1 files changed, 136 insertions, 81 deletions

diff --git a/fs/ceph/inode.c b/fs/ceph/inode.c
index 315ef02f9a3f..2966f88310e3 100644
--- a/fs/ceph/inode.c
+++ b/fs/ceph/inode.c
@@ -55,6 +55,52 @@ static int ceph_set_ino_cb(struct inode *inode, void *data)
 	return 0;
 }
 
+/*
+ * Check if the parent inode matches the vino from directory reply info
+ */
+static inline bool ceph_vino_matches_parent(struct inode *parent,
+					    struct ceph_vino vino)
+{
+	return ceph_ino(parent) == vino.ino && ceph_snap(parent) == vino.snap;
+}
+
+/*
+ * Validate that the directory inode referenced by @req->r_parent matches the
+ * inode number and snapshot id contained in the reply's directory record.  If
+ * they do not match – which can theoretically happen if the parent dentry was
+ * moved between the time the request was issued and the reply arrived – fall
+ * back to looking up the correct inode in the inode cache.
+ *
+ * A reference is *always* returned.  Callers that receive a different inode
+ * than the original @parent are responsible for dropping the extra reference
+ * once the reply has been processed.
+ */
+static struct inode *ceph_get_reply_dir(struct super_block *sb,
+					struct inode *parent,
+					struct ceph_mds_reply_info_parsed *rinfo)
+{
+	struct ceph_vino vino;
+
+	if (unlikely(!rinfo->diri.in))
+		return parent; /* nothing to compare against */
+
+	/* If we didn't have a cached parent inode to begin with, just bail out. */
+	if (!parent)
+		return NULL;
+
+	vino.ino  = le64_to_cpu(rinfo->diri.in->ino);
+	vino.snap = le64_to_cpu(rinfo->diri.in->snapid);
+
+	if (likely(ceph_vino_matches_parent(parent, vino)))
+		return parent; /* matches – use the original reference */
+
+	/* Mismatch – this should be rare.  Emit a WARN and obtain the correct inode. */
+	WARN_ONCE(1, "ceph: reply dir mismatch (parent valid %llx.%llx reply %llx.%llx)\n",
+		  ceph_ino(parent), ceph_snap(parent), vino.ino, vino.snap);
+
+	return ceph_get_inode(sb, vino, NULL);
+}
+
 /**
  * ceph_new_inode - allocate a new inode in advance of an expected create
  * @dir: parent directory for new inode
@@ -86,7 +132,7 @@ struct inode *ceph_new_inode(struct inode *dir, struct dentry *dentry,
 			goto out_err;
 	}
 
-	inode->i_state = 0;
+	inode_state_assign_raw(inode, 0);
 	inode->i_mode = *mode;
 
 	err = ceph_security_init_secctx(dentry, *mode, as_ctx);
@@ -155,12 +201,12 @@ struct inode *ceph_get_inode(struct super_block *sb, struct ceph_vino vino,
 
 	doutc(cl, "on %llx=%llx.%llx got %p new %d\n",
 	      ceph_present_inode(inode), ceph_vinop(inode), inode,
-	      !!(inode->i_state & I_NEW));
+	      !!(inode_state_read_once(inode) & I_NEW));
 	return inode;
 }
 
 /*
- * get/constuct snapdir inode for a given directory
+ * get/construct snapdir inode for a given directory
  */
 struct inode *ceph_get_snapdir(struct inode *parent)
 {
@@ -182,7 +228,7 @@ struct inode *ceph_get_snapdir(struct inode *parent)
 		goto err;
 	}
 
-	if (!(inode->i_state & I_NEW) && !S_ISDIR(inode->i_mode)) {
+	if (!(inode_state_read_once(inode) & I_NEW) && !S_ISDIR(inode->i_mode)) {
 		pr_warn_once_client(cl, "bad snapdir inode type (mode=0%o)\n",
 				    inode->i_mode);
 		goto err;
@@ -215,7 +261,7 @@ struct inode *ceph_get_snapdir(struct inode *parent)
 		}
 	}
 #endif
-	if (inode->i_state & I_NEW) {
+	if (inode_state_read_once(inode) & I_NEW) {
 		inode->i_op = &ceph_snapdir_iops;
 		inode->i_fop = &ceph_snapdir_fops;
 		ci->i_snap_caps = CEPH_CAP_PIN; /* so we can open */
@@ -224,7 +270,7 @@ struct inode *ceph_get_snapdir(struct inode *parent)
 
