Re: [PATCH 27/42] perf tools: Protect dso symbol loading using a mutex

From: Namhyung Kim
Date: Thu Jan 29 2015 - 07:48:58 EST


Hi Arnaldo,

On Thu, Jan 29, 2015 at 9:34 PM, Arnaldo Carvalho de Melo
<acme@xxxxxxxxxx> wrote:
> Em Thu, Jan 29, 2015 at 05:07:08PM +0900, Namhyung Kim escreveu:
>> When multi-thread support for perf report is enabled, it's possible to
>> access a dso concurrently. Add a new pthread_mutex to protect it from
>> concurrent dso__load().
>>
>> Signed-off-by: Namhyung Kim <namhyung@xxxxxxxxxx>
>> ---
>> tools/perf/util/dso.c | 2 ++
>> tools/perf/util/dso.h | 1 +
>> tools/perf/util/symbol.c | 34 ++++++++++++++++++++++++----------
>> 3 files changed, 27 insertions(+), 10 deletions(-)
>>
>> diff --git a/tools/perf/util/dso.c b/tools/perf/util/dso.c
>> index 45be944d450a..3da75816b8f8 100644
>> --- a/tools/perf/util/dso.c
>> +++ b/tools/perf/util/dso.c
>> @@ -888,6 +888,7 @@ struct dso *dso__new(const char *name)
>> RB_CLEAR_NODE(&dso->rb_node);
>> INIT_LIST_HEAD(&dso->node);
>> INIT_LIST_HEAD(&dso->data.open_entry);
>> + pthread_mutex_init(&dso->lock, NULL);
>> }
>>
>> return dso;
>> @@ -917,6 +918,7 @@ void dso__delete(struct dso *dso)
>> dso_cache__free(&dso->data.cache);
>> dso__free_a2l(dso);
>> zfree(&dso->symsrc_filename);
>> + pthread_mutex_destroy(&dso->lock);
>> free(dso);
>> }
>>
>> diff --git a/tools/perf/util/dso.h b/tools/perf/util/dso.h
>> index 3782c82c6e44..ac753594a469 100644
>> --- a/tools/perf/util/dso.h
>> +++ b/tools/perf/util/dso.h
>> @@ -102,6 +102,7 @@ struct dsos {
>> };
>>
>> struct dso {
>> + pthread_mutex_t lock;
>> struct list_head node;
>> struct rb_node rb_node; /* rbtree node sorted by long name */
>> struct rb_root symbols[MAP__NR_TYPES];
>> diff --git a/tools/perf/util/symbol.c b/tools/perf/util/symbol.c
>> index a69066865a55..714e20c99354 100644
>> --- a/tools/perf/util/symbol.c
>> +++ b/tools/perf/util/symbol.c
>> @@ -1357,12 +1357,22 @@ int dso__load(struct dso *dso, struct map *map, symbol_filter_t filter)
>> struct symsrc *syms_ss = NULL, *runtime_ss = NULL;
>> bool kmod;
>>
>> - dso__set_loaded(dso, map->type);
>> + pthread_mutex_lock(&dso->lock);
>> +
>> + /* check again under the dso->lock */
>
> Again? Where was it first checked?

Please see map__load().


> Perhaps we should lock there, so that
> we don't have to do two checks, one unlocked, the other locked?

Hmm.. maybe. I just keep it to avoid locking overhead since it'll be
called whenever it searches symbols during preprocessing. I didn't
measure the overhead but it could be huge IMHO.

Thanks,
Namhyung


>
>> + if (dso__loaded(dso, map->type)) {
>> + ret = 1;
>> + goto out;
>> + }
>> +
>
> - Arnaldo
--
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/