Em Fri, Sep 06, 2013 at 09:48:54AM +0300, Adrian Hunter escreveu:
> Manipulating the sample_type of an evsel requires
> the use of:
> perf_evsel__set_sample_bit()
> and perf_evsel__reset_sample_bit()
Fine, but...
> Manipulating the sample type of an evlist requires
> the id position to be recalculated.
... can't we hide this detail? I.e. we could do it like we do with the
alloc_{mmap,poll}() evlist methods in perf_evlist__mmap(), checking if it was
already done, etc.
This we reduce the contact surface for tools using evlists and evsels.
I.e., this sequence:
+ perf_evlist__set_id_pos(evlist);
+
err = perf_evlist__open(evlist);
Could remain just:
err = perf_evlist__open(evlist);
I'm auditing the other calls to evlist now, to check if there are other cases.
Thanks for the patch!
- Arnaldo
> Signed-off-by: Adrian Hunter <adrian.hun...@intel.com>
> ---
> tools/perf/builtin-kvm.c | 20 +++++++++++---------
> 1 file changed, 11 insertions(+), 9 deletions(-)
>
> diff --git a/tools/perf/builtin-kvm.c b/tools/perf/builtin-kvm.c
> index 47b3540..0b7c5a9 100644
> --- a/tools/perf/builtin-kvm.c
> +++ b/tools/perf/builtin-kvm.c
> @@ -1165,16 +1165,16 @@ static int kvm_live_open_events(struct perf_kvm_stat
> *kvm)
> struct perf_event_attr *attr = &pos->attr;
>
> /* make sure these *are* set */
> - attr->sample_type |= PERF_SAMPLE_TID;
> - attr->sample_type |= PERF_SAMPLE_TIME;
> - attr->sample_type |= PERF_SAMPLE_CPU;
> - attr->sample_type |= PERF_SAMPLE_RAW;
> + perf_evsel__set_sample_bit(pos, TID);
> + perf_evsel__set_sample_bit(pos, TIME);
> + perf_evsel__set_sample_bit(pos, CPU);
> + perf_evsel__set_sample_bit(pos, RAW);
> /* make sure these are *not*; want as small a sample as
> possible */
> - attr->sample_type &= ~PERF_SAMPLE_PERIOD;
> - attr->sample_type &= ~PERF_SAMPLE_IP;
> - attr->sample_type &= ~PERF_SAMPLE_CALLCHAIN;
> - attr->sample_type &= ~PERF_SAMPLE_ADDR;
> - attr->sample_type &= ~PERF_SAMPLE_READ;
> + perf_evsel__reset_sample_bit(pos, PERIOD);
> + perf_evsel__reset_sample_bit(pos, IP);
> + perf_evsel__reset_sample_bit(pos, CALLCHAIN);
> + perf_evsel__reset_sample_bit(pos, ADDR);
> + perf_evsel__reset_sample_bit(pos, READ);
> attr->mmap = 0;
> attr->comm = 0;
> attr->task = 0;
> @@ -1188,6 +1188,8 @@ static int kvm_live_open_events(struct perf_kvm_stat
> *kvm)
> attr->disabled = 1;
> }
>
> + perf_evlist__set_id_pos(evlist);
> +
> err = perf_evlist__open(evlist);
> if (err < 0) {
> printf("Couldn't create the events: %s\n", strerror(errno));
> --
> 1.7.11.7
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
