Re: x86-microcode: get rid of set_cpus_allowed()

[Andreas Herrmann <andreas.herrmann3@amd.com>]

On Thu, Mar 12, 2009 at 06:40:10PM +0100, Andreas Herrmann wrote:
> On Wed, Mar 11, 2009 at 07:44:37AM +0100, Rusty Russell wrote:
> > On Tuesday 10 March 2009 06:08:59 Dmitry Adamushko wrote:
> > > 
> > > Hi,
> > > 
> > > 
> > > here is a possible candidate for Rusty's cpumask-refactored series.
> > > Note the [*] remark below though.
> > 
> > Ah, OK, I'll drop my version then (below) in favor of this, and will
> > push to Ingo with the others if he doesn't take it directly.
> 
> Sorry guys -- for the late reply --
> but I missed Dmitry's mail due to some silly mail filtering and had to
> restore his mail ...
> 
> Now I've tested both patches and both seem to reliably prevent
> microcode updates on CPU1 and CPU2 of an Phenom X3 after
> suspend/resume. (Just CPU0 was updated.)
> 
> Then I've tested mainline kernel w/o your patches and I've observed
> similar problems. I've seen that sometimes ucode of CPU0 was not
> updated and sometimes CPU1 and CPU2 were not updated.
> 
> I'll look into this asap.

Some further testing seem to indicate that suspend/resume does not
work when I have done CPU hotplug before.

During today's tests I did:

(1) set offline/online CPU 1 and 2
(2) perform suspend/resume afterwards

After that microcode update failed on some CPUs when performing
suspend/resume. (When skipping step 1, microcode update during
suspend/resume works.)

Looks strange, but should be debuggable.


Regards,

Andreas

-- 
Operating | Advanced Micro Devices GmbH
  System  | Karl-Hammerschmidt-Str. 34, 85609 Dornach b. München, Germany
 Research | Geschäftsführer: Jochen Polster, Thomas M. McCoy, Giuliano Meroni
  Center  | Sitz: Dornach, Gemeinde Aschheim, Landkreis München
  (OSRC)  | Registergericht München, HRB Nr. 43632