perf tools: Fix record failure when mixed with ARM SPE event
v2 cover perf tools: Fix record failure when mixed with ARM SPE event
v1 -> v2:
- Optimize code in patch 1 as Mathieu adviced.
- Fix memleak in patch 2.
- Detail the commit info to explain the reason.
This patch set fixes perf record failure when we mix arm_spe_x event with other events in specific order.
Wei Li (2):
perf tools: Fix record failure when mixed with ARM SPE event perf tools: ARM SPE code cleanup
tools/perf/arch/arm/util/auxtrace.c | 17 ++++++++--------- 1 file changed, 8 insertions(+), 9 deletions(-)
Signature · 1 line
2.17.1
PATCH v2 1/2 perf tools: Fix record failure when mixed with ARM SPE event
When recording with cache-misses and arm_spe_x event, i found that it will just fail without showing any error info if i put cache-misses after 'arm_spe_x' event.
[root@localhost 0620]# perf record -e cache-misses -e \
arm_spe_0/ts_enable=1,pct_enable=1,pa_enable=1,load_filter=1,\
jitter=1,store_filter=1,min_latency=0/ sleep 1
[ perf record: Woken up 1 times to write data ]
[ perf record: Captured and wrote 0.067 MB perf.data ] [root@localhost 0620]# perf record -e \
arm_spe_0/ts_enable=1,pct_enable=1,pa_enable=1,load_filter=1,jitter=1,\
store_filter=1,min_latency=0/ -e cache-misses sleep 1 [root@localhost 0620]#
The current code can only work if the only event to be traced is an 'arm_spe_x', or if it is the last event to be specified. Otherwise the last event type will be checked against all the arm_spe_pmus[i]->types, none will match and an out of bound 'i' index will be used in arm_spe_recording_init().
We don't support concurrent multiple arm_spe_x events currently, that is checked in arm_spe_recording_options(), and it will show the relevant info. So add the check and record of the first found 'arm_spe_pmu' to fix this issue here.
--- tools/perf/arch/arm/util/auxtrace.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-)
tools/perf/arch/arm/util/auxtrace.c +4 -4
@@ -56,7 +56,7 @@ struct auxtrace_recordstruct perf_pmu *cs_etm_pmu;struct evsel *evsel;bool found_etm = false; -bool found_spe = false; +struct perf_pmu *found_spe =NULL ;static struct perf_pmu **arm_spe_pmus =NULL ;static int nr_spes =0 ;int i =0 ;
@@ -74,12 +74,12 @@ struct auxtrace_record evsel->core.attr.type == cs_etm_pmu->type) found_etm = true; -if (!nr_spes) +if (!nr_spes || found_spe)continue ;for (i =0 ; i < nr_spes; i++) {if (evsel->core.attr.type == arm_spe_pmus[i]->type) { - found_spe = true; + found_spe = arm_spe_pmus[i];break ; } }
@@ -96,7 +96,7 @@ struct auxtrace_record#if defined(__aarch64__)if (found_spe) -return arm_spe_recording_init (err, arm_spe_pmus[i]); +return arm_spe_recording_init (err, found_spe);#endif /*
Signature · 1 line
2.17.1
Quoted 20 lines · Wei Li
When recording with cache-misses and
arm_spe_xevent, i found that it will just fail without showing any error info if i put cache-misses after 'arm_spe_x' event.[root@localhost 0620]# perf record -e cache-misses -e \
arm_spe_0/ts_enable=1,pct_enable=1,pa_enable=1,load_filter=1,\
jitter=1,store_filter=1,min_latency=0/ sleep 1
[ perf record: Woken up 1 times to write data ]
[ perf record: Captured and wrote 0.067 MB perf.data ] [root@localhost 0620]# perf record -e \
arm_spe_0/ts_enable=1,pct_enable=1,pa_enable=1,load_filter=1,jitter=1,\
store_filter=1,min_latency=0/ -e cache-misses sleep 1 [root@localhost 0620]#The current code can only work if the only event to be traced is an '
arm_spe_x', or if it is the last event to be specified. Otherwise the last event type will be checked against all the arm_spe_pmus[i]->types, none will match and an out of bound 'i' index will be used inarm_spe_recording_init().
