09:10 <maz> robher: m1 mini doesn't require anything outside of upstream.

14:17 <Clem> Hello, do you know if testing low speed interface drivers' (uart, spi, i2c) in loopback is sufficient ? I am trying to understand how people are working to validate their developments. + with growing test automation framework is there any mean to test such interfaces ? What I found during my researches is very limited and I think I need some advice from community in this area. Thx !

14:21 <broonie> Clem: For SPI loopback testing is probably fine, though it might be hard to arrange on off the shelf hardware. I2C I don't think physically loops back that easily.

14:22 <broonie> In general I think most people are testing the controllers as part of a full system (ie, if they can control devices connected over the relevant interfaces or use the UART as a console or whatever).

14:55 <Clem> broonie: I understand that in a whole system you test the communication between each component you integrate but when it come to test a SPI/UART/I2C driver you want to upstream and you don't know with wich component it will be in communication with, what is the best way to do so ?

14:57 <broonie> Clem: If you're a controller vendor testing you probably already have some setup for hardware testing?

14:59 <broonie> To a large extent the answer is "whatever makes you comfortable".

15:03 <Clem> broonie: I am trying to identify the best way to do so and try to find the state of the art in term of test for this area. I just discovered about syzkaller/syzbot but it does not cover any of my cases for communication interface so trying to figure out what is the best

15:07 <broonie> I think the "whatever test system you actually have" is broadly the state of the art here, for SoC vendors that'll often be their reference board.

15:11 <Clem> Yes indeed, thank you for your help !