linux-kernel - [PATCHSET wq/for-3.10] workqueue: simplify per-cpu worker rebinding and implement unbound worker CPU affinity restoration

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite for Android: free password hash cracker in your pocket

[<prev] [next>] [thread-next>] [day] [month] [year] [list]

Message-Id: <1363302090-16117-1-git-send-email-tj@kernel.org>
Date:	Thu, 14 Mar 2013 16:01:25 -0700
From:	Tejun Heo <tj@...nel.org>
To:	linux-kernel@...r.kernel.org
Cc:	laijs@...fujitsu.com
Subject: [PATCHSET wq/for-3.10] workqueue: simplify per-cpu worker rebinding and implement unbound worker CPU affinity restoration

Hello,

per-cpu worker rebinding on CPU_ONLINE has always been complex.  A lot
of the complexity comes from the fact that CPU_ONLINE should ask each
worker, which may be in any state, to rebind itself.  With recent
patch to replace PF_THREAD_BOUND with PF_NO_SETAFFINITY (the patch is
pending and included here for review), CPU_ONLINE can directly invoke
set_cpus_allowed_ptr() on the workers.

This patchset simplifies rebinding of per-cpu workers by making
CPU_ONLINE directly call set_cpus_allowed_ptr() on the workers and
implements CPU affinity restoration of unbound workers.  The latter is
necessary because unbound pools can now have custom cpumask and if all
CPUs in the allowed cpumask go down and then later come up, workers
may lose affinity.  This patchset makes workqueue CPU_ONLINE callback
restore CPU affinity when the first allowed CPU comes up for a pool.

This patchset contains the following five patches.

 0001-sched-replace-PF_THREAD_BOUND-with-PF_NO_SETAFFINITY.patch
 0002-workqueue-convert-worker_pool-worker_ida-to-idr-and-.patch
 0003-workqueue-relocate-rebind_workers.patch
 0004-workqueue-directly-restore-CPU-affinity-of-workers-f.patch
 0005-workqueue-restore-CPU-affinity-of-unbound-workers-on.patch

0001 doesn't belong in this series.  It was posted separately[1] and
should be routed through tip.

  wq/for-3.10 e626761691 ("workqueue: implement current_is_workqueue_rescuer()")
+ "workqueue: misc cleanups" patchset [2]
+ "workqueue: workqueue: break up workqueue_lock into multiple locks [3]

and available in the following git branch.

 git://git.kernel.org/pub/scm/linux/kernel/git/tj/wq.git review-restore-affinity

diffstat follows.

 include/linux/sched.h       |    2
 kernel/cgroup.c             |    4
 kernel/cpuset.c             |   16 +-
 kernel/kthread.c            |    2
 kernel/sched/core.c         |    9 -
 kernel/workqueue.c          |  351 ++++++++++++++++++++++----------------------
 kernel/workqueue_internal.h |    3
 7 files changed, 198 insertions(+), 189 deletions(-)

Thanks.

--
tejun

[1] http://article.gmane.org/gmane.linux.kernel/1457229
[2] http://thread.gmane.org/gmane.linux.kernel/1456511
[3] http://thread.gmane.org/gmane.linux.kernel/1456555
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/