[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <uj6railxyazpu6ocl2ygo6lw4lavbsgg26oq57pxxqe5uzxw42@fhnqvq3tia6n>
Date: Mon, 6 Jan 2025 18:48:31 +0100
From: Michal Koutný <mkoutny@...e.com>
To: Maarten Lankhorst <dev@...khorst.se>
Cc: linux-kernel@...r.kernel.org, intel-xe@...ts.freedesktop.org,
dri-devel@...ts.freedesktop.org, Tejun Heo <tj@...nel.org>, Zefan Li <lizefan.x@...edance.com>,
Johannes Weiner <hannes@...xchg.org>, Andrew Morton <akpm@...ux-foundation.org>,
Friedrich Vock <friedrich.vock@....de>, Maxime Ripard <mripard@...nel.org>, cgroups@...r.kernel.org,
linux-mm@...ck.org, Maarten Lankhorst <maarten.lankhorst@...ux.intel.com>
Subject: Re: [PATCH v2 1/7] kernel/cgroup: Add "dmem" memory accounting cgroup
Hello.
On Wed, Dec 04, 2024 at 02:44:01PM +0100, Maarten Lankhorst <dev@...khorst.se> wrote:
> +bool dmem_cgroup_state_evict_valuable(struct dmem_cgroup_pool_state *limit_pool,
> + struct dmem_cgroup_pool_state *test_pool,
> + bool ignore_low, bool *ret_hit_low)
> +{
> + struct dmem_cgroup_pool_state *pool = test_pool;
> + struct page_counter *climit, *ctest;
> + u64 used, min, low;
> +
> + /* Can always evict from current pool, despite limits */
> + if (limit_pool == test_pool)
> + return true;
> +
> + if (limit_pool) {
> + if (!parent_dmemcs(limit_pool->cs))
> + return true;
> +
> + for (pool = test_pool; pool && limit_pool != pool; pool = pool_parent(pool))
> + {}
> +
> + if (!pool)
> + return false;
> + } else {
> + /*
> + * If there is no cgroup limiting memory usage, use the root
> + * cgroup instead for limit calculations.
> + */
> + for (limit_pool = test_pool; pool_parent(limit_pool); limit_pool = pool_parent(limit_pool))
> + {}
> + }
I'm trying to understand the two branches above. If limit_pool is a root
one, eviction is granted and no protection is evaluated.
Then it checks that test_pool is below limit_pool (can this ever fail,
given the limit_pool must have been above when charging in test_pool?).
(OK, this may be called arbitrarily by modules.)
I think it could be simplified and corrected like this:
/* Resolve NULL limit_pool */
if (!limit_pool)
for (limit_pool = test_pool; pool_parent(limit_pool); limit_pool = pool_parent(limit_pool));
/* Check ancestry */
if (!cgroup_is_descendant(test_pool->cs->css.cgroup, limit_pool->cs->css.cgroup))
return false;
HTH,
Michal
Download attachment "signature.asc" of type "application/pgp-signature" (229 bytes)
Powered by blists - more mailing lists