Re: [PATCH 6/7] mempool: fix a wakeup race when sleeping for elements

[Christoph Hellwig <hch@lst.de>]

On Wed, Nov 12, 2025 at 11:53:39AM +0100, Vlastimil Babka wrote:
> > +alloc:
> >  	element = remove_element(pool);
> >  	spin_unlock_irqrestore(&pool->lock, flags);
> >  
> > @@ -406,13 +407,17 @@ static void *mempool_alloc_from_pool(struct mempool *pool, gfp_t gfp_mask)
> >  		DEFINE_WAIT(wait);
> >  
> >  		prepare_to_wait(&pool->wait, &wait, TASK_UNINTERRUPTIBLE);
> > +		if (pool->curr_nr) {
> > +			finish_wait(&pool->wait, &wait);
> > +			goto alloc;
> > +		}
> >  		spin_unlock_irqrestore(&pool->lock, flags);
> I think the race already cannot exist, thanks to the pool->lock being
> unlocked after prepare_to_wait()?
> The freeing path can't bump pool->curr_nr without the lock, so the condition
> you added can't even be true, no?

You're right, I'll drop this again.