 	return inode;
 err:
-	if ((inode->i_state & I_NEW))
+	if ((inode_state_read_once(inode) & I_NEW))
 		discard_new_inode(inode);
 	else
 		iput(inode);
@@ -665,6 +711,7 @@ struct inode *ceph_alloc_inode(struct super_block *sb)
 	ci->i_work_mask = 0;
 	memset(&ci->i_btime, '\0', sizeof(ci->i_btime));
 #ifdef CONFIG_FS_ENCRYPTION
+	ci->i_crypt_info = NULL;
 	ci->fscrypt_auth = NULL;
 	ci->fscrypt_auth_len = 0;
 #endif
@@ -697,7 +744,7 @@ void ceph_evict_inode(struct inode *inode)
 
 	netfs_wait_for_outstanding_io(inode);
 	truncate_inode_pages_final(&inode->i_data);
-	if (inode->i_state & I_PINNING_NETFS_WB)
+	if (inode_state_read_once(inode) & I_PINNING_NETFS_WB)
 		ceph_fscache_unuse_cookie(inode, true);
 	clear_inode(inode);
 
@@ -832,7 +879,9 @@ void ceph_fill_file_time(struct inode *inode, int issued,
 {
 	struct ceph_client *cl = ceph_inode_to_client(inode);
 	struct ceph_inode_info *ci = ceph_inode(inode);
+	struct timespec64 iatime = inode_get_atime(inode);
 	struct timespec64 ictime = inode_get_ctime(inode);
+	struct timespec64 imtime = inode_get_mtime(inode);
 	int warn = 0;
 
 	if (issued & (CEPH_CAP_FILE_EXCL|
@@ -842,39 +891,26 @@ void ceph_fill_file_time(struct inode *inode, int issued,
 		      CEPH_CAP_XATTR_EXCL)) {
 		if (ci->i_version == 0 ||
 		    timespec64_compare(ctime, &ictime) > 0) {
-			doutc(cl, "ctime %lld.%09ld -> %lld.%09ld inc w/ cap\n",
-			     ictime.tv_sec, ictime.tv_nsec,
-			     ctime->tv_sec, ctime->tv_nsec);
+			doutc(cl, "ctime %ptSp -> %ptSp inc w/ cap\n", &ictime, ctime);
 			inode_set_ctime_to_ts(inode, *ctime);
 		}
 		if (ci->i_version == 0 ||
 		    ceph_seq_cmp(time_warp_seq, ci->i_time_warp_seq) > 0) {
 			/* the MDS did a utimes() */
-			doutc(cl, "mtime %lld.%09ld -> %lld.%09ld tw %d -> %d\n",
-			     inode_get_mtime_sec(inode),
-			     inode_get_mtime_nsec(inode),
-			     mtime->tv_sec, mtime->tv_nsec,
-			     ci->i_time_warp_seq, (int)time_warp_seq);
+			doutc(cl, "mtime %ptSp -> %ptSp tw %d -> %d\n", &imtime, mtime,
+			      ci->i_time_warp_seq, (int)time_warp_seq);
 
 			inode_set_mtime_to_ts(inode, *mtime);
 			inode_set_atime_to_ts(inode, *atime);
 			ci->i_time_warp_seq = time_warp_seq;
 		} else if (time_warp_seq == ci->i_time_warp_seq) {
-			struct timespec64	ts;
-
 			/* nobody did utimes(); take the max */
-			ts = inode_get_mtime(inode);
-			if (timespec64_compare(mtime, &ts) > 0) {
-				doutc(cl, "mtime %lld.%09ld -> %lld.%09ld inc\n",
-				     ts.tv_sec, ts.tv_nsec,
-				     mtime->tv_sec, mtime->tv_nsec);
+			if (timespec64_compare(mtime, &imtime) > 0) {
+				doutc(cl, "mtime %ptSp -> %ptSp inc\n", &imtime, mtime);
 				inode_set_mtime_to_ts(inode, *mtime);
 			}
-			ts = inode_get_atime(inode);
-			if (timespec64_compare(atime, &ts) > 0) {
-				doutc(cl, "atime %lld.%09ld -> %lld.%09ld inc\n",
-				     ts.tv_sec, ts.tv_nsec,
-				     atime->tv_sec, atime->tv_nsec);
+			if (timespec64_compare(atime, &iatime) > 0) {
+				doutc(cl, "atime %ptSp -> %ptSp inc\n", &iatime, atime);
 				inode_set_atime_to_ts(inode, *atime);
 			}
 		} else if (issued & CEPH_CAP_FILE_EXCL) {
@@ -911,7 +947,7 @@ static int decode_encrypted_symlink(struct ceph_mds_client *mdsc,
 	if (!sym)
 		return -ENOMEM;
 
