Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
MIME-Version: 1.0
References: <20200831032006.1019978-1-warthog618@gmail.com>
 <20200831032006.1019978-10-warthog618@gmail.com> <CAMpxmJW9EC_ocs8dA3j+QPu8sW4Na9WEvPYEUZjfxk4-ojf+kA@mail.gmail.com>
 <20200903080909.GA17108@sol>
In-Reply-To: <20200903080909.GA17108@sol>
From:   Bartosz Golaszewski <bgolaszewski@baylibre.com>
Date:   Thu, 3 Sep 2020 10:13:43 +0200
Message-ID: <CAMpxmJWMxkg14Jx1VyqcX_rhGWqFo4YTfWtDEQe-NUTtqx2kVA@mail.gmail.com>
Subject: Re: [PATCH v6 09/20] gpiolib: cdev: support edge detection for uAPI v2
To:     Kent Gibson <warthog618@gmail.com>
Cc:     LKML <linux-kernel@vger.kernel.org>,
        linux-gpio <linux-gpio@vger.kernel.org>,
        Linus Walleij <linus.walleij@linaro.org>
Content-Type: text/plain; charset="UTF-8"
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Thu, Sep 3, 2020 at 10:09 AM Kent Gibson <warthog618@gmail.com> wrote:
>
> On Thu, Sep 03, 2020 at 09:55:22AM +0200, Bartosz Golaszewski wrote:
> > On Mon, Aug 31, 2020 at 5:22 AM Kent Gibson <warthog618@gmail.com> wrote:
> > >
> > > Add support for edge detection to lines requested using
> > > GPIO_V2_GET_LINE_IOCTL.
> > >
> > > The edge_detector implementation is based on the v1 lineevent
> > > implementation.
> > >
>
> [snip]
>
> > > +        * close in time as possible to the actual event.
> > > +        */
> > > +       line->timestamp = ktime_get_ns();
> > > +
> > > +       if (lr->num_lines != 1)
> > > +               line->req_seqno = atomic_inc_return(&lr->seqno);
> >
> > Do we never increase req_seqno for a single line?
> >
>
> For a single line we just use line_seqno for both to avoid the
> atomic_inc here and so reduce the time spent in the ISR and any
> SMP sync overheads.
>
> As per the comment in struct linereq:
>  * @seqno: the sequence number for edge events generated on all lines in
>  * this line request.  Note that this is not used when @num_lines is 1, as
>  * the line_seqno is then the same and is cheaper to calculate.
>
> Cheers,
> Kent.

I should have RTFC I guess. :)

Bart