From: Uladzislau Rezki <urezki@gmail.com>
To: Hillf Danton <hdanton@sina.com>
Cc: Uladzislau Rezki <urezki@gmail.com>,
"Paul E . McKenney" <paulmck@kernel.org>,
RCU <rcu@vger.kernel.org>,
Neeraj upadhyay <Neeraj.Upadhyay@amd.com>,
Boqun Feng <boqun.feng@gmail.com>,
LKML <linux-kernel@vger.kernel.org>,
Oleksiy Avramchenko <oleksiy.avramchenko@sony.com>,
Frederic Weisbecker <frederic@kernel.org>
Subject: Re: [PATCH v3 1/1] rcu: Reduce synchronize_rcu() waiting time
Date: Thu, 19 Oct 2023 17:44:23 +0200 [thread overview]
Message-ID: <ZTFO17eNKvDNokpe@pc636> (raw)
In-Reply-To: <20231019114432.1995-1-hdanton@sina.com>
On Thu, Oct 19, 2023 at 07:44:32PM +0800, Hillf Danton wrote:
> On Wed, 18 Oct 2023 13:20:49 +0200 Uladzislau Rezki (Sony) <urezki@gmail.com> wrote:
> > > On Mon, Oct 16, 2023 at 1:30 PM Uladzislau Rezki (Sony) <urezki@gmail.com> wrote:
> > > > +static void rcu_sr_normal_gp_init(void)
> > > > +{
> > > > + struct llist_node *llnode, *rcu;
> > > > + int ret;
> > > > +
> > > > + if (llist_empty(&sr.curr))
> > > > + return;
> > >
> > > This empty check erases the curr_tail race below instead of
> > > atomic_inc_return(&sr.active), because llist_add() will never return true
> > > after this check.
> > >
> > I use "active" counter to guarantee that a tail was updated in the
> > rcu_sr_normal_add_req(), i.e. the list might be not empty whereas the
> > tail updating might be in progress. llist_add() success and the task gets
> > preemted as an example.
>
> You are right - the preempt is what I missed.
>
> Then another question rising - the adding order of sync requests is changed
> at wakeup time, as shown by the two functions below with sr.curr_tail and
> sr.active cut off.
>
> static void rcu_sr_normal_gp_init(void)
> {
> struct llist_node *llnode, *rcu;
>
> llnode = llist_del_all(&sr.curr);
> rcu = llist_reverse_order(llnode);
> if (!rcu)
> return;
> /*
> * A waiting list of GP should be empty on this step,
> * since a GP-kthread, rcu_gp_init() -> gp_cleanup(),
> * rolls it over. If not, it is a BUG, warn a user.
> */
> WARN_ON_ONCE(!llist_empty(&sr.wait));
>
> WRITE_ONCE(sr.wait_tail, llnode);
>
> llist_add_batch(rcu, llnode, &sr.wait);
> }
>
> static void rcu_sr_normal_add_req(struct rcu_synchronize *rs)
> {
> llist_add((struct llist_node *) &rs->head, &sr.curr);
> }
>
This is what i was thinking of in the beginning but i had decided
to make it more complex and maintain the current tail. We will get
a slight penalty in performance on a synthetic test but i have compared
it and this is negligible in fact.
Another good thing with it is we process users in a reverse order,
i.e. the most waiting users to less ones, this is from a waiting time
point of view.
I like it better and it is more simple. Anyway we can improve it
further.
Thank you for the proposal!
--
Uladzislau Rezki
prev parent reply other threads:[~2023-10-19 15:44 UTC|newest]
Thread overview: 7+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-10-16 17:30 [PATCH v3 1/1] rcu: Reduce synchronize_rcu() waiting time Uladzislau Rezki (Sony)
2023-10-17 1:28 ` Joel Fernandes
2023-10-17 14:06 ` Uladzislau Rezki
2023-10-18 14:32 ` Joel Fernandes
2023-10-18 17:36 ` Uladzislau Rezki
[not found] ` <20231017103342.1879-1-hdanton@sina.com>
2023-10-18 11:20 ` Uladzislau Rezki
[not found] ` <20231019114432.1995-1-hdanton@sina.com>
2023-10-19 15:44 ` Uladzislau Rezki [this message]
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=ZTFO17eNKvDNokpe@pc636 \
--to=urezki@gmail.com \
--cc=Neeraj.Upadhyay@amd.com \
--cc=boqun.feng@gmail.com \
--cc=frederic@kernel.org \
--cc=hdanton@sina.com \
--cc=linux-kernel@vger.kernel.org \
--cc=oleksiy.avramchenko@sony.com \
--cc=paulmck@kernel.org \
--cc=rcu@vger.kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).