-	declen = ceph_base64_decode(encsym, enclen, sym);
+	declen = base64_decode(encsym, enclen, sym, false, BASE64_IMAP);
 	if (declen < 0) {
 		pr_err_client(cl,
 			"can't decode symlink (%d). Content: %.*s\n",
@@ -966,7 +1002,7 @@ int ceph_fill_inode(struct inode *inode, struct page *locked_page,
 	      le64_to_cpu(info->version), ci->i_version);
 
 	/* Once I_NEW is cleared, we can't change type or dev numbers */
-	if (inode->i_state & I_NEW) {
+	if (inode_state_read_once(inode) & I_NEW) {
 		inode->i_mode = mode;
 	} else {
 		if (inode_wrong_type(inode, mode)) {
@@ -1043,7 +1079,7 @@ int ceph_fill_inode(struct inode *inode, struct page *locked_page,
 
 #ifdef CONFIG_FS_ENCRYPTION
 	if (iinfo->fscrypt_auth_len &&
-	    ((inode->i_state & I_NEW) || (ci->fscrypt_auth_len == 0))) {
+	    ((inode_state_read_once(inode) & I_NEW) || (ci->fscrypt_auth_len == 0))) {
 		kfree(ci->fscrypt_auth);
 		ci->fscrypt_auth_len = iinfo->fscrypt_auth_len;
 		ci->fscrypt_auth = iinfo->fscrypt_auth;
@@ -1523,6 +1559,7 @@ int ceph_fill_trace(struct super_block *sb, struct ceph_mds_request *req)
 	struct ceph_vino tvino, dvino;
 	struct ceph_fs_client *fsc = ceph_sb_to_fs_client(sb);
 	struct ceph_client *cl = fsc->client;
+	struct inode *parent_dir = NULL;
 	int err = 0;
 
 	doutc(cl, "%p is_dentry %d is_target %d\n", req,
@@ -1536,10 +1573,17 @@ int ceph_fill_trace(struct super_block *sb, struct ceph_mds_request *req)
 	}
 
 	if (rinfo->head->is_dentry) {
-		struct inode *dir = req->r_parent;
-
-		if (dir) {
-			err = ceph_fill_inode(dir, NULL, &rinfo->diri,
+		/*
+		 * r_parent may be stale, in cases when R_PARENT_LOCKED is not set,
+		 * so we need to get the correct inode
+		 */
+		parent_dir = ceph_get_reply_dir(sb, req->r_parent, rinfo);
+		if (unlikely(IS_ERR(parent_dir))) {
+			err = PTR_ERR(parent_dir);
+			goto done;
+		}
+		if (parent_dir) {
+			err = ceph_fill_inode(parent_dir, NULL, &rinfo->diri,
 					      rinfo->dirfrag, session, -1,
 					      &req->r_caps_reservation);
 			if (err < 0)
@@ -1548,14 +1592,14 @@ int ceph_fill_trace(struct super_block *sb, struct ceph_mds_request *req)
 			WARN_ON_ONCE(1);
 		}
 
