Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:18 as permitted sender) client-ip=2620:137:e000::1:18;
Date:   Fri, 6 May 2022 14:09:15 +0100
From:   Will Deacon <will@kernel.org>
To:     Stephane Eranian <eranian@google.com>
Cc:     linux-kernel@vger.kernel.org, peterz@infradead.org,
        mark.rutland@arm.com
Subject: Re: [PATCH v2] perf/arm: adjust hwevents mappings on boot
Message-ID: <20220506130915.GD22892@willie-the-truck>
References: <20220324181458.3216262-1-eranian@google.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20220324181458.3216262-1-eranian@google.com>
User-Agent: Mutt/1.10.1 (2018-07-13)
Precedence: bulk

On Thu, Mar 24, 2022 at 11:14:58AM -0700, Stephane Eranian wrote:
> The mapping of perf_events generic hardware events to actual PMU events on
> ARM PMUv3 may not always be correct. This is in particular true for the
> PERF_COUNT_HW_BRANCH_INSTRUCTIONS event. Although the mapping points to an
> architected event, it may not always be available. This can be seen with a
> simple:
> 
> $ perf stat -e branches sleep 0
>  Performance counter stats for 'sleep 0':
> 
>    <not supported>      branches
> 
>        0.001401081 seconds time elapsed
> 
> Yet the hardware does have an event that could be used for branches.
> This patch fixes the problem by dynamically validating the generic hardware
> events against the supported architected events. If a mapping is wrong it
> can be replaced it with another. This is done for the event above at boot time
> and the kernel will log the remapping:
> 
> armv8_pmuv3: hwevent HW_BRANCH_INSTRUCTIONS remapped  from 0xc to 0x21
> 
> And with that:
> 
> $ perf stat -e branches sleep 0
> 
>  Performance counter stats for 'sleep 0':
> 
>            166,739      branches
> 
>        0.000832163 seconds time elapsed
> 
> Signed-off-by: Stephane Eranian <eranian@google.com>
> ---
>  arch/arm64/kernel/perf_event.c | 41 +++++++++++++++++++++++++++++++++-
>  1 file changed, 40 insertions(+), 1 deletion(-)
> 
> diff --git a/arch/arm64/kernel/perf_event.c b/arch/arm64/kernel/perf_event.c
> index cab678ed6618..d438f5a46bdc 100644
> --- a/arch/arm64/kernel/perf_event.c
> +++ b/arch/arm64/kernel/perf_event.c
> @@ -39,7 +39,7 @@
>   * be supported on any given implementation. Unsupported events will
>   * be disabled at run-time based on the PMCEID registers.
>   */
> -static const unsigned armv8_pmuv3_perf_map[PERF_COUNT_HW_MAX] = {
> +static unsigned armv8_pmuv3_perf_map[PERF_COUNT_HW_MAX] = {
>  	PERF_MAP_ALL_UNSUPPORTED,
>  	[PERF_COUNT_HW_CPU_CYCLES]		= ARMV8_PMUV3_PERFCTR_CPU_CYCLES,
>  	[PERF_COUNT_HW_INSTRUCTIONS]		= ARMV8_PMUV3_PERFCTR_INST_RETIRED,
> @@ -1222,6 +1222,42 @@ static void armv8_pmu_register_sysctl_table(void)
>  		register_sysctl("kernel", armv8_pmu_sysctl_table);
>  }
>  
> +static void armv8pmu_fixup_perf_map(struct arm_pmu *cpu_pmu)
> +{
> +	int i, code;
> +	unsigned *map = armv8_pmuv3_perf_map;
> +
> +	for (i = 0; i < PERF_COUNT_HW_MAX; i++) {
> +retry:
> +		code = map[i];
> +		if (code == HW_OP_UNSUPPORTED)
> +			continue;
> +
> +		if (test_bit(map[i], cpu_pmu->pmceid_bitmap))
> +			continue;
> +		/*
> +		 * mapping does not exist,
> +		 * let's see if we can fix it
> +		 */
> +		switch (i) {
> +		case PERF_COUNT_HW_BRANCH_INSTRUCTIONS:
> +			if (code == ARMV8_PMUV3_PERFCTR_PC_WRITE_RETIRED) {
> +				map[i] = ARMV8_PMUV3_PERFCTR_BR_RETIRED;
> +				pr_info("armv8_pmuv3: hwevent "
> +					"HW_BRANCH_INSTRUCTIONS remapped "
> +					" from 0x%x to 0x%x\n", code, map[i]);
> +				goto retry;
> +			}
> +			break;
> +		default:
> +			pr_info("armv8_pmuv3: hwevent %d not supported\n", i);

If a CPU supports neither ARMV8_PMUV3_PERFCTR_PC_WRITE_RETIRED nor
ARMV8_PMUV3_PERFCTR_BR_RETIRED, won't we get a funny series of messages
here? I think I'd prefer to drop the prints altogether.

Will