This is a note to let you know that I've just added the patch titled gfs2: Fixes to "Implement iomap for block_map" to the 4.15-stable tree which can be found at: http://www.kernel.org/git/?p=linux/kernel/git/stable/stable-queue.git;a=summary The filename of the patch is: gfs2-fixes-to-implement-iomap-for-block_map.patch and it can be found in the queue-4.15 subdirectory. If you, or anyone else, feels it should not be added to the stable tree, please let <stable@xxxxxxxxxxxxxxx> know about it. >From 49edd5bf429c405b3a7f75503845d9f66a47dd4b Mon Sep 17 00:00:00 2001 From: Andreas Gruenbacher <agruenba@xxxxxxxxxx> Date: Tue, 6 Feb 2018 07:20:55 -0700 Subject: gfs2: Fixes to "Implement iomap for block_map" From: Andreas Gruenbacher <agruenba@xxxxxxxxxx> commit 49edd5bf429c405b3a7f75503845d9f66a47dd4b upstream. It turns out that commit 3974320ca6 "Implement iomap for block_map" introduced a few bugs that trigger occasional failures with xfstest generic/476: In gfs2_iomap_begin, we jump to do_alloc when we determine that we are beyond the end of the allocated metadata (height > ip->i_height). There, we can end up calling hole_size with a metapath that doesn't match the current metadata tree, which doesn't make sense. After untangling the code at do_alloc, fix this by checking if the block we are looking for is within the range of allocated metadata. In addition, add a BUG() in case gfs2_iomap_begin is accidentally called for reading stuffed files: this is handled separately. Make sure we don't truncate iomap->length for reads beyond the end of the file; in that case, the entire range counts as a hole. Finally, revert to taking a bitmap write lock when doing allocations. It's unclear why that change didn't lead to any failures during testing. Signed-off-by: Andreas Gruenbacher <agruenba@xxxxxxxxxx> Signed-off-by: Bob Peterson <rpeterso@xxxxxxxxxx> Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx> --- fs/gfs2/bmap.c | 43 +++++++++++++++++++++++-------------------- 1 file changed, 23 insertions(+), 20 deletions(-) --- a/fs/gfs2/bmap.c +++ b/fs/gfs2/bmap.c @@ -736,7 +736,7 @@ int gfs2_iomap_begin(struct inode *inode __be64 *ptr; sector_t lblock; sector_t lend; - int ret; + int ret = 0; int eob; unsigned int len; struct buffer_head *bh; @@ -748,12 +748,14 @@ int gfs2_iomap_begin(struct inode *inode goto out; } - if ((flags & IOMAP_REPORT) && gfs2_is_stuffed(ip)) { - gfs2_stuffed_iomap(inode, iomap); - if (pos >= iomap->length) - return -ENOENT; - ret = 0; - goto out; + if (gfs2_is_stuffed(ip)) { + if (flags & IOMAP_REPORT) { + gfs2_stuffed_iomap(inode, iomap); + if (pos >= iomap->length) + ret = -ENOENT; + goto out; + } + BUG_ON(!(flags & IOMAP_WRITE)); } lblock = pos >> inode->i_blkbits; @@ -764,7 +766,7 @@ int gfs2_iomap_begin(struct inode *inode iomap->type = IOMAP_HOLE; iomap->length = (u64)(lend - lblock) << inode->i_blkbits; iomap->flags = IOMAP_F_MERGED; - bmap_lock(ip, 0); + bmap_lock(ip, flags & IOMAP_WRITE); /* * Directory data blocks have a struct gfs2_meta_header header, so the @@ -807,27 +809,28 @@ int gfs2_iomap_begin(struct inode *inode iomap->flags |= IOMAP_F_BOUNDARY; iomap->length = (u64)len << inode->i_blkbits; - ret = 0; - out_release: release_metapath(&mp); - bmap_unlock(ip, 0); + bmap_unlock(ip, flags & IOMAP_WRITE); out: trace_gfs2_iomap_end(ip, iomap, ret); return ret; do_alloc: - if (!(flags & IOMAP_WRITE)) { - if (pos >= i_size_read(inode)) { + if (flags & IOMAP_WRITE) { + ret = gfs2_iomap_alloc(inode, iomap, flags, &mp); + } else if (flags & IOMAP_REPORT) { + loff_t size = i_size_read(inode); + if (pos >= size) ret = -ENOENT; - goto out_release; - } - ret = 0; - iomap->length = hole_size(inode, lblock, &mp); - goto out_release; + else if (height <= ip->i_height) + iomap->length = hole_size(inode, lblock, &mp); + else + iomap->length = size - pos; + } else { + if (height <= ip->i_height) + iomap->length = hole_size(inode, lblock, &mp); } - - ret = gfs2_iomap_alloc(inode, iomap, flags, &mp); goto out_release; } Patches currently in stable-queue which might be from agruenba@xxxxxxxxxx are queue-4.15/gfs2-fixes-to-implement-iomap-for-block_map.patch