[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: [Xen-devel] Re: [PATCH] CPUIDLE: shorten hpet spin_lock holding time


  • To: Keir Fraser <keir.fraser@xxxxxxxxxxxxx>, "xen-devel@xxxxxxxxxxxxxxxxxxx" <xen-devel@xxxxxxxxxxxxxxxxxxx>
  • From: "Wei, Gang" <gang.wei@xxxxxxxxx>
  • Date: Thu, 29 Apr 2010 19:08:09 +0800
  • Accept-language: zh-CN, en-US
  • Acceptlanguage: zh-CN, en-US
  • Cc:
  • Delivery-date: Thu, 29 Apr 2010 04:11:11 -0700
  • List-id: Xen developer discussion <xen-devel.lists.xensource.com>
  • Thread-index: AcrgS919AC3RHXCET4295IJeXmmgRwAO/3swAAIUDKAAASh2WgADOhngACIUzXUAAMoQoAAB26RvAAAUX9AAANuwEgAAXeGYACOffbAACQ4yYgACATLsAAAdWwcBZQWBwA==
  • Thread-topic: [Xen-devel] Re: [PATCH] CPUIDLE: shorten hpet spin_lock holding time

On Thursday, 2010-4-22 4:23 PM, Keir Fraser wrote:
> On 22/04/2010 09:19, "Keir Fraser" <keir.fraser@xxxxxxxxxxxxx> wrote:
>> How about the attached patch? It MWAITs on a completely new flag
>> field, avoiding the IPI-avoidance semantics of softirq_pending. It
>> also folds in your patch. It also does wakeup-waiting checks on
>> timer_deadline_start, that being the field that initiates wakeup via
>> the MONITORed memory region. 
> 
> ...If you do think it looks okay, could you also test it out on
> relevant hardware? :-)

Did some modification to this patch -- move the per_cpu mait_wakeup flag into 
irq_cpustat_t to make it __cacheline_aligned, and add check for 
timer_deadline_start==0 (means no timer in queue, it took me quite a lot time 
to find out it is necessary) case. Tested ok.

Jimmy

Attachment: 00-mwait-v2.patch
Description: 00-mwait-v2.patch

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-devel

 


Rackspace

Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.