From cf1167d5c1abf3bc42b2a1562bfa7937c05337e2 Mon Sep 17 00:00:00 2001 From: Liu Bo Date: Wed, 20 Sep 2017 17:50:18 -0600 Subject: Btrfs: fix kernel oops while reading compressed data The kernel oops happens at kernel BUG at fs/btrfs/extent_io.c:2104! ... RIP: clean_io_failure+0x263/0x2a0 [btrfs] It's showing that read-repair code is using an improper mirror index. This is due to the fact that compression read's endio hasn't recorded the failed mirror index in %cb->orig_bio. With this, btrfs's read-repair can work properly on reading compressed data. Signed-off-by: Liu Bo Reported-by: Paul Jones Tested-by: Paul Jones Reviewed-by: David Sterba Signed-off-by: David Sterba --- fs/btrfs/compression.c | 9 +++++++++ 1 file changed, 9 insertions(+) (limited to 'fs/btrfs/compression.c') diff --git a/fs/btrfs/compression.c b/fs/btrfs/compression.c index 883ecc58fd0d..c6aa53c4c102 100644 --- a/fs/btrfs/compression.c +++ b/fs/btrfs/compression.c @@ -107,6 +107,7 @@ static void end_compressed_bio_read(struct bio *bio) struct inode *inode; struct page *page; unsigned long index; + unsigned int mirror = btrfs_io_bio(bio)->mirror_num; int ret; if (bio->bi_status) @@ -118,6 +119,14 @@ static void end_compressed_bio_read(struct bio *bio) if (!refcount_dec_and_test(&cb->pending_bios)) goto out; + /* + * Record the correct mirror_num in cb->orig_bio so that + * read-repair can work properly. + */ + ASSERT(btrfs_io_bio(cb->orig_bio)); + btrfs_io_bio(cb->orig_bio)->mirror_num = mirror; + cb->mirror_num = mirror; + inode = cb->inode; ret = check_compressed_csum(BTRFS_I(inode), cb, (u64)bio->bi_iter.bi_sector << 9); -- cgit v1.2.3 From e6311f240c946788131ba2b97e14f37312688072 Mon Sep 17 00:00:00 2001 From: Liu Bo Date: Wed, 20 Sep 2017 17:50:19 -0600 Subject: Btrfs: skip checksum when reading compressed data if some IO have failed Currently even if the underlying disk reports failure on IO, compressed read endio still gets to verify checksum and reports it as a checksum error. In fact, if some IO have failed during reading a compressed data extent , there's no way the checksum could match, therefore, we can skip that in order to return error quickly to the upper layer. Please note that we need to do this after recording the failed mirror index so that read-repair in the upper layer's endio can work properly. Signed-off-by: Liu Bo Tested-by: Paul Jones Reviewed-by: David Sterba Signed-off-by: David Sterba --- fs/btrfs/compression.c | 9 ++++++++- 1 file changed, 8 insertions(+), 1 deletion(-) (limited to 'fs/btrfs/compression.c') diff --git a/fs/btrfs/compression.c b/fs/btrfs/compression.c index c6aa53c4c102..fc31af98a41b 100644 --- a/fs/btrfs/compression.c +++ b/fs/btrfs/compression.c @@ -108,7 +108,7 @@ static void end_compressed_bio_read(struct bio *bio) struct page *page; unsigned long index; unsigned int mirror = btrfs_io_bio(bio)->mirror_num; - int ret; + int ret = 0; if (bio->bi_status) cb->errors = 1; @@ -127,6 +127,13 @@ static void end_compressed_bio_read(struct bio *bio) btrfs_io_bio(cb->orig_bio)->mirror_num = mirror; cb->mirror_num = mirror; + /* + * Some IO in this cb have failed, just skip checksum as there + * is no way it could be correct. + */ + if (cb->errors == 1) + goto csum_failed; + inode = cb->inode; ret = check_compressed_csum(BTRFS_I(inode), cb, (u64)bio->bi_iter.bi_sector << 9); -- cgit v1.2.3