Re: [RFC PATCH 1/2] tools/perf: Don't set attr->disabled when doinggrouping

[David Ahern]

On 10/24/2011 04:56 AM, Deng-Cheng Zhu wrote:
> Events will be created at the state PERF_EVENT_STATE_OFF if attr->disabled
> is set. When these events go to arch level validate_group(), the function
> won't do anything real because they are filtered out by the state check.

Seems like a bug with the group checks. You should be able to enable and
disable events - including creating them disabled.

David


> 
> Signed-off-by: Deng-Cheng Zhu <dczhu@xxxxxxxx>
> Cc: Peter Zijlstra <a.p.zijlstra@xxxxxxxxx>
> Cc: Paul Mackerras <paulus@xxxxxxxxx>
> Cc: Ingo Molnar <mingo@xxxxxxx>
> Cc: Arnaldo Carvalho de Melo <acme@xxxxxxxxxxxxxxxxxx>
> ---
>  tools/perf/builtin-record.c |    6 +++---
>  tools/perf/builtin-stat.c   |    3 ++-
>  2 files changed, 5 insertions(+), 4 deletions(-)
> 
> diff --git a/tools/perf/builtin-record.c b/tools/perf/builtin-record.c
> index f4c3fbe..67e8e4c 100644
> --- a/tools/perf/builtin-record.c
> +++ b/tools/perf/builtin-record.c
> @@ -161,7 +161,6 @@ static void config_attr(struct perf_evsel *evsel, struct perf_evlist *evlist)
>  	struct perf_event_attr *attr = &evsel->attr;
>  	int track = !evsel->idx; /* only the first counter needs these */
>  
> -	attr->disabled		= 1;
>  	attr->inherit		= !no_inherit;
>  	attr->read_format	= PERF_FORMAT_TOTAL_TIME_ENABLED |
>  				  PERF_FORMAT_TOTAL_TIME_RUNNING |
> @@ -222,10 +221,11 @@ static void config_attr(struct perf_evsel *evsel, struct perf_evlist *evlist)
>  	attr->mmap		= track;
>  	attr->comm		= track;
>  
> -	if (target_pid == -1 && target_tid == -1 && !system_wide) {
> +	if (!group)
>  		attr->disabled = 1;
> +
> +	if (target_pid == -1 && target_tid == -1 && !system_wide)
>  		attr->enable_on_exec = 1;
> -	}
>  }
>  
>  static bool perf_evlist__equal(struct perf_evlist *evlist,
> diff --git a/tools/perf/builtin-stat.c b/tools/perf/builtin-stat.c
> index 5deb17d..ca95475 100644
> --- a/tools/perf/builtin-stat.c
> +++ b/tools/perf/builtin-stat.c
> @@ -284,7 +284,8 @@ static int create_perf_stat_counter(struct perf_evsel *evsel)
>  		return perf_evsel__open_per_cpu(evsel, evsel_list->cpus, group);
>  
>  	if (target_pid == -1 && target_tid == -1) {
> -		attr->disabled = 1;
> +		if (!group)
> +			attr->disabled = 1;
>  		attr->enable_on_exec = 1;
>  	}
>  
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/