nlevels %d ptr %d", + TP_printk("dev %d:%d type %s %sbt agno 0x%x agbno 0x%x level %d nlevels %d ptr %d", MAJOR(__entry->dev), MINOR(__entry->dev), __print_symbolic(__entry->type, XFS_SCRUB_TYPE_STRINGS), - __print_symbolic(__entry->btnum, XFS_BTNUM_STRINGS), + __get_str(name), __entry->agno, __entry->bno, __entry->level, @@ -861,18 +900,11 @@ TRACE_EVENT(xfile_destroy, __field(loff_t, size) ), TP_fast_assign( - struct xfile_stat statbuf; - int ret; + struct inode *inode = file_inode(xf->file); - ret = xfile_stat(xf, &statbuf); - if (!ret) { - __entry->bytes = statbuf.bytes; - __entry->size = statbuf.size; - } else { - __entry->bytes = -1; - __entry->size = -1; - } - __entry->ino = file_inode(xf->file)->i_ino; + __entry->ino = inode->i_ino; + __entry->bytes = inode->i_blocks << SECTOR_SHIFT; + __entry->size = i_size_read(inode); ), TP_printk("xfino 0x%lx mem_bytes 0x%llx isize 0x%llx", __entry->ino, @@ -891,19 +923,12 @@ DECLARE_EVENT_CLASS(xfile_class, __field(unsigned long long, bytecount) ), TP_fast_assign( - struct xfile_stat statbuf; - int ret; + struct inode *inode = file_inode(xf->file); - ret = xfile_stat(xf, &statbuf); - if (!ret) { - __entry->bytes_used = statbuf.bytes; - __entry->size = statbuf.size; - } else { - __entry->bytes_used = -1; - __entry->size = -1; - } - __entry->ino = file_inode(xf->file)->i_ino; + __entry->ino = inode->i_ino; + __entry->bytes_used = inode->i_blocks << SECTOR_SHIFT; __entry->pos = pos; + __entry->size = i_size_read(inode); __entry->bytecount = bytecount; ), TP_printk("xfino 0x%lx mem_bytes 0x%llx pos 0x%llx bytecount 0x%llx isize 0x%llx", @@ -917,11 +942,11 @@ DECLARE_EVENT_CLASS(xfile_class, DEFINE_EVENT(xfile_class, name, \ TP_PROTO(struct xfile *xf, loff_t pos, unsigned long long bytecount), \ TP_ARGS(xf, pos, bytecount)) -DEFINE_XFILE_EVENT(xfile_pread); -DEFINE_XFILE_EVENT(xfile_pwrite); +DEFINE_XFILE_EVENT(xfile_load); +DEFINE_XFILE_EVENT(xfile_store); DEFINE_XFILE_EVENT(xfile_seek_data); -DEFINE_XFILE_EVENT(xfile_get_page); -DEFINE_XFILE_EVENT(xfile_put_page); +DEFINE_XFILE_EVENT(xfile_get_folio); +DEFINE_XFILE_EVENT(xfile_put_folio); TRACE_EVENT(xfarray_create, TP_PROTO(struct xfarray *xfa, unsigned long long required_capacity), @@ -968,7 +993,7 @@ TRACE_EVENT(xfarray_isort, __entry->hi - __entry->lo) ); -TRACE_EVENT(xfarray_pagesort, +TRACE_EVENT(xfarray_foliosort, TP_PROTO(struct xfarray_sortinfo *si, uint64_t lo, uint64_t hi), TP_ARGS(si, lo, hi), TP_STRUCT__entry( @@ -1039,6 +1064,47 @@ TRACE_EVENT(xfarray_sort, __entry->bytes) ); +TRACE_EVENT(xfarray_sort_scan, + TP_PROTO(struct xfarray_sortinfo *si, unsigned long long idx), + TP_ARGS(si, idx), + TP_STRUCT__entry( + __field(unsigned long, ino) + __field(unsigned long long, nr) + __field(size_t, obj_size) + __field(unsigned long long, idx) + __field(unsigned long long, folio_pos) + __field(unsigned long, folio_bytes) + __field(unsigned long long, first_idx) + __field(unsigned long long, last_idx) + ), + TP_fast_assign( + __entry->nr = si->array->nr; + __entry->obj_size = si->array->obj_size; + __entry->ino = file_inode(si->array->xfile->file)->i_ino; + __entry->idx = idx; + if (si->folio) { + __entry->folio_pos = folio_pos(si->folio); + __entry->folio_bytes = folio_size(si->folio); + __entry->first_idx = si->first_folio_idx; + __entry->last_idx = si->last_folio_idx; + } else { + __entry->folio_pos = 0; + __entry->folio_bytes = 0; + __entry->first_idx = 0; + __entry->last_idx = 0; + } + ), + TP_printk("xfino 0x%lx nr %llu objsz %zu idx %llu folio_pos 0x%llx folio_bytes 0x%lx first_idx %llu last_idx %llu", + __entry->ino, + __entry->nr, + __entry->obj_size, + __entry->idx, + __entry->folio_pos, + __entry->folio_bytes, + __entry->first_idx, + __entry->last_idx) +); + TRACE_EVENT(xfarray_sort_stats, TP_PROTO(struct xfarray_sortinfo *si, int error), TP_ARGS(si, error), @@ -1119,6 +1185,323 @@ TRACE_EVENT(xchk_rtsum_record_free, ); #endif /* CONFIG_XFS_RT */ +DECLARE_EVENT_CLASS(xchk_iscan_class, + TP_PROTO(struct xchk_iscan *iscan), + TP_ARGS(iscan), + TP_STRUCT__entry( + __field(dev_t, dev) + __field(xfs_ino_t, cursor) + __field(xfs_ino_t, visited) + ), + TP_fast_assign( + __entry->dev = iscan->sc->mp->m_super->s_dev; + __entry->cursor = iscan->cursor_ino; + __entry->visited = iscan->__visited_ino; + ), + TP_printk("dev %d:%d iscan cursor 0x%llx visited 0x%llx", + MAJOR(__entry->dev), MINOR(__entry->dev), + __entry->cursor, + __entry->visited) +) +#define DEFINE_ISCAN_EVENT(name) \ +DEFINE_EVENT(xchk_iscan_class, name, \ + TP_PROTO(struct xchk_iscan *iscan), \ + TP_ARGS(iscan)) +DEFINE_ISCAN_EVENT(xchk_iscan_move_cursor); +DEFINE_ISCAN_EVENT(xchk_iscan_visit); +DEFINE_ISCAN_EVENT(xchk_iscan_skip); +DEFINE_ISCAN_EVENT(xchk_iscan_advance_ag); + +DECLARE_EVENT_CLASS(xchk_iscan_ino_class, + TP_PROTO(struct xchk_iscan *iscan, xfs_ino_t ino), + TP_ARGS(iscan, ino), + TP_STRUCT__entry( + __field(dev_t, dev) + __field(xfs_ino_t, startino) + __field(xfs_ino_t, cursor) + __field(xfs_ino_t, visited) + __field(xfs_ino_t, ino) + ), + TP_fast_assign( + __entry->dev = iscan->sc->mp->m_super->s_dev; + __entry->startino = iscan->scan_start_ino; + __entry->cursor = iscan->cursor_ino; + __entry->visited = iscan->__visited_ino; + __entry->ino = ino; + ), + TP_printk("dev %d:%d iscan start 0x%llx cursor 0x%llx visited 0x%llx ino 0x%llx", + MAJOR(__entry->dev), MINOR(__entry->dev), + __entry->startino, + __entry->cursor, + __entry->visited, + __entry->ino) +) +#define DEFINE_ISCAN_INO_EVENT(name) \ +DEFINE_EVENT(xchk_iscan_ino_class, name, \ + TP_PROTO(struct xchk_iscan *iscan, xfs_ino_t ino), \ + TP_ARGS(iscan, ino)) +DEFINE_ISCAN_INO_EVENT(xchk_iscan_want_live_update); +DEFINE_ISCAN_INO_EVENT(xchk_iscan_start); + +TRACE_EVENT(xchk_iscan_iget, + TP_PROTO(struct xchk_iscan *iscan, int error), + TP_ARGS(iscan, error), + TP_STRUCT__entry( + __field(dev_t, dev) + __field(xfs_ino_t, cursor) + __field(xfs_ino_t, visited) + __field(int, error) + ), + TP_fast_assign( + __entry->dev = iscan->sc->mp->m_super->s_dev; + __entry->cursor = iscan->cursor_ino; + __entry->visited = iscan->__visited_ino; + __entry->error = error; + ), + TP_printk("dev %d:%d iscan cursor 0x%llx visited 0x%llx error %d", + MAJOR(__entry->dev), MINOR(__entry->dev), + __entry->cursor, + __entry->visited, + __entry->error) +); + +TRACE_EVENT(xchk_iscan_iget_batch, + TP_PROTO(struct xfs_mount *mp, struct xchk_iscan *iscan, + unsigned int nr, unsigned int avail), + TP_ARGS(mp, iscan, nr, avail), + TP_STRUCT__entry( + __field(dev_t, dev) + __field(xfs_ino_t, cursor) + __field(xfs_ino_t, visited) + __field(unsigned int, nr) + __field(unsigned int, avail) + __field(unsigned int, unavail) + __field(xfs_ino_t, batch_ino) + __field(unsigned long long, skipmask) + ), + TP_fast_assign( + __entry->dev = mp->m_super->s_dev; + __entry->cursor = iscan->cursor_ino; + __entry->visited = iscan->__visited_ino; + __entry->nr = nr; + __entry->avail = avail; + __entry->unavail = hweight64(iscan->__skipped_inomask); + __entry->batch_ino = iscan->__batch_ino; + __entry->skipmask = iscan->__skipped_inomask; + ), + TP_printk("dev %d:%d iscan cursor 0x%llx visited 0x%llx batchino 0x%llx skipmask 0x%llx nr %u avail %u unavail %u", + MAJOR(__entry->dev), MINOR(__entry->dev), + __entry->cursor, + __entry->visited, + __entry->batch_ino, + __entry->skipmask, + __entry->nr, + __entry->avail, + __entry->unavail) +); + +TRACE_EVENT(xchk_iscan_iget_retry_wait, + TP_PROTO(struct xchk_iscan *iscan), + TP_ARGS(iscan), + TP_STRUCT__entry( + __field(dev_t, dev) + __field(xfs_ino_t, cursor) + __field(xfs_ino_t, visited) + __field(unsigned int, retry_delay) + __field(unsigned long, remaining) + __field(unsigned int, iget_timeout) + ), + TP_fast_assign( + __entry->dev = iscan->sc->mp->m_super->s_dev; + __entry->cursor = iscan->cursor_ino; + __entry->visited = iscan->__visited_ino; + __entry->retry_delay = iscan->iget_retry_delay; + __entry->remaining = jiffies_to_msecs(iscan->__iget_deadline - jiffies); + __entry->iget_timeout = iscan->iget_timeout; + ), + TP_printk("dev %d:%d iscan cursor 0x%llx visited 0x%llx remaining %lu timeout %u delay %u", + MAJOR(__entry->dev), MINOR(__entry->dev), + __entry->cursor, + __entry->visited, + __entry->remaining, + __entry->iget_timeout, + __entry->retry_delay) +); + +TRACE_EVENT(xchk_nlinks_collect_dirent, + TP_PROTO(struct xfs_mount *mp, struct xfs_inode *dp, + xfs_ino_t ino, const struct xfs_name *name), + TP_ARGS(mp, dp, ino, name), + TP_STRUCT__entry( + __field(dev_t, dev) + __field(xfs_ino_t, dir) + __field(xfs_ino_t, ino) + __field(unsigned int, namelen) + __dynamic_array(char, name, name->len) + ), + TP_fast_assign( + __entry->dev = mp->m_super->s_dev; + __entry->dir = dp->i_ino; + __entry->ino = ino; + __entry->namelen = name->len; + memcpy(__get_str(name), name->name, name->len); + ), + TP_printk("dev %d:%d dir 0x%llx -> ino 0x%llx name '%.*s'", + MAJOR(__entry->dev), MINOR(__entry->dev), + __entry->dir, + __entry->ino, + __entry->namelen, + __get_str(name)) +); + +TRACE_EVENT(xchk_nlinks_collect_metafile, + TP_PROTO(struct xfs_mount *mp, xfs_ino_t ino), + TP_ARGS(mp, ino), + TP_STRUCT__entry( + __field(dev_t, dev) + __field(xfs_ino_t, ino) + ), + TP_fast_assign( + __entry->dev = mp->m_super->s_dev; + __entry->ino = ino; + ), + TP_printk("dev %d:%d ino 0x%llx", + MAJOR(__entry->dev), MINOR(__entry->dev), + __entry->ino) +); + +TRACE_EVENT(xchk_nlinks_live_update, + TP_PROTO(struct xfs_mount *mp, const struct xfs_inode *dp, + int action, xfs_ino_t ino, int delta, + const char *name, unsigned int namelen), + TP_ARGS(mp, dp, action, ino, delta, name, namelen), + TP_STRUCT__entry( + __field(dev_t, dev) + __field(xfs_ino_t, dir) + __field(int, action) + __field(xfs_ino_t, ino) + __field(int, delta) + __field(unsigned int, namelen) + __dynamic_array(char, name, namelen) + ), + TP_fast_assign( + __entry->dev = mp->m_super->s_dev; + __entry->dir = dp ? dp->i_ino : NULLFSINO; + __entry->action = action; + __entry->ino = ino; + __entry->delta = delta; + __entry->namelen = namelen; + memcpy(__get_str(name), name, namelen); + ), + TP_printk("dev %d:%d dir 0x%llx ino 0x%llx nlink_delta %d name '%.*s'", + MAJOR(__entry->dev), MINOR(__entry->dev), + __entry->dir, + __entry->ino, + __entry->delta, + __entry->namelen, + __get_str(name)) +); + +TRACE_EVENT(xchk_nlinks_check_zero, + TP_PROTO(struct xfs_mount *mp, xfs_ino_t ino, + const struct xchk_nlink *live), + TP_ARGS(mp, ino, live), + TP_STRUCT__entry( + __field(dev_t, dev) + __field(xfs_ino_t, ino) + __field(xfs_nlink_t, parents) + __field(xfs_nlink_t, backrefs) + __field(xfs_nlink_t, children) + ), + TP_fast_assign( + __entry->dev = mp->m_super->s_dev; + __entry->ino = ino; + __entry->parents = live->parents; + __entry->backrefs = live->backrefs; + __entry->children = live->children; + ), + TP_printk("dev %d:%d ino 0x%llx parents %u backrefs %u children %u", + MAJOR(__entry->dev), MINOR(__entry->dev), + __entry->ino, + __entry->parents, + __entry->backrefs, + __entry->children) +); + +TRACE_EVENT(xchk_nlinks_update_incore, + TP_PROTO(struct xfs_mount *mp, xfs_ino_t ino, + const struct xchk_nlink *live, int parents_delta, + int backrefs_delta, int children_delta), + TP_ARGS(mp, ino, live, parents_delta, backrefs_delta, children_delta), + TP_STRUCT__entry( + __field(dev_t, dev) + __field(xfs_ino_t, ino) + __field(xfs_nlink_t, parents) + __field(xfs_nlink_t, backrefs) + __field(xfs_nlink_t, children) + __field(int, parents_delta) + __field(int, backrefs_delta) + __field(int, children_delta) + ), + TP_fast_assign( + __entry->dev = mp->m_super->s_dev; + __entry->ino = ino; + __entry->parents = live->parents; + __entry->backrefs = live->backrefs; + __entry->children = live->children; + __entry->parents_delta = parents_delta; + __entry->backrefs_delta = backrefs_delta; + __entry->children_delta = children_delta; + ), + TP_printk("dev %d:%d ino 0x%llx parents %d:%u backrefs %d:%u children %d:%u", + MAJOR(__entry->dev), MINOR(__entry->dev), + __entry->ino, + __entry->parents_delta, + __entry->parents, + __entry->backrefs_delta, + __entry->backrefs, + __entry->children_delta, + __entry->children) +); + +DECLARE_EVENT_CLASS(xchk_nlinks_diff_class, + TP_PROTO(struct xfs_mount *mp, struct xfs_inode *ip, + const struct xchk_nlink *live), + TP_ARGS(mp, ip, live), + TP_STRUCT__entry( + __field(dev_t, dev) + __field(xfs_ino_t, ino) + __field(uint8_t, ftype) + __field(xfs_nlink_t, nlink) + __field(xfs_nlink_t, parents) + __field(xfs_nlink_t, backrefs) + __field(xfs_nlink_t, children) + ), + TP_fast_assign( + __entry->dev = mp->m_super->s_dev; + __entry->ino = ip->i_ino; + __entry->ftype = xfs_mode_to_ftype(VFS_I(ip)->i_mode); + __entry->nlink = VFS_I(ip)->i_nlink; + __entry->parents = live->parents; + __entry->backrefs = live->backrefs; + __entry->children = live->children; + ), + TP_printk("dev %d:%d ino 0x%llx ftype %s nlink %u parents %u backrefs %u children %u", + MAJOR(__entry->dev), MINOR(__entry->dev), + __entry->ino, + __print_symbolic(__entry->ftype, XFS_DIR3_FTYPE_STR), + __entry->nlink, + __entry->parents, + __entry->backrefs, + __entry->children) +); +#define DEFINE_SCRUB_NLINKS_DIFF_EVENT(name) \ +DEFINE_EVENT(xchk_nlinks_diff_class, name, \ + TP_PROTO(struct xfs_mount *mp, struct xfs_inode *ip, \ + const struct xchk_nlink *live), \ + TP_ARGS(mp, ip, live)) +DEFINE_SCRUB_NLINKS_DIFF_EVENT(xchk_nlinks_compare_inode); + /* repair tracepoints */ #if IS_ENABLED(CONFIG_XFS_ONLINE_REPAIR) @@ -1223,7 +1606,6 @@ DEFINE_EVENT(xrep_rmap_class, name, \ uint64_t owner, uint64_t offset, unsigned int flags), \ TP_ARGS(mp, agno, agbno, len, owner, offset, flags)) DEFINE_REPAIR_RMAP_EVENT(xrep_ibt_walk_rmap); -DEFINE_REPAIR_RMAP_EVENT(xrep_rmap_extent_fn); DEFINE_REPAIR_RMAP_EVENT(xrep_bmap_walk_rmap); TRACE_EVENT(xrep_abt_found, @@ -1341,6 +1723,38 @@ TRACE_EVENT(xrep_bmap_found, __entry->state) ); +TRACE_EVENT(xrep_rmap_found, + TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno, + const struct xfs_rmap_irec *rec), + TP_ARGS(mp, agno, rec), + TP_STRUCT__entry( + __field(dev_t, dev) + __field(xfs_agnumber_t, agno) + __field(xfs_agblock_t, agbno) + __field(xfs_extlen_t, len) + __field(uint64_t, owner) + __field(uint64_t, offset) + __field(unsigned int, flags) + ), + TP_fast_assign( + __entry->dev = mp->m_super->s_dev; + __entry->agno = agno; + __entry->agbno = rec->rm_startblock; + __entry->len = rec->rm_blockcount; + __entry->owner = rec->rm_owner; + __entry->offset = rec->rm_offset; + __entry->flags = rec->rm_flags; + ), + TP_printk("dev %d:%d agno 0x%x agbno 0x%x fsbcount 0x%x owner 0x%llx fileoff 0x%llx flags 0x%x", + MAJOR(__entry->dev), MINOR(__entry->dev), + __entry->agno, + __entry->agbno, + __entry->len, + __entry->owner, + __entry->offset, + __entry->flags) +); + TRACE_EVENT(xrep_findroot_block, TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno, xfs_agblock_t agbno, uint32_t magic, uint16_t level), @@ -1425,16 +1839,28 @@ TRACE_EVENT(xrep_calc_ag_resblks_btsize, __entry->refcbt_sz) ) TRACE_EVENT(xrep_reset_counters, - TP_PROTO(struct xfs_mount *mp), - TP_ARGS(mp), + TP_PROTO(struct xfs_mount *mp, struct xchk_fscounters *fsc), + TP_ARGS(mp, fsc), TP_STRUCT__entry( __field(dev_t, dev) + __field(uint64_t, icount) + __field(uint64_t, ifree) + __field(uint64_t, fdblocks) + __field(uint64_t, frextents) ), TP_fast_assign( __entry->dev = mp->m_super->s_dev; + __entry->icount = fsc->icount; + __entry->ifree = fsc->ifree; + __entry->fdblocks = fsc->fdblocks; + __entry->frextents = fsc->frextents; ), - TP_printk("dev %d:%d", - MAJOR(__entry->dev), MINOR(__entry->dev)) + TP_printk("dev %d:%d icount %llu ifree %llu fdblocks %llu frextents %llu", + MAJOR(__entry->dev), MINOR(__entry->dev), + __entry->icount, + __entry->ifree, + __entry->fdblocks, + __entry->frextents) ) DECLARE_EVENT_CLASS(xrep_newbt_extent_class, @@ -1645,6 +2071,55 @@ TRACE_EVENT(xrep_dinode_count_rmaps, __entry->attr_extents) ); +TRACE_EVENT(xrep_dinode_findmode_dirent, + TP_PROTO(struct xfs_scrub *sc, struct xfs_inode *dp, + unsigned int ftype), + TP_ARGS(sc, dp, ftype), + TP_STRUCT__entry( + __field(dev_t, dev) + __field(xfs_ino_t, ino) + __field(xfs_ino_t, parent_ino) + __field(unsigned int, ftype) + ), + TP_fast_assign( + __entry->dev = sc->mp->m_super->s_dev; + __entry->ino = sc->sm->sm_ino; + __entry->parent_ino = dp->i_ino; + __entry->ftype = ftype; + ), + TP_printk("dev %d:%d ino 0x%llx parent_ino 0x%llx ftype '%s'", + MAJOR(__entry->dev), MINOR(__entry->dev), + __entry->ino, + __entry->parent_ino, + __print_symbolic(__entry->ftype, XFS_DIR3_FTYPE_STR)) +); + +TRACE_EVENT(xrep_dinode_findmode_dirent_inval, + TP_PROTO(struct xfs_scrub *sc, struct xfs_inode *dp, + unsigned int ftype, unsigned int found_ftype), + TP_ARGS(sc, dp, ftype, found_ftype), + TP_STRUCT__entry( + __field(dev_t, dev) + __field(xfs_ino_t, ino) + __field(xfs_ino_t, parent_ino) + __field(unsigned int, ftype) + __field(unsigned int, found_ftype) + ), + TP_fast_assign( + __entry->dev = sc->mp->m_super->s_dev; + __entry->ino = sc->sm->sm_ino; + __entry->parent_ino = dp->i_ino; + __entry->ftype = ftype; + __entry->found_ftype = found_ftype; + ), + TP_printk("dev %d:%d ino 0x%llx parent_ino 0x%llx ftype '%s' found_ftype '%s'", + MAJOR(__entry->dev), MINOR(__entry->dev), + __entry->ino, + __entry->parent_ino, + __print_symbolic(__entry->ftype, XFS_DIR3_FTYPE_STR), + __print_symbolic(__entry->found_ftype, XFS_DIR3_FTYPE_STR)) +); + TRACE_EVENT(xrep_cow_mark_file_range, TP_PROTO(struct xfs_inode *ip, xfs_fsblock_t startblock, xfs_fileoff_t startoff, xfs_filblks_t blockcount), @@ -1756,8 +2231,48 @@ DEFINE_EVENT(xrep_dquot_class, name, \ DEFINE_XREP_DQUOT_EVENT(xrep_dquot_item); DEFINE_XREP_DQUOT_EVENT(xrep_disk_dquot); DEFINE_XREP_DQUOT_EVENT(xrep_dquot_item_fill_bmap_hole); +DEFINE_XREP_DQUOT_EVENT(xrep_quotacheck_dquot); #endif /* CONFIG_XFS_QUOTA */ +DEFINE_SCRUB_NLINKS_DIFF_EVENT(xrep_nlinks_update_inode); +DEFINE_SCRUB_NLINKS_DIFF_EVENT(xrep_nlinks_unfixable_inode); + +TRACE_EVENT(xrep_rmap_live_update, + TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno, unsigned int op, + const struct xfs_rmap_update_params *p), + TP_ARGS(mp, agno, op, p), + TP_STRUCT__entry( + __field(dev_t, dev) + __field(xfs_agnumber_t, agno) + __field(unsigned int, op) + __field(xfs_agblock_t, agbno) + __field(xfs_extlen_t, len) + __field(uint64_t, owner) + __field(uint64_t, offset) + __field(unsigned int, flags) + ), + TP_fast_assign( + __entry->dev = mp->m_super->s_dev; + __entry->agno = agno; + __entry->op = op; + __entry->agbno = p->startblock; + __entry->len = p->blockcount; + xfs_owner_info_unpack(&p->oinfo, &__entry->owner, + &__entry->offset, &__entry->flags); + if (p->unwritten) + __entry->flags |= XFS_RMAP_UNWRITTEN; + ), + TP_printk("dev %d:%d agno 0x%x op %d agbno 0x%x fsbcount 0x%x owner 0x%llx fileoff 0x%llx flags 0x%x", + MAJOR(__entry->dev), MINOR(__entry->dev), + __entry->agno, + __entry->op, + __entry->agbno, + __entry->len, + __entry->owner, + __entry->offset, + __entry->flags) +); + #endif /* IS_ENABLED(CONFIG_XFS_ONLINE_REPAIR) */ #endif /* _TRACE_XFS_SCRUB_TRACE_H */ diff --git a/fs/xfs/scrub/xfarray.c b/fs/xfs/scrub/xfarray.c index f0f532c10a5a..17c982a4821d 100644 --- a/fs/xfs/scrub/xfarray.c +++ b/fs/xfs/scrub/xfarray.c @@ -16,7 +16,7 @@ * Large Arrays of Fixed-Size Records * ================================== * - * This memory array uses an xfile (which itself is a memfd "file") to store + * This memory array uses an xfile (which itself is a shmem file) to store * large numbers of fixed-size records in memory that can be paged out. This * puts less stress on the memory reclaim algorithms during an online repair * because we don't have to pin so much memory. However, array access is less @@ -136,7 +136,7 @@ xfarray_load( if (idx >= array->nr) return -ENODATA; - return xfile_obj_load(array->xfile, ptr, array->obj_size, + return xfile_load(array->xfile, ptr, array->obj_size, xfarray_pos(array, idx)); } @@ -152,7 +152,7 @@ xfarray_is_unset( if (array->unset_slots == 0) return false; - error = xfile_obj_load(array->xfile, temp, array->obj_size, pos); + error = xfile_load(array->xfile, temp, array->obj_size, pos); if (!error && xfarray_element_is_null(array, temp)) return true; @@ -184,7 +184,7 @@ xfarray_unset( return 0; memset(temp, 0, array->obj_size); - error = xfile_obj_store(array->xfile, temp, array->obj_size, pos); + error = xfile_store(array->xfile, temp, array->obj_size, pos); if (error) return error; @@ -209,7 +209,7 @@ xfarray_store( ASSERT(!xfarray_element_is_null(array, ptr)); - ret = xfile_obj_store(array->xfile, ptr, array->obj_size, + ret = xfile_store(array->xfile, ptr, array->obj_size, xfarray_pos(array, idx)); if (ret) return ret; @@ -245,12 +245,12 @@ xfarray_store_anywhere( for (pos = 0; pos < endpos && array->unset_slots > 0; pos += array->obj_size) { - error = xfile_obj_load(array->xfile, temp, array->obj_size, + error = xfile_load(array->xfile, temp, array->obj_size, pos); if (error || !xfarray_element_is_null(array, temp)) continue; - error = xfile_obj_store(array->xfile, ptr, array->obj_size, + error = xfile_store(array->xfile, ptr, array->obj_size, pos); if (error) return error; @@ -552,7 +552,7 @@ xfarray_isort( trace_xfarray_isort(si, lo, hi); xfarray_sort_bump_loads(si); - error = xfile_obj_load(si->array->xfile, scratch, len, lo_pos); + error = xfile_load(si->array->xfile, scratch, len, lo_pos); if (error) return error; @@ -560,88 +560,45 @@ xfarray_isort( sort(scratch, hi - lo + 1, si->array->obj_size, si->cmp_fn, NULL); xfarray_sort_bump_stores(si); - return xfile_obj_store(si->array->xfile, scratch, len, lo_pos); + return xfile_store(si->array->xfile, scratch, len, lo_pos); } -/* Grab a page for sorting records. */ -static inline int -xfarray_sort_get_page( - struct xfarray_sortinfo *si, - loff_t pos, - uint64_t len) -{ - int error; - - error = xfile_get_page(si->array->xfile, pos, len, &si->xfpage); - if (error) - return error; - - /* - * xfile pages must never be mapped into userspace, so we skip the - * dcache flush when mapping the page. - */ - si->page_kaddr = kmap_local_page(si->xfpage.page); - return 0; -} - -/* Release a page we grabbed for sorting records. */ -static inline int -xfarray_sort_put_page( - struct xfarray_sortinfo *si) -{ - if (!si->page_kaddr) - return 0; - - kunmap_local(si->page_kaddr); - si->page_kaddr = NULL; - - return xfile_put_page(si->array->xfile, &si->xfpage); -} - -/* Decide if these records are eligible for in-page sorting. */ -static inline bool -xfarray_want_pagesort( - struct xfarray_sortinfo *si, - xfarray_idx_t lo, - xfarray_idx_t hi) -{ - pgoff_t lo_page; - pgoff_t hi_page; - loff_t end_pos; - - /* We can only map one page at a time. */ - lo_page = xfarray_pos(si->array, lo) >> PAGE_SHIFT; - end_pos = xfarray_pos(si->array, hi) + si->array->obj_size - 1; - hi_page = end_pos >> PAGE_SHIFT; - - return lo_page == hi_page; -} - -/* Sort a bunch of records that all live in the same memory page. */ +/* + * Sort the records from lo to hi (inclusive) if they are all backed by the + * same memory folio. Returns 1 if it sorted, 0 if it did not, or a negative + * errno. + */ STATIC int -xfarray_pagesort( +xfarray_foliosort( struct xfarray_sortinfo *si, xfarray_idx_t lo, xfarray_idx_t hi) { + struct folio *folio; void *startp; loff_t lo_pos = xfarray_pos(si->array, lo); - uint64_t len = xfarray_pos(si->array, hi - lo); - int error = 0; + uint64_t len = xfarray_pos(si->array, hi - lo + 1); - trace_xfarray_pagesort(si, lo, hi); + /* No single folio could back this many records. */ + if (len > XFILE_MAX_FOLIO_SIZE) + return 0; xfarray_sort_bump_loads(si); - error = xfarray_sort_get_page(si, lo_pos, len); - if (error) - return error; + folio = xfile_get_folio(si->array->xfile, lo_pos, len, XFILE_ALLOC); + if (IS_ERR(folio)) + return PTR_ERR(folio); + if (!folio) + return 0; + + trace_xfarray_foliosort(si, lo, hi); xfarray_sort_bump_heapsorts(si); - startp = si->page_kaddr + offset_in_page(lo_pos); + startp = folio_address(folio) + offset_in_folio(folio, lo_pos); sort(startp, hi - lo + 1, si->array->obj_size, si->cmp_fn, NULL); xfarray_sort_bump_stores(si); - return xfarray_sort_put_page(si); + xfile_put_folio(si->array->xfile, folio); + return 1; } /* Return a pointer to the xfarray pivot record within the sortinfo struct. */ @@ -829,63 +786,78 @@ xfarray_qsort_push( return 0; } +static inline void +xfarray_sort_scan_done( + struct xfarray_sortinfo *si) +{ + if (si->folio) + xfile_put_folio(si->array->xfile, si->folio); + si->folio = NULL; +} + /* - * Load an element from the array into the first scratchpad and cache the page, - * if possible. + * Cache the folio backing the start of the given array element. If the array + * element is contained entirely within the folio, return a pointer to the + * cached folio. Otherwise, load the element into the scratchpad and return a + * pointer to the scratchpad. */ static inline int -xfarray_sort_load_cached( +xfarray_sort_scan( struct xfarray_sortinfo *si, xfarray_idx_t idx, - void *ptr) + void **ptrp) { loff_t idx_pos = xfarray_pos(si->array, idx); - pgoff_t startpage; - pgoff_t endpage; int error = 0; - /* - * If this load would split a page, release the cached page, if any, - * and perform a traditional read. - */ - startpage = idx_pos >> PAGE_SHIFT; - endpage = (idx_pos + si->array->obj_size - 1) >> PAGE_SHIFT; - if (startpage != endpage) { - error = xfarray_sort_put_page(si); - if (error) - return error; + if (xfarray_sort_terminated(si, &error)) + return error; - if (xfarray_sort_terminated(si, &error)) - return error; + trace_xfarray_sort_scan(si, idx); - return xfile_obj_load(si->array->xfile, ptr, - si->array->obj_size, idx_pos); - } + /* If the cached folio doesn't cover this index, release it. */ + if (si->folio && + (idx < si->first_folio_idx || idx > si->last_folio_idx)) + xfarray_sort_scan_done(si); - /* If the cached page is not the one we want, release it. */ - if (xfile_page_cached(&si->xfpage) && - xfile_page_index(&si->xfpage) != startpage) { - error = xfarray_sort_put_page(si); - if (error) - return error; + /* Grab the first folio that backs this array element. */ + if (!si->folio) { + loff_t next_pos; + + si->folio = xfile_get_folio(si->array->xfile, idx_pos, + si->array->obj_size, XFILE_ALLOC); + if (IS_ERR(si->folio)) + return PTR_ERR(si->folio); + + si->first_folio_idx = xfarray_idx(si->array, + folio_pos(si->folio) + si->array->obj_size - 1); + + next_pos = folio_pos(si->folio) + folio_size(si->folio); + si->last_folio_idx = xfarray_idx(si->array, next_pos - 1); + if (xfarray_pos(si->array, si->last_folio_idx + 1) > next_pos) + si->last_folio_idx--; + + trace_xfarray_sort_scan(si, idx); } /* - * If we don't have a cached page (and we know the load is contained - * in a single page) then grab it. + * If this folio still doesn't cover the desired element, it must cross + * a folio boundary. Read into the scratchpad and we're done. */ - if (!xfile_page_cached(&si->xfpage)) { - if (xfarray_sort_terminated(si, &error)) - return error; + if (idx < si->first_folio_idx || idx > si->last_folio_idx) { + void *temp = xfarray_scratch(si->array); - error = xfarray_sort_get_page(si, startpage << PAGE_SHIFT, - PAGE_SIZE); + error = xfile_load(si->array->xfile, temp, si->array->obj_size, + idx_pos); if (error) return error; + + *ptrp = temp; + return 0; } - memcpy(ptr, si->page_kaddr + offset_in_page(idx_pos), - si->array->obj_size); + /* Otherwise return a pointer to the array element in the folio. */ + *ptrp = folio_address(si->folio) + offset_in_folio(si->folio, idx_pos); return 0; } @@ -952,6 +924,8 @@ xfarray_sort( pivot = xfarray_sortinfo_pivot(si); while (si->stack_depth >= 0) { + int ret; + lo = si_lo[si->stack_depth]; hi = si_hi[si->stack_depth]; @@ -964,13 +938,13 @@ xfarray_sort( } /* - * If directly mapping the page and sorting can solve our + * If directly mapping the