Messages in this thread | | | Subject | Re: [PATCH] revert: timers: fix itimer/many thread hang | From | Doug Chapman <> | Date | Fri, 14 Nov 2008 09:10:51 -0500 |
| |
On Thu, 2008-11-13 at 17:08 +0100, Ingo Molnar wrote: > * Doug Chapman <doug.chapman@hp.com> wrote: > > > Peter didn't mention here that the itimer/many thread patches cause > > a race-condition panic which is easily reproduced (in fact > > unavoidable with SLUB debugging on) on hyperthreaded ia64 systems > > (but certainly possible on all systems). Can we get this revert > > patch pulled in? The other attempts to fix this so far have been > > unsuccessful. I have tested with this revert patch and it does > > indeed fix that issue. > > Could you please try with the very latest upstream kernel? It should > be fixed by Oleg's fix patch via: > > ad474ca: fix for account_group_exec_runtime(), make sure ->signal can't be fre > > which is already upstream. > > Ingo
Ingo,
Thanks, somehow I had missed that. I tested with the latest GIT tree and the issue is resolved now.
- Doug
| |