Re: [PATCH 1/2] perf record: Respect --no-switch-events

[Arnaldo Carvalho de Melo]

Em Thu, May 28, 2020 at 03:08:58PM +0300, Adrian Hunter escreveu:
> Context switch events are added automatically by Intel PT and Coresight.
> Make it possible to suppress them. That is useful for tracing the
> scheduler without the disturbance that the switch event processing
> creates.
> 
> Example:
> 
>   Prerequisites:
> 
>     $ which perf
>     ~/bin/perf
>     $ sudo setcap "cap_sys_rawio,cap_sys_admin,cap_sys_ptrace,cap_syslog,cap_ipc_lock=ep" ~/bin/perf
>     $ sudo chmod +r /proc/kcore
> 
>   Before:
> 
>     $ perf record --no-switch-events --kcore -a -e intel_pt//k -- sleep 0.001
>     [ perf record: Woken up 1 times to write data ]
>     [ perf record: Captured and wrote 0.938 MB perf.data ]
>     $ perf script -D | grep PERF_RECORD_SWITCH | wc -l
>     572
> 
>   After:
> 
>     $ perf record --no-switch-events --kcore -a -e intel_pt//k -- sleep 0.001
>     Warning:
>     Intel Processor Trace decoding will not be possible except for kernel tracing!
>     [ perf record: Woken up 1 times to write data ]
>     [ perf record: Captured and wrote 0.838 MB perf.data ]
>     $ perf script -D | grep PERF_RECORD_SWITCH | wc -l
>     0
> 
>     $ sudo chmod go-r /proc/kcore
>     $ sudo setcap -r ~/bin/perf

Thanks, tested as root and applied.

- Arnaldo
 
> Signed-off-by: Adrian Hunter <adrian.hunter@xxxxxxxxx>
> ---
>  tools/perf/Documentation/perf-record.txt | 4 +++-
>  tools/perf/arch/arm/util/cs-etm.c        | 3 ++-
>  tools/perf/arch/x86/util/intel-pt.c      | 3 ++-
>  tools/perf/builtin-record.c              | 5 +++--
>  tools/perf/util/record.h                 | 6 ++++++
>  5 files changed, 16 insertions(+), 5 deletions(-)
> 
> diff --git a/tools/perf/Documentation/perf-record.txt b/tools/perf/Documentation/perf-record.txt
> index 561ef55743e2..97b1a866ab22 100644
> --- a/tools/perf/Documentation/perf-record.txt
> +++ b/tools/perf/Documentation/perf-record.txt
> @@ -458,7 +458,9 @@ This option sets the time out limit. The default value is 500 ms.
>  
>  --switch-events::
>  Record context switch events i.e. events of type PERF_RECORD_SWITCH or
> -PERF_RECORD_SWITCH_CPU_WIDE.
> +PERF_RECORD_SWITCH_CPU_WIDE. In some cases (e.g. Intel PT or CoreSight)
> +switch events will be enabled automatically, which can be suppressed by
> +by the option --no-switch-events.
>  
>  --clang-path=PATH::
>  Path to clang binary to use for compiling BPF scriptlets.
> diff --git a/tools/perf/arch/arm/util/cs-etm.c b/tools/perf/arch/arm/util/cs-etm.c
> index 607499b41bea..cea5e33d61d2 100644
> --- a/tools/perf/arch/arm/util/cs-etm.c
> +++ b/tools/perf/arch/arm/util/cs-etm.c
> @@ -265,7 +265,8 @@ static int cs_etm_recording_options(struct auxtrace_record *itr,
>  	ptr->evlist = evlist;
>  	ptr->snapshot_mode = opts->auxtrace_snapshot_mode;
>  
> -	if (perf_can_record_switch_events())
> +	if (!record_opts__no_switch_events(opts) &&
> +	    perf_can_record_switch_events())
>  		opts->record_switch_events = true;
>  
>  	evlist__for_each_entry(evlist, evsel) {
> diff --git a/tools/perf/arch/x86/util/intel-pt.c b/tools/perf/arch/x86/util/intel-pt.c
> index fdb917fcf511..3fc547769707 100644
> --- a/tools/perf/arch/x86/util/intel-pt.c
> +++ b/tools/perf/arch/x86/util/intel-pt.c
> @@ -779,7 +779,8 @@ static int intel_pt_recording_options(struct auxtrace_record *itr,
>  	 * Per-cpu recording needs sched_switch events to distinguish different
>  	 * threads.
>  	 */
> -	if (have_timing_info && !perf_cpu_map__empty(cpus)) {
> +	if (have_timing_info && !perf_cpu_map__empty(cpus) &&
> +	    !record_opts__no_switch_events(opts)) {
>  		if (perf_can_record_switch_events()) {
>  			bool cpu_wide = !target__none(&opts->target) &&
>  					!target__has_task(&opts->target);
> diff --git a/tools/perf/builtin-record.c b/tools/perf/builtin-record.c
> index 02d85f4e4517..dc6d5ba2fff8 100644
> --- a/tools/perf/builtin-record.c
> +++ b/tools/perf/builtin-record.c
> @@ -2490,8 +2490,9 @@ static struct option __record_options[] = {
>  		    "Record namespaces events"),
>  	OPT_BOOLEAN(0, "all-cgroups", &record.opts.record_cgroup,
>  		    "Record cgroup events"),
> -	OPT_BOOLEAN(0, "switch-events", &record.opts.record_switch_events,
> -		    "Record context switch events"),
> +	OPT_BOOLEAN_SET(0, "switch-events", &record.opts.record_switch_events,
> +			&record.opts.record_switch_events_set,
> +			"Record context switch events"),
>  	OPT_BOOLEAN_FLAG(0, "all-kernel", &record.opts.all_kernel,
>  			 "Configure all used events to run in kernel space.",
>  			 PARSE_OPT_EXCLUSIVE),
> diff --git a/tools/perf/util/record.h b/tools/perf/util/record.h
> index dce6332f5071..84dbbc3f0204 100644
> --- a/tools/perf/util/record.h
> +++ b/tools/perf/util/record.h
> @@ -36,6 +36,7 @@ struct record_opts {
>  	bool	      record_namespaces;
>  	bool	      record_cgroup;
>  	bool	      record_switch_events;
> +	bool	      record_switch_events_set;
>  	bool	      all_kernel;
>  	bool	      all_user;
>  	bool	      kernel_callchains;
> @@ -77,4 +78,9 @@ extern struct option *record_options;
>  
>  int record__parse_freq(const struct option *opt, const char *str, int unset);
>  
> +static inline bool record_opts__no_switch_events(const struct record_opts *opts)
> +{
> +	return opts->record_switch_events_set && !opts->record_switch_events;
> +}
> +
>  #endif // _PERF_RECORD_H
> -- 
> 2.17.1
> 

-- 

- Arnaldo