linux-perf-users.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Madadi Vineeth Reddy <vineethr@linux.ibm.com>
To: yangjihong <yangjihong@bytedance.com>
Cc: peterz@infradead.org, mingo@redhat.com, acme@kernel.org,
	namhyung@kernel.org, mark.rutland@arm.com,
	alexander.shishkin@linux.intel.com, jolsa@kernel.org,
	irogers@google.com, adrian.hunter@intel.com,
	kan.liang@linux.intel.com, james.clark@arm.com,
	linux-perf-users@vger.kernel.org, linux-kernel@vger.kernel.org,
	Madadi Vineeth Reddy <vineethr@linux.ibm.com>
Subject: RE: [PATCH v2] perf sched timehist: Fixed timestamp error when unable to confirm event sched_in time
Date: Wed, 21 Aug 2024 15:58:15 +0530	[thread overview]
Message-ID: <ab337b83-ac74-43ce-83f0-25885c136574@linux.ibm.com> (raw)
In-Reply-To: <CAPePreJCahzH0LwfXN_X+UdxU54_UMRPZgDgGqt7+PC+ky997Q@mail.gmail.com>

Hi Yang Jihong,

On 21/08/24 11:46, yangjihong wrote:
> Hello,
> 
> On 2024/8/21 2:00, Madadi Vineeth Reddy wrote:
>> Hi Yang Jihong,
>>
>> On 19/08/24 08:17, Yang Jihong wrote:
>>> If sched_in event for current task is not recorded, sched_in timestamp
>>> will be set to end_time of time window interest, causing an error in
>>> timestamp show. In this case, we choose to ignore this event.
>>>
>>
>> Why sched-in is not happening for the perf task continuously
>>
>> perf sched record -a sleep 5
>>
>> perf sched script
>> perf 19593 [000] 42668.450179: sched:sched_stat_runtime: comm=perf
> pid=19593 runtime=69449 [ns] vruntime=7303687 [ns]
>> perf 19593 [000] 42668.450190: sched:sched_stat_runtime: comm=perf
> pid=19593 runtime=6039 [ns] vruntime=7309726 [ns]
>> perf 19593 [000] 42668.450191: sched:sched_waking: comm=migration/0
> pid=19 prio=0 target_cpu=000
>> perf 19593 [000] 42668.450192: sched:sched_stat_runtime: comm=perf
> pid=19593 runtime=2397 [ns] vruntime=7312123 [ns]
>> perf 19593 [000] 42668.450193: sched:sched_switch: perf:19593 [120] R ==>
> migration/0:19 [0]
>> migration/0 19 [000] 42668.450195: sched:sched_migrate_task: comm=perf
> pid=19593 prio=120 orig_cpu=0 dest_cpu=1
>> migration/0 19 [000] 42668.450200: sched:sched_switch: migration/0:19 [0]
> S ==> swapper/0:0 [120]
>> perf 19593 [001] 42668.450308: sched:sched_stat_runtime: comm=perf
> pid=19593 runtime=112515 [ns] vruntime=9972397 [ns]
>> perf 19593 [001] 42668.450311: sched:sched_waking: comm=migration/1
> pid=24 prio=0 target_cpu=001
>> perf 19593 [001] 42668.450313: sched:sched_stat_runtime: comm=perf
> pid=19593 runtime=4690 [ns] vruntime=9977087 [ns]
>> perf 19593 [001] 42668.450314: sched:sched_switch: perf:19593 [120] R ==>
> migration/1:24 [0]
>> migration/1 24 [001] 42668.450316: sched:sched_migrate_task: comm=perf
> pid=19593 prio=120 orig_cpu=1 dest_cpu=2
>> migration/1 24 [001] 42668.450322: sched:sched_switch: migration/1:24 [0]
> S ==> swapper/1:0 [120]
>> perf 19593 [002] 42668.450385: sched:sched_stat_runtime: comm=perf
> pid=19593 runtime=67379 [ns] vruntime=2885994 [ns]
>> perf 19593 [002] 42668.450389: sched:sched_waking: comm=migration/2
> pid=30 prio=0 target_cpu=002
>> perf 19593 [002] 42668.450391: sched:sched_stat_runtime: comm=perf
> pid=19593 runtime=5526 [ns] vruntime=2891520 [ns]
>> perf 19593 [002] 42668.450392: sched:sched_switch: perf:19593 [120] R ==>
> migration/2:30 [0]
>>
>> The perf task (PID 19593) is being switched out without being switched in
> after CPU migration. The task migrates to a different CPU,
>> and then immediately switches out without a sched_in event in between.
>>
> Because the perf tool sets CPU affinity first (which causes perf process
> to migrate to the target core) and then enables events, by the time
> starts recording, task is likely to have migrated to the target core and
> started running. In this case, the sched_in event is not recorded.
> 

Got it. Thanks for the explanation!

Thanks and Regards
Madadi Vineeth Reddy

> Thanks,
> Yang
> 


      parent reply	other threads:[~2024-08-21 10:28 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2024-08-19  2:47 [PATCH v2] perf sched timehist: Fixed timestamp error when unable to confirm event sched_in time Yang Jihong
2024-08-19 20:08 ` Namhyung Kim
2024-08-30 13:32   ` Arnaldo Carvalho de Melo
2024-08-20 18:00 ` Madadi Vineeth Reddy
     [not found]   ` <CAPePreJCahzH0LwfXN_X+UdxU54_UMRPZgDgGqt7+PC+ky997Q@mail.gmail.com>
2024-08-21 10:28     ` Madadi Vineeth Reddy [this message]

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=ab337b83-ac74-43ce-83f0-25885c136574@linux.ibm.com \
    --to=vineethr@linux.ibm.com \
    --cc=acme@kernel.org \
    --cc=adrian.hunter@intel.com \
    --cc=alexander.shishkin@linux.intel.com \
    --cc=irogers@google.com \
    --cc=james.clark@arm.com \
    --cc=jolsa@kernel.org \
    --cc=kan.liang@linux.intel.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-perf-users@vger.kernel.org \
    --cc=mark.rutland@arm.com \
    --cc=mingo@redhat.com \
    --cc=namhyung@kernel.org \
    --cc=peterz@infradead.org \
    --cc=yangjihong@bytedance.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).