-		if (dir && req->r_op == CEPH_MDS_OP_LOOKUPNAME &&
+		if (parent_dir && req->r_op == CEPH_MDS_OP_LOOKUPNAME &&
 		    test_bit(CEPH_MDS_R_PARENT_LOCKED, &req->r_req_flags) &&
 		    !test_bit(CEPH_MDS_R_ABORTED, &req->r_req_flags)) {
 			bool is_nokey = false;
 			struct qstr dname;
 			struct dentry *dn, *parent;
 			struct fscrypt_str oname = FSTR_INIT(NULL, 0);
-			struct ceph_fname fname = { .dir	= dir,
+			struct ceph_fname fname = { .dir	= parent_dir,
 						    .name	= rinfo->dname,
 						    .ctext	= rinfo->altname,
 						    .name_len	= rinfo->dname_len,
@@ -1564,10 +1608,10 @@ int ceph_fill_trace(struct super_block *sb, struct ceph_mds_request *req)
 			BUG_ON(!rinfo->head->is_target);
 			BUG_ON(req->r_dentry);
 
-			parent = d_find_any_alias(dir);
+			parent = d_find_any_alias(parent_dir);
 			BUG_ON(!parent);
 
-			err = ceph_fname_alloc_buffer(dir, &oname);
+			err = ceph_fname_alloc_buffer(parent_dir, &oname);
 			if (err < 0) {
 				dput(parent);
 				goto done;
@@ -1576,7 +1620,7 @@ int ceph_fill_trace(struct super_block *sb, struct ceph_mds_request *req)
 			err = ceph_fname_to_usr(&fname, NULL, &oname, &is_nokey);
 			if (err < 0) {
 				dput(parent);
-				ceph_fname_free_buffer(dir, &oname);
+				ceph_fname_free_buffer(parent_dir, &oname);
 				goto done;
 			}
 			dname.name = oname.name;
@@ -1595,7 +1639,7 @@ retry_lookup:
 				      dname.len, dname.name, dn);
 				if (!dn) {
 					dput(parent);
-					ceph_fname_free_buffer(dir, &oname);
+					ceph_fname_free_buffer(parent_dir, &oname);
 					err = -ENOMEM;
 					goto done;
 				}
@@ -1610,12 +1654,12 @@ retry_lookup:
 				    ceph_snap(d_inode(dn)) != tvino.snap)) {
 				doutc(cl, " dn %p points to wrong inode %p\n",
 				      dn, d_inode(dn));
-				ceph_dir_clear_ordered(dir);
+				ceph_dir_clear_ordered(parent_dir);
 				d_delete(dn);
 				dput(dn);
 				goto retry_lookup;
 			}
-			ceph_fname_free_buffer(dir, &oname);
+			ceph_fname_free_buffer(parent_dir, &oname);
 
 			req->r_dentry = dn;
 			dput(parent);
@@ -1637,13 +1681,13 @@ retry_lookup:
 			pr_err_client(cl, "badness %p %llx.%llx\n", in,
 				      ceph_vinop(in));
 			req->r_target_inode = NULL;
-			if (in->i_state & I_NEW)
+			if (inode_state_read_once(in) & I_NEW)
 				discard_new_inode(in);
 			else
 				iput(in);
 			goto done;
 		}
-		if (in->i_state & I_NEW)
+		if (inode_state_read_once(in) & I_NEW)
 			unlock_new_inode(in);
 	}
 
@@ -1739,6 +1783,11 @@ retry_lookup:
 			goto done;
 		}
 
+		if (unlikely(!in)) {
+			err = -EINVAL;
+			goto done;
+		}
+
 		/* attach proper inode */
 		if (d_really_is_negative(dn)) {
 			ceph_dir_clear_ordered(dir);
@@ -1774,6 +1823,12 @@ retry_lookup:
 		doutc(cl, " linking snapped dir %p to dn %p\n", in,
 		      req->r_dentry);
 		ceph_dir_clear_ordered(dir);
+
+		if (unlikely(!in)) {
+			err = -EINVAL;
+			goto done;
+		}
+
 		ihold(in);
 		err = splice_dentry(&req->r_dentry, in);
 		if (err < 0)
@@ -1794,6 +1849,9 @@ retry_lookup:
 					    &dvino, ptvino);
 	}
 done:
+	/* Drop extra ref from ceph_get_reply_dir() if it returned a new inode */
+	if (unlikely(!IS_ERR_OR_NULL(parent_dir) && parent_dir != req->r_parent))
+		iput(parent_dir);
 	doutc(cl, "done err=%d\n", err);
 	return err;
 }
@@ -1829,11 +1887,11 @@ static int readdir_prepopulate_inodes_only(struct ceph_mds_request *req,
 			pr_err_client(cl, "inode badness on %p got %d\n", in,
 				      rc);
 			err = rc;
-			if (in->i_state & I_NEW) {
+			if (inode_state_read_once(in) & I_NEW) {
 				ihold(in);
 				discard_new_inode(in);
 			}
-		} else if (in->i_state & I_NEW) {
+		} else if (inode_state_read_once(in) & I_NEW) {
 			unlock_new_inode(in);
 		}
 
