perf evlist: Fix the warning logic about warn_user_request_cpu
PATCH 1/1 perf evlist: Fix the warning logic about warn_user_request_cpu
The old logic in evlist__warn_user_requested_cpus incorrectly output the warning message when I run 'perf record -C xxx' command on my context(the cpu has 8 performance-cores and 8 efficient-cores.).
The old warning like this:
# perf record -C 17 WARNING: A requested CPU in '17' is not supported by PMU 'cpu_core' (CPUs 0-15) for event 'cycles:P' # perf record -C 14 WARNING: A requested CPU in '14' is not supported by PMU 'cpu_atom' (CPUs 16-23) for event 'cycles:P' # perf record -C 14-17 WARNING: A requested CPU in '15-16' is not supported by PMU 'cpu_atom' (CPUs 16-23) for event 'cycles:P' WARNING: A requested CPU in '15-16' is not supported by PMU 'cpu_core' (CPUs 0-15) for event 'cycles:P'
After patching, the warning is as follows
# perf record -C 17 WARNING: A requested CPU '17' in '17' is not supported by PMU 'cpu_atom' (CPUs 16-23) for event 'cycles:P' # perf record -C 14 WARNING: A requested CPU '14' in '14' is not supported by PMU 'cpu_core' (CPUs 0-15) for event 'cycles:P' # perf record -C 15-18 WARNING: A requested CPU '16-18' in '15-18' is not supported by PMU 'cpu_atom' (CPUs 16-23) for event 'cycles:P' WARNING: A requested CPU '15' in '15-18' is not supported by PMU 'cpu_core' (CPUs 0-15) for event 'cycles:P'
--- tools/perf/util/evlist.c | 14 +++++++++----- 1 file changed, 9 insertions(+), 5 deletions(-)
tools/perf/util/evlist.c +9 -5
@@ -2514,12 +2514,16 @@ void evlist__warn_user_requested_cpus(struct evlist *evlist, const char *cpu_lis to_test = pmu && pmu->is_core ? pmu->cpus :cpu_map__online (); intersect =perf_cpu_map__intersect (to_test, user_requested_cpus); -if (!perf_cpu_map__equal (intersect, user_requested_cpus)) { -char buf[128 ]; -cpu_map__snprint (to_test, buf,sizeof (buf)); -pr_warning ("WARNING: A requested CPU in '%s' is not supported by PMU '%s' (CPUs %s) for event '%s'\n" , - cpu_list, pmu ? pmu->name :"cpu" , buf,evsel__name (pos)); +if (intersect &&perf_cpu_map__is_subset (user_requested_cpus, intersect)) { +char buf_test[128 ]; +char buf_intersect[128 ]; + +cpu_map__snprint (to_test, buf_test,sizeof (buf_test)); +cpu_map__snprint (intersect, buf_intersect,sizeof (buf_intersect)); +pr_warning ("WARNING: A requested CPU '%s' in '%s' is not supported by " +"PMU '%s' (CPUs %s) for event '%s'\n" , buf_intersect, cpu_list, + pmu ? pmu->name :"cpu" , buf_test,evsel__name (pos)); }perf_cpu_map__put (intersect); }
Signature · 1 line
2.34.1
Arnaldo Carvalho de Melo
maintainerQuoted 13 lines · Chunxin Zang
The old logic in
evlist__warn_user_requested_cpusincorrectly output the warning message when I run 'perf record -C xxx' command on my context(the cpu has 8 performance-cores and 8 efficient-cores.).The old warning like this:
# perf record -C 17 WARNING: A requested CPU in '17' is not supported by PMU 'cpu_core' (CPUs 0-15) for event 'cycles:P' # perf record -C 14 WARNING: A requested CPU in '14' is not supported by PMU 'cpu_atom' (CPUs 16-23) for event 'cycles:P' # perf record -C 14-17 WARNING: A requested CPU in '15-16' is not supported by PMU 'cpu_atom' (CPUs 16-23) for event 'cycles:P' WARNING: A requested CPU in '15-16' is not supported by PMU 'cpu_core' (CPUs 0-15) for event 'cycles:P'
Quoted 3 lines
After patching, the warning is as follows
# perf record -C 17 WARNING: A requested CPU '17' in '17' is not supported by PMU 'cpu_atom' (CPUs 16-23) for event 'cycles:P'
Looks wrong, i.e. I ask for CPU 17 and it says it is not in the range 16-23?
Ditto for the rest, no?
