summaryrefslogtreecommitdiff
path: root/fs/ubifs
diff options
context:
space:
mode:
Diffstat (limited to 'fs/ubifs')
-rw-r--r--fs/ubifs/compress.c2
-rw-r--r--fs/ubifs/debug.c38
-rw-r--r--fs/ubifs/debug.h7
-rw-r--r--fs/ubifs/dir.c154
-rw-r--r--fs/ubifs/file.c450
-rw-r--r--fs/ubifs/find.c40
-rw-r--r--fs/ubifs/ioctl.c6
-rw-r--r--fs/ubifs/journal.c198
-rw-r--r--fs/ubifs/lprops.c8
-rw-r--r--fs/ubifs/lpt.c1
-rw-r--r--fs/ubifs/lpt_commit.c6
-rw-r--r--fs/ubifs/master.c5
-rw-r--r--fs/ubifs/orphan.c157
-rw-r--r--fs/ubifs/replay.c1
-rw-r--r--fs/ubifs/super.c416
-rw-r--r--fs/ubifs/sysfs.c6
-rw-r--r--fs/ubifs/tnc.c11
-rw-r--r--fs/ubifs/tnc_commit.c2
-rw-r--r--fs/ubifs/tnc_misc.c22
-rw-r--r--fs/ubifs/ubifs.h22
-rw-r--r--fs/ubifs/xattr.c47
21 files changed, 829 insertions, 770 deletions
diff --git a/fs/ubifs/compress.c b/fs/ubifs/compress.c
index 75461777c466..0b48cbab8a3d 100644
--- a/fs/ubifs/compress.c
+++ b/fs/ubifs/compress.c
@@ -82,6 +82,7 @@ struct ubifs_compressor *ubifs_compressors[UBIFS_COMPR_TYPES_CNT];
/**
* ubifs_compress - compress data.
+ * @c: UBIFS file-system description object
* @in_buf: data to compress
* @in_len: length of the data to compress
* @out_buf: output buffer where compressed data should be stored
@@ -140,6 +141,7 @@ no_compr:
/**
* ubifs_decompress - decompress data.
+ * @c: UBIFS file-system description object
* @in_buf: data to decompress
* @in_len: length of the data to decompress
* @out_buf: output buffer where decompressed data should
diff --git a/fs/ubifs/debug.c b/fs/ubifs/debug.c
index d013c5b3f1ed..b01f382ce8db 100644
--- a/fs/ubifs/debug.c
+++ b/fs/ubifs/debug.c
@@ -863,7 +863,6 @@ void ubifs_dump_leb(const struct ubifs_info *c, int lnum)
out:
vfree(buf);
- return;
}
void ubifs_dump_znode(const struct ubifs_info *c,
@@ -946,16 +945,20 @@ void ubifs_dump_tnc(struct ubifs_info *c)
pr_err("\n");
pr_err("(pid %d) start dumping TNC tree\n", current->pid);
- znode = ubifs_tnc_levelorder_next(c, c->zroot.znode, NULL);
- level = znode->level;
- pr_err("== Level %d ==\n", level);
- while (znode) {
- if (level != znode->level) {
- level = znode->level;
- pr_err("== Level %d ==\n", level);
+ if (c->zroot.znode) {
+ znode = ubifs_tnc_levelorder_next(c, c->zroot.znode, NULL);
+ level = znode->level;
+ pr_err("== Level %d ==\n", level);
+ while (znode) {
+ if (level != znode->level) {
+ level = znode->level;
+ pr_err("== Level %d ==\n", level);
+ }
+ ubifs_dump_znode(c, znode);
+ znode = ubifs_tnc_levelorder_next(c, c->zroot.znode, znode);
}
- ubifs_dump_znode(c, znode);
- znode = ubifs_tnc_levelorder_next(c, c->zroot.znode, znode);
+ } else {
+ pr_err("empty TNC tree in memory\n");
}
pr_err("(pid %d) finish dumping TNC tree\n", current->pid);
}
@@ -1742,17 +1745,22 @@ 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;
}
/**
@@ -2802,7 +2810,6 @@ static const struct file_operations dfs_fops = {
.read = dfs_file_read,
.write = dfs_file_write,
.owner = THIS_MODULE,
- .llseek = no_llseek,
};
/**
@@ -2822,9 +2829,9 @@ void dbg_debugfs_init_fs(struct ubifs_info *c)
const char *fname;
struct ubifs_debug_info *d = c->dbg;
- n = snprintf(d->dfs_dir_name, UBIFS_DFS_DIR_LEN + 1, UBIFS_DFS_DIR_NAME,
+ n = snprintf(d->dfs_dir_name, UBIFS_DFS_DIR_LEN, UBIFS_DFS_DIR_NAME,
c->vi.ubi_num, c->vi.vol_id);
- if (n > UBIFS_DFS_DIR_LEN) {
+ if (n >= UBIFS_DFS_DIR_LEN) {
/* The array size is too small */
return;
}
@@ -2947,7 +2954,6 @@ static const struct file_operations dfs_global_fops = {
.read = dfs_global_file_read,
.write = dfs_global_file_write,
.owner = THIS_MODULE,
- .llseek = no_llseek,
};
/**
diff --git a/fs/ubifs/debug.h b/fs/ubifs/debug.h
index ed966108da80..d425861e6b82 100644
--- a/fs/ubifs/debug.h
+++ b/fs/ubifs/debug.h
@@ -19,10 +19,11 @@ typedef int (*dbg_znode_callback)(struct ubifs_info *c,
/*
* The UBIFS debugfs directory name pattern and maximum name length (3 for "ubi"
- * + 1 for "_" and plus 2x2 for 2 UBI numbers and 1 for the trailing zero byte.
+ * + 1 for "_" and 2 for UBI device numbers and 3 for volume number and 1 for
+ * the trailing zero byte.
*/
#define UBIFS_DFS_DIR_NAME "ubi%d_%d"
-#define UBIFS_DFS_DIR_LEN (3 + 1 + 2*2 + 1)
+#define UBIFS_DFS_DIR_LEN (3 + 1 + 2 + 3 + 1)
/**
* ubifs_debug_info - per-FS debugging information.
@@ -103,7 +104,7 @@ struct ubifs_debug_info {
unsigned int chk_fs:1;
unsigned int tst_rcvry:1;
- char dfs_dir_name[UBIFS_DFS_DIR_LEN + 1];
+ char dfs_dir_name[UBIFS_DFS_DIR_LEN];
struct dentry *dfs_dir;
struct dentry *dfs_dump_lprops;
struct dentry *dfs_dump_budg;
diff --git a/fs/ubifs/dir.c b/fs/ubifs/dir.c
index e413a9cf8ee3..fda82f3e16e8 100644
--- a/fs/ubifs/dir.c
+++ b/fs/ubifs/dir.c
@@ -71,8 +71,13 @@ static int inherit_flags(const struct inode *dir, umode_t mode)
* @is_xattr: whether the inode is xattr inode
*
* This function finds an unused inode number, allocates new inode and
- * initializes it. Returns new inode in case of success and an error code in
- * case of failure.
+ * initializes it. Non-xattr new inode may be written with xattrs(selinux/
+ * encryption) before writing dentry, which could cause inconsistent problem
+ * when powercut happens between two operations. To deal with it, non-xattr
+ * new inode is initialized with zero-nlink and added into orphan list, caller
+ * should make sure that inode is relinked later, and make sure that orphan
+ * removing and journal writing into an committing atomic operation. Returns
+ * new inode in case of success and an error code in case of failure.
*/
struct inode *ubifs_new_inode(struct ubifs_info *c, struct inode *dir,
umode_t mode, bool is_xattr)
@@ -163,9 +168,25 @@ struct inode *ubifs_new_inode(struct ubifs_info *c, struct inode *dir,
ui->creat_sqnum = ++c->max_sqnum;
spin_unlock(&c->cnt_lock);
+ if (!is_xattr) {
+ set_nlink(inode, 0);
+ err = ubifs_add_orphan(c, inode->i_ino);
+ if (err) {
+ ubifs_err(c, "ubifs_add_orphan failed: %i", err);
+ goto out_iput;
+ }
+ down_read(&c->commit_sem);
+ ui->del_cmtno = c->cmt_no;
+ up_read(&c->commit_sem);
+ }
+
if (encrypted) {
err = fscrypt_set_context(inode, NULL);
if (err) {
+ if (!is_xattr) {
+ set_nlink(inode, 1);
+ ubifs_delete_orphan(c, inode->i_ino);
+ }
ubifs_err(c, "fscrypt_set_context failed: %i", err);
goto out_iput;
}
@@ -205,7 +226,6 @@ static struct dentry *ubifs_lookup(struct inode *dir, struct dentry *dentry,
dbg_gen("'%pd' in dir ino %lu", dentry, dir->i_ino);
err = fscrypt_prepare_lookup(dir, dentry, &nm);
- generic_set_encrypted_ci_d_ops(dentry);
if (err == -ENOENT)
return d_splice_alias(NULL, dentry);
if (err)
@@ -321,12 +341,13 @@ static int ubifs_create(struct mnt_idmap *idmap, struct inode *dir,
if (err)
goto out_inode;
+ set_nlink(inode, 1);
mutex_lock(&dir_ui->ui_mutex);
dir->i_size += sz_change;
dir_ui->ui_size = dir->i_size;
inode_set_mtime_to_ts(dir,
inode_set_ctime_to_ts(dir, inode_get_ctime(inode)));
- err = ubifs_jnl_update(c, dir, &nm, inode, 0, 0);
+ err = ubifs_jnl_update(c, dir, &nm, inode, 0, 0, 1);
if (err)
goto out_cancel;
mutex_unlock(&dir_ui->ui_mutex);
@@ -341,8 +362,8 @@ out_cancel:
dir->i_size -= sz_change;
dir_ui->ui_size = dir->i_size;
mutex_unlock(&dir_ui->ui_mutex);
+ set_nlink(inode, 0);
out_inode:
- make_bad_inode(inode);
iput(inode);
out_fname:
fscrypt_free_filename(&nm);
@@ -387,7 +408,6 @@ static struct inode *create_whiteout(struct inode *dir, struct dentry *dentry)
return inode;
out_inode:
- make_bad_inode(inode);
iput(inode);
out_free:
ubifs_err(c, "cannot create whiteout file, error %d", err);
@@ -471,6 +491,7 @@ static int ubifs_tmpfile(struct mnt_idmap *idmap, struct inode *dir,
if (err)
goto out_inode;
+ set_nlink(inode, 1);
mutex_lock(&ui->ui_mutex);
insert_inode_hash(inode);
d_tmpfile(file, inode);
@@ -480,7 +501,7 @@ static int ubifs_tmpfile(struct mnt_idmap *idmap, struct inode *dir,
mutex_unlock(&ui->ui_mutex);
lock_2_inodes(dir, inode);
- err = ubifs_jnl_update(c, dir, &nm, inode, 1, 0);
+ err = ubifs_jnl_update(c, dir, &nm, inode, 1, 0, 1);
if (err)
goto out_cancel;
unlock_2_inodes(dir, inode);
@@ -493,7 +514,6 @@ static int ubifs_tmpfile(struct mnt_idmap *idmap, struct inode *dir,
out_cancel:
unlock_2_inodes(dir, inode);
out_inode:
- make_bad_inode(inode);
if (!instantiated)
iput(inode);
out_budg:
@@ -535,6 +555,11 @@ static unsigned int vfs_dent_type(uint8_t type)
return 0;
}
+struct ubifs_dir_data {
+ struct ubifs_dent_node *dent;
+ u64 cookie;
+};
+
/*
* The classical Unix view for directory is that it is a linear array of
* (name, inode number) entries. Linux/VFS assumes this model as well.
