From mboxrd@z Thu Jan 1 00:00:00 1970 From: Li Zefan Subject: Re: [PATCHSET] perf, cgroup: implement hierarchy support for perf_event controller Date: Tue, 9 Apr 2013 15:07:56 +0800 Message-ID: <5163BE4C.5040301@huawei.com> References: <1365474213-13354-1-git-send-email-tj@kernel.org> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <1365474213-13354-1-git-send-email-tj-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org> List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: containers-bounces-cunTk1MwBs9QetFLy7KEm3xJsTq8ys+cHZ5vskTnxNA@public.gmane.org Errors-To: containers-bounces-cunTk1MwBs9QetFLy7KEm3xJsTq8ys+cHZ5vskTnxNA@public.gmane.org To: Tejun Heo Cc: namhyung.kim-Hm3cg6mZ9cc@public.gmane.org, a.p.zijlstra-/NLkJaSkS4VmR6Xm/wNWPw@public.gmane.org, containers-cunTk1MwBs9QetFLy7KEm3xJsTq8ys+cHZ5vskTnxNA@public.gmane.org, linux-kernel-u79uwXL29TY76Z2rM5mHXA@public.gmane.org, eranian-hpIqsD4AKlfQT0dZR+AlfA@public.gmane.org, mhocko-AlSwsSmVLrQ@public.gmane.org, mingo-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org, paulus-eUNUBHrolfbYtjvyW6yDsg@public.gmane.org, acme-f8uhVLnGfZaxAyOMLChx1axOck334EZe@public.gmane.org, cgroups-u79uwXL29TY76Z2rM5mHXA@public.gmane.org List-Id: containers.vger.kernel.org On 2013/4/9 10:23, Tejun Heo wrote: > perf_event cgroup controller is one of the remaining few with broken > hierarchy support. It turns out it's pretty easy to implement - the > only thing necessary is making perf_cgroup_match() return %true also > when the cgroup of the current task is a descendant of the event's > cgroup. This patchset implements cgroup_is_descendant() and uses it > to implement hierarchy support in perf_event controller. > > This patchset contains the following three patches. > > 0001-cgroup-make-sure-parent-won-t-be-destroyed-before-it.patch > 0002-cgroup-implement-cgroup_is_descendant.patch > 0003-perf-make-perf_event-cgroup-hierarchical.patch > > The patches are also available in the following git branch, which is > based on top of cgroup/for-3.10. It's currently based on top of > cgroup/for-3.10 as the first patch causes non-trivial conflict with it > otherwise, which is not difficult to resolve but still nice to avoid > anyway. > > Li, Michal, I picked the first two patches from Li's memcg patchset. > Can we push the first two through cgroup/for-3.10 and put the rest in > -mm? > Sure. Andrew asked me to resend the memcg patchset when we are at 3.10-rc1, because 3.9-rc6 is a bit too late and he's too busy. > Ingo, how should these be routed?