folio and sorting can solve our * problems, we're done. */ - if (xfarray_want_pagesort(si, lo, hi)) { - error = xfarray_pagesort(si, lo, hi); - if (error) - goto out_free; + ret = xfarray_foliosort(si, lo, hi); + if (ret < 0) + goto out_free; + if (ret == 1) { si->stack_depth--; continue; } @@ -995,25 +969,24 @@ xfarray_sort( * than the pivot is on the right side of the range. */ while (lo < hi) { + void *p; + /* * Decrement hi until it finds an a[hi] less than the * pivot value. */ - error = xfarray_sort_load_cached(si, hi, scratch); + error = xfarray_sort_scan(si, hi, &p); if (error) goto out_free; - while (xfarray_sort_cmp(si, scratch, pivot) >= 0 && - lo < hi) { + while (xfarray_sort_cmp(si, p, pivot) >= 0 && lo < hi) { hi--; - error = xfarray_sort_load_cached(si, hi, - scratch); + error = xfarray_sort_scan(si, hi, &p); if (error) goto out_free; } - error = xfarray_sort_put_page(si); - if (error) - goto out_free; - + if (p != scratch) + memcpy(scratch, p, si->array->obj_size); + xfarray_sort_scan_done(si); if (xfarray_sort_terminated(si, &error)) goto out_free; @@ -1028,21 +1001,18 @@ xfarray_sort( * Increment lo until it finds an a[lo] greater than * the pivot value. */ - error = xfarray_sort_load_cached(si, lo, scratch); + error = xfarray_sort_scan(si, lo, &p); if (error) goto out_free; - while (xfarray_sort_cmp(si, scratch, pivot) <= 0 && - lo < hi) { + while (xfarray_sort_cmp(si, p, pivot) <= 0 && lo < hi) { lo++; - error = xfarray_sort_load_cached(si, lo, - scratch); + error = xfarray_sort_scan(si, lo, &p); if (error) goto out_free; } - error = xfarray_sort_put_page(si); - if (error) - goto out_free; - + if (p != scratch) + memcpy(scratch, p, si->array->obj_size); + xfarray_sort_scan_done(si); if (xfarray_sort_terminated(si, &error)) goto out_free; diff --git a/fs/xfs/scrub/xfarray.h b/fs/xfs/scrub/xfarray.h index 62b9c506fdd1..acb2f94c56c1 100644 --- a/fs/xfs/scrub/xfarray.h +++ b/fs/xfs/scrub/xfarray.h @@ -45,6 +45,25 @@ int xfarray_store(struct xfarray *array, xfarray_idx_t idx, const void *ptr); int xfarray_store_anywhere(struct xfarray *array, const void *ptr); bool xfarray_element_is_null(struct xfarray *array, const void *ptr); +/* + * Load an array element, but zero the buffer if there's no data because we + * haven't stored to that array element yet. + */ +static inline int +xfarray_load_sparse( + struct xfarray *array, + uint64_t idx, + void *rec) +{ + int error = xfarray_load(array, idx, rec); + + if (error == -ENODATA) { + memset(rec, 0, array->obj_size); + return 0; + } + return error; +} + /* Append an element to the array. */ static inline int xfarray_append(struct xfarray *array, const void *ptr) { @@ -105,9 +124,14 @@ struct xfarray_sortinfo { /* XFARRAY_SORT_* flags; see below. */ unsigned int flags; - /* Cache a page here for faster access. */ - struct xfile_page xfpage; - void *page_kaddr; + /* Cache a folio here for faster scanning for pivots */ + struct folio *folio; + + /* First array index in folio that is completely readable */ + xfarray_idx_t first_folio_idx; + + /* Last array index in folio that is completely readable */ + xfarray_idx_t last_folio_idx; #ifdef DEBUG /* Performance statistics. */ diff --git a/fs/xfs/scrub/xfile.c b/fs/xfs/scrub/xfile.c index 090c3ead43fd..8cdd863db585 100644 --- a/fs/xfs/scrub/xfile.c +++ b/fs/xfs/scrub/xfile.c @@ -34,13 +34,6 @@ * xfiles assume that the caller will handle all required concurrency * management; standard vfs locks (freezer and inode) are not taken. Reads * and writes are satisfied directly from the page cache. - * - * NOTE: The current shmemfs implementation has a quirk that in-kernel reads - * of a hole cause a page to be mapped into the file. If you are going to - * create a sparse xfile, please be careful about reading from uninitialized - * parts of the file. These pages are !Uptodate and will eventually be - * reclaimed if not written, but in the short term this boosts memory - * consumption. */ /* @@ -62,38 +55,27 @@ xfile_create( { struct inode *inode; struct xfile *xf; - int error = -ENOMEM; + int error; xf = kmalloc(sizeof(struct xfile), XCHK_GFP_FLAGS); if (!xf) return -ENOMEM; - xf->file = shmem_file_setup(description, isize, 0); - if (!xf->file) - goto out_xfile; + xf->file = shmem_kernel_file_setup(description, isize, VM_NORESERVE); if (IS_ERR(xf->file)) { error = PTR_ERR(xf->file); goto out_xfile; } - /* - * We want a large sparse file that we can pread, pwrite, and seek. - * xfile users are responsible for keeping the xfile hidden away from - * all other callers, so we skip timestamp updates and security checks. - * Make the inode only accessible by root, just in case the xfile ever - * escapes. - */ - xf->file->f_mode |= FMODE_PREAD | FMODE_PWRITE | FMODE_NOCMTIME | - FMODE_LSEEK; - xf->file->f_flags |= O_RDWR | O_LARGEFILE | O_NOATIME; inode = file_inode(xf->file); - inode->i_flags |= S_PRIVATE | S_NOCMTIME | S_NOATIME; - inode->i_mode &= ~0177; - inode->i_uid = GLOBAL_ROOT_UID; - inode->i_gid = GLOBAL_ROOT_GID; - lockdep_set_class(&inode->i_rwsem, &xfile_i_mutex_key); + /* + * We don't want to bother with kmapping data during repair, so don't + * allow highmem pages to back this mapping. + */ + mapping_set_gfp_mask(inode->i_mapping, GFP_KERNEL); + trace_xfile_create(xf); *xfilep = xf; @@ -118,164 +100,128 @@ xfile_destroy( } /* - * Read a memory object directly from the xfile's page cache. Unlike regular - * pread, we return -E2BIG and -EFBIG for reads that are too large or at too - * high an offset, instead of truncating the read. Otherwise, we return - * bytes read or an error code, like regular pread. + * Load an object. Since we're treating this file as "memory", any error or + * short IO is treated as a failure to allocate memory. */ -ssize_t -xfile_pread( +int +xfile_load( struct xfile *xf, void *buf, size_t count, loff_t pos) { struct inode *inode = file_inode(xf->file); - struct address_space *mapping = inode->i_mapping; - struct page *page = NULL; - ssize_t read = 0; unsigned int pflags; - int error = 0; if (count > MAX_RW_COUNT) - return -E2BIG; + return -ENOMEM; if (inode->i_sb->s_maxbytes - pos < count) - return -EFBIG; + return -ENOMEM; - trace_xfile_pread(xf, pos, count); + trace_xfile_load(xf, pos, count); pflags = memalloc_nofs_save(); while (count > 0) { - void *p, *kaddr; + struct folio *folio; unsigned int len; + unsigned int offset; - len = min_t(ssize_t, count, PAGE_SIZE - offset_in_page(pos)); - - /* - * In-kernel reads of a shmem file cause it to allocate a page - * if the mapping shows a hole. Therefore, if we hit ENOMEM - * we can continue by zeroing the caller's buffer. - */ - page = shmem_read_mapping_page_gfp(mapping, pos >> PAGE_SHIFT, - __GFP_NOWARN); - if (IS_ERR(page)) { - error = PTR_ERR(page); - if (error != -ENOMEM) - break; - - memset(buf, 0, len); - goto advance; - } - - if (PageUptodate(page)) { + if (shmem_get_folio(inode, pos >> PAGE_SHIFT, &folio, + SGP_READ) < 0) + break; + if (!folio) { /* - * xfile pages must never be mapped into userspace, so - * we skip the dcache flush. + * No data stored at this offset, just zero the output + * buffer until the next page boundary. */ - kaddr = kmap_local_page(page); - p = kaddr + offset_in_page(pos); - memcpy(buf, p, len); - kunmap_local(kaddr); - } else { + len = min_t(ssize_t, count, + PAGE_SIZE - offset_in_page(pos)); memset(buf, 0, len); - } - put_page(page); + } else { + if (filemap_check_wb_err(inode->i_mapping, 0)) { + folio_unlock(folio); + folio_put(folio); + break; + } + + offset = offset_in_folio(folio, pos); + len = min_t(ssize_t, count, folio_size(folio) - offset); + memcpy(buf, folio_address(folio) + offset, len); -advance: + folio_unlock(folio); + folio_put(folio); + } count -= len; pos += len; buf += len; - read += len; } memalloc_nofs_restore(pflags); - if (read > 0) - return read; - return error; + if (count) + return -ENOMEM; + return 0; } /* - * Write a memory object directly to the xfile's page cache. Unlike regular - * pwrite, we return -E2BIG and -EFBIG for writes that are too large or at too - * high an offset, instead of truncating the write. Otherwise, we return - * bytes written or an error code, like regular pwrite. + * Store an object. Since we're treating this file as "memory", any error or + * short IO is treated as a failure to allocate memory. */ -ssize_t -xfile_pwrite( +int +xfile_store( struct xfile *xf, const void *buf, size_t count, loff_t pos) { struct inode *inode = file_inode(xf->file); - struct address_space *mapping = inode->i_mapping; - const struct address_space_operations *aops = mapping->a_ops; - struct page *page = NULL; - ssize_t written = 0; unsigned int pflags; - int error = 0; if (count > MAX_RW_COUNT) - return -E2BIG; + return -ENOMEM; if (inode->i_sb->s_maxbytes - pos < count) - return -EFBIG; + return -ENOMEM; - trace_xfile_pwrite(xf, pos, count); + trace_xfile_store(xf, pos, count); + + /* + * Increase the file size first so that shmem_get_folio(..., SGP_CACHE), + * actually allocates a folio instead of erroring out. + */ + if (pos + count > i_size_read(inode)) + i_size_write(inode, pos + count); pflags = memalloc_nofs_save(); while (count > 0) { - void *fsdata = NULL; - void *p, *kaddr; + struct folio *folio; unsigned int len; - int ret; - - len = min_t(ssize_t, count, PAGE_SIZE - offset_in_page(pos)); - - /* - * We call write_begin directly here to avoid all the freezer - * protection lock-taking that happens in the normal path. - * shmem doesn't support fs freeze, but lockdep doesn't know - * that and will trip over that. - */ - error = aops->write_begin(NULL, mapping, pos, len, &page, - &fsdata); - if (error) - break; + unsigned int offset; - /* - * xfile pages must never be mapped into userspace, so we skip - * the dcache flush. If the page is not uptodate, zero it - * before writing data. - */ - kaddr = kmap_local_page(page); - if (!PageUptodate(page)) { - memset(kaddr, 0, PAGE_SIZE); - SetPageUptodate(page); - } - p = kaddr + offset_in_page(pos); - memcpy(p, buf, len); - kunmap_local(kaddr); - - ret = aops->write_end(NULL, mapping, pos, len, len, page, - fsdata); - if (ret < 0) { - error = ret; + if (shmem_get_folio(inode, pos >> PAGE_SHIFT, &folio, + SGP_CACHE) < 0) + break; + if (filemap_check_wb_err(inode->i_mapping, 0)) { + folio_unlock(folio); + folio_put(folio); break; } - written += ret; - if (ret != len) - break; + offset = offset_in_folio(folio, pos); + len = min_t(ssize_t, count, folio_size(folio) - offset); + memcpy(folio_address(folio) + offset, buf, len); + + folio_mark_dirty(folio); + folio_unlock(folio); + folio_put(folio); - count -= ret; - pos += ret; - buf += ret; + count -= len; + pos += len; + buf += len; } memalloc_nofs_restore(pflags); - if (written > 0) - return written; - return error; + if (count) + return -ENOMEM; + return 0; } /* Find the next written area in the xfile data for a given offset. */ @@ -291,129 +237,76 @@ xfile_seek_data( return ret; } -/* Query stat information for an xfile. */ -int -xfile_stat( - struct xfile *xf, - struct xfile_stat *statbuf) -{ - struct kstat ks; - int error; - - error = vfs_getattr_nosec(&xf->file->f_path, &ks, - STATX_SIZE | STATX_BLOCKS, AT_STATX_DONT_SYNC); - if (error) - return error; - - statbuf->size = ks.size; - statbuf->bytes = ks.blocks << SECTOR_SHIFT; - return 0; -} - /* - * Grab the (locked) page for a memory object. The object cannot span a page - * boundary. Returns 0 (and a locked page) if successful, -ENOTBLK if we - * cannot grab the page, or the usual negative errno. + * Grab the (locked) folio for a memory object. The object cannot span a folio + * boundary. Returns the locked folio if successful, NULL if there was no + * folio or it didn't cover the range requested, or an ERR_PTR on failure. */ -int -xfile_get_page( +struct folio * +xfile_get_folio( struct xfile *xf, loff_t pos, - unsigned int len, - struct xfile_page *xfpage) + size_t len, + unsigned int flags) { struct inode *inode = file_inode(xf->file); - struct address_space *mapping = inode->i_mapping; - const struct address_space_operations *aops = mapping->a_ops; - struct page *page = NULL; - void *fsdata = NULL; - loff_t key = round_down(pos, PAGE_SIZE); + struct folio *folio = NULL; unsigned int pflags; int error; if (inode->i_sb->s_maxbytes - pos < len) - return -ENOMEM; - if (len > PAGE_SIZE - offset_in_page(pos)) - return -ENOTBLK; - - trace_xfile_get_page(xf, pos, len); + return ERR_PTR(-ENOMEM); - pflags = memalloc_nofs_save(); + trace_xfile_get_folio(xf, pos, len); /* - * We call write_begin directly here to avoid all the freezer - * protection lock-taking that happens in the normal path. shmem - * doesn't support fs freeze, but lockdep doesn't know that and will - * trip over that. + * Increase the file size first so that shmem_get_folio(..., SGP_CACHE), + * actually allocates a folio instead of erroring out. */ - error = aops->write_begin(NULL, mapping, key, PAGE_SIZE, &page, - &fsdata); + if ((flags & XFILE_ALLOC) && pos + len > i_size_read(inode)) + i_size_write(inode, pos + len); + + pflags = memalloc_nofs_save(); + error = shmem_get_folio(inode, pos >> PAGE_SHIFT, &folio, + (flags & XFILE_ALLOC) ? SGP_CACHE : SGP_READ); + memalloc_nofs_restore(pflags); if (error) - goto out_pflags; + return ERR_PTR(error); - /* We got the page, so make sure we push out EOF. */ - if (i_size_read(inode) < pos + len) - i_size_write(inode, pos + len); + if (!folio) + return NULL; - /* - * If the page isn't up to date, fill it with zeroes before we hand it - * to the caller and make sure the backing store will hold on to them. - */ - if (!PageUptodate(page)) { - void *kaddr; + if (len > folio_size(folio) - offset_in_folio(folio, pos)) { + folio_unlock(folio); + folio_put(folio); + return NULL; + } - kaddr = kmap_local_page(page); - memset(kaddr, 0, PAGE_SIZE); - kunmap_local(kaddr); - SetPageUptodate(page); + if (filemap_check_wb_err(inode->i_mapping, 0)) { + folio_unlock(folio); + folio_put(folio); + return ERR_PTR(-EIO); } /* - * Mark each page dirty so that the contents are written to some - * backing store when we drop this buffer, and take an extra reference - * to prevent the xfile page from being swapped or removed from the - * page cache by reclaim if the caller unlocks the page. + * Mark the folio dirty so that it won't be reclaimed once we drop the + * (potentially last) reference in xfile_put_folio. */ - set_page_dirty(page); - get_page(page); - - xfpage->page = page; - xfpage->fsdata = fsdata; - xfpage->pos = key; -out_pflags: - memalloc_nofs_restore(pflags); - return error; + if (flags & XFILE_ALLOC) + folio_set_dirty(folio); + return folio; } /* - * Release the (locked) page for a memory object. Returns 0 or a negative - * errno. + * Release the (locked) folio for a memory object. */ -int -xfile_put_page( +void +xfile_put_folio( struct xfile *xf, - struct xfile_page *xfpage) + struct folio *folio) { - struct inode *inode = file_inode(xf->file); - struct address_space *mapping = inode->i_mapping; - const struct address_space_operations *aops = mapping->a_ops; - unsigned int pflags; - int ret; - - trace_xfile_put_page(xf, xfpage->pos, PAGE_SIZE); - - /* Give back the reference that we took in xfile_get_page. */ - put_page(xfpage->page); + trace_xfile_put_folio(xf, folio_pos(folio), folio_size(folio)); - pflags = memalloc_nofs_save(); - ret = aops->write_end(NULL, mapping, xfpage->pos, PAGE_SIZE, PAGE_SIZE, - xfpage->page, xfpage->fsdata); - memalloc_nofs_restore(pflags); - memset(xfpage, 0, sizeof(struct xfile_page)); - - if (ret < 0) - return ret; - if (ret != PAGE_SIZE) - return -EIO; - return 0; + folio_unlock(folio); + folio_put(folio); } diff --git a/fs/xfs/scrub/xfile.h b/fs/xfs/scrub/xfile.h index d56643b0f429..76d78dba7e34 100644 --- a/fs/xfs/scrub/xfile.h +++ b/fs/xfs/scrub/xfile.h @@ -6,22 +6,6 @@ #ifndef __XFS_SCRUB_XFILE_H__ #define __XFS_SCRUB_XFILE_H__ -struct xfile_page { - struct page *page; - void *fsdata; - loff_t pos; -}; - -static inline bool xfile_page_cached(const struct xfile_page *xfpage) -{ - return xfpage->page != NULL; -} - -static inline pgoff_t xfile_page_index(const struct xfile_page *xfpage) -{ - return xfpage->page->index; -} - struct xfile { struct file *file; }; @@ -29,49 +13,17 @@ struct xfile { int xfile_create(const char *description, loff_t isize, struct xfile **xfilep); void xfile_destroy(struct xfile *xf); -ssize_t xfile_pread(struct xfile *xf, void *buf, size_t count, loff_t pos); -ssize_t xfile_pwrite(struct xfile *xf, const void *buf, size_t count, +int xfile_load(struct xfile *xf, void *buf, size_t count, loff_t pos); +int xfile_store(struct xfile *xf, const void *buf, size_t count, loff_t pos); -/* - * Load an object. Since we're treating this file as "memory", any error or - * short IO is treated as a failure to allocate memory. - */ -static inline int -xfile_obj_load(struct xfile *xf, void *buf, size_t count, loff_t pos) -{ - ssize_t ret = xfile_pread(xf, buf, count, pos); - - if (ret < 0 || ret != count) - return -ENOMEM; - return 0; -} - -/* - * Store an object. Since we're treating this file as "memory", any error or - * short IO is treated as a failure to allocate memory. - */ -static inline int -xfile_obj_store(struct xfile *xf, const void *buf, size_t count, loff_t pos) -{ - ssize_t ret = xfile_pwrite(xf, buf, count, pos); - - if (ret < 0 || ret != count) - return -ENOMEM; - return 0; -} - loff_t xfile_seek_data(struct xfile *xf, loff_t pos); -struct xfile_stat { - loff_t size; - unsigned long long bytes; -}; - -int xfile_stat(struct xfile *xf, struct xfile_stat *statbuf); +#define XFILE_MAX_FOLIO_SIZE (PAGE_SIZE << MAX_PAGECACHE_ORDER) -int xfile_get_page(struct xfile *xf, loff_t offset, unsigned int len, - struct xfile_page *xbuf); -int xfile_put_page(struct xfile *xf, struct xfile_page *xbuf); +#define XFILE_ALLOC (1 << 0) /* allocate folio if not present */ +struct folio *xfile_get_folio(struct xfile *xf, loff_t offset, size_t len, + unsigned int flags); +void xfile_put_folio(struct xfile *xf, struct folio *folio); #endif /* __XFS_SCRUB_XFILE_H__ */ diff --git a/fs/xfs/xfs_acl.c b/fs/xfs/xfs_acl.c index 6b840301817a..4bf69c9c088e 100644 --- a/fs/xfs/xfs_acl.c +++ b/fs/xfs/xfs_acl.c @@ -167,7 +167,7 @@ xfs_get_acl(struct inode *inode, int type, bool rcu) acl = ERR_PTR(error); } - kmem_free(args.value); + kvfree(args.value); return acl; } @@ -204,7 +204,7 @@ __xfs_set_acl(struct inode *inode, struct posix_acl *acl, int type) } error = xfs_attr_change(&args); - kmem_free(args.value); + kvfree(args.value); /* * If the attribute didn't exist to start with that's fine. diff --git a/fs/xfs/xfs_attr_inactive.c b/fs/xfs/xfs_attr_inactive.c index 89c7a9f4f930..24fb12986a56 100644 --- a/fs/xfs/xfs_attr_inactive.c +++ b/fs/xfs/xfs_attr_inactive.c @@ -23,6 +23,7 @@ #include "xfs_quota.h" #include "xfs_dir2.h" #include "xfs_error.h" +#include "xfs_health.h" /* * Invalidate any incore buffers associated with this remote attribute value @@ -147,6 +148,7 @@ xfs_attr3_node_inactive( if (level > XFS_DA_NODE_MAXDEPTH) { xfs_buf_mark_corrupt(bp); xfs_trans_brelse(*trans, bp); /* no locks for later trans */ + xfs_dirattr_mark_sick(dp, XFS_ATTR_FORK); return -EFSCORRUPTED; } @@ -197,6 +199,7 @@ xfs_attr3_node_inactive( default: xfs_buf_mark_corrupt(child_bp); xfs_trans_brelse(*trans, child_bp); + xfs_dirattr_mark_sick(dp, XFS_ATTR_FORK); error = -EFSCORRUPTED; break; } @@ -286,6 +289,7 @@ xfs_attr3_root_inactive( error = xfs_attr3_leaf_inactive(trans, dp, bp); break; default: + xfs_dirattr_mark_sick(dp, XFS_ATTR_FORK); error = -EFSCORRUPTED; xfs_buf_mark_corrupt(bp); xfs_trans_brelse(*trans, bp); diff --git a/fs/xfs/xfs_attr_item.c b/fs/xfs/xfs_attr_item.c index 9e02111bd890..9b4c61e1c22e 100644 --- a/fs/xfs/xfs_attr_item.c +++ b/fs/xfs/xfs_attr_item.c @@ -108,7 +108,7 @@ STATIC void xfs_attri_item_free( struct xfs_attri_log_item *attrip) { - kmem_free(attrip->attri_item.li_lv_shadow); + kvfree(attrip->attri_item.li_lv_shadow); xfs_attri_log_nameval_put(attrip->attri_nameval); kmem_cache_free(xfs_attri_cache, attrip); } @@ -226,7 +226,7 @@ xfs_attri_init( { struct xfs_attri_log_item *attrip; - attrip = kmem_cache_zalloc(xfs_attri_cache, GFP_NOFS | __GFP_NOFAIL); + attrip = kmem_cache_zalloc(xfs_attri_cache, GFP_KERNEL | __GFP_NOFAIL); /* * Grab an extra reference to the name/value buffer for this log item. @@ -251,7 +251,7 @@ static inline struct xfs_attrd_log_item *ATTRD_ITEM(struct xfs_log_item *lip) STATIC void xfs_attrd_item_free(struct xfs_attrd_log_item *attrdp) { - kmem_free(attrdp->attrd_item.li_lv_shadow); + kvfree(attrdp->attrd_item.li_lv_shadow); kmem_cache_free(xfs_attrd_cache, attrdp); } @@ -386,11 +386,16 @@ xfs_attr_free_item( xfs_da_state_free(attr->xattri_da_state); xfs_attri_log_nameval_put(attr->xattri_nameval); if (attr->xattri_da_args->op_flags & XFS_DA_OP_RECOVERY) - kmem_free(attr); + kfree(attr); else kmem_cache_free(xfs_attr_intent_cache, attr); } +static inline struct xfs_attr_intent *attri_entry(const struct list_head *e) +{ + return list_entry(e, struct xfs_attr_intent, xattri_list); +} + /* Process an attr. */ STATIC int xfs_attr_finish_item( @@ -399,11 +404,10 @@ xfs_attr_finish_item( struct list_head *item, struct xfs_btree_cur **state) { - struct xfs_attr_intent *attr; + struct xfs_attr_intent *attr = attri_entry(item); struct xfs_da_args *args; int error; - attr = container_of(item, struct xfs_attr_intent, xattri_list); args = attr->xattri_da_args; /* Reset trans after EAGAIN cycle since the transaction is new */ @@ -443,9 +447,8 @@ STATIC void xfs_attr_cancel_item( struct list_head *item) { - struct xfs_attr_intent *attr; + struct xfs_attr_intent *attr = attri_entry(item); - attr = container_of(item, struct xfs_attr_intent, xattri_list); xfs_attr_free_item(attr); } @@ -512,8 +515,8 @@ xfs_attri_recover_work( if (error) return ERR_PTR(error); - attr = kmem_zalloc(sizeof(struct xfs_attr_intent) + - sizeof(struct xfs_da_args), KM_NOFS); + attr = kzalloc(sizeof(struct xfs_attr_intent) + + sizeof(struct xfs_da_args), GFP_KERNEL | __GFP_NOFAIL); args = (struct xfs_da_args *)(attr + 1); attr->xattri_da_args = args; @@ -666,7 +669,7 @@ xfs_attr_create_done( attrip = ATTRI_ITEM(intent); - attrdp = kmem_cache_zalloc(xfs_attrd_cache, GFP_NOFS | __GFP_NOFAIL); + attrdp = kmem_cache_zalloc(xfs_attrd_cache, GFP_KERNEL | __GFP_NOFAIL); xfs_log_item_init(tp->t_mountp, &attrdp->attrd_item, XFS_LI_ATTRD, &xfs_attrd_item_ops); diff --git a/fs/xfs/xfs_attr_list.c b/fs/xfs/xfs_attr_list.c index e368ad671e26..a6819a642cc0 100644 --- a/fs/xfs/xfs_attr_list.c +++ b/fs/xfs/xfs_attr_list.c @@ -22,6 +22,7 @@ #include "xfs_error.h" #include "xfs_trace.h" #include "xfs_dir2.h" +#include "xfs_health.h" STATIC int xfs_attr_shortform_compare(const void *a, const void *b) @@ -82,8 +83,10 @@ xfs_attr_shortform_list( for (i = 0, sfe = xfs_attr_sf_firstentry(sf); i < sf->count; i++) { if (XFS_IS_CORRUPT(context->dp->i_mount, !xfs_attr_namecheck(sfe->nameval, - sfe->namelen))) + sfe->namelen))) { + xfs_dirattr_mark_sick(context->dp, XFS_ATTR_FORK); return -EFSCORRUPTED; + } context->put_listent(context, sfe->flags, sfe->nameval, @@ -109,7 +112,7 @@ xfs_attr_shortform_list( * It didn't all fit, so we have to sort everything on hashval. */ sbsize = sf->count * sizeof(*sbuf); - sbp = sbuf = kmem_alloc(sbsize, KM_NOFS); + sbp = sbuf = kmalloc(sbsize, GFP_KERNEL | __GFP_NOFAIL); /* * Scan the attribute list for the rest of the entries, storing @@ -124,7 +127,8 @@ xfs_attr_shortform_list( XFS_ERRLEVEL_LOW, context->dp->i_mount, sfe, sizeof(*sfe)); - kmem_free(sbuf); + kfree(sbuf); + xfs_dirattr_mark_sick(dp, XFS_ATTR_FORK); return -EFSCORRUPTED; } @@ -175,6 +179,7 @@ xfs_attr_shortform_list( if (XFS_IS_CORRUPT(context->dp->i_mount, !xfs_attr_namecheck(sbp->name, sbp->namelen))) { + xfs_dirattr_mark_sick(context->dp, XFS_ATTR_FORK); error = -EFSCORRUPTED; goto out; } @@ -188,7 +193,7 @@ xfs_attr_shortform_list( cursor->offset++; } out: - kmem_free(sbuf); + kfree(sbuf); return error; } @@ -262,8 +267,10 @@ xfs_attr_node_list_lookup( return 0; /* We can't point back to the root. */ - if (XFS_IS_CORRUPT(mp, cursor->blkno == 0)) + if (XFS_IS_CORRUPT(mp, cursor->blkno == 0)) { + xfs_dirattr_mark_sick(dp, XFS_ATTR_FORK); return -EFSCORRUPTED; + } } if (expected_level != 0) @@ -275,6 +282,7 @@ xfs_attr_node_list_lookup( out_corruptbuf: xfs_buf_mark_corrupt(bp); xfs_trans_brelse(tp, bp); + xfs_dirattr_mark_sick(dp, XFS_ATTR_FORK); return -EFSCORRUPTED; } @@ -304,6 +312,8 @@ xfs_attr_node_list( if (cursor->blkno > 0) { error = xfs_da3_node_read(context->tp, dp, cursor->blkno, &bp, XFS_ATTR_FORK); + if (xfs_metadata_is_sick(error)) + xfs_dirattr_mark_sick(dp, XFS_ATTR_FORK); if ((error != 0) && (error != -EFSCORRUPTED)) return error; if (bp) { @@ -464,8 +474,10 @@ xfs_attr3_leaf_list_int( } if (XFS_IS_CORRUPT(context->dp->i_mount, - !xfs_attr_namecheck(name, namelen))) + !xfs_attr_namecheck(name, namelen))) { + xfs_dirattr_mark_sick(context->dp, XFS_ATTR_FORK); return -EFSCORRUPTED; + } context->put_listent(context, entry->flags, name, namelen, valuelen); if (context->seen_enough) @@ -504,7 +516,7 @@ xfs_attr_list_ilocked( { struct xfs_inode *dp = context->dp; - ASSERT(xfs_isilocked(dp, XFS_ILOCK_SHARED | XFS_ILOCK_EXCL)); + xfs_assert_ilocked(dp, XFS_ILOCK_SHARED | XFS_ILOCK_EXCL); /* * Decide on what work routines to call based on the inode size. diff --git a/fs/xfs/xfs_bmap_item.c b/fs/xfs/xfs_bmap_item.c index 52fb8a148b7d..d27859a684aa 100644 --- a/fs/xfs/xfs_bmap_item.c +++ b/fs/xfs/xfs_bmap_item.c @@ -25,6 +25,7 @@ #include "xfs_log_priv.h" #include "xfs_log_recover.h" #include "xfs_ag.h" +#include "xfs_trace.h" struct kmem_cache *xfs_bui_cache; struct kmem_cache *xfs_bud_cache; @@ -40,7 +41,7 @@ STATIC void xfs_bui_item_free( struct xfs_bui_log_item *buip) { - kmem_free(buip->bui_item.li_lv_shadow); + kvfree(buip->bui_item.li_lv_shadow); kmem_cache_free(xfs_bui_cache, buip); } @@ -201,7 +202,7 @@ xfs_bud_item_release( struct xfs_bud_log_item *budp = BUD_ITEM(lip); xfs_bui_release(budp->bud_buip); - kmem_free(budp->bud_item.li_lv_shadow); + kvfree(budp->bud_item.li_lv_shadow); kmem_cache_free(xfs_bud_cache, budp); } @@ -221,6 +222,11 @@ static const struct xfs_item_ops xfs_bud_item_ops = { .iop_intent = xfs_bud_item_intent, }; +static inline struct xfs_bmap_intent *bi_entry(const struct list_head *e) +{ + return list_entry(e, struct xfs_bmap_intent, bi_list); +} + /* Sort bmap intents by inode. */ static int xfs_bmap_update_diff_items( @@ -228,37 +234,12 @@ xfs_bmap_update_diff_items( const struct list_head *a, const struct list_head *b) { - struct xfs_bmap_intent *ba; - struct xfs_bmap_intent *bb; + struct xfs_bmap_intent *ba = bi_entry(a); + struct xfs_bmap_intent *bb = bi_entry(b); - ba = container_of(a, struct xfs_bmap_intent, bi_list); - bb = container_of(b, struct xfs_bmap_intent, bi_list); return ba->bi_owner->i_ino - bb->bi_owner->i_ino; } -/* Set the map extent flags for this mapping. */ -static void -xfs_trans_set_bmap_flags( - struct xfs_map_extent *map, - enum xfs_bmap_intent_type type, - int whichfork, - xfs_exntst_t state) -{ - map->me_flags = 0; - switch (type) { - case XFS_BMAP_MAP: - case XFS_BMAP_UNMAP: - map->me_flags = type; - break; - default: - ASSERT(0); - } - if (state == XFS_EXT_UNWRITTEN) - map->me_flags |= XFS_BMAP_EXTENT_UNWRITTEN; - if (whichfork == XFS_ATTR_FORK) - map->me_flags |= XFS_BMAP_EXTENT_ATTR_FORK; -} - /* Log bmap updates in the intent item. */ STATIC void xfs_bmap_update_log_item( @@ -281,8 +262,21 @@ xfs_bmap_update_log_item( map->me_startblock = bi->bi_bmap.br_startblock; map->me_startoff = bi->bi_bmap.br_startoff; map->me_len = bi->bi_bmap.br_blockcount; - xfs_trans_set_bmap_flags(map, bi->bi_type, bi->bi_whichfork, - bi->bi_bmap.br_state); + + switch (bi->bi_type) { + case XFS_BMAP_MAP: + case XFS_BMAP_UNMAP: + map->me_flags = bi->bi_type; + break; + default: + ASSERT(0); + } + if (bi->bi_bmap.br_state == XFS_EXT_UNWRITTEN) + map->me_flags |= XFS_BMAP_EXTENT_UNWRITTEN; + if (bi->bi_whichfork == XFS_ATTR_FORK) + map->me_flags |= XFS_BMAP_EXTENT_ATTR_FORK; + if (xfs_ifork_is_realtime(bi->bi_owner, bi->bi_whichfork)) + map->me_flags |= XFS_BMAP_EXTENT_REALTIME; } static struct xfs_log_item * @@ -325,13 +319,16 @@ xfs_bmap_update_create_done( } /* Take a passive ref to the AG containing the space we're mapping. */ -void +static inline void xfs_bmap_update_get_group( struct xfs_mount *mp, struct xfs_bmap_intent *bi) { xfs_agnumber_t agno; + if (xfs_ifork_is_realtime(bi->bi_owner, bi->bi_whichfork)) + return; + agno = XFS_FSB_TO_AGNO(mp, bi->bi_bmap.br_startblock); /* @@ -344,14 +341,40 @@ xfs_bmap_update_get_group( bi->bi_pag = xfs_perag_intent_get(mp, agno); } +/* Add this deferred BUI to the transaction. */ +void +xfs_bmap_defer_add( + struct xfs_trans *tp, + struct xfs_bmap_intent *bi) +{ + trace_xfs_bmap_defer(bi); + + xfs_bmap_update_get_group(tp->t_mountp, bi); + xfs_defer_add(tp, &bi->bi_list, &xfs_bmap_update_defer_type); +} + /* Release a passive AG ref after finishing mapping work. */ static inline void xfs_bmap_update_put_group( struct xfs_bmap_intent *bi) { + if (xfs_ifork_is_realtime(bi->bi_owner, bi->bi_whichfork)) + return; + xfs_perag_intent_put(bi->bi_pag); } +/* Cancel a deferred bmap update. */ +STATIC void +xfs_bmap_update_cancel_item( + struct list_head *item) +{ + struct xfs_bmap_intent *bi = bi_entry(item); + + xfs_bmap_update_put_group(bi); + kmem_cache_free(xfs_bmap_intent_cache, bi); +} + /* Process a deferred bmap update. */ STATIC int xfs_bmap_update_finish_item( @@ -360,19 +383,16 @@ xfs_bmap_update_finish_item( struct list_head *item, struct xfs_btree_cur **state) { - struct xfs_bmap_intent *bi; + struct xfs_bmap_intent *bi = bi_entry(item); int error; - bi = container_of(item, struct xfs_bmap_intent, bi_list); - error = xfs_bmap_finish_one(tp, bi); if (!error && bi->bi_bmap.br_blockcount > 0) { ASSERT(bi->bi_type == XFS_BMAP_UNMAP); return -EAGAIN; } - xfs_bmap_update_put_group(bi); - kmem_cache_free(xfs_bmap_intent_cache, bi); + xfs_bmap_update_cancel_item(item); return error; } @@ -384,19 +404,6 @@ xfs_bmap_update_abort_intent( xfs_bui_release(BUI_ITEM(intent)); } -/* Cancel a deferred bmap update. */ -STATIC void -xfs_bmap_update_cancel_item( - struct list_head *item) -{ - struct xfs_bmap_intent *bi; - - bi = container_of(item, struct xfs_bmap_intent, bi_list); - - xfs_bmap_update_put_group(bi); - kmem_cache_free(xfs_bmap_intent_cache, bi); -} - /* Is this recovered BUI ok? */ static inline bool xfs_bui_validate( @@ -428,6 +435,9 @@ xfs_bui_validate( if (!xfs_verify_fileext(mp, map->me_startoff, map->me_len)) return false; + if (map->me_flags & XFS_BMAP_EXTENT_REALTIME) + return xfs_verify_rtbext(mp, map->me_startblock, map->me_len); + return xfs_verify_fsbext(mp, map->me_startblock, map->me_len); } @@ -445,7 +455,8 @@ xfs_bui_recover_work( if (error) return ERR_PTR(error); - bi = kmem_cache_zalloc(xfs_bmap_intent_cache, GFP_NOFS | __GFP_NOFAIL); + bi = kmem_cache_zalloc(xfs_bmap_intent_cache, + GFP_KERNEL | __GFP_NOFAIL); bi->bi_whichfork = (map->me_flags & XFS_BMAP_EXTENT_ATTR_FORK) ? XFS_ATTR_FORK : XFS_DATA_FORK; bi->bi_type = map->me_flags & XFS_BMAP_EXTENT_TYPE_MASK; @@ -502,6 +513,12 @@ xfs_bmap_recover_work( xfs_ilock(ip, XFS_ILOCK_EXCL); xfs_trans_ijoin(tp, ip, 0); + if (!!