@@ -562,6 +587,7 @@ static int ubifs_readdir(struct file *file, struct dir_context *ctx)
struct inode *dir = file_inode(file);
struct ubifs_info *c = dir->i_sb->s_fs_info;
bool encrypted = IS_ENCRYPTED(dir);
+ struct ubifs_dir_data *data = file->private_data;
dbg_gen("dir ino %lu, f_pos %#llx", dir->i_ino, ctx->pos);
@@ -584,27 +610,27 @@ static int ubifs_readdir(struct file *file, struct dir_context *ctx)
fstr_real_len = fstr.len;
}
- if (file->f_version == 0) {
+ if (data->cookie == 0) {
/*
- * The file was seek'ed, which means that @file->private_data
+ * The file was seek'ed, which means that @data->dent
* is now invalid. This may also be just the first
* 'ubifs_readdir()' invocation, in which case
- * @file->private_data is NULL, and the below code is
+ * @data->dent is NULL, and the below code is
* basically a no-op.
*/
- kfree(file->private_data);
- file->private_data = NULL;
+ kfree(data->dent);
+ data->dent = NULL;
}
/*
- * 'generic_file_llseek()' unconditionally sets @file->f_version to
- * zero, and we use this for detecting whether the file was seek'ed.
+ * 'ubifs_dir_llseek()' sets @data->cookie to zero, and we use this
+ * for detecting whether the file was seek'ed.
*/
- file->f_version = 1;
+ data->cookie = 1;
/* File positions 0 and 1 correspond to "." and ".." */
if (ctx->pos < 2) {
- ubifs_assert(c, !file->private_data);
+ ubifs_assert(c, !data->dent);
if (!dir_emit_dots(file, ctx)) {
if (encrypted)
fscrypt_fname_free_buffer(&fstr);
@@ -621,10 +647,10 @@ static int ubifs_readdir(struct file *file, struct dir_context *ctx)
}
ctx->pos = key_hash_flash(c, &dent->key);
- file->private_data = dent;
+ data->dent = dent;
}
- dent = file->private_data;
+ dent = data->dent;
if (!dent) {
/*
* The directory was seek'ed to and is now readdir'ed.
@@ -638,7 +664,7 @@ static int ubifs_readdir(struct file *file, struct dir_context *ctx)
goto out;
}
ctx->pos = key_hash_flash(c, &dent->key);
- file->private_data = dent;
+ data->dent = dent;
}
while (1) {
@@ -681,15 +707,15 @@ static int ubifs_readdir(struct file *file, struct dir_context *ctx)
goto out;
}
- kfree(file->private_data);
+ kfree(data->dent);
ctx->pos = key_hash_flash(c, &dent->key);
- file->private_data = dent;
+ data->dent = dent;
cond_resched();
}
out:
- kfree(file->private_data);
- file->private_data = NULL;
+ kfree(data->dent);
+ data->dent = NULL;
if (encrypted)
fscrypt_fname_free_buffer(&fstr);
@@ -713,7 +739,10 @@ out:
/* Free saved readdir() state when the directory is closed */
static int ubifs_dir_release(struct inode *dir, struct file *file)
{
- kfree(file->private_data);
+ struct ubifs_dir_data *data = file->private_data;
+
+ kfree(data->dent);
+ kfree(data);
file->private_data = NULL;
return 0;
}
@@ -761,10 +790,6 @@ static int ubifs_link(struct dentry *old_dentry, struct inode *dir,
lock_2_inodes(dir, inode);
- /* Handle O_TMPFILE corner case, it is allowed to link a O_TMPFILE. */
- if (inode->i_nlink == 0)
- ubifs_delete_orphan(c, inode->i_ino);
-
inc_nlink(inode);
ihold(inode);
inode_set_ctime_current(inode);
@@ -772,7 +797,7 @@ static int ubifs_link(struct dentry *old_dentry, struct inode *dir,
dir_ui->ui_size = dir->i_size;
inode_set_mtime_to_ts(dir,
inode_set_ctime_to_ts(dir, inode_get_ctime(inode)));
- err = ubifs_jnl_update(c, dir, &nm, inode, 0, 0);
+ err = ubifs_jnl_update(c, dir, &nm, inode, 0, 0, inode->i_nlink == 1);
if (err)
goto out_cancel;
unlock_2_inodes(dir, inode);
@@ -786,8 +811,6 @@ out_cancel:
dir->i_size -= sz_change;
dir_ui->ui_size = dir->i_size;
drop_nlink(inode);
- if (inode->i_nlink == 0)
- ubifs_add_orphan(c, inode->i_ino);
unlock_2_inodes(dir, inode);
ubifs_release_budget(c, &req);
iput(inode);
@@ -847,7 +870,7 @@ static int ubifs_unlink(struct inode *dir, struct dentry *dentry)
dir_ui->ui_size = dir->i_size;
inode_set_mtime_to_ts(dir,
inode_set_ctime_to_ts(dir, inode_get_ctime(inode)));
- err = ubifs_jnl_update(c, dir, &nm, inode, 1, 0);
+ err = ubifs_jnl_update(c, dir, &nm, inode, 1, 0, 0);
if (err)
goto out_cancel;
unlock_2_inodes(dir, inode);
@@ -951,7 +974,7 @@ static int ubifs_rmdir(struct inode *dir, struct dentry *dentry)
dir_ui->ui_size = dir->i_size;
inode_set_mtime_to_ts(dir,
inode_set_ctime_to_ts(dir, inode_get_ctime(inode)));
- err = ubifs_jnl_update(c, dir, &nm, inode, 1, 0);
+ err = ubifs_jnl_update(c, dir, &nm, inode, 1, 0, 0);
if (err)
goto out_cancel;
unlock_2_inodes(dir, inode);
@@ -1018,6 +1041,7 @@ static int ubifs_mkdir(struct mnt_idmap *idmap, struct inode *dir,
if (err)
goto out_inode;
+ set_nlink(inode, 1);
mutex_lock(&dir_ui->ui_mutex);
insert_inode_hash(inode);
inc_nlink(inode);
@@ -1026,7 +1050,7 @@ static int ubifs_mkdir(struct mnt_idmap *idmap, struct inode *dir,
dir_ui->ui_size = dir->i_size;
inode_set_mtime_to_ts(dir,
inode_set_ctime_to_ts(dir, inode_get_ctime(inode)));
- err = ubifs_jnl_update(c, dir, &nm, inode, 0, 0);
+ err = ubifs_jnl_update(c, dir, &nm, inode, 0, 0, 1);
if (err) {
ubifs_err(c, "cannot create directory, error %d", err);
goto out_cancel;
@@ -1043,8 +1067,8 @@ out_cancel:
dir_ui->ui_size = dir->i_size;
drop_nlink(dir);
mutex_unlock(&dir_ui->ui_mutex);
+ set_nlink(inode, 0);
out_inode:
- make_bad_inode(inode);
iput(inode);
out_fname:
fscrypt_free_filename(&nm);
@@ -1103,22 +1127,25 @@ static int ubifs_mknod(struct mnt_idmap *idmap, struct inode *dir,
goto out_fname;
}
+ err = ubifs_init_security(dir, inode, &dentry->d_name);
+ if (err) {
+ kfree(dev);
+ goto out_inode;
+ }
+
init_special_inode(inode, inode->i_mode, rdev);
inode->i_size = ubifs_inode(inode)->ui_size = devlen;
ui = ubifs_inode(inode);
ui->data = dev;
ui->data_len = devlen;
-
- err = ubifs_init_security(dir, inode, &dentry->d_name);
- if (err)
- goto out_inode;
+ set_nlink(inode, 1);
mutex_lock(&dir_ui->ui_mutex);
dir->i_size += sz_change;
dir_ui->ui_size = dir->i_size;
inode_set_mtime_to_ts(dir,
inode_set_ctime_to_ts(dir, inode_get_ctime(inode)));
- err = ubifs_jnl_update(c, dir, &nm, inode, 0, 0);
+ err = ubifs_jnl_update(c, dir, &nm, inode, 0, 0, 1);
if (err)
goto out_cancel;
mutex_unlock(&dir_ui->ui_mutex);
@@ -1133,8 +1160,8 @@ out_cancel:
dir->i_size -= sz_change;
dir_ui->ui_size = dir->i_size;
mutex_unlock(&dir_ui->ui_mutex);
+ set_nlink(inode, 0);
out_inode:
- make_bad_inode(inode);
iput(inode);
out_fname:
fscrypt_free_filename(&nm);
@@ -1185,6 +1212,10 @@ static int ubifs_symlink(struct mnt_idmap *idmap, struct inode *dir,
goto out_fname;
}
+ err = ubifs_init_security(dir, inode, &dentry->d_name);
+ if (err)
+ goto out_inode;
+
ui = ubifs_inode(inode);
ui->data = kmalloc(disk_link.len, GFP_NOFS);
if (!ui->data) {
@@ -1209,17 +1240,14 @@ static int ubifs_symlink(struct mnt_idmap *idmap, struct inode *dir,
*/
ui->data_len = disk_link.len - 1;
inode->i_size = ubifs_inode(inode)->ui_size = disk_link.len - 1;
-
- err = ubifs_init_security(dir, inode, &dentry->d_name);
- if (err)
- goto out_inode;
+ set_nlink(inode, 1);
mutex_lock(&dir_ui->ui_mutex);
dir->i_size += sz_change;
dir_ui->ui_size = dir->i_size;
inode_set_mtime_to_ts(dir,
inode_set_ctime_to_ts(dir, inode_get_ctime(inode)));
- err = ubifs_jnl_update(c, dir, &nm, inode, 0, 0);
+ err = ubifs_jnl_update(c, dir, &nm, inode, 0, 0, 1);
if (err)
goto out_cancel;
mutex_unlock(&dir_ui->ui_mutex);
@@ -1233,10 +1261,10 @@ out_cancel:
dir->i_size -= sz_change;
dir_ui->ui_size = dir->i_size;
mutex_unlock(&dir_ui->ui_mutex);
+ set_nlink(inode, 0);
out_inode:
/* Free inode->i_link before inode is marked as bad. */
fscrypt_free_inode(inode);
- make_bad_inode(inode);
iput(inode);
out_fname:
fscrypt_free_filename(&nm);
@@ -1404,14 +1432,10 @@ static int do_rename(struct inode *old_dir, struct dentry *old_dentry,
*/
err = ubifs_budget_space(c, &wht_req);
if (err) {
- /*
- * Whiteout inode can not be written on flash by
- * ubifs_jnl_write_inode(), because it's neither
- * dirty nor zero-nlink.
