qemu-ppc
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Qemu-ppc] Status of mac99?


From: ardi
Subject: Re: [Qemu-ppc] Status of mac99?
Date: Sun, 8 Mar 2015 11:07:00 +0100

On Sun, Mar 8, 2015 at 9:59 AM, Alexander Graf <address@hidden> wrote:
>
>
> On 07.03.15 10:56, ardi wrote:
>> But, however, I just had an idea that
>> (I believe) would work for my purposes of debugging G5 code on an
>> emulated Tiger machine: would it be possible to boot mac99 with a G4
>> CPU that identifies itself as G4 but that can execute (32 bit)
>> specific G5 instructions?
>
> Which instructions do you have in mind? But yes, you can modify the code
> generator whichever way you like.

Unfortunately, I just realized that OS X knows about the minimum
instruction set of an executable (the "file" command won't tell it -it
just says if it's ppc or ppc64-, but "lipo" can tell if a binary was
compiled for G3, G4, or G5). So, I didn't try it yet, but I guess
Tiger will reject to run a 32bit G5 executable if the OS detects it's
running on a G4).

My interest is being able to run 32bit ppc executables that are
compiled for G5. I don't know what specific G5 instructions uses GCC
on Tiger when you tell it to build specifically for G5. In the worst
case it would be all G5 userland instructions except the
ppc64-specific stuff. IIRC, the Tiger marketing said that if you
optimized for G5, you got 64bit integer operations even when you
compiled for the 32bit ppc target).

Maybe my only hope would be to investigate how to fool OS X so that it
sees my G5 binaries as if they were compiled for G4, and then build
qemu for a G5 that identifies itself as a G4.

Either that, or wait for the U4 emulation, but that looks as a heavy
long term goal, and I believe there must be some quicker way of doing
what I need.

ardi



reply via email to

[Prev in Thread] Current Thread [Next in Thread]