From: Dongsheng Yang <yangds.fnst@cn.fujitsu.com>
To: <rostedt@goodmis.org>, <fweisbec@gmail.com>, <mingo@redhat.com>,
<peterz@infradead.org>, <acme@ghostprotocols.net>
Cc: <linux-kernel@vger.kernel.org>,
Dongsheng Yang <yangds.fnst@cn.fujitsu.com>
Subject: [PATCH 0/8] perf sched: Add trace event for sched wait.
Date: Tue, 15 Apr 2014 21:32:49 +0900 [thread overview]
Message-ID: <cover.1397562542.git.yangds.fnst@cn.fujitsu.com> (raw)
Hi all:
This is to solve the bug message shown in perf sched latency.
# perf sched latency|tail
ksoftirqd/0:3 | 0.597 ms | 57 | avg: 0.004 ms | max: 0.054 ms | max at: 19681.546204 s
ksoftirqd/1:14 | 0.637 ms | 58 | avg: 0.004 ms | max: 0.066 ms | max at: 19674.687734 s
irqbalance:349 | 0.429 ms | 1 | avg: 0.004 ms | max: 0.004 ms | max at: 19675.791528 s
ksoftirqd/3:24 | 0.527 ms | 67 | avg: 0.003 ms | max: 0.011 ms | max at: 19673.285019 s
migration/3:23 | 0.000 ms | 1 | avg: 0.002 ms | max: 0.002 ms | max at: 19672.055354 s
-----------------------------------------------------------------------------------------
TOTAL: | 4384.616 ms | 36879 |
---------------------------------------------------
INFO: 0.030% state machine bugs (11 out of 36684)
After some investigation, there are two reasons cause this problem.
(1). Sometimes, scheduler will wake up a running task, it is not necessary,
then I skip the wakeup if task->state is TASK_RUNNING. [4/8]
(2). No tracing for sched wait.
This is a simple graph for task state changing.
---------------- 1 ----------------
| TASK_RUNNING | ------------------------------>| TASK_RUNNING |
| (running) |<------------------------------ | (wait cpu) |
---------------- 2 ----------------
^ |
|4 ------------------------- 3 |
|-------|TASK_{UN}INTERRUPTABLE |<--------------|
| in wait_rq |
-------------------------
As the graph shown above, there are four event in scheduling, and
we currently are tracing 3 of them.
1 & 2: sched:sched_switch
4: sched:sched_wakeup|sched:sched_wakeup_new
But about 3, we have no trace event for it.
This patchset add a new trace event for sched wait, and add a trace point
before any task added into wait queue. [1/8] & [3/8]
BTW, other patchs in this thread are about some little fix and enhancement
in the development, please help to review at the same time.
Thanx
Dongsheng (8):
sched & trace: Add a trace event for wait.
sched/wait: Add trace point before add task into wait queue.
sched/wait: Use __add_wait_queue{_tail}_exclusive() as possible.
sched/core: Skip wakeup when task is already running.
perf tools: record and process sched:sched_wait event.
perf tools: add missing event for perf sched record.
perf tools: Adapt the TASK_STATE_TO_CHAR_STR to new value in kernel
space.
perf tools: Clarify the output of perf sched map.
include/trace/events/sched.h | 20 +++++++++
kernel/sched/core.c | 3 ++
kernel/sched/wait.c | 13 ++++--
tools/perf/builtin-sched.c | 100 +++++++++++++++++++++++++++++++++++--------
4 files changed, 115 insertions(+), 21 deletions(-)
--
1.8.2.1
next reply other threads:[~2014-04-15 13:32 UTC|newest]
Thread overview: 42+ messages / expand[flat|nested] mbox.gz Atom feed top
2014-04-15 12:32 Dongsheng Yang [this message]
2014-04-15 12:32 ` [PATCH 1/8] sched & trace: Add a trace event for wait Dongsheng Yang
2014-04-15 13:49 ` Peter Zijlstra
2014-04-16 14:23 ` Steven Rostedt
2014-04-15 12:32 ` [PATCH 2/8] sched/wait: Add trace point before add task into wait queue Dongsheng Yang
2014-04-15 12:32 ` [PATCH 3/8] sched/wait: Use __add_wait_queue{_tail}_exclusive() as possible Dongsheng Yang
2014-04-15 13:49 ` Peter Zijlstra
2014-04-16 9:51 ` Dongsheng Yang
2014-04-15 12:32 ` [PATCH 4/8] sched/core: Skip wakeup when task is already running Dongsheng Yang
2014-04-15 13:53 ` Peter Zijlstra
2014-04-16 10:22 ` Dongsheng Yang
2014-04-22 11:56 ` Dongsheng Yang
2014-04-22 13:23 ` Peter Zijlstra
2014-04-22 17:10 ` bsegall
2014-04-22 17:53 ` Steven Rostedt
2014-04-22 18:18 ` Peter Zijlstra
2014-05-05 6:32 ` Dongsheng Yang
2014-05-05 6:34 ` [PATCH] sched: Move the wakeup tracepoint from ttwu_do_wakeup() to ttwu_activate() Dongsheng Yang
2014-05-05 14:00 ` Steven Rostedt
2014-05-06 0:19 ` Dongsheng Yang
2014-05-06 0:26 ` Dongsheng Yang
2014-05-06 2:06 ` Steven Rostedt
2014-05-06 1:29 ` Dongsheng Yang
2014-05-06 1:52 ` [PATCH] sched: Distinguish sched_wakeup event when wake up a task which did schedule out or not Dongsheng Yang
2014-05-09 0:16 ` Dongsheng Yang
2014-05-09 1:27 ` Steven Rostedt
2014-05-10 15:29 ` Peter Zijlstra
[not found] ` <536F90BE.2080806@gmail.com>
2014-05-11 15:24 ` Fwd: " Dongsheng Yang
2014-05-11 16:35 ` Peter Zijlstra
2014-05-11 18:52 ` Steven Rostedt
2014-05-12 6:47 ` Peter Zijlstra
2014-05-12 8:58 ` Dongsheng Yang
2014-05-12 14:09 ` Steven Rostedt
2014-05-12 15:09 ` Peter Zijlstra
2014-05-12 15:17 ` Steven Rostedt
2014-05-12 15:28 ` Peter Zijlstra
2014-04-15 12:32 ` [PATCH 5/8] perf tools: record and process sched:sched_wait event Dongsheng Yang
2014-04-15 12:32 ` [PATCH 6/8] perf tools: add missing event for perf sched record Dongsheng Yang
2014-04-15 12:32 ` [PATCH 7/8] perf tools: Adapt the TASK_STATE_TO_CHAR_STR to new value in kernel space Dongsheng Yang
2014-04-15 12:32 ` [PATCH 8/8] perf tools: Clarify the output of perf sched map Dongsheng Yang
2014-04-15 13:54 ` [PATCH 0/8] perf sched: Add trace event for sched wait Peter Zijlstra
2014-04-16 10:28 ` Dongsheng Yang
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=cover.1397562542.git.yangds.fnst@cn.fujitsu.com \
--to=yangds.fnst@cn.fujitsu.com \
--cc=acme@ghostprotocols.net \
--cc=fweisbec@gmail.com \
--cc=linux-kernel@vger.kernel.org \
--cc=mingo@redhat.com \
--cc=peterz@infradead.org \
--cc=rostedt@goodmis.org \
/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