From mboxrd@z Thu Jan 1 00:00:00 1970 From: Tejun Heo Subject: Re: [lxc-devel] [RFC PATCH] cgroup, cpuset: add cpuset.remap_cpus Date: Wed, 30 Aug 2017 17:41:31 -0700 Message-ID: <20170831004131.GZ491396@devbig577.frc2.facebook.com> References: <1482419271-15410-1-git-send-email-w.bumiller@proxmox.com> <20170830132755.tnqmuttodexc3oh6@mailbox.org> Mime-Version: 1.0 Return-path: DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20161025; h=sender:date:from:to:cc:subject:message-id:references:mime-version :content-disposition:in-reply-to:user-agent; bh=qMFzUnWBOvthJDrlYxfqDAQgIKhgh37RBEclaBJVb+I=; b=InXQSSNK+172b92cm6Spy1A31gly2E6fRMUaUyjE3Lsovluf7lhj/2wwGDoWTLQ7Kz KHxj3a0ftW4S+lKIhxMElWGLd5HToQPqOspypXicsWj0g+EjzxWCxkOX4SuV0PICW7cu WAnHVL8gPMBmbjf/LZ6HJp0MA42bxB611j+cobBoxAyqIBc9X7/VHfLFKftyBZ+7fnfo tAM8CUeum70CYJheuDG8oloHkHHVzbYoOCKR6wHZXIHZA3JUqfIz9SLYlZ4Eo1LIq2lG SZ4GKms0h86znB+VXX0nzk8zStFnjCMm7vpWsqlFcyInBhbUrxBa14qWmItvAHCKvHlg di/w== Content-Disposition: inline In-Reply-To: <20170830132755.tnqmuttodexc3oh6-cl+VPiYnx/1AfugRpC6u6w@public.gmane.org> Sender: cgroups-owner-u79uwXL29TY76Z2rM5mHXA@public.gmane.org List-ID: Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: christian.brauner-GeWIH/nMZzLQT0dZR+AlfA@public.gmane.org Cc: Li Zefan , cgroups-u79uwXL29TY76Z2rM5mHXA@public.gmane.org, linux-kernel-u79uwXL29TY76Z2rM5mHXA@public.gmane.org, w.bumiller-YTcQvvOqK21BDgjK7y7TUQ@public.gmane.org, stgraber-GeWIH/nMZzLQT0dZR+AlfA@public.gmane.org, serge-A9i7LUbDfNHQT0dZR+AlfA@public.gmane.org Hello, On Wed, Aug 30, 2017 at 03:27:55PM +0200, Christian Brauner wrote: > The following patch was sent a while back by Wolfgang Bumiller to remap cpusets > for a whole subtree in a cgroup v1 cpuset hierarchy. The fact that currently > this is not possible in a non-racy why is a pretty big limitation. This is > especially true for nested containers. Where the nested containers often create > additional subcgroups in the cpuset controller at will. The fact that you can't > *easily* and in a non-racy way tighten the restriction on them after having > created the parent container's cpuset cgroup seems really troubling. There was a recent patch to enable v2 behavior on v1, which feels like the better approach at this point. I'm not sure about adding a whole new interface for this. Thanks. -- tejun