(map->me_flags & XFS_BMAP_EXTENT_REALTIME) != + xfs_ifork_is_realtime(ip, work->bi_whichfork)) { + error = -EFSCORRUPTED; + goto err_cancel; + } + if (work->bi_type == XFS_BMAP_MAP) iext_delta = XFS_IEXT_ADD_NOSPLIT_CNT; else diff --git a/fs/xfs/xfs_bmap_item.h b/fs/xfs/xfs_bmap_item.h index 3fafd3881a0b..6fee6a508343 100644 --- a/fs/xfs/xfs_bmap_item.h +++ b/fs/xfs/xfs_bmap_item.h @@ -68,4 +68,8 @@ struct xfs_bud_log_item { extern struct kmem_cache *xfs_bui_cache; extern struct kmem_cache *xfs_bud_cache; +struct xfs_bmap_intent; + +void xfs_bmap_defer_add(struct xfs_trans *tp, struct xfs_bmap_intent *bi); + #endif /* __XFS_BMAP_ITEM_H__ */ diff --git a/fs/xfs/xfs_bmap_util.c b/fs/xfs/xfs_bmap_util.c index c2531c28905c..19e11d1da660 100644 --- a/fs/xfs/xfs_bmap_util.c +++ b/fs/xfs/xfs_bmap_util.c @@ -66,7 +66,7 @@ xfs_zero_extent( return blkdev_issue_zeroout(target->bt_bdev, block << (mp->m_super->s_blocksize_bits - 9), count_fsb << (mp->m_super->s_blocksize_bits - 9), - GFP_NOFS, 0); + GFP_KERNEL, 0); } /* @@ -508,8 +508,8 @@ xfs_can_free_eofblocks( * Caller must either hold the exclusive io lock; or be inactivating * the inode, which guarantees there are no other users of the inode. */ - ASSERT(xfs_isilocked(ip, XFS_IOLOCK_EXCL) || - (VFS_I(ip)->i_state & I_FREEING)); + if (!(VFS_I(ip)->i_state & I_FREEING)) + xfs_assert_ilocked(ip, XFS_IOLOCK_EXCL); /* prealloc/delalloc exists only on regular files */ if (!S_ISREG(VFS_I(ip)->i_mode)) @@ -965,8 +965,7 @@ xfs_collapse_file_space( xfs_fileoff_t shift_fsb = XFS_B_TO_FSB(mp, len); bool done = false; - ASSERT(xfs_isilocked(ip, XFS_IOLOCK_EXCL)); - ASSERT(xfs_isilocked(ip, XFS_MMAPLOCK_EXCL)); + xfs_assert_ilocked(ip, XFS_IOLOCK_EXCL | XFS_MMAPLOCK_EXCL); trace_xfs_collapse_file_space(ip); @@ -1035,8 +1034,7 @@ xfs_insert_file_space( xfs_fileoff_t shift_fsb = XFS_B_TO_FSB(mp, len); bool done = false; - ASSERT(xfs_isilocked(ip, XFS_IOLOCK_EXCL)); - ASSERT(xfs_isilocked(ip, XFS_MMAPLOCK_EXCL)); + xfs_assert_ilocked(ip, XFS_IOLOCK_EXCL | XFS_MMAPLOCK_EXCL); trace_xfs_insert_file_space(ip); @@ -1307,16 +1305,16 @@ xfs_swap_extent_rmap( } /* Remove the mapping from the donor file. */ - xfs_bmap_unmap_extent(tp, tip, &uirec); + xfs_bmap_unmap_extent(tp, tip, XFS_DATA_FORK, &uirec); /* Remove the mapping from the source file. */ - xfs_bmap_unmap_extent(tp, ip, &irec); + xfs_bmap_unmap_extent(tp, ip, XFS_DATA_FORK, &irec); /* Map the donor file's blocks into the source file. */ - xfs_bmap_map_extent(tp, ip, &uirec); + xfs_bmap_map_extent(tp, ip, XFS_DATA_FORK, &uirec); /* Map the source file's blocks into the donor file. */ - xfs_bmap_map_extent(tp, tip, &irec); + xfs_bmap_map_extent(tp, tip, XFS_DATA_FORK, &irec); error = xfs_defer_finish(tpp); tp = *tpp; diff --git a/fs/xfs/xfs_buf.c b/fs/xfs/xfs_buf.c index 01b41fabbe3c..1a18c381127e 100644 --- a/fs/xfs/xfs_buf.c +++ b/fs/xfs/xfs_buf.c @@ -21,6 +21,7 @@ #include "xfs_errortag.h" #include "xfs_error.h" #include "xfs_ag.h" +#include "xfs_buf_mem.h" struct kmem_cache *xfs_buf_cache; @@ -60,6 +61,11 @@ xfs_buf_submit( return __xfs_buf_submit(bp, !(bp->b_flags & XBF_ASYNC)); } +static inline bool xfs_buf_is_uncached(struct xfs_buf *bp) +{ + return bp->b_rhash_key == XFS_BUF_DADDR_NULL; +} + static inline int xfs_buf_is_vmapped( struct xfs_buf *bp) @@ -189,8 +195,8 @@ xfs_buf_get_maps( return 0; } - bp->b_maps = kmem_zalloc(map_count * sizeof(struct xfs_buf_map), - KM_NOFS); + bp->b_maps = kzalloc(map_count * sizeof(struct xfs_buf_map), + GFP_KERNEL | __GFP_NOLOCKDEP | __GFP_NOFAIL); if (!bp->b_maps) return -ENOMEM; return 0; @@ -204,7 +210,7 @@ xfs_buf_free_maps( struct xfs_buf *bp) { if (bp->b_maps != &bp->__b_map) { - kmem_free(bp->b_maps); + kfree(bp->b_maps); bp->b_maps = NULL; } } @@ -222,7 +228,8 @@ _xfs_buf_alloc( int i; *bpp = NULL; - bp = kmem_cache_zalloc(xfs_buf_cache, GFP_NOFS | __GFP_NOFAIL); + bp = kmem_cache_zalloc(xfs_buf_cache, + GFP_KERNEL | __GFP_NOLOCKDEP | __GFP_NOFAIL); /* * We don't want certain flags to appear in b_flags unless they are @@ -289,7 +296,7 @@ xfs_buf_free_pages( mm_account_reclaimed_pages(bp->b_page_count); if (bp->b_pages != bp->b_page_array) - kmem_free(bp->b_pages); + kfree(bp->b_pages); bp->b_pages = NULL; bp->b_flags &= ~_XBF_PAGES; } @@ -312,10 +319,12 @@ xfs_buf_free( ASSERT(list_empty(&bp->b_lru)); - if (bp->b_flags & _XBF_PAGES) + if (xfs_buftarg_is_mem(bp->b_target)) + xmbuf_unmap_page(bp); + else if (bp->b_flags & _XBF_PAGES) xfs_buf_free_pages(bp); else if (bp->b_flags & _XBF_KMEM) - kmem_free(bp->b_addr); + kfree(bp->b_addr); call_rcu(&bp->b_rcu, xfs_buf_free_callback); } @@ -325,21 +334,21 @@ xfs_buf_alloc_kmem( struct xfs_buf *bp, xfs_buf_flags_t flags) { - xfs_km_flags_t kmflag_mask = KM_NOFS; + gfp_t gfp_mask = GFP_KERNEL | __GFP_NOLOCKDEP | __GFP_NOFAIL; size_t size = BBTOB(bp->b_length); /* Assure zeroed buffer for non-read cases. */ if (!(flags & XBF_READ)) - kmflag_mask |= KM_ZERO; + gfp_mask |= __GFP_ZERO; - bp->b_addr = kmem_alloc(size, kmflag_mask); + bp->b_addr = kmalloc(size, gfp_mask); if (!bp->b_addr) return -ENOMEM; if (((unsigned long)(bp->b_addr + size - 1) & PAGE_MASK) != ((unsigned long)bp->b_addr & PAGE_MASK)) { /* b_addr spans two pages - use alloc_page instead */ - kmem_free(bp->b_addr); + kfree(bp->b_addr); bp->b_addr = NULL; return -ENOMEM; } @@ -356,13 +365,11 @@ xfs_buf_alloc_pages( struct xfs_buf *bp, xfs_buf_flags_t flags) { - gfp_t gfp_mask = __GFP_NOWARN; + gfp_t gfp_mask = GFP_KERNEL | __GFP_NOLOCKDEP | __GFP_NOWARN; long filled = 0; if (flags & XBF_READ_AHEAD) gfp_mask |= __GFP_NORETRY; - else - gfp_mask |= GFP_NOFS; /* Make sure that we have a page list */ bp->b_page_count = DIV_ROUND_UP(BBTOB(bp->b_length), PAGE_SIZE); @@ -429,11 +436,18 @@ _xfs_buf_map_pages( /* * vm_map_ram() will allocate auxiliary structures (e.g. - * pagetables) with GFP_KERNEL, yet we are likely to be under - * GFP_NOFS context here. Hence we need to tell memory reclaim - * that we are in such a context via PF_MEMALLOC_NOFS to prevent - * memory reclaim re-entering the filesystem here and - * potentially deadlocking. + * pagetables) with GFP_KERNEL, yet we often under a scoped nofs + * context here. Mixing GFP_KERNEL with GFP_NOFS allocations + * from the same call site that can be run from both above and + * below memory reclaim causes lockdep false positives. Hence we + * always need to force this allocation to nofs context because + * we can't pass __GFP_NOLOCKDEP down to auxillary structures to + * prevent false positive lockdep reports. + * + * XXX(dgc): I think dquot reclaim is the only place we can get + * to this function from memory reclaim context now. If we fix + * that like we've fixed inode reclaim to avoid writeback from + * reclaim, this nofs wrapping can go away. */ nofs_flag = memalloc_nofs_save(); do { @@ -499,18 +513,18 @@ static const struct rhashtable_params xfs_buf_hash_params = { }; int -xfs_buf_hash_init( - struct xfs_perag *pag) +xfs_buf_cache_init( + struct xfs_buf_cache *bch) { - spin_lock_init(&pag->pag_buf_lock); - return rhashtable_init(&pag->pag_buf_hash, &xfs_buf_hash_params); + spin_lock_init(&bch->bc_lock); + return rhashtable_init(&bch->bc_hash, &xfs_buf_hash_params); } void -xfs_buf_hash_destroy( - struct xfs_perag *pag) +xfs_buf_cache_destroy( + struct xfs_buf_cache *bch) { - rhashtable_destroy(&pag->pag_buf_hash); + rhashtable_destroy(&bch->bc_hash); } static int @@ -573,7 +587,7 @@ xfs_buf_find_lock( static inline int xfs_buf_lookup( - struct xfs_perag *pag, + struct xfs_buf_cache *bch, struct xfs_buf_map *map, xfs_buf_flags_t flags, struct xfs_buf **bpp) @@ -582,7 +596,7 @@ xfs_buf_lookup( int error; rcu_read_lock(); - bp = rhashtable_lookup(&pag->pag_buf_hash, map, xfs_buf_hash_params); + bp = rhashtable_lookup(&bch->bc_hash, map, xfs_buf_hash_params); if (!bp || !atomic_inc_not_zero(&bp->b_hold)) { rcu_read_unlock(); return -ENOENT; @@ -607,6 +621,7 @@ xfs_buf_lookup( static int xfs_buf_find_insert( struct xfs_buftarg *btp, + struct xfs_buf_cache *bch, struct xfs_perag *pag, struct xfs_buf_map *cmap, struct xfs_buf_map *map, @@ -622,31 +637,33 @@ xfs_buf_find_insert( if (error) goto out_drop_pag; - /* - * For buffers that fit entirely within a single page, first attempt to - * allocate the memory from the heap to minimise memory usage. If we - * can't get heap memory for these small buffers, we fall back to using - * the page allocator. - */ - if (BBTOB(new_bp->b_length) >= PAGE_SIZE || - xfs_buf_alloc_kmem(new_bp, flags) < 0) { + if (xfs_buftarg_is_mem(new_bp->b_target)) { + error = xmbuf_map_page(new_bp); + } else if (BBTOB(new_bp->b_length) >= PAGE_SIZE || + xfs_buf_alloc_kmem(new_bp, flags) < 0) { + /* + * For buffers that fit entirely within a single page, first + * attempt to allocate the memory from the heap to minimise + * memory usage. If we can't get heap memory for these small + * buffers, we fall back to using the page allocator. + */ error = xfs_buf_alloc_pages(new_bp, flags); - if (error) - goto out_free_buf; } + if (error) + goto out_free_buf; - spin_lock(&pag->pag_buf_lock); - bp = rhashtable_lookup_get_insert_fast(&pag->pag_buf_hash, + spin_lock(&bch->bc_lock); + bp = rhashtable_lookup_get_insert_fast(&bch->bc_hash, &new_bp->b_rhash_head, xfs_buf_hash_params); if (IS_ERR(bp)) { error = PTR_ERR(bp); - spin_unlock(&pag->pag_buf_lock); + spin_unlock(&bch->bc_lock); goto out_free_buf; } if (bp) { /* found an existing buffer */ atomic_inc(&bp->b_hold); - spin_unlock(&pag->pag_buf_lock); + spin_unlock(&bch->bc_lock); error = xfs_buf_find_lock(bp, flags); if (error) xfs_buf_rele(bp); @@ -657,17 +674,40 @@ xfs_buf_find_insert( /* The new buffer keeps the perag reference until it is freed. */ new_bp->b_pag = pag; - spin_unlock(&pag->pag_buf_lock); + spin_unlock(&bch->bc_lock); *bpp = new_bp; return 0; out_free_buf: xfs_buf_free(new_bp); out_drop_pag: - xfs_perag_put(pag); + if (pag) + xfs_perag_put(pag); return error; } +static inline struct xfs_perag * +xfs_buftarg_get_pag( + struct xfs_buftarg *btp, + const struct xfs_buf_map *map) +{ + struct xfs_mount *mp = btp->bt_mount; + + if (xfs_buftarg_is_mem(btp)) + return NULL; + return xfs_perag_get(mp, xfs_daddr_to_agno(mp, map->bm_bn)); +} + +static inline struct xfs_buf_cache * +xfs_buftarg_buf_cache( + struct xfs_buftarg *btp, + struct xfs_perag *pag) +{ + if (pag) + return &pag->pag_bcache; + return btp->bt_cache; +} + /* * Assembles a buffer covering the specified range. The code is optimised for * cache hits, as metadata intensive workloads will see 3 orders of magnitude @@ -681,6 +721,7 @@ xfs_buf_get_map( xfs_buf_flags_t flags, struct xfs_buf **bpp) { + struct xfs_buf_cache *bch; struct xfs_perag *pag; struct xfs_buf *bp = NULL; struct xfs_buf_map cmap = { .bm_bn = map[0].bm_bn }; @@ -696,10 +737,10 @@ xfs_buf_get_map( if (error) return error; - pag = xfs_perag_get(btp->bt_mount, - xfs_daddr_to_agno(btp->bt_mount, cmap.bm_bn)); + pag = xfs_buftarg_get_pag(btp, &cmap); + bch = xfs_buftarg_buf_cache(btp, pag); - error = xfs_buf_lookup(pag, &cmap, flags, &bp); + error = xfs_buf_lookup(bch, &cmap, flags, &bp); if (error && error != -ENOENT) goto out_put_perag; @@ -711,13 +752,14 @@ xfs_buf_get_map( goto out_put_perag; /* xfs_buf_find_insert() consumes the perag reference. */ - error = xfs_buf_find_insert(btp, pag, &cmap, map, nmaps, + error = xfs_buf_find_insert(btp, bch, pag, &cmap, map, nmaps, flags, &bp); if (error) return error; } else { XFS_STATS_INC(btp->bt_mount, xb_get_locked); - xfs_perag_put(pag); + if (pag) + xfs_perag_put(pag); } /* We do not hold a perag reference anymore. */ @@ -745,7 +787,8 @@ xfs_buf_get_map( return 0; out_put_perag: - xfs_perag_put(pag); + if (pag) + xfs_perag_put(pag); return error; } @@ -892,6 +935,13 @@ xfs_buf_readahead_map( { struct xfs_buf *bp; + /* + * Currently we don't have a good means or justification for performing + * xmbuf_map_page asynchronously, so we don't do readahead. + */ + if (xfs_buftarg_is_mem(target)) + return; + xfs_buf_read_map(target, map, nmaps, XBF_TRYLOCK | XBF_ASYNC | XBF_READ_AHEAD, &bp, ops, __this_address); @@ -957,7 +1007,10 @@ xfs_buf_get_uncached( if (error) return error; - error = xfs_buf_alloc_pages(bp, flags); + if (xfs_buftarg_is_mem(bp->b_target)) + error = xmbuf_map_page(bp); + else + error = xfs_buf_alloc_pages(bp, flags); if (error) goto fail_free_buf; @@ -990,29 +1043,29 @@ xfs_buf_hold( atomic_inc(&bp->b_hold); } -/* - * Release a hold on the specified buffer. If the hold count is 1, the buffer is - * placed on LRU or freed (depending on b_lru_ref). - */ -void -xfs_buf_rele( +static void +xfs_buf_rele_uncached( + struct xfs_buf *bp) +{ + ASSERT(list_empty(&bp->b_lru)); + if (atomic_dec_and_test(&bp->b_hold)) { + xfs_buf_ioacct_dec(bp); + xfs_buf_free(bp); + } +} + +static void +xfs_buf_rele_cached( struct xfs_buf *bp) { + struct xfs_buftarg *btp = bp->b_target; struct xfs_perag *pag = bp->b_pag; + struct xfs_buf_cache *bch = xfs_buftarg_buf_cache(btp, pag); bool release; bool freebuf = false; trace_xfs_buf_rele(bp, _RET_IP_); - if (!pag) { - ASSERT(list_empty(&bp->b_lru)); - if (atomic_dec_and_test(&bp->b_hold)) { - xfs_buf_ioacct_dec(bp); - xfs_buf_free(bp); - } - return; - } - ASSERT(atomic_read(&bp->b_hold) > 0); /* @@ -1026,7 +1079,7 @@ xfs_buf_rele( * leading to a use-after-free scenario. */ spin_lock(&bp->b_lock); - release = atomic_dec_and_lock(&bp->b_hold, &pag->pag_buf_lock); + release = atomic_dec_and_lock(&bp->b_hold, &bch->bc_lock); if (!release) { /* * Drop the in-flight state if the buffer is already on the LRU @@ -1047,11 +1100,11 @@ xfs_buf_rele( * buffer for the LRU and clear the (now stale) dispose list * state flag */ - if (list_lru_add_obj(&bp->b_target->bt_lru, &bp->b_lru)) { + if (list_lru_add_obj(&btp->bt_lru, &bp->b_lru)) { bp->b_state &= ~XFS_BSTATE_DISPOSE; atomic_inc(&bp->b_hold); } - spin_unlock(&pag->pag_buf_lock); + spin_unlock(&bch->bc_lock); } else { /* * most of the time buffers will already be removed from the @@ -1060,16 +1113,17 @@ xfs_buf_rele( * was on was the disposal list */ if (!(bp->b_state & XFS_BSTATE_DISPOSE)) { - list_lru_del_obj(&bp->b_target->bt_lru, &bp->b_lru); + list_lru_del_obj(&btp->bt_lru, &bp->b_lru); } else { ASSERT(list_empty(&bp->b_lru)); } ASSERT(!(bp->b_flags & _XBF_DELWRI_Q)); - rhashtable_remove_fast(&pag->pag_buf_hash, &bp->b_rhash_head, - xfs_buf_hash_params); - spin_unlock(&pag->pag_buf_lock); - xfs_perag_put(pag); + rhashtable_remove_fast(&bch->bc_hash, &bp->b_rhash_head, + xfs_buf_hash_params); + spin_unlock(&bch->bc_lock); + if (pag) + xfs_perag_put(pag); freebuf = true; } @@ -1080,6 +1134,19 @@ out_unlock: xfs_buf_free(bp); } +/* + * Release a hold on the specified buffer. + */ +void +xfs_buf_rele( + struct xfs_buf *bp) +{ + trace_xfs_buf_rele(bp, _RET_IP_); + if (xfs_buf_is_uncached(bp)) + xfs_buf_rele_uncached(bp); + else + xfs_buf_rele_cached(bp); +} /* * Lock a buffer object, if it is not already locked. @@ -1585,6 +1652,12 @@ _xfs_buf_ioapply( /* we only use the buffer cache for meta-data */ op |= REQ_META; + /* in-memory targets are directly mapped, no IO required. */ + if (xfs_buftarg_is_mem(bp->b_target)) { + xfs_buf_ioend(bp); + return; + } + /* * Walk all the vectors issuing IO on them. Set up the initial offset * into the buffer and the desired IO size before we start - @@ -1940,25 +2013,30 @@ xfs_buftarg_shrink_count( } void -xfs_free_buftarg( +xfs_destroy_buftarg( struct xfs_buftarg *btp) { shrinker_free(btp->bt_shrinker); ASSERT(percpu_counter_sum(&btp->bt_io_count) == 0); percpu_counter_destroy(&btp->bt_io_count); list_lru_destroy(&btp->bt_lru); +} +void +xfs_free_buftarg( + struct xfs_buftarg *btp) +{ + xfs_destroy_buftarg(btp); fs_put_dax(btp->bt_daxdev, btp->bt_mount); /* the main block device is closed by kill_block_super */ if (btp->bt_bdev != btp->bt_mount->m_super->s_bdev) fput(btp->bt_bdev_file); - - kmem_free(btp); + kfree(btp); } int xfs_setsize_buftarg( - xfs_buftarg_t *btp, + struct xfs_buftarg *btp, unsigned int sectorsize) { /* Set up metadata sector size info */ @@ -1972,23 +2050,46 @@ xfs_setsize_buftarg( return -EINVAL; } - /* Set up device logical sector size mask */ - btp->bt_logical_sectorsize = bdev_logical_block_size(btp->bt_bdev); - btp->bt_logical_sectormask = bdev_logical_block_size(btp->bt_bdev) - 1; - return 0; } -/* - * When allocating the initial buffer target we have not yet - * read in the superblock, so don't know what sized sectors - * are being used at this early stage. Play safe. - */ -STATIC int -xfs_setsize_buftarg_early( - xfs_buftarg_t *btp) +int +xfs_init_buftarg( + struct xfs_buftarg *btp, + size_t logical_sectorsize, + const char *descr) { - return xfs_setsize_buftarg(btp, bdev_logical_block_size(btp->bt_bdev)); + /* Set up device logical sector size mask */ + btp->bt_logical_sectorsize = logical_sectorsize; + btp->bt_logical_sectormask = logical_sectorsize - 1; + + /* + * Buffer IO error rate limiting. Limit it to no more than 10 messages + * per 30 seconds so as to not spam logs too much on repeated errors. + */ + ratelimit_state_init(&btp->bt_ioerror_rl, 30 * HZ, + DEFAULT_RATELIMIT_BURST); + + if (list_lru_init(&btp->bt_lru)) + return -ENOMEM; + if (percpu_counter_init(&btp->bt_io_count, 0, GFP_KERNEL)) + goto out_destroy_lru; + + btp->bt_shrinker = + shrinker_alloc(SHRINKER_NUMA_AWARE, "xfs-buf:%s", descr); + if (!btp->bt_shrinker) + goto out_destroy_io_count; + btp->bt_shrinker->count_objects = xfs_buftarg_shrink_count; + btp->bt_shrinker->scan_objects = xfs_buftarg_shrink_scan; + btp->bt_shrinker->private_data = btp; + shrinker_register(btp->bt_shrinker); + return 0; + +out_destroy_io_count: + percpu_counter_destroy(&btp->bt_io_count); +out_destroy_lru: + list_lru_destroy(&btp->bt_lru); + return -ENOMEM; } struct xfs_buftarg * @@ -1996,13 +2097,13 @@ xfs_alloc_buftarg( struct xfs_mount *mp, struct file *bdev_file) { - xfs_buftarg_t *btp; + struct xfs_buftarg *btp; const struct dax_holder_operations *ops = NULL; #if defined(CONFIG_FS_DAX) && defined(CONFIG_MEMORY_FAILURE) ops = &xfs_dax_holder_operations; #endif - btp = kmem_zalloc(sizeof(*btp), KM_NOFS); + btp = kzalloc(sizeof(*btp), GFP_KERNEL | __GFP_NOFAIL); btp->bt_mount = mp; btp->bt_bdev_file = bdev_file; @@ -2012,40 +2113,19 @@ xfs_alloc_buftarg( mp, ops); /* - * Buffer IO error rate limiting. Limit it to no more than 10 messages - * per 30 seconds so as to not spam logs too much on repeated errors. + * When allocating the buftargs we have not yet read the super block and + * thus don't know the file system sector size yet. */ - ratelimit_state_init(&btp->bt_ioerror_rl, 30 * HZ, - DEFAULT_RATELIMIT_BURST); - - if (xfs_setsize_buftarg_early(btp)) + if (xfs_setsize_buftarg(btp, bdev_logical_block_size(btp->bt_bdev))) goto error_free; - - if (list_lru_init(&btp->bt_lru)) + if (xfs_init_buftarg(btp, bdev_logical_block_size(btp->bt_bdev), + mp->m_super->s_id)) goto error_free; - if (percpu_counter_init(&btp->bt_io_count, 0, GFP_KERNEL)) - goto error_lru; - - btp->bt_shrinker = shrinker_alloc(SHRINKER_NUMA_AWARE, "xfs-buf:%s", - mp->m_super->s_id); - if (!btp->bt_shrinker) - goto error_pcpu; - - btp->bt_shrinker->count_objects = xfs_buftarg_shrink_count; - btp->bt_shrinker->scan_objects = xfs_buftarg_shrink_scan; - btp->bt_shrinker->private_data = btp; - - shrinker_register(btp->bt_shrinker); - return btp; -error_pcpu: - percpu_counter_destroy(&btp->bt_io_count); -error_lru: - list_lru_destroy(&btp->bt_lru); error_free: - kmem_free(btp); + kfree(btp); return NULL; } diff --git a/fs/xfs/xfs_buf.h b/fs/xfs/xfs_buf.h index 304e858d04fb..b1580644501f 100644 --- a/fs/xfs/xfs_buf.h +++ b/fs/xfs/xfs_buf.h @@ -83,6 +83,14 @@ typedef unsigned int xfs_buf_flags_t; #define XFS_BSTATE_DISPOSE (1 << 0) /* buffer being discarded */ #define XFS_BSTATE_IN_FLIGHT (1 << 1) /* I/O in flight */ +struct xfs_buf_cache { + spinlock_t bc_lock; + struct rhashtable bc_hash; +}; + +int xfs_buf_cache_init(struct xfs_buf_cache *bch); +void xfs_buf_cache_destroy(struct xfs_buf_cache *bch); + /* * The xfs_buftarg contains 2 notions of "sector size" - * @@ -96,11 +104,12 @@ typedef unsigned int xfs_buf_flags_t; * The latter is derived from the underlying device, and controls direct IO * alignment constraints. */ -typedef struct xfs_buftarg { +struct xfs_buftarg { dev_t bt_dev; struct file *bt_bdev_file; struct block_device *bt_bdev; struct dax_device *bt_daxdev; + struct file *bt_file; u64 bt_dax_part_off; struct xfs_mount *bt_mount; unsigned int bt_meta_sectorsize; @@ -114,7 +123,10 @@ typedef struct xfs_buftarg { struct percpu_counter bt_io_count; struct ratelimit_state bt_ioerror_rl; -} xfs_buftarg_t; + + /* built-in cache, if we're not using the perag one */ + struct xfs_buf_cache bt_cache[]; +}; #define XB_PAGES 2 @@ -379,4 +391,9 @@ int xfs_buf_reverify(struct xfs_buf *bp, const struct xfs_buf_ops *ops); bool xfs_verify_magic(struct xfs_buf *bp, __be32 dmagic); bool xfs_verify_magic16(struct xfs_buf *bp, __be16 dmagic); +/* for xfs_buf_mem.c only: */ +int xfs_init_buftarg(struct xfs_buftarg *btp, size_t logical_sectorsize, + const char *descr); +void xfs_destroy_buftarg(struct xfs_buftarg *btp); + #endif /* __XFS_BUF_H__ */ diff --git a/fs/xfs/xfs_buf_item.c b/fs/xfs/xfs_buf_item.c index 023d4e0385dd..43031842341a 100644 --- a/fs/xfs/xfs_buf_item.c +++ b/fs/xfs/xfs_buf_item.c @@ -805,8 +805,8 @@ xfs_buf_item_get_format( return; } - bip->bli_formats = kmem_zalloc(count * sizeof(struct xfs_buf_log_format), - 0); + bip->bli_formats = kzalloc(count * sizeof(struct xfs_buf_log_format), + GFP_KERNEL | __GFP_NOFAIL); } STATIC void @@ -814,7 +814,7 @@ xfs_buf_item_free_format( struct xfs_buf_log_item *bip) { if (bip->bli_formats != &bip->__bli_format) { - kmem_free(bip->bli_formats); + kfree(bip->bli_formats); bip->bli_formats = NULL; } } @@ -1044,7 +1044,7 @@ xfs_buf_item_free( struct xfs_buf_log_item *bip) { xfs_buf_item_free_format(bip); - kmem_free(bip->bli_item.li_lv_shadow); + kvfree(bip->bli_item.li_lv_shadow); kmem_cache_free(xfs_buf_item_cache, bip); } diff --git a/fs/xfs/xfs_buf_item_recover.c b/fs/xfs/xfs_buf_item_recover.c index 43167f543afc..09e893cf563c 100644 --- a/fs/xfs/xfs_buf_item_recover.c +++ b/fs/xfs/xfs_buf_item_recover.c @@ -85,7 +85,7 @@ xlog_add_buffer_cancelled( return false; } - bcp = kmem_alloc(sizeof(struct xfs_buf_cancel), 0); + bcp = kmalloc(sizeof(struct xfs_buf_cancel), GFP_KERNEL | __GFP_NOFAIL); bcp->bc_blkno = blkno; bcp->bc_len = len; bcp->bc_refcount = 1; @@ -129,7 +129,7 @@ xlog_put_buffer_cancelled( if (--bcp->bc_refcount == 0) { list_del(&bcp->bc_list); - kmem_free(bcp); + kfree(bcp); } return true; } @@ -1062,10 +1062,10 @@ xlog_free_buf_cancel_table( &log->l_buf_cancel_table[i], struct xfs_buf_cancel, bc_list))) { list_del(&bc->bc_list); - kmem_free(bc); + kfree(bc); } } - kmem_free(log->l_buf_cancel_table); + kfree(log->l_buf_cancel_table); log->l_buf_cancel_table = NULL; } diff --git a/fs/xfs/xfs_buf_mem.c b/fs/xfs/xfs_buf_mem.c new file mode 100644 index 000000000000..8ad38c64708e --- /dev/null +++ b/fs/xfs/xfs_buf_mem.c @@ -0,0 +1,270 @@ +// SPDX-License-Identifier: GPL-2.0-or-later +/* + * Copyright (c) 2023-2024 Oracle. All Rights Reserved. + * Author: Darrick J. Wong <djwong@kernel.org> + */ +#include "xfs.h" +#include "xfs_fs.h" +#include "xfs_buf.h" +#include "xfs_buf_mem.h" +#include "xfs_trace.h" +#include <linux/shmem_fs.h> +#include "xfs_log_format.h" +#include "xfs_trans.h" +#include "xfs_buf_item.h" +#include "xfs_error.h" + +/* + * Buffer Cache for In-Memory Files + * ================================ + * + * Online fsck wants to create ephemeral ordered recordsets. The existing + * btree infrastructure can do this, but we need the buffer cache to target + * memory instead of block devices. + * + * When CONFIG_TMPFS=y, shmemfs is enough of a filesystem to meet those + * requirements. Therefore, the xmbuf mechanism uses an unlinked shmem file to + * store our staging data. This file is not installed in the file descriptor + * table so that user programs cannot access the data, which means that the + * xmbuf must be freed with xmbuf_destroy. + * + * xmbufs assume that the caller will handle all required concurrency + * management; standard vfs locks (freezer and inode) are not taken. Reads + * and writes are satisfied directly from the page cache. + * + * The only supported block size is PAGE_SIZE, and we cannot use highmem. + */ + +/* + * shmem files used to back an in-memory buffer cache must not be exposed to + * userspace. Upper