Re:Re: [PATCH] md/raid5: serialize plug list add with device_lock

["李佑鸿 " <dayou5941@163.com>]

At 2026-09-12 22:04:12, "yu kuai" <yukuai@fygo.io> wrote:
>Hi,
>
>在 2026/9/7 11:24, 李佑鸿 写道:
>>
>>
>>
>> Hi Kuai,
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>> At 2026-09-05 11:17:58, "yu kuai" <yukuai@fygo.io> wrote:
>>> Hi,
>>>
>>> 在 2026/9/2 17:53, Li Youhong 写道:
>>>> From: Li Youhong <liyouhong@kylinos.cn>
>>>>
>>>> raid5_unplug() can spin forever under conf->device_lock when
>>>> raid5_plug_cb.list still points at a stripe whose sh->lru has
>>>> already been reinitialized (self-looped). That disables IRQs on
>>>> the holder CPU and causes multi-CPU hard lockups on waiters of
>>>> the same lock.
>>>>
>>>> This happens because release_stripe_plug() sets
>>>> STRIPE_ON_UNPLUG_LIST and list_add_tail(sh->lru) without
>>>> device_lock, while do_release_stripe() may concurrently move the
>>>> same lru onto handle/inactive when the last reference drops.
>>>>
>>>> Note: a 2020 proposal tried extra refs / checking
>>>> STRIPE_ON_UNPLUG_LIST in do_release_stripe() without serializing
>>>> the plug enqueue:
>>>> https://lore.kernel.org/linux-raid/20200108163023.9301-1-guoqing.jiang@cloud.ionos.com/
>>>> That still leaves a TOCTOU window where the bit is clear during
>>>> the check and set afterwards, allowing two list_add on the same
>>>> lru. It was not merged.
>>>>
>>>> Serialize the bit update and list_add with device_lock. If
>>>> do_release_stripe() still sees STRIPE_ON_UNPLUG_LIST, restore the
>>>> reference and let raid5_unplug() own the final release.
>>>>
>>>> Observed on production 9-disk NVMe RAID5 under MySQL AIO
>>>> (io_submit -> blk_finish_plug -> raid5_unplug).
>>>>
>>>> Fixes: 8811b5968f62 ("raid5: make_request use batch stripe release")
>>>> Cc: stable@vger.kernel.org
>>>> Signed-off-by: Li Youhong <liyouhong@kylinos.cn>
>>>> ---
>>>>    drivers/md/raid5.c | 33 ++++++++++++++++++++++++++++++---
>>>>    1 file changed, 30 insertions(+), 3 deletions(-)
>>>>
>>>> diff --git a/drivers/md/raid5.c b/drivers/md/raid5.c
>>>> index b91545ce090d..c9197b63b1c4 100644
>>>> --- a/drivers/md/raid5.c
>>>> +++ b/drivers/md/raid5.c
>>>> @@ -229,6 +229,17 @@ static void do_release_stripe(struct r5conf *conf, struct stripe_head *sh,
>>>>    	int i;
>>>>    	int injournal = 0;	/* number of date pages with R5_InJournal */
>>>>    
>>>> +	/*
>>>> +	 * Stripe is owned by release_stripe_plug()'s cb->list. A concurrent
>>>> +	 * last-ref release can reach here after the stripe was queued for
>>>> +	 * unplug (lru may already be non-empty). Do not re-add lru elsewhere;
>>>> +	 * restore the reference and let raid5_unplug() finish the release.
>>>> +	 */
>>>> +	if (test_bit(STRIPE_ON_UNPLUG_LIST, &sh->state)) {
>>>> +		atomic_inc(&sh->count);
>>>> +		return;
>>>> +	}
>>>> +
>>>>    	BUG_ON(!list_empty(&sh->lru));
>>>>    	BUG_ON(atomic_read(&conf->active_stripes)==0);
>>>>    
>>>> @@ -5760,6 +5771,9 @@ static void release_stripe_plug(struct mddev *mddev,
>>>>    		raid5_unplug, mddev,
>>>>    		sizeof(struct raid5_plug_cb));
>>>>    	struct raid5_plug_cb *cb;
>>>> +	struct r5conf *conf = mddev->private;
>>>> +	unsigned long flags;
>>>> +	bool queued = false;
>>>>    
>>>>    	if (!blk_cb) {
>>>>    		raid5_release_stripe(sh);
>>>> @@ -5775,9 +5789,22 @@ static void release_stripe_plug(struct mddev *mddev,
>>>>    			INIT_LIST_HEAD(cb->temp_inactive_list + i);
>>>>    	}
>>>>    
>>>> -	if (!test_and_set_bit(STRIPE_ON_UNPLUG_LIST, &sh->state))
>>>> -		list_add_tail(&sh->lru, &cb->list);
>>>> -	else
>>>> +	/*
>>>> +	 * Serialize with do_release_stripe() on device_lock so sh->lru cannot
>>>> +	 * be added to handle/inactive and cb->list at the same time.
>>>> +	 */
>>>> +	spin_lock_irqsave(&conf->device_lock, flags);
>>>> +	if (!test_and_set_bit(STRIPE_ON_UNPLUG_LIST, &sh->state)) {
>>>> +		if (unlikely(!list_empty(&sh->lru))) {
>>>> +			clear_bit(STRIPE_ON_UNPLUG_LIST, &sh->state);
>>>> +		} else {
>>>> +			list_add_tail(&sh->lru, &cb->list);
>>>> +			queued = true;
>>>> +		}
>>>> +	}
>>>> +	spin_unlock_irqrestore(&conf->device_lock, flags);
>>> Do you run some performance test on this patch? I believe this is from IO hot path,
>>> a spinlock is not acceptable. I know there is already some spinlock for raid5, but
>>> I'd like not to introduce new lock contention.
>>>
>>> Can this problem be fixed by a new llist in stripe_head?
>>
>> I agree that adding device_lock on this path can bring some
>> overhead; I will follow up with performance numbers.
>>
>>
>> Regarding the alternative: did you mean adding a dedicated
>> llist_node in stripe_head for the plug/unplug path (instead of
>> reusing sh->lru), so that release_stripe_plug() and
>> do_release_stripe() no longer share the same list node?
>>
>>
>> If we still queue the plug list via sh->lru, switching that list
>> to an llist would not remove the race, because do_release_stripe()
>> can still list_add the same lru onto handle/inactive concurrently.
>> A separate node would avoid that, but unplug would then need to
>> move stripes from the plug llist onto the existing lru-based
>> release path, which is a larger change.
>
>I mean a new list for unplug in sh like unplug_llist. And I think it's
>fine to use the llist directly for unplug, there is no need to move
>sh from unplug llist to lru list first.
>

