From: Frederic Weisbecker <frederic@kernel.org>
To: Marcelo Tosatti <mtosatti@redhat.com>
Cc: linux-kernel@vger.kernel.org, Christoph Lameter <cl@linux.com>,
Thomas Gleixner <tglx@linutronix.de>,
Juri Lelli <juri.lelli@redhat.com>, Nitesh Lal <nilal@redhat.com>
Subject: Re: [patch 0/5] optionally sync per-CPU vmstats counter on return to userspace
Date: Fri, 2 Jul 2021 14:30:32 +0200 [thread overview]
Message-ID: <20210702123032.GA72061@lothringen> (raw)
In-Reply-To: <20210701210336.358118649@fuller.cnet>
On Thu, Jul 01, 2021 at 06:03:36PM -0300, Marcelo Tosatti wrote:
> The logic to disable vmstat worker thread, when entering
> nohz full, does not cover all scenarios. For example, it is possible
> for the following to happen:
>
> 1) enter nohz_full, which calls refresh_cpu_vm_stats, syncing the stats.
> 2) app runs mlock, which increases counters for mlock'ed pages.
> 3) start -RT loop
>
> Since refresh_cpu_vm_stats from nohz_full logic can happen _before_
> the mlock, vmstat shepherd can restart vmstat worker thread on
> the CPU in question.
>
> To fix this, optionally sync the vmstat counters when returning
> from userspace, controllable by a new "vmstat_sync" isolcpus
> flags (default off).
Wasn't the plan for such finegrained isolation features to do it at
the per task level using prctl()?
Thanks.
>
> See individual patches for details.
>
>
next prev parent reply other threads:[~2021-07-02 12:30 UTC|newest]
Thread overview: 27+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-07-01 21:03 [patch 0/5] optionally sync per-CPU vmstats counter on return to userspace Marcelo Tosatti
2021-07-01 21:03 ` [patch 1/5] sched: isolation: introduce vmstat_sync isolcpu flags Marcelo Tosatti
2021-07-01 21:03 ` [patch 2/5] common entry: add hook for isolation to __syscall_exit_to_user_mode_work Marcelo Tosatti
2021-07-01 21:03 ` [patch 3/5] mm: vmstat: optionally flush per-CPU vmstat counters on return to userspace Marcelo Tosatti
2021-07-01 23:11 ` kernel test robot
2021-07-01 23:11 ` kernel test robot
2021-07-02 6:50 ` kernel test robot
2021-07-02 6:50 ` kernel test robot
2021-07-01 21:03 ` [patch 4/5] mm: vmstat: move need_update Marcelo Tosatti
2021-07-01 21:03 ` [patch 5/5] mm: vmstat_refresh: avoid queueing work item if cpu stats are clean Marcelo Tosatti
2021-07-02 4:10 ` kernel test robot
2021-07-02 4:10 ` kernel test robot
2021-07-02 4:43 ` kernel test robot
2021-07-02 4:43 ` kernel test robot
2021-07-02 8:00 ` [patch 0/5] optionally sync per-CPU vmstats counter on return to userspace Christoph Lameter
2021-07-02 11:52 ` Marcelo Tosatti
2021-07-02 11:59 ` Marcelo Tosatti
2021-07-05 14:26 ` Christoph Lameter
2021-07-05 14:45 ` Marcelo Tosatti
2021-07-02 12:30 ` Frederic Weisbecker [this message]
2021-07-02 15:28 ` Marcelo Tosatti
2021-07-06 13:09 ` Frederic Weisbecker
2021-07-06 14:05 ` Marcelo Tosatti
2021-07-06 14:09 ` Marcelo Tosatti
2021-07-06 14:17 ` Marcelo Tosatti
2021-07-06 16:15 ` Peter Zijlstra
2021-07-06 16:53 ` Marcelo Tosatti
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=20210702123032.GA72061@lothringen \
--to=frederic@kernel.org \
--cc=cl@linux.com \
--cc=juri.lelli@redhat.com \
--cc=linux-kernel@vger.kernel.org \
--cc=mtosatti@redhat.com \
--cc=nilal@redhat.com \
--cc=tglx@linutronix.de \
/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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.