Re: [block:for-3.14/core] kernel BUG at fs/bio.c:1748

From: Chris Mason
Date: Tue Jan 07 2014 - 15:30:10 EST

On Tue, 2014-01-07 at 12:15 -0800, Muthu Kumar wrote:
+AD4- Thanks Fengguang. Final patch with added comment. BTW, fengguang
+AD4- mentioned that git-am has trouble with the inline patch and +ACI-quilt
+AD4- import+ACI- worked fine for him...
+AD4- ------------
+AD4- In btrfs+AF8-end+AF8-bio(), we increment bi+AF8-remaining if is+AF8-orig+AF8-bio. If not,
+AD4- we restore the orig+AF8-bio but failed to increment bi+AF8-remaining for
+AD4- orig+AF8-bio, which triggers a BUG+AF8-ON later when bio+AF8-endio is called. Fix
+AD4- is to increment bi+AF8-remaining when we restore the orig bio as well.

Hi everyone,

Which git tree is this against? Just Jens or some extra code too?

I'll run some tests here. My original patch is below (it's slightly
different from Muthu's).

Btrfs is sometimes calling bio+AF8-endio twice on the same bio while
we chain things. This makes sure we don't trip over new assertions in

Signed-off-by: Chris Mason

diff --git a/fs/btrfs/check-integrity.c b/fs/btrfs/check-integrity.c
index 7fcac70..5b30360 100644
--- a/fs/btrfs/check-integrity.c
+-+-+- b/fs/btrfs/check-integrity.c
+AEAAQA- -2289,6 +-2289,10 +AEAAQA- static void btrfsic+AF8-bio+AF8-end+AF8-io(struct bio +ACo-bp, int bio+AF8-error+AF8-status)
block +AD0- next+AF8-block+ADs-
+AH0- while (NULL +ACEAPQ- block)+ADs-

+- /+ACo-
+- +ACo- since we're not using bio+AF8-endio here, we don't need to worry about
+- +ACo- the remaining count
+- +ACo-/
bp-+AD4-bi+AF8-end+AF8-io(bp, bio+AF8-error+AF8-status)+ADs-

diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c
index 62176ad..786ddac 100644
--- a/fs/btrfs/disk-io.c
+-+-+- b/fs/btrfs/disk-io.c
+AEAAQA- -1684,7 +-1684,7 +AEAAQA- static void end+AF8-workqueue+AF8-fn(struct btrfs+AF8-work +ACo-work)
bio-+AD4-bi+AF8-private +AD0- end+AF8-io+AF8-wq-+AD4-private+ADs-
bio-+AD4-bi+AF8-end+AF8-io +AD0- end+AF8-io+AF8-wq-+AD4-end+AF8-io+ADs-
- bio+AF8-endio(bio, error)+ADs-
+- bio+AF8-endio+AF8-nodec(bio, error)+ADs-

static int cleaner+AF8-kthread(void +ACo-arg)
diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c
index ef48947..a31448f 100644
--- a/fs/btrfs/volumes.c
+-+-+- b/fs/btrfs/volumes.c
+AEAAQA- -5284,9 +-5284,17 +AEAAQA- static void btrfs+AF8-end+AF8-bio(struct bio +ACo-bio, int err)

- if (bio +AD0APQ- bbio-+AD4-orig+AF8-bio)
+- if (bio +AD0APQ- bbio-+AD4-orig+AF8-bio) +AHs-
is+AF8-orig+AF8-bio +AD0- 1+ADs-

+- /+ACo-
+- +ACo- eventually we will call the bi+AF8-endio for the original bio,
+- +ACo- make sure that we've properly bumped bi+AF8-remaining to reflect
+- +ACo- our chain of endios here
+- +ACo-/
+- atomic+AF8-inc(+ACY-bio-+AD4-bi+AF8-remaining)+ADs-
+- +AH0-
if (atomic+AF8-dec+AF8-and+AF8-test(+ACY-bbio-+AD4-stripes+AF8-pending)) +AHs-
if (+ACE-is+AF8-orig+AF8-bio) +AHs-
To unsubscribe from this list: send the line +ACI-unsubscribe linux-fsdevel+ACI- in
the body of a message to
More majordomo info at

To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at
Please read the FAQ at