From: Oleg Nesterov <oleg@redhat.com>
To: Andrew Morton <akpm@linux-foundation.org>
Cc: Aaron Tomlin <atomlin@redhat.com>,
"Eric W. Biederman" <ebiederm@xmission.com>,
Sterling Alexander <stalexan@redhat.com>,
linux-kernel@vger.kernel.org
Subject: Re: [PATCH -mm 1/3] exit: reparent: avoid find_new_reaper() if no children
Date: Fri, 21 Nov 2014 21:01:38 +0100 [thread overview]
Message-ID: <20141121200138.GA21656@redhat.com> (raw)
In-Reply-To: <20141120143722.af15074e6922108962e84649@linux-foundation.org>
On 11/20, Andrew Morton wrote:
>
> On Thu, 20 Nov 2014 19:34:23 +0100 Oleg Nesterov <oleg@redhat.com> wrote:
>
> > $ time ./test 16 16536 shows:
> >
> > real user sys
> > - 5m37.628s 0m4.437s 8m5.560s
> > + 0m50.032s 0m7.130s 1m4.927s
>
> Is that the best you can do?
Unfortunately these changes do not even try to solve the main problem,
tasklist_lock doesn't scale simply because it is global. These changes
make sense (I hope) anyway, even if/when we redesign the locking. But
so far I do not have a good plan.
> (I assume the increase in user time was a glitch?)
To be honest, I didn't even notice this change. I repeated the testing
before/after this patch and (to my surprize) the "user" numbers are more
or less stable, and /usr/bin/time reports the increase.
1. First of all: this is impossible ;)
Note that this test-case uses SIGTRAP to trigger the coredumping.
This means that exit_notify() can only be called when all threads
are already in kernel mode, the coredumping thread sleeps until
they all are parked in exit_mm(). Until then this patch has no
effect.
2. With this patch applied, I added mdelay(2) into forget_original_parent(),
right after find_child_reaper(). And yes, this changes the numbers too:
real user sys
10m1.225s 0m5.443s 17m25.797s
note that "user time" goes down.
3. So I think that this just reminds that utime/stime accounting isn't
precise. sum_exec_runtime is accurate and thus we can more or less
trust utime + stime, but utime/stime is random. Plus scale_stime()
doesn't look very accurate too.
4. In this particular case the accounting is even more impresize, this
test-case spends a lot of time in kernel mode with irqs disabled and
this "freezes" task->stime.
5. That said, I still can't really understand why "user" grows. If I
understand the calculations in cputime_adjust() correctly (probably
I don't), it should not.
In short, I am a bit confused but I still don't think that this increase
is real.
Oleg.
next prev parent reply other threads:[~2014-11-21 20:01 UTC|newest]
Thread overview: 55+ messages / expand[flat|nested] mbox.gz Atom feed top
2014-11-07 20:14 [PATCH 0/4] proc: deuglify task_state() Oleg Nesterov
2014-11-07 20:14 ` [PATCH 1/4] proc: task_state: read cred->group_info outside of task_lock() Oleg Nesterov
2014-11-07 20:14 ` [PATCH 2/4] proc: task_state: deuglify the max_fds calculation Oleg Nesterov
2014-11-07 20:14 ` [PATCH 3/4] proc: task_state: move the main seq_printf() outside of rcu_read_lock() Oleg Nesterov
2014-11-13 18:04 ` Paul E. McKenney
2014-11-07 20:14 ` [PATCH 4/4] proc: task_state: ptrace_parent() doesn't need pid_alive() check Oleg Nesterov
2014-11-10 21:59 ` [PATCH 0/5] exit: reparent zombie fix + cleanups/optimizations Oleg Nesterov
2014-11-10 22:00 ` [PATCH 1/5] sched_show_task: fix unsafe usage of ->real_parent Oleg Nesterov
2014-11-11 10:39 ` Peter Zijlstra
2014-11-10 22:00 ` [PATCH 2/5] exit: reparent: use ->ptrace_entry rather than ->sibling for EXIT_DEAD tasks Oleg Nesterov
2014-11-10 22:00 ` [PATCH 3/5] exit: reparent: cleanup the changing of ->parent Oleg Nesterov