Quoted 7 lines
We don't support concurrent multiple
arm_spe_xevents currently, that is checked inarm_spe_recording_options(), and it will show the relevant info. So add the check and record of the first found 'arm_spe_pmu' to fix this issue here.Fixes ffd3d18c20b8d ("perf tools: Add ARM Statistical Profiling Extensions (SPE) support")Signed-off-by Wei Li <[email protected]>
Thanks for the patch, Wei. I have tested this series on Arm64 D06 platform:
I'd like to wait for Mathieu's ACK.
Thanks,
Leo
Quoted 44 lines
--- tools/perf/arch/arm/util/auxtrace.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-)tools/perf/arch/arm/util/auxtrace.c +4
-4@@ -56,7 +56,7 @@ struct auxtrace_recordstruct perf_pmu *cs_etm_pmu;struct evsel *evsel;bool found_etm = false; -bool found_spe = false; +struct perf_pmu *found_spe =NULL ;static struct perf_pmu **arm_spe_pmus =NULL ;static int nr_spes =0 ;int i =0 ;@@ -74,12 +74,12 @@ struct auxtrace_record evsel->core.attr.type == cs_etm_pmu->type) found_etm = true; -if (!nr_spes) +if (!nr_spes || found_spe)continue ;for (i =0 ; i < nr_spes; i++) {if (evsel->core.attr.type == arm_spe_pmus[i]->type) { - found_spe = true; + found_spe = arm_spe_pmus[i];break ; } }@@ -96,7 +96,7 @@ struct auxtrace_record#if defined(__aarch64__)if (found_spe) -return arm_spe_recording_init (err, arm_spe_pmus[i]); +return arm_spe_recording_init (err, found_spe);#endif /* Signature · 1 line
2.17.1
Quoted 20 lines · Wei Li
When recording with cache-misses and
arm_spe_xevent, i found that it will just fail without showing any error info if i put cache-misses after 'arm_spe_x' event.[root@localhost 0620]# perf record -e cache-misses -e \
arm_spe_0/ts_enable=1,pct_enable=1,pa_enable=1,load_filter=1,\
jitter=1,store_filter=1,min_latency=0/ sleep 1
[ perf record: Woken up 1 times to write data ]
[ perf record: Captured and wrote 0.067 MB perf.data ] [root@localhost 0620]# perf record -e \
arm_spe_0/ts_enable=1,pct_enable=1,pa_enable=1,load_filter=1,jitter=1,\
store_filter=1,min_latency=0/ -e cache-misses sleep 1 [root@localhost 0620]#The current code can only work if the only event to be traced is an '
arm_spe_x', or if it is the last event to be specified. Otherwise the last event type will be checked against all the arm_spe_pmus[i]->types, none will match and an out of bound 'i' index will be used inarm_spe_recording_init().
Quoted 6 lines
We don't support concurrent multiple
arm_spe_xevents currently, that is checked inarm_spe_recording_options(), and it will show the relevant info. So add the check and record of the first found 'arm_spe_pmu' to fix this issue here.Fixes ffd3d18c20b8d ("perf tools: Add ARM Statistical Profiling Extensions (SPE) support")
Usually SHA1 are 12 character long rather than 13. Depending on what Arnaldo wants to do you may have to resend.
Quoted 1 line
Signed-off-by Wei Li <[email protected]>
Quoted 44 lines
--- tools/perf/arch/arm/util/auxtrace.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-)tools/perf/arch/arm/util/auxtrace.c +4
-4@@ -56,7 +56,7 @@ struct auxtrace_recordstruct perf_pmu *cs_etm_pmu;struct evsel *evsel;bool found_etm = false; -bool found_spe = false; +struct perf_pmu *found_spe =NULL ;static struct perf_pmu **arm_spe_pmus =NULL ;static int nr_spes =0 ;int i =0 ;@@ -74,12 +74,12 @@ struct auxtrace_record evsel->core.attr.type == cs_etm_pmu->type) found_etm = true; -if (!nr_spes) +if (!nr_spes || found_spe)continue ;for (i =0 ; i < nr_spes; i++) {if (evsel->core.attr.type == arm_spe_pmus[i]->type) { - found_spe = true; + found_spe = arm_spe_pmus[i];break ; } }@@ -96,7 +96,7 @@ struct auxtrace_record#if defined(__aarch64__)if (found_spe) -return arm_spe_recording_init (err, arm_spe_pmus[i]); +return arm_spe_recording_init (err, found_spe);#endif /* Signature · 1 line
2.17.1
PATCH v2 2/2 perf tools: ARM SPE code cleanup
In reply to Wei Li- Firstly, the function auxtrace_record__init() will be invoked only once, the variable "arm_spe_pmus" will not be used afterwards, thus we don't need to check "arm_spe_pmus" is NULL or not;
- Another reason is, even though SPE is micro-architecture dependent, but so far it only supports "statistical-profiling-extension-v1" and we have no chance to use multiple SPE's PMU events in Perf command.
