lkml.org 
[lkml]   [2014]   [Apr]   [9]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: [tip:x86/urgent] x86 idle: Repair large-server 50-watt idle-power regression
On Tue, Apr 08, 2014 at 09:43:54PM +0000, Brown, Len wrote:
> #1: skip flush for C1
> #2: skip flush for C1 and C1E

One stray thought I had: IFF the erratum is due to the cache flushing of
the higher C states, these two patches should still be correct while
also (possibly) avoiding the performance penalty, since C1* doesn't
flush caches yet.

Of course, I've no clue if this is indeed the root cause of this
particular erratum and I'm not sure you guys are allowed to say even if
you would know :/


\
 
 \ /
  Last update: 2014-04-09 11:01    [W:0.085 / U:0.932 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site