From: Gireesh Hiremath <Gireesh.Hiremath@xxxxxxxxxxxx> Hi Krzysztof, >>> You wrote pretty long message explaining how the device works, but I >>> still do not see the answer to questions - why it cannot be part of >>> matrix keypad? >> >> Following are the difference between matrix keypad and Bosch keypad >> make us to add another keypad driver. >> >> matrix keypad: >> - By hardware schematic, a column GPIO line will intersect only >> with row GPIO lines, not with the other column GPIO lines >> - so, row and column GPIO property are fixed, because of this >> - key scanning work based on interrupt mode >> - and key press is determined based on setting column as output, >> row GPIO as input and set interrupt to monitor the changes in state, >> serve the key pressed in ISR >> >> Bosch keypad: >> - By hardware schematic column GPIO line can intersect with row GPIO line >> as well as other column GPIO lines >> - so, all GPIO act as row as well as column, because of this >> - key scanning based on polling mode >> - a key pressed is determined by setting one of GPIO line as output and >> other as input and poll for change in the state of input GPIO lines. >> Setting one of a GPIO line as output and remaining GPIO lines as input is on >> round robin bases. > >Which is still not the answer "why it cannot be part of matrix keypad?". >To me looks similar enough, although maybe not exactly superset of the >other. I have merged the Bosch keypad in to the matrix keypad and sending version v3 patches to support that. v3-0001-driver-input-matric-keypad-switch-to-gpiod.patch v3-0002-driver-input-matric-keypad-add-reduced-matrix-sup.patch v3-0003-dt-bindings-input-gpio-matrix-keypad-add-reduced-.patch > >>> >>> "It looks like this driver has smaller number of features than >>> matrix-keypad, so it should be integrated into the matrix-keypad. >>> matrix-keypad features are superset to this one." >>> >>> "But anyway this should be just merged into matrix-keypad. It's a >>> simpler set of that binding." >> >> This keypad driver specific to Bosch measuring tool or similar devices. >> Please let me know to send latest patch which resolves build warning >> and gpiod API support. > >That's a poor reason not to merge into existing driver... I am sorry, >but our entire Linux kernel concept is to integrate, not duplicate. If >each of vendors wanted their own feature, we would have unmanageable >monstrosity with millions of drivers doing almost the same... > > >Best regards, >Krzysztof > Best regards, Gireesh Hiremath
