[PATCH][next] Input: resistive-adc-touch: Fix uninitialized variable 'press'

[PATCH][next] Input: resistive-adc-touch: Fix uninitialized variable 'press'

[Colin King]

From: Colin Ian King <colin.king@canonical.com>

In the case where st->ch_map[GRTS_CH_PRESSURE] < GRTS_MAX_CHANNELS is false
and also st->ch_map[GRTS_CH_Z1] < GRTS_MAX_CHANNELS is false the variable
press is not initialized and contains garbage. This affects a later
comparison of press < st->pressure_min.  Fix this by initializing press
to 0 and allows us to also remove an else clause that sets press to 0.

Addresses-Coverity: ("Uninitialized scalar variable")
Fixes: 60b7db914ddd ("Input: resistive-adc-touch - rework mapping of channels")
Signed-off-by: Colin Ian King <colin.king@canonical.com>
---
 drivers/input/touchscreen/resistive-adc-touch.c | 4 +---
 1 file changed, 1 insertion(+), 3 deletions(-)

diff --git a/drivers/input/touchscreen/resistive-adc-touch.c b/drivers/input/touchscreen/resistive-adc-touch.c
index ea7dd9d2b2ac..744544a723b7 100644
--- a/drivers/input/touchscreen/resistive-adc-touch.c
+++ b/drivers/input/touchscreen/resistive-adc-touch.c
@@ -59,7 +59,7 @@ static int grts_cb(const void *data, void *private)
 {
 	const u16 *touch_info = data;
 	struct grts_state *st = private;
-	unsigned int x, y, press;
+	unsigned int x, y, press = 0;
 
 	x = touch_info[st->ch_map[GRTS_CH_X]];
 	y = touch_info[st->ch_map[GRTS_CH_Y]];
@@ -84,8 +84,6 @@ static int grts_cb(const void *data, void *private)
 		 */
 		if (Rt < GRTS_DEFAULT_PRESSURE_MAX)
 			press = GRTS_DEFAULT_PRESSURE_MAX - Rt;
-		else
-			press = 0;
 	}
 
 	if ((!x && !y) || (st->pressure && (press < st->pressure_min))) {
-- 
2.31.1

Re: [PATCH][next] Input: resistive-adc-touch: Fix uninitialized variable 'press'

[Dmitry Torokhov]

Hi Colin,

On Thu, Jun 03, 2021 at 11:08:09PM +0100, Colin King wrote:
> From: Colin Ian King <colin.king@canonical.com>
> 
> In the case where st->ch_map[GRTS_CH_PRESSURE] < GRTS_MAX_CHANNELS is false
> and also st->ch_map[GRTS_CH_Z1] < GRTS_MAX_CHANNELS is false the variable
> press is not initialized and contains garbage. This affects a later
> comparison of press < st->pressure_min.

If neither of the conditions is true, then st->pressure is also false,
and we will not be evaluating condition involving "press". However it is
impossible for the compiler/Coverity to figure this out, so I'll apply
the patch adjusting the description a bit.

Thanks.

-- 
Dmitry