On Wed 24-04-24 05:59:39, Breno Leitao wrote:
> A data-race issue in memcg rstat occurs when two distinct code paths
> access the same 4-byte region concurrently. KCSAN detection triggers the
> following BUG as a result.
>
> BUG: KCSAN: data-race in __count_memcg_events / mem_cgroup_css_rstat_flush
>
> write to 0xffffe8ffff98e300 of 4 bytes by task 5274 on cpu 17:
> mem_cgroup_css_rstat_flush (mm/memcontrol.c:5850)
> cgroup_rstat_flush_locked (kernel/cgroup/rstat.c:243 (discriminator 7))
> cgroup_rstat_flush (./include/linux/spinlock.h:401 kernel/cgroup/rstat.c:278)
> mem_cgroup_flush_stats.part.0 (mm/memcontrol.c:767)
> memory_numa_stat_show (mm/memcontrol.c:6911)
> <snip>
>
> read to 0xffffe8ffff98e300 of 4 bytes by task 410848 on cpu 27:
> __count_memcg_events (mm/memcontrol.c:725 mm/memcontrol.c:962)
> count_memcg_event_mm.part.0 (./include/linux/memcontrol.h:1097 ./include/linux/memcontrol.h:1120)
> handle_mm_fault (mm/memory.c:5483 mm/memory.c:5622)
> <snip>
>
> value changed: 0x00000029 -> 0x00000000
>
> The race occurs because two code paths access the same "stats_updates"
> location. Although "stats_updates" is a per-CPU variable, it is remotely
> accessed by another CPU at
> cgroup_rstat_flush_locked()->mem_cgroup_css_rstat_flush(), leading to
> the data race mentioned.
>
> Considering that memcg_rstat_updated() is in the hot code path, adding
> a lock to protect it may not be desirable, especially since this
> variable pertains solely to statistics.
>
> Therefore, annotating accesses to stats_updates with READ/WRITE_ONCE()
> can prevent KCSAN splats and potential partial reads/writes.
>
> Suggested-by: Shakeel Butt <shakeel.butt@xxxxxxxxx>
> Signed-off-by: Breno Leitao <leitao@xxxxxxxxxx>
Acked-by: Michal Hocko <mhocko@xxxxxxxx>
It is worth mentioning that the race is harmless.
Thanks!
> ---
> mm/memcontrol.c | 12 +++++++-----
> 1 file changed, 7 insertions(+), 5 deletions(-)
>
> diff --git a/mm/memcontrol.c b/mm/memcontrol.c
> index fabce2b50c69..3c99457b36a1 100644
> --- a/mm/memcontrol.c
> +++ b/mm/memcontrol.c
> @@ -715,6 +715,7 @@ static inline void memcg_rstat_updated(struct mem_cgroup *memcg, int val)
> {
> struct memcg_vmstats_percpu *statc;
> int cpu = smp_processor_id();
> + unsigned int stats_updates;
>
> if (!val)
> return;
> @@ -722,8 +723,9 @@ static inline void memcg_rstat_updated(struct mem_cgroup *memcg, int val)
> cgroup_rstat_updated(memcg->css.cgroup, cpu);
> statc = this_cpu_ptr(memcg->vmstats_percpu);
> for (; statc; statc = statc->parent) {
> - statc->stats_updates += abs(val);
> - if (statc->stats_updates < MEMCG_CHARGE_BATCH)
> + stats_updates = READ_ONCE(statc->stats_updates) + abs(val);
> + WRITE_ONCE(statc->stats_updates, stats_updates);
> + if (stats_updates < MEMCG_CHARGE_BATCH)
> continue;
>
> /*
> @@ -731,9 +733,9 @@ static inline void memcg_rstat_updated(struct mem_cgroup *memcg, int val)
> * redundant. Avoid the overhead of the atomic update.
> */
> if (!memcg_vmstats_needs_flush(statc->vmstats))
> - atomic64_add(statc->stats_updates,
> + atomic64_add(stats_updates,
> &statc->vmstats->stats_updates);
> - statc->stats_updates = 0;
> + WRITE_ONCE(statc->stats_updates, 0);
> }
> }
>
> @@ -5845,7 +5847,7 @@ static void mem_cgroup_css_rstat_flush(struct cgroup_subsys_state *css, int cpu)
> }
> }
> }
> - statc->stats_updates = 0;
> + WRITE_ONCE(statc->stats_updates, 0);
> /* We are in a per-cpu loop here, only do the atomic write once */
> if (atomic64_read(&memcg->vmstats->stats_updates))
> atomic64_set(&memcg->vmstats->stats_updates, 0);
> --
> 2.43.0
>
--
Michal Hocko
SUSE Labs