layers must coordinate access to the one handle returned + * by the constructor, so establish a separate lock class for xmbufs to avoid + * confusing lockdep. + */ +static struct lock_class_key xmbuf_i_mutex_key; + +/* + * Allocate a buffer cache target for a memory-backed file and set up the + * buffer target. + */ +int +xmbuf_alloc( + struct xfs_mount *mp, + const char *descr, + struct xfs_buftarg **btpp) +{ + struct file *file; + struct inode *inode; + struct xfs_buftarg *btp; + int error; + + btp = kzalloc(struct_size(btp, bt_cache, 1), GFP_KERNEL); + if (!btp) + return -ENOMEM; + + file = shmem_kernel_file_setup(descr, 0, 0); + if (IS_ERR(file)) { + error = PTR_ERR(file); + goto out_free_btp; + } + inode = file_inode(file); + + /* private file, private locking */ + lockdep_set_class(&inode->i_rwsem, &xmbuf_i_mutex_key); + + /* + * We don't want to bother with kmapping data during repair, so don't + * allow highmem pages to back this mapping. + */ + mapping_set_gfp_mask(inode->i_mapping, GFP_KERNEL); + + /* ensure all writes are below EOF to avoid pagecache zeroing */ + i_size_write(inode, inode->i_sb->s_maxbytes); + + trace_xmbuf_create(btp); + + error = xfs_buf_cache_init(btp->bt_cache); + if (error) + goto out_file; + + /* Initialize buffer target */ + btp->bt_mount = mp; + btp->bt_dev = (dev_t)-1U; + btp->bt_bdev = NULL; /* in-memory buftargs have no bdev */ + btp->bt_file = file; + btp->bt_meta_sectorsize = XMBUF_BLOCKSIZE; + btp->bt_meta_sectormask = XMBUF_BLOCKSIZE - 1; + + error = xfs_init_buftarg(btp, XMBUF_BLOCKSIZE, descr); + if (error) + goto out_bcache; + + *btpp = btp; + return 0; + +out_bcache: + xfs_buf_cache_destroy(btp->bt_cache); +out_file: + fput(file); +out_free_btp: + kfree(btp); + return error; +} + +/* Free a buffer cache target for a memory-backed buffer cache. */ +void +xmbuf_free( + struct xfs_buftarg *btp) +{ + ASSERT(xfs_buftarg_is_mem(btp)); + ASSERT(percpu_counter_sum(&btp->bt_io_count) == 0); + + trace_xmbuf_free(btp); + + xfs_destroy_buftarg(btp); + xfs_buf_cache_destroy(btp->bt_cache); + fput(btp->bt_file); + kfree(btp); +} + +/* Directly map a shmem page into the buffer cache. */ +int +xmbuf_map_page( + struct xfs_buf *bp) +{ + struct inode *inode = file_inode(bp->b_target->bt_file); + struct folio *folio = NULL; + struct page *page; + loff_t pos = BBTOB(xfs_buf_daddr(bp)); + int error; + + ASSERT(xfs_buftarg_is_mem(bp->b_target)); + + if (bp->b_map_count != 1) + return -ENOMEM; + if (BBTOB(bp->b_length) != XMBUF_BLOCKSIZE) + return -ENOMEM; + if (offset_in_page(pos) != 0) { + ASSERT(offset_in_page(pos)); + return -ENOMEM; + } + + error = shmem_get_folio(inode, pos >> PAGE_SHIFT, &folio, SGP_CACHE); + if (error) + return error; + + if (filemap_check_wb_err(inode->i_mapping, 0)) { + folio_unlock(folio); + folio_put(folio); + return -EIO; + } + + page = folio_file_page(folio, pos >> PAGE_SHIFT); + + /* + * Mark the page dirty so that it won't be reclaimed once we drop the + * (potentially last) reference in xmbuf_unmap_page. + */ + set_page_dirty(page); + unlock_page(page); + + bp->b_addr = page_address(page); + bp->b_pages = bp->b_page_array; + bp->b_pages[0] = page; + bp->b_page_count = 1; + return 0; +} + +/* Unmap a shmem page that was mapped into the buffer cache. */ +void +xmbuf_unmap_page( + struct xfs_buf *bp) +{ + struct page *page = bp->b_pages[0]; + + ASSERT(xfs_buftarg_is_mem(bp->b_target)); + + put_page(page); + + bp->b_addr = NULL; + bp->b_pages[0] = NULL; + bp->b_pages = NULL; + bp->b_page_count = 0; +} + +/* Is this a valid daddr within the buftarg? */ +bool +xmbuf_verify_daddr( + struct xfs_buftarg *btp, + xfs_daddr_t daddr) +{ + struct inode *inode = file_inode(btp->bt_file); + + ASSERT(xfs_buftarg_is_mem(btp)); + + return daddr < (inode->i_sb->s_maxbytes >> BBSHIFT); +} + +/* Discard the page backing this buffer. */ +static void +xmbuf_stale( + struct xfs_buf *bp) +{ + struct inode *inode = file_inode(bp->b_target->bt_file); + loff_t pos; + + ASSERT(xfs_buftarg_is_mem(bp->b_target)); + + pos = BBTOB(xfs_buf_daddr(bp)); + shmem_truncate_range(inode, pos, pos + BBTOB(bp->b_length) - 1); +} + +/* + * Finalize a buffer -- discard the backing page if it's stale, or run the + * write verifier to detect problems. + */ +int +xmbuf_finalize( + struct xfs_buf *bp) +{ + xfs_failaddr_t fa; + int error = 0; + + if (bp->b_flags & XBF_STALE) { + xmbuf_stale(bp); + return 0; + } + + /* + * Although this btree is ephemeral, validate the buffer structure so + * that we can detect memory corruption errors and software bugs. + */ + fa = bp->b_ops->verify_struct(bp); + if (fa) { + error = -EFSCORRUPTED; + xfs_verifier_error(bp, error, fa); + } + + return error; +} + +/* + * Detach this xmbuf buffer from the transaction by any means necessary. + * All buffers are direct-mapped, so they do not need bwrite. + */ +void +xmbuf_trans_bdetach( + struct xfs_trans *tp, + struct xfs_buf *bp) +{ + struct xfs_buf_log_item *bli = bp->b_log_item; + + ASSERT(bli != NULL); + + bli->bli_flags &= ~(XFS_BLI_DIRTY | XFS_BLI_ORDERED | + XFS_BLI_LOGGED | XFS_BLI_STALE); + clear_bit(XFS_LI_DIRTY, &bli->bli_item.li_flags); + + while (bp->b_log_item != NULL) + xfs_trans_bdetach(tp, bp); +} diff --git a/fs/xfs/xfs_buf_mem.h b/fs/xfs/xfs_buf_mem.h new file mode 100644 index 000000000000..eed4a7b63232 --- /dev/null +++ b/fs/xfs/xfs_buf_mem.h @@ -0,0 +1,34 @@ +// SPDX-License-Identifier: GPL-2.0-or-later +/* + * Copyright (c) 2023-2024 Oracle. All Rights Reserved. + * Author: Darrick J. Wong <djwong@kernel.org> + */ +#ifndef __XFS_BUF_MEM_H__ +#define __XFS_BUF_MEM_H__ + +#define XMBUF_BLOCKSIZE (PAGE_SIZE) +#define XMBUF_BLOCKSHIFT (PAGE_SHIFT) + +#ifdef CONFIG_XFS_MEMORY_BUFS +static inline bool xfs_buftarg_is_mem(const struct xfs_buftarg *btp) +{ + return btp->bt_bdev == NULL; +} + +int xmbuf_alloc(struct xfs_mount *mp, const char *descr, + struct xfs_buftarg **btpp); +void xmbuf_free(struct xfs_buftarg *btp); + +int xmbuf_map_page(struct xfs_buf *bp); +void xmbuf_unmap_page(struct xfs_buf *bp); +bool xmbuf_verify_daddr(struct xfs_buftarg *btp, xfs_daddr_t daddr); +void xmbuf_trans_bdetach(struct xfs_trans *tp, struct xfs_buf *bp); +int xmbuf_finalize(struct xfs_buf *bp); +#else +# define xfs_buftarg_is_mem(...) (false) +# define xmbuf_map_page(...) (-ENOMEM) +# define xmbuf_unmap_page(...) ((void)0) +# define xmbuf_verify_daddr(...) (false) +#endif /* CONFIG_XFS_MEMORY_BUFS */ + +#endif /* __XFS_BUF_MEM_H__ */ diff --git a/fs/xfs/xfs_dir2_readdir.c b/fs/xfs/xfs_dir2_readdir.c index cc6dc56f455d..cf9296b7e06f 100644 --- a/fs/xfs/xfs_dir2_readdir.c +++ b/fs/xfs/xfs_dir2_readdir.c @@ -118,8 +118,10 @@ xfs_dir2_sf_getdents( ctx->pos = off & 0x7fffffff; if (XFS_IS_CORRUPT(dp->i_mount, !xfs_dir2_namecheck(sfep->name, - sfep->namelen))) + sfep->namelen))) { + xfs_dirattr_mark_sick(dp, XFS_DATA_FORK); return -EFSCORRUPTED; + } if (!dir_emit(ctx, (char *)sfep->name, sfep->namelen, ino, xfs_dir3_get_dtype(mp, filetype))) return 0; @@ -211,6 +213,7 @@ xfs_dir2_block_getdents( if (XFS_IS_CORRUPT(dp->i_mount, !xfs_dir2_namecheck(dep->name, dep->namelen))) { + xfs_dirattr_mark_sick(dp, XFS_DATA_FORK); error = -EFSCORRUPTED; goto out_rele; } @@ -465,6 +468,7 @@ xfs_dir2_leaf_getdents( if (XFS_IS_CORRUPT(dp->i_mount, !xfs_dir2_namecheck(dep->name, dep->namelen))) { + xfs_dirattr_mark_sick(dp, XFS_DATA_FORK); error = -EFSCORRUPTED; break; } @@ -522,7 +526,7 @@ xfs_readdir( return -EIO; ASSERT(S_ISDIR(VFS_I(dp)->i_mode)); - ASSERT(xfs_isilocked(dp, XFS_IOLOCK_SHARED | XFS_IOLOCK_EXCL)); + xfs_assert_ilocked(dp, XFS_IOLOCK_SHARED | XFS_IOLOCK_EXCL); XFS_STATS_INC(dp->i_mount, xs_dir_getdents); args.dp = dp; diff --git a/fs/xfs/xfs_discard.c b/fs/xfs/xfs_discard.c index d5787991bb5b..268bb734dc0a 100644 --- a/fs/xfs/xfs_discard.c +++ b/fs/xfs/xfs_discard.c @@ -8,6 +8,7 @@ #include "xfs_format.h" #include "xfs_log_format.h" #include "xfs_trans_resv.h" +#include "xfs_trans.h" #include "xfs_mount.h" #include "xfs_btree.h" #include "xfs_alloc_btree.h" @@ -18,6 +19,7 @@ #include "xfs_trace.h" #include "xfs_log.h" #include "xfs_ag.h" +#include "xfs_health.h" /* * Notes on an efficient, low latency fstrim algorithm @@ -79,7 +81,7 @@ xfs_discard_endio_work( container_of(work, struct xfs_busy_extents, endio_work); xfs_extent_busy_clear(extents->mount, &extents->extent_list, false); - kmem_free(extents->owner); + kfree(extents->owner); } /* @@ -120,7 +122,7 @@ xfs_discard_extents( error = __blkdev_issue_discard(mp->m_ddev_targp->bt_bdev, XFS_AGB_TO_DADDR(mp, busyp->agno, busyp->bno), XFS_FSB_TO_BB(mp, busyp->length), - GFP_NOFS, &bio); + GFP_KERNEL, &bio); if (error && error != -EOPNOTSUPP) { xfs_info(mp, "discard failed for extent [0x%llx,%u], error %d", @@ -155,6 +157,7 @@ xfs_trim_gather_extents( uint64_t *blocks_trimmed) { struct xfs_mount *mp = pag->pag_mount; + struct xfs_trans *tp; struct xfs_btree_cur *cur; struct xfs_buf *agbp; int error; @@ -168,11 +171,15 @@ xfs_trim_gather_extents( */ xfs_log_force(mp, XFS_LOG_SYNC); - error = xfs_alloc_read_agf(pag, NULL, 0, &agbp); + error = xfs_trans_alloc_empty(mp, &tp); if (error) return error; - cur = xfs_allocbt_init_cursor(mp, NULL, agbp, pag, XFS_BTNUM_CNT); + error = xfs_alloc_read_agf(pag, tp, 0, &agbp); + if (error) + goto out_trans_cancel; + + cur = xfs_cntbt_init_cursor(mp, tp, agbp, pag); /* * Look up the extent length requested in the AGF and start with it. @@ -204,6 +211,7 @@ xfs_trim_gather_extents( if (error) break; if (XFS_IS_CORRUPT(mp, i != 1)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; break; } @@ -279,7 +287,8 @@ next_extent: xfs_extent_busy_clear(mp, &extents->extent_list, false); out_del_cursor: xfs_btree_del_cursor(cur, error); - xfs_buf_relse(agbp); +out_trans_cancel: + xfs_trans_cancel(tp); return error; } diff --git a/fs/xfs/xfs_dquot.c b/fs/xfs/xfs_dquot.c index b4f20d9c8f98..30d36596a2e4 100644 --- a/fs/xfs/xfs_dquot.c +++ b/fs/xfs/xfs_dquot.c @@ -24,6 +24,7 @@ #include "xfs_log.h" #include "xfs_bmap_btree.h" #include "xfs_error.h" +#include "xfs_health.h" /* * Lock order: @@ -44,6 +45,29 @@ static struct kmem_cache *xfs_dquot_cache; static struct lock_class_key xfs_dquot_group_class; static struct lock_class_key xfs_dquot_project_class; +/* Record observations of quota corruption with the health tracking system. */ +static void +xfs_dquot_mark_sick( + struct xfs_dquot *dqp) +{ + struct xfs_mount *mp = dqp->q_mount; + + switch (dqp->q_type) { + case XFS_DQTYPE_USER: + xfs_fs_mark_sick(mp, XFS_SICK_FS_UQUOTA); + break; + case XFS_DQTYPE_GROUP: + xfs_fs_mark_sick(mp, XFS_SICK_FS_GQUOTA); + break; + case XFS_DQTYPE_PROJ: + xfs_fs_mark_sick(mp, XFS_SICK_FS_PQUOTA); + break; + default: + ASSERT(0); + break; + } +} + /* * This is called to free all the memory associated with a dquot */ @@ -53,7 +77,7 @@ xfs_qm_dqdestroy( { ASSERT(list_empty(&dqp->q_lru)); - kmem_free(dqp->q_logitem.qli_item.li_lv_shadow); + kvfree(dqp->q_logitem.qli_item.li_lv_shadow); mutex_destroy(&dqp->q_qlock); XFS_STATS_DEC(dqp->q_mount, xs_qm_dquot); @@ -451,6 +475,8 @@ xfs_dquot_disk_read( error = xfs_trans_read_buf(mp, NULL, mp->m_ddev_targp, dqp->q_blkno, mp->m_quotainfo->qi_dqchunklen, 0, &bp, &xfs_dquot_buf_ops); + if (xfs_metadata_is_sick(error)) + xfs_dquot_mark_sick(dqp); if (error) { ASSERT(bp == NULL); return error; @@ -574,6 +600,7 @@ xfs_dquot_from_disk( "Metadata corruption detected at %pS, quota %u", __this_address, dqp->q_id); xfs_alert(bp->b_mount, "Unmount and run xfs_repair"); + xfs_dquot_mark_sick(dqp); return -EFSCORRUPTED; } @@ -950,7 +977,7 @@ xfs_qm_dqget_inode( if (error) return error; - ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL)); + xfs_assert_ilocked(ip, XFS_ILOCK_EXCL); ASSERT(xfs_inode_dquot(ip, type) == NULL); id = xfs_qm_id_for_quotatype(ip, type); @@ -1007,7 +1034,7 @@ restart: } dqret: - ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL)); + xfs_assert_ilocked(ip, XFS_ILOCK_EXCL); trace_xfs_dqget_miss(dqp); *O_dqpp = dqp; return 0; @@ -1238,6 +1265,8 @@ xfs_qm_dqflush( &bp, &xfs_dquot_buf_ops); if (error == -EAGAIN) goto out_unlock; + if (xfs_metadata_is_sick(error)) + xfs_dquot_mark_sick(dqp); if (error) goto out_abort; @@ -1246,6 +1275,7 @@ xfs_qm_dqflush( xfs_alert(mp, "corrupt dquot ID 0x%x in memory at %pS", dqp->q_id, fa); xfs_buf_relse(bp); + xfs_dquot_mark_sick(dqp); error = -EFSCORRUPTED; goto out_abort; } diff --git a/fs/xfs/xfs_error.c b/fs/xfs/xfs_error.c index b2cbbba3e15a..7ad0e92c6b5b 100644 --- a/fs/xfs/xfs_error.c +++ b/fs/xfs/xfs_error.c @@ -240,15 +240,15 @@ xfs_errortag_init( { int ret; - mp->m_errortag = kmem_zalloc(sizeof(unsigned int) * XFS_ERRTAG_MAX, - KM_MAYFAIL); + mp->m_errortag = kzalloc(sizeof(unsigned int) * XFS_ERRTAG_MAX, + GFP_KERNEL | __GFP_RETRY_MAYFAIL); if (!mp->m_errortag) return -ENOMEM; ret = xfs_sysfs_init(&mp->m_errortag_kobj, &xfs_errortag_ktype, &mp->m_kobj, "errortag"); if (ret) - kmem_free(mp->m_errortag); + kfree(mp->m_errortag); return ret; } @@ -257,7 +257,7 @@ xfs_errortag_del( struct xfs_mount *mp) { xfs_sysfs_del(&mp->m_errortag_kobj); - kmem_free(mp->m_errortag); + kfree(mp->m_errortag); } static bool diff --git a/fs/xfs/xfs_extent_busy.c b/fs/xfs/xfs_extent_busy.c index 2ccde32c9a9e..56cfa1498571 100644 --- a/fs/xfs/xfs_extent_busy.c +++ b/fs/xfs/xfs_extent_busy.c @@ -32,7 +32,8 @@ xfs_extent_busy_insert_list( struct rb_node **rbp; struct rb_node *parent = NULL; - new = kmem_zalloc(sizeof(struct xfs_extent_busy), 0); + new = kzalloc(sizeof(struct xfs_extent_busy), + GFP_KERNEL | __GFP_NOFAIL); new->agno = pag->pag_agno; new->bno = bno; new->length = len; @@ -530,7 +531,7 @@ xfs_extent_busy_clear_one( } list_del_init(&busyp->list); - kmem_free(busyp); + kfree(busyp); } static void diff --git a/fs/xfs/xfs_extfree_item.c b/fs/xfs/xfs_extfree_item.c index 1d1185fca6a5..8c382f092332 100644 --- a/fs/xfs/xfs_extfree_item.c +++ b/fs/xfs/xfs_extfree_item.c @@ -40,9 +40,9 @@ STATIC void xfs_efi_item_free( struct xfs_efi_log_item *efip) { - kmem_free(efip->efi_item.li_lv_shadow); + kvfree(efip->efi_item.li_lv_shadow); if (efip->efi_format.efi_nextents > XFS_EFI_MAX_FAST_EXTENTS) - kmem_free(efip); + kfree(efip); else kmem_cache_free(xfs_efi_cache, efip); } @@ -229,9 +229,9 @@ static inline struct xfs_efd_log_item *EFD_ITEM(struct xfs_log_item *lip) STATIC void xfs_efd_item_free(struct xfs_efd_log_item *efdp) { - kmem_free(efdp->efd_item.li_lv_shadow); + kvfree(efdp->efd_item.li_lv_shadow); if (efdp->efd_format.efd_nextents > XFS_EFD_MAX_FAST_EXTENTS) - kmem_free(efdp); + kfree(efdp); else kmem_cache_free(xfs_efd_cache, efdp); } diff --git a/fs/xfs/xfs_file.c b/fs/xfs/xfs_file.c index e33e5e13b95f..632653e00906 100644 --- a/fs/xfs/xfs_file.c +++ b/fs/xfs/xfs_file.c @@ -879,7 +879,7 @@ xfs_break_dax_layouts( { struct page *page; - ASSERT(xfs_isilocked(XFS_I(inode), XFS_MMAPLOCK_EXCL)); + xfs_assert_ilocked(XFS_I(inode), XFS_MMAPLOCK_EXCL); page = dax_layout_busy_page(inode->i_mapping); if (!page) @@ -900,7 +900,7 @@ xfs_break_layouts( bool retry; int error; - ASSERT(xfs_isilocked(XFS_I(inode), XFS_IOLOCK_SHARED|XFS_IOLOCK_EXCL)); + xfs_assert_ilocked(XFS_I(inode), XFS_IOLOCK_SHARED | XFS_IOLOCK_EXCL); do { retry = false; diff --git a/fs/xfs/xfs_filestream.c b/fs/xfs/xfs_filestream.c index 2fc98d313708..e3aaa0555597 100644 --- a/fs/xfs/xfs_filestream.c +++ b/fs/xfs/xfs_filestream.c @@ -44,7 +44,7 @@ xfs_fstrm_free_func( atomic_dec(&pag->pagf_fstrms); xfs_perag_rele(pag); - kmem_free(item); + kfree(item); } /* @@ -313,7 +313,7 @@ xfs_filestream_create_association( * we return a referenced AG, the allocation can still go ahead just * fine. */ - item = kmem_alloc(sizeof(*item), KM_MAYFAIL); + item = kmalloc(sizeof(*item), GFP_KERNEL | __GFP_RETRY_MAYFAIL); if (!item) goto out_put_fstrms; @@ -326,7 +326,7 @@ xfs_filestream_create_association( out_free_item: xfs_perag_rele(item->pag); - kmem_free(item); + kfree(item); out_put_fstrms: atomic_dec(&args->pag->pagf_fstrms); return 0; diff --git a/fs/xfs/xfs_fsmap.c b/fs/xfs/xfs_fsmap.c index 5a72217f5feb..de59eec74765 100644 --- a/fs/xfs/xfs_fsmap.c +++ b/fs/xfs/xfs_fsmap.c @@ -763,8 +763,8 @@ xfs_getfsmap_datadev_bnobt_query( return xfs_getfsmap_datadev_bnobt_helper(*curpp, &key[1], info); /* Allocate cursor for this AG and query_range it. */ - *curpp = xfs_allocbt_init_cursor(tp->t_mountp, tp, info->agf_bp, - info->pag, XFS_BTNUM_BNO); + *curpp = xfs_bnobt_init_cursor(tp->t_mountp, tp, info->agf_bp, + info->pag); key->ar_startblock = info->low.rm_startblock; key[1].ar_startblock = info->high.rm_startblock; return xfs_alloc_query_range(*curpp, key, &key[1], diff --git a/fs/xfs/xfs_health.c b/fs/xfs/xfs_health.c index 9a57afee9338..b39f959146bc 100644 --- a/fs/xfs/xfs_health.c +++ b/fs/xfs/xfs_health.c @@ -14,6 +14,10 @@ #include "xfs_trace.h" #include "xfs_health.h" #include "xfs_ag.h" +#include "xfs_btree.h" +#include "xfs_da_format.h" +#include "xfs_da_btree.h" +#include "xfs_quota_defs.h" /* * Warn about metadata corruption that we detected but haven't fixed, and @@ -93,11 +97,25 @@ xfs_fs_mark_sick( struct xfs_mount *mp, unsigned int mask) { - ASSERT(!(mask & ~XFS_SICK_FS_PRIMARY)); + ASSERT(!(mask & ~XFS_SICK_FS_ALL)); trace_xfs_fs_mark_sick(mp, mask); spin_lock(&mp->m_sb_lock); mp->m_fs_sick |= mask; + spin_unlock(&mp->m_sb_lock); +} + +/* Mark per-fs metadata as having been checked and found unhealthy by fsck. */ +void +xfs_fs_mark_corrupt( + struct xfs_mount *mp, + unsigned int mask) +{ + ASSERT(!(mask & ~XFS_SICK_FS_ALL)); + trace_xfs_fs_mark_corrupt(mp, mask); + + spin_lock(&mp->m_sb_lock); + mp->m_fs_sick |= mask; mp->m_fs_checked |= mask; spin_unlock(&mp->m_sb_lock); } @@ -108,11 +126,13 @@ xfs_fs_mark_healthy( struct xfs_mount *mp, unsigned int mask) { - ASSERT(!(mask & ~XFS_SICK_FS_PRIMARY)); + ASSERT(!(mask & ~XFS_SICK_FS_ALL)); trace_xfs_fs_mark_healthy(mp, mask); spin_lock(&mp->m_sb_lock); mp->m_fs_sick &= ~mask; + if (!(mp->m_fs_sick & XFS_SICK_FS_PRIMARY)) + mp->m_fs_sick &= ~XFS_SICK_FS_SECONDARY; mp->m_fs_checked |= mask; spin_unlock(&mp->m_sb_lock); } @@ -136,11 +156,25 @@ xfs_rt_mark_sick( struct xfs_mount *mp, unsigned int mask) { - ASSERT(!(mask & ~XFS_SICK_RT_PRIMARY)); + ASSERT(!(mask & ~XFS_SICK_RT_ALL)); trace_xfs_rt_mark_sick(mp, mask); spin_lock(&mp->m_sb_lock); mp->m_rt_sick |= mask; + spin_unlock(&mp->m_sb_lock); +} + +/* Mark realtime metadata as having been checked and found unhealthy by fsck. */ +void +xfs_rt_mark_corrupt( + struct xfs_mount *mp, + unsigned int mask) +{ + ASSERT(!(mask & ~XFS_SICK_RT_ALL)); + trace_xfs_rt_mark_corrupt(mp, mask); + + spin_lock(&mp->m_sb_lock); + mp->m_rt_sick |= mask; mp->m_rt_checked |= mask; spin_unlock(&mp->m_sb_lock); } @@ -151,11 +185,13 @@ xfs_rt_mark_healthy( struct xfs_mount *mp, unsigned int mask) { - ASSERT(!(mask & ~XFS_SICK_RT_PRIMARY)); + ASSERT(!(mask & ~XFS_SICK_RT_ALL)); trace_xfs_rt_mark_healthy(mp, mask); spin_lock(&mp->m_sb_lock); mp->m_rt_sick &= ~mask; + if (!(mp->m_rt_sick & XFS_SICK_RT_PRIMARY)) + mp->m_rt_sick &= ~XFS_SICK_RT_SECONDARY; mp->m_rt_checked |= mask; spin_unlock(&mp->m_sb_lock); } @@ -173,17 +209,48 @@ xfs_rt_measure_sickness( spin_unlock(&mp->m_sb_lock); } +/* Mark unhealthy per-ag metadata given a raw AG number. */ +void +xfs_agno_mark_sick( + struct xfs_mount *mp, + xfs_agnumber_t agno, + unsigned int mask) +{ + struct xfs_perag *pag = xfs_perag_get(mp, agno); + + /* per-ag structure not set up yet? */ + if (!pag) + return; + + xfs_ag_mark_sick(pag, mask); + xfs_perag_put(pag); +} + /* Mark unhealthy per-ag metadata. */ void xfs_ag_mark_sick( struct xfs_perag *pag, unsigned int mask) { - ASSERT(!(mask & ~XFS_SICK_AG_PRIMARY)); + ASSERT(!(mask & ~XFS_SICK_AG_ALL)); trace_xfs_ag_mark_sick(pag->pag_mount, pag->pag_agno, mask); spin_lock(&pag->pag_state_lock); pag->pag_sick |= mask; + spin_unlock(&pag->pag_state_lock); +} + +/* Mark per-ag metadata as having been checked and found unhealthy by fsck. */ +void +xfs_ag_mark_corrupt( + struct xfs_perag *pag, + unsigned int mask) +{ + ASSERT(!(mask & ~XFS_SICK_AG_ALL)); + trace_xfs_ag_mark_corrupt(pag->pag_mount, pag->pag_agno, mask); + + spin_lock(&pag->pag_state_lock); + pag->pag_sick |= mask; pag->pag_checked |= mask; spin_unlock(&pag->pag_state_lock); } @@ -194,11 +261,13 @@ xfs_ag_mark_healthy( struct xfs_perag *pag, unsigned int mask) { - ASSERT(!(mask & ~XFS_SICK_AG_PRIMARY)); + ASSERT(!(mask & ~XFS_SICK_AG_ALL)); trace_xfs_ag_mark_healthy(pag->pag_mount, pag->pag_agno, mask); spin_lock(&pag->pag_state_lock); pag->pag_sick &= ~mask; + if (!(pag->pag_sick & XFS_SICK_AG_PRIMARY)) + pag->pag_sick &= ~XFS_SICK_AG_SECONDARY; pag->pag_checked |= mask; spin_unlock(&pag->pag_state_lock); } @@ -222,11 +291,34 @@ xfs_inode_mark_sick( struct xfs_inode *ip, unsigned int mask) { - ASSERT(!(mask & ~(XFS_SICK_INO_PRIMARY | XFS_SICK_INO_ZAPPED))); + ASSERT(!(mask & ~XFS_SICK_INO_ALL)); trace_xfs_inode_mark_sick(ip, mask); spin_lock(&ip->i_flags_lock); ip->i_sick |= mask; + spin_unlock(&ip->i_flags_lock); + + /* + * Keep this inode around so we don't lose the sickness report. Scrub + * grabs inodes with DONTCACHE assuming that most inode are ok, which + * is not the case here. + */ + spin_lock(&VFS_I(ip)->i_lock); + VFS_I(ip)->i_state &= ~I_DONTCACHE; + spin_unlock(&VFS_I(ip)->i_lock); +} + +/* Mark inode metadata as having been checked and found unhealthy by fsck. */ +void +xfs_inode_mark_corrupt( + struct xfs_inode *ip, + unsigned int mask) +{ + ASSERT(!(mask & ~XFS_SICK_INO_ALL)); + trace_xfs_inode_mark_corrupt(ip, mask); + + spin_lock(&ip->i_flags_lock); + ip->i_sick |= mask; ip->i_checked |= mask; spin_unlock(&ip->i_flags_lock); @@ -246,11 +338,13 @@ xfs_inode_mark_healthy( struct xfs_inode *ip, unsigned int mask) { - ASSERT(!(mask & ~(XFS_SICK_INO_PRIMARY | XFS_SICK_INO_ZAPPED))); + ASSERT(!(mask & ~XFS_SICK_INO_ALL)); trace_xfs_inode_mark_healthy(ip, mask); spin_lock(&ip->i_flags_lock); ip->i_sick &= ~mask; + if (!(ip->i_sick & XFS_SICK_INO_PRIMARY)) + ip->i_sick &= ~XFS_SICK_INO_SECONDARY; ip->i_checked |= mask; spin_unlock(&ip->i_flags_lock); } @@ -280,6 +374,8 @@ static const struct ioctl_sick_map fs_map[] = { { XFS_SICK_FS_UQUOTA, XFS_FSOP_GEOM_SICK_UQUOTA }, { XFS_SICK_FS_GQUOTA, XFS_FSOP_GEOM_SICK_GQUOTA }, { XFS_SICK_FS_PQUOTA, XFS_FSOP_GEOM_SICK_PQUOTA }, + { XFS_SICK_FS_QUOTACHECK, XFS_FSOP_GEOM_SICK_QUOTACHECK }, + { XFS_SICK_FS_NLINKS, XFS_FSOP_GEOM_SICK_NLINKS }, { 0, 0 }, }; @@ -335,6 +431,7 @@ static const struct ioctl_sick_map ag_map[] = { { XFS_SICK_AG_FINOBT, XFS_AG_GEOM_SICK_FINOBT }, { XFS_SICK_AG_RMAPBT, XFS_AG_GEOM_SICK_RMAPBT }, { XFS_SICK_AG_REFCNTBT, XFS_AG_GEOM_SICK_REFCNTBT }, + { XFS_SICK_AG_INODES, XFS_AG_GEOM_SICK_INODES }, { 0, 0 }, }; @@ -397,3 +494,92 @@ xfs_bulkstat_health( bs->bs_sick |= m->ioctl_mask; } } + +/* Mark a block mapping sick. */ +void +xfs_bmap_mark_sick( + struct xfs_inode *ip, + int whichfork) +{ + unsigned int mask; + + switch (whichfork) { + case XFS_DATA_FORK: + mask = XFS_SICK_INO_BMBTD; + break; + case XFS_ATTR_FORK: + mask = XFS_SICK_INO_BMBTA; + break; + case XFS_COW_FORK: + mask = XFS_SICK_INO_BMBTC; + break; + default: + ASSERT(0); + return; + } + + xfs_inode_mark_sick(ip, mask); +} + +/* Record observations of btree corruption with the health tracking system. */ +void +xfs_btree_mark_sick( + struct xfs_btree_cur *cur) +{ + switch (cur->bc_ops->type) { + case XFS_BTREE_TYPE_MEM: + /* no health state tracking for ephemeral btrees */ + return; + case XFS_BTREE_TYPE_AG: + ASSERT(cur->bc_ops->sick_mask); + xfs_ag_mark_sick(cur->bc_ag.pag, cur->bc_ops->sick_mask); + return; + case XFS_BTREE_TYPE_INODE: + if (xfs_btree_is_bmap(cur->bc_ops)) { + xfs_bmap_mark_sick(cur->bc_ino.ip, + cur->bc_ino.whichfork); + return; + } + fallthrough; + default: + ASSERT(0); + return; + } +} + +/* + * Record observations of dir/attr btree corruption with the health tracking + * system. + */ +void +xfs_dirattr_mark_sick( + struct xfs_inode *ip, + int whichfork) +{ + unsigned int mask; + + switch (whichfork) { + case XFS_DATA_FORK: + mask = XFS_SICK_INO_DIR; + break; + case XFS_ATTR_FORK: + mask = XFS_SICK_INO_XATTR; + break; + default: + ASSERT(0); + return; + } + + xfs_inode_mark_sick(ip, mask); +} + +/* + * Record observations of dir/attr btree corruption with the health tracking + * system. + */ +void +xfs_da_mark_sick( + struct xfs_da_args *args) +{ + xfs_dirattr_mark_sick(args->dp, args->whichfork); +} diff --git a/fs/xfs/xfs_hooks.c b/fs/xfs/xfs_hooks.c new file mode 100644 index 000000000000..a58d1de2d37d --- /dev/null +++ b/fs/xfs/xfs_hooks.c @@ -0,0 +1,52 @@ +// SPDX-License-Identifier: GPL-2.0-or-later +/* + * Copyright (c) 2022-2024 Oracle. All Rights Reserved. + * Author: Darrick J. Wong <djwong@kernel.org> + */ +#include "xfs.h" +#include "xfs_fs.h" +#include "xfs_shared.h" +#include "xfs_format.h" +#include "xfs_trans_resv.h" +#include "xfs_mount.h" +#include "xfs_ag.h" +#include "xfs_trace.h" + +/* Initialize a notifier chain. */ +void +xfs_hooks_init( + struct xfs_hooks *chain) +{ + BLOCKING_INIT_NOTIFIER_HEAD(&chain->head); +} + +/* Make it so a function gets called whenever we hit a certain hook point. */ +int +xfs_hooks_add( + struct xfs_hooks *chain, + struct xfs_hook *hook) +{ + ASSERT(hook->nb.notifier_call != NULL); + BUILD_BUG_ON(offsetof(struct xfs_hook, nb) != 0); + + return blocking_notifier_chain_register(&chain->head, &hook->nb); +} + +/* Remove a previously installed hook. */ +void +xfs_hooks_del( + struct xfs_hooks *chain, + struct xfs_hook *hook) +{ + blocking_notifier_chain_unregister(&chain->head, &hook->nb); +} + +/* Call a hook. Returns the NOTIFY_* value returned by the last hook. */ +int +xfs_hooks_call( + struct xfs_hooks *chain, + unsigned long val, + void *priv) +{ + return blocking_notifier_call_chain(&chain->head, val, priv); +} diff --git a/fs/xfs/xfs_hooks.h b/fs/xfs/xfs_hooks.h new file mode 100644 index 000000000000..60b8a5831536 --- /dev/null +++ b/fs/xfs/xfs_hooks.h @@ -0,0 +1,65 @@ +// SPDX-License-Identifier: GPL-2.0-or-later +/* + * Copyright (c) 2022-2024 Oracle. All Rights Reserved. + * Author: Darrick J. Wong <djwong@kernel.org> + */ +#ifndef XFS_HOOKS_H_ +#define XFS_HOOKS_H_ + +#ifdef CONFIG_XFS_LIVE_HOOKS +struct xfs_hooks { + struct blocking_notifier_head head; +}; + +/* + * If jump labels are enabled in Kconfig, the static key uses nop sleds and + * code patching to eliminate the overhead of taking the rwsem in + * blocking_notifier_call_chain when there are no hooks configured. If not, + * the static key per-call overhead is an atomic read. Most arches that can + * handle XFS also support jump labels. + * + * Note: Patching the kernel code requires taking the cpu hotplug lock. Other + * parts of the kernel allocate memory with that lock held, which means that + * XFS callers cannot hold any locks that might be used by memory reclaim or + * writeback when calling the static_branch_{inc,dec} functions. + */ +# define DEFINE_STATIC_XFS_HOOK_SWITCH(name) \ + static DEFINE_STATIC_KEY_FALSE(name) +# define xfs_hooks_switch_on(name) static_branch_inc(name) +# define xfs_hooks_switch_off(name) static_branch_dec(name) +# define xfs_hooks_switched_on(name) static_branch_unlikely(name) + +struct xfs_hook { + /* This must come at the start of the structure. */ + struct notifier_block nb; +}; + +typedef int (*xfs_hook_fn_t)(struct xfs_hook *hook, unsigned long action, + void *data); + +void xfs_hooks_init(struct xfs_hooks *chain); +int xfs_hooks_add(struct xfs_hooks *chain, struct xfs_hook *hook); +void xfs_hooks_del(struct xfs_hooks *chain, struct xfs_hook *hook); +int xfs_hooks_call(struct xfs_hooks *chain, unsigned long action, + void *priv); + +static inline void xfs_hook_setup(struct xfs_hook *hook, notifier_fn_t fn) +{ + hook->nb.notifier_call = fn; + hook->nb.priority = 0; +} + +#else + +struct xfs_hooks { /* empty */ }; + +# define DEFINE_STATIC_XFS_HOOK_SWITCH(name) +# define