On Fri, Jun 10, 2016 at 7:37 PM, Hannes Reinecke [off-list ref] wrote:

On 06/10/2016 01:07 PM, Ming Lei wrote:

quoted

After arbitrary bio size is supported, the incoming bio may
be very big. We have to split the bio into small bios so that
each holds at most BIO_MAX_PAGES bvecs for safety reason, such
as bio_clone().

This patch fixes the following kernel crash:

quoted

[  172.660142] BUG: unable to handle kernel NULL pointer dereference at=

 0000000000000028

quoted

quoted

[  172.660229] IP: [<ffffffff811e53b4>] bio_trim+0xf/0x2a
[  172.660289] PGD 7faf3e067 PUD 7f9279067 PMD 0
[  172.660399] Oops: 0000 [#1] SMP
[...]
[  172.664780] Call Trace:
[  172.664813]  [<ffffffffa007f3be>] ? raid1_make_request+0x2e8/0xad7 [=

raid1]

quoted

quoted

[  172.664846]  [<ffffffff811f07da>] ? blk_queue_split+0x377/0x3d4
[  172.664880]  [<ffffffffa005fb5f>] ? md_make_request+0xf6/0x1e9 [md_m=

od]

quoted

quoted

[  172.664912]  [<ffffffff811eb860>] ? generic_make_request+0xb5/0x155
[  172.664947]  [<ffffffffa0445c89>] ? prio_io+0x85/0x95 [bcache]
[  172.664981]  [<ffffffffa0448252>] ? register_cache_set+0x355/0x8d0 [=

bcache]

quoted

quoted

[  172.665016]  [<ffffffffa04497d3>] ? register_bcache+0x1006/0x1174 [b=

cache]

quoted

The issue can be reproduced by the following steps:
      - create one raid1 over two virtio-blk
      - build bcache device over the above raid1 and another cache devic=

e

quoted

      and bucket size is set as 2Mbytes
      - set cache mode as writeback
      - run random write over ext4 on the bcache device

Fixes: 54efd50(block: make generic_make_request handle arbitrarily sized=

 bios)

quoted

Reported-by: Sebastian Roesner <redacted>
Reported-by: Eric Wheeler <redacted>
Cc: stable@vger.kernel.org (4.3+)
Cc: Shaohua Li <redacted>
Acked-by: Kent Overstreet <redacted>
Signed-off-by: Ming Lei <redacted>
---
V2:
      - don't mark as REQ_NOMERGE in case the bio is splitted
      for reaching the limit of bvecs count
V1:
        - Kent pointed out that using max io size can't cover
        the case of non-full bvecs/pages
 block/blk-merge.c | 35 ++++++++++++++++++++++++++++++++---
 1 file changed, 32 insertions(+), 3 deletions(-)

Hmm. So everybody is suffering because someone _might_ be using bio_clone=

?

I believe most of usages are involved with <=3D 256 bvecs in one bio, so
only few(such as bcache) will 'suffer', not everybody, :-)

Why can't we fixup bio_clone() (or the callers of which) to correctly
set the queue limits?

IMO there isn't a good solution to fix the issue in bio_clone.

Firstly one page can held at most 256 bvecs, and not safe to allocate
multi-pages in I/O path.

Secondaly as said in the comment of the patch it can't be a queue limit
now because bio_clone() is used inside bio bounce.

But it should be possible to use bio splitting to deal with bio bounce, and
it can be a following up job, and of course that change can be a bit too bi=
g
for backporting.

That is why I suggest to fix the issue with this patch. Or other ideas?


Thanks,
Ming

Cheers,

Hannes
--
Dr. Hannes Reinecke                Teamlead Storage & Networking
hare@suse.de                                   +49 911 74053 688
SUSE LINUX GmbH, Maxfeldstr. 5, 90409 N=C3=BCrnberg
GF: F. Imend=C3=B6rffer, J. Smithard, J. Guild, D. Upmanyu, G. Norton
HRB 21284 (AG N=C3=BCrnberg)
--
To unsubscribe from this list: send the line "unsubscribe linux-block" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html