Sorry, been MIA for a bit with $WORK

What's the path forwards for this? This is still fixing and outstanding bug for me but requires manual effort to apply.

Very exciting! Probably want to split the ports out from the Mk to keep the initial diff simpler though?

Dmitry - appologies for not replying, I got distracted by The Witcher 3 being playable again

Added the same reorderings to sys/arm64/arm64/memcpy.S

At least in Linux, the drm drivers create a fb0 device and things like mesa may talk directly to that via memcpy.

@allanjude not necessary, but "nice to have" as it resolves an outstanding bug in sysutils/hw-probe on ACPI-based arm64 systems

Ping - is there any work left to do here or can this be merged and closed?

Unless they are going to completely handle all DesignWare quirks in firmware this doesn't seem correct…

SR changed the HID and CID which is why the new version didn't actually attach properly: https://github.com/SolidRun/edk2-platforms/commit/7de27183f473debe05e77cf0f0aa4b14479e4152

I have not tested stability or actual bandwidth yet, but on my Honeycomb I am finally seeing USB3.0 devices negotiate to superspeed. This seems a bit inconsistent however as sometimes superspeed devices still only negotiate to highspeed, but my initial testing leads me to believe that my cabling is at least partially at fault.

You are correct - D28737 should accomplish the same thing and is already accepted. This can be rejected in favor of that.

Related bug here as well: https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=253606

Using AcpiOsGetRootPointer() instead of FADT or SPCR

You're right, that's much easier to read and makes more sense long term. And I just retested and it works fine in all combinations of serial and video.

We could do that, but it seems like more code than necessary to do the same exact thing.

AcpiOsGetRootPointer just wraps resource_long_value in sys/bus.h which we already include in sys/arm64/arm64/machdep.c. Since that returns 0 when successful, what about something like this?

long acpi_root;
has_acpi = (resource_long_value("acpi", 0, "rsdp", &acpi_root) == 0);

Already accepted, but just wanted to confirm this solved interrupt mapping problems we were having on LX2160a.