- */
iput(whiteout);
goto out_release;
}
+ set_nlink(whiteout, 1);
/* Add the old_dentry size to the old_dir size. */
old_sz -= CALC_DENT_SIZE(fname_len(&old_nm));
@@ -1490,7 +1514,7 @@ static int do_rename(struct inode *old_dir, struct dentry *old_dentry,
}
err = ubifs_jnl_rename(c, old_dir, old_inode, &old_nm, new_dir,
- new_inode, &new_nm, whiteout, sync);
+ new_inode, &new_nm, whiteout, sync, !!whiteout);
if (err)
goto out_cancel;
@@ -1543,6 +1567,7 @@ out_cancel:
unlock_4_inodes(old_dir, new_dir, new_inode, whiteout);
if (whiteout) {
ubifs_release_budget(c, &wht_req);
+ set_nlink(whiteout, 0);
iput(whiteout);
}
out_release:
@@ -1696,6 +1721,24 @@ int ubifs_getattr(struct mnt_idmap *idmap, const struct path *path,
return 0;
}
+static int ubifs_dir_open(struct inode *inode, struct file *file)
+{
+ struct ubifs_dir_data *data;
+
+ data = kzalloc(sizeof(struct ubifs_dir_data), GFP_KERNEL);
+ if (!data)
+ return -ENOMEM;
+ file->private_data = data;
+ return 0;
+}
+
+static loff_t ubifs_dir_llseek(struct file *file, loff_t offset, int whence)
+{
+ struct ubifs_dir_data *data = file->private_data;
+
+ return generic_llseek_cookie(file, offset, whence, &data->cookie);
+}
+
const struct inode_operations ubifs_dir_inode_operations = {
.lookup = ubifs_lookup,
.create = ubifs_create,
@@ -1716,7 +1759,8 @@ const struct inode_operations ubifs_dir_inode_operations = {
};
const struct file_operations ubifs_dir_operations = {
- .llseek = generic_file_llseek,
+ .open = ubifs_dir_open,
+ .llseek = ubifs_dir_llseek,
.release = ubifs_dir_release,
.read = generic_read_dir,
.iterate_shared = ubifs_readdir,
diff --git a/fs/ubifs/file.c b/fs/ubifs/file.c
index 5029eb3390a5..5130123005e4 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;
}
@@ -215,23 +211,23 @@ static void release_existing_page_budget(struct ubifs_info *c)
}
static int write_begin_slow(struct address_space *mapping,
- loff_t pos, unsigned len, struct page **pagep)
+ loff_t pos, unsigned len, struct folio **foliop)
{
struct inode *inode = mapping->host;
struct ubifs_info *c = inode->i_sb->s_fs_info;
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;
+ *foliop = folio;
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.
*/
@@ -420,7 +414,7 @@ static int allocate_budget(struct ubifs_info *c, struct page *page,
*/
static int ubifs_write_begin(struct file *file, struct address_space *mapping,
loff_t pos, unsigned len,
- struct page **pagep, void **fsdata)
+ struct folio **foliop, void **fsdata)
{
struct inode *inode = mapping->host;
struct ubifs_info *c = inode->i_sb->s_fs_info;
@@ -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
@@ -490,10 +480,10 @@ 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);
+ return write_begin_slow(mapping, pos, len, foliop);
}
/*
@@ -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;
+ *foliop = folio;
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);
@@ -535,7 +524,7 @@ static void cancel_budget(struct ubifs_info *c, struct page *page,
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, void *fsdata)
{
struct inode *inode = mapping->host;
struct ubifs_inode *ui = ubifs_inode(inode);
@@ -544,44 +533,47 @@ 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 +581,43 @@ 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 +666,13 @@ 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 +681,19 @@ 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 +703,15 @@ 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 +761,11 @@ 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 +775,19 @@ 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 +810,7 @@ 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 +819,12 @@ 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 +872,8 @@ 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 +887,71 @@ 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 +1001,21 @@ 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 +1024,9 @@ 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) + len > synced_i_size) {
err = inode->i_sb->s_op->write_inode(inode, NULL);
if (err)
goto out_redirty;
@@ -1045,20 +1039,18 @@ 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 +1058,25 @@ 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 +1153,11 @@ 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 +1166,14 @@ 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 +1186,8 @@ 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 +1510,14 @@ 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 +1525,17 @@ 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 +1558,21 @@ 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 +1588,11 @@ 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 +1646,7 @@ 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..643718906b9f 100644
--- a/fs/ubifs/find.c
+++ b/fs/ubifs/find.c
@@ -73,7 +73,7 @@ static int valuable(struct ubifs_info *c, const struct ubifs_lprops *lprops)
* @c: the UBIFS file-system description object
* @lprops: LEB properties to scan
* @in_tree: whether the LEB properties are in main memory
- * @data: information passed to and from the caller of the scan
+ * @arg: information passed to and from the caller of the scan
*
* This function returns a code that indicates whether the scan should continue
* (%LPT_SCAN_CONTINUE), whether the LEB properties should be added to the tree
@@ -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);
@@ -340,7 +340,7 @@ out:
* @c: the UBIFS file-system description object
* @lprops: LEB properties to scan
* @in_tree: whether the LEB properties are in main memory
- * @data: information passed to and from the caller of the scan
+ * @arg: information passed to and from the caller of the scan
*
* This function returns a code that indicates whether the scan should continue
* (%LPT_SCAN_CONTINUE), whether the LEB properties should be added to the tree
@@ -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);
@@ -580,7 +581,7 @@ out:
* @c: the UBIFS file-system description object
* @lprops: LEB properties to scan
* @in_tree: whether the LEB properties are in main memory
- * @data: information passed to and from the caller of the scan
+ * @arg: information passed to and from the caller of the scan
*
* This function returns a code that indicates whether the scan should continue
* (%LPT_SCAN_CONTINUE), whether the LEB properties should be added to the tree
@@ -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",
@@ -773,7 +773,7 @@ int ubifs_save_dirty_idx_lnums(struct ubifs_info *c)
* @c: the UBIFS file-system description object
* @lprops: LEB properties to scan
* @in_tree: whether the LEB properties are in main memory
- * @data: information passed to and from the caller of the scan
+ * @arg: information passed to and from the caller of the scan
*
* This function returns a code that indicates whether the scan should continue
* (%LPT_SCAN_CONTINUE), whether the LEB properties should be added to the tree
@@ -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/ioctl.c b/fs/ubifs/ioctl.c
index d79cabe193c3..2c99349cf537 100644
--- a/fs/ubifs/ioctl.c
+++ b/fs/ubifs/ioctl.c
@@ -213,12 +213,6 @@ long ubifs_ioctl(struct file *file, unsigned int cmd, unsigned long arg)
long ubifs_compat_ioctl(struct file *file, unsigned int cmd, unsigned long arg)
{
switch (cmd) {
- case FS_IOC32_GETFLAGS:
- cmd = FS_IOC_GETFLAGS;
- break;
- case FS_IOC32_SETFLAGS:
- cmd = FS_IOC_SETFLAGS;
- break;
case FS_IOC_SET_ENCRYPTION_POLICY:
case FS_IOC_GET_ENCRYPTION_POLICY:
case FS_IOC_GET_ENCRYPTION_POLICY_EX:
diff --git a/fs/ubifs/journal.c b/fs/ubifs/journal.c
index f0a5538c84b0..36ba79fbd2ff 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);
+}
+
+/**
+ * add_or_start_queue - 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;
}
@@ -516,6 +643,7 @@ static void set_dent_cookie(struct ubifs_info *c, struct ubifs_dent_node *dent)
* @inode: inode to update
* @deletion: indicates a directory entry deletion i.e unlink or rmdir
* @xent: non-zero if the directory entry is an extended attribute entry
+ * @in_orphan: indicates whether the @inode is in orphan list
*
* This function updates an inode by writing a directory entry (or extended
* attribute entry), the inode itself, and the parent directory inode (or the
@@ -537,7 +665,7 @@ static void set_dent_cookie(struct ubifs_info *c, struct ubifs_dent_node *dent)
*/
int ubifs_jnl_update(struct ubifs_info *c, const struct inode *dir,
const struct fscrypt_name *nm, const struct inode *inode,
- int deletion, int xent)
+ int deletion, int xent, int in_orphan)
{
int err, dlen, ilen, len, lnum, ino_offs, dent_offs, orphan_added = 0;
int aligned_dlen, aligned_ilen, sync = IS_DIRSYNC(dir);
@@ -623,7 +751,7 @@ int ubifs_jnl_update(struct ubifs_info *c, const struct inode *dir,
if (err)
goto out_release;
- if (last_reference) {
+ if (last_reference && !in_orphan) {
err = ubifs_add_orphan(c, inode->i_ino);
if (err) {
release_head(c, BASEHD);
@@ -679,6 +807,9 @@ int ubifs_jnl_update(struct ubifs_info *c, const struct inode *dir,
if (err)
goto out_ro;
+ if (in_orphan && inode->i_nlink)
+ ubifs_delete_orphan(c, inode->i_ino);
+
finish_reservation(c);
spin_lock(&ui->ui_lock);
ui->synced_i_size = ui->ui_size;
@@ -850,6 +981,13 @@ int ubifs_jnl_write_inode(struct ubifs_info *c, const struct inode *inode)
dbg_jnl("ino %lu, nlink %u", inode->i_ino, inode->i_nlink);
+ if (kill_xattrs && ui->xattr_cnt > ubifs_xattr_max_cnt(c)) {
+ ubifs_err(c, "Cannot delete inode, it has too much xattrs!");
+ err = -EPERM;
+ ubifs_ro_mode(c, err);
+ return err;
+ }
+
/*
* If the inode is being deleted, do not write the attached data. No
* need to synchronize the write-buffer either.
