[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <27c8f898-93f3-6a2d-5e9e-b5b1225df1c9@arm.com>
Date: Mon, 11 Jan 2021 09:47:31 +0000
From: Suzuki K Poulose <suzuki.poulose@....com>
To: Leo Yan <leo.yan@...aro.org>,
Arnaldo Carvalho de Melo <acme@...nel.org>,
Mathieu Poirier <mathieu.poirier@...aro.org>,
Mike Leach <mike.leach@...aro.org>,
Alexander Shishkin <alexander.shishkin@...ux.intel.com>,
John Garry <john.garry@...wei.com>,
Will Deacon <will@...nel.org>,
Peter Zijlstra <peterz@...radead.org>,
Ingo Molnar <mingo@...hat.com>,
Mark Rutland <mark.rutland@....com>,
Jiri Olsa <jolsa@...hat.com>,
Namhyung Kim <namhyung@...nel.org>,
Daniel Kiss <Daniel.Kiss@....com>,
Denis Nikitin <denik@...omium.org>, coresight@...ts.linaro.org,
linux-arm-kernel@...ts.infradead.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH v1 5/7] perf cs-etm: Fixup PID_FMT when it is zero
On 1/9/21 7:44 AM, Leo Yan wrote:
> If the metadata item CS_ETM_PID_FMT/CS_ETMV4_PID_FMT is zero, this means
> the perf data file is recorded with old version tool and the tool has
> not extended to support the item.
>
> For this case, this patch fixes up PID_FMT entry to set the value as
> BIT(ETM_OPT_CTXTID), this info will be delivered to the decoder to
> extract PID from packet's field "context_id".
>
> Signed-off-by: Leo Yan <leo.yan@...aro.org>
Acked-by: Suzuki K Poulose <suzuki.poulose@....com>
> ---
> tools/perf/util/cs-etm.c | 19 +++++++++++++++++++
> 1 file changed, 19 insertions(+)
>
> diff --git a/tools/perf/util/cs-etm.c b/tools/perf/util/cs-etm.c
> index 763085db29ae..8c125134a756 100644
> --- a/tools/perf/util/cs-etm.c
> +++ b/tools/perf/util/cs-etm.c
> @@ -7,6 +7,7 @@
> */
>
> #include <linux/bitops.h>
> +#include <linux/coresight-pmu.h>
> #include <linux/err.h>
> #include <linux/kernel.h>
> #include <linux/log2.h>
> @@ -2577,6 +2578,15 @@ int cs_etm__process_auxtrace_info(union perf_event *event,
> for (k = 0; k < metadata_cpu_array_size; k++)
> metadata[j][k] = ptr[i + k];
>
> + /*
> + * If the data in CS_ETM_PID_FMT is zero, means the
> + * information isn't stored in the data file, this is
> + * because the old perf tool hasn't yet supported
> + * CS_ETM_PID_FMT. Fixup the item to option "CTXTID".
> + */
> + if (!metadata[j][CS_ETM_PID_FMT])
> + metadata[j][CS_ETM_PID_FMT] = BIT(ETM_OPT_CTXTID);
> +
> /* The traceID is our handle */
> idx = metadata[j][CS_ETM_ETMTRACEIDR];
> i += metadata_cpu_array_size;
> @@ -2590,6 +2600,15 @@ int cs_etm__process_auxtrace_info(union perf_event *event,
> for (k = 0; k < metadata_cpu_array_size; k++)
> metadata[j][k] = ptr[i + k];
>
> + /*
> + * If the data in CS_ETMV4_PID_FMT is zero, means the
> + * information isn't stored in the data file, this is
> + * because the old perf tool hasn't yet supported
> + * CS_ETMV4_PID_FMT. Fixup the item to option "CTXTID".
> + */
> + if (!metadata[j][CS_ETMV4_PID_FMT])
> + metadata[j][CS_ETMV4_PID_FMT] = BIT(ETM_OPT_CTXTID);
> +
> /* The traceID is our handle */
> idx = metadata[j][CS_ETMV4_TRCTRACEIDR];
> i += metadata_cpu_array_size;
>
Powered by blists - more mailing lists