@@ -1845,10 +1903,9 @@ static int readdir_prepopulate_inodes_only(struct ceph_mds_request *req,
 
 void ceph_readdir_cache_release(struct ceph_readdir_cache_control *ctl)
 {
-	if (ctl->page) {
-		kunmap(ctl->page);
-		put_page(ctl->page);
-		ctl->page = NULL;
+	if (ctl->folio) {
+		folio_release_kmap(ctl->folio, ctl->dentries);
+		ctl->folio = NULL;
 	}
 }
 
@@ -1862,20 +1919,26 @@ static int fill_readdir_cache(struct inode *dir, struct dentry *dn,
 	unsigned idx = ctl->index % nsize;
 	pgoff_t pgoff = ctl->index / nsize;
 
-	if (!ctl->page || pgoff != ctl->page->index) {
+	if (!ctl->folio || pgoff != ctl->folio->index) {
 		ceph_readdir_cache_release(ctl);
+		fgf_t fgf = FGP_LOCK;
+
 		if (idx == 0)
-			ctl->page = grab_cache_page(&dir->i_data, pgoff);
-		else
-			ctl->page = find_lock_page(&dir->i_data, pgoff);
-		if (!ctl->page) {
+			fgf |= FGP_ACCESSED | FGP_CREAT;
+
+		ctl->folio = __filemap_get_folio(&dir->i_data, pgoff,
+				fgf, mapping_gfp_mask(&dir->i_data));
+		if (IS_ERR(ctl->folio)) {
+			int err = PTR_ERR(ctl->folio);
+
+			ctl->folio = NULL;
 			ctl->index = -1;
-			return idx == 0 ? -ENOMEM : 0;
+			return idx == 0 ? err : 0;
 		}
 		/* reading/filling the cache are serialized by
-		 * i_rwsem, no need to use page lock */
-		unlock_page(ctl->page);
-		ctl->dentries = kmap(ctl->page);
+		 * i_rwsem, no need to use folio lock */
+		folio_unlock(ctl->folio);
+		ctl->dentries = kmap_local_folio(ctl->folio, 0);
 		if (idx == 0)
 			memset(ctl->dentries, 0, PAGE_SIZE);
 	}
@@ -2040,7 +2103,7 @@ retry_lookup:
 			pr_err_client(cl, "badness on %p %llx.%llx\n", in,
 				      ceph_vinop(in));
 			if (d_really_is_negative(dn)) {
-				if (in->i_state & I_NEW) {
+				if (inode_state_read_once(in) & I_NEW) {
 					ihold(in);
 					discard_new_inode(in);
 				}
@@ -2050,7 +2113,7 @@ retry_lookup:
 			err = ret;
 			goto next_item;
 		}
-		if (in->i_state & I_NEW)
+		if (inode_state_read_once(in) & I_NEW)
 			unlock_new_inode(in);
 
 		if (d_really_is_negative(dn)) {
@@ -2362,7 +2425,7 @@ static int fill_fscrypt_truncate(struct inode *inode,
 
