[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-Id: <20220701160947.e4902e5b0484ed084db5d41f@linux-foundation.org>
Date: Fri, 1 Jul 2022 16:09:47 -0700
From: Andrew Morton <akpm@...ux-foundation.org>
To: Yosry Ahmed <yosryahmed@...gle.com>
Cc: Johannes Weiner <hannes@...xchg.org>,
Michal Hocko <mhocko@...nel.org>,
Roman Gushchin <roman.gushchin@...ux.dev>,
Shakeel Butt <shakeelb@...gle.com>,
Muchun Song <songmuchun@...edance.com>,
Matthew Wilcox <willy@...radead.org>,
Vlastimil Babka <vbabka@...e.cz>,
David Hildenbrand <david@...hat.com>,
Miaohe Lin <linmiaohe@...wei.com>, NeilBrown <neilb@...e.de>,
Alistair Popple <apopple@...dia.com>,
Suren Baghdasaryan <surenb@...gle.com>,
Peter Xu <peterx@...hat.com>, linux-kernel@...r.kernel.org,
cgroups@...r.kernel.org, linux-mm@...ck.org
Subject: Re: [PATCH v3] mm: vmpressure: don't count proactive reclaim in
vmpressure
On Thu, 30 Jun 2022 08:30:44 +0000 Yosry Ahmed <yosryahmed@...gle.com> wrote:
> vmpressure is used in cgroup v1 to notify userspace of reclaim
> efficiency events, and is also used in both cgroup v1 and v2 as a signal
> for memory pressure for networking, see
> mem_cgroup_under_socket_pressure().
>
> Proactive reclaim intends to probe memcgs for cold memory, without
> affecting their performance. Hence, reclaim caused by writing to
> memory.reclaim should not trigger vmpressure.
>
> ...
>
> --- a/mm/memcontrol.c
> +++ b/mm/memcontrol.c
> @@ -2319,6 +2319,7 @@ static unsigned long reclaim_high(struct mem_cgroup *memcg,
> gfp_t gfp_mask)
> {
> unsigned long nr_reclaimed = 0;
> + unsigned int reclaim_options = MEMCG_RECLAIM_MAY_SWAP;
>
> do {
> unsigned long pflags;
> @@ -2331,7 +2332,8 @@ static unsigned long reclaim_high(struct mem_cgroup *memcg,
>
> psi_memstall_enter(&pflags);
> nr_reclaimed += try_to_free_mem_cgroup_pages(memcg, nr_pages,
> - gfp_mask, true);
> + gfp_mask,
> + reclaim_options);
It's a bit irksome to create all these unneeded local variables. Why
not simply add the constant arg to the try_to_free_mem_cgroup_pages()
call?
> psi_memstall_leave(&pflags);
> } while ((memcg = parent_mem_cgroup(memcg)) &&
> !mem_cgroup_is_root(memcg));
> @@ -2576,7 +2578,7 @@ static int try_charge_memcg(struct mem_cgroup *memcg, gfp_t gfp_mask,
> struct page_counter *counter;
> unsigned long nr_reclaimed;
> bool passed_oom = false;
> - bool may_swap = true;
> + unsigned int reclaim_options = MEMCG_RECLAIM_MAY_SWAP;
> bool drained = false;
> unsigned long pflags;
>
> @@ -2593,7 +2595,7 @@ static int try_charge_memcg(struct mem_cgroup *memcg, gfp_t gfp_mask,
> mem_over_limit = mem_cgroup_from_counter(counter, memory);
> } else {
> mem_over_limit = mem_cgroup_from_counter(counter, memsw);
> - may_swap = false;
> + reclaim_options &= ~MEMCG_RECLAIM_MAY_SWAP;
reclaim_options = 0
would be clearer?
Powered by blists - more mailing lists