Re: intel_pstate divide error with v3.13-rc4-256-gb7000ad

From: Rafael J. Wysocki
Date: Fri Dec 27 2013 - 09:22:03 EST


On Tuesday, December 24, 2013 09:36:01 PM Viresh Kumar wrote:
> Adding Dirk..
>
> On 24 December 2013 20:06, Josh Boyer <jwboyer@xxxxxxxxxxxxxxxxx> wrote:
> > Hi All,
> >
> > We've had a report [1] that the pstate driver causes KVM guests to
> > fail to boot because of a divide error. See the backtrace below.
> >
> > 4.839784] Intel P-state driver initializing.
> > [ 4.859972] Intel pstate controlling: cpu 0
> > [ 4.867653] cpufreq: __cpufreq_add_dev: ->get() failed
>
> After a call to ->init(), ->get() is supposed to work.
> @Dirk: Any idea why it failed?

Well, it looks like sample->freq is 0 in intel_pstate_get().

> And then I don't know what made this divide by zero to happen :)

>From code inspection it looks like that is caused by the
intel_pstate_get_scaled_busy() called from intel_pstate_adjust_busy_pstate(),
so it appears that cpu->pstate.current_state is 0 at that point.

I'm wondering if something like the (untested) patch below helps, then?

Rafael


---
drivers/cpufreq/intel_pstate.c | 5 +++++
1 file changed, 5 insertions(+)

Index: linux-pm/drivers/cpufreq/intel_pstate.c
===================================================================
--- linux-pm.orig/drivers/cpufreq/intel_pstate.c
+++ linux-pm/drivers/cpufreq/intel_pstate.c
@@ -652,6 +652,11 @@ static int intel_pstate_init_cpu(unsigne
cpu = all_cpu_data[cpunum];

intel_pstate_get_cpu_pstates(cpu);
+ if (!cpu->pstate.current_state) {
+ all_cpu_data[cpunum] = NULL;
+ kfree(cpu);
+ return -ENODATA;
+ }

cpu->cpu = cpunum;


--
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/