The key press detection logic was using ptr instead of !ptr (where ptr is
the result of memchr searching for the new key in the old keys).

If ptr is not NULL, it means the key was already pressed in the previous
report, so it is not a new press. If ptr is NULL, it means the key was
not pressed before, so it is a new press.

Using ptr instead of !ptr caused new presses to be ignored in the first
poll cycle (only reported in the second cycle if still held, because
then it was in 'old'), and caused redundant press events to be sent
on every subsequent poll while the key was held.

Fix this by using !ptr for press detection, matching the release
detection logic.

Fixes: b11d2127c489 ("Input: add support for SEGA Dreamcast keyboard")
Assisted-by: Antigravity:gemini-3.5-flash
Signed-off-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>
---
 drivers/input/keyboard/maple_keyb.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/input/keyboard/maple_keyb.c b/drivers/input/keyboard/maple_keyb.c
index 3d5538dd4f23..623d7951405b 100644
--- a/drivers/input/keyboard/maple_keyb.c
+++ b/drivers/input/keyboard/maple_keyb.c

@@ -107,7 +107,7 @@ static void dc_scan_kbd(struct dc_kbd *kbd)
 		}
 		ptr = memchr(kbd->old + 2, kbd->new[i], 6);
 		code = kbd->new[i];
-		if (code > 3 && ptr) {
+		if (code > 3 && !ptr) {
 			keycode = kbd->keycode[code];
 			if (keycode) {
 				input_event(dev, EV_MSC, MSC_SCAN, code);

-- 
2.55.0.rc0.799.gd6f94ed593-goog