From mboxrd@z Thu Jan 1 00:00:00 1970 From: Zefan Li Subject: Re: [PATCH 2/3] sched, cgroup: replace signal_struct->group_rwsem with a global percpu_rwsem Date: Sun, 24 May 2015 10:35:48 +0800 Message-ID: <55613904.7060502@huawei.com> References: <1431549318-16756-1-git-send-email-tj@kernel.org> <1431549318-16756-3-git-send-email-tj@kernel.org> <20150519151659.GF3644@twins.programming.kicks-ass.net> <20150519155133.GM24861@htj.duckdns.org> <555C5C71.80200@huawei.com> <20150521203943.GS24861@htj.duckdns.org> Mime-Version: 1.0 Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <20150521203943.GS24861-piEFEHQLUPpN0TnZuCh8vA@public.gmane.org> Sender: cgroups-owner-u79uwXL29TY76Z2rM5mHXA@public.gmane.org List-ID: Content-Type: text/plain; charset="us-ascii" To: Tejun Heo Cc: Peter Zijlstra , cgroups-u79uwXL29TY76Z2rM5mHXA@public.gmane.org, mingo-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org, linux-kernel-u79uwXL29TY76Z2rM5mHXA@public.gmane.org On 2015/5/22 4:39, Tejun Heo wrote: > Hello, Li. > > On Wed, May 20, 2015 at 06:05:37PM +0800, Zefan Li wrote: >>> The latency is bound by synchronize_sched_expedited(). Given the way >>> cgroups are used in majority of setups (process migration happening >>> only during service / session setups), I think this should be okay. >> >> Actually process migration can happen quite frequently, for example in >> Android phones, and that's why Google had an out-of-tree patch to remove >> the synchronize_rcu() in that path, which turned out to be buggy. > > It's still not a very frequent operation tho. We're talking about > users switching fore/background jobs here and the expedited > synchronization w/ preemption enabled doesn't take much time. In > addition, as it currently stands, android is doing memory charge > immigration on each fore/background switches. I'm pretty doubtful > this would make any difference. > I did some testing with my laptop. Moving a task between 2 cgroups for 10W times with one or two threads: 1T 2T orig 3.36s 3.65s orig+tj 3.55s 6.31s orig+sync_rcu 16.69s 28.47s (only 1000 times) The overhead looks acceptable.