Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
Date:   Thu, 22 Feb 2018 21:19:55 -0600 (CST)
From:   Christopher Lameter <cl@linux.com>
To:     Andrew Morton <akpm@linux-foundation.org>
cc:     Shakeel Butt <shakeelb@google.com>, Jan Kara <jack@suse.cz>,
        Amir Goldstein <amir73il@gmail.com>,
        Pekka Enberg <penberg@kernel.org>,
        David Rientjes <rientjes@google.com>,
        Joonsoo Kim <iamjoonsoo.kim@lge.com>,
        Greg Thelen <gthelen@google.com>,
        Johannes Weiner <hannes@cmpxchg.org>,
        Michal Hocko <mhocko@kernel.org>,
        Vladimir Davydov <vdavydov.dev@gmail.com>,
        Mel Gorman <mgorman@suse.de>, Vlastimil Babka <vbabka@suse.cz>,
        linux-fsdevel <linux-fsdevel@vger.kernel.org>,
        Linux MM <linux-mm@kvack.org>,
        Cgroups <cgroups@vger.kernel.org>,
        LKML <linux-kernel@vger.kernel.org>
Subject: Re: [PATCH v2 0/3] Directed kmem charging
In-Reply-To: <20180221125426.464f894d29a0b6e525b2e3be@linux-foundation.org>
Message-ID: <alpine.DEB.2.20.1802222117140.6687@nuc-kabylake>
References: <20180221030101.221206-1-shakeelb@google.com> <alpine.DEB.2.20.1802211002200.12567@nuc-kabylake> <CALvZod68LD-wnbm2+MQks=bd_D2zY64uScUBp28hyug_vaGyDA@mail.gmail.com> <20180221125426.464f894d29a0b6e525b2e3be@linux-foundation.org>
User-Agent: Alpine 2.20 (DEB 67 2015-01-07)
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Wed, 21 Feb 2018, Andrew Morton wrote:

> What do others think?

I think the changes to the hotpaths of the slab allocators increasing
register pressure in some of the hotttest paths of the kernel are
problematic.

Its better to do the allocation properly in the task context to which it
is finally charged. There may be other restrictions that emerge from other
fields in the task_struct that also may influence allocation and reclaim
behavior. It is cleanest to do this in the proper task context instead of
taking a piece (like the cgroup) out of context.