Update:

> Project Qmiga:
> https://lists.gnu.org/archive/html/qemu-ppc/2019-06/msg00105.html
> http://qmiga.codeberg.page (was: http://qmiga.osdn.io)

Current status of Pegasos II emulation:

"In its current state the board firmware ROM runs and also can boot operating systems but only MorphOS >=2.6 can usefully boot yet. Linux [...] also boots but it does not seem to have driver for the VGA card (expects Radeon VGA) so can only be used over serial console. Similar to Linux, AmigaOS4.1 PegasosII version also boots but can't be seen due to no graphics driver. Maybe the driver from Sam460EX version could be used somehow but I did not attempt that. A lot of MV64361 might still be missing and there are also missing pieces of VT8231 but not sure how much of these are needed."
https://codeberg.org/qmiga/pages/wiki/SubprojectPegasos2
(was: https://osdn.net/projects/qmiga/wiki/SubprojectPegasos2)


Edit:

"It is basically working now, can boot MorphOS (and also AmigaOS but that has no display driver for VGA so can't be seen [...]) but to be able to upstream it I'll need to clean it up and have some firmware to avoid needing a non-distributable ROM image. VOF [Virtual Open Firmware] might be the simplest way for this to just get the Amiga like bootloaders [...] start which only need some CI fuctions."
https://lists.gnu.org/archive/html/qemu-ppc/2020-12/msg00077.html

[ Edited by Andreas_Wolf 06.12.2023 - 11:33 ]

Sadly in next version of QEMU will be not present the new Pegasos II board on which Zoltan Balaton is working on. Meanwhile there was a great fix for Sam460ex board to prevents random gurus and and improved emulation od via-ide chipset.
For a complete Pegasos II usage there will be still needed ATI Radeon full emulation and some other stuff.
Zoltan said me maybe it could be done in a 5.1 or 5.2 version of QEMU, so on 3th or 4th quarter of 2020. I tried to support him with debug logs from a real Pegasos II, but I hope there will other people interested on it who can support in any way the project.
Remember a virtual board will be never broken and time after time it will be faster and faster with release of new cpus.. It can be considered also a good way to preserve from time our beloved Amigas and our history.

Update:

> QEMU Sam460 emulation for running OS4 and [...]
> QEMU mac99 emulation for running MorphOS

It seems that QEMU v6.0 will come with a Sam460 emulation bug-fixed to be MorphOS-compatible:

https://lists.gnu.org/archive/html/qemu-ppc/2021-01/msg00006.html
https://lists.gnu.org/archive/html/qemu-ppc/2021-01/msg00029.html
https://lists.gnu.org/archive/html/qemu-ppc/2021-01/msg00060.html

[ Edited by Andreas_Wolf 06.01.2021 - 14:20 ]

Yes I hope Balaton's patches will be reviewed and approved for main tree in next days.
Meanwhile, as written in my other recent post, also Pegasos II board could be present in Qemu 6.0
With this new addition, main goal should be to virtualize powerpc on powermacs, and run OS4 at full speed.

Anyway this seems to be a great period for Amiga systems, there are happening a lot of nice things.

Quote:

Hmm, does this make any difference in the emulation over mac99? I mean, it wouldn't be faster, and the gfxchipset is still the emulated ST 502?

> does this make any difference in the emulation over mac99?

I can't think of any advantage of running MorphOS inside Sam460 emulation to running it inside PowerMac3,1/mac99 emulation, if that's what you mean.

> the gfxchipset is still the emulated ST 502?

Of course, it's still the SM502 (or rather emulated SM501, but that's apparently similar enough for the OS drivers not to notice) because that's what happens to be on the real Sam460ex board after all. However, I can't see why the emulated Rage 128 Pro (see comment #151) wouldn't be usable with Sam460 emulation (or the upcoming Pegasos II emulation for that matter). As I understand, Radeon R100 emulation is still being worked on.

> Pegasos II board could be present in Qemu 6.0

https://morph.zone/modules/newbb_plus/viewtopic.php?topic_id=13043&forum=3&start=19

Update:

> "I'll need to […] have some firmware to avoid needing a non-distributable ROM
> image. VOF [Virtual Open Firmware] might be the simplest way for this to just
> get the Amiga like bootloaders [...] start which only need some CI fuctions."
> https://lists.gnu.org/archive/html/qemu-ppc/2020-12/msg00077.html

MorphOS and Linux can now boot on emulated Pegasos II with VOF (built into coming QEMU v6.1) instead of the original SmartFirmware:

https://lists.gnu.org/archive/html/qemu-ppc/2021-06/msg00119.html
https://lists.gnu.org/archive/html/qemu-ppc/2021-06/msg00452.html
https://lists.gnu.org/archive/html/qemu-ppc/2021-06/msg00454.html

"The QEMU command line […] with the built in Virtual Open Firmware (VOF): […]
AmigaOS cannot boot with VOF and needs firmware ROM as it uses firmware to read files from boot volume that is not supported by VOF yet."
https://osdn.net/projects/qmiga/wiki/SubprojectPegasos2


Edit: added newer mailing list messages

[ Edited by Andreas_Wolf 27.06.2021 - 22:22 ]

I wonder if some in MOS/OS4 developer team would help Zoltan to reach the goal and have a virtual NG platform ready for next Qemu 6.1 release.
Zoltan could be also asked to be (at least partially) involved in x86 porting of MOS, because he has proved to be a very skilled lowlevel programmer.
QEMU could really be a valid alternative to UAE for AmigaNG platforms.
Next steps are about correctly emulating ATI radeon gfx cards.

