On Powermacs I use linuxppc to run qemu.
For latest 5.2 release I had to resolve many new dependencies to let it compile, anyway it works but it's too slow in sam460ex and powerpc emulation.
With pegasos2 board "should" be finally possible to virtualize cpu via KVM and run at full speed.

NB:
Both powermacs and pegasos2 share the same cpu (book3s) so virtualizations is possible, sadly with sam460ex the emulated cpu belongs to a different family (book3e) and thus it had to be emulated with a huge slowdown.

> do you mean there will be no QEMU emulator dropdowns in speed
> even if it will use the host's CPU? I mean, it still will be slower than
> native IMHO. [...] EDIT: I read that KVM can run a guest OS at 95-98%
> of the host’s native performance

Running a guest OS in QEMU KVM-PR mode makes the user mode instructions run directly on the host CPU and the supervisor/privileged mode instructions run emulated. Thus, the overall performance depends on the distribution of the executed instructions, i.e. the less supervisor/privileged instructions the better the performance.

> QEMU should be run from something, so it will mean MacOS
> probably necessary too.

I'm not sure that current/future QEMU versions can run on 11½ years old MacOS.

> it probably lucks audio driver and RadeonR92xx driver (with all compositing
> things working) [...]? Or there already some audio/video drivers support being done?

Audio: QEMU can emulate Ensoniq AudioPCI ES1370 chipset (deployed on Sound Blaster PCI128 card), which OS4 has drivers for. So if QEMU can't emulate the Pegasos II onboard audio, you can always insert a virtual SB128 sound card into the virtual Pegasos II.
Graphics: Correct, no Radeon or compositing etc. yet. However, I think it should be possible to insert a virtual SM501/SM502 graphics card into the virtual Pegasos II (as had to be done for running MorphOS on QEMU PowerMac3,1/mac99 emulation before the Rage 128 Pro emulation emerged).

Qemu may gain Pegasos II support, whereas the kernel may loose Pegasos II support (klick).

Another reason to go on pegasos2 integration in qemu.
Virtual board will never broke and will improve speed, compatibility and features over years.
Better than a good old wine

> the kernel may loose Pegasos II support

THE kernel? Ah, fortunately just the Linux kernel, not the MorphOS/PPC kernel (Quark) ;-)

> klick

I read this a little differently:

"Some of these look like they are no longer actively maintained or used, but I'm not doing anything about those unless the maintainers would like me to: [...]
* powerpc/chrp (32-bit rs6000, pegasos2): last updated in 2009"

As I read it, this means that support will stay in for the time being unless the maintainer (who?) wants it removed. The supporters are already starting to line up :-)
Overall, I don't think that the future (non-)ability to run a then-current Linux on QEMU-emulated/virtualized Pegasos II has a relevance. Why emulate/virtualize a Pegasos II in QEMU to run Linux on it when QEMU is already running on Linux on the host?

>> the kernel may loose Pegasos II support

> Another reason to go on pegasos2 integration in qemu.

This I don't get. Why should an inability to run new Linux versions on a system be "another reason" to have that system emulated?

Time always consumes and try to kickoff all things what are considered obsolete.
I understand but really dont' like it.
The stop of drop of new linux kernels for pegasos2 is an example in that direction and IMHO a sort of mistake.
It can be rebalanced in some way by adding pegasos2 in qemu, letting it still survive in a next future.

Quote:
True.

Quote:
True. But no one lined up for the AOne or the Peg2 yet, only for some of the ARM boards and for the Alpha.

My guess is that this 'deprecation request' will be made again in some time, annually (?) perhaps. PPC Maintainer btw is Michael Ellerman. Most probably all CCed people in the 1st post are arch maintainers. Recently PowerPC 601 support got thrown out, which I don't miss - but some people will probably miss the AOne or Peg2.

Quote:
A good use of such QEMU-emulated machines is for bisecting/working on bugs on specific archs. Not every maintainer of some software has ppc/ppc64/arm/aarch64/mips/etc. hardware lying around to test his stuff on, but he can use QEMU.

This has been quite handy for some Linux bugs I reported on ppc/ppc64. The devs often use ppc64 workstations in Little Endian mode and may not be able to reproduce many bugs which happen on my PowerMac G4. Most certainly not when it's about (Linux) kernel bugs.

