Re: [Qemu-ppc] MorphOS 4.x on QEMU

[BALATON Zoltan]

On Fri, 14 Mar 2014, Mark Cave-Ayland wrote:
> Ah on the basis of this I think I understand the alerts now:
>
> > > > Alert: SYS_MMUAddPage: Page 0x80000 EndPage 0x81000 already exists
> > > > Alert: SYS_MMUAddPage: Page 0xf2000 EndPage 0xf2001 already exists
>
> Based upon a page size of 0x1000, my guess is that this is trying to tell you 
> that ranges 0x80000000-0x81000000 and 0xf2000000-0xf2001000 already have 
> virtual to physical address mappings; the former seems to correspond with the 
> VGA BAR0 (VGA vram) region which is likely mapped by OpenBIOS while the 
> latter which is just a single page could be control registers for the 
> UniNorth PCI host bridge.

The latter seems to be related to the serial card I've added for 
debugging. If running with only VGA it dumps the first alert in the memory 
debug buffer and running with the serial card gives only the second. The 
serial card has this BAR0: I/O at 0x0400 [0x0407] which probably end up on 
the isa_mmio from f2000000. But unfortunately removing these cards is not 
enough to prevent the crash even if it's enough to prevent the alerts. 
Something else is failing here for some reason which I yet need to find 
out.

By the way, for the serial card OpenBIOS prints this at start:
> > Cannot manage '16550 serial controller' PCI device type '<NULL>':
> >  1b36 2 (7 0 2)

I add it with these options:
-chardev file,id=tS0,path=debug-serial.out -device pci-serial,chardev=tS0

but it is detected and used by MorphOS nevertheless.

> One thing that does confuse me is how does MorphOS know these mapping exist?

Yes, that's mistery. For sure not by looking at the translations in the 
CPU properties because it seems to skip that. Instead it seems to look at 
the properties of the pci cards and maybe reconstructs mappings from that. 
It reads these properties: #address-cells #size-cells ranges bus-range reg 
assigned-addresses and those related to irq mapping too.

> Since the OS would presumably manage its own MMU hash entries then maybe 
> these entries are locked in the MMU TLB?

Do you mean it probably has built in knowledge about mappings already 
present on a certain model of Apple hardware? I'm not sure I understand 
this sentence correctly.

> Perhaps as a starting point you could try moving the VGA card to 0x90000000 
> similar to the sample device trees you listed in case the MorphOS guys have 
> hardcoded this region? It seems that the ATI rage card is on a separate 
> (presumably AGP) bus so you would need to create another PCI bus/PCI address 
> space to model this in the same way.

I've seen some commented out code in the mac_newworld file. I'll try to 
enable it to see if that works but I'm afraid I don't really know what am 
I doing there.

Regards,
BALATON Zoltan