A dedicated llist would avoid two list_adds on the same sh->lru from plug and do_release_stripe(). 
Last-ref enqueue still does not drop sh->count; the decrement happens later in __release_stripe(). 
So at count == 1 the same stripe can sit on both the plug list and released_stripes. What follows is 
that one undropped last-ref being released twice by __release_stripe(), not a separate bug.

raid5_unplug() and release_stripe_list() both take device_lock, so they do not run __release_stripe() in parallel. 
They run one after the other. The only difference is whether another get happens between the two drains.

1. No get in between
The first __release_stripe() does 1→0 and puts the stripe on inactive.
The second does 0→-1. dec_and_test is false, so it does not enter do_release_stripe() 
and there is no second list_add(lru). Count is already -1, so a later get_free_stripe() → init_stripe() hits the BUG.

2. A get in between
The first __release_stripe() does 1→0.
The stripe is gotten again and count is 1; that is the new I/O's reference.
The second __release_stripe() drops that new 1 to 0.

The second drain is not necessarily unplug; release_stripe_list() running second is the same. 
At drain time count == 1 does not tell you whether this is the parked last-ref or a ref after reuse.

Enqueue to released_stripes happens only at count == 1. When count > 1, raid5_release_stripe() only does add_unless and returns; 
it does not enqueue. If the stripe is already on released_stripes, a get raises count above 1, and then plug runs, both lists can be occupied. 
That is two references and two drains each dropping one; only the drop to 0 does list_add(lru). Cases 1 and 2 are only about parking twice at count == 1.

So an llist only fixes the lru collision. It does not fix last-ref ownership. I do not see a lockless way to fix that.

Thanks,
Li Youhong