[PATCH v1 0/2] perf mem: Support multiple Arm SPE PMUs

[PATCH v1 0/2] perf mem: Support multiple Arm SPE PMUs

[Leo Yan]

This patch series is to enable multiple Arm SPE PMUs.

The patch 01 is to enable multiple Arm SPE PMUs. The second patch is to
print out warning if not all CPUs support memory events, this can give
users a hint that the memory profiling is absent on some CPUs.


Leo Yan (2):
  perf arm-spe: Support multiple Arm SPE PMUs
  perf mem: Warn if memory events are not supported on all CPUs

 tools/perf/arch/arm/util/pmu.c |  2 +-
 tools/perf/util/mem-events.c   | 14 ++++++++++++++
 2 files changed, 15 insertions(+), 1 deletion(-)

-- 
2.34.1

[PATCH v1 1/2] perf arm-spe: Support multiple Arm SPE PMUs

[Leo Yan]

A platform can have more than one Arm SPE PMU. For example, a system
with multiple clusters may have each cluster enabled with its own Arm
SPE instance. In such case, the PMU devices will be named 'arm_spe_0',
'arm_spe_1', and so on.

Currently, the tool only supports 'arm_spe_0'. This commit extends
support to multiple Arm SPE PMUs by detecting the substring 'arm_spe'.

Signed-off-by: Leo Yan <leo.yan@arm.com>
---
 tools/perf/arch/arm/util/pmu.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/tools/perf/arch/arm/util/pmu.c b/tools/perf/arch/arm/util/pmu.c
index 8b7cb68ba1a8..29cfa1e427ed 100644
--- a/tools/perf/arch/arm/util/pmu.c
+++ b/tools/perf/arch/arm/util/pmu.c
@@ -27,7 +27,7 @@ void perf_pmu__arch_init(struct perf_pmu *pmu __maybe_unused)
 		pmu->selectable = true;
 		pmu->is_uncore = false;
 		pmu->perf_event_attr_init_default = arm_spe_pmu_default_config;
-		if (!strcmp(pmu->name, "arm_spe_0"))
+		if (strstr(pmu->name, "arm_spe"))
 			pmu->mem_events = perf_mem_events_arm;
 	} else if (strstarts(pmu->name, HISI_PTT_PMU_NAME)) {
 		pmu->selectable = true;
-- 
2.34.1

[PATCH v1 2/2] perf mem: Warn if memory events are not supported on all CPUs

[Leo Yan]

It is possible that memory events are not supported on all CPUs.

Prints a warning by dumping the enabled CPU maps in this case.

Signed-off-by: Leo Yan <leo.yan@arm.com>
---
 tools/perf/util/mem-events.c | 14 ++++++++++++++
 1 file changed, 14 insertions(+)

diff --git a/tools/perf/util/mem-events.c b/tools/perf/util/mem-events.c
index 6dda47bb774f..8aff2ca8bbd5 100644
--- a/tools/perf/util/mem-events.c
+++ b/tools/perf/util/mem-events.c
@@ -8,6 +8,7 @@
 #include <unistd.h>
 #include <api/fs/fs.h>
 #include <linux/kernel.h>
+#include "cpumap.h"
 #include "map_symbol.h"
 #include "mem-events.h"
 #include "mem-info.h"
@@ -242,6 +243,7 @@ int perf_mem_events__record_args(const char **rec_argv, int *argv_nr)
 	int i = *argv_nr;
 	const char *s;
 	char *copy;
+	struct perf_cpu_map *cpu_map = NULL;
 
 	while ((pmu = perf_pmus__scan_mem(pmu)) != NULL) {
 		for (int j = 0; j < PERF_MEM_EVENTS__MAX; j++) {
@@ -266,7 +268,19 @@ int perf_mem_events__record_args(const char **rec_argv, int *argv_nr)
 
 			rec_argv[i++] = "-e";
 			rec_argv[i++] = copy;
+
+			cpu_map = perf_cpu_map__merge(cpu_map, pmu->cpus);
+		}
+	}
+
+	if (cpu_map) {
+		if (!perf_cpu_map__equal(cpu_map, cpu_map__online())) {
+			char buf[200];
+
+			cpu_map__snprint(cpu_map, buf, sizeof(buf));
+			pr_warning("Memory events are enabled on a subset of CPUs: %s\n", buf);
 		}
+		perf_cpu_map__put(cpu_map);
 	}
 
 	*argv_nr = i;
-- 
2.34.1

Re: [PATCH v1 0/2] perf mem: Support multiple Arm SPE PMUs

[James Clark]

On 23/06/2024 14:34, Leo Yan wrote:
> This patch series is to enable multiple Arm SPE PMUs.
> 
> The patch 01 is to enable multiple Arm SPE PMUs. The second patch is to
> print out warning if not all CPUs support memory events, this can give
> users a hint that the memory profiling is absent on some CPUs.
> 
> 
> Leo Yan (2):
>   perf arm-spe: Support multiple Arm SPE PMUs
>   perf mem: Warn if memory events are not supported on all CPUs
> 
>  tools/perf/arch/arm/util/pmu.c |  2 +-
>  tools/perf/util/mem-events.c   | 14 ++++++++++++++
>  2 files changed, 15 insertions(+), 1 deletion(-)
> 

Reviewed-by: James Clark <james.clark@arm.com>

Re: [PATCH v1 1/2] perf arm-spe: Support multiple Arm SPE PMUs

[Ian Rogers]