xfs_hooks_switch_on(name) ((void)0) +# define xfs_hooks_switch_off(name) ((void)0) +# define xfs_hooks_switched_on(name) (false) + +# define xfs_hooks_init(chain) ((void)0) +# define xfs_hooks_call(chain, val, priv) (NOTIFY_DONE) +#endif + +#endif /* XFS_HOOKS_H_ */ diff --git a/fs/xfs/xfs_icache.c b/fs/xfs/xfs_icache.c index dba514a2c84d..e64265bc0b33 100644 --- a/fs/xfs/xfs_icache.c +++ b/fs/xfs/xfs_icache.c @@ -24,6 +24,7 @@ #include "xfs_ialloc.h" #include "xfs_ag.h" #include "xfs_log_priv.h" +#include "xfs_health.h" #include <linux/iversion.h> @@ -415,6 +416,9 @@ xfs_iget_check_free_state( xfs_warn(ip->i_mount, "Corruption detected! Free inode 0x%llx not marked free! (mode 0x%x)", ip->i_ino, VFS_I(ip)->i_mode); + xfs_agno_mark_sick(ip->i_mount, + XFS_INO_TO_AGNO(ip->i_mount, ip->i_ino), + XFS_SICK_AG_INOBT); return -EFSCORRUPTED; } @@ -422,6 +426,9 @@ xfs_iget_check_free_state( xfs_warn(ip->i_mount, "Corruption detected! Free inode 0x%llx has blocks allocated!", ip->i_ino); + xfs_agno_mark_sick(ip->i_mount, + XFS_INO_TO_AGNO(ip->i_mount, ip->i_ino), + XFS_SICK_AG_INOBT); return -EFSCORRUPTED; } return 0; @@ -640,6 +647,8 @@ xfs_iget_cache_miss( xfs_buf_offset(bp, ip->i_imap.im_boffset)); if (!error) xfs_buf_set_ref(bp, XFS_INO_REF); + else + xfs_inode_mark_sick(ip, XFS_SICK_INO_CORE); xfs_trans_brelse(tp, bp); if (error) @@ -659,10 +668,9 @@ xfs_iget_cache_miss( /* * Preload the radix tree so we can insert safely under the * write spinlock. Note that we cannot sleep inside the preload - * region. Since we can be called from transaction context, don't - * recurse into the file system. + * region. */ - if (radix_tree_preload(GFP_NOFS)) { + if (radix_tree_preload(GFP_KERNEL | __GFP_NOLOCKDEP)) { error = -EAGAIN; goto out_destroy; } diff --git a/fs/xfs/xfs_icreate_item.c b/fs/xfs/xfs_icreate_item.c index b05314d48176..4345db501714 100644 --- a/fs/xfs/xfs_icreate_item.c +++ b/fs/xfs/xfs_icreate_item.c @@ -63,7 +63,7 @@ STATIC void xfs_icreate_item_release( struct xfs_log_item *lip) { - kmem_free(ICR_ITEM(lip)->ic_item.li_lv_shadow); + kvfree(ICR_ITEM(lip)->ic_item.li_lv_shadow); kmem_cache_free(xfs_icreate_cache, ICR_ITEM(lip)); } diff --git a/fs/xfs/xfs_inode.c b/fs/xfs/xfs_inode.c index 1fd94958aa97..ea48774f6b76 100644 --- a/fs/xfs/xfs_inode.c +++ b/fs/xfs/xfs_inode.c @@ -203,9 +203,9 @@ xfs_ilock( } if (lock_flags & XFS_ILOCK_EXCL) - mrupdate_nested(&ip->i_lock, XFS_ILOCK_DEP(lock_flags)); + down_write_nested(&ip->i_lock, XFS_ILOCK_DEP(lock_flags)); else if (lock_flags & XFS_ILOCK_SHARED) - mraccess_nested(&ip->i_lock, XFS_ILOCK_DEP(lock_flags)); + down_read_nested(&ip->i_lock, XFS_ILOCK_DEP(lock_flags)); } /* @@ -246,10 +246,10 @@ xfs_ilock_nowait( } if (lock_flags & XFS_ILOCK_EXCL) { - if (!mrtryupdate(&ip->i_lock)) + if (!down_write_trylock(&ip->i_lock)) goto out_undo_mmaplock; } else if (lock_flags & XFS_ILOCK_SHARED) { - if (!mrtryaccess(&ip->i_lock)) + if (!down_read_trylock(&ip->i_lock)) goto out_undo_mmaplock; } return 1; @@ -298,9 +298,9 @@ xfs_iunlock( up_read(&VFS_I(ip)->i_mapping->invalidate_lock); if (lock_flags & XFS_ILOCK_EXCL) - mrunlock_excl(&ip->i_lock); + up_write(&ip->i_lock); else if (lock_flags & XFS_ILOCK_SHARED) - mrunlock_shared(&ip->i_lock); + up_read(&ip->i_lock); trace_xfs_iunlock(ip, lock_flags, _RET_IP_); } @@ -319,7 +319,7 @@ xfs_ilock_demote( ~(XFS_IOLOCK_EXCL|XFS_MMAPLOCK_EXCL|XFS_ILOCK_EXCL)) == 0); if (lock_flags & XFS_ILOCK_EXCL) - mrdemote(&ip->i_lock); + downgrade_write(&ip->i_lock); if (lock_flags & XFS_MMAPLOCK_EXCL) downgrade_write(&VFS_I(ip)->i_mapping->invalidate_lock); if (lock_flags & XFS_IOLOCK_EXCL) @@ -328,52 +328,30 @@ xfs_ilock_demote( trace_xfs_ilock_demote(ip, lock_flags, _RET_IP_); } -#if defined(DEBUG) || defined(XFS_WARN) -static inline bool -__xfs_rwsem_islocked( - struct rw_semaphore *rwsem, - bool shared) -{ - if (!debug_locks) - return rwsem_is_locked(rwsem); - - if (!shared) - return lockdep_is_held_type(rwsem, 0); - - /* - * We are checking that the lock is held at least in shared - * mode but don't care that it might be held exclusively - * (i.e. shared | excl). Hence we check if the lock is held - * in any mode rather than an explicit shared mode. - */ - return lockdep_is_held_type(rwsem, -1); -} - -bool -xfs_isilocked( +void +xfs_assert_ilocked( struct xfs_inode *ip, uint lock_flags) { - if (lock_flags & (XFS_ILOCK_EXCL|XFS_ILOCK_SHARED)) { - if (!(lock_flags & XFS_ILOCK_SHARED)) - return !!ip->i_lock.mr_writer; - return rwsem_is_locked(&ip->i_lock.mr_lock); - } - - if (lock_flags & (XFS_MMAPLOCK_EXCL|XFS_MMAPLOCK_SHARED)) { - return __xfs_rwsem_islocked(&VFS_I(ip)->i_mapping->invalidate_lock, - (lock_flags & XFS_MMAPLOCK_SHARED)); - } + /* + * Sometimes we assert the ILOCK is held exclusively, but we're in + * a workqueue, so lockdep doesn't know we're the owner. + */ + if (lock_flags & XFS_ILOCK_SHARED) + rwsem_assert_held(&ip->i_lock); + else if (lock_flags & XFS_ILOCK_EXCL) + rwsem_assert_held_write_nolockdep(&ip->i_lock); - if (lock_flags & (XFS_IOLOCK_EXCL | XFS_IOLOCK_SHARED)) { - return __xfs_rwsem_islocked(&VFS_I(ip)->i_rwsem, - (lock_flags & XFS_IOLOCK_SHARED)); - } + if (lock_flags & XFS_MMAPLOCK_SHARED) + rwsem_assert_held(&VFS_I(ip)->i_mapping->invalidate_lock); + else if (lock_flags & XFS_MMAPLOCK_EXCL) + rwsem_assert_held_write(&VFS_I(ip)->i_mapping->invalidate_lock); - ASSERT(0); - return false; + if (lock_flags & XFS_IOLOCK_SHARED) + rwsem_assert_held(&VFS_I(ip)->i_rwsem); + else if (lock_flags & XFS_IOLOCK_EXCL) + rwsem_assert_held_write(&VFS_I(ip)->i_rwsem); } -#endif /* * xfs_lockdep_subclass_ok() is only used in an ASSERT, so is only called when @@ -671,7 +649,7 @@ xfs_lookup( out_free_name: if (ci_name) - kmem_free(ci_name->name); + kfree(ci_name->name); out_unlock: *ipp = NULL; return error; @@ -802,6 +780,8 @@ xfs_init_new_inode( */ if ((pip && ino == pip->i_ino) || !xfs_verify_dir_ino(mp, ino)) { xfs_alert(mp, "Allocated a known in-use inode 0x%llx!", ino); + xfs_agno_mark_sick(mp, XFS_INO_TO_AGNO(mp, ino), + XFS_SICK_AG_INOBT); return -EFSCORRUPTED; } @@ -947,6 +927,81 @@ xfs_bumplink( xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE); } +#ifdef CONFIG_XFS_LIVE_HOOKS +/* + * Use a static key here to reduce the overhead of directory live update hooks. + * If the compiler supports jump labels, the static branch will be replaced by + * a nop sled when there are no hook users. Online fsck is currently the only + * caller, so this is a reasonable tradeoff. + * + * Note: Patching the kernel code requires taking the cpu hotplug lock. Other + * parts of the kernel allocate memory with that lock held, which means that + * XFS callers cannot hold any locks that might be used by memory reclaim or + * writeback when calling the static_branch_{inc,dec} functions. + */ +DEFINE_STATIC_XFS_HOOK_SWITCH(xfs_dir_hooks_switch); + +void +xfs_dir_hook_disable(void) +{ + xfs_hooks_switch_off(&xfs_dir_hooks_switch); +} + +void +xfs_dir_hook_enable(void) +{ + xfs_hooks_switch_on(&xfs_dir_hooks_switch); +} + +/* Call hooks for a directory update relating to a child dirent update. */ +inline void +xfs_dir_update_hook( + struct xfs_inode *dp, + struct xfs_inode *ip, + int delta, + const struct xfs_name *name) +{ + if (xfs_hooks_switched_on(&xfs_dir_hooks_switch)) { + struct xfs_dir_update_params p = { + .dp = dp, + .ip = ip, + .delta = delta, + .name = name, + }; + struct xfs_mount *mp = ip->i_mount; + + xfs_hooks_call(&mp->m_dir_update_hooks, 0, &p); + } +} + +/* Call the specified function during a directory update. */ +int +xfs_dir_hook_add( + struct xfs_mount *mp, + struct xfs_dir_hook *hook) +{ + return xfs_hooks_add(&mp->m_dir_update_hooks, &hook->dirent_hook); +} + +/* Stop calling the specified function during a directory update. */ +void +xfs_dir_hook_del( + struct xfs_mount *mp, + struct xfs_dir_hook *hook) +{ + xfs_hooks_del(&mp->m_dir_update_hooks, &hook->dirent_hook); +} + +/* Configure directory update hook functions. */ +void +xfs_dir_hook_setup( + struct xfs_dir_hook *hook, + notifier_fn_t mod_fn) +{ + xfs_hook_setup(&hook->dirent_hook, mod_fn); +} +#endif /* CONFIG_XFS_LIVE_HOOKS */ + int xfs_create( struct mnt_idmap *idmap, @@ -1058,6 +1113,12 @@ xfs_create( } /* + * Create ip with a reference from dp, and add '.' and '..' references + * if it's a directory. + */ + xfs_dir_update_hook(dp, ip, 1, name); + + /* * If this is a synchronous mount, make sure that the * create transaction goes to disk before returning to * the user. @@ -1271,6 +1332,7 @@ xfs_link( xfs_trans_log_inode(tp, tdp, XFS_ILOG_CORE); xfs_bumplink(tp, sip); + xfs_dir_update_hook(tdp, sip, 1, target_name); /* * If this is a synchronous mount, make sure that the @@ -1342,9 +1404,9 @@ xfs_itruncate_extents_flags( xfs_fileoff_t first_unmap_block; int error = 0; - ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL)); - ASSERT(!atomic_read(&VFS_I(ip)->i_count) || - xfs_isilocked(ip, XFS_IOLOCK_EXCL)); + xfs_assert_ilocked(ip, XFS_ILOCK_EXCL); + if (atomic_read(&VFS_I(ip)->i_count)) + xfs_assert_ilocked(ip, XFS_IOLOCK_EXCL); ASSERT(new_size <= XFS_ISIZE(ip)); ASSERT(tp->t_flags & XFS_TRANS_PERM_LOG_RES); ASSERT(ip->i_itemp != NULL); @@ -1596,7 +1658,7 @@ xfs_inactive_ifree( xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL); error = xfs_ifree(tp, ip); - ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL)); + xfs_assert_ilocked(ip, XFS_ILOCK_EXCL); if (error) { /* * If we fail to free the inode, shut down. The cancel @@ -1677,6 +1739,39 @@ xfs_inode_needs_inactive( } /* + * Save health status somewhere, if we're dumping an inode with uncorrected + * errors and online repair isn't running. + */ +static inline void +xfs_inactive_health( + struct xfs_inode *ip) +{ + struct xfs_mount *mp = ip->i_mount; + struct xfs_perag *pag; + unsigned int sick; + unsigned int checked; + + xfs_inode_measure_sickness(ip, &sick, &checked); + if (!sick) + return; + + trace_xfs_inode_unfixed_corruption(ip, sick); + + if (sick & XFS_SICK_INO_FORGET) + return; + + pag = xfs_perag_get(mp, XFS_INO_TO_AGNO(mp, ip->i_ino)); + if (!pag) { + /* There had better still be a perag structure! */ + ASSERT(0); + return; + } + + xfs_ag_mark_sick(pag, XFS_SICK_AG_INODES); + xfs_perag_put(pag); +} + +/* * xfs_inactive * * This is called when the vnode reference count for the vnode @@ -1704,6 +1799,8 @@ xfs_inactive( mp = ip->i_mount; ASSERT(!xfs_iflags_test(ip, XFS_IRECOVERY)); + xfs_inactive_health(ip); + /* * If this is a read-only mount, don't do this (would generate I/O) * unless we're in log recovery and cleaning the iunlinked list. @@ -1910,6 +2007,7 @@ xfs_iunlink_update_bucket( */ if (old_value == new_agino) { xfs_buf_mark_corrupt(agibp); + xfs_ag_mark_sick(pag, XFS_SICK_AG_AGI); return -EFSCORRUPTED; } @@ -1959,11 +2057,14 @@ xfs_iunlink_reload_next( */ ino = XFS_AGINO_TO_INO(mp, pag->pag_agno, next_agino); error = xfs_iget(mp, tp, ino, XFS_IGET_UNTRUSTED, 0, &next_ip); - if (error) + if (error) { + xfs_ag_mark_sick(pag, XFS_SICK_AG_AGI); return error; + } /* If this is not an unlinked inode, something is very wrong. */ if (VFS_I(next_ip)->i_nlink != 0) { + xfs_ag_mark_sick(pag, XFS_SICK_AG_AGI); error = -EFSCORRUPTED; goto rele; } @@ -2001,6 +2102,7 @@ xfs_iunlink_insert_inode( if (next_agino == agino || !xfs_verify_agino_or_null(pag, next_agino)) { xfs_buf_mark_corrupt(agibp); + xfs_ag_mark_sick(pag, XFS_SICK_AG_AGI); return -EFSCORRUPTED; } @@ -2088,6 +2190,7 @@ xfs_iunlink_remove_inode( if (!xfs_verify_agino(pag, head_agino)) { XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, mp, agi, sizeof(*agi)); + xfs_ag_mark_sick(pag, XFS_SICK_AG_AGI); return -EFSCORRUPTED; } @@ -2116,8 +2219,10 @@ xfs_iunlink_remove_inode( struct xfs_inode *prev_ip; prev_ip = xfs_iunlink_lookup(pag, ip->i_prev_unlinked); - if (!prev_ip) + if (!prev_ip) { + xfs_inode_mark_sick(ip, XFS_SICK_INO_CORE); return -EFSCORRUPTED; + } error = xfs_iunlink_log_inode(tp, prev_ip, pag, ip->i_next_unlinked); @@ -2350,7 +2455,7 @@ xfs_ifree( struct xfs_inode_log_item *iip = ip->i_itemp; int error; - ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL)); + xfs_assert_ilocked(ip, XFS_ILOCK_EXCL); ASSERT(VFS_I(ip)->i_nlink == 0); ASSERT(ip->i_df.if_nextents == 0); ASSERT(ip->i_disk_size == 0 || !S_ISREG(VFS_I(ip)->i_mode)); @@ -2378,7 +2483,7 @@ xfs_ifree( * already been freed by xfs_attr_inactive. */ if (ip->i_df.if_format == XFS_DINODE_FMT_LOCAL) { - kmem_free(ip->i_df.if_data); + kfree(ip->i_df.if_data); ip->i_df.if_data = NULL; ip->i_df.if_bytes = 0; } @@ -2419,7 +2524,7 @@ static void xfs_iunpin( struct xfs_inode *ip) { - ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL|XFS_ILOCK_SHARED)); + xfs_assert_ilocked(ip, XFS_ILOCK_EXCL | XFS_ILOCK_SHARED); trace_xfs_inode_unpin_nowait(ip, _RET_IP_); @@ -2585,6 +2690,12 @@ xfs_remove( } /* + * Drop the link from dp to ip, and if ip was a directory, remove the + * '.' and '..' references since we freed the directory. + */ + xfs_dir_update_hook(dp, ip, -1, name); + + /* * If this is a synchronous mount, make sure that the * remove transaction goes to disk before returning to * the user. @@ -2774,6 +2885,20 @@ xfs_cross_rename( } xfs_trans_ichgtime(tp, dp1, XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG); xfs_trans_log_inode(tp, dp1, XFS_ILOG_CORE); + + /* + * Inform our hook clients that we've finished an exchange operation as + * follows: removed the source and target files from their directories; + * added the target to the source directory; and added the source to + * the target directory. All inodes are locked, so it's ok to model a + * rename this way so long as we say we deleted entries before we add + * new ones. + */ + xfs_dir_update_hook(dp1, ip1, -1, name1); + xfs_dir_update_hook(dp2, ip2, -1, name2); + xfs_dir_update_hook(dp1, ip2, 1, name1); + xfs_dir_update_hook(dp2, ip1, 1, name2); + return xfs_finish_rename(tp); out_trans_abort: @@ -3157,6 +3282,21 @@ retry: if (new_parent) xfs_trans_log_inode(tp, target_dp, XFS_ILOG_CORE); + /* + * Inform our hook clients that we've finished a rename operation as + * follows: removed the source and target files from their directories; + * that we've added the source to the target directory; and finally + * that we've added the whiteout, if there was one. All inodes are + * locked, so it's ok to model a rename this way so long as we say we + * deleted entries before we add new ones. + */ + if (target_ip) + xfs_dir_update_hook(target_dp, target_ip, -1, target_name); + xfs_dir_update_hook(src_dp, src_ip, -1, src_name); + xfs_dir_update_hook(target_dp, src_ip, 1, target_name); + if (wip) + xfs_dir_update_hook(src_dp, wip, 1, src_name); + error = xfs_finish_rename(tp); if (wip) xfs_irele(wip); @@ -3182,7 +3322,7 @@ xfs_iflush( struct xfs_mount *mp = ip->i_mount; int error; - ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL|XFS_ILOCK_SHARED)); + xfs_assert_ilocked(ip, XFS_ILOCK_EXCL | XFS_ILOCK_SHARED); ASSERT(xfs_iflags_test(ip, XFS_IFLUSHING)); ASSERT(ip->i_df.if_format != XFS_DINODE_FMT_BTREE || ip->i_df.if_nextents > XFS_IFORK_MAXEXT(ip, XFS_DATA_FORK)); @@ -3317,6 +3457,8 @@ flush_out: /* generate the checksum. */ xfs_dinode_calc_crc(mp, dip); + if (error) + xfs_inode_mark_sick(ip, XFS_SICK_INO_CORE); return error; } @@ -3777,3 +3919,19 @@ xfs_ifork_zapped( return false; } } + +/* Compute the number of data and realtime blocks used by a file. */ +void +xfs_inode_count_blocks( + struct xfs_trans *tp, + struct xfs_inode *ip, + xfs_filblks_t *dblocks, + xfs_filblks_t *rblocks) +{ + struct xfs_ifork *ifp = xfs_ifork_ptr(ip, XFS_DATA_FORK); + + *rblocks = 0; + if (XFS_IS_REALTIME_INODE(ip)) + xfs_bmap_count_leaves(ifp, rblocks); + *dblocks = ip->i_nblocks - *rblocks; +} diff --git a/fs/xfs/xfs_inode.h b/fs/xfs/xfs_inode.h index 97f63bacd4c2..ab46ffb3ac19 100644 --- a/fs/xfs/xfs_inode.h +++ b/fs/xfs/xfs_inode.h @@ -39,7 +39,7 @@ typedef struct xfs_inode { /* Transaction and locking information. */ struct xfs_inode_log_item *i_itemp; /* logging information */ - mrlock_t i_lock; /* inode lock */ + struct rw_semaphore i_lock; /* inode lock */ atomic_t i_pincount; /* inode pin count */ struct llist_node i_gclist; /* deferred inactivation list */ @@ -171,6 +171,12 @@ static inline struct inode *VFS_I(struct xfs_inode *ip) return &ip->i_vnode; } +/* convert from const xfs inode to const vfs inode */ +static inline const struct inode *VFS_IC(const struct xfs_inode *ip) +{ + return &ip->i_vnode; +} + /* * For regular files we only update the on-disk filesize when actually * writing data back to disk. Until then only the copy in the VFS inode @@ -523,7 +529,7 @@ void xfs_ilock(xfs_inode_t *, uint); int xfs_ilock_nowait(xfs_inode_t *, uint); void xfs_iunlock(xfs_inode_t *, uint); void xfs_ilock_demote(xfs_inode_t *, uint); -bool xfs_isilocked(struct xfs_inode *, uint); +void xfs_assert_ilocked(struct xfs_inode *, uint); uint xfs_ilock_data_map_shared(struct xfs_inode *); uint xfs_ilock_attr_map_shared(struct xfs_inode *); @@ -623,5 +629,32 @@ int xfs_inode_reload_unlinked_bucket(struct xfs_trans *tp, struct xfs_inode *ip) int xfs_inode_reload_unlinked(struct xfs_inode *ip); bool xfs_ifork_zapped(const struct xfs_inode *ip, int whichfork); +void xfs_inode_count_blocks(struct xfs_trans *tp, struct xfs_inode *ip, + xfs_filblks_t *dblocks, xfs_filblks_t *rblocks); + +struct xfs_dir_update_params { + const struct xfs_inode *dp; + const struct xfs_inode *ip; + const struct xfs_name *name; + int delta; +}; + +#ifdef CONFIG_XFS_LIVE_HOOKS +void xfs_dir_update_hook(struct xfs_inode *dp, struct xfs_inode *ip, + int delta, const struct xfs_name *name); + +struct xfs_dir_hook { + struct xfs_hook dirent_hook; +}; + +void xfs_dir_hook_disable(void); +void xfs_dir_hook_enable(void); + +int xfs_dir_hook_add(struct xfs_mount *mp, struct xfs_dir_hook *hook); +void xfs_dir_hook_del(struct xfs_mount *mp, struct xfs_dir_hook *hook); +void xfs_dir_hook_setup(struct xfs_dir_hook *hook, notifier_fn_t mod_fn); +#else +# define xfs_dir_update_hook(dp, ip, delta, name) ((void)0) +#endif /* CONFIG_XFS_LIVE_HOOKS */ #endif /* __XFS_INODE_H__ */ diff --git a/fs/xfs/xfs_inode_item.c b/fs/xfs/xfs_inode_item.c index 0aee97ba0be8..f28d653300d1 100644 --- a/fs/xfs/xfs_inode_item.c +++ b/fs/xfs/xfs_inode_item.c @@ -650,7 +650,7 @@ xfs_inode_item_pin( { struct xfs_inode *ip = INODE_ITEM(lip)->ili_inode; - ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL)); + xfs_assert_ilocked(ip, XFS_ILOCK_EXCL); ASSERT(lip->li_buf); trace_xfs_inode_pin(ip, _RET_IP_); @@ -756,7 +756,7 @@ xfs_inode_item_release( unsigned short lock_flags; ASSERT(ip->i_itemp != NULL); - ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL)); + xfs_assert_ilocked(ip, XFS_ILOCK_EXCL); lock_flags = iip->ili_lock_flags; iip->ili_lock_flags = 0; @@ -856,7 +856,7 @@ xfs_inode_item_destroy( ASSERT(iip->ili_item.li_buf == NULL); ip->i_itemp = NULL; - kmem_free(iip->ili_item.li_lv_shadow); + kvfree(iip->ili_item.li_lv_shadow); kmem_cache_free(xfs_ili_cache, iip); } diff --git a/fs/xfs/xfs_inode_item_recover.c b/fs/xfs/xfs_inode_item_recover.c index 144198a6b270..dbdab4ce7c44 100644 --- a/fs/xfs/xfs_inode_item_recover.c +++ b/fs/xfs/xfs_inode_item_recover.c @@ -291,7 +291,8 @@ xlog_recover_inode_commit_pass2( if (item->ri_buf[0].i_len == sizeof(struct xfs_inode_log_format)) { in_f = item->ri_buf[0].i_addr; } else { - in_f = kmem_alloc(sizeof(struct xfs_inode_log_format), 0); + in_f = kmalloc(sizeof(struct xfs_inode_log_format), + GFP_KERNEL | __GFP_NOFAIL); need_free = 1; error = xfs_inode_item_format_convert(&item->ri_buf[0], in_f); if (error) @@ -553,7 +554,7 @@ out_release: xfs_buf_relse(bp); error: if (need_free) - kmem_free(in_f); + kfree(in_f); return error; } diff --git a/fs/xfs/xfs_ioctl.c b/fs/xfs/xfs_ioctl.c index f02b6e558af5..d0e2cec6210d 100644 --- a/fs/xfs/xfs_ioctl.c +++ b/fs/xfs/xfs_ioctl.c @@ -435,7 +435,7 @@ xfs_ioc_attr_list( copy_to_user(ucursor, &context.cursor, sizeof(context.cursor))) error = -EFAULT; out_free: - kmem_free(buffer); + kvfree(buffer); return error; } @@ -493,7 +493,7 @@ xfs_attrmulti_attr_get( error = -EFAULT; out_kfree: - kmem_free(args.value); + kvfree(args.value); return error; } @@ -1506,7 +1506,7 @@ xfs_ioc_getbmap( error = 0; out_free_buf: - kmem_free(buf); + kvfree(buf); return error; } @@ -1636,7 +1636,7 @@ xfs_ioc_getfsmap( } out_free: - kmem_free(recs); + kvfree(recs); return error; } diff --git a/fs/xfs/xfs_iomap.c b/fs/xfs/xfs_iomap.c index 18c8f168b153..4087af7f3c9f 100644 --- a/fs/xfs/xfs_iomap.c +++ b/fs/xfs/xfs_iomap.c @@ -27,6 +27,7 @@ #include "xfs_dquot_item.h" #include "xfs_dquot.h" #include "xfs_reflink.h" +#include "xfs_health.h" #define XFS_ALLOC_ALIGN(mp, off) \ (((off) >> mp->m_allocsize_log) << mp->m_allocsize_log) @@ -45,6 +46,7 @@ xfs_alert_fsblock_zero( (unsigned long long)imap->br_startoff, (unsigned long long)imap->br_blockcount, imap->br_state); + xfs_bmap_mark_sick(ip, XFS_DATA_FORK); return -EFSCORRUPTED; } @@ -99,8 +101,10 @@ xfs_bmbt_to_iomap( struct xfs_mount *mp = ip->i_mount; struct xfs_buftarg *target = xfs_inode_buftarg(ip); - if (unlikely(!xfs_valid_startblock(ip, imap->br_startblock))) + if (unlikely(!xfs_valid_startblock(ip, imap->br_startblock))) { + xfs_bmap_mark_sick(ip, XFS_DATA_FORK); return xfs_alert_fsblock_zero(ip, imap); + } if (imap->br_startblock == HOLESTARTBLOCK) { iomap->addr = IOMAP_NULL_ADDR; @@ -325,8 +329,10 @@ xfs_iomap_write_direct( goto out_unlock; } - if (unlikely(!xfs_valid_startblock(ip, imap->br_startblock))) + if (unlikely(!xfs_valid_startblock(ip, imap->br_startblock))) { + xfs_bmap_mark_sick(ip, XFS_DATA_FORK); error = xfs_alert_fsblock_zero(ip, imap); + } out_unlock: *seq = xfs_iomap_inode_sequence(ip, 0); @@ -639,8 +645,10 @@ xfs_iomap_write_unwritten( if (error) return error; - if (unlikely(!xfs_valid_startblock(ip, imap.br_startblock))) + if (unlikely(!xfs_valid_startblock(ip, imap.br_startblock))) { + xfs_bmap_mark_sick(ip, XFS_DATA_FORK); return xfs_alert_fsblock_zero(ip, &imap); + } if ((numblks_fsb = imap.br_blockcount) == 0) { /* @@ -986,6 +994,7 @@ xfs_buffered_write_iomap_begin( if (XFS_IS_CORRUPT(mp, !xfs_ifork_has_extents(&ip->i_df)) || XFS_TEST_ERROR(false, mp, XFS_ERRTAG_BMAPIFORMAT)) { + xfs_bmap_mark_sick(ip, XFS_DATA_FORK); error = -EFSCORRUPTED; goto out_unlock; } @@ -1323,7 +1332,7 @@ xfs_seek_iomap_begin( if (cow_fsb != NULLFILEOFF && cow_fsb <= offset_fsb) { if (data_fsb < cow_fsb + cmap.br_blockcount) end_fsb = min(end_fsb, data_fsb); - xfs_trim_extent(&cmap, offset_fsb, end_fsb); + xfs_trim_extent(&cmap, offset_fsb, end_fsb - offset_fsb); seq = xfs_iomap_inode_sequence(ip, IOMAP_F_SHARED); error = xfs_bmbt_to_iomap(ip, iomap, &cmap, flags, IOMAP_F_SHARED, seq); @@ -1348,7 +1357,7 @@ xfs_seek_iomap_begin( imap.br_state = XFS_EXT_NORM; done: seq = xfs_iomap_inode_sequence(ip, 0); - xfs_trim_extent(&imap, offset_fsb, end_fsb); + xfs_trim_extent(&imap, offset_fsb, end_fsb - offset_fsb); error = xfs_bmbt_to_iomap(ip, iomap, &imap, flags, 0, seq); out_unlock: xfs_iunlock(ip, lockmode); diff --git a/fs/xfs/xfs_iops.c b/fs/xfs/xfs_iops.c index a0d77f5f512e..66f8c47642e8 100644 --- a/fs/xfs/xfs_iops.c +++ b/fs/xfs/xfs_iops.c @@ -346,7 +346,7 @@ xfs_vn_ci_lookup( dname.name = ci_name.name; dname.len = ci_name.len; dentry = d_add_ci(dentry, VFS_I(ip), &dname); - kmem_free(ci_name.name); + kfree(ci_name.name); return dentry; } @@ -796,8 +796,7 @@ xfs_setattr_size( uint lock_flags = 0; bool did_zeroing = false; - ASSERT(xfs_isilocked(ip, XFS_IOLOCK_EXCL)); - ASSERT(xfs_isilocked(ip, XFS_MMAPLOCK_EXCL)); + xfs_assert_ilocked(ip, XFS_IOLOCK_EXCL | XFS_MMAPLOCK_EXCL); ASSERT(S_ISREG(inode->i_mode)); ASSERT((iattr->ia_valid & (ATTR_UID|ATTR_GID|ATTR_ATIME|ATTR_ATIME_SET| ATTR_MTIME_SET|ATTR_TIMES_SET)) == 0); @@ -1285,9 +1284,9 @@ xfs_setup_inode( */ lockdep_set_class(&inode->i_rwsem, &inode->i_sb->s_type->i_mutex_dir_key); - lockdep_set_class(&ip->i_lock.mr_lock, &xfs_dir_ilock_class); + lockdep_set_class(&ip->i_lock, &xfs_dir_ilock_class); } else { - lockdep_set_class(&ip->i_lock.mr_lock, &xfs_nondir_ilock_class); + lockdep_set_class(&ip->i_lock, &xfs_nondir_ilock_class); } /* diff --git a/fs/xfs/xfs_itable.c b/fs/xfs/xfs_itable.c index 14462614fcc8..95fc31b9f87d 100644 --- a/fs/xfs/xfs_itable.c +++ b/fs/xfs/xfs_itable.c @@ -197,8 +197,8 @@ xfs_bulkstat_one( ASSERT(breq->icount == 1); - bc.buf = kmem_zalloc(sizeof(struct xfs_bulkstat), - KM_MAYFAIL); + bc.buf = kzalloc(sizeof(struct xfs_bulkstat), + GFP_KERNEL | __GFP_RETRY_MAYFAIL); if (!bc.buf) return -ENOMEM; @@ -214,7 +214,7 @@ xfs_bulkstat_one( breq->startino, &bc); xfs_trans_cancel(tp); out: - kmem_free(bc.buf); + kfree(bc.buf); /* * If we reported one inode to userspace then we abort because we hit @@ -289,8 +289,8 @@ xfs_bulkstat( if (xfs_bulkstat_already_done(breq->mp, breq->startino)) return 0; - bc.buf = kmem_zalloc(sizeof(struct xfs_bulkstat), - KM_MAYFAIL); + bc.buf = kzalloc(sizeof(struct xfs_bulkstat), + GFP_KERNEL | __GFP_RETRY_MAYFAIL); if (!bc.buf) return -ENOMEM; @@ -309,7 +309,7 @@ xfs_bulkstat( xfs_bulkstat_iwalk, breq->icount, &bc); xfs_trans_cancel(tp); out: - kmem_free(bc.buf); + kfree(bc.buf); /* * We found some inodes, so clear the error status and return them. diff --git a/fs/xfs/xfs_iwalk.c b/fs/xfs/xfs_iwalk.c index b3275e8d47b6..01b55f03a102 100644 --- a/fs/xfs/xfs_iwalk.c +++ b/fs/xfs/xfs_iwalk.c @@ -22,6 +22,7 @@ #include "xfs_trans.h" #include "xfs_pwork.h" #include "xfs_ag.h" +#include "xfs_bit.h" /* * Walking Inodes in the Filesystem @@ -99,6 +100,7 @@ xfs_iwalk_ichunk_ra( struct xfs_inobt_rec_incore *irec) { struct xfs_ino_geometry *igeo = M_IGEO(mp); + xfs_agnumber_t agno = pag->pag_agno; xfs_agblock_t agbno; struct blk_plug plug; int i; /* inode chunk index */ @@ -111,8 +113,9 @@ xfs_iwalk_ichunk_ra( imask = xfs_inobt_maskn(i, igeo->inodes_per_cluster); if (imask & ~irec->ir_free) { - xfs_btree_reada_bufs(mp, pag->pag_agno, agbno, - igeo->blocks_per_cluster, + xfs_buf_readahead(mp->m_ddev_targp, + XFS_AGB_TO_DADDR(mp, agno, agbno), + igeo->blocks_per_cluster * mp->m_bsize, &xfs_inode_buf_ops); } agbno += igeo->blocks_per_cluster; @@ -131,21 +134,11 @@ xfs_iwalk_adjust_start( struct xfs_inobt_rec_incore *irec) /* btree record */ { int idx; /* index into inode chunk */ - int i; idx = agino - irec->ir_startino; - /* - * We got a right chunk with some left inodes allocated at it. Grab - * the chunk record. Mark all the uninteresting inodes free because - * they're before our start point. - */ - for (i = 0; i < idx; i++) { - if (XFS_INOBT_MASK(i) & ~irec->ir_free) - irec->ir_freecount++; - } - irec->ir_free |= xfs_inobt_maskn(0, idx); + irec->ir_freecount = hweight64(irec->ir_free); } /* Allocate memory for a walk. */ @@ -160,7 +153,7 @@ xfs_iwalk_alloc( /* Allocate a prefetch buffer for inobt records. */ size = iwag->sz_recs * sizeof(struct xfs_inobt_rec_incore); - iwag->recs = kmem_alloc(size, KM_MAYFAIL); + iwag->recs = kmalloc(size, GFP_KERNEL | __GFP_RETRY_MAYFAIL); if (iwag->recs == NULL) return -ENOMEM; @@ -172,7 +165,7 @@ STATIC void xfs_iwalk_free( struct xfs_iwalk_ag *iwag) { - kmem_free(iwag->recs); + kfree(iwag->recs); iwag->recs = NULL; } @@ -275,9 +268,10 @@ xfs_iwalk_ag_start( /* Set up a fresh cursor and empty the inobt cache. */ iwag->nr_recs = 0; - error = xfs_inobt_cur(pag, tp, XFS_BTNUM_INO, curpp, agi_bpp); + error = xfs_ialloc_read_agi(pag, tp, agi_bpp); if (error) return error; + *curpp = xfs_inobt_init_cursor(pag, tp, *agi_bpp); /* Starting at the beginning of the AG? That's easy! */ if (agino == 0) @@ -306,8 +300,10 @@ xfs_iwalk_ag_start( error = xfs_inobt_get_rec(*curpp, irec, has_more); if (error) return error; - if (XFS_IS_CORRUPT(mp, *has_more != 1)) + if (XFS_IS_CORRUPT(mp, *has_more != 1)) { + xfs_btree_mark_sick(*curpp); return -EFSCORRUPTED; + } iwag->lastino = XFS_AGINO_TO_INO(mp, pag->pag_agno, irec->ir_startino + XFS_INODES_PER_CHUNK - 1); @@ -390,11 +386,10 @@ xfs_iwalk_run_callbacks( } /* ...and recreate the cursor just past where we left off. */ - error = xfs_inobt_cur(iwag->pag, iwag->tp, XFS_BTNUM_INO, curpp, - agi_bpp); + error = xfs_ialloc_read_agi(iwag->pag, iwag->tp, agi_bpp); if (error) return error; - + *curpp = xfs_inobt_init_cursor(iwag->pag, iwag->tp, *agi_bpp); return xfs_inobt_lookup(*curpp, next_agino, XFS_LOOKUP_GE, has_more); } @@ -434,6 +429,7 @@ xfs_iwalk_ag( rec_fsino = XFS_AGINO_TO_INO(mp, pag->pag_agno, irec->ir_startino); if (iwag->lastino != NULLFSINO && XFS_IS_CORRUPT(mp, iwag->lastino >= rec_fsino)) { + xfs_btree_mark_sick(cur); error = -EFSCORRUPTED; goto out; } @@ -627,7 +623,7 @@ xfs_iwalk_ag_work( xfs_iwalk_free(iwag); out: xfs_perag_put(iwag->pag); - kmem_free(iwag); + kfree(iwag); return error; } @@ -663,7 +659,8 @@ xfs_iwalk_threaded( if (xfs_pwork_ctl_want_abort(&pctl)) break; - iwag = kmem_zalloc(sizeof(struct xfs_iwalk_ag), 0); + iwag = kzalloc(sizeof(struct xfs_iwalk_ag), + GFP_KERNEL | __GFP_NOFAIL); iwag->mp = mp; /* diff --git a/fs/xfs/xfs_linux.h b/fs/xfs/xfs_linux.h index d7873e0360f0..8f07c9f6157f 100644 --- a/fs/xfs/xfs_linux.h +++ b/fs/xfs/xfs_linux.h @@ -21,15 +21,13 @@ typedef __u32 xfs_nlink_t; #include "xfs_types.h" -#include "kmem.h" -#include "mrlock.h" - #include <linux/semaphore.h> #include <linux/mm.h> #include <linux/sched/mm.h> #include <linux/kernel.h> #include <linux/blkdev.h> #include <linux/slab.h> +#include <linux/vmalloc.h> #include <linux/crc32c.h> #include <linux/module.h> #include <linux/mutex.h> @@ -51,6 +49,7 @@ typedef __u32 xfs_nlink_t; #include <linux/notifier.h> #include <linux/delay.h> #include <linux/log2.h> +#include <linux/rwsem.h> #include <linux/spinlock.h> #include <linux/random.h> #include <linux/ctype.h> @@ -82,6 +81,7 @@ typedef __u32 xfs_nlink_t; #include "xfs_buf.h" #include "xfs_message.h" #include "xfs_drain.h" +#include "xfs_hooks.h" #ifdef __BIG_ENDIAN #define XFS_NATIVE_HOST 1 @@ -269,4 +269,15 @@ int xfs_rw_bdev(struct block_device *bdev, sector_t sector, unsigned int count, # define PTR_FMT "%p" #endif +/* + * Helper for IO routines to grab backing pages from allocated kernel memory. + */ +static inline struct page * +kmem_to_page(void *addr) +{ + if (is_vmalloc_addr(addr)) + return vmalloc_to_page(addr); + return virt_to_page(addr); +} + #endif /* __XFS_LINUX__ */ diff --git a/fs/xfs/xfs_log.c b/fs/xfs/xfs_log.c index a1650fc81382..5004f23d344e 100644 --- a/fs/xfs/xfs_log.c +++ b/fs/xfs/xfs_log.c @@ -633,14 +633,14 @@ xlog_state_release_iclog( */ int xfs_log_mount( - xfs_mount_t *mp, - xfs_buftarg_t *log_target, - xfs_daddr_t blk_offset, - int num_bblks) + xfs_mount_t *mp, + struct xfs_buftarg *log_target, + xfs_daddr_t blk_offset, + int num_bblks) { - struct xlog *log; - int error = 0; - int min_logfsbs; + struct xlog *log; + int error = 0; + int min_logfsbs; if (!xfs_has_norecovery(mp)) { xfs_notice(mp, "Mounting V%d Filesystem %pU", @@ -1528,7 +1528,7 @@ xlog_alloc_log( int error = -ENOMEM; uint log2_size = 0; - log = kmem_zalloc(sizeof(struct xlog), KM_MAYFAIL); + log = kzalloc(sizeof(struct xlog), GFP_KERNEL | __GFP_RETRY_MAYFAIL); if (!log) { xfs_warn(mp, "Log allocation failed: No memory!"); goto out; @@ -1605,7 +1605,8 @@ xlog_alloc_log( size_t bvec_size = howmany(log->l_iclog_size, PAGE_SIZE) * sizeof(struct bio_vec); - iclog = kmem_zalloc(sizeof(*iclog) + bvec_size, KM_MAYFAIL); + iclog = kzalloc(sizeof(*iclog) + bvec_size, + GFP_KERNEL | __GFP_RETRY_MAYFAIL); if (!iclog) goto out_free_iclog; @@ -1661,13 +1662,13 @@ out_destroy_workqueue: out_free_iclog: for (iclog = log->l_iclog; iclog; iclog = prev_iclog) { prev_iclog = iclog->ic_next; - kmem_free(iclog->ic_data); - kmem_free(iclog); + kvfree(iclog->ic_data); + kfree(iclog); if (prev_iclog == log->l_iclog) break; } out_free_log: - kmem_free(log); + kfree(log); out: return ERR_PTR(error); } /* xlog_alloc_log */ @@ -2118,14 +2119,14 @@ xlog_dealloc_log( iclog = log->l_iclog; for (i = 0; i < log->l_iclog_bufs; i++) { next_iclog = iclog->ic_next; - kmem_free(iclog->ic_data); - kmem_free(iclog); + kvfree(iclog->ic_data); + kfree(iclog); iclog = next_iclog; } log->l_mp->m_log = NULL; destroy_workqueue(log->l_ioend_workqueue); - kmem_free(log); + kfree(log); } /* @@ -3517,7 +3518,8 @@ xlog_ticket_alloc( struct xlog_ticket *tic; int unit_res; - tic = kmem_cache_zalloc(xfs_log_ticket_cache, GFP_NOFS | __GFP_NOFAIL); + tic = kmem_cache_zalloc(xfs_log_ticket_cache, + GFP_KERNEL | __GFP_NOFAIL); unit_res = xlog_calc_unit_res(log, unit_bytes, &tic->t_iclog_hdrs); diff --git a/fs/xfs/xfs_log_cil.c b/fs/xfs/xfs_log_cil.c index 67a99d94701e..73f5b7f628f4 100644 --- a/fs/xfs/xfs_log_cil.c +++ b/fs/xfs/xfs_log_cil.c @@ -100,7 +100,7 @@ xlog_cil_ctx_alloc(void) { struct xfs_cil_ctx *ctx; - ctx = kmem_zalloc(sizeof(*ctx), KM_NOFS); + ctx = kzalloc(sizeof(*ctx), GFP_KERNEL | __GFP_NOFAIL); INIT_LIST_HEAD(&ctx->committing); INIT_LIST_HEAD(&ctx->busy_extents.extent_list); INIT_LIST_HEAD(&ctx->log_items); @@ -339,7 +339,7 @@ xlog_cil_alloc_shadow_bufs( * the buffer, only the log vector header and the iovec * storage. */ - kmem_free(lip->li_lv_shadow); + kvfree(lip->li_lv_shadow); lv = xlog_kvmalloc(buf_size); memset(lv, 0, xlog_cil_iovec_space(niovecs)); @@ -703,7 +703,7 @@ xlog_cil_free_logvec( while (!list_empty(lv_chain)) { lv = list_first_entry(lv_chain, struct xfs_log_vec, lv_list); list_del_init(&lv->lv_list); - kmem_free(lv); + kvfree(lv); } } @@ -753,7 +753,7 @@ xlog_cil_committed( return; } - kmem_free(ctx); + kfree(ctx); } void @@ -1116,11 +1116,18 @@ xlog_cil_cleanup_whiteouts( * same sequence twice. If we get a race between multiple pushes for the same * sequence they will block on the first one and then abort, hence avoiding * needless pushes. + * + * This runs from a workqueue so it does not inherent any specific memory + * allocation context. However, we do not want to block on memory reclaim + * recursing back into the filesystem because this push may have been triggered + * by memory reclaim itself. Hence we really need to run under full GFP_NOFS + * contraints here. */ static void xlog_cil_push_work( struct work_struct *work) { + unsigned int nofs_flags = memalloc_nofs_save(); struct xfs_cil_ctx *ctx = container_of(work, struct xfs_cil_ctx, push_work); struct xfs_cil *cil = ctx->cil; @@ -1334,12 +1341,14 @@ xlog_cil_push_work( spin_unlock(&log->l_icloglock); xlog_cil_cleanup_whiteouts(&whiteouts); xfs_log_ticket_ungrant(log, ticket); + memalloc_nofs_restore(nofs_flags); return; out_skip: up_write(&cil->xc_ctx_lock); xfs_log_ticket_put(new_ctx->ticket); - kmem_free(new_ctx); + kfree(new_ctx); + memalloc_nofs_restore(nofs_flags); return; out_abort_free_ticket: @@ -1348,6 +1357,7 @@ out_abort_free_ticket: if (!ctx->commit_iclog) { xfs_log_ticket_ungrant(log, ctx->ticket); xlog_cil_committed(ctx); + memalloc_nofs_restore(nofs_flags); return; } spin_lock(&log->l_icloglock); @@ -1356,6 +1366,7 @@ out_abort_free_ticket: /* Not safe to reference ctx now! */ spin_unlock(&log->l_icloglock); xfs_log_ticket_ungrant(log, ticket); + memalloc_nofs_restore(nofs_flags); } /* @@ -1533,7 +1544,7 @@ xlog_cil_process_intents( set_bit(XFS_LI_WHITEOUT, &ilip->li_flags); trace_xfs_cil_whiteout_mark(ilip); len += ilip->li_lv->lv_bytes; - kmem_free(ilip->li_lv); + kvfree(ilip->li_lv); ilip->li_lv = NULL; xfs_trans_del_item(lip); @@ -1747,7 +1758,7 @@ xlog_cil_init( struct xlog_cil_pcp *cilpcp; int cpu; - cil = kmem_zalloc(sizeof(*cil), KM_MAYFAIL); + cil = kzalloc(sizeof(*cil), GFP_KERNEL | __GFP_RETRY_MAYFAIL); if (!cil) return -ENOMEM; /* @@ -1786,7 +1797,7 @@ xlog_cil_init( out_destroy_wq: destroy_workqueue(cil->xc_push_wq); out_destroy_cil: - kmem_free(cil); + kfree(cil); return -ENOMEM; } @@ -1799,12 +1810,12 @@ xlog_cil_destroy( if (cil->xc_ctx) { if (cil->xc_ctx->ticket) xfs_log_ticket_put(cil->xc_ctx->ticket); - kmem_free(cil->xc_ctx); + kfree(cil->xc_ctx); } ASSERT(test_bit(XLOG_CIL_EMPTY, &cil->xc_flags)); free_percpu(cil->xc_pcp); destroy_workqueue(cil->xc_push_wq); - kmem_free(cil); + kfree(cil); } diff --git a/fs/xfs/xfs_log_recover.c b/fs/xfs/xfs_log_recover.c index 1251c81e55f9..13f1d2e91540 100644 --- a/fs/xfs/xfs_log_recover.c +++ b/fs/xfs/xfs_log_recover.c @@ -361,7 +361,7 @@ xlog_find_verify_cycle( *new_blk = -1; out: - kmem_free(buffer); + kvfree(buffer); return error; } @@ -477,7 +477,7 @@ xlog_find_verify_log_record( *last_blk = i; out: - kmem_free(buffer); + kvfree(buffer); return error; } @@ -731,7 +731,7 @@ validate_head: goto out_free_buffer; } - kmem_free(buffer); + kvfree(buffer); if (head_blk == log_bbnum) *return_head_blk = 0; else @@ -745,7 +745,7 @@ validate_head: return 0; out_free_buffer: - kmem_free(buffer); + kvfree(buffer); if (error) xfs_warn(log->l_mp, "failed to find log head"); return error; @@ -999,7 +999,7 @@ xlog_verify_tail( "Tail block (0x%llx) overwrite detected. Updated to 0x%llx", orig_tail, *tail_blk); out: - kmem_free(buffer); + kvfree(buffer); return error; } @@ -1046,7 +1046,7 @@ xlog_verify_head( error = xlog_rseek_logrec_hdr(log, *head_blk, *tail_blk, XLOG_MAX_ICLOGS, tmp_buffer, &tmp_rhead_blk, &tmp_rhead, &tmp_wrapped); - kmem_free(tmp_buffer); + kvfree(tmp_buffer); if (error < 0) return error; @@ -1365,7 +1365,7 @@ xlog_find_tail( error = xlog_clear_stale_blocks(log, tail_lsn); done: - kmem_free(buffer); + kvfree(buffer); if (error) xfs_warn(log->l_mp, "failed to locate log tail"); @@ -1399,6 +1399,7 @@ xlog_find_zeroed( xfs_daddr_t new_blk, last_blk, start_blk; xfs_daddr_t num_scan_bblks; int error, log_bbnum = log->l_logBBsize; + int ret = 1; *blk_no = 0; @@ -1413,8 +1414,7 @@ xlog_find_zeroed( first_cycle = xlog_get_cycle(offset); if (first_cycle == 0) { /* completely zeroed log */ *blk_no = 0; - kmem_free(buffer); - return 1; + goto out_free_buffer; } /* check partially zeroed log */ @@ -1424,8 +1424,8 @@ xlog_find_zeroed( last_cycle = xlog_get_cycle(offset); if (last_cycle != 0) { /* log completely written to */ - kmem_free(buffer); - return 0; + ret = 0; + goto out_free_buffer; } /* we have a partially zeroed log */ @@ -1471,10 +1471,10 @@ xlog_find_zeroed( *blk_no = last_blk; out_free_buffer: - kmem_free(buffer); + kvfree(buffer); if (error) return error; - return 1; + return ret; } /* @@ -1583,7 +1583,7 @@ xlog_write_log_records( } out_free_buffer: - kmem_free(buffer); + kvfree(buffer); return error; } @@ -2057,7 +2057,8 @@ xlog_recover_add_item( { struct xlog_recover_item *item; - item = kmem_zalloc(sizeof(struct xlog_recover_item), 0); + item = kzalloc(sizeof(struct xlog_recover_item), + GFP_KERNEL | __GFP_NOFAIL); INIT_LIST_HEAD(&item->ri_list); list_add_tail(&item->ri_list, head); } @@ -2160,7 +2161,7 @@ xlog_recover_add_to_trans( return 0; } - ptr = kmem_alloc(len, 0); + ptr = xlog_kvmalloc(len); memcpy(ptr, dp, len); in_f = (struct xfs_inode_log_format *)ptr; @@ -2182,14 +2183,13 @@ xlog_recover_add_to_trans( "bad number of regions (%d) in inode log format", in_f->ilf_size); ASSERT(0); - kmem_free(ptr); + kvfree(ptr); return -EFSCORRUPTED; } item->ri_total = in_f->ilf_size; - item->ri_buf = - kmem_zalloc(item->ri_total * sizeof(xfs_log_iovec_t), - 0); + item->ri_buf = kzalloc(item->ri_total * sizeof(xfs_log_iovec_t), + GFP_KERNEL | __GFP_NOFAIL); } if (item->ri_total <= item->ri_cnt) { @@ -2197,7 +2197,7 @@ xlog_recover_add_to_trans( "log item region count (%d) overflowed size (%d)", item->ri_cnt, item->ri_total); ASSERT(0); - kmem_free(ptr); + kvfree(ptr); return -EFSCORRUPTED; } @@ -2227,13 +2227,13 @@ xlog_recover_free_trans( /* Free the regions in the item. */ list_del(&item->ri_list); for (i = 0; i < item->ri_cnt; i++) - kmem_free(item->ri_buf[i].i_addr); + kvfree(item->ri_buf[i].i_addr); /* Free the item itself */ - kmem_free(item->ri_buf); - kmem_free(item); + kfree(item->ri_buf); + kfree(item); } /* Free the transaction recover structure */ - kmem_free(trans); + kfree(trans); } /* @@ -2332,7 +2332,7 @@ xlog_recover_ophdr_to_trans( * This is a new transaction so allocate a new recovery container to * hold the recovery ops that will follow. */ - trans = kmem_zalloc(sizeof(struct xlog_recover), 0); + trans = kzalloc(sizeof(struct xlog_recover), GFP_KERNEL | __GFP_NOFAIL); trans->r_log_tid = tid; trans->r_lsn = be64_to_cpu(rhead->h_lsn); INIT_LIST_HEAD(&trans->r_itemq); @@ -3024,7 +3024,7 @@ xlog_do_recovery_pass( hblks = xlog_logrec_hblks(log, rhead); if (hblks != 1) { - kmem_free(hbp); + kvfree(hbp); hbp = xlog_alloc_buffer(log, hblks); } } else { @@ -3038,7 +3038,7 @@ xlog_do_recovery_pass( return -ENOMEM; dbp = xlog_alloc_buffer(log, BTOBB(h_size)); if (!dbp) { - kmem_free(hbp); + kvfree(hbp); return -ENOMEM; } @@ -3199,16 +3199,33 @@ xlog_do_recovery_pass( } bread_err2: - kmem_free(dbp); + kvfree(dbp); bread_err1: - kmem_free(hbp); + kvfree(hbp); /* - * Submit buffers that have been added from the last record processed, - * regardless of error status. + * Submit buffers that have been dirtied by the last record recovered. */ - if (!list_empty(&buffer_list)) + if (!list_empty(&buffer_list)) { + if (error) { + /* + * If there has been an item recovery error then we + * cannot allow partial checkpoint writeback to + * occur. We might have multiple checkpoints with the + * same start LSN in this buffer list, and partial + * writeback of a checkpoint in this situation can + * prevent future recovery of all the changes in the + * checkpoints at this start LSN. + * + * Note: Shutting down the filesystem will result in the + * delwri submission marking all the buffers stale, + * completing them and cleaning up _XBF_LOGRECOVERY + * state without doing any IO. + */ + xlog_force_shutdown(log, SHUTDOWN_LOG_IO_ERROR); + } error2 = xfs_buf_delwri_submit(&buffer_list); + } if (error && first_bad) *first_bad = rhead_blk; @@ -3443,12 +3460,19 @@ xlog_recover( * part of recovery so that the root and real-time bitmap inodes can be read in * from disk in between the two stages. This is necessary so that we can free * space in the real-time portion of the file system. + * + * We run this whole process under GFP_NOFS allocation context. We do a + * combination of non-transactional and transactional work, yet we really don't + * want to recurse into the filesystem from direct reclaim during any of this + * processing. This allows all the recovery code run here not to care about the + * memory allocation context it is running in. */ int xlog_recover_finish( struct xlog *log) { - int error; + unsigned int nofs_flags = memalloc_nofs_save(); + int error; error = xlog_recover_process_intents(log); if (error) { @@ -3462,7 +3486,7 @@ xlog_recover_finish( xlog_recover_cancel_intents(log); xfs_alert(log->l_mp, "Failed to recover intents"); xlog_force_shutdown(log, SHUTDOWN_LOG_IO_ERROR); - return error; + goto out_error; } /* @@ -3483,7 +3507,7 @@ xlog_recover_finish( if (error < 0) { xfs_alert(log->l_mp, "Failed to clear log incompat features on recovery"); - return error; + goto out_error; } } @@ -3508,9 +3532,13 @@ xlog_recover_finish( * and AIL. */ xlog_force_shutdown(log, SHUTDOWN_LOG_IO_ERROR); + error = 0; + goto out_error; } - return 0; +out_error: + memalloc_nofs_restore(nofs_flags); + return error; } void diff --git a/fs/xfs/xfs_mount.c b/fs/xfs/xfs_mount.c index 57fa21ad7912..df370eb5dc15 100644 --- a/fs/xfs/xfs_mount.c +++ b/fs/xfs/xfs_mount.c @@ -45,7 +45,7 @@ xfs_uuid_table_free(void) { if (xfs_uuid_table_size == 0) return; - kmem_free(xfs_uuid_table); + kfree(xfs_uuid_table); xfs_uuid_table = NULL; xfs_uuid_table_size = 0; } diff --git a/fs/xfs/xfs_mount.h b/fs/xfs/xfs_mount.h index 503fe3c7edbf..e880aa48de68 100644 --- a/fs/xfs/xfs_mount.h +++ b/fs/xfs/xfs_mount.h @@ -94,9 +94,9 @@ typedef struct xfs_mount { struct xfs_inode *m_rsumip; /* pointer to summary inode */ struct xfs_inode *m_rootip; /* pointer to root directory */ struct xfs_quotainfo *m_quotainfo; /* disk quota information */ - xfs_buftarg_t *m_ddev_targp; /* saves taking the address */ - xfs_buftarg_t *m_logdev_targp;/* ptr to log device */ - xfs_buftarg_t *m_rtdev_targp; /* ptr to rt device */ + struct xfs_buftarg *m_ddev_targp; /* data device */ + struct xfs_buftarg *m_logdev_targp;/* log device */ + struct xfs_buftarg *m_rtdev_targp; /* rt device */ void __percpu *m_inodegc; /* percpu inodegc structures */ /* @@ -252,6 +252,9 @@ typedef struct xfs_mount { /* cpus that have inodes queued for inactivation */ struct cpumask m_inodegc_cpumask; + + /* Hook to feed dirent updates to an active online repair. */ + struct xfs_hooks m_dir_update_hooks; } xfs_mount_t; #define M_IGEO(mp) (&(mp)->m_ino_geo) @@ -502,9 +505,6 @@ xfs_daddr_to_agbno(struct xfs_mount *mp, xfs_daddr_t d) return (xfs_agblock_t) do_div(ld, mp->m_sb.sb_agblocks); } -int xfs_buf_hash_init(struct xfs_perag *pag); -void xfs_buf_hash_destroy(struct xfs_perag *pag); - extern void xfs_uuid_table_free(void); extern uint64_t xfs_default_resblks(xfs_mount_t *mp); extern int xfs_mountfs(xfs_mount_t *mp); diff --git a/fs/xfs/xfs_mru_cache.c b/fs/xfs/xfs_mru_cache.c index f85e3b07ab44..7443debaffd6 100644 --- a/fs/xfs/xfs_mru_cache.c +++ b/fs/xfs/xfs_mru_cache.c @@ -333,13 +333,14 @@ xfs_mru_cache_create( if (!(grp_time = msecs_to_jiffies(lifetime_ms) / grp_count)) return -EINVAL; - if (!(mru = kmem_zalloc(sizeof(*mru), 0))) + mru = kzalloc(sizeof(*mru), GFP_KERNEL | __GFP_NOFAIL); + if (!mru) return -ENOMEM; /* An extra list is needed to avoid reaping up to a grp_time early. */ mru->grp_count = grp_count + 1; - mru->lists = kmem_zalloc(mru->grp_count * sizeof(*mru->lists), 0); - + mru->lists = kzalloc(mru->grp_count * sizeof(*mru->lists), + GFP_KERNEL | __GFP_NOFAIL); if (!mru->lists) { err = -ENOMEM; goto exit; @@ -364,9 +365,9 @@ xfs_mru_cache_create( exit: if (err && mru && mru->lists) - kmem_free(mru->lists); + kfree(mru->lists); if (err && mru) - kmem_free(mru); + kfree(mru); return err; } @@ -406,8 +407,8 @@ xfs_mru_cache_destroy( xfs_mru_cache_flush(mru); - kmem_free(mru->lists); - kmem_free(mru); + kfree(mru->lists); + kfree(mru); } /* @@ -427,7 +428,7 @@ xfs_mru_cache_insert( if (!mru || !mru->lists) return -EINVAL; - if (radix_tree_preload(GFP_NOFS)) + if (radix_tree_preload(GFP_KERNEL)) return -ENOMEM; INIT_LIST_HEAD(&elem->list_node); diff --git a/fs/xfs/xfs_qm.c b/fs/xfs/xfs_qm.c index 67d0a8564ff3..0f4cf4170c35 100644 --- a/fs/xfs/xfs_qm.c +++ b/fs/xfs/xfs_qm.c @@ -26,6 +26,7 @@ #include "xfs_ag.h" #include "xfs_ialloc.h" #include "xfs_log_priv.h" +#include "xfs_health.h" /* * The global quota manager. There is only one of these for the entire @@ -254,7 +255,7 @@ xfs_qm_dqattach_one( struct xfs_dquot *dqp; int error; - ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL)); + xfs_assert_ilocked(ip, XFS_ILOCK_EXCL); error = 0; /* @@ -322,7 +323,7 @@ xfs_qm_dqattach_locked( if (!xfs_qm_need_dqattach(ip)) return 0; - ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL)); + xfs_assert_ilocked(ip, XFS_ILOCK_EXCL); if (XFS_IS_UQUOTA_ON(mp) && !ip->i_udquot) { error = xfs_qm_dqattach_one(ip, XFS_DQTYPE_USER, @@ -353,7 +354,7 @@ done: * Don't worry about the dquots that we may have attached before any * error - they'll get detached later if it has not already been done. */ - ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL)); + xfs_assert_ilocked(ip, XFS_ILOCK_EXCL); return error; } @@ -628,7 +629,8 @@ xfs_qm_init_quotainfo( ASSERT(XFS_IS_QUOTA_ON(mp)); - qinf = mp->m_quotainfo = kmem_zalloc(sizeof(struct xfs_quotainfo), 0); + qinf = mp->m_quotainfo = kzalloc(sizeof(struct xfs_quotainfo), + GFP_KERNEL | __GFP_NOFAIL); error = list_lru_init(&qinf->qi_lru); if (error) @@ -642,9 +644,9 @@ xfs_qm_init_quotainfo( if (error) goto out_free_lru; - INIT_RADIX_TREE(&qinf->qi_uquota_tree, GFP_NOFS); - INIT_RADIX_TREE(&qinf->qi_gquota_tree, GFP_NOFS); - INIT_RADIX_TREE(&qinf->qi_pquota_tree, GFP_NOFS); + INIT_RADIX_TREE(&qinf->qi_uquota_tree, GFP_KERNEL); + INIT_RADIX_TREE(&qinf->qi_gquota_tree, GFP_KERNEL); + INIT_RADIX_TREE(&qinf->qi_pquota_tree, GFP_KERNEL); mutex_init(&qinf->qi_tree_lock); /* mutex used to serialize quotaoffs */ @@ -691,6 +693,9 @@ xfs_qm_init_quotainfo( shrinker_register(qinf->qi_shrinker); + xfs_hooks_init(&qinf->qi_mod_ino_dqtrx_hooks); + xfs_hooks_init(&qinf->qi_apply_dqtrx_hooks); + return 0; out_free_inos: @@ -700,7 +705,7 @@ out_free_inos: out_free_lru: list_lru_destroy(&qinf->qi_lru); out_free_qinf: - kmem_free(qinf); + kfree(qinf); mp->m_quotainfo = NULL; return error; } @@ -724,7 +729,7 @@ xfs_qm_destroy_quotainfo( xfs_qm_destroy_quotainos(qi); mutex_destroy(&qi->qi_tree_lock); mutex_destroy(&qi->qi_quotaofflock); - kmem_free(qi); + kfree(qi); mp->m_quotainfo = NULL; } @@ -758,14 +763,18 @@ xfs_qm_qino_alloc( (mp->m_sb.sb_gquotino != NULLFSINO)) { ino = mp->m_sb.sb_gquotino; if (XFS_IS_CORRUPT(mp, - mp->m_sb.sb_pquotino != NULLFSINO)) + mp->m_sb.sb_pquotino != NULLFSINO)) { + xfs_fs_mark_sick(mp, XFS_SICK_FS_PQUOTA); return -EFSCORRUPTED; + } } else if ((flags & XFS_QMOPT_GQUOTA) && (mp->m_sb.sb_pquotino != NULLFSINO)) { ino = mp->m_sb.sb_pquotino; if (XFS_IS_CORRUPT(mp, - mp->m_sb.sb_gquotino != NULLFSINO)) + mp->m_sb.sb_gquotino != NULLFSINO)) { + xfs_fs_mark_sick(mp, XFS_SICK_FS_GQUOTA); return -EFSCORRUPTED; + } } if (ino != NULLFSINO) { error = xfs_iget(mp, NULL, ino, 0, 0, ipp); @@ -996,7 +1005,8 @@ xfs_qm_reset_dqcounts_buf( if (qip->i_nblocks == 0) return 0; - map = kmem_alloc(XFS_DQITER_MAP_SIZE * sizeof(*map), 0); + map = kmalloc(XFS_DQITER_MAP_SIZE * sizeof(*map), + GFP_KERNEL | __GFP_NOFAIL); lblkno = 0; maxlblkcnt = XFS_B_TO_FSB(mp, mp->m_super->s_maxbytes); @@ -1058,7 +1068,7 @@ xfs_qm_reset_dqcounts_buf( } while (nmaps > 0); out: - kmem_free(map); + kfree(map); return error; } @@ -1406,8 +1416,12 @@ error_return: xfs_warn(mp, "Quotacheck: Failed to reset quota flags."); } - } else + xfs_fs_mark_sick(mp, XFS_SICK_FS_QUOTACHECK); + } else { xfs_notice(mp, "Quotacheck: Done."); + xfs_fs_mark_healthy(mp, XFS_SICK_FS_QUOTACHECK); + } + return error; error_purge: @@ -1809,7 +1823,7 @@ xfs_qm_vop_chown( XFS_TRANS_DQ_RTBCOUNT : XFS_TRANS_DQ_BCOUNT; - ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL)); + xfs_assert_ilocked(ip, XFS_ILOCK_EXCL); ASSERT(XFS_IS_QUOTA_ON(ip->i_mount)); /* old dquot */ @@ -1817,12 +1831,12 @@ xfs_qm_vop_chown( ASSERT(prevdq); ASSERT(prevdq != newdq); - xfs_trans_mod_dquot(tp, prevdq, bfield, -(ip->i_nblocks)); - xfs_trans_mod_dquot(tp, prevdq, XFS_TRANS_DQ_ICOUNT, -1); + xfs_trans_mod_ino_dquot(tp, ip, prevdq, bfield, -(ip->i_nblocks)); + xfs_trans_mod_ino_dquot(tp, ip, prevdq, XFS_TRANS_DQ_ICOUNT, -1); /* the sparkling new dquot */ - xfs_trans_mod_dquot(tp, newdq, bfield, ip->i_nblocks); - xfs_trans_mod_dquot(tp, newdq, XFS_TRANS_DQ_ICOUNT, 1); + xfs_trans_mod_ino_dquot(tp, ip, newdq, bfield, ip->i_nblocks); + xfs_trans_mod_ino_dquot(tp, ip, newdq, XFS_TRANS_DQ_ICOUNT, 1); /* * Back when we made quota reservations for the chown, we reserved the @@ -1897,29 +1911,28 @@ xfs_qm_vop_create_dqattach( if (!XFS_IS_QUOTA_ON(mp)) return; - ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL)); + xfs_assert_ilocked(ip, XFS_ILOCK_EXCL); if (udqp && XFS_IS_UQUOTA_ON(mp)) { ASSERT(ip->i_udquot == NULL); ASSERT(i_uid_read(VFS_I(ip)) == udqp->q_id); ip->i_udquot = xfs_qm_dqhold(udqp); - xfs_trans_mod_dquot(tp, udqp, XFS_TRANS_DQ_ICOUNT, 1); } if (gdqp && XFS_IS_GQUOTA_ON(mp)) { ASSERT(ip->i_gdquot == NULL); ASSERT(i_gid_read(VFS_I(ip)) == gdqp->q_id); ip->i_gdquot = xfs_qm_dqhold(gdqp); - xfs_trans_mod_dquot(tp, gdqp, XFS_TRANS_DQ_ICOUNT, 1); } if (pdqp && XFS_IS_PQUOTA_ON(mp)) { ASSERT(ip->i_pdquot == NULL); ASSERT(ip->i_projid == pdqp->q_id); ip->i_pdquot = xfs_qm_dqhold(pdqp); - xfs_trans_mod_dquot(tp, pdqp, XFS_TRANS_DQ_ICOUNT, 1); } + + xfs_trans_mod_dquot_byino(tp, ip, XFS_TRANS_DQ_ICOUNT, 1); } /* Decide if this inode's dquot is near an enforcement boundary. */ diff --git a/fs/xfs/xfs_qm.h b/fs/xfs/xfs_qm.h index d5c9fc4ba591..f5993012bf98 100644 --- a/fs/xfs/xfs_qm.h +++ b/fs/xfs/xfs_qm.h @@ -68,6 +68,10 @@ struct xfs_quotainfo { /* Minimum and maximum quota expiration timestamp values. */ time64_t qi_expiry_min; time64_t qi_expiry_max; + + /* Hook to feed quota counter updates to an active online repair. */ + struct xfs_hooks qi_mod_ino_dqtrx_hooks; + struct xfs_hooks qi_apply_dqtrx_hooks; }; static inline struct radix_tree_root * @@ -104,6 +108,18 @@ xfs_quota_inode(struct xfs_mount *mp, xfs_dqtype_t type) return NULL; } +/* + * Parameters for tracking dqtrx changes on behalf of an inode. The hook + * function arg parameter is the field being updated. + */ +struct xfs_mod_ino_dqtrx_params { + uintptr_t tx_id; + xfs_ino_t ino; + xfs_dqtype_t q_type; + xfs_dqid_t q_id; + int64_t delta; +}; + extern void xfs_trans_mod_dquot(struct xfs_trans *tp, struct xfs_dquot *dqp, uint field, int64_t delta); extern void xfs_trans_dqjoin(struct xfs_trans *, struct xfs_dquot *); diff --git a/fs/xfs/xfs_qm_bhv.c b/fs/xfs/xfs_qm_bhv.c index b77673dd0558..271c1021c733 100644 --- a/fs/xfs/xfs_qm_bhv.c +++ b/fs/xfs/xfs_qm_bhv.c @@ -9,6 +9,7 @@ #include "xfs_format.h" #include "xfs_log_format.h" #include "xfs_trans_resv.h" +#include "xfs_mount.h" #include "xfs_quota.h" #include "xfs_mount.h" #include "xfs_inode.h" diff --git a/fs/xfs/xfs_quota.h b/fs/xfs/xfs_quota.h index e0d56489f3b2..85a4ae1a17f6 100644 --- a/fs/xfs/xfs_quota.h +++ b/fs/xfs/xfs_quota.h @@ -74,6 +74,22 @@ struct xfs_dqtrx { int64_t qt_icount_delta; /* dquot inode count changes */ }; +enum xfs_apply_dqtrx_type { + XFS_APPLY_DQTRX_COMMIT = 0, + XFS_APPLY_DQTRX_UNRESERVE, +}; + +/* + * Parameters for applying dqtrx changes to a dquot. The hook function arg + * parameter is enum xfs_apply_dqtrx_type. + */ +struct xfs_apply_dqtrx_params { + uintptr_t tx_id; + xfs_ino_t ino; + xfs_dqtype_t q_type; + xfs_dqid_t q_id; +}; + #ifdef CONFIG_XFS_QUOTA extern void xfs_trans_dup_dqinfo(struct xfs_trans *, struct xfs_trans *); extern void xfs_trans_free_dqinfo(struct xfs_trans *); @@ -114,6 +130,30 @@ xfs_quota_reserve_blkres(struct xfs_inode *ip, int64_t blocks) return xfs_trans_reserve_quota_nblks(NULL, ip, blocks, 0, false); } bool xfs_inode_near_dquot_enforcement(struct xfs_inode *ip, xfs_dqtype_t type); + +# ifdef CONFIG_XFS_LIVE_HOOKS +void xfs_trans_mod_ino_dquot(struct xfs_trans *tp, struct xfs_inode *ip, + struct xfs_dquot *dqp, unsigned int field, int64_t delta); + +struct xfs_quotainfo; + +struct xfs_dqtrx_hook { + struct xfs_hook mod_hook; + struct xfs_hook apply_hook; +}; + +void xfs_dqtrx_hook_disable(void); +void xfs_dqtrx_hook_enable(void); + +int xfs_dqtrx_hook_add(struct xfs_quotainfo *qi, struct xfs_dqtrx_hook *hook); +void xfs_dqtrx_hook_del(struct xfs_quotainfo *qi, struct xfs_dqtrx_hook *hook); +void xfs_dqtrx_hook_setup(struct xfs_dqtrx_hook *hook, notifier_fn_t mod_fn, + notifier_fn_t apply_fn); +# else +# define xfs_trans_mod_ino_dquot(tp, ip, dqp, field, delta) \ + xfs_trans_mod_dquot((tp), (dqp), (field), (delta)) +# endif /* CONFIG_XFS_LIVE_HOOKS */ + #else static inline int xfs_qm_vop_dqalloc(struct xfs_inode *ip, kuid_t kuid, kgid_t kgid, @@ -173,6 +213,12 @@ xfs_trans_reserve_quota_icreate(struct xfs_trans *tp, struct xfs_dquot *udqp, #define xfs_qm_unmount(mp) #define xfs_qm_unmount_quotas(mp) #define xfs_inode_near_dquot_enforcement(ip, type) (false) + +# ifdef CONFIG_XFS_LIVE_HOOKS +# define xfs_dqtrx_hook_enable() ((void)0) +# define xfs_dqtrx_hook_disable() ((void)0) +# endif /* CONFIG_XFS_LIVE_HOOKS */ + #endif /* CONFIG_XFS_QUOTA */ static inline int diff --git a/fs/xfs/xfs_refcount_item.c b/fs/xfs/xfs_refcount_item.c index 20ad8086da60..14919b33e4fe 100644 --- a/fs/xfs/xfs_refcount_item.c +++ b/fs/xfs/xfs_refcount_item.c @@ -36,9 +36,9 @@ STATIC void xfs_cui_item_free( struct xfs_cui_log_item *cuip) { - kmem_free(cuip->cui_item.li_lv_shadow); + kvfree(cuip->cui_item.li_lv_shadow); if (cuip->cui_format.cui_nextents > XFS_CUI_MAX_FAST_EXTENTS) - kmem_free(cuip); + kfree(cuip); else kmem_cache_free(xfs_cui_cache, cuip); } @@ -143,8 +143,8 @@ xfs_cui_init( ASSERT(nextents > 0); if (nextents > XFS_CUI_MAX_FAST_EXTENTS) - cuip = kmem_zalloc(xfs_cui_log_item_sizeof(nextents), - 0); + cuip = kzalloc(xfs_cui_log_item_sizeof(nextents), + GFP_KERNEL | __GFP_NOFAIL); else cuip = kmem_cache_zalloc(xfs_cui_cache, GFP_KERNEL | __GFP_NOFAIL); @@ -207,7 +207,7 @@ xfs_cud_item_release( struct xfs_cud_log_item *cudp = CUD_ITEM(lip); xfs_cui_release(cudp->cud_cuip); - kmem_free(cudp->cud_item.li_lv_shadow); + kvfree(cudp->cud_item.li_lv_shadow); kmem_cache_free(xfs_cud_cache, cudp); } @@ -425,7 +425,7 @@ xfs_cui_recover_work( struct xfs_refcount_intent *ri; ri = kmem_cache_alloc(xfs_refcount_intent_cache, - GFP_NOFS | __GFP_NOFAIL); + GFP_KERNEL | __GFP_NOFAIL); ri->ri_type = pmap->pe_flags & XFS_REFCOUNT_EXTENT_TYPE_MASK; ri->ri_startblock = pmap->pe_startblock; ri->ri_blockcount = pmap->pe_len; diff --git a/fs/xfs/xfs_reflink.c b/fs/xfs/xfs_reflink.c index d5ca8bcae65b..7da0e8f961d3 100644 --- a/fs/xfs/xfs_reflink.c +++ b/fs/xfs/xfs_reflink.c @@ -29,6 +29,7 @@ #include "xfs_iomap.h" #include "xfs_ag.h" #include "xfs_ag_resv.h" +#include "xfs_health.h" /* * Copy on Write of Shared Blocks @@ -527,7 +528,7 @@ xfs_reflink_allocate_cow( int error; bool found; - ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL)); + xfs_assert_ilocked(ip, XFS_ILOCK_EXCL); if (!ip->i_cowfp) { ASSERT(!xfs_is_reflink_inode(ip)); xfs_ifork_init_cow(ip); @@ -805,7 +806,7 @@ xfs_reflink_end_cow_extent( * If the extent we're remapping is backed by storage (written * or not), unmap the extent and drop its refcount. */ - xfs_bmap_unmap_extent(tp, ip, &data); + xfs_bmap_unmap_extent(tp, ip, XFS_DATA_FORK, &data); xfs_refcount_decrease_extent(tp, &data); xfs_trans_mod_dquot_byino(tp, ip, XFS_TRANS_DQ_BCOUNT, -data.br_blockcount); @@ -829,7 +830,7 @@ xfs_reflink_end_cow_extent( xfs_refcount_free_cow_extent(tp, del.br_startblock, del.br_blockcount); /* Map the new blocks into the data fork. */ - xfs_bmap_map_extent(tp, ip, &del); + xfs_bmap_map_extent(tp, ip, XFS_DATA_FORK, &del); /* Charge this new data fork mapping to the on-disk quota. */ xfs_trans_mod_dquot_byino(tp, ip, XFS_TRANS_DQ_DELBCOUNT, @@ -1227,8 +1228,10 @@ xfs_reflink_remap_extent( * extent if they're both holes or both the same physical extent. */ if (dmap->br_startblock == smap.br_startblock) { - if (dmap->br_state != smap.br_state) + if (dmap->br_state != smap.br_state) { + xfs_bmap_mark_sick(ip, XFS_DATA_FORK); error = -EFSCORRUPTED; + } goto out_cancel; } @@ -1291,7 +1294,7 @@ xfs_reflink_remap_extent( * If the extent we're unmapping is backed by storage (written * or not), unmap the extent and drop its refcount. */ - xfs_bmap_unmap_extent(tp, ip, &smap); + xfs_bmap_unmap_extent(tp, ip, XFS_DATA_FORK, &smap); xfs_refcount_decrease_extent(tp, &smap); qdelta -= smap.br_blockcount; } else if (smap.br_startblock == DELAYSTARTBLOCK) { @@ -1316,7 +1319,7 @@ xfs_reflink_remap_extent( */ if (dmap_written) { xfs_refcount_increase_extent(tp, dmap); - xfs_bmap_map_extent(tp, ip, dmap); + xfs_bmap_map_extent(tp, ip, XFS_DATA_FORK, dmap); qdelta += dmap->br_blockcount; } @@ -1391,6 +1394,7 @@ xfs_reflink_remap_blocks( ASSERT(nimaps == 1 && imap.br_startoff == srcoff); if (imap.br_startblock == DELAYSTARTBLOCK) { ASSERT(imap.br_startblock != DELAYSTARTBLOCK); + xfs_bmap_mark_sick(src, XFS_DATA_FORK); error = -EFSCORRUPTED; break; } diff --git a/fs/xfs/xfs_rmap_item.c b/fs/xfs/xfs_rmap_item.c index 79ad0087aeca..e473124e29cc 100644 --- a/fs/xfs/xfs_rmap_item.c +++ b/fs/xfs/xfs_rmap_item.c @@ -36,9 +36,9 @@ STATIC void xfs_rui_item_free( struct xfs_rui_log_item *ruip) { - kmem_free(ruip->rui_item.li_lv_shadow); + kvfree(ruip->rui_item.li_lv_shadow); if (ruip->rui_format.rui_nextents > XFS_RUI_MAX_FAST_EXTENTS) - kmem_free(ruip); + kfree(ruip); else kmem_cache_free(xfs_rui_cache, ruip); } @@ -142,7 +142,8 @@ xfs_rui_init( ASSERT(nextents > 0); if (nextents > XFS_RUI_MAX_FAST_EXTENTS) - ruip = kmem_zalloc(xfs_rui_log_item_sizeof(nextents), 0); + ruip = kzalloc(xfs_rui_log_item_sizeof(nextents), + GFP_KERNEL | __GFP_NOFAIL); else ruip = kmem_cache_zalloc(xfs_rui_cache, GFP_KERNEL | __GFP_NOFAIL); @@ -205,7 +206,7 @@ xfs_rud_item_release( struct xfs_rud_log_item *rudp = RUD_ITEM(lip); xfs_rui_release(rudp->rud_ruip); - kmem_free(rudp->rud_item.li_lv_shadow); + kvfree(rudp->rud_item.li_lv_shadow); kmem_cache_free(xfs_rud_cache, rudp); } @@ -454,7 +455,7 @@ xfs_rui_recover_work( { struct xfs_rmap_intent *ri; - ri = kmem_cache_alloc(xfs_rmap_intent_cache, GFP_NOFS | __GFP_NOFAIL); + ri = kmem_cache_alloc(xfs_rmap_intent_cache, GFP_KERNEL | __GFP_NOFAIL); switch (map->me_flags & XFS_RMAP_EXTENT_TYPE_MASK) { case XFS_RMAP_EXTENT_MAP: diff --git a/fs/xfs/xfs_rtalloc.c b/fs/xfs/xfs_rtalloc.c index 8649d981a097..e66f9bd5de5c 100644 --- a/fs/xfs/xfs_rtalloc.c +++ b/fs/xfs/xfs_rtalloc.c @@ -22,6 +22,8 @@ #include "xfs_sb.h" #include "xfs_rtbitmap.h" #include "xfs_quota.h" +#include "xfs_log_priv.h" +#include "xfs_health.h" /* * Return whether there are any free extents in the size range given @@ -903,7 +905,7 @@ xfs_growfs_rt( /* * Allocate a new (fake) mount/sb. */ - nmp = kmem_alloc(sizeof(*nmp), 0); + nmp = kmalloc(sizeof(*nmp), GFP_KERNEL | __GFP_NOFAIL); /* * Loop over the bitmap blocks. * We will do everything one bitmap block at a time. @@ -1050,7 +1052,7 @@ out_free: /* * Free the fake mp structure. */ - kmem_free(nmp); + kfree(nmp); /* * If we had to allocate a new rsum_cache, we either need to free the @@ -1059,10 +1061,10 @@ out_free: */ if (rsum_cache != mp->m_rsum_cache) { if (error) { - kmem_free(mp->m_rsum_cache); + kvfree(mp->m_rsum_cache); mp->m_rsum_cache = rsum_cache; } else { - kmem_free(rsum_cache); + kvfree(rsum_cache); } } @@ -1202,6 +1204,8 @@ xfs_rtmount_inodes( sbp = &mp->m_sb; error = xfs_iget(mp, NULL, sbp->sb_rbmino, 0, 0, &mp->m_rbmip); + if (xfs_metadata_is_sick(error)) + xfs_rt_mark_sick(mp, XFS_SICK_RT_BITMAP); if (error) return error; ASSERT(mp->m_rbmip != NULL); @@ -1211,6 +1215,8 @@ xfs_rtmount_inodes( goto out_rele_bitmap; error = xfs_iget(mp, NULL, sbp->sb_rsumino, 0, 0, &mp->m_rsumip); + if (xfs_metadata_is_sick(error)) + xfs_rt_mark_sick(mp, XFS_SICK_RT_SUMMARY); if (error) goto out_rele_bitmap; ASSERT(mp->m_rsumip != NULL); @@ -1233,7 +1239,7 @@ void xfs_rtunmount_inodes( struct xfs_mount *mp) { - kmem_free(mp->m_rsum_cache); + kvfree(mp->m_rsum_cache); if (mp->m_rbmip) xfs_irele(mp->m_rbmip); if (mp->m_rsumip) @@ -1260,7 +1266,7 @@ xfs_rtpick_extent( uint64_t seq; /* sequence number of file creation */ struct timespec64 ts; /* timespec in inode */ - ASSERT(xfs_isilocked(mp->m_rbmip, XFS_ILOCK_EXCL)); + xfs_assert_ilocked(mp->m_rbmip, XFS_ILOCK_EXCL); ts = inode_get_atime(VFS_I(mp->m_rbmip)); if (!(mp->m_rbmip->i_diflags & XFS_DIFLAG_NEWRTBM)) { diff --git a/fs/xfs/xfs_stats.c b/fs/xfs/xfs_stats.c index 90a77cd3ebad..ed97d72caa66 100644 --- a/fs/xfs/xfs_stats.c +++ b/fs/xfs/xfs_stats.c @@ -50,7 +50,9 @@ int xfs_stats_format(struct xfsstats __percpu *stats, char *buf) { "ibt2", xfsstats_offset(xs_fibt_2) }, { "fibt2", xfsstats_offset(xs_rmap_2) }, { "rmapbt", xfsstats_offset(xs_refcbt_2) }, - { "refcntbt", xfsstats_offset(xs_qm_dqreclaims)}, + { "refcntbt", xfsstats_offset(xs_rmap_mem_2) }, + { "rmapbt_mem", xfsstats_offset(xs_rcbag_2) }, + { "rcbagbt", xfsstats_offset(xs_qm_dqreclaims)}, /* we print both series of quota information together */ { "qm", xfsstats_offset(xs_xstrat_bytes)}, }; diff --git a/fs/xfs/xfs_stats.h b/fs/xfs/xfs_stats.h index 43ffba74f045..a61fb56ed2e6 100644 --- a/fs/xfs/xfs_stats.h +++ b/fs/xfs/xfs_stats.h @@ -125,6 +125,8 @@ struct __xfsstats { uint32_t xs_fibt_2[__XBTS_MAX]; uint32_t xs_rmap_2[__XBTS_MAX]; uint32_t xs_refcbt_2[__XBTS_MAX]; + uint32_t xs_rmap_mem_2[__XBTS_MAX]; + uint32_t xs_rcbag_2[__XBTS_MAX]; uint32_t xs_qm_dqreclaims; uint32_t xs_qm_dqreclaim_misses; uint32_t xs_qm_dquot_dups; diff --git a/fs/xfs/xfs_super.c b/fs/xfs/xfs_super.c index 59c8c0541bdd..c21f10ab0f5d 100644 --- a/fs/xfs/xfs_super.c +++ b/fs/xfs/xfs_super.c @@ -44,6 +44,7 @@ #include "xfs_dahash_test.h" #include "xfs_rtbitmap.h" #include "scrub/stats.h" +#include "scrub/rcbag_btree.h" #include <linux/magic.h> #include <linux/fs_context.h> @@ -715,9 +716,7 @@ xfs_fs_inode_init_once( /* xfs inode */ atomic_set(&ip->i_pincount, 0); spin_lock_init(&ip->i_flags_lock); - - mrlock_init(&ip->i_lock, MRLOCK_ALLOW_EQUAL_PRI|MRLOCK_BARRIER, - "xfsino", ip->i_ino); + init_rwsem(&ip->i_lock); } /* @@ -760,7 +759,7 @@ xfs_mount_free( debugfs_remove(mp->m_debugfs); kfree(mp->m_rtname); kfree(mp->m_logname); - kmem_free(mp); + kfree(mp); } STATIC int @@ -1986,7 +1985,7 @@ static int xfs_init_fs_context( { struct xfs_mount *mp; - mp = kmem_alloc(sizeof(struct xfs_mount), KM_ZERO); + mp = kzalloc(sizeof(struct xfs_mount), GFP_KERNEL | __GFP_NOFAIL); if (!mp) return -ENOMEM; @@ -2012,6 +2011,8 @@ static int xfs_init_fs_context( mp->m_logbsize = -1; mp->m_allocsize_log = 16; /* 64k */ + xfs_hooks_init(&mp->m_dir_update_hooks); + fc->s_fs_info = mp; fc->ops = &xfs_context_ops; @@ -2058,10 +2059,14 @@ xfs_init_caches(void) if (error) goto out_destroy_log_ticket_cache; - error = xfs_defer_init_item_caches(); + error = rcbagbt_init_cur_cache(); if (error) goto out_destroy_btree_cur_cache; + error = xfs_defer_init_item_caches(); + if (error) + goto out_destroy_rcbagbt_cur_cache; + xfs_da_state_cache = kmem_cache_create("xfs_da_state", sizeof(struct xfs_da_state), 0, 0, NULL); @@ -2218,6 +2223,8 @@ xfs_init_caches(void) kmem_cache_destroy(xfs_da_state_cache); out_destroy_defer_item_cache: xfs_defer_destroy_item_caches(); + out_destroy_rcbagbt_cur_cache: + rcbagbt_destroy_cur_cache(); out_destroy_btree_cur_cache: xfs_btree_destroy_cur_caches(); out_destroy_log_ticket_cache: @@ -2255,6 +2262,7 @@ xfs_destroy_caches(void) kmem_cache_destroy(xfs_ifork_cache); kmem_cache_destroy(xfs_da_state_cache); xfs_defer_destroy_item_caches(); + rcbagbt_destroy_cur_cache(); xfs_btree_destroy_cur_caches(); kmem_cache_destroy(xfs_log_ticket_cache); kmem_cache_destroy(xfs_buf_cache); diff --git a/fs/xfs/xfs_symlink.c b/fs/xfs/xfs_symlink.c index 92974a4414c8..3e376d24c7c1 100644 --- a/fs/xfs/xfs_symlink.c +++ b/fs/xfs/xfs_symlink.c @@ -24,77 +24,7 @@ #include "xfs_ialloc.h" #include "xfs_error.h" #include "xfs_health.h" - -/* ----- Kernel only functions below ----- */ -int -xfs_readlink_bmap_ilocked( - struct xfs_inode *ip, - char *link) -{ - struct xfs_mount *mp = ip->i_mount; - struct xfs_bmbt_irec mval[XFS_SYMLINK_MAPS]; - struct xfs_buf *bp; - xfs_daddr_t d; - char *cur_chunk; - int pathlen = ip->i_disk_size; - int nmaps = XFS_SYMLINK_MAPS; - int byte_cnt; - int n; - int error = 0; - int fsblocks = 0; - int offset; - - ASSERT(xfs_isilocked(ip, XFS_ILOCK_SHARED | XFS_ILOCK_EXCL)); - - fsblocks = xfs_symlink_blocks(mp, pathlen); - error = xfs_bmapi_read(ip, 0, fsblocks, mval, &nmaps, 0); - if (error) - goto out; - - offset = 0; - for (n = 0; n < nmaps; n++) { - d = XFS_FSB_TO_DADDR(mp, mval[n].br_startblock); - byte_cnt = XFS_FSB_TO_B(mp, mval[n].br_blockcount); - - error = xfs_buf_read(mp->m_ddev_targp, d, BTOBB(byte_cnt), 0, - &bp, &xfs_symlink_buf_ops); - if (error) - return error; - byte_cnt = XFS_SYMLINK_BUF_SPACE(mp, byte_cnt); - if (pathlen < byte_cnt) - byte_cnt = pathlen; - - cur_chunk = bp->b_addr; - if (xfs_has_crc(mp)) { - if (!xfs_symlink_hdr_ok(ip->i_ino, offset, - byte_cnt, bp)) { - error = -EFSCORRUPTED; - xfs_alert(mp, -"symlink header does not match required off/len/owner (0x%x/Ox%x,0x%llx)", - offset, byte_cnt, ip->i_ino); - xfs_buf_relse(bp); - goto out; - - } - - cur_chunk += sizeof(struct xfs_dsymlink_hdr); - } - - memcpy(link + offset, cur_chunk, byte_cnt); - - pathlen -= byte_cnt; - offset += byte_cnt; - - xfs_buf_relse(bp); - } - ASSERT(pathlen == 0); - - link[ip->i_disk_size] = '\0'; - error = 0; - - out: - return error; -} +#include "xfs_symlink_remote.h" int xfs_readlink( @@ -103,7 +33,7 @@ xfs_readlink( { struct xfs_mount *mp = ip->i_mount; xfs_fsize_t pathlen; - int error = -EFSCORRUPTED; + int error; trace_xfs_readlink(ip); @@ -116,14 +46,14 @@ xfs_readlink( pathlen = ip->i_disk_size; if (!pathlen) - goto out; + goto out_corrupt; if (pathlen < 0 || pathlen > XFS_SYMLINK_MAXLEN) { xfs_alert(mp, "%s: inode (%llu) bad symlink length (%lld)", __func__, (unsigned long long) ip->i_ino, (long long) pathlen); ASSERT(0); - goto out; + goto out_corrupt; } if (ip->i_df.if_format == XFS_DINODE_FMT_LOCAL) { @@ -132,17 +62,20 @@ xfs_readlink( * if if_data is junk. */ if (XFS_IS_CORRUPT(ip->i_mount, !ip->i_df.if_data)) - goto out; + goto out_corrupt; memcpy(link, ip->i_df.if_data, pathlen + 1); error = 0; } else { - error = xfs_readlink_bmap_ilocked(ip, link); + error = xfs_symlink_remote_read(ip, link); } - out: xfs_iunlock(ip, XFS_ILOCK_SHARED); return error; + out_corrupt: + xfs_iunlock(ip, XFS_ILOCK_SHARED); + xfs_inode_mark_sick(ip, XFS_SICK_INO_SYMLINK); + return -EFSCORRUPTED; } int @@ -160,15 +93,7 @@ xfs_symlink( int error = 0; int pathlen; bool unlock_dp_on_error = false; - xfs_fileoff_t first_fsb; xfs_filblks_t fs_blocks; - int nmaps; - struct xfs_bmbt_irec mval[XFS_SYMLINK_MAPS]; - xfs_daddr_t d; - const char *cur_chunk; - int byte_cnt; - int n; - struct xfs_buf *bp; prid_t prid; struct xfs_dquot *udqp = NULL; struct xfs_dquot *gdqp = NULL; @@ -256,62 +181,11 @@ xfs_symlink( xfs_qm_vop_create_dqattach(tp, ip, udqp, gdqp, pdqp); resblks -= XFS_IALLOC_SPACE_RES(mp); - /* - * If the symlink will fit into the inode, write it inline. - */ - if (pathlen <= xfs_inode_data_fork_size(ip)) { - xfs_init_local_fork(ip, XFS_DATA_FORK, target_path, pathlen); - - ip->i_disk_size = pathlen; - ip->i_df.if_format = XFS_DINODE_FMT_LOCAL; - xfs_trans_log_inode(tp, ip, XFS_ILOG_DDATA | XFS_ILOG_CORE); - } else { - int offset; - - first_fsb = 0; - nmaps = XFS_SYMLINK_MAPS; - - error = xfs_bmapi_write(tp, ip, first_fsb, fs_blocks, - XFS_BMAPI_METADATA, resblks, mval, &nmaps); - if (error) - goto out_trans_cancel; - - resblks -= fs_blocks; - ip->i_disk_size = pathlen; - xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE); - - cur_chunk = target_path; - offset = 0; - for (n = 0; n < nmaps; n++) { - char *buf; - - d = XFS_FSB_TO_DADDR(mp, mval[n].br_startblock); - byte_cnt = XFS_FSB_TO_B(mp, mval[n].br_blockcount); - error = xfs_trans_get_buf(tp, mp->m_ddev_targp, d, - BTOBB(byte_cnt), 0, &bp); - if (error) - goto out_trans_cancel; - bp->b_ops = &xfs_symlink_buf_ops; - - byte_cnt = XFS_SYMLINK_BUF_SPACE(mp, byte_cnt); - byte_cnt = min(byte_cnt, pathlen); - - buf = bp->b_addr; - buf += xfs_symlink_hdr_set(mp, ip->i_ino, offset, - byte_cnt, bp); - - memcpy(buf, cur_chunk, byte_cnt); - - cur_chunk += byte_cnt; - pathlen -= byte_cnt; - offset += byte_cnt; - - xfs_trans_buf_set_type(tp, bp, XFS_BLFT_SYMLINK_BUF); - xfs_trans_log_buf(tp, bp, 0, (buf + byte_cnt - 1) - - (char *)bp->b_addr); - } - ASSERT(pathlen == 0); - } + error = xfs_symlink_write_target(tp, ip, target_path, pathlen, + fs_blocks, resblks); + if (error) + goto out_trans_cancel; + resblks -= fs_blocks; i_size_write(VFS_I(ip), ip->i_disk_size); /* @@ -322,6 +196,7 @@ xfs_symlink( goto out_trans_cancel; xfs_trans_ichgtime(tp, dp, XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG); xfs_trans_log_inode(tp, dp, XFS_ILOG_CORE); + xfs_dir_update_hook(dp, ip, 1, link_name); /* * If this is a synchronous mount, make sure that the @@ -496,6 +371,7 @@ xfs_inactive_symlink( __func__, (unsigned long long)ip->i_ino, pathlen); xfs_iunlock(ip, XFS_ILOCK_EXCL); ASSERT(0); + xfs_inode_mark_sick(ip, XFS_SICK_INO_SYMLINK); return -EFSCORRUPTED; } diff --git a/fs/xfs/xfs_symlink.h b/fs/xfs/xfs_symlink.h index d1ca1ce62a93..0d29a50e66fd 100644 --- a/fs/xfs/xfs_symlink.h +++ b/fs/xfs/xfs_symlink.h @@ -10,7 +10,6 @@ int xfs_symlink(struct mnt_idmap *idmap, struct xfs_inode *dp, struct xfs_name *link_name, const char *target_path, umode_t mode, struct xfs_inode **ipp); -int xfs_readlink_bmap_ilocked(struct xfs_inode *ip, char *link); int xfs_readlink(struct xfs_inode *ip, char *link); int xfs_inactive_symlink(struct xfs_inode *ip); diff --git a/fs/xfs/xfs_sysfs.c b/fs/xfs/xfs_sysfs.c index 17485666b672..d2391eec37fe 100644 --- a/fs/xfs/xfs_sysfs.c +++ b/fs/xfs/xfs_sysfs.c @@ -193,7 +193,6 @@ always_cow_show( } XFS_SYSFS_ATTR_RW(always_cow); -#ifdef DEBUG /* * Override how many threads the parallel work queue is allowed to create. * This has to be a debug-only global (instead of an errortag) because one of @@ -260,7 +259,6 @@ larp_show( return snprintf(buf, PAGE_SIZE, "%d\n", xfs_globals.larp); } XFS_SYSFS_ATTR_RW(larp); -#endif /* DEBUG */ STATIC ssize_t bload_leaf_slack_store( @@ -319,10 +317,8 @@ static struct attribute *xfs_dbg_attrs[] = { ATTR_LIST(log_recovery_delay), ATTR_LIST(mount_delay), ATTR_LIST(always_cow), -#ifdef DEBUG ATTR_LIST(pwork_threads), ATTR_LIST(larp), -#endif ATTR_LIST(bload_leaf_slack), ATTR_LIST(bload_node_slack), NULL, diff --git a/fs/xfs/xfs_trace.c b/fs/xfs/xfs_trace.c index 8a5dc1538aa8..1a963382e5e9 100644 --- a/fs/xfs/xfs_trace.c +++ b/fs/xfs/xfs_trace.c @@ -36,6 +36,9 @@ #include "xfs_error.h" #include <linux/iomap.h> #include "xfs_iomap.h" +#include "xfs_buf_mem.h" +#include "xfs_btree_mem.h" +#include "xfs_bmap.h" /* * We include this last to have the helpers above available for the trace diff --git a/fs/xfs/xfs_trace.h b/fs/xfs/xfs_trace.h index 0984a1c884c7..56b07d8ed431 100644 --- a/fs/xfs/xfs_trace.h +++ b/fs/xfs/xfs_trace.h @@ -79,6 +79,9 @@ union xfs_btree_ptr; struct xfs_dqtrx; struct xfs_icwalk; struct xfs_perag; +struct xfbtree; +struct xfs_btree_ops; +struct xfs_bmap_intent; #define XFS_ATTR_FILTER_FLAGS \ { XFS_ATTR_ROOT, "ROOT" }, \ @@ -640,6 +643,7 @@ DEFINE_BUF_ITEM_EVENT(xfs_trans_read_buf); DEFINE_BUF_ITEM_EVENT(xfs_trans_read_buf_recur); DEFINE_BUF_ITEM_EVENT(xfs_trans_log_buf); DEFINE_BUF_ITEM_EVENT(xfs_trans_brelse); +DEFINE_BUF_ITEM_EVENT(xfs_trans_bdetach); DEFINE_BUF_ITEM_EVENT(xfs_trans_bjoin); DEFINE_BUF_ITEM_EVENT(xfs_trans_bhold); DEFINE_BUF_ITEM_EVENT(xfs_trans_bhold_release); @@ -1710,12 +1714,10 @@ DECLARE_EVENT_CLASS(xfs_agf_class, __entry->agno = be32_to_cpu(agf->agf_seqno), __entry->flags = flags; __entry->length = be32_to_cpu(agf->agf_length), - __entry->bno_root = be32_to_cpu(agf->agf_roots[XFS_BTNUM_BNO]), - __entry->cnt_root = be32_to_cpu(agf->agf_roots[XFS_BTNUM_CNT]), - __entry->bno_level = - be32_to_cpu(agf->agf_levels[XFS_BTNUM_BNO]), - __entry->cnt_level = - be32_to_cpu(agf->agf_levels[XFS_BTNUM_CNT]), + __entry->bno_root = be32_to_cpu(agf->agf_bno_root), + __entry->cnt_root = be32_to_cpu(agf->agf_cnt_root), + __entry->bno_level = be32_to_cpu(agf->agf_bno_level), + __entry->cnt_level = be32_to_cpu(agf->agf_cnt_level), __entry->flfirst = be32_to_cpu(agf->agf_flfirst), __entry->fllast = be32_to_cpu(agf->agf_fllast), __entry->flcount = be32_to_cpu(agf->agf_flcount), @@ -1890,28 +1892,28 @@ DEFINE_ALLOC_EVENT(xfs_alloc_vextent_near_bno); DEFINE_ALLOC_EVENT(xfs_alloc_vextent_finish); TRACE_EVENT(xfs_alloc_cur_check, - TP_PROTO(struct xfs_mount *mp, xfs_btnum_t btnum, xfs_agblock_t bno, + TP_PROTO(struct xfs_btree_cur *cur, xfs_agblock_t bno, xfs_extlen_t len, xfs_extlen_t diff, bool new), - TP_ARGS(mp, btnum, bno, len, diff, new), + TP_ARGS(cur, bno, len, diff, new), TP_STRUCT__entry( __field(dev_t, dev) - __field(xfs_btnum_t, btnum) + __string(name, cur->bc_ops->name) __field(xfs_agblock_t, bno) __field(xfs_extlen_t, len) __field(xfs_extlen_t, diff) __field(bool, new) ), TP_fast_assign( - __entry->dev = mp->m_super->s_dev; - __entry->btnum = btnum; + __entry->dev = cur->bc_mp->m_super->s_dev; + __assign_str(name, cur->bc_ops->name); __entry->bno = bno; __entry->len = len; __entry->diff = diff; __entry->new = new; ), - TP_printk("dev %d:%d btree %s agbno 0x%x fsbcount 0x%x diff 0x%x new %d", + TP_printk("dev %d:%d %sbt agbno 0x%x fsbcount 0x%x diff 0x%x new %d", MAJOR(__entry->dev), MINOR(__entry->dev), - __print_symbolic(__entry->btnum, XFS_BTNUM_STRINGS), + __get_str(name), __entry->bno, __entry->len, __entry->diff, __entry->new) ) @@ -2452,21 +2454,12 @@ DEFINE_DISCARD_EVENT(xfs_discard_toosmall); DEFINE_DISCARD_EVENT(xfs_discard_exclude); DEFINE_DISCARD_EVENT(xfs_discard_busy); -/* btree cursor events */ -TRACE_DEFINE_ENUM(XFS_BTNUM_BNOi); -TRACE_DEFINE_ENUM(XFS_BTNUM_CNTi); -TRACE_DEFINE_ENUM(XFS_BTNUM_BMAPi); -TRACE_DEFINE_ENUM(XFS_BTNUM_INOi); -TRACE_DEFINE_ENUM(XFS_BTNUM_FINOi); -TRACE_DEFINE_ENUM(XFS_BTNUM_RMAPi); -TRACE_DEFINE_ENUM(XFS_BTNUM_REFCi); - DECLARE_EVENT_CLASS(xfs_btree_cur_class, TP_PROTO(struct xfs_btree_cur *cur, int level, struct xfs_buf *bp), TP_ARGS(cur, level, bp), TP_STRUCT__entry( __field(dev_t, dev) - __field(xfs_btnum_t, btnum) + __string(name, cur->bc_ops->name) __field(int, level) __field(int, nlevels) __field(int, ptr) @@ -2474,15 +2467,15 @@ DECLARE_EVENT_CLASS(xfs_btree_cur_class, ), TP_fast_assign( __entry->dev = cur->bc_mp->m_super->s_dev; - __entry->btnum = cur->bc_btnum; + __assign_str(name, cur->bc_ops->name); __entry->level = level; __entry->nlevels = cur->bc_nlevels; __entry->ptr = cur->bc_levels[level].ptr; __entry->daddr = bp ? xfs_buf_daddr(bp) : -1; ), - TP_printk("dev %d:%d btree %s level %d/%d ptr %d daddr 0x%llx", + TP_printk("dev %d:%d %sbt level %d/%d ptr %d daddr 0x%llx", MAJOR(__entry->dev), MINOR(__entry->dev), - __print_symbolic(__entry->btnum, XFS_BTNUM_STRINGS), + __get_str(name), __entry->level, __entry->nlevels, __entry->ptr, @@ -2496,6 +2489,90 @@ DEFINE_EVENT(xfs_btree_cur_class, name, \ DEFINE_BTREE_CUR_EVENT(xfs_btree_updkeys); DEFINE_BTREE_CUR_EVENT(xfs_btree_overlapped_query_range); +TRACE_EVENT(xfs_btree_alloc_block, + TP_PROTO(struct xfs_btree_cur *cur, union xfs_btree_ptr *ptr, int stat, + int error), + TP_ARGS(cur, ptr, stat, error), + TP_STRUCT__entry( + __field(dev_t, dev) + __field(xfs_agnumber_t, agno) + __field(xfs_ino_t, ino) + __string(name, cur->bc_ops->name) + __field(int, error) + __field(xfs_agblock_t, agbno) + ), + TP_fast_assign( + __entry->dev = cur->bc_mp->m_super->s_dev; + switch (cur->bc_ops->type) { + case XFS_BTREE_TYPE_INODE: + __entry->agno = 0; + __entry->ino = cur->bc_ino.ip->i_ino; + break; + case XFS_BTREE_TYPE_AG: + __entry->agno = cur->bc_ag.pag->pag_agno; + __entry->ino = 0; + break; + case XFS_BTREE_TYPE_MEM: + __entry->agno = 0; + __entry->ino = 0; + break; + } + __assign_str(name, cur->bc_ops->name); + __entry->error = error; + if (!error && stat) { + if (cur->bc_ops->ptr_len == XFS_BTREE_LONG_PTR_LEN) { + xfs_fsblock_t fsb = be64_to_cpu(ptr->l); + + __entry->agno = XFS_FSB_TO_AGNO(cur->bc_mp, + fsb); + __entry->agbno = XFS_FSB_TO_AGBNO(cur->bc_mp, + fsb); + } else { + __entry->agbno = be32_to_cpu(ptr->s); + } + } else { + __entry->agbno = NULLAGBLOCK; + } + ), + TP_printk("dev %d:%d %sbt agno 0x%x ino 0x%llx agbno 0x%x error %d", + MAJOR(__entry->dev), MINOR(__entry->dev), + __get_str(name), + __entry->agno, + __entry->ino, + __entry->agbno, + __entry->error) +); + +TRACE_EVENT(xfs_btree_free_block, + TP_PROTO(struct xfs_btree_cur *cur, struct xfs_buf *bp), + TP_ARGS(cur, bp), + TP_STRUCT__entry( + __field(dev_t, dev) + __field(xfs_agnumber_t, agno) + __field(xfs_ino_t, ino) + __string(name, cur->bc_ops->name) + __field(xfs_agblock_t, agbno) + ), + TP_fast_assign( + __entry->dev = cur->bc_mp->m_super->s_dev; + __entry->agno = xfs_daddr_to_agno(cur->bc_mp, + xfs_buf_daddr(bp)); + if (cur->bc_ops->type == XFS_BTREE_TYPE_INODE) + __entry->ino = cur->bc_ino.ip->i_ino; + else + __entry->ino = 0; + __assign_str(name, cur->bc_ops->name); + __entry->agbno = xfs_daddr_to_agbno(cur->bc_mp, + xfs_buf_daddr(bp)); + ), + TP_printk("dev %d:%d %sbt agno 0x%x ino 0x%llx agbno 0x%x", + MAJOR(__entry->dev), MINOR(__entry->dev), + __get_str(name), + __entry->agno, + __entry->ino, + __entry->agbno) +); + /* deferred ops */ struct xfs_defer_pending; @@ -2579,7 +2656,25 @@ DEFINE_EVENT(xfs_defer_pending_class, name, \ TP_PROTO(struct xfs_mount *mp, struct xfs_defer_pending *dfp), \ TP_ARGS(mp, dfp)) -DECLARE_EVENT_CLASS(xfs_phys_extent_deferred_class, +DEFINE_DEFER_EVENT(xfs_defer_cancel); +DEFINE_DEFER_EVENT(xfs_defer_trans_roll); +DEFINE_DEFER_EVENT(xfs_defer_trans_abort); +DEFINE_DEFER_EVENT(xfs_defer_finish); +DEFINE_DEFER_EVENT(xfs_defer_finish_done); + +DEFINE_DEFER_ERROR_EVENT(xfs_defer_trans_roll_error); +DEFINE_DEFER_ERROR_EVENT(xfs_defer_finish_error); + +DEFINE_DEFER_PENDING_EVENT(xfs_defer_create_intent); +DEFINE_DEFER_PENDING_EVENT(xfs_defer_cancel_list); +DEFINE_DEFER_PENDING_EVENT(xfs_defer_pending_finish); +DEFINE_DEFER_PENDING_EVENT(xfs_defer_pending_abort); +DEFINE_DEFER_PENDING_EVENT(xfs_defer_relog_intent); +DEFINE_DEFER_PENDING_EVENT(xfs_defer_isolate_paused); +DEFINE_DEFER_PENDING_EVENT(xfs_defer_item_pause); +DEFINE_DEFER_PENDING_EVENT(xfs_defer_item_unpause); + +DECLARE_EVENT_CLASS(xfs_free_extent_deferred_class, TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno, int type, xfs_agblock_t agbno, xfs_extlen_t len), TP_ARGS(mp, agno, type, agbno, len), @@ -2604,92 +2699,17 @@ DECLARE_EVENT_CLASS(xfs_phys_extent_deferred_class, __entry->agbno, __entry->len) ); -#define DEFINE_PHYS_EXTENT_DEFERRED_EVENT(name) \ -DEFINE_EVENT(xfs_phys_extent_deferred_class, name, \ +#define DEFINE_FREE_EXTENT_DEFERRED_EVENT(name) \ +DEFINE_EVENT(xfs_free_extent_deferred_class, name, \ TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno, \ int type, \ xfs_agblock_t bno, \ xfs_extlen_t len), \ TP_ARGS(mp, agno, type, bno, len)) - -DECLARE_EVENT_CLASS(xfs_map_extent_deferred_class, - TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno, - int op, - xfs_agblock_t agbno, - xfs_ino_t ino, - int whichfork, - xfs_fileoff_t offset, - xfs_filblks_t len, - xfs_exntst_t state), - TP_ARGS(mp, agno, op, agbno, ino, whichfork, offset, len, state), - TP_STRUCT__entry( - __field(dev_t, dev) - __field(xfs_agnumber_t, agno) - __field(xfs_ino_t, ino) - __field(xfs_agblock_t, agbno) - __field(int, whichfork) - __field(xfs_fileoff_t, l_loff) - __field(xfs_filblks_t, l_len) - __field(xfs_exntst_t, l_state) - __field(int, op) - ), - TP_fast_assign( - __entry->dev = mp->m_super->s_dev; - __entry->agno = agno; - __entry->ino = ino; - __entry->agbno = agbno; - __entry->whichfork = whichfork; - __entry->l_loff = offset; - __entry->l_len = len; - __entry->l_state = state; - __entry->op = op; - ), - TP_printk("dev %d:%d op %d agno 0x%x agbno 0x%x owner 0x%llx %s fileoff 0x%llx fsbcount 0x%llx state %d", - MAJOR(__entry->dev), MINOR(__entry->dev), - __entry->op, - __entry->agno, - __entry->agbno, - __entry->ino, - __print_symbolic(__entry->whichfork, XFS_WHICHFORK_STRINGS), - __entry->l_loff, - __entry->l_len, - __entry->l_state) -); -#define DEFINE_MAP_EXTENT_DEFERRED_EVENT(name) \ -DEFINE_EVENT(xfs_map_extent_deferred_class, name, \ - TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno, \ - int op, \ - xfs_agblock_t agbno, \ - xfs_ino_t ino, \ - int whichfork, \ - xfs_fileoff_t offset, \ - xfs_filblks_t len, \ - xfs_exntst_t state), \ - TP_ARGS(mp, agno, op, agbno, ino, whichfork, offset, len, state)) - -DEFINE_DEFER_EVENT(xfs_defer_cancel); -DEFINE_DEFER_EVENT(xfs_defer_trans_roll); -DEFINE_DEFER_EVENT(xfs_defer_trans_abort); -DEFINE_DEFER_EVENT(xfs_defer_finish); -DEFINE_DEFER_EVENT(xfs_defer_finish_done); - -DEFINE_DEFER_ERROR_EVENT(xfs_defer_trans_roll_error); -DEFINE_DEFER_ERROR_EVENT(xfs_defer_finish_error); - -DEFINE_DEFER_PENDING_EVENT(xfs_defer_create_intent); -DEFINE_DEFER_PENDING_EVENT(xfs_defer_cancel_list); -DEFINE_DEFER_PENDING_EVENT(xfs_defer_pending_finish); -DEFINE_DEFER_PENDING_EVENT(xfs_defer_pending_abort); -DEFINE_DEFER_PENDING_EVENT(xfs_defer_relog_intent); -DEFINE_DEFER_PENDING_EVENT(xfs_defer_isolate_paused); -DEFINE_DEFER_PENDING_EVENT(xfs_defer_item_pause); -DEFINE_DEFER_PENDING_EVENT(xfs_defer_item_unpause); - -#define DEFINE_BMAP_FREE_DEFERRED_EVENT DEFINE_PHYS_EXTENT_DEFERRED_EVENT -DEFINE_BMAP_FREE_DEFERRED_EVENT(xfs_bmap_free_defer); -DEFINE_BMAP_FREE_DEFERRED_EVENT(xfs_bmap_free_deferred); -DEFINE_BMAP_FREE_DEFERRED_EVENT(xfs_agfl_free_defer); -DEFINE_BMAP_FREE_DEFERRED_EVENT(xfs_agfl_free_deferred); +DEFINE_FREE_EXTENT_DEFERRED_EVENT(xfs_bmap_free_defer); +DEFINE_FREE_EXTENT_DEFERRED_EVENT(xfs_bmap_free_deferred); +DEFINE_FREE_EXTENT_DEFERRED_EVENT(xfs_agfl_free_defer); +DEFINE_FREE_EXTENT_DEFERRED_EVENT(xfs_agfl_free_deferred); DECLARE_EVENT_CLASS(xfs_defer_pending_item_class, TP_PROTO(struct xfs_mount *mp, struct xfs_defer_pending *dfp, @@ -2854,12 +2874,63 @@ DEFINE_EVENT(xfs_rmapbt_class, name, \ uint64_t owner, uint64_t offset, unsigned int flags), \ TP_ARGS(mp, agno, agbno, len, owner, offset, flags)) -#define DEFINE_RMAP_DEFERRED_EVENT DEFINE_MAP_EXTENT_DEFERRED_EVENT +DECLARE_EVENT_CLASS(xfs_rmap_deferred_class, + TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno, + int op, + xfs_agblock_t agbno, + xfs_ino_t ino, + int whichfork, + xfs_fileoff_t offset, + xfs_filblks_t len, + xfs_exntst_t state), + TP_ARGS(mp, agno, op, agbno, ino, whichfork, offset, len, state), + TP_STRUCT__entry( + __field(dev_t, dev) + __field(xfs_agnumber_t, agno) + __field(xfs_ino_t, ino) + __field(xfs_agblock_t, agbno) + __field(int, whichfork) + __field(xfs_fileoff_t, l_loff) + __field(xfs_filblks_t, l_len) + __field(xfs_exntst_t, l_state) + __field(int, op) + ), + TP_fast_assign( + __entry->dev = mp->m_super->s_dev; + __entry->agno = agno; + __entry->ino = ino; + __entry->agbno = agbno; + __entry->whichfork = whichfork; + __entry->l_loff = offset; + __entry->l_len = len; + __entry->l_state = state; + __entry->op = op; + ), + TP_printk("dev %d:%d op %d agno 0x%x agbno 0x%x owner 0x%llx %s fileoff 0x%llx fsbcount 0x%llx state %d", + MAJOR(__entry->dev), MINOR(__entry->dev), + __entry->op, + __entry->agno, + __entry->agbno, + __entry->ino, + __print_symbolic(__entry->whichfork, XFS_WHICHFORK_STRINGS), + __entry->l_loff, + __entry->l_len, + __entry->l_state) +); +#define DEFINE_RMAP_DEFERRED_EVENT(name) \ +DEFINE_EVENT(xfs_rmap_deferred_class, name, \ + TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno, \ + int op, \ + xfs_agblock_t agbno, \ + xfs_ino_t ino, \ + int whichfork, \ + xfs_fileoff_t offset, \ + xfs_filblks_t len, \ + xfs_exntst_t state), \ + TP_ARGS(mp, agno, op, agbno, ino, whichfork, offset, len, state)) DEFINE_RMAP_DEFERRED_EVENT(xfs_rmap_defer); DEFINE_RMAP_DEFERRED_EVENT(xfs_rmap_deferred); -DEFINE_BUSY_EVENT(xfs_rmapbt_alloc_block); -DEFINE_BUSY_EVENT(xfs_rmapbt_free_block); DEFINE_RMAPBT_EVENT(xfs_rmap_update); DEFINE_RMAPBT_EVENT(xfs_rmap_insert); DEFINE_RMAPBT_EVENT(xfs_rmap_delete); @@ -2876,7 +2947,66 @@ DEFINE_RMAPBT_EVENT(xfs_rmap_find_right_neighbor_result); DEFINE_RMAPBT_EVENT(xfs_rmap_find_left_neighbor_result); /* deferred bmbt updates */ -#define DEFINE_BMAP_DEFERRED_EVENT DEFINE_RMAP_DEFERRED_EVENT +TRACE_DEFINE_ENUM(XFS_BMAP_MAP); +TRACE_DEFINE_ENUM(XFS_BMAP_UNMAP); + +DECLARE_EVENT_CLASS(xfs_bmap_deferred_class, + TP_PROTO(struct xfs_bmap_intent *bi), + TP_ARGS(bi), + TP_STRUCT__entry( + __field(dev_t, dev) + __field(dev_t, opdev) + __field(xfs_agnumber_t, agno) + __field(xfs_ino_t, ino) + __field(xfs_agblock_t, agbno) + __field(xfs_fsblock_t, rtbno) + __field(int, whichfork) + __field(xfs_fileoff_t, l_loff) + __field(xfs_filblks_t, l_len) + __field(xfs_exntst_t, l_state) + __field(int, op) + ), + TP_fast_assign( + struct xfs_inode *ip = bi->bi_owner; + + __entry->dev = ip->i_mount->m_super->s_dev; + if (xfs_ifork_is_realtime(ip, bi->bi_whichfork)) { + __entry->agno = 0; + __entry->agbno = 0; + __entry->rtbno = bi->bi_bmap.br_startblock; + __entry->opdev = ip->i_mount->m_rtdev_targp->bt_dev; + } else { + __entry->agno = XFS_FSB_TO_AGNO(ip->i_mount, + bi->bi_bmap.br_startblock); + __entry->agbno = XFS_FSB_TO_AGBNO(ip->i_mount, + bi->bi_bmap.br_startblock); + __entry->rtbno = 0; + __entry->opdev = __entry->dev; + } + __entry->ino = ip->i_ino; + __entry->whichfork = bi->bi_whichfork; + __entry->l_loff = bi->bi_bmap.br_startoff; + __entry->l_len = bi->bi_bmap.br_blockcount; + __entry->l_state = bi->bi_bmap.br_state; + __entry->op = bi->bi_type; + ), + TP_printk("dev %d:%d op %s opdev %d:%d ino 0x%llx agno 0x%x agbno 0x%x rtbno 0x%llx %s fileoff 0x%llx fsbcount 0x%llx state %d", + MAJOR(__entry->dev), MINOR(__entry->dev), + __print_symbolic(__entry->op, XFS_BMAP_INTENT_STRINGS), + MAJOR(__entry->opdev), MINOR(__entry->opdev), + __entry->ino, + __entry->agno, + __entry->agbno, + __entry->rtbno, + __print_symbolic(__entry->whichfork, XFS_WHICHFORK_STRINGS), + __entry->l_loff, + __entry->l_len, + __entry->l_state) +); +#define DEFINE_BMAP_DEFERRED_EVENT(name) \ +DEFINE_EVENT(xfs_bmap_deferred_class, name, \ + TP_PROTO(struct xfs_bmap_intent *bi), \ + TP_ARGS(bi)) DEFINE_BMAP_DEFERRED_EVENT(xfs_bmap_defer); DEFINE_BMAP_DEFERRED_EVENT(xfs_bmap_deferred); @@ -3217,8 +3347,6 @@ DEFINE_EVENT(xfs_refcount_triple_extent_class, name, \ TP_ARGS(mp, agno, i1, i2, i3)) /* refcount btree tracepoints */ -DEFINE_BUSY_EVENT(xfs_refcountbt_alloc_block); -DEFINE_BUSY_EVENT(xfs_refcountbt_free_block); DEFINE_AG_BTREE_LOOKUP_EVENT(xfs_refcount_lookup); DEFINE_REFCOUNT_EXTENT_EVENT(xfs_refcount_get); DEFINE_REFCOUNT_EXTENT_EVENT(xfs_refcount_update); @@ -3255,7 +3383,39 @@ DEFINE_AG_ERROR_EVENT(xfs_refcount_find_right_extent_error); DEFINE_AG_EXTENT_EVENT(xfs_refcount_find_shared); DEFINE_AG_EXTENT_EVENT(xfs_refcount_find_shared_result); DEFINE_AG_ERROR_EVENT(xfs_refcount_find_shared_error); -#define DEFINE_REFCOUNT_DEFERRED_EVENT DEFINE_PHYS_EXTENT_DEFERRED_EVENT + +DECLARE_EVENT_CLASS(xfs_refcount_deferred_class, + TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno, + int type, xfs_agblock_t agbno, xfs_extlen_t len), + TP_ARGS(mp, agno, type, agbno, len), + TP_STRUCT__entry( + __field(dev_t, dev) + __field(xfs_agnumber_t, agno) + __field(int, type) + __field(xfs_agblock_t, agbno) + __field(xfs_extlen_t, len) + ), + TP_fast_assign( + __entry->dev = mp->m_super->s_dev; + __entry->agno = agno; + __entry->type = type; + __entry->agbno = agbno; + __entry->len = len; + ), + TP_printk("dev %d:%d op %d agno 0x%x agbno 0x%x fsbcount 0x%x", + MAJOR(__entry->dev), MINOR(__entry->dev), + __entry->type, + __entry->agno, + __entry->agbno, + __entry->len) +); +#define DEFINE_REFCOUNT_DEFERRED_EVENT(name) \ +DEFINE_EVENT(xfs_refcount_deferred_class, name, \ + TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno, \ + int type, \ + xfs_agblock_t bno, \ + xfs_extlen_t len), \ + TP_ARGS(mp, agno, type, bno, len)) DEFINE_REFCOUNT_DEFERRED_EVENT(xfs_refcount_defer); DEFINE_REFCOUNT_DEFERRED_EVENT(xfs_refcount_deferred); @@ -3926,9 +4086,11 @@ DEFINE_EVENT(xfs_fs_corrupt_class, name, \ TP_PROTO(struct xfs_mount *mp, unsigned int flags), \ TP_ARGS(mp, flags)) DEFINE_FS_CORRUPT_EVENT(xfs_fs_mark_sick); +DEFINE_FS_CORRUPT_EVENT(xfs_fs_mark_corrupt); DEFINE_FS_CORRUPT_EVENT(xfs_fs_mark_healthy); DEFINE_FS_CORRUPT_EVENT(xfs_fs_unfixed_corruption); DEFINE_FS_CORRUPT_EVENT(xfs_rt_mark_sick); +DEFINE_FS_CORRUPT_EVENT(xfs_rt_mark_corrupt); DEFINE_FS_CORRUPT_EVENT(xfs_rt_mark_healthy); DEFINE_FS_CORRUPT_EVENT(xfs_rt_unfixed_corruption); @@ -3955,6 +4117,7 @@ DEFINE_EVENT(xfs_ag_corrupt_class, name, \ unsigned int flags), \ TP_ARGS(mp, agno, flags)) DEFINE_AG_CORRUPT_EVENT(xfs_ag_mark_sick); +DEFINE_AG_CORRUPT_EVENT(xfs_ag_mark_corrupt); DEFINE_AG_CORRUPT_EVENT(xfs_ag_mark_healthy); DEFINE_AG_CORRUPT_EVENT(xfs_ag_unfixed_corruption); @@ -3980,7 +4143,9 @@ DEFINE_EVENT(xfs_inode_corrupt_class, name, \ TP_PROTO(struct xfs_inode *ip, unsigned int flags), \ TP_ARGS(ip, flags)) DEFINE_INODE_CORRUPT_EVENT(xfs_inode_mark_sick); +DEFINE_INODE_CORRUPT_EVENT(xfs_inode_mark_corrupt); DEFINE_INODE_CORRUPT_EVENT(xfs_inode_mark_healthy); +DEFINE_INODE_CORRUPT_EVENT(xfs_inode_unfixed_corruption); TRACE_EVENT(xfs_iwalk_ag, TP_PROTO(struct xfs_mount *mp, xfs_agnumber_t agno, @@ -4040,31 +4205,6 @@ TRACE_EVENT(xfs_pwork_init, __entry->nr_threads, __entry->pid) ) -DECLARE_EVENT_CLASS(xfs_kmem_class, - TP_PROTO(ssize_t size, int flags, unsigned long caller_ip), - TP_ARGS(size, flags, caller_ip), - TP_STRUCT__entry( - __field(ssize_t, size) - __field(int, flags) - __field(unsigned long, caller_ip) - ), - TP_fast_assign( - __entry->size = size; - __entry->flags = flags; - __entry->caller_ip = caller_ip; - ), - TP_printk("size %zd flags 0x%x caller %pS", - __entry->size, - __entry->flags, - (char *)__entry->caller_ip) -) - -#define DEFINE_KMEM_EVENT(name) \ -DEFINE_EVENT(xfs_kmem_class, name, \ - TP_PROTO(ssize_t size, int flags, unsigned long caller_ip), \ - TP_ARGS(size, flags, caller_ip)) -DEFINE_KMEM_EVENT(kmem_alloc); - TRACE_EVENT(xfs_check_new_dalign, TP_PROTO(struct xfs_mount *mp, int new_dalign, xfs_ino_t calc_rootino), TP_ARGS(mp, new_dalign, calc_rootino), @@ -4091,7 +4231,7 @@ TRACE_EVENT(xfs_btree_commit_afakeroot, TP_ARGS(cur), TP_STRUCT__entry( __field(dev_t, dev) - __field(xfs_btnum_t, btnum) + __string(name, cur->bc_ops->name) __field(xfs_agnumber_t, agno) __field(xfs_agblock_t, agbno) __field(unsigned int, levels) @@ -4099,15 +4239,15 @@ TRACE_EVENT(xfs_btree_commit_afakeroot, ), TP_fast_assign( __entry->dev = cur->bc_mp->m_super->s_dev; - __entry->btnum = cur->bc_btnum; + __assign_str(name, cur->bc_ops->name); __entry->agno = cur->bc_ag.pag->pag_agno; __entry->agbno = cur->bc_ag.afake->af_root; __entry->levels = cur->bc_ag.afake->af_levels; __entry->blocks = cur->bc_ag.afake->af_blocks; ), - TP_printk("dev %d:%d btree %s agno 0x%x levels %u blocks %u root %u", + TP_printk("dev %d:%d %sbt agno 0x%x levels %u blocks %u root %u", MAJOR(__entry->dev), MINOR(__entry->dev), - __print_symbolic(__entry->btnum, XFS_BTNUM_STRINGS), + __get_str(name), __entry->agno, __entry->levels, __entry->blocks, @@ -4119,7 +4259,7 @@ TRACE_EVENT(xfs_btree_commit_ifakeroot, TP_ARGS(cur), TP_STRUCT__entry( __field(dev_t, dev) - __field(xfs_btnum_t, btnum) + __string(name, cur->bc_ops->name) __field(xfs_agnumber_t, agno) __field(xfs_agino_t, agino) __field(unsigned int, levels) @@ -4128,7 +4268,7 @@ TRACE_EVENT(xfs_btree_commit_ifakeroot, ), TP_fast_assign( __entry->dev = cur->bc_mp->m_super->s_dev; - __entry->btnum = cur->bc_btnum; + __assign_str(name, cur->bc_ops->name); __entry->agno = XFS_INO_TO_AGNO(cur->bc_mp, cur->bc_ino.ip->i_ino); __entry->agino = XFS_INO_TO_AGINO(cur->bc_mp, @@ -4137,9 +4277,9 @@ TRACE_EVENT(xfs_btree_commit_ifakeroot, __entry->blocks = cur->bc_ino.ifake->if_blocks; __entry->whichfork = cur->bc_ino.whichfork; ), - TP_printk("dev %d:%d btree %s agno 0x%x agino 0x%x whichfork %s levels %u blocks %u", + TP_printk("dev %d:%d %sbt agno 0x%x agino 0x%x whichfork %s levels %u blocks %u", MAJOR(__entry->dev), MINOR(__entry->dev), - __print_symbolic(__entry->btnum, XFS_BTNUM_STRINGS), + __get_str(name), __entry->agno, __entry->agino, __print_symbolic(__entry->whichfork, XFS_WHICHFORK_STRINGS), @@ -4156,7 +4296,7 @@ TRACE_EVENT(xfs_btree_bload_level_geometry, blocks_with_extra), TP_STRUCT__entry( __field(dev_t, dev) - __field(xfs_btnum_t, btnum) + __string(name, cur->bc_ops->name) __field(unsigned int, level) __field(unsigned int, nlevels) __field(uint64_t, nr_this_level) @@ -4167,7 +4307,7 @@ TRACE_EVENT(xfs_btree_bload_level_geometry, ), TP_fast_assign( __entry->dev = cur->bc_mp->m_super->s_dev; - __entry->btnum = cur->bc_btnum; + __assign_str(name, cur->bc_ops->name); __entry->level = level; __entry->nlevels = cur->bc_nlevels; __entry->nr_this_level = nr_this_level; @@ -4176,9 +4316,9 @@ TRACE_EVENT(xfs_btree_bload_level_geometry, __entry->blocks = blocks; __entry->blocks_with_extra = blocks_with_extra; ), - TP_printk("dev %d:%d btree %s level %u/%u nr_this_level %llu nr_per_block %u desired_npb %u blocks %llu blocks_with_extra %llu", + TP_printk("dev %d:%d %sbt level %u/%u nr_this_level %llu nr_per_block %u desired_npb %u blocks %llu blocks_with_extra %llu", MAJOR(__entry->dev), MINOR(__entry->dev), - __print_symbolic(__entry->btnum, XFS_BTNUM_STRINGS), + __get_str(name), __entry->level, __entry->nlevels, __entry->nr_this_level, @@ -4195,7 +4335,7 @@ TRACE_EVENT(xfs_btree_bload_block, TP_ARGS(cur, level, block_idx, nr_blocks, ptr, nr_records), TP_STRUCT__entry( __field(dev_t, dev) - __field(xfs_btnum_t, btnum) + __string(name, cur->bc_ops->name) __field(unsigned int, level) __field(unsigned long long, block_idx) __field(unsigned long long, nr_blocks) @@ -4205,11 +4345,11 @@ TRACE_EVENT(xfs_btree_bload_block, ), TP_fast_assign( __entry->dev = cur->bc_mp->m_super->s_dev; - __entry->btnum = cur->bc_btnum; + __assign_str(name, cur->bc_ops->name); __entry->level = level; __entry->block_idx = block_idx; __entry->nr_blocks = nr_blocks; - if (cur->bc_flags & XFS_BTREE_LONG_PTRS) { + if (cur->bc_ops->ptr_len == XFS_BTREE_LONG_PTR_LEN) { xfs_fsblock_t fsb = be64_to_cpu(ptr->l); __entry->agno = XFS_FSB_TO_AGNO(cur->bc_mp, fsb); @@ -4220,9 +4360,9 @@ TRACE_EVENT(xfs_btree_bload_block, } __entry->nr_records = nr_records; ), - TP_printk("dev %d:%d btree %s level %u block %llu/%llu agno 0x%x agbno 0x%x recs %u", + TP_printk("dev %d:%d %sbt level %u block %llu/%llu agno 0x%x agbno 0x%x recs %u", MAJOR(__entry->dev), MINOR(__entry->dev), - __print_symbolic(__entry->btnum, XFS_BTNUM_STRINGS), + __get_str(name), __entry->level, __entry->block_idx, __entry->nr_blocks, @@ -4472,6 +4612,159 @@ DEFINE_PERAG_INTENTS_EVENT(xfs_perag_wait_intents); #endif /* CONFIG_XFS_DRAIN_INTENTS */ +#ifdef CONFIG_XFS_MEMORY_BUFS +TRACE_EVENT(xmbuf_create, + TP_PROTO(struct xfs_buftarg *btp), + TP_ARGS(btp), + TP_STRUCT__entry( + __field(dev_t, dev) + __field(unsigned long, ino) + __array(char, pathname, 256) + ), + TP_fast_assign( + char pathname[257]; + char *path; + struct file *file = btp->bt_file; + + __entry->ino = file_inode(file)->i_ino; + memset(pathname, 0, sizeof(pathname)); + path = file_path(file, pathname, sizeof(pathname) - 1); + if (IS_ERR(path)) + path = "(unknown)"; + strncpy(__entry->pathname, path, sizeof(__entry->pathname)); + ), + TP_printk("xmino 0x%lx path '%s'", + __entry->ino, + __entry->pathname) +); + +TRACE_EVENT(xmbuf_free, + TP_PROTO(struct xfs_buftarg *btp), + TP_ARGS(btp), + TP_STRUCT__entry( + __field(unsigned long, ino) + __field(unsigned long long, bytes) + __field(loff_t, size) + ), + TP_fast_assign( + struct file *file = btp->bt_file; + struct inode *inode = file_inode(file); + + __entry->size = i_size_read(inode); + __entry->bytes = (inode->i_blocks << SECTOR_SHIFT) + inode->i_bytes; + __entry->ino = inode->i_ino; + ), + TP_printk("xmino 0x%lx mem_bytes 0x%llx isize 0x%llx", + __entry->ino, + __entry->bytes, + __entry->size) +); +#endif /* CONFIG_XFS_MEMORY_BUFS */ + +#ifdef CONFIG_XFS_BTREE_IN_MEM +TRACE_EVENT(xfbtree_init, + TP_PROTO(struct xfs_mount *mp, struct xfbtree *xfbt, + const struct xfs_btree_ops *ops), + TP_ARGS(mp, xfbt, ops), + TP_STRUCT__entry( + __field(const void *, btree_ops) + __field(unsigned long, xfino) + __field(unsigned int, leaf_mxr) + __field(unsigned int, leaf_mnr) + __field(unsigned int, node_mxr) + __field(unsigned int, node_mnr) + __field(unsigned long long, owner) + ), + TP_fast_assign( + __entry->btree_ops = ops; + __entry->xfino = file_inode(xfbt->target->bt_file)->i_ino; + __entry->leaf_mxr = xfbt->maxrecs[0]; + __entry->node_mxr = xfbt->maxrecs[1]; + __entry->leaf_mnr = xfbt->minrecs[0]; + __entry->node_mnr = xfbt->minrecs[1]; + __entry->owner = xfbt->owner; + ), + TP_printk("xfino 0x%lx btree_ops %pS owner 0x%llx leaf_mxr %u leaf_mnr %u node_mxr %u node_mnr %u", + __entry->xfino, + __entry->btree_ops, + __entry->owner, + __entry->leaf_mxr, + __entry->leaf_mnr, + __entry->node_mxr, + __entry->node_mnr) +); + +DECLARE_EVENT_CLASS(xfbtree_buf_class, + TP_PROTO(struct xfbtree *xfbt, struct xfs_buf *bp), + TP_ARGS(xfbt, bp), + TP_STRUCT__entry( + __field(unsigned long, xfino) + __field(xfs_daddr_t, bno) + __field(int, nblks) + __field(int, hold) + __field(int, pincount) + __field(unsigned int, lockval) + __field(unsigned int, flags) + ), + TP_fast_assign( + __entry->xfino = file_inode(xfbt->target->bt_file)->i_ino; + __entry->bno = xfs_buf_daddr(bp); + __entry->nblks = bp->b_length; + __entry->hold = atomic_read(&bp->b_hold); + __entry->pincount = atomic_read(&bp->b_pin_count); + __entry->lockval = bp->b_sema.count; + __entry->flags = bp->b_flags; + ), + TP_printk("xfino 0x%lx daddr 0x%llx bbcount 0x%x hold %d pincount %d lock %d flags %s", + __entry->xfino, + (unsigned long long)__entry->bno, + __entry->nblks, + __entry->hold, + __entry->pincount, + __entry->lockval, + __print_flags(__entry->flags, "|", XFS_BUF_FLAGS)) +) + +#define DEFINE_XFBTREE_BUF_EVENT(name) \ +DEFINE_EVENT(xfbtree_buf_class, name, \ + TP_PROTO(struct xfbtree *xfbt, struct xfs_buf *bp), \ + TP_ARGS(xfbt, bp)) +DEFINE_XFBTREE_BUF_EVENT(xfbtree_create_root_buf); +DEFINE_XFBTREE_BUF_EVENT(xfbtree_trans_commit_buf); +DEFINE_XFBTREE_BUF_EVENT(xfbtree_trans_cancel_buf); + +DECLARE_EVENT_CLASS(xfbtree_freesp_class, + TP_PROTO(struct xfbtree *xfbt, struct xfs_btree_cur *cur, + xfs_fileoff_t fileoff), + TP_ARGS(xfbt, cur, fileoff), + TP_STRUCT__entry( + __field(unsigned long, xfino) + __string(btname, cur->bc_ops->name) + __field(int, nlevels) + __field(xfs_fileoff_t, fileoff) + ), + TP_fast_assign( + __entry->xfino = file_inode(xfbt->target->bt_file)->i_ino; + __assign_str(btname, cur->bc_ops->name); + __entry->nlevels = cur->bc_nlevels; + __entry->fileoff = fileoff; + ), + TP_printk("xfino 0x%lx %sbt nlevels %d fileoff 0x%llx", + __entry->xfino, + __get_str(btname), + __entry->nlevels, + (unsigned long long)__entry->fileoff) +) + +#define DEFINE_XFBTREE_FREESP_EVENT(name) \ +DEFINE_EVENT(xfbtree_freesp_class, name, \ + TP_PROTO(struct xfbtree *xfbt, struct xfs_btree_cur *cur, \ + xfs_fileoff_t fileoff), \ + TP_ARGS(xfbt, cur, fileoff)) +DEFINE_XFBTREE_FREESP_EVENT(xfbtree_alloc_block); +DEFINE_XFBTREE_FREESP_EVENT(xfbtree_free_block); +#endif /* CONFIG_XFS_BTREE_IN_MEM */ + #endif /* _TRACE_XFS_H */ #undef TRACE_INCLUDE_PATH diff --git a/fs/xfs/xfs_trans.c b/fs/xfs/xfs_trans.c index 12d45e93f07d..7350640059cc 100644 --- a/fs/xfs/xfs_trans.c +++ b/fs/xfs/xfs_trans.c @@ -1273,7 +1273,7 @@ xfs_trans_reserve_more_inode( unsigned int rtx = xfs_extlen_to_rtxlen(mp, rblocks); int error; - ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL)); + xfs_assert_ilocked(ip, XFS_ILOCK_EXCL); error = xfs_trans_reserve(tp, &resv, dblocks, rtx); if (error) diff --git a/fs/xfs/xfs_trans.h b/fs/xfs/xfs_trans.h index 08ce757c7454..3f7e3a09a49f 100644 --- a/fs/xfs/xfs_trans.h +++ b/fs/xfs/xfs_trans.h @@ -215,6 +215,7 @@ struct xfs_buf *xfs_trans_getsb(struct xfs_trans *); void xfs_trans_brelse(xfs_trans_t *, struct xfs_buf *); void xfs_trans_bjoin(xfs_trans_t *, struct xfs_buf *); +void xfs_trans_bdetach(struct xfs_trans *tp, struct xfs_buf *bp); void xfs_trans_bhold(xfs_trans_t *, struct xfs_buf *); void xfs_trans_bhold_release(xfs_trans_t *, struct xfs_buf *); void xfs_trans_binval(xfs_trans_t *, struct xfs_buf *); diff --git a/fs/xfs/xfs_trans_ail.c b/fs/xfs/xfs_trans_ail.c index 1098452e7f95..e4c343096f95 100644 --- a/fs/xfs/xfs_trans_ail.c +++ b/fs/xfs/xfs_trans_ail.c @@ -901,7 +901,8 @@ xfs_trans_ail_init( { struct xfs_ail *ailp; - ailp = kmem_zalloc(sizeof(struct xfs_ail), KM_MAYFAIL); + ailp = kzalloc(sizeof(struct xfs_ail), + GFP_KERNEL | __GFP_RETRY_MAYFAIL); if (!ailp) return -ENOMEM; @@ -921,7 +922,7 @@ xfs_trans_ail_init( return 0; out_free_ailp: - kmem_free(ailp); + kfree(ailp); return -ENOMEM; } @@ -932,5 +933,5 @@ xfs_trans_ail_destroy( struct xfs_ail *ailp = mp->m_ail; kthread_stop(ailp->ail_task); - kmem_free(ailp); + kfree(ailp); } diff --git a/fs/xfs/xfs_trans_buf.c b/fs/xfs/xfs_trans_buf.c index 6549e50d852c..e28ab74af4f0 100644 --- a/fs/xfs/xfs_trans_buf.c +++ b/fs/xfs/xfs_trans_buf.c @@ -393,6 +393,48 @@ xfs_trans_brelse( } /* + * Forcibly detach a buffer previously joined to the transaction. The caller + * will retain its locked reference to the buffer after this function returns. + * The buffer must be completely clean and must not be held to the transaction. + */ +void +xfs_trans_bdetach( + struct xfs_trans *tp, + struct xfs_buf *bp) +{ + struct xfs_buf_log_item *bip = bp->b_log_item; + + ASSERT(tp != NULL); + ASSERT(bp->b_transp == tp); + ASSERT(bip->bli_item.li_type == XFS_LI_BUF); + ASSERT(atomic_read(&bip->bli_refcount) > 0); + + trace_xfs_trans_bdetach(bip); + + /* + * Erase all recursion count, since we're removing this buffer from the + * transaction. + */ + bip->bli_recur = 0; + + /* + * The buffer must be completely clean. Specifically, it had better + * not be dirty, stale, logged, ordered, or held to the transaction. + */ + ASSERT(!test_bit(XFS_LI_DIRTY, &bip->bli_item.li_flags)); + ASSERT(!(bip->bli_flags & XFS_BLI_DIRTY)); + ASSERT(!(bip->bli_flags & XFS_BLI_HOLD)); + ASSERT(!(bip->bli_flags & XFS_BLI_LOGGED)); + ASSERT(!(bip->bli_flags & XFS_BLI_ORDERED)); + ASSERT(!(bip->bli_flags & XFS_BLI_STALE)); + + /* Unlink the log item from the transaction and drop the log item. */ + xfs_trans_del_item(&bip->bli_item); + xfs_buf_item_put(bip); + bp->b_transp = NULL; +} + +/* * Mark the buffer as not needing to be unlocked when the buf item's * iop_committing() routine is called. The buffer must already be locked * and associated with the given transaction. diff --git a/fs/xfs/xfs_trans_dquot.c b/fs/xfs/xfs_trans_dquot.c index aa00cf67ad72..577b535a595c 100644 --- a/fs/xfs/xfs_trans_dquot.c +++ b/fs/xfs/xfs_trans_dquot.c @@ -17,6 +17,7 @@ #include "xfs_qm.h" #include "xfs_trace.h" #include "xfs_error.h" +#include "xfs_health.h" STATIC void xfs_trans_alloc_dqinfo(xfs_trans_t *); @@ -120,6 +121,116 @@ xfs_trans_dup_dqinfo( } } +#ifdef CONFIG_XFS_LIVE_HOOKS +/* + * Use a static key here to reduce the overhead of quota live updates. If the + * compiler supports jump labels, the static branch will be replaced by a nop + * sled when there are no hook users. Online fsck is currently the only + * caller, so this is a reasonable tradeoff. + * + * Note: Patching the kernel code requires taking the cpu hotplug lock. Other + * parts of the kernel allocate memory with that lock held, which means that + * XFS callers cannot hold any locks that might be used by memory reclaim or + * writeback when calling the static_branch_{inc,dec} functions. + */ +DEFINE_STATIC_XFS_HOOK_SWITCH(xfs_dqtrx_hooks_switch); + +void +xfs_dqtrx_hook_disable(void) +{ + xfs_hooks_switch_off(&xfs_dqtrx_hooks_switch); +} + +void +xfs_dqtrx_hook_enable(void) +{ + xfs_hooks_switch_on(&xfs_dqtrx_hooks_switch); +} + +/* Schedule a transactional dquot update on behalf of an inode. */ +void +xfs_trans_mod_ino_dquot( + struct xfs_trans *tp, + struct xfs_inode *ip, + struct xfs_dquot *dqp, + unsigned int field, + int64_t delta) +{ + xfs_trans_mod_dquot(tp, dqp, field, delta); + + if (xfs_hooks_switched_on(&xfs_dqtrx_hooks_switch)) { + struct xfs_mod_ino_dqtrx_params p = { + .tx_id = (uintptr_t)tp, + .ino = ip->i_ino, + .q_type = xfs_dquot_type(dqp), + .q_id = dqp->q_id, + .delta = delta + }; + struct xfs_quotainfo *qi = tp->t_mountp->m_quotainfo; + + xfs_hooks_call(&qi->qi_mod_ino_dqtrx_hooks, field, &p); + } +} + +/* Call the specified functions during a dquot counter update. */ +int +xfs_dqtrx_hook_add( + struct xfs_quotainfo *qi, + struct xfs_dqtrx_hook *hook) +{ + int error; + + /* + * Transactional dquot updates first call the mod hook when changes + * are attached to the transaction and then call the apply hook when + * those changes are committed (or canceled). + * + * The apply hook must be installed before the mod hook so that we + * never fail to catch the end of a quota update sequence. + */ + error = xfs_hooks_add(&qi->qi_apply_dqtrx_hooks, &hook->apply_hook); + if (error) + goto out; + + error = xfs_hooks_add(&qi->qi_mod_ino_dqtrx_hooks, &hook->mod_hook); + if (error) + goto out_apply; + + return 0; + +out_apply: + xfs_hooks_del(&qi->qi_apply_dqtrx_hooks, &hook->apply_hook); +out: + return error; +} + +/* Stop calling the specified function during a dquot counter update. */ +void +xfs_dqtrx_hook_del( + struct xfs_quotainfo *qi, + struct xfs_dqtrx_hook *hook) +{ + /* + * The mod hook must be removed before apply hook to avoid giving the + * hook consumer with an incomplete update. No hooks should be running + * after these functions return. + */ + xfs_hooks_del(&qi->qi_mod_ino_dqtrx_hooks, &hook->mod_hook); + xfs_hooks_del(&qi->qi_apply_dqtrx_hooks, &hook->apply_hook); +} + +/* Configure dquot update hook functions. */ +void +xfs_dqtrx_hook_setup( + struct xfs_dqtrx_hook *hook, + notifier_fn_t mod_fn, + notifier_fn_t apply_fn) +{ + xfs_hook_setup(&hook->mod_hook, mod_fn); + xfs_hook_setup(&hook->apply_hook, apply_fn); +} +#endif /* CONFIG_XFS_LIVE_HOOKS */ + /* * Wrap around mod_dquot to account for both user and group quotas. */ @@ -137,11 +248,11 @@ xfs_trans_mod_dquot_byino( return; if (XFS_IS_UQUOTA_ON(mp) && ip->i_udquot) - (void) xfs_trans_mod_dquot(tp, ip->i_udquot, field, delta); + xfs_trans_mod_ino_dquot(tp, ip, ip->i_udquot, field, delta); if (XFS_IS_GQUOTA_ON(mp) && ip->i_gdquot) - (void) xfs_trans_mod_dquot(tp, ip->i_gdquot, field, delta); + xfs_trans_mod_ino_dquot(tp, ip, ip->i_gdquot, field, delta); if (XFS_IS_PQUOTA_ON(mp) && ip->i_pdquot) - (void) xfs_trans_mod_dquot(tp, ip->i_pdquot, field, delta); + xfs_trans_mod_ino_dquot(tp, ip, ip->i_pdquot, field, delta); } STATIC struct xfs_dqtrx * @@ -321,6 +432,29 @@ xfs_apply_quota_reservation_deltas( } } +#ifdef CONFIG_XFS_LIVE_HOOKS +/* Call downstream hooks now that it's time to apply dquot deltas. */ +static inline void +xfs_trans_apply_dquot_deltas_hook( + struct xfs_trans *tp, + struct xfs_dquot *dqp) +{ + if (xfs_hooks_switched_on(&xfs_dqtrx_hooks_switch)) { + struct xfs_apply_dqtrx_params p = { + .tx_id = (uintptr_t)tp, + .q_type = xfs_dquot_type(dqp), + .q_id = dqp->q_id, + }; + struct xfs_quotainfo *qi = tp->t_mountp->m_quotainfo; + + xfs_hooks_call(&qi->qi_apply_dqtrx_hooks, + XFS_APPLY_DQTRX_COMMIT, &p); + } +} +#else +# define xfs_trans_apply_dquot_deltas_hook(tp, dqp) ((void)0) +#endif /* CONFIG_XFS_LIVE_HOOKS */ + /* * Called by xfs_trans_commit() and similar in spirit to * xfs_trans_apply_sb_deltas(). @@ -366,6 +500,8 @@ xfs_trans_apply_dquot_deltas( ASSERT(XFS_DQ_IS_LOCKED(dqp)); + xfs_trans_apply_dquot_deltas_hook(tp, dqp); + /* * adjust the actual number of blocks used */ @@ -465,6 +601,29 @@ xfs_trans_apply_dquot_deltas( } } +#ifdef CONFIG_XFS_LIVE_HOOKS +/* Call downstream hooks now that it's time to cancel dquot deltas. */ +static inline void +xfs_trans_unreserve_and_mod_dquots_hook( + struct xfs_trans *tp, + struct xfs_dquot *dqp) +{ + if (xfs_hooks_switched_on(&xfs_dqtrx_hooks_switch)) { + struct xfs_apply_dqtrx_params p = { + .tx_id = (uintptr_t)tp, + .q_type = xfs_dquot_type(dqp), + .q_id = dqp->q_id, + }; + struct xfs_quotainfo *qi = tp->t_mountp->m_quotainfo; + + xfs_hooks_call(&qi->qi_apply_dqtrx_hooks, + XFS_APPLY_DQTRX_UNRESERVE, &p); + } +} +#else +# define xfs_trans_unreserve_and_mod_dquots_hook(tp, dqp) ((void)0) +#endif /* CONFIG_XFS_LIVE_HOOKS */ + /* * Release the reservations, and adjust the dquots accordingly. * This is called only when the transaction is being aborted. If by @@ -495,6 +654,9 @@ xfs_trans_unreserve_and_mod_dquots( */ if ((dqp = qtrx->qt_dquot) == NULL) break; + + xfs_trans_unreserve_and_mod_dquots_hook(tp, dqp); + /* * Unreserve the original reservation. We don't care * about the number of blocks used field, or deltas. @@ -706,6 +868,7 @@ error_return: error_corrupt: xfs_dqunlock(dqp); xfs_force_shutdown(mp, SHUTDOWN_CORRUPT_INCORE); + xfs_fs_mark_sick(mp, XFS_SICK_FS_QUOTACHECK); return -EFSCORRUPTED; } @@ -796,7 +959,7 @@ xfs_trans_reserve_quota_nblks( return 0; ASSERT(!xfs_is_quota_inode(&mp->m_sb, ip->i_ino)); - ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL)); + xfs_assert_ilocked(ip, XFS_ILOCK_EXCL); if (force) qflags |= XFS_QMOPT_FORCE_RES; |