Re: [PATCH 3/3] md/raid5: per hash value and exclusive wait_for_stripe

[NeilBrown]

On Thu, 14 May 2015 13:55:08 +0800 Yuanhan Liu <yuanhan.liu@xxxxxxxxxxxxxxx>
wrote:

> On Thu, May 14, 2015 at 03:45:11PM +1000, NeilBrown wrote:
> > On Wed, 29 Apr 2015 10:48:55 +0800 Yuanhan Liu <yuanhan.liu@xxxxxxxxxxxxxxx>
> > wrote:
> > 
> > > diff --git a/drivers/md/raid5.c b/drivers/md/raid5.c
> > > index 64d5bea..697d77a 100644
> > > --- a/drivers/md/raid5.c
> > > +++ b/drivers/md/raid5.c
> > > @@ -344,7 +344,8 @@ static void release_inactive_stripe_list(struct r5conf *conf,
> > >  					 int hash)
> > >  {
> > >  	int size;
> > > -	bool do_wakeup = false;
> > > +	unsigned long do_wakeup = 0;
> > > +	int i = 0;
> > >  	unsigned long flags;
> > >  
> > >  	if (hash == NR_STRIPE_HASH_LOCKS) {
> > > @@ -365,15 +366,19 @@ static void release_inactive_stripe_list(struct r5conf *conf,
> > >  			    !list_empty(list))
> > >  				atomic_dec(&conf->empty_inactive_list_nr);
> > >  			list_splice_tail_init(list, conf->inactive_list + hash);
> > > -			do_wakeup = true;
> > > +			do_wakeup |= 1 << (size - 1);
> > >  			spin_unlock_irqrestore(conf->hash_locks + hash, flags);
> > >  		}
> > >  		size--;
> > >  		hash--;
> > >  	}
> > >  
> > > +	for (i = 0; i < NR_STRIPE_HASH_LOCKS; i++) {
> > > +		if (do_wakeup & (1 << i))
> > > +			wake_up(&conf->wait_for_stripe[i]);
> > > +	}
> > > +
> > 
> > hi,
> >  I've been doing some testing and got a lock-up in resize_stripes, waiting
> >  on wait_for_stripe[].
> > 
> >  Looking at the above code,  I think
> >       do_wakeup |= 1 << (size - 1);
> >  should be
> >       do_wakeup |= 1 << hash;
> > 
> >  do you agree?  Or am I missing something?
> 
> Right. Sorry for the careless mistake.

Thanks for confirming.
Mistakes happen - that's why we test :-)

I've merged that fix into your patch.

NeilBrown

Attachment: pgpX7zOApOCUV.pgp
Description: OpenPGP digital signature