MIME-Version: 1.0
References: <cover.c238416f0e82377b449846dbb2459ae9d7030c8e.1675669136.git-series.apopple@nvidia.com>
 <c7b5e502d1a3b9b8f6e96cbf9ca553b143c327e0.1675669136.git-series.apopple@nvidia.com>
 <Y+Fttp1ozejoSQzl@slm.duckdns.org>
In-Reply-To: <Y+Fttp1ozejoSQzl@slm.duckdns.org>
From:   Yosry Ahmed <yosryahmed@google.com>
Date:   Mon, 6 Feb 2023 14:32:10 -0800
Message-ID: <CAJD7tkb_Cr7rTTpKc1VBpS8h=n3Hu+nGiV8dkLH-NdC1bSG9mg@mail.gmail.com>
Subject: Re: [PATCH 14/19] mm: Introduce a cgroup for pinned memory
To:     Tejun Heo <tj@kernel.org>
Cc:     Alistair Popple <apopple@nvidia.com>, linux-mm@kvack.org,
        cgroups@vger.kernel.org, linux-kernel@vger.kernel.org,
        jgg@nvidia.com, jhubbard@nvidia.com, tjmercier@google.com,
        hannes@cmpxchg.org, surenb@google.com, mkoutny@suse.com,
        daniel@ffwll.ch, "Daniel P . Berrange" <berrange@redhat.com>,
        Alex Williamson <alex.williamson@redhat.com>,
        Zefan Li <lizefan.x@bytedance.com>,
        Andrew Morton <akpm@linux-foundation.org>
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

On Mon, Feb 6, 2023 at 1:14 PM Tejun Heo <tj@kernel.org> wrote:
>
> On Mon, Feb 06, 2023 at 06:47:51PM +1100, Alistair Popple wrote:
> > If too much memory in a system is pinned or locked it can lead to
> > problems such as performance degradation or in the worst case
> > out-of-memory errors as such memory cannot be moved or paged out.
> >
> > In order to prevent users without CAP_IPC_LOCK from causing these
> > issues the amount of memory that can be pinned is typically limited by
> > RLIMIT_MEMLOCK. However this is inflexible as limits can't be shared
> > between tasks and the enforcement of these limits is inconsistent
> > between in-kernel users of pinned memory such as mlock() and device
> > drivers which may also pin pages with pin_user_pages().
> >
> > To allow for a single limit to be set introduce a cgroup controller
> > which can be used to limit the number of pages being pinned by all
> > tasks in the cgroup.
>
> As I wrote before, I think this might fit better as a part of memcg than as
> its own controller.

I guess it boils down to which we want:
(a) Limit the amount of memory processes in a cgroup can be pinned/locked.
(b) Limit the amount of memory charged to a cgroup that can be pinned/locked.

The proposal is doing (a), I suppose if this was part of memcg it
would be (b), right?

I am not saying it should be one or the other, I am just making sure
my understanding is clear.

>
> Thanks.
>
> --
> tejun