lkml.org 
[lkml]   [2011]   [Jan]   [20]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: [PATCH 13/14] x86/ticketlock: add slowpath logic
>>> On 19.01.11 at 19:55, Jeremy Fitzhardinge <jeremy@goop.org> wrote:
> On 01/19/2011 10:39 AM, Srivatsa Vaddagiri wrote:
>> I have tested quite extensively with booting a 16-vcpu guest (on a 16-pcpu
> host)
>> and running kernel compine (with 32-threads). Without this patch, I had
>> difficulty booting/shutting-down successfully (it would hang mid-way).
>
> Sounds good. But I like to test with "make -j 100-200" to really give
> things a workout.

Hmm, in my experience, heavily over-committing CPUs (e.g. a
domain with double or more the vCPU-s that the system has
pCPU-s, or the pCPU-s the domain is allowed to run on) is a
much better test for eventual problems in the spin lock code
paths.

Jan



\
 
 \ /
  Last update: 2011-01-20 10:55    [W:0.065 / U:5.656 seconds]
©2003-2018 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site