Re: [PATCH v2 02/17] perf tools: Configure timestsamp generation in CPU-wide mode

On Fri, 7 Jun 2019 at 03:41, Suzuki K Poulose suzuki.poulose@arm.com wrote:

On 24/05/2019 18:34, Mathieu Poirier wrote:

...

When operating in CPU-wide mode tracers need to generate timestamps in order to correlate the code being traced on one CPU with what is executed on other CPUs.

Signed-off-by: Mathieu Poirier mathieu.poirier@linaro.org

tools/perf/arch/arm/util/cs-etm.c | 57 +++++++++++++++++++++++++++++++ 1 file changed, 57 insertions(+)

diff --git a/tools/perf/arch/arm/util/cs-etm.c b/tools/perf/arch/arm/util/cs-etm.c index 3912f0bf04ed..be1e4f20affa 100644 --- a/tools/perf/arch/arm/util/cs-etm.c +++ b/tools/perf/arch/arm/util/cs-etm.c @@ -99,6 +99,54 @@ static int cs_etm_set_context_id(struct auxtrace_record *itr, return err; }

+static int cs_etm_set_timestamp(struct auxtrace_record *itr,

•                         struct perf_evsel *evsel, int cpu)
  

+{

• struct cs_etm_recording *ptr;
  

• struct perf_pmu *cs_etm_pmu;
  

• char path[PATH_MAX];
  

• int err = -EINVAL;
  

• u32 val;
  

• ptr = container_of(itr, struct cs_etm_recording, itr);
  

• cs_etm_pmu = ptr->cs_etm_pmu;
  

• if (!cs_etm_is_etmv4(itr, cpu))
  

•         goto out;
  

• /* Get a handle on TRCIRD0 */
  

• snprintf(path, PATH_MAX, "cpu%d/%s",
  

•          cpu, metadata_etmv4_ro[CS_ETMV4_TRCIDR0]);
  

• err = perf_pmu__scan_file(cs_etm_pmu, path, "%x", &val);
  

• /* There was a problem reading the file, bailing out */
  

• if (err != 1) {
  

•         pr_err("%s: can't read file %s\n",
  

•                CORESIGHT_ETM_PMU_NAME, path);
  

•         goto out;
  

• }
  

• /*
  

•  * TRCIDR0.TSSIZE, bit [28-24], indicates whether global timestamping
  

•  * is supported:
  

•  *  0b00000 Global timestamping is not implemented
  

•  *  0b00110 Implementation supports a maximum timestamp of 48bits.
  

•  *  0b01000 Implementation supports a maximum timestamp of 64bits.
  

•  */
  

• val &= GENMASK(28, 24);
  

• if (!val) {
  

•         err = -EINVAL;
  

•         goto out;
  

• }
  

• /* All good, let the kernel know */
  

• evsel->attr.config |= (1 << ETM_OPT_TS);
  

• err = 0;
  

+out:

• return err;
  

+}

• static int cs_etm_set_option(struct auxtrace_record *itr, struct perf_evsel *evsel, u32 option) {

@@ -118,6 +166,11 @@ static int cs_etm_set_option(struct auxtrace_record *itr, if (err) goto out; break;

•         case ETM_OPT_TS:
  

•                 err = cs_etm_set_timestamp(itr, evsel, i);
  

•                 if (err)
  

•                         goto out;
  

•                 break;
          default:
                  goto out;
          }
  

@@ -343,6 +396,10 @@ static int cs_etm_recording_options(struct auxtrace_record *itr, err = cs_etm_set_option(itr, cs_etm_evsel, ETM_OPT_CTXTID); if (err) goto out;

•         err = cs_etm_set_option(itr, cs_etm_evsel, ETM_OPT_TS);
  

•         if (err)
  

•                 goto out;
  

nit: Could we not do this in one shot, say :

    cs_etm_set_option(itr, cs_etm_evsel, ETM_OPT_TS | ETM_OPT_CTXTID) ?

rather than iterating over the per-CPU events twice ? The cs_etm_set_option() could simply replace the switch() to :

    if (option & ETM_OPT_1)
            do_something_for_1()
    if (option & ETM_OPT_2)
            do_something_for_2();
    if (option & ~(ETM_OPT_1 | ETM_OPT_2 |...))
            /* do unsupported option */

Yes, that is a good optimization.

Arnaldo, do you prefer a new set or another patch on top of this one?

Thanks, Mathieu

Cheers Suzuki