Re: SPR implementation request.

[BALATON Zoltan]

Hello,

On Sun, 21 Jun 2020, Jd Lyons wrote:
> Would some knowledgeable person implement a few SPR’s to get OS X booting on 
> our emulated G5 CPU’s
>
> qemu-system-ppc64 -M mac99 -m 1024 -hda ~/Leopard.img -serial stdio  -d 
> guest_errors
>
> > > =============================================================
> > > OpenBIOS 1.1 [Oct 28 2019 17:09]
> > > Configuration device id QEMU version 1 machine id 3
> > > CPUs: 1
> > > Memory: 1024M
> > > UUID: 00000000-0000-0000-0000-000000000000
> > > CPU type PowerPC,970FX
> milliseconds isn't unique.
> > > switching to new context:
> > > call-method slw_update_keymap failed with error ffffffdf
> > > call-method slw_update_keymap failed with error ffffffdf
> Trying to write invalid spr 276 (0x114) at 00000000000b6db4
> Trying to read invalid spr 277 (0x115) at 00000000000b6db8
> Trying to read invalid spr 276 (0x114) at 00000000000b6dbc
> Trying to write invalid spr 277 (0x115) at 00000000000b6dd8
> Trying to write invalid spr 276 (0x114) at 00000000000b6ddc
> Trying to read invalid spr 276 (0x114) at 00000000000b6de0
> Trying to write invalid spr 277 (0x115) at 00000000000b6e8c
> Trying to write invalid spr 276 (0x114) at 00000000000b6e90
> Trying to read invalid spr 276 (0x114) at 00000000000b6e94

Before implementing these have you checked in the G5 manual what are they 
and if they are actually needed? I know we have similar messages with 
MacOS on mac99 and it was said they are harmless because those don't 
actually exist on G4 but MacOS probes for older CPUs by accessing them. 
Could this be a similar case here?

> invalid/unsupported opcode: 00 - 00 - 00 - 00 (00000000) 0000000000000000 0
> Trying to write invalid spr 304 (0x130) at 0000000000003e14
> Trying to read invalid spr 304 (0x130) at 0000000000003e38
> invalid/unsupported opcode: 00 - 00 - 00 - 00 (00000000) 0000000000000008 0
> invalid/unsupported opcode: 00 - 00 - 00 - 00 (00000000) 0000000000000008 0

This looks like it jumps off somewhere it should not go, maybe a spurious 
exception or it would want to use a value read from one of those SPRs? 
Maybe enable a few more debug options like -d int,in_asm (or look for 
debugs for exceptions in target/ppc) and try to find out what actually 
happens then you're sure what needs to be implemented. It may be missing 
SPR or something else.

Regards,
BALATON Zoltan