From mboxrd@z Thu Jan 1 00:00:00 1970 From: Shaohua Li Subject: Re: [patch]raid5: get_active_stripe avoids device_lock Date: Wed, 9 Apr 2014 11:27:42 +0800 Message-ID: <20140409032742.GA16051@kernel.org> References: <20140408040553.GB20886@kernel.org> <20140409121729.18955461@notabene.brown> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Content-Disposition: inline In-Reply-To: <20140409121729.18955461@notabene.brown> Sender: linux-raid-owner@vger.kernel.org To: NeilBrown Cc: linux-raid@vger.kernel.org List-Id: linux-raid.ids On Wed, Apr 09, 2014 at 12:17:29PM +1000, NeilBrown wrote: > On Tue, 8 Apr 2014 12:05:53 +0800 Shaohua Li wrote: > > > > > For sequential workload (or request size big workload), get_active_stripe can > > find cached stripe. In this case, we always hold device_lock, which exposes a > > lot of lock contention for such workload. If stripe count isn't 0, we don't > > need hold the lock actually, since we just increase its count. And this is the > > hot code path for such workload. Unfortunately we must delete the BUG_ON. > > > > Signed-off-by: Shaohua Li > > --- > > drivers/md/raid5.c | 9 ++------- > > 1 file changed, 2 insertions(+), 7 deletions(-) > > > > Index: linux/drivers/md/raid5.c > > =================================================================== > > --- linux.orig/drivers/md/raid5.c 2014-04-08 09:16:39.377536607 +0800 > > +++ linux/drivers/md/raid5.c 2014-04-08 09:16:39.369536607 +0800 > > @@ -679,14 +679,9 @@ get_active_stripe(struct r5conf *conf, s > > init_stripe(sh, sector, previous); > > atomic_inc(&sh->count); > > } > > - } else { > > + } else if (!atomic_add_unless(&sh->count, 1, 0)) { > > Can I be really fussy and ask you to use "atomic_inc_not_zero" rather than > "atomic_add_unless" ?? > I feel it makes the code a bit clearer. Missed this API. Yes, it's better. Subject: raid5: get_active_stripe avoids device_lock For sequential workload (or request size big workload), get_active_stripe can find cached stripe. In this case, we always hold device_lock, which exposes a lot of lock contention for such workload. If stripe count isn't 0, we don't need hold the lock actually, since we just increase its count. And this is the hot code path for such workload. Unfortunately we must delete the BUG_ON. Signed-off-by: Shaohua Li --- drivers/md/raid5.c | 9 ++------- 1 file changed, 2 insertions(+), 7 deletions(-) Index: linux/drivers/md/raid5.c =================================================================== --- linux.orig/drivers/md/raid5.c 2014-04-09 11:07:01.165231305 +0800 +++ linux/drivers/md/raid5.c 2014-04-09 11:07:26.492913067 +0800 @@ -679,14 +679,9 @@ get_active_stripe(struct r5conf *conf, s init_stripe(sh, sector, previous); atomic_inc(&sh->count); } - } else { + } else if (!atomic_inc_not_zero(&sh->count)) { spin_lock(&conf->device_lock); - if (atomic_read(&sh->count)) { - BUG_ON(!list_empty(&sh->lru) - && !test_bit(STRIPE_EXPANDING, &sh->state) - && !test_bit(STRIPE_ON_UNPLUG_LIST, &sh->state) - ); - } else { + if (!atomic_read(&sh->count)) { if (!test_bit(STRIPE_HANDLE, &sh->state)) atomic_inc(&conf->active_stripes); BUG_ON(list_empty(&sh->lru) &&