@@ -881,12 +1019,6 @@ int ubifs_jnl_write_inode(struct ubifs_info *c, const struct inode *inode)
struct inode *xino;
struct ubifs_dent_node *xent, *pxent = NULL;
- if (ui->xattr_cnt > ubifs_xattr_max_cnt(c)) {
- err = -EPERM;
- ubifs_err(c, "Cannot delete inode, it has too much xattrs!");
- goto out_release;
- }
-
lowest_xent_key(c, &key, inode->i_ino);
while (1) {
xent = ubifs_tnc_next_ent(c, &key, &nm);
@@ -1209,6 +1341,7 @@ out_free:
* @new_nm: new name of the new directory entry
* @whiteout: whiteout inode
* @sync: non-zero if the write-buffer has to be synchronized
+ * @delete_orphan: indicates an orphan entry deletion for @whiteout
*
* This function implements the re-name operation which may involve writing up
* to 4 inodes(new inode, whiteout inode, old and new parent directory inodes)
@@ -1221,7 +1354,7 @@ int ubifs_jnl_rename(struct ubifs_info *c, const struct inode *old_dir,
const struct inode *new_dir,
const struct inode *new_inode,
const struct fscrypt_name *new_nm,
- const struct inode *whiteout, int sync)
+ const struct inode *whiteout, int sync, int delete_orphan)
{
void *p;
union ubifs_key key;
@@ -1438,6 +1571,9 @@ int ubifs_jnl_rename(struct ubifs_info *c, const struct inode *old_dir,
goto out_ro;
}
+ if (delete_orphan)
+ ubifs_delete_orphan(c, whiteout->i_ino);
+
finish_reservation(c);
if (new_inode) {
mark_inode_clean(c, new_ui);
diff --git a/fs/ubifs/lprops.c b/fs/ubifs/lprops.c
index 6d6cd85c2b4c..8788740ec57f 100644
--- a/fs/ubifs/lprops.c
+++ b/fs/ubifs/lprops.c
@@ -1005,7 +1005,7 @@ out:
* @c: the UBIFS file-system description object
* @lp: LEB properties to scan
* @in_tree: whether the LEB properties are in main memory
- * @lst: lprops statistics to update
+ * @arg: lprops statistics to update
*
* This function returns a code that indicates whether the scan should continue
* (%LPT_SCAN_CONTINUE), whether the LEB properties should be added to the tree
@@ -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.c b/fs/ubifs/lpt.c
index 778a22bf9a92..441d0beca4cf 100644
--- a/fs/ubifs/lpt.c
+++ b/fs/ubifs/lpt.c
@@ -1918,6 +1918,7 @@ out_err:
* @pnode: where to keep a pnode
* @cnode: where to keep a cnode
* @in_tree: is the node in the tree in memory
+ * @ptr: union of node pointers
* @ptr.nnode: pointer to the nnode (if it is an nnode) which may be here or in
* the tree
* @ptr.pnode: ditto for pnode
diff --git a/fs/ubifs/lpt_commit.c b/fs/ubifs/lpt_commit.c
index c4d079328b92..f2cb214581fd 100644
--- a/fs/ubifs/lpt_commit.c
+++ b/fs/ubifs/lpt_commit.c
@@ -577,7 +577,7 @@ static struct ubifs_pnode *next_pnode_to_dirty(struct ubifs_info *c,
/* Go right */
nnode = ubifs_get_nnode(c, nnode, iip);
if (IS_ERR(nnode))
- return (void *)nnode;
+ return ERR_CAST(nnode);
/* Go down to level 1 */
while (nnode->level > 1) {
@@ -594,7 +594,7 @@ static struct ubifs_pnode *next_pnode_to_dirty(struct ubifs_info *c,
}
nnode = ubifs_get_nnode(c, nnode, iip);
if (IS_ERR(nnode))
- return (void *)nnode;
+ return ERR_CAST(nnode);
}
for (iip = 0; iip < UBIFS_LPT_FANOUT; iip++)
@@ -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;
@@ -1933,7 +1932,6 @@ static void dump_lpt_leb(const struct ubifs_info *c, int lnum)
pr_err("(pid %d) finish dumping LEB %d\n", current->pid, lnum);
out:
vfree(buf);
- return;
}
/**
diff --git a/fs/ubifs/master.c b/fs/ubifs/master.c
index 7adc37c10b6a..a148760fa49e 100644
--- a/fs/ubifs/master.c
+++ b/fs/ubifs/master.c
@@ -67,10 +67,13 @@ static int mst_node_check_hash(const struct ubifs_info *c,
{
u8 calc[UBIFS_MAX_HASH_LEN];
const void *node = mst;
+ int ret;
- crypto_shash_tfm_digest(c->hash_tfm, node + sizeof(struct ubifs_ch),
+ ret = crypto_shash_tfm_digest(c->hash_tfm, node + sizeof(struct ubifs_ch),
UBIFS_MST_NODE_SZ - sizeof(struct ubifs_ch),
calc);
+ if (ret)
+ return ret;
if (ubifs_check_hash(c, expected, calc))
return -EPERM;
diff --git a/fs/ubifs/orphan.c b/fs/ubifs/orphan.c
index 4909321d84cf..5555dd740889 100644
--- a/fs/ubifs/orphan.c
+++ b/fs/ubifs/orphan.c
@@ -42,24 +42,30 @@
static int dbg_check_orphans(struct ubifs_info *c);
-static struct ubifs_orphan *orphan_add(struct ubifs_info *c, ino_t inum,
- struct ubifs_orphan *parent_orphan)
+/**
+ * ubifs_add_orphan - add an orphan.
+ * @c: UBIFS file-system description object
+ * @inum: orphan inode number
+ *
+ * Add an orphan. This function is called when an inodes link count drops to
+ * zero.
+ */
+int ubifs_add_orphan(struct ubifs_info *c, ino_t inum)
{
struct ubifs_orphan *orphan, *o;
struct rb_node **p, *parent = NULL;
orphan = kzalloc(sizeof(struct ubifs_orphan), GFP_NOFS);
if (!orphan)
- return ERR_PTR(-ENOMEM);
+ return -ENOMEM;
orphan->inum = inum;
orphan->new = 1;
- INIT_LIST_HEAD(&orphan->child_list);
spin_lock(&c->orphan_lock);
if (c->tot_orphans >= c->max_orphans) {
spin_unlock(&c->orphan_lock);
kfree(orphan);
- return ERR_PTR(-ENFILE);
+ return -ENFILE;
}
p = &c->orph_tree.rb_node;
while (*p) {
@@ -70,10 +76,10 @@ static struct ubifs_orphan *orphan_add(struct ubifs_info *c, ino_t inum,
else if (inum > o->inum)
p = &(*p)->rb_right;
else {
- ubifs_err(c, "orphaned twice");
+ ubifs_err(c, "ino %lu orphaned twice", (unsigned long)inum);
spin_unlock(&c->orphan_lock);
kfree(orphan);
- return ERR_PTR(-EINVAL);
+ return -EINVAL;
}
}
c->tot_orphans += 1;
@@ -83,14 +89,9 @@ static struct ubifs_orphan *orphan_add(struct ubifs_info *c, ino_t inum,
list_add_tail(&orphan->list, &c->orph_list);
list_add_tail(&orphan->new_list, &c->orph_new);
- if (parent_orphan) {
- list_add_tail(&orphan->child_list,
- &parent_orphan->child_list);
- }
-
spin_unlock(&c->orphan_lock);
dbg_gen("ino %lu", (unsigned long)inum);
- return orphan;
+ return 0;
}
static struct ubifs_orphan *lookup_orphan(struct ubifs_info *c, ino_t inum)
@@ -135,6 +136,7 @@ static void orphan_delete(struct ubifs_info *c, struct ubifs_orphan *orph)
if (orph->cmt) {
orph->del = 1;
+ rb_erase(&orph->rb, &c->orph_tree);
orph->dnext = c->orph_dnext;
c->orph_dnext = orph;
dbg_gen("delete later ino %lu", (unsigned long)orph->inum);
@@ -145,59 +147,6 @@ static void orphan_delete(struct ubifs_info *c, struct ubifs_orphan *orph)
}
/**
- * ubifs_add_orphan - add an orphan.
