Date: Wed, 14 Dec 2016 19:07:30 +0100
From: Peter Zijlstra <peterz@infradead.org>
To: Jiri Olsa <jolsa@redhat.com>
Cc: Andi Kleen <andi@firstfloor.org>, lkml <linux-kernel@vger.kernel.org>,
        Alexander Shishkin <alexander.shishkin@linux.intel.com>,
        Vince Weaver <vince@deater.net>, Ingo Molnar <mingo@kernel.org>
Subject: Re: [RFC] perf/x86/intel: Account interrupts for PEBS errors
Message-ID: <20161214180730.GR3124@twins.programming.kicks-ass.net>
References: <20161214165036.GB9180@krava>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20161214165036.GB9180@krava>
User-Agent: Mutt/1.5.23.1 (2014-03-12)
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 1367
Lines: 51

On Wed, Dec 14, 2016 at 05:50:36PM +0100, Jiri Olsa wrote:
> 
> I also fail to reproduce on other than snb_x (model 45) server

reproduces on my ivb-ep as well model 62.

> thoughts?

cute find :-)

> +++ b/arch/x86/events/intel/ds.c
> @@ -1389,9 +1389,13 @@ static void intel_pmu_drain_pebs_nhm(struct pt_regs *iregs)
>  			continue;
>  
>  		/* log dropped samples number */
> -		if (error[bit])
> +		if (error[bit]) {
>  			perf_log_lost_samples(event, error[bit]);
>  
> +			if (perf_event_account_interrupt(event, 1))

Seems a bit daft to expose the .throttle argument, since that would be
the only point of calling this.


> +static int __perf_event_overflow(struct perf_event *event,
> +				   int throttle, struct perf_sample_data *data,
> +				   struct pt_regs *regs)
> +{
> +	int events = atomic_read(&event->event_limit);
> +	struct hw_perf_event *hwc = &event->hw;
> +	int ret = 0;
> +
> +	/*
> +	 * Non-sampling counters might still use the PMI to fold short
> +	 * hardware counters, ignore those.
> +	 */
> +	if (unlikely(!is_sampling_event(event)))
> +		return 0;
> +
> +	ret = perf_event_account_interrupt(event, throttle);
> +
>  	if (event->attr.freq) {
>  		u64 now = perf_clock();
>  		s64 delta = now - hwc->freq_time_stamp;

Arguably, everything in __perf_event_overflow() except for calling of
->overflow_handler() should be done I think.