Re: [PATCH V2 2/3] perf/imx_ddr: Add system PMU identifier for userspace

From: Will Deacon <will@kernel.org>
To: Joakim Zhang <qiangqing.zhang@nxp.com>
Cc: mark.rutland@arm.com, devicetree@vger.kernel.org,
	john.garry@huawei.com, linux-kernel@vger.kernel.org,
	robh+dt@kernel.org, linux-imx@nxp.com, shawnguo@kernel.org,
	linux-arm-kernel@lists.infradead.org
Subject: Re: [PATCH V2 2/3] perf/imx_ddr: Add system PMU identifier for userspace
Date: Wed, 20 May 2020 08:41:54 +0100	[thread overview]
Message-ID: <20200520074154.GA23818@willie-the-truck> (raw)
In-Reply-To: <20200520025619.687-3-qiangqing.zhang@nxp.com>

On Wed, May 20, 2020 at 10:56:18AM +0800, Joakim Zhang wrote:
> The DDR Perf for i.MX8 is a system PMU whose axi id would different from
> SoC to SoC. Need expose system PMU identifier for userspace which refer
> to /sys/bus/event_source/devices/<PMU DEVICE>/identifier.
> 
> Signed-off-by: Joakim Zhang <qiangqing.zhang@nxp.com>
> ---
>  drivers/perf/fsl_imx8_ddr_perf.c | 29 +++++++++++++++++++++++++++++
>  1 file changed, 29 insertions(+)
> 
> diff --git a/drivers/perf/fsl_imx8_ddr_perf.c b/drivers/perf/fsl_imx8_ddr_perf.c
> index 90884d14f95f..ba523a94f4d7 100644
> --- a/drivers/perf/fsl_imx8_ddr_perf.c
> +++ b/drivers/perf/fsl_imx8_ddr_perf.c
> @@ -76,6 +76,7 @@ struct ddr_pmu {
>  	unsigned int cpu;
>  	struct	hlist_node node;
>  	struct	device *dev;
> +	const char *identifier;
>  	struct perf_event *events[NUM_COUNTERS];
>  	int active_events;
>  	enum cpuhp_state cpuhp_state;
> @@ -84,6 +85,27 @@ struct ddr_pmu {
>  	int id;
>  };
>  
> +static ssize_t ddr_perf_identifier_show(struct device *dev,
> +					struct device_attribute *attr,
> +					char *page)
> +{
> +	struct ddr_pmu *pmu = dev_get_drvdata(dev);
> +
> +	return sprintf(page, "%s\n", pmu->identifier);
> +}
> +
> +static struct device_attribute ddr_perf_identifier_attr =
> +	__ATTR(identifier, 0444, ddr_perf_identifier_show, NULL);
> +
> +static struct attribute *ddr_perf_identifier_attrs[] = {
> +	&ddr_perf_identifier_attr.attr,
> +	NULL,
> +};
> +
> +static struct attribute_group ddr_perf_identifier_attr_group = {
> +	.attrs = ddr_perf_identifier_attrs,
> +};
> +
>  enum ddr_perf_filter_capabilities {
>  	PERF_CAP_AXI_ID_FILTER = 0,
>  	PERF_CAP_AXI_ID_FILTER_ENHANCED,
> @@ -237,6 +259,7 @@ static const struct attribute_group *attr_groups[] = {
>  	&ddr_perf_format_attr_group,
>  	&ddr_perf_cpumask_attr_group,
>  	&ddr_perf_filter_cap_attr_group,
> +	&ddr_perf_identifier_attr_group,
>  	NULL,
>  };
>  
> @@ -601,6 +624,7 @@ static int ddr_perf_probe(struct platform_device *pdev)
>  	struct ddr_pmu *pmu;
>  	struct device_node *np;
>  	void __iomem *base;
> +	const char *identifier = NULL;
>  	char *name;
>  	int num;
>  	int ret;
> @@ -620,6 +644,11 @@ static int ddr_perf_probe(struct platform_device *pdev)
>  
>  	platform_set_drvdata(pdev, pmu);
>  
> +	ret = of_property_read_string(np, "identifier", &identifier);
> +	if (ret < 0)
> +		dev_warn(&pdev->dev, "Failed to get identifier\n");
> +	pmu->identifier = identifier;

I think this is exactly what Rob was objecting to when he said "yet another
way to identify the SoC from userspace". I've asked him on the other thread
as to what the best way to do this is.

Will

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel