Re: kernel BUG at fs/xfs/support/debug.c:106!

[Steve Lord]

Mihai RUSU wrote:

> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> Hi
>
> Another problem now, on another system. This one is a 2xP3 1.1 Ghz, 3 GB 
> RAM, MB Intel SCB2, Adaptec 7899 Controller onboard having one 18 GB SCSI 
> disk connected to it (for XFS external journal, swap and / partition which 
> is on ext3), Mylex 170 RAID connected to external storage enclosure with 3 
> x 70 GB SCSI RAID5. The kernel error message is:
>
>  

Mihai,

You missed one thing out of your report, the console message xfs output 
before
this.

I suspect it would have been this: xfs_iget_core: ambiguous vns: vp/0x .....
but it would be good to confirm it. This was supposed to be a dead code
path which there was no longer a route to, it is possible something in the
NFS interface in 2.6 has changed to cause this though. Basically a race
between two threads looking up the same inode, xfs has it cached already
and two threads raced to setup the mapping from the linux inode.

The use of iget_locked when looking up new inodes is supposed to protect
against just this condition.

Steve


-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/