diff options
| author | Hendrik Brueckner <brueckner@linux.vnet.ibm.com> | 2013-12-12 16:47:00 +0100 |
|---|---|---|
| committer | Martin Schwidefsky <schwidefsky@de.ibm.com> | 2013-12-16 14:37:52 +0100 |
| commit | 55baa2f831ae4a41da9617ab9e7cef5ebc991ec9 (patch) | |
| tree | 700ba9a81a60e1c4eaff56f57d38fd818f6de6a8 /arch | |
| parent | 8c069ff4bd6063a3f15e606c882e03f75c7e7711 (diff) | |
| download | linux-55baa2f831ae4a41da9617ab9e7cef5ebc991ec9.tar.bz2 | |
s390/perf: Improve PMU selection for PERF_COUNT_HW_CPU_CYCLES events
The cpum_cf (counter facility) PMU does not support sampling events.
With cpum_sf (sampling facility), a PMU for sampling CPU cycles is
available.
Make cpum_sf the "default" PMU for PERF_COUNT_HW_CPU_CYCLES sampling
events but use the more precise cpum_cf PMU for non-sampling events.
Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Diffstat (limited to 'arch')
| -rw-r--r-- | arch/s390/kernel/perf_cpum_sf.c | 26 |
1 files changed, 22 insertions, 4 deletions
diff --git a/arch/s390/kernel/perf_cpum_sf.c b/arch/s390/kernel/perf_cpum_sf.c index 141eca0917f4..52bf36ee91aa 100644 --- a/arch/s390/kernel/perf_cpum_sf.c +++ b/arch/s390/kernel/perf_cpum_sf.c @@ -468,11 +468,29 @@ static int cpumsf_pmu_event_init(struct perf_event *event) { int err; - if (event->attr.type != PERF_TYPE_RAW) - return -ENOENT; - - if (event->attr.config != PERF_EVENT_CPUM_SF) + /* No support for taken branch sampling */ + if (has_branch_stack(event)) + return -EOPNOTSUPP; + + switch (event->attr.type) { + case PERF_TYPE_RAW: + if (event->attr.config != PERF_EVENT_CPUM_SF) + return -ENOENT; + break; + case PERF_TYPE_HARDWARE: + /* Support sampling of CPU cycles in addition to the + * counter facility. However, the counter facility + * is more precise and, hence, restrict this PMU to + * sampling events only. + */ + if (event->attr.config != PERF_COUNT_HW_CPU_CYCLES) + return -ENOENT; + if (!is_sampling_event(event)) + return -ENOENT; + break; + default: return -ENOENT; + } if (event->cpu >= nr_cpumask_bits || (event->cpu >= 0 && !cpu_online(event->cpu))) |