Re: [PATCH v3 8/9] perf record: implement control commands handling

From: Alexey Budankov <alexey.budankov@linux.intel.com>
To: Jiri Olsa <jolsa@redhat.com>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>,
	Namhyung Kim <namhyung@kernel.org>,
	Alexander Shishkin <alexander.shishkin@linux.intel.com>,
	Peter Zijlstra <peterz@infradead.org>,
	Ingo Molnar <mingo@redhat.com>, Andi Kleen <ak@linux.intel.com>,
	linux-kernel <linux-kernel@vger.kernel.org>
Subject: Re: [PATCH v3 8/9] perf record: implement control commands handling
Date: Wed, 20 May 2020 21:06:49 +0300	[thread overview]
Message-ID: <df2cd76d-a9ae-7092-1ee2-83fee8f557c8@linux.intel.com> (raw)
In-Reply-To: <20200520123833.GG157452@krava>

On 20.05.2020 15:38, Jiri Olsa wrote:
> On Wed, May 13, 2020 at 11:04:25AM +0300, Alexey Budankov wrote:
>>
>> Implement handling of 'enable' and 'disable' control commands
>> coming from control file descriptor.
>>
>> Signed-off-by: Alexey Budankov <alexey.budankov@linux.intel.com>
>> ---
>>  tools/perf/builtin-record.c | 18 +++++++++++++++++-
>>  1 file changed, 17 insertions(+), 1 deletion(-)
>>
>> diff --git a/tools/perf/builtin-record.c b/tools/perf/builtin-record.c
>> index 50dc2fe626e5..72f388623364 100644
>> --- a/tools/perf/builtin-record.c
>> +++ b/tools/perf/builtin-record.c
>> @@ -1505,6 +1505,7 @@ static int __cmd_record(struct record *rec, int argc, const char **argv)
>>  	bool disabled = false, draining = false;
>>  	int fd;
>>  	float ratio = 0;
>> +	enum evlist_ctl_cmd cmd = EVLIST_CTL_CMD_UNSUPPORTED;
>>  
>>  	atexit(record__sig_exit);
>>  	signal(SIGCHLD, sig_handler);
>> @@ -1802,8 +1803,23 @@ static int __cmd_record(struct record *rec, int argc, const char **argv)
>>  			 * Propagate error, only if there's any. Ignore positive
>>  			 * number of returned events and interrupt error.
>>  			 */
>> -			if (err > 0 || (err < 0 && errno == EINTR))
>> +			if (err > 0 || (err < 0 && errno == EINTR)) {
>>  				err = 0;
>> +				if (evlist__ctlfd_process(rec->evlist, &cmd) > 0) {
>> +					switch (cmd) {
>> +					case EVLIST_CTL_CMD_ENABLE:
>> +						pr_info(EVLIST_ENABLED_MSG);
>> +						break;
>> +					case EVLIST_CTL_CMD_DISABLE:
>> +						pr_info(EVLIST_DISABLED_MSG);
>> +						break;
>> +					case EVLIST_CTL_CMD_ACK:
>> +					case EVLIST_CTL_CMD_UNSUPPORTED:
>> +					default:
>> +						break;
>> +					}
>> +				}
>> +			}
> 
> so this will be processed only when:
> 
> 	if (hits == rec->samples) {
> 
> what if there's always somethign in the buffer? will this stall?
> 
>>  			waking++;
>>  
>>  			if (evlist__filter_pollfd(rec->evlist, POLLERR | POLLHUP) == 0)
> 
> evlist__filter_pollfd will trigger draining if all the maps are closed,
> but there's one more fd in pollfd now, will this prevent draining?

Good catch. This needs redesign to notice pending commands during 
active trace streaming. The same is regarding draining.

> 
> I wonder this would fit better to the sideband thread (util/sideband_evlist.c)
> so we don't disturb the main thread with another check

~Alexey