MATRIX_KB in 2.8.0

Hi guys,

I control my WABECO F1410 LF HS with LinuxCNC and MESA 7i76e + 7i85 + 7i73 and am actually very satisfied.
I have loaded a 4x8 matrix_kb for the 7i73. Works in principle, but I think I found a bug in version 2.8.0 pre1 6085:

The output signal of the key.r1c1 button is 'ON' after starting LinuxCNC even though the button is not pressed. If I press it once and release it again, the output of key.r1c1 goes 'LOW' as expected and behaves like all other outputs. Can this be an error during initialization?
I have tested this configuration with version 2.7.14. Here this error does not occur and everything behaves as expected.
I know that LinuxCNC 2.8 is not released yet. Can someone confirm this bug before I downgrade to 2.7.15?

# >>>> Keypad >>>
loadrt matrix_kb config=4x8
addf matrix_kb.0 servo thread -2
net keypad hm2_7i76e.0.7i73.0.2.keycode => matrix_kb.0.keycode
# <<< Keypad #<

Best regards

I cannot duplicate this in master
Your might want to upgrade to the latest 2.8 version

Thanks for trying and the quick answer.

I have now updated to 2.8.0-pre1-6125 and got the same error.
I had actually excluded the 7i73 as the cause, since it behaves correctly with the same hardware and version 2.7.14.
I have also tried 10k pull-up resistors on the scan inputs of the 7i73, without success.

I start LinuxCNC and add all key-outputs of the 7i73 to the observation window. All are FALSE except r1c1. If I now press r1c1, it stays HIGH and changes to LOW when I release it. After that it behaves normally. When I restart LinuxCNC it is the same game.

Does anyone have any idea what else I could try? Especially after PCW could not reproduce this error I am helpless.

Thanks

I just created a new configuration with pncconf and this is also the problem.

I have attached the configurations of the MESA boards.

Your thread order is not right, the hardware read should be the first addf
and the hardware write should be the last, something like this:

Ohhh, very good point, thank you very much. I changed that... but unfortunately it doesn't fix anything on this button problem

Do you have the matrix_kb addf between read and write?

Jep, rearranged the HAL a little bit for better readability .

Hmm I just tried 2.8 and don't see a problem either
Do all the keys work?
(4x8 mode scanner columns are P1 pins 7,8,9,10)

Yes, all keys are wirking fine. Key r1c1 works too, but it starts in pressed state.
If I use 2.7.14 key r1c1 starts in inactive state like all the others do. Because of that i thought it is an issue in 2.8.