So remove the useless check code to make it clear.
--- tools/perf/arch/arm/util/auxtrace.c | 9 ++++----- 1 file changed, 4 insertions(+), 5 deletions(-)
tools/perf/arch/arm/util/auxtrace.c +4 -5
@@ -57,17 +57,15 @@ struct auxtrace_recordstruct evsel *evsel;bool found_etm = false;struct perf_pmu *found_spe =NULL ; -static struct perf_pmu **arm_spe_pmus =NULL ; -static int nr_spes =0 ; +struct perf_pmu **arm_spe_pmus =NULL ; +int nr_spes =0 ;int i =0 ;if (!evlist)return NULL ; cs_etm_pmu =perf_pmu__find (CORESIGHT_ETM_PMU_NAME ); - -if (!arm_spe_pmus) - arm_spe_pmus =find_all_arm_spe_pmus (&nr_spes, err); + arm_spe_pmus =find_all_arm_spe_pmus (&nr_spes, err);evlist__for_each_entry (evlist, evsel) { if (cs_etm_pmu &&
@@ -84,6 +82,7 @@ struct auxtrace_record } } } +free (arm_spe_pmus);if (found_etm && found_spe) {pr_err ("Concurrent ARM Coresight ETM and SPE operation not currently supported\n" );
Signature · 1 line
2.17.1
Re: [PATCH v2 2/2] perf tools: ARM SPE code cleanup
Quoted 6 lines · Wei Li
- Firstly, the function
auxtrace_record__init()will be invoked only once, the variable "arm_spe_pmus" will not be used afterwards, thus we don't need to check "arm_spe_pmus" is NULL or not;
- Another reason is, even though SPE is micro-architecture dependent, but so far it only supports "statistical-profiling-extension-v1" and we have no chance to use multiple SPE's PMU events in Perf command.
I find the above changelog somewhat out of touch with the patch itself. The only thing that is happening here is the removal of a useless check and a fix for a memory leak.
Once again whether Arnaldo wants to make the changes by hand or not you may have to resubmit.
Quoted 44 lines
So remove the useless check code to make it clear.
Signed-off-by Wei Li <[email protected]>--- tools/perf/arch/arm/util/auxtrace.c | 9 ++++----- 1 file changed, 4 insertions(+), 5 deletions(-)tools/perf/arch/arm/util/auxtrace.c +4
-5@@ -57,17 +57,15 @@ struct auxtrace_recordstruct evsel *evsel;bool found_etm = false;struct perf_pmu *found_spe =NULL ; -static struct perf_pmu **arm_spe_pmus =NULL ; -static int nr_spes =0 ; +struct perf_pmu **arm_spe_pmus =NULL ; +int nr_spes =0 ;int i =0 ;if (!evlist)return NULL ; cs_etm_pmu =perf_pmu__find (CORESIGHT_ETM_PMU_NAME ); - -if (!arm_spe_pmus) - arm_spe_pmus =find_all_arm_spe_pmus (&nr_spes, err); + arm_spe_pmus =find_all_arm_spe_pmus (&nr_spes, err);evlist__for_each_entry (evlist, evsel) { if (cs_etm_pmu &&@@ -84,6 +82,7 @@ struct auxtrace_record } } } +free (arm_spe_pmus);if (found_etm && found_spe) {pr_err ("Concurrent ARM Coresight ETM and SPE operation not currently supported\n" );Signature · 1 line
2.17.1
Arnaldo Carvalho de Melo
maintainerRe: [PATCH v2 2/2] perf tools: ARM SPE code cleanup
Em Mon, Jul 27, 2020 at 02:34:36PM -0600, Mathieu Poirier escreveu:
Quoted 8 lines · Wei Li
Quoted 6 lines
- Firstly, the function
auxtrace_record__init()will be invoked only once, the variable "arm_spe_pmus" will not be used afterwards, thus we don't need to check "arm_spe_pmus" is NULL or not;
- Another reason is, even though SPE is micro-architecture dependent, but so far it only supports "statistical-profiling-extension-v1" and we have no chance to use multiple SPE's PMU events in Perf command.
