On Tue, Jan 11, 2022 at 02:59:28AM PST, Joel Stanley wrote:
On Wed, 5 Jan 2022 at 23:10, Zev Weiss <zev@xxxxxxxxxxxxxxxxx> wrote:
This is a half-width, single-socket Epyc server board with an AST2500
BMC. This device tree is sufficient for basic OpenBMC functionality,
but we'll need to add a few more devices (as driver support becomes
available) before it's fully usable.
Signed-off-by: Zev Weiss <zev@xxxxxxxxxxxxxxxxx>
Reviewed-by: Joel Stanley <joel@xxxxxxxxx>
Thanks!
Have you considered using the openbmc gpio naming scheme for the
gpio-line-names?
I looked at it, but decided not to for a few reasons:
- For systems that are in the early stages of a porting effort (like
this one currently is), I'm still referring to hardware schematics
fairly often, and using the same identifiers in software that are
used in the schematics simplifies things by avoiding an extra
translation step between the two.
- Most of the GPIO-related userspace components (that I'm dealing with
anyway, e.g. x86-power-control and host-error-monitor) already have
their own GPIO line-name configuration/remapping mechanisms that need
to be set up anyway.
- There's a solid mix of GPIOs that would be covered by the naming
guidelines and others that aren't; having a mix of the two styles
seems a bit awkward to me.
That said, I sympathize with the motivation behind it and I'm not
vehemently opposed on the whole, so if there's a strong preference to
follow that scheme I could probably be talked into changing it.
Zev