- * @c: UBIFS file-system description object
- * @inum: orphan inode number
- *
- * Add an orphan. This function is called when an inodes link count drops to
- * zero.
- */
-int ubifs_add_orphan(struct ubifs_info *c, ino_t inum)
-{
- int err = 0;
- ino_t xattr_inum;
- union ubifs_key key;
- struct ubifs_dent_node *xent, *pxent = NULL;
- struct fscrypt_name nm = {0};
- struct ubifs_orphan *xattr_orphan;
- struct ubifs_orphan *orphan;
-
- orphan = orphan_add(c, inum, NULL);
- if (IS_ERR(orphan))
- return PTR_ERR(orphan);
-
- lowest_xent_key(c, &key, inum);
- while (1) {
- xent = ubifs_tnc_next_ent(c, &key, &nm);
- if (IS_ERR(xent)) {
- err = PTR_ERR(xent);
- if (err == -ENOENT)
- break;
- kfree(pxent);
- return err;
- }
-
- fname_name(&nm) = xent->name;
- fname_len(&nm) = le16_to_cpu(xent->nlen);
- xattr_inum = le64_to_cpu(xent->inum);
-
- xattr_orphan = orphan_add(c, xattr_inum, orphan);
- if (IS_ERR(xattr_orphan)) {
- kfree(pxent);
- kfree(xent);
- return PTR_ERR(xattr_orphan);
- }
-
- kfree(pxent);
- pxent = xent;
- key_read(c, &xent->key, &key);
- }
- kfree(pxent);
-
- return 0;
-}
-
-/**
* ubifs_delete_orphan - delete an orphan.
* @c: UBIFS file-system description object
* @inum: orphan inode number
@@ -206,7 +155,7 @@ int ubifs_add_orphan(struct ubifs_info *c, ino_t inum)
*/
void ubifs_delete_orphan(struct ubifs_info *c, ino_t inum)
{
- struct ubifs_orphan *orph, *child_orph, *tmp_o;
+ struct ubifs_orphan *orph;
spin_lock(&c->orphan_lock);
@@ -219,11 +168,6 @@ void ubifs_delete_orphan(struct ubifs_info *c, ino_t inum)
return;
}
- list_for_each_entry_safe(child_orph, tmp_o, &orph->child_list, child_list) {
- list_del(&child_orph->child_list);
- orphan_delete(c, child_orph);
- }
-
orphan_delete(c, orph);
spin_unlock(&c->orphan_lock);
@@ -518,7 +462,6 @@ static void erase_deleted(struct ubifs_info *c)
dnext = orphan->dnext;
ubifs_assert(c, !orphan->new);
ubifs_assert(c, orphan->del);
- rb_erase(&orphan->rb, &c->orph_tree);
list_del(&orphan->list);
c->tot_orphans -= 1;
dbg_gen("deleting orphan ino %lu", (unsigned long)orphan->inum);
@@ -571,51 +514,6 @@ int ubifs_clear_orphans(struct ubifs_info *c)
}
/**
- * insert_dead_orphan - insert an orphan.
- * @c: UBIFS file-system description object
- * @inum: orphan inode number
- *
- * This function is a helper to the 'do_kill_orphans()' function. The orphan
- * must be kept until the next commit, so it is added to the rb-tree and the
- * deletion list.
- */
-static int insert_dead_orphan(struct ubifs_info *c, ino_t inum)
-{
- struct ubifs_orphan *orphan, *o;
- struct rb_node **p, *parent = NULL;
-
- orphan = kzalloc(sizeof(struct ubifs_orphan), GFP_KERNEL);
- if (!orphan)
- return -ENOMEM;
- orphan->inum = inum;
-
- p = &c->orph_tree.rb_node;
- while (*p) {
- parent = *p;
- o = rb_entry(parent, struct ubifs_orphan, rb);
- if (inum < o->inum)
- p = &(*p)->rb_left;
- else if (inum > o->inum)
- p = &(*p)->rb_right;
- else {
- /* Already added - no problem */
- kfree(orphan);
- return 0;
- }
- }
- c->tot_orphans += 1;
- rb_link_node(&orphan->rb, parent, p);
- rb_insert_color(&orphan->rb, &c->orph_tree);
- list_add_tail(&orphan->list, &c->orph_list);
- orphan->del = 1;
- orphan->dnext = c->orph_dnext;
- c->orph_dnext = orphan;
- dbg_mnt("ino %lu, new %d, tot %d", (unsigned long)inum,
- c->new_orphans, c->tot_orphans);
- return 0;
-}
-
-/**
* do_kill_orphans - remove orphan inodes from the index.
* @c: UBIFS file-system description object
* @sleb: scanned LEB
@@ -691,12 +589,12 @@ static int do_kill_orphans(struct ubifs_info *c, struct ubifs_scan_leb *sleb,
n = (le32_to_cpu(orph->ch.len) - UBIFS_ORPH_NODE_SZ) >> 3;
for (i = 0; i < n; i++) {
- union ubifs_key key1, key2;
+ union ubifs_key key;
inum = le64_to_cpu(orph->inos[i]);
- ino_key_init(c, &key1, inum);
- err = ubifs_tnc_lookup(c, &key1, ino);
+ ino_key_init(c, &key, inum);
+ err = ubifs_tnc_lookup(c, &key, ino);
if (err && err != -ENOENT)
goto out_free;
@@ -708,17 +606,10 @@ static int do_kill_orphans(struct ubifs_info *c, struct ubifs_scan_leb *sleb,
dbg_rcvry("deleting orphaned inode %lu",
(unsigned long)inum);
- lowest_ino_key(c, &key1, inum);
- highest_ino_key(c, &key2, inum);
-
- err = ubifs_tnc_remove_range(c, &key1, &key2);
+ err = ubifs_tnc_remove_ino(c, inum);
if (err)
goto out_ro;
}
-
- err = insert_dead_orphan(c, inum);
- if (err)
- goto out_free;
}
*last_cmt_no = cmt_no;
@@ -925,8 +816,12 @@ static int dbg_orphan_check(struct ubifs_info *c, struct ubifs_zbranch *zbr,
inum = key_inum(c, &zbr->key);
if (inum != ci->last_ino) {
- /* Lowest node type is the inode node, so it comes first */
- if (key_type(c, &zbr->key) != UBIFS_INO_KEY)
+ /*
+ * Lowest node type is the inode node or xattr entry(when
+ * selinux/encryption is enabled), so it comes first
+ */
+ if (key_type(c, &zbr->key) != UBIFS_INO_KEY &&
+ key_type(c, &zbr->key) != UBIFS_XENT_KEY)
ubifs_err(c, "found orphan node ino %lu, type %d",
(unsigned long)inum, key_type(c, &zbr->key));
ci->last_ino = inum;
diff --git a/fs/ubifs/replay.c b/fs/ubifs/replay.c
index 17da28d6247a..a950c5f2560e 100644
--- a/fs/ubifs/replay.c
+++ b/fs/ubifs/replay.c
@@ -29,6 +29,7 @@
* @lnum: logical eraseblock number of the node
* @offs: node offset
* @len: node length
+ * @hash: node hash
* @deletion: non-zero if this entry corresponds to a node deletion
* @sqnum: node sequence number
* @list: links the replay list
diff --git a/fs/ubifs/super.c b/fs/ubifs/super.c
index 09e270d6ed02..f3e3b2068608 100644
--- a/fs/ubifs/super.c
+++ b/fs/ubifs/super.c
@@ -19,9 +19,9 @@
#include <linux/module.h>
#include <linux/ctype.h>
#include <linux/kthread.h>
-#include <linux/parser.h>
+#include <linux/fs_context.h>
+#include <linux/fs_parser.h>
#include <linux/seq_file.h>
-#include <linux/mount.h>
#include <linux/math64.h>
#include <linux/writeback.h>
#include "ubifs.h"
@@ -773,10 +773,10 @@ static void init_constants_master(struct ubifs_info *c)
* necessary to report something for the 'statfs()' call.
*
* Subtract the LEB reserved for GC, the LEB which is reserved for
- * deletions, minimum LEBs for the index, and assume only one journal
- * head is available.
+ * deletions, minimum LEBs for the index, the LEBs which are reserved
+ * for each journal head.
*/
- tmp64 = c->main_lebs - 1 - 1 - MIN_INDEX_LEBS - c->jhead_cnt + 1;
+ tmp64 = c->main_lebs - 1 - 1 - MIN_INDEX_LEBS - c->jhead_cnt;
tmp64 *= (long long)c->leb_size - c->leb_overhead;
tmp64 = ubifs_reported_space(c, tmp64);
c->block_cnt = tmp64 >> UBIFS_BLOCK_SHIFT;
@@ -981,177 +981,120 @@ enum {
Opt_auth_key,
Opt_auth_hash_name,
Opt_ignore,
- Opt_err,
};
-static const match_table_t tokens = {
- {Opt_fast_unmount, "fast_unmount"},
- {Opt_norm_unmount, "norm_unmount"},
- {Opt_bulk_read, "bulk_read"},
- {Opt_no_bulk_read, "no_bulk_read"},
- {Opt_chk_data_crc, "chk_data_crc"},
- {Opt_no_chk_data_crc, "no_chk_data_crc"},
- {Opt_override_compr, "compr=%s"},
- {Opt_auth_key, "auth_key=%s"},
- {Opt_auth_hash_name, "auth_hash_name=%s"},
- {Opt_ignore, "ubi=%s"},
- {Opt_ignore, "vol=%s"},
- {Opt_assert, "assert=%s"},
- {Opt_err, NULL},
+static const struct constant_table ubifs_param_compr[] = {
+ { "none", UBIFS_COMPR_NONE },
+ { "lzo", UBIFS_COMPR_LZO },
+ { "zlib", UBIFS_COMPR_ZLIB },
+ { "zstd", UBIFS_COMPR_ZSTD },
+ {}
};
-/**
- * parse_standard_option - parse a standard mount option.
- * @option: the option to parse
- *
- * Normally, standard mount options like "sync" are passed to file-systems as
- * flags. However, when a "rootflags=" kernel boot parameter is used, they may
- * be present in the options string. This function tries to deal with this
- * situation and parse standard options. Returns 0 if the option was not
- * recognized, and the corresponding integer flag if it was.