Quoted 6 lines
I find the above changelog somewhat out of touch with the patch itself. The only thing that is happening here is the removal of a useless check and a fix for a memory leak.
Once again whether Arnaldo wants to make the changes by hand or not you may have to resubmit.
I'll fix it now, thanks for reviewing.
- Arnaldo
Quoted 46 lines
Reviewed-by Mathieu Poirier <[email protected]>Quoted 44 lines
So remove the useless check code to make it clear.
Signed-off-by Wei Li <[email protected]>--- tools/perf/arch/arm/util/auxtrace.c | 9 ++++----- 1 file changed, 4 insertions(+), 5 deletions(-)tools/perf/arch/arm/util/auxtrace.c +4
-5@@ -57,17 +57,15 @@ struct auxtrace_recordstruct evsel *evsel;bool found_etm = false;struct perf_pmu *found_spe =NULL ; -static struct perf_pmu **arm_spe_pmus =NULL ; -static int nr_spes =0 ; +struct perf_pmu **arm_spe_pmus =NULL ; +int nr_spes =0 ;int i =0 ;if (!evlist)return NULL ; cs_etm_pmu =perf_pmu__find (CORESIGHT_ETM_PMU_NAME ); - -if (!arm_spe_pmus) - arm_spe_pmus =find_all_arm_spe_pmus (&nr_spes, err); + arm_spe_pmus =find_all_arm_spe_pmus (&nr_spes, err);evlist__for_each_entry (evlist, evsel) { if (cs_etm_pmu &&@@ -84,6 +82,7 @@ struct auxtrace_record } } } +free (arm_spe_pmus);if (found_etm && found_spe) {pr_err ("Concurrent ARM Coresight ETM and SPE operation not currently supported\n" );Signature · 1 line
2.17.1
Signature · 1 line
- Arnaldo
Arnaldo Carvalho de Melo
maintainerRe: [PATCH v2 2/2] perf tools: ARM SPE code cleanup
In reply to Mathieu PoirierEm Mon, Jul 27, 2020 at 02:34:36PM -0600, Mathieu Poirier escreveu:
Quoted 11 lines · Wei Li
Quoted 6 lines
- Firstly, the function
auxtrace_record__init()will be invoked only once, the variable "arm_spe_pmus" will not be used afterwards, thus we don't need to check "arm_spe_pmus" is NULL or not;
- Another reason is, even though SPE is micro-architecture dependent, but so far it only supports "statistical-profiling-extension-v1" and we have no chance to use multiple SPE's PMU events in Perf command.I find the above changelog somewhat out of touch with the patch itself. The only thing that is happening here is the removal of a useless check and a fix for a memory leak.
Humm, I think the original intent of that code was to cache the results of find_all_arm_spe_pmus(), as the variable it is assigned to is static.
So not a leak, as there was that static reference to it to reuse it later, but that is strange in a function named "__init()" which usually is called only once, anyway, so I think that the paragraph with "Firstly" is kinda ok, but confusing, I think it should read:
- auxtrace_record__init() is called only once, so there is no point in using a static variable to cache the results of find_all_arm_spe_pmus(), make it local and free the results after use.
The second paragraph is SPE specific, so I'm not qualified to judge on it.
I'm replacing the first paragraph with the version I wrote and keep it in my local branch, please holler if you think I misunderstood.
- Arnaldo
Quoted 49 lines
Once again whether Arnaldo wants to make the changes by hand or not you may have to resubmit.
