[LUCID][PATCH] fix performance regression for SMT and MC

[Tim Gardner]

On 03/04/2010 11:48 PM, Yingying Zhao wrote:
> There're two patches need to get backported to Lucid which address scheduler
> regression for SMT and MC.
>
> Both of them are in mainline now and also copied to stable tree:
>
> SMT fix:
> http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=9000f05c6d1607f79c0deacf42b09693be673f4c
> commit 9000f05c6d1607f79c0deacf42b09693be673f4c
> Author: Suresh Siddha<suresh.b.siddha at intel.com>
> Date:   Fri Feb 12 17:14:22 2010 -0800
>
>      sched: Fix SMT scheduler regression in find_busiest_queue()
>
> MC fix:
> http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=dd5feea14a7de4edbd9f36db1a2db785de91b88d
> commit dd5feea14a7de4edbd9f36db1a2db785de91b88d
> Author: Suresh Siddha<suresh.b.siddha at intel.com>
> Date:   Tue Feb 23 16:13:52 2010 -0800
>
>      sched: Fix SCHED_MC regression caused by change in sched cpu_power
>
> These patches are critical for SMT and MC performance, please consider to pick
> up for Lucid.
>
> Thanks
> Yingying
>
>

Yingying - the first patch 'sched: Fix SMT scheduler regression in 
find_busiest_queue()' applies cleanly and ought to show up in stable 
pretty soon, but the second patch appears to have some prerequisite 
commits. Can you have Suresh research the minimum set required in order 
to apply 'sched: Fix SCHED_MC regression caused by change in sched 
cpu_power' ?

rtg
-- 
Tim Gardner tim.gardner at canonical.com