- *
- * UBIFS is only interested in the "sync" option, so do not check for anything
- * else.
- */
-static int parse_standard_option(const char *option)
-{
+static const struct constant_table ubifs_param_assert[] = {
+ { "report", ASSACT_REPORT },
+ { "read-only", ASSACT_RO },
+ { "panic", ASSACT_PANIC },
+ {}
+};
- pr_notice("UBIFS: parse %s\n", option);
- if (!strcmp(option, "sync"))
- return SB_SYNCHRONOUS;
- return 0;
-}
+static const struct fs_parameter_spec ubifs_fs_param_spec[] = {
+ fsparam_flag ("fast_unmount", Opt_fast_unmount),
+ fsparam_flag ("norm_unmount", Opt_norm_unmount),
+ fsparam_flag ("bulk_read", Opt_bulk_read),
+ fsparam_flag ("no_bulk_read", Opt_no_bulk_read),
+ fsparam_flag ("chk_data_crc", Opt_chk_data_crc),
+ fsparam_flag ("no_chk_data_crc", Opt_no_chk_data_crc),
+ fsparam_enum ("compr", Opt_override_compr, ubifs_param_compr),
+ fsparam_enum ("assert", Opt_assert, ubifs_param_assert),
+ fsparam_string ("auth_key", Opt_auth_key),
+ fsparam_string ("auth_hash_name", Opt_auth_hash_name),
+ fsparam_string ("ubi", Opt_ignore),
+ fsparam_string ("vol", Opt_ignore),
+ {}
+};
+
+struct ubifs_fs_context {
+ struct ubifs_mount_opts mount_opts;
+ char *auth_key_name;
+ char *auth_hash_name;
+ unsigned int no_chk_data_crc:1;
+ unsigned int bulk_read:1;
+ unsigned int default_compr:2;
+ unsigned int assert_action:2;
+};
/**
- * ubifs_parse_options - parse mount parameters.
- * @c: UBIFS file-system description object
- * @options: parameters to parse
- * @is_remount: non-zero if this is FS re-mount
+ * ubifs_parse_param - parse a parameter.
+ * @fc: the filesystem context
+ * @param: the parameter to parse
*
* This function parses UBIFS mount options and returns zero in case success
* and a negative error code in case of failure.
*/
-static int ubifs_parse_options(struct ubifs_info *c, char *options,
- int is_remount)
+static int ubifs_parse_param(struct fs_context *fc, struct fs_parameter *param)
{
- char *p;
- substring_t args[MAX_OPT_ARGS];
-
- if (!options)
- return 0;
+ struct ubifs_fs_context *ctx = fc->fs_private;
+ struct fs_parse_result result;
+ bool is_remount = (fc->purpose & FS_CONTEXT_FOR_RECONFIGURE);
+ int opt;
- while ((p = strsep(&options, ","))) {
- int token;
+ opt = fs_parse(fc, ubifs_fs_param_spec, param, &result);
+ if (opt < 0)
+ return opt;
- if (!*p)
- continue;
-
- token = match_token(p, tokens, args);
- switch (token) {
+ switch (opt) {
/*
* %Opt_fast_unmount and %Opt_norm_unmount options are ignored.
* We accept them in order to be backward-compatible. But this
* should be removed at some point.
*/
- case Opt_fast_unmount:
- c->mount_opts.unmount_mode = 2;
- break;
- case Opt_norm_unmount:
- c->mount_opts.unmount_mode = 1;
- break;
- case Opt_bulk_read:
- c->mount_opts.bulk_read = 2;
- c->bulk_read = 1;
- break;
- case Opt_no_bulk_read:
- c->mount_opts.bulk_read = 1;
- c->bulk_read = 0;
- break;
- case Opt_chk_data_crc:
- c->mount_opts.chk_data_crc = 2;
- c->no_chk_data_crc = 0;
- break;
- case Opt_no_chk_data_crc:
- c->mount_opts.chk_data_crc = 1;
- c->no_chk_data_crc = 1;
- break;
- case Opt_override_compr:
- {
- char *name = match_strdup(&args[0]);
-
- if (!name)
- return -ENOMEM;
- if (!strcmp(name, "none"))
- c->mount_opts.compr_type = UBIFS_COMPR_NONE;
- else if (!strcmp(name, "lzo"))
- c->mount_opts.compr_type = UBIFS_COMPR_LZO;
- else if (!strcmp(name, "zlib"))
- c->mount_opts.compr_type = UBIFS_COMPR_ZLIB;
- else if (!strcmp(name, "zstd"))
- c->mount_opts.compr_type = UBIFS_COMPR_ZSTD;
- else {
- ubifs_err(c, "unknown compressor \"%s\"", name); //FIXME: is c ready?
- kfree(name);
- return -EINVAL;
- }
- kfree(name);
- c->mount_opts.override_compr = 1;
- c->default_compr = c->mount_opts.compr_type;
- break;
- }
- case Opt_assert:
- {
- char *act = match_strdup(&args[0]);
-
- if (!act)
- return -ENOMEM;
- if (!strcmp(act, "report"))
- c->assert_action = ASSACT_REPORT;
- else if (!strcmp(act, "read-only"))
- c->assert_action = ASSACT_RO;
- else if (!strcmp(act, "panic"))
- c->assert_action = ASSACT_PANIC;
- else {
- ubifs_err(c, "unknown assert action \"%s\"", act);
- kfree(act);
- return -EINVAL;
- }
- kfree(act);
- break;
- }
- case Opt_auth_key:
- if (!is_remount) {
- c->auth_key_name = kstrdup(args[0].from,
- GFP_KERNEL);
- if (!c->auth_key_name)
- return -ENOMEM;
- }
- break;
- case Opt_auth_hash_name:
- if (!is_remount) {
- c->auth_hash_name = kstrdup(args[0].from,
- GFP_KERNEL);
- if (!c->auth_hash_name)
- return -ENOMEM;
- }
- break;
- case Opt_ignore:
- break;
- default:
- {
- unsigned long flag;
- struct super_block *sb = c->vfs_sb;
-
- flag = parse_standard_option(p);
- if (!flag) {
- ubifs_err(c, "unrecognized mount option \"%s\" or missing value",
- p);
- return -EINVAL;
- }
- sb->s_flags |= flag;
- break;
+ case Opt_fast_unmount:
+ ctx->mount_opts.unmount_mode = 2;
+ break;
+ case Opt_norm_unmount:
+ ctx->mount_opts.unmount_mode = 1;
+ break;
+ case Opt_bulk_read:
+ ctx->mount_opts.bulk_read = 2;
+ ctx->bulk_read = 1;
+ break;
+ case Opt_no_bulk_read:
+ ctx->mount_opts.bulk_read = 1;
+ ctx->bulk_read = 0;
+ break;
+ case Opt_chk_data_crc:
+ ctx->mount_opts.chk_data_crc = 2;
+ ctx->no_chk_data_crc = 0;
+ break;
+ case Opt_no_chk_data_crc:
+ ctx->mount_opts.chk_data_crc = 1;
+ ctx->no_chk_data_crc = 1;
+ break;
+ case Opt_override_compr:
+ ctx->mount_opts.compr_type = result.uint_32;
+ ctx->mount_opts.override_compr = 1;
+ ctx->default_compr = ctx->mount_opts.compr_type;
+ break;
+ case Opt_assert:
+ ctx->assert_action = result.uint_32;
+ break;
+ case Opt_auth_key:
+ if (!is_remount) {
+ kfree(ctx->auth_key_name);
+ ctx->auth_key_name = param->string;
+ param->string = NULL;
}
+ break;
+ case Opt_auth_hash_name:
+ if (!is_remount) {
+ kfree(ctx->auth_hash_name);
+ ctx->auth_hash_name = param->string;
+ param->string = NULL;
}
+ break;
+ case Opt_ignore:
+ break;
}
return 0;
@@ -2003,21 +1946,27 @@ static void ubifs_put_super(struct super_block *sb)
mutex_unlock(&c->umount_mutex);
}
-static int ubifs_remount_fs(struct super_block *sb, int *flags, char *data)
+static int ubifs_reconfigure(struct fs_context *fc)
{
+ struct ubifs_fs_context *ctx = fc->fs_private;
+ struct super_block *sb = fc->root->d_sb;
int err;
struct ubifs_info *c = sb->s_fs_info;
sync_filesystem(sb);
- dbg_gen("old flags %#lx, new flags %#x", sb->s_flags, *flags);
+ dbg_gen("old flags %#lx, new flags %#x", sb->s_flags, fc->sb_flags);
- err = ubifs_parse_options(c, data, 1);
- if (err) {
- ubifs_err(c, "invalid or unknown remount parameter");
- return err;
- }
+ /*
+ * Apply the mount option changes.
+ * auth_key_name and auth_hash_name are ignored on remount.
+ */
+ c->mount_opts = ctx->mount_opts;
+ c->bulk_read = ctx->bulk_read;
+ c->no_chk_data_crc = ctx->no_chk_data_crc;
+ c->default_compr = ctx->default_compr;
+ c->assert_action = ctx->assert_action;
- if (c->ro_mount && !(*flags & SB_RDONLY)) {
+ if (c->ro_mount && !(fc->sb_flags & SB_RDONLY)) {
if (c->ro_error) {
ubifs_msg(c, "cannot re-mount R/W due to prior errors");
return -EROFS;
@@ -2029,7 +1978,7 @@ static int ubifs_remount_fs(struct super_block *sb, int *flags, char *data)
err = ubifs_remount_rw(c);
if (err)
return err;
- } else if (!c->ro_mount && (*flags & SB_RDONLY)) {
+ } else if (!c->ro_mount && (fc->sb_flags & SB_RDONLY)) {
if (c->ro_error) {
ubifs_msg(c, "cannot re-mount R/O due to prior errors");
return -EROFS;
@@ -2062,14 +2011,13 @@ const struct super_operations ubifs_super_operations = {
.evict_inode = ubifs_evict_inode,
.statfs = ubifs_statfs,
.dirty_inode = ubifs_dirty_inode,
- .remount_fs = ubifs_remount_fs,
.show_options = ubifs_show_options,
.sync_fs = ubifs_sync_fs,
};
/**
* open_ubi - parse UBI device name string and open the UBI device.
