ceph: fix handling of "meta" errors

Currently, we check the wb_err too early for directories, before all of
the unsafe child requests have been waited on. In order to fix that we
need to check the mapping->wb_err later nearer to the end of ceph_fsync.

We also have an overly-complex method for tracking errors after
blocklisting. The errors recorded in cleanup_session_requests go to a
completely separate field in the inode, but we end up reporting them the
same way we would for any other error (in fsync).

There's no real benefit to tracking these errors in two different
places, since the only reporting mechanism for them is in fsync, and
we'd need to advance them both every time.

Given that, we can just remove i_meta_err, and convert the places that
used it to instead just use mapping->wb_err instead. That also fixes
the original problem by ensuring that we do a check_and_advance of the
wb_err at the end of the fsync op.

Cc: stable@vger.kernel.org
URL: https://tracker.ceph.com/issues/52864
Reported-by: Patrick Donnelly <pdonnell@redhat.com>
Signed-off-by: Jeff Layton <jlayton@kernel.org>
Reviewed-by: Xiubo Li <xiubli@redhat.com>
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>

5 files changed, 8 insertions, 27 deletions

diff --git a/fs/ceph/caps.c b/fs/ceph/caps.c
index 3e42d0466521..8f537f1d9d1d 100644
--- a/fs/ceph/caps.c
+++ b/fs/ceph/caps.c
@@ -2330,7 +2330,6 @@ retry:
 
 int ceph_fsync(struct file *file, loff_t start, loff_t end, int datasync)
 {
-	struct ceph_file_info *fi = file->private_data;
 	struct inode *inode = file->f_mapping->host;
 	struct ceph_inode_info *ci = ceph_inode(inode);
 	u64 flush_tid;
@@ -2365,14 +2364,9 @@ int ceph_fsync(struct file *file, loff_t start, loff_t end, int datasync)
 	if (err < 0)
 		ret = err;
 
-	if (errseq_check(&ci->i_meta_err, READ_ONCE(fi->meta_err))) {
-		spin_lock(&file->f_lock);
-		err = errseq_check_and_advance(&ci->i_meta_err,
-					       &fi->meta_err);
-		spin_unlock(&file->f_lock);
-		if (err < 0)
-			ret = err;
-	}
+	err = file_check_and_advance_wb_err(file);
+	if (err < 0)
+		ret = err;
 out:
 	dout("fsync %p%s result=%d\n", inode, datasync ? " datasync" : "", ret);
 	return ret;
diff --git a/fs/ceph/file.c b/fs/ceph/file.c
index d16fd2d5fd42..e61018d9764e 100644
--- a/fs/ceph/file.c
+++ b/fs/ceph/file.c
@@ -233,7 +233,6 @@ static int ceph_init_file_info(struct inode *inode, struct file *file,
 
 	spin_lock_init(&fi->rw_contexts_lock);
 	INIT_LIST_HEAD(&fi->rw_contexts);
-	fi->meta_err = errseq_sample(&ci->i_meta_err);
 	fi->filp_gen = READ_ONCE(ceph_inode_to_client(inode)->filp_gen);
 
 	return 0;
diff --git a/fs/ceph/inode.c b/fs/ceph/inode.c
index 2df1e1284451..1c7574105478 100644
--- a/fs/ceph/inode.c
+++ b/fs/ceph/inode.c
@@ -541,8 +541,6 @@ struct inode *ceph_alloc_inode(struct super_block *sb)
 
 	ceph_fscache_inode_init(ci);
 
-	ci->i_meta_err = 0;
-
 	return &ci->vfs_inode;
 }
 
diff --git a/fs/ceph/mds_client.c b/fs/ceph/mds_client.c
index 7cad180d6deb..d64413adc0fd 100644
--- a/fs/ceph/mds_client.c
+++ b/fs/ceph/mds_client.c
@@ -1493,7 +1493,6 @@ static void cleanup_session_requests(struct ceph_mds_client *mdsc,
 {
 	struct ceph_mds_request *req;
 	struct rb_node *p;
-	struct ceph_inode_info *ci;
 
 	dout("cleanup_session_requests mds%d\n", session->s_mds);
 	mutex_lock(&mdsc->mutex);
@@ -1502,16 +1501,10 @@ static void cleanup_session_requests(struct ceph_mds_client *mdsc,
 				       struct ceph_mds_request, r_unsafe_item);
 		pr_warn_ratelimited(" dropping unsafe request %llu\n",
 				    req->r_tid);
-		if (req->r_target_inode) {
-			/* dropping unsafe change of inode's attributes */
-			ci = ceph_inode(req->r_target_inode);
-			errseq_set(&ci->i_meta_err, -EIO);
-		}
-		if (req->r_unsafe_dir) {
-			/* dropping unsafe directory operation */
-			ci = ceph_inode(req->r_unsafe_dir);
-			errseq_set(&ci->i_meta_err, -EIO);
-		}
+		if (req->r_target_inode)
+			mapping_set_error(req->r_target_inode->i_mapping, -EIO);
+		if (req->r_unsafe_dir)
+			mapping_set_error(req->r_unsafe_dir->i_mapping, -EIO);
 		__unregister_request(mdsc, req);
 	}
 	/* zero r_attempts, so kick_requests() will re-send requests */
@@ -1678,7 +1671,7 @@ static int remove_session_caps_cb(struct inode *inode, struct ceph_cap *cap,
 		spin_unlock(&mdsc->cap_dirty_lock);
 
 		if (dirty_dropped) {
-			errseq_set(&ci->i_meta_err, -EIO);
+			mapping_set_error(inode->i_mapping, -EIO);
 
 			if (ci->i_wrbuffer_ref_head == 0 &&
 			    ci->i_wr_ref == 0 &&
diff --git a/fs/ceph/super.h b/fs/ceph/super.h
index a40eb14c282a..14f951cd5b61 100644
--- a/fs/ceph/super.h
+++ b/fs/ceph/super.h
@@ -429,8 +429,6 @@ struct ceph_inode_info {
 #ifdef CONFIG_CEPH_FSCACHE
 	struct fscache_cookie *fscache;
 #endif
-	errseq_t i_meta_err;
-
 	struct inode vfs_inode; /* at end */
 };
 
@@ -774,7 +772,6 @@ struct ceph_file_info {
 	spinlock_t rw_contexts_lock;
 	struct list_head rw_contexts;
 
-	errseq_t meta_err;
 	u32 filp_gen;
 	atomic_t num_locks;
 };