The question is how much longer the Peg2 'hardware' will stay in QEMU when there is no Linux kernel for the Peg2 to boot? I dunno. Are there existing conventions for that case?

>> The supporters are already starting to line up :-)

> no one lined up for the AOne or the Peg2 yet [...]. [...]
> some people will probably miss the AOne or Peg2. ;-)

Maybe you should have clicked my link to encounter one of them. Here's another supporter.

> A good use of such QEMU-emulated machines is for bisecting/working
> on bugs on specific archs. Not every maintainer of some software has
> ppc [...] hardware lying around to test his stuff on, but he can use QEMU.

I doubt Pegasos II emulation is needed for this purpose as QEMU already emulates well other Linux-supported PPC32 hardware, among them also other G4-based hardware (like PowerMac3,1/mac99). Zoltan Balaton has been making no secret of his motivation to create a Pegasos II emulation, and it's not Linux-related. Furthermore, bug hunting of Linux software doesn't require QEMU system emulation as QEMU user-mode emulation is to be preferred for this.
Bottom line: I'm still not convinced regarding the correlation between QEMU's Pegasos II emulation and the (non-)ability of the Pegasos II (emulation) to run Linux.

> The question is how much longer the Peg2 'hardware' will stay in
> QEMU when there is no Linux kernel for the Peg2 to boot? I dunno.
> Are there existing conventions for that case?

I'm not aware of Linux capability being a prerequisite for a machine to be allowed in QEMU. Zoltan always recommends the most recent MorphOS ISO image for testing the emulation anyway :-)

Quote:
Not regular software bugs yes, but Linux kernel bugs.

Didn't want to convince you. Anyway, meanwhile several devs stood up for the Peg2 so it is safe to stay for the time being.

Quote:
Ah, didn't knew that. Nice!

>> QEMU already emulates well other Linux-supported PPC32 hardware,
>> among them also other G4-based hardware (like PowerMac3,1/mac99).
>> [...] Furthermore, bug hunting of Linux software doesn't require QEMU
>> system emulation as QEMU user-mode emulation is to be preferred for this.

> Not regular software bugs yes, but Linux kernel bugs.

As said, that's what long-existing PowerMac3,1/mac99 emulation or other PPC32 system emulations are perfectly fine for.

> meanwhile several devs stood up for the Peg2
> so it is safe to stay for the time being.

Yes, Arnd Bergmann's summary for "powerpc/chrp (32-bit rs6000, pegasos2)" and others:

"a couple of users replied, but no architecture maintainer added any information, so I won't take any action."
https://www.lkml.org/lkml/2021/1/13/746

Update:

> the relevant ML messages:
> https://lists.gnu.org/archive/html/qemu-ppc/2021-01/msg00072.html
> https://lists.gnu.org/archive/html/qemu-ppc/2021-01/msg00073.html
> https://lists.gnu.org/archive/html/qemu-ppc/2021-01/msg00074.html

Next steps:

https://lists.gnu.org/archive/html/qemu-ppc/2021-03/msg00638.html
https://lists.gnu.org/archive/html/qemu-ppc/2021-03/msg00639.html
https://lists.gnu.org/archive/html/qemu-ppc/2021-03/msg00641.html
https://lists.gnu.org/archive/html/qemu-ppc/2021-03/msg00644.html
https://lists.gnu.org/archive/html/qemu-ppc/2021-03/msg00645.html

With those, QEMU's Pegasos II emulation seems to contain the essential parts now.

Interesting comment from the author:

"The whole pegasos2 seems to be like a MIPS board with a PPC CPU even down to the firmware which looks quite like PMON just has SmartFirmware instead of the commands of PMON. So whoever designed it, might have taken inspiration from some MIPS hardware."
https://lists.gnu.org/archive/html/qemu-ppc/2021-02/msg00252.html


Edit: changed links to more recent versions of the patches

[ Edited by Andreas_Wolf 25.03.2021 - 22:39 ]

> All should be ready for next QEmu 6.0 release.

Seems it won't come before v6.1:
https://lists.gnu.org/archive/html/qemu-ppc/2021-03/msg00546.html
https://lists.gnu.org/archive/html/qemu-ppc/2021-03/msg00686.html


Edit: added another link
Edit2: QEMU v6.1 released

[ Edited by Andreas_Wolf 26.08.2021 - 13:11 ]