- * @name: UBI volume name
+ * @fc: The filesystem context
* @mode: UBI volume open mode
*
* The primary method of mounting UBIFS is by specifying the UBI volume
@@ -2086,15 +2034,13 @@ const struct super_operations ubifs_super_operations = {
* returns UBI volume description object in case of success and a negative
* error code in case of failure.
*/
-static struct ubi_volume_desc *open_ubi(const char *name, int mode)
+static struct ubi_volume_desc *open_ubi(struct fs_context *fc, int mode)
{
struct ubi_volume_desc *ubi;
+ const char *name = fc->source;
int dev, vol;
char *endptr;
- if (!name || !*name)
- return ERR_PTR(-EINVAL);
-
/* First, try to open using the device node path method */
ubi = ubi_open_volume_path(name, mode);
if (!IS_ERR(ubi))
@@ -2102,14 +2048,14 @@ static struct ubi_volume_desc *open_ubi(const char *name, int mode)
/* Try the "nodev" method */
if (name[0] != 'u' || name[1] != 'b' || name[2] != 'i')
- return ERR_PTR(-EINVAL);
+ goto invalid_source;
/* ubi:NAME method */
if ((name[3] == ':' || name[3] == '!') && name[4] != '\0')
return ubi_open_volume_nm(0, name + 4, mode);
if (!isdigit(name[3]))
- return ERR_PTR(-EINVAL);
+ goto invalid_source;
dev = simple_strtoul(name + 3, &endptr, 0);
@@ -2121,7 +2067,7 @@ static struct ubi_volume_desc *open_ubi(const char *name, int mode)
if (*endptr == '_' && isdigit(endptr[1])) {
vol = simple_strtoul(endptr + 1, &endptr, 0);
if (*endptr != '\0')
- return ERR_PTR(-EINVAL);
+ goto invalid_source;
return ubi_open_volume(dev, vol, mode);
}
@@ -2129,7 +2075,8 @@ static struct ubi_volume_desc *open_ubi(const char *name, int mode)
if ((*endptr == ':' || *endptr == '!') && endptr[1] != '\0')
return ubi_open_volume_nm(dev, ++endptr, mode);
- return ERR_PTR(-EINVAL);
+invalid_source:
+ return ERR_PTR(invalf(fc, "Invalid source name"));
}
static struct ubifs_info *alloc_ubifs_info(struct ubi_volume_desc *ubi)
@@ -2151,6 +2098,8 @@ 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;
@@ -2179,9 +2128,10 @@ static struct ubifs_info *alloc_ubifs_info(struct ubi_volume_desc *ubi)
return c;
}
-static int ubifs_fill_super(struct super_block *sb, void *data, int silent)
+static int ubifs_fill_super(struct super_block *sb, struct fs_context *fc)
{
struct ubifs_info *c = sb->s_fs_info;
+ struct ubifs_fs_context *ctx = fc->fs_private;
struct inode *root;
int err;
@@ -2193,9 +2143,18 @@ static int ubifs_fill_super(struct super_block *sb, void *data, int silent)
goto out;
}
- err = ubifs_parse_options(c, data, 0);
- if (err)
- goto out_close;
+ /* Copy in parsed mount options */
+ c->mount_opts = ctx->mount_opts;
+ c->auth_key_name = ctx->auth_key_name;
+ c->auth_hash_name = ctx->auth_hash_name;
+ c->no_chk_data_crc = ctx->no_chk_data_crc;
+ c->bulk_read = ctx->bulk_read;
+ c->default_compr = ctx->default_compr;
+ c->assert_action = ctx->assert_action;
+
+ /* ubifs_info owns auth strings now */
+ ctx->auth_key_name = NULL;
+ ctx->auth_hash_name = NULL;
/*
* UBIFS provides 'backing_dev_info' in order to disable read-ahead. For
@@ -2239,13 +2198,16 @@ static int ubifs_fill_super(struct super_block *sb, void *data, int silent)
goto out_umount;
}
+ generic_set_sb_d_ops(sb);
sb->s_root = d_make_root(root);
if (!sb->s_root) {
err = -ENOMEM;
goto out_umount;
}
- import_uuid(&sb->s_uuid, c->uuid);
+ super_set_uuid(sb, c->uuid, sizeof(c->uuid));
+ super_set_sysfs_name_generic(sb, UBIFS_DFS_DIR_NAME,
+ c->vi.ubi_num, c->vi.vol_id);
mutex_unlock(&c->umount_mutex);
return 0;
@@ -2261,41 +2223,38 @@ out:
return err;
}
-static int sb_test(struct super_block *sb, void *data)
+static int sb_test(struct super_block *sb, struct fs_context *fc)
{
- struct ubifs_info *c1 = data;
+ struct ubifs_info *c1 = fc->s_fs_info;
struct ubifs_info *c = sb->s_fs_info;
return c->vi.cdev == c1->vi.cdev;
}
-static int sb_set(struct super_block *sb, void *data)
-{
- sb->s_fs_info = data;
- return set_anon_super(sb, NULL);
-}
-
-static struct dentry *ubifs_mount(struct file_system_type *fs_type, int flags,
- const char *name, void *data)
+static int ubifs_get_tree(struct fs_context *fc)
{
struct ubi_volume_desc *ubi;
struct ubifs_info *c;
struct super_block *sb;
int err;
- dbg_gen("name %s, flags %#x", name, flags);
+ if (!fc->source || !*fc->source)
+ return invalf(fc, "No source specified");
+
+ dbg_gen("name %s, flags %#x", fc->source, fc->sb_flags);
/*
* Get UBI device number and volume ID. Mount it read-only so far
* because this might be a new mount point, and UBI allows only one
* read-write user at a time.
*/
- ubi = open_ubi(name, UBI_READONLY);
+ ubi = open_ubi(fc, UBI_READONLY);
if (IS_ERR(ubi)) {
- if (!(flags & SB_SILENT))
+ err = PTR_ERR(ubi);
+ if (!(fc->sb_flags & SB_SILENT))
pr_err("UBIFS error (pid: %d): cannot open \"%s\", error %d",
- current->pid, name, (int)PTR_ERR(ubi));
- return ERR_CAST(ubi);
+ current->pid, fc->source, err);
+ return err;
}
c = alloc_ubifs_info(ubi);
@@ -2303,10 +2262,11 @@ static struct dentry *ubifs_mount(struct file_system_type *fs_type, int flags,
err = -ENOMEM;
goto out_close;
}
+ fc->s_fs_info = c;
dbg_gen("opened ubi%d_%d", c->vi.ubi_num, c->vi.vol_id);
- sb = sget(fs_type, sb_test, sb_set, flags, c);
+ sb = sget_fc(fc, sb_test, set_anon_super_fc);
if (IS_ERR(sb)) {
err = PTR_ERR(sb);
kfree(c);
@@ -2318,12 +2278,12 @@ static struct dentry *ubifs_mount(struct file_system_type *fs_type, int flags,
kfree(c);
/* A new mount point for already mounted UBIFS */
dbg_gen("this ubi volume is already mounted");
- if (!!(flags & SB_RDONLY) != c1->ro_mount) {
+ if (!!(fc->sb_flags & SB_RDONLY) != c1->ro_mount) {
err = -EBUSY;
goto out_deact;
}
} else {
- err = ubifs_fill_super(sb, data, flags & SB_SILENT ? 1 : 0);
+ err = ubifs_fill_super(sb, fc);
if (err)
goto out_deact;
/* We do not support atime */
@@ -2337,13 +2297,14 @@ static struct dentry *ubifs_mount(struct file_system_type *fs_type, int flags,
/* 'fill_super()' opens ubi again so we must close it here */
ubi_close_volume(ubi);
- return dget(sb->s_root);
+ fc->root = dget(sb->s_root);
+ return 0;
out_deact:
deactivate_locked_super(sb);
out_close:
ubi_close_volume(ubi);
- return ERR_PTR(err);
+ return err;
}
static void kill_ubifs_super(struct super_block *s)
@@ -2353,10 +2314,61 @@ static void kill_ubifs_super(struct super_block *s)
kfree(c);
}
+static void ubifs_free_fc(struct fs_context *fc)
+{
+ struct ubifs_fs_context *ctx = fc->fs_private;
+
+ if (ctx) {
+ kfree(ctx->auth_key_name);
+ kfree(ctx->auth_hash_name);
+ kfree(ctx);
+ }
+}
+
+static const struct fs_context_operations ubifs_context_ops = {
+ .free = ubifs_free_fc,
+ .parse_param = ubifs_parse_param,
+ .get_tree = ubifs_get_tree,
+ .reconfigure = ubifs_reconfigure,
+};
+
+static int ubifs_init_fs_context(struct fs_context *fc)
+{
+ struct ubifs_fs_context *ctx;
+
+ ctx = kzalloc(sizeof(struct ubifs_fs_context), GFP_KERNEL);
+ if (!ctx)
+ return -ENOMEM;
+
+ if (fc->purpose != FS_CONTEXT_FOR_RECONFIGURE) {
+ /* Iniitialize for first mount */
+ ctx->no_chk_data_crc = 1;
+ ctx->assert_action = ASSACT_RO;
+ } else {
+ struct ubifs_info *c = fc->root->d_sb->s_fs_info;
+
+ /*
+ * Preserve existing options across remounts.
+ * auth_key_name and auth_hash_name are not remountable.
