md/raid5: don't let shrink_slab shrink too far. (49895bcc) · Commits · 戴 / test

I have a report of drop_one_stripe() called from raid5_cache_scan() apparently finding ->max_nr_stripes == 0. This should not be allowed. So add a test to keep max_nr_stripes above min_nr_stripes. Also use a 'mask' rather than a 'mod' in drop_one_stripe to ensure 'hash' is valid even if max_nr_stripes does reach zero. Fixes: ("md/raid5: allow the stripe_cache to grow and shrink.") Cc: stable@vger.kernel.org (4.1 - please release with ) Reported-by:

Tomas Papan <tomas.papan@gmail.com> Signed-off-by:

NeilBrown <neilb@suse.com>

drivers/md/raid5.c

+3 −2

Original line number	Original line	Diff line number	Diff line
	@@ -2256,7 +2256,7 @@ static int resize_stripes(struct r5conf *conf, int newsize)
	static int drop_one_stripe(struct r5conf *conf)		static int drop_one_stripe(struct r5conf *conf)
	{		{
	struct stripe_head *sh;		struct stripe_head *sh;
	int hash = (conf->max_nr_stripes - 1) % NR_STRIPE_HASH_LOCKS;		int hash = (conf->max_nr_stripes - 1) & STRIPE_HASH_LOCKS_MASK;

	spin_lock_irq(conf->hash_locks + hash);		spin_lock_irq(conf->hash_locks + hash);
	sh = get_free_stripe(conf, hash);		sh = get_free_stripe(conf, hash);
	@@ -6388,7 +6388,8 @@ static unsigned long raid5_cache_scan(struct shrinker *shrink,

	if (mutex_trylock(&conf->cache_size_mutex)) {		if (mutex_trylock(&conf->cache_size_mutex)) {
	ret= 0;		ret= 0;
	while (ret < sc->nr_to_scan) {		while (ret < sc->nr_to_scan &&
			conf->max_nr_stripes > conf->min_nr_stripes) {
	if (drop_one_stripe(conf) == 0) {		if (drop_one_stripe(conf) == 0) {
	ret = SHRINK_STOP;		ret = SHRINK_STOP;
	break;		break;

Admin message