g_vfs_done():da3s1a[READ(offset=81064794762854400,
length=8192)]error = 5
Scott Long
scottl at samsco.org
Sun Oct 14 17:14:05 PDT 2007
Jan Mikkelsen wrote:
> Scott Long wrote:
>> Ivan Voras wrote:
>>> Either that or file system errors. Does fsck run ok or does
>> it say
>>> anything unusual?
>>>
>> No, filesystem corruption has nothing to do with g_vfs_done
>> messages.
>
> Well, perhaps not directly but I think filesystem corruption can
> indirectly cause g_vfs_done messages.
>
> If a filesystem is corrupt, the filesystem might attempt to read an
> out-of-range block, leading to a g_vfs_done error. This was the
> case for some of the arcmsr problems last year.
>
> In this case, I think the original poster said that the block
> number was out of range for the device.
>
> Regards,
>
> Jan Mikkelsen
>
>
Yeah, you're right, the block number is absurd, and it could well be
caused by a bad block pointer in the filesystem. It sounds like he's
getting this problem even on fresh installs, which ordinarily would
point to a bad driver. If it's happening with both TWA and ATA, it's
hard to blame both of those drivers.
Scott
More information about the freebsd-stable
mailing list