- Arnaldo
Quoted 40 lines
# perf record -C 14 WARNING: A requested CPU '14' in '14' is not supported by PMU 'cpu_core' (CPUs 0-15) for event 'cycles:P' # perf record -C 15-18 WARNING: A requested CPU '16-18' in '15-18' is not supported by PMU 'cpu_atom' (CPUs 16-23) for event 'cycles:P' WARNING: A requested CPU '15' in '15-18' is not supported by PMU 'cpu_core' (CPUs 0-15) for event 'cycles:P'Signed-off-by Chunxin Zang <[email protected]>Reviewed-by Chen Yang <[email protected]>--- tools/perf/util/evlist.c | 14 +++++++++----- 1 file changed, 9 insertions(+), 5 deletions(-)tools/perf/util/evlist.c +9
-5@@ -2514,12 +2514,16 @@ void evlist__warn_user_requested_cpus(struct evlist *evlist, const char *cpu_lis to_test = pmu && pmu->is_core ? pmu->cpus :cpu_map__online (); intersect =perf_cpu_map__intersect (to_test, user_requested_cpus); -if (!perf_cpu_map__equal (intersect, user_requested_cpus)) { -char buf[128 ]; -cpu_map__snprint (to_test, buf,sizeof (buf)); -pr_warning ("WARNING: A requested CPU in '%s' is not supported by PMU '%s' (CPUs %s) for event '%s'\n" , - cpu_list, pmu ? pmu->name :"cpu" , buf,evsel__name (pos)); +if (intersect &&perf_cpu_map__is_subset (user_requested_cpus, intersect)) { +char buf_test[128 ]; +char buf_intersect[128 ]; + +cpu_map__snprint (to_test, buf_test,sizeof (buf_test)); +cpu_map__snprint (intersect, buf_intersect,sizeof (buf_intersect)); +pr_warning ("WARNING: A requested CPU '%s' in '%s' is not supported by " +"PMU '%s' (CPUs %s) for event '%s'\n" , buf_intersect, cpu_list, + pmu ? pmu->name :"cpu" , buf_test,evsel__name (pos)); }perf_cpu_map__put (intersect); }Signature · 1 line
2.34.1
James Clark
reviewerQuoted 18 lines · Arnaldo Carvalho de Melo
Quoted 13 lines · Chunxin Zang
The old logic in
evlist__warn_user_requested_cpusincorrectly output the warning message when I run 'perf record -C xxx' command on my context(the cpu has 8 performance-cores and 8 efficient-cores.).The old warning like this:
# perf record -C 17 WARNING: A requested CPU in '17' is not supported by PMU 'cpu_core' (CPUs 0-15) for event 'cycles:P' # perf record -C 14 WARNING: A requested CPU in '14' is not supported by PMU 'cpu_atom' (CPUs 16-23) for event 'cycles:P' # perf record -C 14-17 WARNING: A requested CPU in '15-16' is not supported by PMU 'cpu_atom' (CPUs 16-23) for event 'cycles:P' WARNING: A requested CPU in '15-16' is not supported by PMU 'cpu_core' (CPUs 0-15) for event 'cycles:P'Quoted 3 lines
After patching, the warning is as follows
# perf record -C 17 WARNING: A requested CPU '17' in '17' is not supported by PMU 'cpu_atom' (CPUs 16-23) for event 'cycles:P'
Quoted 7 lines
Looks wrong, i.e. I ask for CPU 17 and it says it is not in the range 16-23?
Ditto for the rest, no?
- Arnaldo
Is this fix needed if we have https://lore.kernel.org/linux-perf-users/[email protected]/T/#t ?
I think some of these warnings might not be shown in that case due to fixing the other issue.
Quoted 41 lines
Quoted 40 lines
# perf record -C 14 WARNING: A requested CPU '14' in '14' is not supported by PMU 'cpu_core' (CPUs 0-15) for event 'cycles:P' # perf record -C 15-18 WARNING: A requested CPU '16-18' in '15-18' is not supported by PMU 'cpu_atom' (CPUs 16-23) for event 'cycles:P' WARNING: A requested CPU '15' in '15-18' is not supported by PMU 'cpu_core' (CPUs 0-15) for event 'cycles:P'Signed-off-by Chunxin Zang <[email protected]>Reviewed-by Chen Yang <[email protected]>--- tools/perf/util/evlist.c | 14 +++++++++----- 1 file changed, 9 insertions(+), 5 deletions(-)tools/perf/util/evlist.c +9
-5@@ -2514,12 +2514,16 @@ void evlist__warn_user_requested_cpus(struct evlist *evlist, const char *cpu_lis to_test = pmu && pmu->is_core ? pmu->cpus :cpu_map__online (); intersect =perf_cpu_map__intersect (to_test, user_requested_cpus); -if (!perf_cpu_map__equal (intersect, user_requested_cpus)) { -char buf[128 ]; -cpu_map__snprint (to_test, buf,sizeof (buf)); -pr_warning ("WARNING: A requested CPU in '%s' is not supported by PMU '%s' (CPUs %s) for event '%s'\n" , - cpu_list, pmu ? pmu->name :"cpu" , buf,evsel__name (pos)); +if (intersect &&perf_cpu_map__is_subset (user_requested_cpus, intersect)) { +char buf_test[128 ]; +char buf_intersect[128 ]; + +cpu_map__snprint (to_test, buf_test,sizeof (buf_test)); +cpu_map__snprint (intersect, buf_intersect,sizeof (buf_intersect)); +pr_warning ("WARNING: A requested CPU '%s' in '%s' is not supported by " +"PMU '%s' (CPUs %s) for event '%s'\n" , buf_intersect, cpu_list, + pmu ? pmu->name :"cpu" , buf_test,evsel__name (pos)); }perf_cpu_map__put (intersect); }Signature · 1 line
2.34.1
Quoted 5 lines · James Clark
.
Hi James, Arnaldo
At that time, I also released a v2 version, and Ian Rogers provided some suggestions for modifications. However, I'm very sorry that later, because I was busy with other things, I forgot about this issue.
Now I'm very glad to hear that this problem has been fixed. I'll verify it in my environment when I have time this weekend.
https://lore.kernel.org/lkml/[email protected]/
Best wishes
Chunxin