[OpenAFS-devel] Kernel panic with openafs 1.4.0

[Ulrich Seidl]

Jeffrey Hutzelman wrote:
>
>
> On Thursday, August 24, 2006 01:14:11 PM +0200 Harald Barth 
> <haba@pdc.kth.se> wrote:
>
>>
>>> Today I've encountered the following kernel panic with openafs 1.4.0 on
>>> a Scientific Linux 4 system (RedHat Enterprise 4 compatible). Will this
>>> be fixed in 1.4.2?
>>
>> That's a bit difficult to tell before we know why it broke.
>> I have not seen this particular bug before, someone else
>> might.
>>
>> Let me see... We are in afs_NewVCache() and call osi_Panic().
>> Unfortunately osi_Panic() did not write the message string,
>> but only called BUG(). (Can we fix that for the future?)
>
> There's nothing to fix.  osi_Panic _does_ print out the message 
> string, but necessarily does so _before_ it calls BUG().  Which means 
> that users habitually don't actually copy the message when sending in 
> a report. Fixing this would require changes to Linux to allow 
> something calling BUG to include a message that gets included _below_ 
> the "cut here" line.
>
>
>> Normally afs_NewVCache() works, so why did it not work for
>> you here?
>>
>> So the standard questions are:
>
> ... all the ones you mentioned, plus "what were the last few messages 
> right _before_ the 'cut here' line.
The line before the 'cut here' was:

Aug 24 11:08:31 gromit kernel: NewVCache: warning none freed, using 300 
of 300

>
>
> Also, the standard cause of NewVCache failing in 1.4.0 on a Linux 2.6 
> system is that the default parameters include only 300 vcache entries, 
> which is nowhere near enough.
Even though the vcache is exhausted, is this a reason to oops the kernel?