Bug#579005: XFS internal error xfs_da_do_buf(2) at line 2085 of file fs/xfs/xfs_da_btree.c.

[Jonathan Nieder <jrnieder@gmail.com>]

found 579005 linux-2.6/2.6.32-38
quit

Rudy Gevaert wrote:

> I'm seeing the same issue.
[...]
>
> ffff880aa6ac4000: 49 4e 00 00 02 02 00 00 00 00 03 d4 00 00 00 08  IN..............
> Filesystem "dm-29": XFS internal error xfs_da_do_buf(2) at line 2112 of file /build/buildd-linux-2.6_2.6.32-38-amd64-bk66e4/linux-2.6-2.6.32/debian/build/source_amd64_none/fs/xfs/xfs_da_btree.c.  Caller 0xffffffffa036c9b4
> 
> Pid: 32675, comm: bpbkar Not tainted 2.6.32-5-amd64 #1
> Call Trace:
>  [<ffffffffa036c9b4>] ? xfs_da_read_buf+0x24/0x29 [xfs]
>  [<ffffffffa036c884>] ? xfs_da_do_buf+0x558/0x61e [xfs]
>  [<ffffffffa036c9b4>] ? xfs_da_read_buf+0x24/0x29 [xfs]
>  [<ffffffffa0396e82>] ? xfs_dir_open+0x0/0x55 [xfs]
>  [<ffffffffa036c97b>] ? xfs_da_reada_buf+0x31/0x46 [xfs]
>  [<ffffffff810ed641>] ? __dentry_open+0x1c4/0x2bf
>  [<ffffffff8111fd6b>] ? compat_filldir64+0x0/0xb8
>  [<ffffffffa036c9b4>] ? xfs_da_read_buf+0x24/0x29 [xfs]
>  [<ffffffffa037025d>] ? xfs_dir2_block_getdents+0x66/0x1ab [xfs]
>  [<ffffffffa037025d>] ? xfs_dir2_block_getdents+0x66/0x1ab [xfs]
>  [<ffffffff8111fd6b>] ? compat_filldir64+0x0/0xb8
>  [<ffffffff8111fd6b>] ? compat_filldir64+0x0/0xb8
>  [<ffffffffa036f082>] ? xfs_readdir+0x8b/0xb0 [xfs]
>  [<ffffffff8111fd6b>] ? compat_filldir64+0x0/0xb8
>  [<ffffffffa0396f0b>] ? xfs_file_readdir+0x34/0x43 [xfs]
>  [<ffffffff810fb6b8>] ? vfs_readdir+0x75/0xa7
>  [<ffffffff81121a6c>] ? compat_sys_getdents64+0x7a/0xbe
>  [<ffffffff810371e0>] ? sysenter_dispatch+0x7/0x2e

Nice.  Please provide a few lines before and after, too.

I would also be interested in the basic information from "reportbug
--template linux-image-$(uname -r)" (as an attachment), even though I
do not expect to get any earth-shattering revelations from it.

Could you also say a little about the workload at the time this was
experienced?  Is this easy to reproduce?  Does it happen often, and
if not, any common patterns between moments when it happens?

If this is reproducible, would you be able to test patches to debug
this if provided?

If you get a chance to test a 3.x.y kernel from sid or experimental,
that would also be very helpful.  (The only packages from outside
squeeze that should be needed in order to do so aside from the kernel
image itself are linux-base and initramfs-tools.)  If it is
reproducible there, we can get help from upstream, and if not, we can
try some kernels halfway between to narrow down when the fix was
introduced and try applying the same fix to squeeze.

Thanks much,
Jonathan