Change in coreboot[master]: mb/google/brya: Reset XHCI controller while preparing for S5

[Subrata Banik (Code Review)]

Attention is currently required from: Nico Huber, Maulik V Vaghela, Paul Menzel, Sridhar Siricilla, Arthur Heymans, Nick Vaccaro, Eric Lai, Felix Held.
Subrata Banik has posted comments on this change. ( https://review.coreboot.org/c/coreboot/+/63552 )

Change subject: mb/google/brya: Reset XHCI controller while preparing for S5
......................................................................

Patch Set 6:

(1 comment)

Commit Message:

https://review.coreboot.org/c/coreboot/+/63552/comment/c3138fc7_e9e4fbb7 
PS4, Line 37: SMI handler
> > > Sorry, thinking error on my part. coreboot of course does not run anymore during shutdown (besides the installed handlers).
> > >
> > > But, why can’t this be worked around/fixed in the OS driver?
> >
> > fixing in FW is comparatively easier 😊 and could apply MB specific W/As as well.
>
> How so?

- the fixes in FW are effective in timely manner where one can't land those changes easily in kernel.
- FW qualification is less costly compared to OS.
- we could add device specific W/A as we have mainboard smi handler (originally designed for the same).
> You had to add a whole new API function (and should have added
> a new file for it, AFAICS; also, you use PCH_DEV_XHCI which contains a
> workaround for NULL pointers which should be fixed). In Linux, OTOH,
> you probably just need to set a correct quirk flag.
>
> Doing it in SMM makes OS developers expect that they don't have to
> handle it, even provides them an excuse not to do it, and that weakens
> the ecosystem. IIRC, SMI handlers were originally added to coreboot
> because of a similar story. Now we risk to repeat the 10y+ old mistake.

Getting rid of SMI is really a task IMO.