On Sun, Jun 23, 2024 at 6:34 AM Leo Yan <leo.yan@arm.com> wrote:
>
> A platform can have more than one Arm SPE PMU. For example, a system
> with multiple clusters may have each cluster enabled with its own Arm
> SPE instance. In such case, the PMU devices will be named 'arm_spe_0',
> 'arm_spe_1', and so on.
>
> Currently, the tool only supports 'arm_spe_0'. This commit extends
> support to multiple Arm SPE PMUs by detecting the substring 'arm_spe'.
>
> Signed-off-by: Leo Yan <leo.yan@arm.com>
> ---
>  tools/perf/arch/arm/util/pmu.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/tools/perf/arch/arm/util/pmu.c b/tools/perf/arch/arm/util/pmu.c
> index 8b7cb68ba1a8..29cfa1e427ed 100644
> --- a/tools/perf/arch/arm/util/pmu.c
> +++ b/tools/perf/arch/arm/util/pmu.c
> @@ -27,7 +27,7 @@ void perf_pmu__arch_init(struct perf_pmu *pmu __maybe_unused)
>                 pmu->selectable = true;
>                 pmu->is_uncore = false;
>                 pmu->perf_event_attr_init_default = arm_spe_pmu_default_config;
> -               if (!strcmp(pmu->name, "arm_spe_0"))
> +               if (strstr(pmu->name, "arm_spe"))

Why not use strstarts?

Thanks,
Ian


>                         pmu->mem_events = perf_mem_events_arm;
>         } else if (strstarts(pmu->name, HISI_PTT_PMU_NAME)) {
>                 pmu->selectable = true;
> --
> 2.34.1
>

Re: [PATCH v1 1/2] perf arm-spe: Support multiple Arm SPE PMUs

[Leo Yan]

Hi Ian,

On 6/24/24 17:16, Ian Rogers wrote:
> On Sun, Jun 23, 2024 at 6:34 AM Leo Yan <leo.yan@arm.com> wrote:
>>
>> A platform can have more than one Arm SPE PMU. For example, a system
>> with multiple clusters may have each cluster enabled with its own Arm
>> SPE instance. In such case, the PMU devices will be named 'arm_spe_0',
>> 'arm_spe_1', and so on.
>>
>> Currently, the tool only supports 'arm_spe_0'. This commit extends
>> support to multiple Arm SPE PMUs by detecting the substring 'arm_spe'.
>>
>> Signed-off-by: Leo Yan <leo.yan@arm.com>
>> ---
>>   tools/perf/arch/arm/util/pmu.c | 2 +-
>>   1 file changed, 1 insertion(+), 1 deletion(-)
>>
>> diff --git a/tools/perf/arch/arm/util/pmu.c b/tools/perf/arch/arm/util/pmu.c
>> index 8b7cb68ba1a8..29cfa1e427ed 100644
>> --- a/tools/perf/arch/arm/util/pmu.c
>> +++ b/tools/perf/arch/arm/util/pmu.c
>> @@ -27,7 +27,7 @@ void perf_pmu__arch_init(struct perf_pmu *pmu __maybe_unused)
>>                  pmu->selectable = true;
>>                  pmu->is_uncore = false;
>>                  pmu->perf_event_attr_init_default = arm_spe_pmu_default_config;
>> -               if (!strcmp(pmu->name, "arm_spe_0"))
>> +               if (strstr(pmu->name, "arm_spe"))
> 
> Why not use strstarts?

Indeed, strstarts() is better, will spin for this.

Thank for suggestion.
Leo

Re: [PATCH v1 1/2] perf arm-spe: Support multiple Arm SPE PMUs

[Namhyung Kim]

Hello Leo,

On Tue, Jun 25, 2024 at 05:49:16PM +0100, Leo Yan wrote:
> Hi Ian,
> 
> On 6/24/24 17:16, Ian Rogers wrote:
> > On Sun, Jun 23, 2024 at 6:34 AM Leo Yan <leo.yan@arm.com> wrote:
> > > 
> > > A platform can have more than one Arm SPE PMU. For example, a system
> > > with multiple clusters may have each cluster enabled with its own Arm
> > > SPE instance. In such case, the PMU devices will be named 'arm_spe_0',
> > > 'arm_spe_1', and so on.
> > > 
> > > Currently, the tool only supports 'arm_spe_0'. This commit extends
> > > support to multiple Arm SPE PMUs by detecting the substring 'arm_spe'.
> > > 
> > > Signed-off-by: Leo Yan <leo.yan@arm.com>
> > > ---
> > >   tools/perf/arch/arm/util/pmu.c | 2 +-
> > >   1 file changed, 1 insertion(+), 1 deletion(-)
> > > 
> > > diff --git a/tools/perf/arch/arm/util/pmu.c b/tools/perf/arch/arm/util/pmu.c
> > > index 8b7cb68ba1a8..29cfa1e427ed 100644
> > > --- a/tools/perf/arch/arm/util/pmu.c
> > > +++ b/tools/perf/arch/arm/util/pmu.c
> > > @@ -27,7 +27,7 @@ void perf_pmu__arch_init(struct perf_pmu *pmu __maybe_unused)
> > >                  pmu->selectable = true;
> > >                  pmu->is_uncore = false;
> > >                  pmu->perf_event_attr_init_default = arm_spe_pmu_default_config;
> > > -               if (!strcmp(pmu->name, "arm_spe_0"))
> > > +               if (strstr(pmu->name, "arm_spe"))
> > 
> > Why not use strstarts?
> 
> Indeed, strstarts() is better, will spin for this.
> 
> Thank for suggestion.

Probably we need to check the last underscore too to prevent potential
name clashes..

Thanks,
Namhyung