 	/* Try to writeback the dirty pagecaches */
 	if (issued & (CEPH_CAP_FILE_BUFFER)) {
-		loff_t lend = orig_pos + CEPH_FSCRYPT_BLOCK_SHIFT - 1;
+		loff_t lend = orig_pos + CEPH_FSCRYPT_BLOCK_SIZE - 1;
 
 		ret = filemap_write_and_wait_range(inode->i_mapping,
 						   orig_pos, lend);
@@ -2431,8 +2494,7 @@ static int fill_fscrypt_truncate(struct inode *inode,
 		/* encrypt the last block */
 		ret = ceph_fscrypt_encrypt_block_inplace(inode, page,
 						    CEPH_FSCRYPT_BLOCK_SIZE,
-						    0, block,
-						    GFP_KERNEL);
+						    0, block);
 		if (ret)
 			goto out;
 	}
@@ -2483,22 +2545,21 @@ int __ceph_setattr(struct mnt_idmap *idmap, struct inode *inode,
 	int truncate_retry = 20; /* The RMW will take around 50ms */
 	struct dentry *dentry;
 	char *path;
-	int pathlen;
-	u64 pathbase;
 	bool do_sync = false;
 
 	dentry = d_find_alias(inode);
 	if (!dentry) {
 		do_sync = true;
 	} else {
-		path = ceph_mdsc_build_path(mdsc, dentry, &pathlen, &pathbase, 0);
+		struct ceph_path_info path_info;
+		path = ceph_mdsc_build_path(mdsc, dentry, &path_info, 0);
 		if (IS_ERR(path)) {
 			do_sync = true;
 			err = 0;
 		} else {
 			err = ceph_mds_check_access(mdsc, path, MAY_WRITE);
 		}
-		ceph_mdsc_free_path(path, pathlen);
+		ceph_mdsc_free_path_info(&path_info);
 		dput(dentry);
 
 		/* For none EACCES cases will let the MDS do the mds auth check */
@@ -2631,10 +2692,8 @@ retry:
 	if (ia_valid & ATTR_ATIME) {
 		struct timespec64 atime = inode_get_atime(inode);
 
-		doutc(cl, "%p %llx.%llx atime %lld.%09ld -> %lld.%09ld\n",
-		      inode, ceph_vinop(inode),
-		      atime.tv_sec, atime.tv_nsec,
-		      attr->ia_atime.tv_sec, attr->ia_atime.tv_nsec);
+		doutc(cl, "%p %llx.%llx atime %ptSp -> %ptSp\n",
+		      inode, ceph_vinop(inode), &atime, &attr->ia_atime);
 		if (!do_sync && (issued & CEPH_CAP_FILE_EXCL)) {
 			ci->i_time_warp_seq++;
 			inode_set_atime_to_ts(inode, attr->ia_atime);
@@ -2708,10 +2767,8 @@ retry:
 	if (ia_valid & ATTR_MTIME) {
 		struct timespec64 mtime = inode_get_mtime(inode);
 
-		doutc(cl, "%p %llx.%llx mtime %lld.%09ld -> %lld.%09ld\n",
-		      inode, ceph_vinop(inode),
-		      mtime.tv_sec, mtime.tv_nsec,
-		      attr->ia_mtime.tv_sec, attr->ia_mtime.tv_nsec);
+		doutc(cl, "%p %llx.%llx mtime %ptSp -> %ptSp\n",
+		      inode, ceph_vinop(inode), &mtime, &attr->ia_mtime);
 		if (!do_sync && (issued & CEPH_CAP_FILE_EXCL)) {
 			ci->i_time_warp_seq++;
 			inode_set_mtime_to_ts(inode, attr->ia_mtime);
@@ -2732,13 +2789,11 @@ retry:
 
 	/* these do nothing */
 	if (ia_valid & ATTR_CTIME) {
+		struct timespec64 ictime = inode_get_ctime(inode);
 		bool only = (ia_valid & (ATTR_SIZE|ATTR_MTIME|ATTR_ATIME|
 					 ATTR_MODE|ATTR_UID|ATTR_GID)) == 0;
-		doutc(cl, "%p %llx.%llx ctime %lld.%09ld -> %lld.%09ld (%s)\n",
-		      inode, ceph_vinop(inode),
-		      inode_get_ctime_sec(inode),
-		      inode_get_ctime_nsec(inode),
-		      attr->ia_ctime.tv_sec, attr->ia_ctime.tv_nsec,
+		doutc(cl, "%p %llx.%llx ctime %ptSp -> %ptSp (%s)\n",
+		      inode, ceph_vinop(inode), &ictime, &attr->ia_ctime,
 		      only ? "ctime only" : "ignored");
 		if (only) {
 			/*