[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <rwb5lqorevlmywo6da623oc4vacy3wjfxaokfx3it52fkvhrvx@qyef5cmd5sgr>
Date: Mon, 27 May 2024 13:54:29 +0200
From: Michal Koutný <mkoutny@...e.com>
To: "T.J. Mercier" <tjmercier@...gle.com>
Cc: Tejun Heo <tj@...nel.org>, Zefan Li <lizefan.x@...edance.com>,
Johannes Weiner <hannes@...xchg.org>, shakeel.butt@...ux.dev, cgroups@...r.kernel.org,
linux-kernel@...r.kernel.org
Subject: Re: [RFC] cgroup: Fix /proc/cgroups count for v2
On Fri, May 24, 2024 at 10:36:45AM GMT, "T.J. Mercier" <tjmercier@...gle.com> wrote:
> On Fri, May 24, 2024 at 7:23 AM Michal Koutný <mkoutny@...e.com> wrote:
> Sometimes? Take freezer as an example. If you don't mount it on v1
> then /proc/cgroups currently advertises the total number of v2
> cgroups. I thought that was reasonable since there exists a
> cgroup.freeze in every cgroup, but does freezer really count as a
> controller in this case?
v1 freezer controller and freezing implementation in v2 are different
things.
Before v1 mounting, the freezer* entry points to the v2 hierarchy (which
causes listing it as realized for each (v2) cgroup but that's not true).
> There's no freezer css for each cgroup
Exactly.
> so I guess the better answer is just to report 1 like you suggest.
It matches better the reality of alloc'd css objects.
Michal
*) Same for any v1-only controller.
Download attachment "signature.asc" of type "application/pgp-signature" (229 bytes)
Powered by blists - more mailing lists