[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <875zkdgf72.fsf@ashishki-desk.ger.corp.intel.com>
Date: Fri, 25 Oct 2019 15:52:33 +0300
From: Alexander Shishkin <alexander.shishkin@...ux.intel.com>
To: Peter Zijlstra <peterz@...radead.org>
Cc: Arnaldo Carvalho de Melo <acme@...hat.com>,
Ingo Molnar <mingo@...hat.com>, linux-kernel@...r.kernel.org,
jolsa@...hat.com, adrian.hunter@...el.com,
mathieu.poirier@...aro.org, mark.rutland@....com,
alexander.shishkin@...ux.intel.com
Subject: Re: [PATCH v2 1/4] perf: Allow using AUX data in perf samples
Peter Zijlstra <peterz@...radead.org> writes:
> On Tue, Oct 22, 2019 at 12:58:09PM +0300, Alexander Shishkin wrote:
>> @@ -11213,6 +11367,9 @@ SYSCALL_DEFINE5(perf_event_open,
>> if (event->attr.aux_output && !perf_get_aux_event(event, group_leader))
>> goto err_locked;
>>
>> + if (event->attr.aux_sample_size && !perf_get_aux_event(event, group_leader))
>> + goto err_locked;
>> +
>
> Either aux_sample_size and aux_output are mutually exclusive, or you're
> leaking a refcount on group_leader. The first wants a check, the second
> wants error path fixes.
Now you mention it, it should be possible to have both. Whether that's
useful is another question. But we only need one reference either way.
Thanks,
--
Alex
Powered by blists - more mailing lists