Re: [PATCH v4] md/raid5: fix atomicity violation in raid5_cache_count
From: Song Liu <song@kernel.org>
Date: 2024-01-30 07:37:24
Also in:
lkml, stable
On Thu, Jan 11, 2024 at 11:10 PM Gui-Dong Han [off-list ref] wrote:
[...]
raid5_release_stripe(sh);
- conf->max_nr_stripes++;
+ WRITE_ONCE(conf->max_nr_stripes, conf->max_nr_stripes + 1);This is weird. We are reading max_nr_stripes without READ_ONCE.
quoted hunk ↗ jump to hunk
return 1; }@@ -2707,7 +2707,7 @@ static int drop_one_stripe(struct r5conf *conf) shrink_buffers(sh); free_stripe(conf->slab_cache, sh); atomic_dec(&conf->active_stripes); - conf->max_nr_stripes--; + WRITE_ONCE(conf->max_nr_stripes, conf->max_nr_stripes - 1);
Same here.
quoted hunk ↗ jump to hunk
return 1; }@@ -6820,7 +6820,7 @@ raid5_set_cache_size(struct mddev *mddev, int size) if (size <= 16 || size > 32768) return -EINVAL; - conf->min_nr_stripes = size; + WRITE_ONCE(conf->min_nr_stripes, size); mutex_lock(&conf->cache_size_mutex); while (size < conf->max_nr_stripes && drop_one_stripe(conf))@@ -6832,7 +6832,7 @@ raid5_set_cache_size(struct mddev *mddev, int size) mutex_lock(&conf->cache_size_mutex); while (size > conf->max_nr_stripes) if (!grow_one_stripe(conf, GFP_KERNEL)) { - conf->min_nr_stripes = conf->max_nr_stripes; + WRITE_ONCE(conf->min_nr_stripes, conf->max_nr_stripes);
And here.
result = -ENOMEM;
break;
}Thanks, Song