Messages in this thread | | | Date | Mon, 08 Jun 2009 10:36:23 +0800 | From | Lai Jiangshan <> | Subject | Re: [PATCH] cpuhotplug: introduce try_get_online_cpus() take 2 |
| |
Paul E. McKenney wrote: >> Add CC to Peter Zijlstra <peterz@infradead.org> >> >> (This patch is against mainline but not for inclusion, it will adapted >> against -mm when request) >> >> Requst For Comments and Reviewing Hungeringly. >> >> - Lockless for get_online_cpus()'s fast path >> - Introduce try_get_online_cpus() > > One question for Gautham Shenoy -- are non-atomic CPU-hotplug notifiers > always invoked from the task that did the cpu_hotplug_begin()? > > If so, well and good. If not, then it would not be possible to > expedite RCU grace periods from within CPU-hotplug notifiers. > > Thanx, Paul >
I have no proper machine for several days to test the next version of this patch, so it is still being delayed a while.
Lai.
| |