Closed neuschaefer closed 2 years ago
We can get it clarified but I believe every one of our definitions is as an input. Typically we describe signals as from the BMCs perspective, so calling an output a button wouldn’t make much sense.
Thanks.
Well, as I mentioned, there are BMC-outputs that act just like virtual buttons, "pressed" by code on the BMC. Hence my confusion. To be more explicit: I've seen BMC-output signals that act like the familiar PC power button: "Press" shortly to power on or request power-off, "press" for 4-5 seconds to force power-off.
Yep, I'm aware. Today most of that is buried in a repository called 'x86-power-control' (even though it also supports some other architectures as long as they use the "standard" power sequence). There is a pending commit to better document those GPIOs you might want to take a look at. https://gerrit.openbmc-project.xyz/c/openbmc/docs/+/49863
Attempted a clarification:
Attempted a clarification:
* https://gerrit.openbmc-project.xyz/c/openbmc/docs/+/51527 gpio-naming: clarify button definition [NEW]
This answers the question for me. Thanks!
Currently, the Device Tree GPIO Naming in OpenBMC document is rather short on words about buttons. However, I think there is potential for ambiguity here.
In my experience, a GPIO labeled a "button" could mean either of two things:
I think both cases can occur in hardware designs and are useful, but they are quite different. The DT GPIO name document should, IMHO, be extended to clarify what
*-button
currently means, and potentially also to accomodate the other meaning of "button".