Re: [PATCH] x86, smp: Fix topology checks on AMD MCM

From: Borislav Petkov
Date: Mon Jun 04 2012 - 09:36:56 EST


On Mon, Jun 04, 2012 at 02:43:50PM +0200, Peter Zijlstra wrote:
> On Mon, 2012-06-04 at 14:41 +0200, Borislav Petkov wrote:
> > The warning below triggers on AMD MCM packages because physical
> > package
> > IDs on the cores of a _physical_ socket are the same. I.e., this field
> > says which CPUs belong to the same physical package.
> >
> Thanks!

This is still crap.

It breaks tests like those in powernow-k8:

if (cpumask_first(cpu_core_mask(data->cpu)) == data->cpu)
print_basics(data);

and it calls print_basics() on each core filling up dmesg with:

[ 5.216925] powernow-k8: 0 : pstate 0 (1700 MHz)
[ 5.222139] powernow-k8: 1 : pstate 1 (1300 MHz)
[ 5.227354] powernow-k8: 2 : pstate 2 (1100 MHz)
[ 5.232568] powernow-k8: 3 : pstate 3 (1000 MHz)
[ 5.237783] powernow-k8: 4 : pstate 4 (800 MHz)
[ 5.242988] powernow-k8: 0 : pstate 0 (1700 MHz)
[ 5.248242] powernow-k8: 1 : pstate 1 (1300 MHz)
[ 5.253462] powernow-k8: 2 : pstate 2 (1100 MHz)
[ 5.258677] powernow-k8: 3 : pstate 3 (1000 MHz)
[ 5.263892] powernow-k8: 4 : pstate 4 (800 MHz)
[ 5.269094] powernow-k8: 0 : pstate 0 (1700 MHz)
[ 5.274348] powernow-k8: 1 : pstate 1 (1300 MHz)
[ 5.279566] powernow-k8: 2 : pstate 2 (1100 MHz)
[ 5.284787] powernow-k8: 3 : pstate 3 (1000 MHz)
[ 5.298387] powernow-k8: 4 : pstate 4 (800 MHz)
...

while it should dump the Pstates only once per physical socket.

The proper fix should be:

static bool __cpuinit match_mc(struct cpuinfo_x86 *c, struct cpuinfo_x86 *o)
{
if (c->phys_proc_id == o->phys_proc_id) {
if (cpu_has(c, X86_FEATURE_AMD_DCM))
return true;

return topology_sane(c, o, "mc");
}
return false;
}

which basically circumvents the topology check on MCM modules. I'll run
this a bit longer to verify it doesn't break anything else.

Thanks.

--
Regards/Gruss,
Boris.

Advanced Micro Devices GmbH
Einsteinring 24, 85609 Dornach
GM: Alberto Bozzo
Reg: Dornach, Landkreis Muenchen
HRB Nr. 43632 WEEE Registernr: 129 19551
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/