Messages in this thread |  | | | From | Nick Piggin <> | | Subject | Re: [Bug #10815] 2.6.26-rc4: RIP find_pid_ns+0x6b/0xa0 | | Date | Tue, 24 Jun 2008 11:51:10 +1000 |
| |
On Tuesday 24 June 2008 11:31, Linus Torvalds wrote: > On Tue, 24 Jun 2008, Alexey Dobriyan wrote: > > Status update: > > * bug is reproduced on another box with the very same symptoms: > > SMP=y, maxcpus=1 kernel occasionally fails, SMP=n is fine. > > Also Core 2 Duo, x86_64 [1] > > > > Race is wide -- 60 seconds of rcutorture is enough. > > Very odd how you can reproduce it - now on two machines - but it doesn't > seem to happen for others. You've tried different compilers, you've tried > different config options, what the heck is left?
The PREEMPT_RCU thing? I've reproduced it. It hits lockless pagecache, but I've also reproduced the same problem in dentry cache shrinking. I'm sure it is a free-before-grace bug.
I've been using DEBUG_PAGEALLOC to make it easier to spot.
|  |