page fault panic tracked down (selwakeuppri())

[Scott Long]

Stefan Ehmann wrote:
> On Tue, 2003-12-30 at 17:20, Bruce Evans wrote:
> 
>>On Tue, 30 Dec 2003, Stefan Ehmann wrote:
>>
>>
>>>On Tue, 2003-12-30 at 07:56, Bruce Evans wrote:
>>>
>>>>I have farily large patches which do buffering in ext2fs in a different
>>>>way so that the 2003/08/28 changes are irrelevant.  I will send these
>>>>in private mail.
>>>
>>>Applied the patches and rebuilt kernel. Unfortunately I got the same
>>>panic after 2 hours.
>>
>>At least it's easy to duplicate, and another possible cause is ruled out.
>>
>>Do you do anything special with ext2fs or have a special configuration?
> 
> On this partition my music collection is stored. Normally I run cplay
> (CLI music player) in shuffle mode. So there is a low but constant load
> on the drive. I'm not sure about this, but at least sometimes the panics
> seems to happen immediately after a new file is opened.
> 
> The only special thing I can think about is that I created it with the
> largefile4 (one inode per 4 megabytes) option in knoppix. As I mentioned
> above it is mounted read-only.
> 
> 
>>It should be easy to turn off the selwakeuppri() chnages by editing a
>>line or two of selwakeuppri().  From sys_generic.c:
> 
> 
> Thanks. So I can (hopefully) run a recent CURRENT somewhat stable at
> least until this is resolved.
> 

Does Bruce's hack actually work for you?  I'm trying to decide on
whether to make this a show-stopper for the 5.2 release.

Scott