linux-kernel - Re: [PATCH 2/4] Input: adc-keys - support types that aren't just keyboard keys

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <2411362.ElGaqSPkdT@workhorse>
Date: Mon, 15 Dec 2025 12:10:21 +0100
From: Nicolas Frattaroli <nicolas.frattaroli@...labora.com>
To: Dragan Simic <dsimic@...jaro.org>
Cc: alexandre.belloni@...tlin.com, conor+dt@...nel.org,
 devicetree@...r.kernel.org, dmitry.torokhov@...il.com, heiko@...ech.de,
 kernel@...labora.com, krzk+dt@...nel.org,
 linux-arm-kernel@...ts.infradead.org, linux-input@...r.kernel.org,
 linux-kernel@...r.kernel.org, linux-rockchip@...ts.infradead.org,
 robh@...nel.org
Subject:
 Re: [PATCH 2/4] Input: adc-keys - support types that aren't just keyboard
 keys

On Tuesday, 28 October 2025 22:32:09 Central European Standard Time Dragan Simic wrote:
> Hello Nicolas,
> 
> On Monday, June 30, 2025, 12:19:24 Nicolas Frattaroli wrote:
> > Instead of doing something like what gpio-keys is doing, adc-keys
> > hardcodes that all keycodes must be of type EV_KEY.
> > 
> > This limits the usefulness of adc-keys, and overcomplicates the code
> > with manual bit-setting logic.
> > 
> > Instead, refactor the code to read the linux,input-type fwnode property,
> > and get rid of the custom bit setting logic, replacing it with
> > input_set_capability instead. input_report_key is replaced with
> > input_event, which allows us to explicitly pass the type.
> 
> Thanks for this patch, it's indeed very useful!  Please see some
> comments below.
> 
> > Signed-off-by: Nicolas Frattaroli <nicolas.frattaroli@...labora.com>
> > ---
> >  drivers/input/keyboard/adc-keys.c | 16 ++++++++++++----
> >  1 file changed, 12 insertions(+), 4 deletions(-)
> > 
> > diff --git a/drivers/input/keyboard/adc-keys.c b/drivers/input/keyboard/adc-keys.c
> > index f1753207429db02ce6510e5ec0da9b24d9edb61d..339dd4d4a0842108da2c6136b1e0098cd1f6a3cd 100644
> > --- a/drivers/input/keyboard/adc-keys.c
> > +++ b/drivers/input/keyboard/adc-keys.c
> > @@ -19,12 +19,14 @@
> >  struct adc_keys_button {
> >  	u32 voltage;
> >  	u32 keycode;
> > +	u32 type;
> >  };
> >  
> >  struct adc_keys_state {
> >  	struct iio_channel *channel;
> >  	u32 num_keys;
> >  	u32 last_key;
> > +	u32 last_type;
> >  	u32 keyup_voltage;
> >  	const struct adc_keys_button *map;
> >  };
> > @@ -35,6 +37,7 @@ static void adc_keys_poll(struct input_dev *input)
> >  	int i, value, ret;
> >  	u32 diff, closest = 0xffffffff;
> >  	int keycode = 0;
> > +	u32 type = EV_KEY;
> >  
> >  	ret = iio_read_channel_processed(st->channel, &value);
> >  	if (unlikely(ret < 0)) {
> > @@ -46,6 +49,7 @@ static void adc_keys_poll(struct input_dev *input)
> >  			if (diff < closest) {
> >  				closest = diff;
> >  				keycode = st->map[i].keycode;
> > +				type = st->map[i].type;
> >  			}
> >  		}
> >  	}
> > @@ -54,13 +58,14 @@ static void adc_keys_poll(struct input_dev *input)
> >  		keycode = 0;
> >  
> >  	if (st->last_key && st->last_key != keycode)
> > -		input_report_key(input, st->last_key, 0);
> > +		input_event(input, st->last_type, st->last_key, 0);
> >  
> >  	if (keycode)
> > -		input_report_key(input, keycode, 1);
> > +		input_event(input, type, keycode, 1);
> 
> When EV_ABS is defined in the DT as the key type, which happens with
> joysticks and whatnot, separate handling is needed, by requiring the
> actual associated button values to be reported in the input_event()
> invocations, more specifically on the keypresses only.

Analogue control devices like joysticks are handled in adc-joystick.c.
adc-keys is specifically for resistor-ladder key devices. A driver or
binding for some absolute/relative positioning device that is not a
joystick should get its own binding first. Overloading the meaning
of the adc-keys binding and then adding functionality to the driver
is not the way to go.

I'll instead add code to only allow EV_KEY and EV_SW. Switches
function like keys for all intents and purposes, so they fit here.

> 
> That's also visible in the gpio_keys_gpio_report_event() function in
> drivers/input/keyboard/gpio_keys.c.
> 
> >  	input_sync(input);
> >  	st->last_key = keycode;
> > +	st->last_type = type;
> >  }
> >  
> >  static int adc_keys_load_keymap(struct device *dev, struct adc_keys_state *st)
> > @@ -93,6 +98,10 @@ static int adc_keys_load_keymap(struct device *dev, struct adc_keys_state *st)
> >  			return -EINVAL;
> >  		}
> >  
> > +		if (fwnode_property_read_u32(child, "linux,input-type",
> > +					     &map[i].type))
> > +			map[i].type = EV_KEY;
> 
> Going along with the remarks above, it will also be needed to read
> and record the values of "linux,input-value" DT properties here, and
> to extend the associated binding to define their presence.
> 
> >  		i++;
> >  	}
> >  
> > @@ -156,9 +165,8 @@ static int adc_keys_probe(struct platform_device *pdev)
> >  	input->id.product = 0x0001;
> >  	input->id.version = 0x0100;
> >  
> > -	__set_bit(EV_KEY, input->evbit);
> >  	for (i = 0; i < st->num_keys; i++)
> > -		__set_bit(st->map[i].keycode, input->keybit);
> > +		input_set_capability(input, st->map[i].type, st->map[i].keycode);
> >  
> >  	if (device_property_read_bool(dev, "autorepeat"))
> >  		__set_bit(EV_REP, input->evbit);
>