kernel BUG at page_alloc.c:185! with 2.5.31 + akpm stuff

kernel BUG at page_alloc.c:185! with 2.5.31 + akpm stuff

[Steven Cole]

With this patch applied to 2.5.31,

http://www.zip.com.au/~akpm/linux/patches/2.5/2.5.31/stuff-sent-to-linus/everything.gz

I got these BUGs,

kernel BUG at page_alloc.c:185!
 kernel BUG at page_alloc.c:98!

I was running dbench as a stress tester.  I had run dbench with
up to 128 clients with no problems, and was running my stress test a
second time when the BUGs occurred with dbench 6.

I have run the dbench 1..128 stress test with plain vanilla 2.5.31 four
times without these failures.

This is on a 2-way p-III box, CONFIG_SMP=y.

Steven


--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/

Re: kernel BUG at page_alloc.c:185! with 2.5.31 + akpm stuff

[Andrew Morton]

Steven Cole wrote:
> 
> With this patch applied to 2.5.31,
> 
> http://www.zip.com.au/~akpm/linux/patches/2.5/2.5.31/stuff-sent-to-linus/everything.gz
> 
> I got these BUGs,
> 
> kernel BUG at page_alloc.c:185!

This one's a worry.  A PageLRU(page) came off the buddy list.
This would tend to indicate that I goofed.

>  kernel BUG at page_alloc.c:98!

That's the BUG_ON(page->pte.chain).  Numerous people have been
getting that.  I had it a single time, about three kernels ago.
It's proving elusive.
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/