Reviewed-by Mathieu Poirier <[email protected]>Quoted 44 lines
So remove the useless check code to make it clear.
Signed-off-by Wei Li <[email protected]>--- tools/perf/arch/arm/util/auxtrace.c | 9 ++++----- 1 file changed, 4 insertions(+), 5 deletions(-)tools/perf/arch/arm/util/auxtrace.c +4
-5@@ -57,17 +57,15 @@ struct auxtrace_recordstruct evsel *evsel;bool found_etm = false;struct perf_pmu *found_spe =NULL ; -static struct perf_pmu **arm_spe_pmus =NULL ; -static int nr_spes =0 ; +struct perf_pmu **arm_spe_pmus =NULL ; +int nr_spes =0 ;int i =0 ;if (!evlist)return NULL ; cs_etm_pmu =perf_pmu__find (CORESIGHT_ETM_PMU_NAME ); - -if (!arm_spe_pmus) - arm_spe_pmus =find_all_arm_spe_pmus (&nr_spes, err); + arm_spe_pmus =find_all_arm_spe_pmus (&nr_spes, err);evlist__for_each_entry (evlist, evsel) { if (cs_etm_pmu &&@@ -84,6 +82,7 @@ struct auxtrace_record } } } +free (arm_spe_pmus);if (found_etm && found_spe) {pr_err ("Concurrent ARM Coresight ETM and SPE operation not currently supported\n" );Signature · 1 line
2.17.1
Signature · 1 line
- Arnaldo
Re: [PATCH v2 2/2] perf tools: ARM SPE code cleanup
Hi Arnaldo,
Quoted 25 lines · Arnaldo Carvalho de Melo
Em Mon, Jul 27, 2020 at 02:34:36PM -0600, Mathieu Poirier escreveu:
Quoted 11 lines · Wei Li
Quoted 6 lines
- Firstly, the function
auxtrace_record__init()will be invoked only once, the variable "arm_spe_pmus" will not be used afterwards, thus we don't need to check "arm_spe_pmus" is NULL or not;
- Another reason is, even though SPE is micro-architecture dependent, but so far it only supports "statistical-profiling-extension-v1" and we have no chance to use multiple SPE's PMU events in Perf command.I find the above changelog somewhat out of touch with the patch itself. The only thing that is happening here is the removal of a useless check and a fix for a memory leak.
Humm, I think the original intent of that code was to cache the results of
find_all_arm_spe_pmus(), as the variable it is assigned to is static.So not a leak, as there was that static reference to it to reuse it later, but that is strange in a function named "
__init()" which usually is called only once, anyway, so I think that the paragraph with "Firstly" is kinda ok, but confusing, I think it should read:-
auxtrace_record__init()is called only once, so there is no point in using a static variable to cache the results offind_all_arm_spe_pmus(), make it local and free the results after use.
Quoted 5 lines
The second paragraph is SPE specific, so I'm not qualified to judge on it.
I'm replacing the first paragraph with the version I wrote and keep it in my local branch, please holler if you think I misunderstood.
Thanks a lot for this. These two paragraphs were coming from reviewing and comments, but I think your rephrasing is very sufficient to describe what this patch is doing :)
Thanks,
Leo
Quoted 53 lines
Quoted 49 lines
Once again whether Arnaldo wants to make the changes by hand or not you may have to resubmit.
Reviewed-by Mathieu Poirier <[email protected]>Quoted 44 lines
So remove the useless check code to make it clear.
