Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
MIME-Version: 1.0
References: <20221022220508.37489-1-maurinbe@gmail.com>
In-Reply-To: <20221022220508.37489-1-maurinbe@gmail.com>
From:   Roderick Colenbrander <thunderbird2k@gmail.com>
Date:   Sat, 22 Oct 2022 21:00:02 -0700
Message-ID: <CAEc3jaDjGBUspf5zPUfjPKgP+nk4zVFqJ0pgR8LAmX35qmx+fA@mail.gmail.com>
Subject: Re: [PATCH] input/xpad: LED controllable through input events
To:     Benoit Maurin <maurinbe@gmail.com>
Cc:     Dmitry Torokhov <dmitry.torokhov@gmail.com>,
        linux-input@vger.kernel.org, linux-kernel@vger.kernel.org
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

Hi Benoit,

Thanks for your patch. I'm thinking out loud whether this is a
direction we want to go in regards to EV_LED. I vaguely recall some
discussions a while ago that EV_LED was really from a different era
for legacy reasons (keyboard LEDs). Since the introduction of the LED
framework that was really the way to go.

Thanks,
Roderick

On Sat, Oct 22, 2022 at 3:08 PM Benoit Maurin <maurinbe@gmail.com> wrote:
>
> (EV_LED, LED_MISC, #VALUE) can now be used to control leds on the
> xpad gamepad (was only possible through /sys/class/leds/xpad0/brightness)
> with permissions of /dev/input/xxx
>
> To test the code (xpad can be compiled out-of-tree with some slight
> tweaks):
>
> ```
> import evdev
> device = evdev.InputDevice('/dev/input/event15') # not js0
> device.set_led(8, 2)
> device.set_led(8, 0) # this won't be delivered
> device.set_led(8, 16) # must do this instead
> device.set_led(8, 15)
> ```
>
> Signed-off-by: Benoit Maurin <maurinbe@gmail.com>
> ---
>  drivers/input/joystick/xpad.c | 24 ++++++++++++++++++++++++
>  1 file changed, 24 insertions(+)
>
> diff --git a/drivers/input/joystick/xpad.c b/drivers/input/joystick/xpad.c
> index 2959d80f7..fcf4d2c8f 100644
> --- a/drivers/input/joystick/xpad.c
> +++ b/drivers/input/joystick/xpad.c
> @@ -1646,6 +1646,7 @@ static int xpad_led_probe(struct usb_xpad *xpad)
>                 goto err_free_mem;
>         }
>
> +       input_set_capability(xpad->dev, EV_LED, LED_MISC);
>         snprintf(led->name, sizeof(led->name), "xpad%d", xpad->pad_nr);
>         led->xpad = xpad;
>
> @@ -1824,6 +1825,28 @@ static void xpad_deinit_input(struct usb_xpad *xpad)
>         }
>  }
>
> +static int xpad_event(struct input_dev *dev, unsigned int type,
> +                     unsigned int code, int value)
> +{
> +       struct usb_xpad *xpad = input_get_drvdata(dev);
> +
> +       if (type != EV_LED || xpad->led == NULL)
> +               return 0;
> +       xpad_send_led_command(xpad, value);
> +       xpad->led->led_cdev.brightness = value;
> +       /* Bit clearing is necessary otherwise two events with
> +        * different non-null values will deliver only the first one.
> +        * To work around this, we clear the bit to indicate that the
> +        * current value is zero. The downside is that events with zero
> +        * value won't be delivered. It's not a big deal since a value of
> +        * 16 can be sent which is the same as 0
> +        * See xpad_send_led_command, command %= 16
> +        */
> +
> +       clear_bit(code, xpad->dev->led);
> +       return 0;
> +}
> +
>  static int xpad_init_input(struct usb_xpad *xpad)
>  {
>         struct input_dev *input_dev;
> @@ -1851,6 +1874,7 @@ static int xpad_init_input(struct usb_xpad *xpad)
>                 input_dev->open = xpad_open;
>                 input_dev->close = xpad_close;
>         }
> +       input_dev->event = xpad_event;
>
>         if (!(xpad->mapping & MAP_STICKS_TO_NULL)) {
>                 /* set up axes */
> --
> 2.38.1
>