2014-11-10 22:00 ` [PATCH 4/5] exit: reparent: cleanup the usage of reparent_leader() Oleg Nesterov
2014-11-10 22:00 ` [PATCH 5/5] exit: ptrace: shift "reap dead" code from exit_ptrace() to forget_original_parent() Oleg Nesterov
2014-11-14 1:37 ` [PATCH 0/5] exit: more cleanups/optimizations Oleg Nesterov
2014-11-14 1:38 ` [PATCH 1/5] exit: wait: cleanup the ptrace_reparented() checks Oleg Nesterov
2014-11-14 1:38 ` [PATCH 2/5] exit: wait: don't use zombie->real_parent Oleg Nesterov
2014-11-14 1:38 ` [PATCH 3/5] exit: wait: drop tasklist_lock before psig->c* accounting Oleg Nesterov
2014-11-14 1:38 ` [PATCH 4/5] exit: release_task: fix the comment about group leader accounting Oleg Nesterov
2014-11-14 1:38 ` [PATCH 5/5] exit: proc: don't try to flush /proc/tgid/task/tgid Oleg Nesterov
2014-11-18 21:29 ` [PATCH 0/6] exit: find_new_reaper() fixes/cleanups Oleg Nesterov
2014-11-18 21:30 ` [PATCH 1/6] exit: reparent: fix the dead-parent PR_SET_CHILD_SUBREAPER reparenting Oleg Nesterov
2014-11-18 21:30 ` [PATCH 2/6] exit: reparent: fix the cross-namespace " Oleg Nesterov
2014-11-18 21:30 ` [PATCH 3/6] exit: reparent: s/while_each_thread/for_each_thread/ in find_new_reaper() Oleg Nesterov
2014-11-18 21:30 ` [PATCH 4/6] exit: reparent: document the ->has_child_subreaper checks Oleg Nesterov
2014-11-18 21:30 ` [PATCH 5/6] exit: reparent: introduce find_child_reaper() Oleg Nesterov
2014-11-18 21:30 ` [PATCH 6/6] exit: reparent: introduce find_alive_thread() Oleg Nesterov
2014-11-20 18:34 ` [PATCH 0/3] exit: avoid O(n ** 2) thread-list scan on group-exit if possible Oleg Nesterov
2014-11-20 18:34 ` [PATCH -mm 1/3] exit: reparent: avoid find_new_reaper() if no children Oleg Nesterov
2014-11-20 22:37 ` Andrew Morton
2014-11-21 20:01 ` Oleg Nesterov [this message]
2014-11-20 18:34 ` [PATCH -mm 2/3] exit: reparent: call forget_original_parent() under tasklist_lock Oleg Nesterov
2014-11-20 18:34 ` [PATCH -mm 3/3] exit: exit_notify: re-use "dead" list to autoreap current Oleg Nesterov
2014-11-24 20:06 ` [PATCH 0/2] exit/pid_ns: comments + simple fix Oleg Nesterov
2014-11-24 20:06 ` [PATCH 1/2] exit: reparent: fix/update the comments in zap_pid_ns_processes() Oleg Nesterov
2014-11-24 20:14 ` Oleg Nesterov
2014-11-24 22:07 ` Eric W. Biederman
2014-11-25 16:57 ` Oleg Nesterov
2014-11-25 17:17 ` Oleg Nesterov
2014-11-24 20:06 ` [PATCH 2/2] exit: pidns: alloc_pid() leaks pid_namespace if child_reaper is exiting Oleg Nesterov
2014-11-24 21:46 ` Eric W. Biederman
2014-11-25 17:07 ` Oleg Nesterov
2014-11-25 17:50 ` Eric W. Biederman
2014-11-25 18:15 ` Oleg Nesterov
2014-11-25 18:43 ` Eric W. Biederman
2014-11-25 18:59 ` Oleg Nesterov
2014-11-24 21:27 ` [PATCH 0/2] exit/pid_ns: comments + simple fix Eric W. Biederman
2014-11-24 21:38 ` Oleg Nesterov
2014-11-24 21:48 ` Eric W. Biederman
2014-11-25 16:57 ` Oleg Nesterov
2014-11-26 23:54 ` [PATCH v2 " Oleg Nesterov
2014-11-26 23:54 ` [PATCH v2 1/2] exit: pidns: alloc_pid() leaks pid_namespace if child_reaper is exiting Oleg Nesterov
2014-11-27 15:44 ` Eric W. Biederman
2014-11-26 23:54 ` [PATCH v2 2/2] exit: pidns: fix/update the comments in zap_pid_ns_processes() Oleg Nesterov
2014-12-01 22:39 ` Andrew Morton
2014-12-01 23:24 ` Oleg Nesterov
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=20141121200138.GA21656@redhat.com \
--to=oleg@redhat.com \
--cc=akpm@linux-foundation.org \
--cc=atomlin@redhat.com \
--cc=ebiederm@xmission.com \
--cc=linux-kernel@vger.kernel.org \
--cc=stalexan@redhat.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).