Signed-off-by Wei Li <[email protected]>--- tools/perf/arch/arm/util/auxtrace.c | 9 ++++----- 1 file changed, 4 insertions(+), 5 deletions(-)tools/perf/arch/arm/util/auxtrace.c +4
-5@@ -57,17 +57,15 @@ struct auxtrace_recordstruct evsel *evsel;bool found_etm = false;struct perf_pmu *found_spe =NULL ; -static struct perf_pmu **arm_spe_pmus =NULL ; -static int nr_spes =0 ; +struct perf_pmu **arm_spe_pmus =NULL ; +int nr_spes =0 ;int i =0 ;if (!evlist)return NULL ; cs_etm_pmu =perf_pmu__find (CORESIGHT_ETM_PMU_NAME ); - -if (!arm_spe_pmus) - arm_spe_pmus =find_all_arm_spe_pmus (&nr_spes, err); + arm_spe_pmus =find_all_arm_spe_pmus (&nr_spes, err);evlist__for_each_entry (evlist, evsel) { if (cs_etm_pmu &&@@ -84,6 +82,7 @@ struct auxtrace_record } } } +free (arm_spe_pmus);if (found_etm && found_spe) {pr_err ("Concurrent ARM Coresight ETM and SPE operation not currently supported\n" );Signature · 1 line
2.17.1Signature · 1 line
- Arnaldo
Re: [PATCH v2 2/2] perf tools: ARM SPE code cleanup
In reply to Arnaldo Carvalho de MeloQuoted 10 lines · Arnaldo Carvalho de Melo
Em Mon, Jul 27, 2020 at 02:34:36PM -0600, Mathieu Poirier escreveu:
Quoted 8 lines · Wei Li
Quoted 6 lines
- Firstly, the function
auxtrace_record__init()will be invoked only once, the variable "arm_spe_pmus" will not be used afterwards, thus we don't need to check "arm_spe_pmus" is NULL or not;
- Another reason is, even though SPE is micro-architecture dependent, but so far it only supports "statistical-profiling-extension-v1" and we have no chance to use multiple SPE's PMU events in Perf command.
Quoted 6 lines
Quoted 3 lines
I find the above changelog somewhat out of touch with the patch itself. The only thing that is happening here is the removal of a useless check and a fix for a memory leak.
Humm, I think the original intent of that code was to cache the results of
find_all_arm_spe_pmus(), as the variable it is assigned to is static.
Correct, but as you pointed out below the function is called only once. And there is still a leak as that memory is never freed.
Quoted 9 lines
So not a leak, as there was that static reference to it to reuse it later, but that is strange in a function named "
__init()" which usually is called only once, anyway, so I think that the paragraph with "Firstly" is kinda ok, but confusing, I think it should read:-
auxtrace_record__init()is called only once, so there is no point in using a static variable to cache the results offind_all_arm_spe_pmus(), make it local and free the results after use.
This is exactly what this patch does and what the changelog should read.
Quoted 7 lines
The second paragraph is SPE specific, so I'm not qualified to judge on it.
I'm replacing the first paragraph with the version I wrote and keep it in my local branch, please holler if you think I misunderstood.
There is no point for the next paragraph, it has no relevance to what the code is doing.
Thanks for the editing.
Quoted 55 lines
- Arnaldo
Quoted 44 lines
Once again whether Arnaldo wants to make the changes by hand or not you may have to resubmit.
Reviewed-by Mathieu Poirier <[email protected]>Quoted 39 lines
So remove the useless check code to make it clear.
Signed-off-by Wei Li <[email protected]>--- tools/perf/arch/arm/util/auxtrace.c | 9 ++++----- 1 file changed, 4 insertions(+), 5 deletions(-)tools/perf/arch/arm/util/auxtrace.c +4
-5@@ -57,17 +57,15 @@ struct auxtrace_recordstruct evsel *evsel;bool found_etm = false;struct perf_pmu *found_spe =NULL ; -static struct perf_pmu **arm_spe_pmus =NULL ; -static int nr_spes =0 ; +struct perf_pmu **arm_spe_pmus =NULL ; +int nr_spes =0 ;int i =0 ;if (!evlist)return NULL ; cs_etm_pmu =perf_pmu__find (CORESIGHT_ETM_PMU_NAME ); - -if (!arm_spe_pmus) - arm_spe_pmus =find_all_arm_spe_pmus (&nr_spes, err); + arm_spe_pmus =find_all_arm_spe_pmus (&nr_spes, err);evlist__for_each_entry (evlist, evsel) { if (cs_etm_pmu &&@@ -84,6 +82,7 @@ struct auxtrace_record } } } +free (arm_spe_pmus);Quoted 5 lines
Quoted 5 lines
tools/perf/arch/arm/util/auxtrace.c +0
-1if (found_etm && found_spe) {pr_err ("Concurrent ARM Coresight ETM and SPE operation not currently supported\n" ); --2.17.1
--
- Arnaldo