Received-SPF: pass (google.com: domain of linux-kernel+bounces-171767-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:40f1:3f00::1 as permitted sender) client-ip=2604:1380:40f1:3f00::1;
Precedence: bulk
MIME-Version: 1.0
References: <20240507121346.16969-1-brgl@bgdev.pl> <597f5da2-71be-4144-a570-fdc4f06c4cc6@paulmck-laptop>
 <CAMRc=MexihBpLBcY-8aX06buUYmtE07ZpkMq0Ho3jrHb6VE7Sw@mail.gmail.com> <1e9042a9-f8ba-4ca9-a5e3-5eed1a02e1ca@paulmck-laptop>
In-Reply-To: <1e9042a9-f8ba-4ca9-a5e3-5eed1a02e1ca@paulmck-laptop>
From: Bartosz Golaszewski <brgl@bgdev.pl>
Date: Tue, 7 May 2024 18:03:41 +0200
Message-ID: <CAMRc=Mc0FLXtmiD0GhYbc8Wna+rAM+9dUF6JrdtZDQttBMgXBw@mail.gmail.com>
Subject: Re: [PATCH] gpiolib: fix the speed of descriptor label setting with SRCU
To: paulmck@kernel.org
Cc: Kent Gibson <warthog618@gmail.com>, Linus Walleij <linus.walleij@linaro.org>, 
	linux-gpio@vger.kernel.org, linux-kernel@vger.kernel.org, 
	Bartosz Golaszewski <bartosz.golaszewski@linaro.org>, Neil Armstrong <neil.armstrong@linaro.org>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

On Tue, May 7, 2024 at 6:01=E2=80=AFPM Paul E. McKenney <paulmck@kernel.org=
> wrote:
>
> On Tue, May 07, 2024 at 04:48:04PM +0200, Bartosz Golaszewski wrote:
> > On Tue, May 7, 2024 at 4:24=E2=80=AFPM Paul E. McKenney <paulmck@kernel=
org> wrote:
> > >
> > > On Tue, May 07, 2024 at 02:13:46PM +0200, Bartosz Golaszewski wrote:
> > > > From: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>
> > > >
> > > > Commit 1f2bcb8c8ccd ("gpio: protect the descriptor label with SRCU"=
)
> > > > caused a massive drop in performance of requesting GPIO lines due t=
o the
> > > > call to synchronize_srcu() on each label change. Rework the code to=
 not
> > > > wait until all read-only users are done with reading the label but
> > > > instead atomically replace the label pointer and schedule its relea=
se
> > > > after all read-only critical sections are done.
> > > >
> > > > To that end wrap the descriptor label in a struct that also contain=
s the
> > > > rcu_head struct required for deferring tasks using call_srcu() and =
stop
> > > > using kstrdup_const() as we're required to allocate memory anyway. =
Just
> > > > allocate enough for the label string and rcu_head in one go.
> > > >
> > > > Reported-by: Neil Armstrong <neil.armstrong@linaro.org>
> > > > Closes: https://lore.kernel.org/linux-gpio/CAMRc=3DMfig2oooDQYTqo23=
W3PXSdzhVO4p=3DG4+P8y1ppBOrkrJQ@mail.gmail.com/
> > > > Fixes: 1f2bcb8c8ccd ("gpio: protect the descriptor label with SRCU"=
)
> > > > Suggested-by: Paul E. McKenney <paulmck@kernel.org>
> > > > Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org>
> > >
> > > Looks good to me!
> > >
> > > Acked-by: Paul E. McKenney <paulmck@kernel.org>
> > >
> > > One semi-related question...  Why the per-descriptor srcu_struct?
> > >
> > > If the srcu_struct was shared among all of these, you could just do o=
ne
> > > synchronize_srcu() and one cleanup_srcu_struct() instead of needing t=
o
> > > do one per gdev->desc[] entry.
> > >
> > > You might be able to go further and have one srcu_struct for all the
> > > gpio devices.
> > >
> > > Or did you guys run tests and find some performance problem with shar=
ing
> > > srcu_struct structures?   (I wouldn't expect one, but sometimes the
> > > hardware has a better imagination than I do.)
> > >
> >
> > I guess my goal was not to make synchronize_srcu() for descriptor X
> > wait for read-only operations on descriptor Y. But with that gone, I
> > suppose you're right, we can improve this patch further by switching
> > to a single SRCU descriptor.
> >
> > I'll send a v2.
>
> My guess is that a separate patch for each of the two changes would be
> best, but I must defer to you guys on that.
>
>                                                         Thanx, Paul

Ok, makes sense, I'll apply this and send a follow-up.

Bart