Re: [PATCH] perf, x86: catch spurious interrupts after disabling counters

[Don Zickus <dzickus@redhat.com>]

On Wed, Sep 15, 2010 at 06:20:34PM +0200, Robert Richter wrote:
> On 14.09.10 19:41:32, Robert Richter wrote:
> > I found the reason why we get the unknown nmi. For some reason
> > cpuc->active_mask in x86_pmu_handle_irq() is zero. Thus, no counters
> > are handled when we get an nmi. It seems there is somewhere a race
> > accessing the active_mask. So far I don't have a fix available.
> > Changing x86_pmu_stop() did not help:
> 
> The patch below for tip/perf/urgent fixes this.
> 
> -Robert

I was able to duplicate the problem and can confirm this patch fixes the
issue for me.  I tried poking around (similar to things Robert probably
did) and had no luck.  Something just doesn't make sense, but I guess for
now this patch is good enough for me. :-)

Cheers,
Don

> 
> From 4206a086f5b37efc1b4d94f1d90b55802b299ca0 Mon Sep 17 00:00:00 2001
> From: Robert Richter <robert.richter@amd.com>
> Date: Wed, 15 Sep 2010 16:12:59 +0200
> Subject: [PATCH] perf, x86: catch spurious interrupts after disabling counters
> 
> Some cpus still deliver spurious interrupts after disabling a counter.
> This caused 'undelivered NMI' messages. This patch fixes this.
> 
> Signed-off-by: Robert Richter <robert.richter@amd.com>
> ---
>  arch/x86/kernel/cpu/perf_event.c |   13 ++++++++++++-
>  1 files changed, 12 insertions(+), 1 deletions(-)
> 
> diff --git a/arch/x86/kernel/cpu/perf_event.c b/arch/x86/kernel/cpu/perf_event.c
> index 3efdf28..df7aabd 100644
> --- a/arch/x86/kernel/cpu/perf_event.c
> +++ b/arch/x86/kernel/cpu/perf_event.c
> @@ -102,6 +102,7 @@ struct cpu_hw_events {
>  	 */
>  	struct perf_event	*events[X86_PMC_IDX_MAX]; /* in counter order */
>  	unsigned long		active_mask[BITS_TO_LONGS(X86_PMC_IDX_MAX)];
> +	unsigned long		running[BITS_TO_LONGS(X86_PMC_IDX_MAX)];
>  	int			enabled;
>  
>  	int			n_events;
> @@ -1010,6 +1011,7 @@ static int x86_pmu_start(struct perf_event *event)
>  	x86_perf_event_set_period(event);
>  	cpuc->events[idx] = event;
>  	__set_bit(idx, cpuc->active_mask);
> +	__set_bit(idx, cpuc->running);
>  	x86_pmu.enable(event);
>  	perf_event_update_userpage(event);
>  
> @@ -1141,8 +1143,17 @@ static int x86_pmu_handle_irq(struct pt_regs *regs)
>  	cpuc = &__get_cpu_var(cpu_hw_events);
>  
>  	for (idx = 0; idx < x86_pmu.num_counters; idx++) {
> -		if (!test_bit(idx, cpuc->active_mask))
> +		if (!test_bit(idx, cpuc->active_mask)) {
> +			if (__test_and_clear_bit(idx, cpuc->running))
> +				/*
> +				 * Though we deactivated the counter
> +				 * some cpus might still deliver
> +				 * spurious interrupts. Catching them
> +				 * here.
> +				 */
> +				handled++;
>  			continue;
> +		}
>  
>  		event = cpuc->events[idx];
>  		hwc = &event->hw;
> -- 
> 1.7.2.2
> 
> 
> 
> 
> 
> -- 
> Advanced Micro Devices, Inc.
> Operating System Research Center
>