From: Michal Hocko Subject: Re: [PATCH 2/6] mm, memcg: Prevent memory.max load tearing Date: Mon, 16 Mar 2020 15:56:30 +0100 Message-ID: <20200316145630.GN11482@dhcp22.suse.cz> References: <50a31e5f39f8ae6c8fb73966ba1455f0924e8f44.1584034301.git.chris@chrisdown.name> Mime-Version: 1.0 Return-path: Content-Disposition: inline In-Reply-To: <50a31e5f39f8ae6c8fb73966ba1455f0924e8f44.1584034301.git.chris-6Bi1550iOqEnzZ6mRAm98g@public.gmane.org> Sender: cgroups-owner-u79uwXL29TY76Z2rM5mHXA@public.gmane.org List-ID: Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: Chris Down Cc: Andrew Morton , Johannes Weiner , Tejun Heo , Roman Gushchin , linux-mm-Bw31MaZKKs3YtjvyW6yDsg@public.gmane.org, cgroups-u79uwXL29TY76Z2rM5mHXA@public.gmane.org, linux-kernel-u79uwXL29TY76Z2rM5mHXA@public.gmane.org, kernel-team-b10kYP2dOMg@public.gmane.org On Thu 12-03-20 17:32:56, Chris Down wrote: > This one is a bit more nuanced because we have memcg_max_mutex, which is > mostly just used for enforcing invariants, but we still need to > READ_ONCE since (despite its name) it doesn't really protect memory.max > access. > > On write (page_counter_set_max() and memory_max_write()) we use xchg(), > which uses smp_mb(), so that's already fine. > > Signed-off-by: Chris Down > Cc: Andrew Morton > Cc: Johannes Weiner > Cc: Roman Gushchin > Cc: Tejun Heo > Cc: linux-mm-Bw31MaZKKs3YtjvyW6yDsg@public.gmane.org > Cc: cgroups-u79uwXL29TY76Z2rM5mHXA@public.gmane.org > Cc: linux-kernel-u79uwXL29TY76Z2rM5mHXA@public.gmane.org > Cc: kernel-team-b10kYP2dOMg@public.gmane.org Acked-by: Michal Hocko > --- > mm/memcontrol.c | 12 ++++++------ > 1 file changed, 6 insertions(+), 6 deletions(-) > > diff --git a/mm/memcontrol.c b/mm/memcontrol.c > index d32d3c0a16d4..aca2964ea494 100644 > --- a/mm/memcontrol.c > +++ b/mm/memcontrol.c > @@ -1507,7 +1507,7 @@ void mem_cgroup_print_oom_meminfo(struct mem_cgroup *memcg) > > pr_info("memory: usage %llukB, limit %llukB, failcnt %lu\n", > K((u64)page_counter_read(&memcg->memory)), > - K((u64)memcg->memory.max), memcg->memory.failcnt); > + K((u64)READ_ONCE(memcg->memory.max)), memcg->memory.failcnt); > if (cgroup_subsys_on_dfl(memory_cgrp_subsys)) > pr_info("swap: usage %llukB, limit %llukB, failcnt %lu\n", > K((u64)page_counter_read(&memcg->swap)), > @@ -1538,7 +1538,7 @@ unsigned long mem_cgroup_get_max(struct mem_cgroup *memcg) > { > unsigned long max; > > - max = memcg->memory.max; > + max = READ_ONCE(memcg->memory.max); > if (mem_cgroup_swappiness(memcg)) { > unsigned long memsw_max; > unsigned long swap_max; > @@ -3006,7 +3006,7 @@ static int mem_cgroup_resize_max(struct mem_cgroup *memcg, > * Make sure that the new limit (memsw or memory limit) doesn't > * break our basic invariant rule memory.max <= memsw.max. > */ > - limits_invariant = memsw ? max >= memcg->memory.max : > + limits_invariant = memsw ? max >= READ_ONCE(memcg->memory.max) : > max <= memcg->memsw.max; > if (!limits_invariant) { > mutex_unlock(&memcg_max_mutex); > @@ -3753,8 +3753,8 @@ static int memcg_stat_show(struct seq_file *m, void *v) > /* Hierarchical information */ > memory = memsw = PAGE_COUNTER_MAX; > for (mi = memcg; mi; mi = parent_mem_cgroup(mi)) { > - memory = min(memory, mi->memory.max); > - memsw = min(memsw, mi->memsw.max); > + memory = min(memory, READ_ONCE(mi->memory.max)); > + memsw = min(memsw, READ_ONCE(mi->memsw.max)); > } > seq_printf(m, "hierarchical_memory_limit %llu\n", > (u64)memory * PAGE_SIZE); > @@ -4257,7 +4257,7 @@ void mem_cgroup_wb_stats(struct bdi_writeback *wb, unsigned long *pfilepages, > *pheadroom = PAGE_COUNTER_MAX; > > while ((parent = parent_mem_cgroup(memcg))) { > - unsigned long ceiling = min(memcg->memory.max, > + unsigned long ceiling = min(READ_ONCE(memcg->memory.max), > READ_ONCE(memcg->high)); > unsigned long used = page_counter_read(&memcg->memory); > > -- > 2.25.1 -- Michal Hocko SUSE Labs