Re: kernel preemption and segfault

2004-10-08 Thread Mathieu Virbel
Benjamin Herrenschmidt wrote: On Fri, 2004-10-08 at 17:36, Vincent Bernat wrote: OoO Vers la fin de l'après-midi du jeudi 07 octobre 2004, vers 16:53, Mathieu Virbel <[EMAIL PROTECTED]> disait: With kernel preemption, all work perfectly. Without kernel preemption, they are many bugs: - und

Re: kernel preemption and segfault

2004-10-08 Thread Stefano Zacchiroli
On Fri, Oct 08, 2004 at 06:05:10PM +1000, Benjamin Herrenschmidt wrote: > > The situation is going better but it is still not usable. > Can you try isolating a given activity or driver that cause the > stability problem ? It should work, really... I'm not able to isolate a problematic scenario. Bu

Re: kernel preemption and segfault

2004-10-08 Thread Benjamin Herrenschmidt
On Fri, 2004-10-08 at 17:36, Vincent Bernat wrote: > OoO Vers la fin de l'après-midi du jeudi 07 octobre 2004, vers 16:53, > Mathieu Virbel <[EMAIL PROTECTED]> disait: > > > With kernel preemption, all work perfectly. > > Without kernel preemption, they are many bugs: > > - under medium/high load

Re: kernel preemption and segfault

2004-10-08 Thread Vincent Bernat
OoO Vers la fin de l'après-midi du jeudi 07 octobre 2004, vers 16:53, Mathieu Virbel <[EMAIL PROTECTED]> disait: > With kernel preemption, all work perfectly. > Without kernel preemption, they are many bugs: > - under medium/high load, gcc crash randomly > - mozilla have trouble to render (each r