kern/189355: zfs panic on 10-stable

[Steven Hartland]

----- Original Message ----- 
From: "Radim Kolar" <hsn at sendmail.cz>
snip
> #11 0xc0cb801e in vdev_queue_io_to_issue (vq=3D0xc46a4b00)
>     at /usr/src/sys/modules/zfs/../../cddl/contrib/opensolaris/uts/common/f=
> s/zfs/vdev_queue.c:471
> #12 0xc0cb7fb8 in vdev_queue_io (zio=3D0xc4855000)
>     at /usr/src/sys/modules/zfs/../../cddl/contrib/opensolaris/uts/common/f=
> s/zfs/vdev_queue.c:744
> #13 0xc0cd84ee in zio_vdev_io_start (ziop=3D<value optimized out>)
>     at /usr/src/sys/modules/zfs/../../cddl/contrib/opensolaris/uts/common/f=
> s/zfs/zio.c:2607
> #14 0xc0cd4c18 in zio_execute (zio=3D0xc4855000)
>     at /usr/src/sys/modules/zfs/../../cddl/contrib/opensolaris/uts/common/f=
> s/zfs/zio.c:1350
> #15 0xc0cb74e4 in vdev_mirror_io_start (zio=3D0xc4823894)
>     at /usr/src/sys/modules/zfs/../../cddl/contrib/opensolaris/uts/common/f=
> s/zfs/vdev_mirror.c:284

Its not totally clear as the faulting line shouldn't ever do so but I suspect
you may be seeing corruption earlier on due to some edge cases in the new
queing code.

This could be something that has already has a fixed in head but wasn't expected
to get triggered without TRIM queueing which isn't in 10-stable yet. The fix
is also likely to change based on feedback from the openzfs guys, hence isn't
in 10-stable yet.

Can you try appling the following patches from head to your 10-stable and
see if that helps:-
http://svnweb.freebsd.org/base?view=revision&revision=265046
http://svnweb.freebsd.org/base?view=revision&revision=265321

    Regards
    Steve