lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
Open Source and information security mailing list archives
| ||
|
Date: Thu, 7 Oct 2021 19:05:58 +0200 From: Jiri Olsa <jolsa@...hat.com> To: Shunsuke Nakamura <nakamura.shun@...itsu.com> Cc: peterz@...radead.org, mingo@...hat.com, acme@...nel.org, mark.rutland@....com, alexander.shishkin@...ux.intel.com, namhyung@...nel.org, linux-kernel@...r.kernel.org, linux-perf-users@...r.kernel.org Subject: Re: [PATCH v2] libperf tests: Fix test_stat_cpu On Wed, Oct 06, 2021 at 06:48:17PM +0900, Shunsuke Nakamura wrote: > `cpu` of perf_evsel__read() must be specified the cpu index. > perf_cpu_map__for_each_cpu is for iterating the cpu number (not index) > and is not appropriate. > So, if there is an offline CPU, the cpu number specified in the argument > may point out of range because the cpu number and the cpu index are > different. nice catch > > Fix test_stat_cpu. > > Committer testing: > > # make tests -C tools/lib/perf/ > make: Entering directory '/home/nakamura/kernel_src/linux-5.15-rc4_fix/tools/lib/perf' > running static: > - running tests/test-cpumap.c...OK > - running tests/test-threadmap.c...OK > - running tests/test-evlist.c...OK > - running tests/test-evsel.c...OK > running dynamic: > - running tests/test-cpumap.c...OK > - running tests/test-threadmap.c...OK > - running tests/test-evlist.c...OK > - running tests/test-evsel.c...OK > make: Leaving directory '/home/nakamura/kernel_src/linux-5.15-rc4_fix/tools/lib/perf' > > > Signed-off-by: Shunsuke Nakamura <nakamura.shun@...itsu.com> > --- > Previous version at: > https://lore.kernel.org/lkml/20211006080456.474273-1-nakamura.shun@fujitsu.com/ > > Changes in v2: > - Remove "2/2" from Patch Subject > > tools/lib/perf/tests/test-evlist.c | 6 +++--- > tools/lib/perf/tests/test-evsel.c | 6 +++--- > 2 files changed, 6 insertions(+), 6 deletions(-) > > diff --git a/tools/lib/perf/tests/test-evlist.c b/tools/lib/perf/tests/test-evlist.c > index c67c83399170..47badd7eabf2 100644 > --- a/tools/lib/perf/tests/test-evlist.c > +++ b/tools/lib/perf/tests/test-evlist.c > @@ -40,7 +40,7 @@ static int test_stat_cpu(void) > .type = PERF_TYPE_SOFTWARE, > .config = PERF_COUNT_SW_TASK_CLOCK, > }; > - int err, cpu, tmp; > + int err, idx; > > cpus = perf_cpu_map__new(NULL); > __T("failed to create cpus", cpus); > @@ -70,10 +70,10 @@ static int test_stat_cpu(void) > perf_evlist__for_each_evsel(evlist, evsel) { > cpus = perf_evsel__cpus(evsel); > > - perf_cpu_map__for_each_cpu(cpu, tmp, cpus) { > + for (idx = 0, idx < perf_cpu_map__nr(cpus); idx++) { s/,/;/ ^ tests/test-evlist.c: In function ‘test_stat_cpu’: tests/test-evlist.c:73:52: error: expected ‘;’ before ‘)’ token 73 | for (idx = 0, idx < perf_cpu_map__nr(cpus); idx++) { | ^ | ; perf_cpu_map__for_each_cpu also returns the cpu index (tmp), maybe we could use that instead? thanks, jirka > struct perf_counts_values counts = { .val = 0 }; > > - perf_evsel__read(evsel, cpu, 0, &counts); > + perf_evsel__read(evsel, idx, 0, &counts); > __T("failed to read value for evsel", counts.val != 0); > } > } > diff --git a/tools/lib/perf/tests/test-evsel.c b/tools/lib/perf/tests/test-evsel.c > index 9abd4c0bf6db..33ae9334861a 100644 > --- a/tools/lib/perf/tests/test-evsel.c > +++ b/tools/lib/perf/tests/test-evsel.c > @@ -22,7 +22,7 @@ static int test_stat_cpu(void) > .type = PERF_TYPE_SOFTWARE, > .config = PERF_COUNT_SW_CPU_CLOCK, > }; > - int err, cpu, tmp; > + int err, idx; > > cpus = perf_cpu_map__new(NULL); > __T("failed to create cpus", cpus); > @@ -33,10 +33,10 @@ static int test_stat_cpu(void) > err = perf_evsel__open(evsel, cpus, NULL); > __T("failed to open evsel", err == 0); > > - perf_cpu_map__for_each_cpu(cpu, tmp, cpus) { > + for (idx = 0; idx < perf_cpu_map__nr(cpus); idx++) { > struct perf_counts_values counts = { .val = 0 }; > > - perf_evsel__read(evsel, cpu, 0, &counts); > + perf_evsel__read(evsel, idx, 0, &counts); > __T("failed to read value for evsel", counts.val != 0); > } > > -- > 2.25.1 >
Powered by blists - more mailing lists