+ */
+ ctx->mount_opts = c->mount_opts;
+ ctx->bulk_read = c->bulk_read;
+ ctx->no_chk_data_crc = c->no_chk_data_crc;
+ ctx->default_compr = c->default_compr;
+ ctx->assert_action = c->assert_action;
+ }
+
+ fc->ops = &ubifs_context_ops;
+ fc->fs_private = ctx;
+
+ return 0;
+}
+
static struct file_system_type ubifs_fs_type = {
.name = "ubifs",
.owner = THIS_MODULE,
- .mount = ubifs_mount,
+ .init_fs_context = ubifs_init_fs_context,
+ .parameters = ubifs_fs_param_spec,
.kill_sb = kill_ubifs_super,
};
MODULE_ALIAS_FS("ubifs");
@@ -2433,8 +2445,8 @@ static int __init ubifs_init(void)
ubifs_inode_slab = kmem_cache_create("ubifs_inode_slab",
sizeof(struct ubifs_inode), 0,
- SLAB_MEM_SPREAD | SLAB_RECLAIM_ACCOUNT |
- SLAB_ACCOUNT, &inode_slab_ctor);
+ SLAB_RECLAIM_ACCOUNT | SLAB_ACCOUNT,
+ &inode_slab_ctor);
if (!ubifs_inode_slab)
return -ENOMEM;
diff --git a/fs/ubifs/sysfs.c b/fs/ubifs/sysfs.c
index 1c958148bb87..aae32222f11b 100644
--- a/fs/ubifs/sysfs.c
+++ b/fs/ubifs/sysfs.c
@@ -91,17 +91,17 @@ static struct kset ubifs_kset = {
int ubifs_sysfs_register(struct ubifs_info *c)
{
int ret, n;
- char dfs_dir_name[UBIFS_DFS_DIR_LEN+1];
+ char dfs_dir_name[UBIFS_DFS_DIR_LEN];
c->stats = kzalloc(sizeof(struct ubifs_stats_info), GFP_KERNEL);
if (!c->stats) {
ret = -ENOMEM;
goto out_last;
}
- n = snprintf(dfs_dir_name, UBIFS_DFS_DIR_LEN + 1, UBIFS_DFS_DIR_NAME,
+ n = snprintf(dfs_dir_name, UBIFS_DFS_DIR_LEN, UBIFS_DFS_DIR_NAME,
c->vi.ubi_num, c->vi.vol_id);
- if (n > UBIFS_DFS_DIR_LEN) {
+ if (n >= UBIFS_DFS_DIR_LEN) {
/* The array size is too small */
ret = -EINVAL;
goto out_free;
diff --git a/fs/ubifs/tnc.c b/fs/ubifs/tnc.c
index f4728e65d1bd..33946b518148 100644
--- a/fs/ubifs/tnc.c
+++ b/fs/ubifs/tnc.c
@@ -2930,8 +2930,6 @@ int ubifs_tnc_remove_ino(struct ubifs_info *c, ino_t inum)
dbg_tnc("xent '%s', ino %lu", xent->name,
(unsigned long)xattr_inum);
- ubifs_evict_xattr_inode(c, xattr_inum);
-
fname_name(&nm) = xent->name;
fname_len(&nm) = le16_to_cpu(xent->nlen);
err = ubifs_tnc_remove_nm(c, &key1, &nm);
@@ -3116,14 +3114,7 @@ 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_commit.c b/fs/ubifs/tnc_commit.c
index a55e04822d16..7c43e0ccf6d4 100644
--- a/fs/ubifs/tnc_commit.c
+++ b/fs/ubifs/tnc_commit.c
@@ -657,6 +657,8 @@ static int get_znodes_to_commit(struct ubifs_info *c)
znode->alt = 0;
cnext = find_next_dirty(znode);
if (!cnext) {
+ ubifs_assert(c, !znode->parent);
+ znode->cparent = NULL;
znode->cnext = c->cnext;
break;
}
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..3375bbe0508c 100644
--- a/fs/ubifs/ubifs.h
+++ b/fs/ubifs/ubifs.h
@@ -158,13 +158,6 @@
#endif
/*
- * The UBIFS sysfs directory name pattern and maximum name length (3 for "ubi"
- * + 1 for "_" and plus 2x2 for 2 UBI numbers and 1 for the trailing zero byte.
- */
-#define UBIFS_DFS_DIR_NAME "ubi%d_%d"
-#define UBIFS_DFS_DIR_LEN (3 + 1 + 2*2 + 1)
-
-/*
* Lockdep classes for UBIFS inode @ui_mutex.
*/
enum {
@@ -923,8 +916,6 @@ struct ubifs_budget_req {
* @rb: rb-tree node of rb-tree of orphans sorted by inode number
* @list: list head of list of orphans in order added
* @new_list: list head of list of orphans added since the last commit
- * @child_list: list of xattr children if this orphan hosts xattrs, list head
- * if this orphan is a xattr, not used otherwise.
* @cnext: next orphan to commit
* @dnext: next orphan to delete
* @inum: inode number
@@ -936,7 +927,6 @@ struct ubifs_orphan {
struct rb_node rb;
struct list_head list;
struct list_head new_list;
- struct list_head child_list;
struct ubifs_orphan *cnext;
struct ubifs_orphan *dnext;
ino_t inum;
@@ -1047,6 +1037,8 @@ 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 +1297,8 @@ 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;
@@ -1799,7 +1793,7 @@ int ubifs_consolidate_log(struct ubifs_info *c);
/* journal.c */
int ubifs_jnl_update(struct ubifs_info *c, const struct inode *dir,
const struct fscrypt_name *nm, const struct inode *inode,
- int deletion, int xent);
+ int deletion, int xent, int in_orphan);
int ubifs_jnl_write_data(struct ubifs_info *c, const struct inode *inode,
const union ubifs_key *key, const void *buf, int len);
int ubifs_jnl_write_inode(struct ubifs_info *c, const struct inode *inode);
@@ -1816,7 +1810,7 @@ int ubifs_jnl_rename(struct ubifs_info *c, const struct inode *old_dir,
const struct inode *new_dir,
const struct inode *new_inode,
const struct fscrypt_name *new_nm,
- const struct inode *whiteout, int sync);
+ const struct inode *whiteout, int sync, int delete_orphan);
int ubifs_jnl_truncate(struct ubifs_info *c, const struct inode *inode,
loff_t old_size, loff_t new_size);
int ubifs_jnl_delete_xattr(struct ubifs_info *c, const struct inode *host,
@@ -1903,6 +1897,7 @@ 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);
@@ -2045,13 +2040,10 @@ ssize_t ubifs_xattr_get(struct inode *host, const char *name, void *buf,
#ifdef CONFIG_UBIFS_FS_XATTR
extern const struct xattr_handler * const ubifs_xattr_handlers[];
ssize_t ubifs_listxattr(struct dentry *dentry, char *buffer, size_t size);
-void ubifs_evict_xattr_inode(struct ubifs_info *c, ino_t xattr_inum);
int ubifs_purge_xattrs(struct inode *host);
#else
#define ubifs_listxattr NULL
#define ubifs_xattr_handlers NULL
-static inline void ubifs_evict_xattr_inode(struct ubifs_info *c,
- ino_t xattr_inum) { }
static inline int ubifs_purge_xattrs(struct inode *host)
{
return 0;
diff --git a/fs/ubifs/xattr.c b/fs/ubifs/xattr.c
index 0847db521984..c21a0c2b3e90 100644
--- a/fs/ubifs/xattr.c
+++ b/fs/ubifs/xattr.c
@@ -48,19 +48,6 @@
#include <linux/slab.h>
#include <linux/xattr.h>
-/*
- * Extended attribute type constants.
- *
- * USER_XATTR: user extended attribute ("user.*")
- * TRUSTED_XATTR: trusted extended attribute ("trusted.*)
- * SECURITY_XATTR: security extended attribute ("security.*")
- */
-enum {
- USER_XATTR,
- TRUSTED_XATTR,
- SECURITY_XATTR,
-};
-
static const struct inode_operations empty_iops;
static const struct file_operations empty_fops;
@@ -149,7 +136,7 @@ static int create_xattr(struct ubifs_info *c, struct inode *host,
if (strcmp(fname_name(nm), UBIFS_XATTR_NAME_ENCRYPTION_CONTEXT) == 0)
host_ui->flags |= UBIFS_CRYPT_FL;
- err = ubifs_jnl_update(c, host, nm, inode, 0, 1);
+ err = ubifs_jnl_update(c, host, nm, inode, 0, 1, 0);
if (err)
goto out_cancel;
ubifs_set_inode_flags(host);
@@ -532,8 +519,6 @@ int ubifs_purge_xattrs(struct inode *host)
ubifs_err(c, "dead directory entry '%s', error %d",
xent->name, err);
ubifs_ro_mode(c, err);
- kfree(pxent);
- kfree(xent);
goto out_err;
}
@@ -541,16 +526,12 @@ int ubifs_purge_xattrs(struct inode *host)
clear_nlink(xino);
err = remove_xattr(c, host, xino, &nm);
+ iput(xino);
if (err) {
- kfree(pxent);
- kfree(xent);
- iput(xino);
ubifs_err(c, "cannot remove xattr, error %d", err);
goto out_err;
}
- iput(xino);
-
kfree(pxent);
pxent = xent;
key_read(c, &xent->key, &key);
@@ -566,32 +547,12 @@ int ubifs_purge_xattrs(struct inode *host)
return 0;
out_err:
+ kfree(pxent);
+ kfree(xent);
up_write(&ubifs_inode(host)->xattr_sem);
return err;
}
-/**
- * ubifs_evict_xattr_inode - Evict an xattr inode.
- * @c: UBIFS file-system description object
- * @xattr_inum: xattr inode number
- *
- * When an inode that hosts xattrs is being removed we have to make sure
- * that cached inodes of the xattrs also get removed from the inode cache
- * otherwise we'd waste memory. This function looks up an inode from the
- * inode cache and clears the link counter such that iput() will evict
- * the inode.
- */
-void ubifs_evict_xattr_inode(struct ubifs_info *c, ino_t xattr_inum)
-{
- struct inode *inode;
-
- inode = ilookup(c->vfs_sb, xattr_inum);
- if (inode) {
- clear_nlink(inode);
- iput(inode);
- }
-}
-
static int ubifs_xattr_remove(struct inode *host, const char *name)
{
struct inode *inode;
generated by cgit (git 2.34.1) at 2025-07-30 18:51:58 +0000