From: Filipe Manana fdmanana@suse.com
[ Upstream commit 418b09027743d9a9fb39116bed46a192f868a3c3 ]
When FIEMAP_FLAG_SYNC is given to fiemap the expectation is that that are no concurrent writes and we get a stable view of the inode's extent layout.
When the flag is given we flush all IO (and wait for ordered extents to complete) and then lock the inode in shared mode, however that leaves open the possibility that a write might happen right after the flushing and before locking the inode. So fix this by flushing again after locking the inode - we leave the initial flushing before locking the inode to avoid holding the lock and blocking other RO operations while waiting for IO and ordered extents to complete. The second flushing while holding the inode's lock will most of the time do nothing or very little since the time window for new writes to have happened is small.
CC: stable@vger.kernel.org # 6.1 Reviewed-by: Josef Bacik josef@toxicpanda.com Signed-off-by: Filipe Manana fdmanana@suse.com Signed-off-by: David Sterba dsterba@suse.com [ backport to v6.1.158 ] Signed-off-by: Andrey Kalachev kalachev@swemel.ru --- fs/btrfs/extent_io.c | 20 ++++++++------------ fs/btrfs/inode.c | 22 +++++++++++++++++++++- 2 files changed, 29 insertions(+), 13 deletions(-)
diff --git a/fs/btrfs/extent_io.c b/fs/btrfs/extent_io.c index 2c2c66c26869..4f5e5b6380f7 100644 --- a/fs/btrfs/extent_io.c +++ b/fs/btrfs/extent_io.c @@ -4050,17 +4050,15 @@ int extent_fiemap(struct btrfs_inode *inode, struct fiemap_extent_info *fieinfo, range_end = round_up(start + len, sectorsize); prev_extent_end = range_start;
- btrfs_inode_lock(&inode->vfs_inode, BTRFS_ILOCK_SHARED); - ret = fiemap_find_last_extent_offset(inode, path, &last_extent_end); if (ret < 0) - goto out_unlock; + goto out; btrfs_release_path(path);
path->reada = READA_FORWARD; ret = fiemap_search_slot(inode, path, range_start); if (ret < 0) { - goto out_unlock; + goto out; } else if (ret > 0) { /* * No file extent item found, but we may have delalloc between @@ -4105,7 +4103,7 @@ int extent_fiemap(struct btrfs_inode *inode, struct fiemap_extent_info *fieinfo, roots, tmp_ulist, prev_extent_end, hole_end); if (ret < 0) { - goto out_unlock; + goto out; } else if (ret > 0) { /* fiemap_fill_next_extent() told us to stop. */ stopped = true; @@ -4162,7 +4160,7 @@ int extent_fiemap(struct btrfs_inode *inode, struct fiemap_extent_info *fieinfo, tmp_ulist, backref_cache); if (ret < 0) - goto out_unlock; + goto out; else if (ret > 0) flags |= FIEMAP_EXTENT_SHARED; } @@ -4173,7 +4171,7 @@ int extent_fiemap(struct btrfs_inode *inode, struct fiemap_extent_info *fieinfo, }
if (ret < 0) { - goto out_unlock; + goto out; } else if (ret > 0) { /* fiemap_fill_next_extent() told us to stop. */ stopped = true; @@ -4184,12 +4182,12 @@ int extent_fiemap(struct btrfs_inode *inode, struct fiemap_extent_info *fieinfo, next_item: if (fatal_signal_pending(current)) { ret = -EINTR; - goto out_unlock; + goto out; }
ret = fiemap_next_leaf_item(inode, path); if (ret < 0) { - goto out_unlock; + goto out; } else if (ret > 0) { /* No more file extent items for this inode. */ break; @@ -4213,7 +4211,7 @@ int extent_fiemap(struct btrfs_inode *inode, struct fiemap_extent_info *fieinfo, 0, 0, 0, roots, tmp_ulist, prev_extent_end, range_end - 1); if (ret < 0) - goto out_unlock; + goto out; prev_extent_end = range_end; }
@@ -4251,8 +4249,6 @@ int extent_fiemap(struct btrfs_inode *inode, struct fiemap_extent_info *fieinfo,
ret = emit_last_fiemap_cache(fieinfo, &cache);
-out_unlock: - btrfs_inode_unlock(&inode->vfs_inode, BTRFS_ILOCK_SHARED); out: kfree(backref_cache); btrfs_free_path(path); diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c index 78cd7b8ccfc8..a389d0b06fea 100644 --- a/fs/btrfs/inode.c +++ b/fs/btrfs/inode.c @@ -8269,6 +8269,7 @@ struct iomap_dio *btrfs_dio_write(struct kiocb *iocb, struct iov_iter *iter, static int btrfs_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo, u64 start, u64 len) { + struct btrfs_inode *btrfs_inode = BTRFS_I(inode); int ret;
ret = fiemap_prep(inode, fieinfo, start, &len, 0); @@ -8294,7 +8295,26 @@ static int btrfs_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo, return ret; }
- return extent_fiemap(BTRFS_I(inode), fieinfo, start, len); + btrfs_inode_lock(inode, BTRFS_ILOCK_SHARED); + + /* + * We did an initial flush to avoid holding the inode's lock while + * triggering writeback and waiting for the completion of IO and ordered + * extents. Now after we locked the inode we do it again, because it's + * possible a new write may have happened in between those two steps. + */ + if (fieinfo->fi_flags & FIEMAP_FLAG_SYNC) { + ret = btrfs_wait_ordered_range(inode, 0, LLONG_MAX); + if (ret) { + btrfs_inode_unlock(inode, BTRFS_ILOCK_SHARED); + return ret; + } + } + + ret = extent_fiemap(btrfs_inode, fieinfo, start, len); + btrfs_inode_unlock(inode, BTRFS_ILOCK_SHARED); + + return ret; }
static int btrfs_writepages(struct address_space *mapping,