> to reach the goal and have a virtual NG platform ready

Isn't this already there in form of the Sam460 emulation? What's intended with the Pegasos II emulation is a common QEMU platform that can be virtualized with KVM-PR on Book III-S hardware. This cannot be done with the Book III-E platform that is the Sam460, as you correctly wrote 5 months back.

> QEMU could really be a valid alternative to UAE
> for AmigaNG platforms.

UAE cannot run current MorphOS 2.0+ because there hasn't been released any MorphOS 2.0+ for PowerUP. So in terms of running MorphOS 2.0+, and thus also in terms of running both OS4 and MorphOS 2.0+ via the same software solution, QEMU isn't just a valid alternative to UAE but very much the only way to do it.

Quote:
I don't think he needs any help really, he seems to be doing just fine. Also, QEMU has been working fine for some time already. I've for example ran most of my MorphOS testing on QEMU for several years now (no need to carry multiple laptops around).

QEMU Pegasos II emulation gives no benefit over Sam460 emulation, to be honest. At least I see no point in switching to it (note that I care about MorphOS emulation).

QEMU has a very limited usability really: The performance is so limited it mainly is usable for quick testing. Compile and similar tasks are better suited for amd64 linux cross-compile environment.

Please also note that QEMU isn't an officially supported platform for MorphOS and you cannot register MorphOS for QEMU.

[ Edited by Piru 07.06.2021 - 22:15 ]

> I don't think he needs any help really

If desired, I can compile a list of Zoltan's messages to mailing lists (including the MorphOS one) where he (told he) asked the MorphOS devs for help and reported that sadly he did not receive any answers from them to technical questions ;-) Fortunately, you are right that so far he was able to solve crucial problems without the help of the MorphOS team, even if it took him longer to finally figure out the solutions or workarounds himself to make MorphOS run.

> QEMU Pegasos II emulation gives no benefit over Sam460 emulation

Maybe not so much for MorphOS, yes.

@Andreas_Wolf

At least me personally I have no desire to get involved.

Maybe could be sufficient give hits to some questions raised by Zoltan in qemu ml, as example about ide legacy implementation or firmware load. It’s not needed to code for qemu, Zoltan can do it for all us.

Update:

> MorphOS and Linux can now boot on emulated Pegasos II with VOF
> (built into coming QEMU v6.1) instead of the original SmartFirmware

QEMU v6.1 now released:

"new board support for Genesi/bPlan Pegasos II (pegasos2)"
https://lists.gnu.org/archive/html/qemu-devel/2021-08/msg04108.html
https://www.qemu.org/2021/08/24/qemu-6-1-0/

"- Experimental "Virtual Open Firmware" option for pseries (and Pegasos2) which implements most of the firmware behaviour inside qemu. […]
- Added 'pegasos2' machine type emulating the Genesi/bPlan Pegasos II board"
https://wiki.qemu.org/ChangeLog/6.1#PowerPC

I'm using an emulated Pegasos with MorphOS 3.15 rc1 to rc4. It can even be used, although it lacks a good graphics card. Lack of support for Lan and sound cards bothers a bit. It is also impossible to save autostart from setenv. Time to try the full version 6.1

> I'm using an emulated Pegasos with MorphOS 3.15 rc1 to rc4.

I guess you mean rc1 to rc4 of QEMU v6.1 ;-)

> Lack of support for Lan

Why not simply use an emulated RTL8139 like in comment #140?

> It is also impossible to save autostart from setenv.

As a workaround until NVRAM emulation is added, you can extract the boot.img from the ISO image and use '-kernel boot.img'.

Quote:

Yes, sorry, my mistake.:)

Quote:

Parameter which i used in Mac99 emulation, -net nic,model=rtl8139 -net user not work, dos console message: warning:requested NIC (#net033, model rtl8139) was not created.
(not supported by this machine?)


Quote:

Will try it tomorrow. Thanks for the help!

Is there some reason why anyone would use Pegasos II instead of Mac99 or Sam460 system in QEmu when running MorphOS? At least I don't see any reason.

[ Edited by Piru 28.08.2021 - 07:59 ]

Maybe because Pegasos was first MorphOS computer?
And what about Altivec - how is emulated/used in Qemu PPC ? I think there are no options such "cpu mode='host-passthrough" in PPC Qemu+KVM version ?..

Quote:

The goal was to use kvm-pr mode and use powerpc at full speed.
With a Sam460ex it was not possible due it's book3e cpu.

I repeat, any help is much appreciated and MOS Team could elaborate another way to release licences for these virtual boards.

Quote:

Once again, there are no plans to allow registering emulators.

> what about Altivec - how is emulated/used in Qemu PPC ?

On x86-64, which is what Piru runs QEMU on, AltiVec/VMX instructions are translated to AVX instructions (or to scalar instructions on old x86-64 CPUs that lack AVX).

> I think there are no options such "cpu mode='host-passthrough"
> in PPC Qemu+KVM version ?

In theory this should be possible with Book III-S CPU virtualized on Book III-S CPU (see comment #161). Some years ago, there were reports on POWER9 virtualizing G4 at 2 GHz running Mac OS X. I'm not aware of anyone successfully running MorphOS or OS4 this way, though.

Quote:
You can configure arbitrary CPU while using QEmu. I often use 750cxe and 7448 when testing (this is in combination with -machine mac99). I don't know how kvm-pr works exactly but it'd be mighty silly if it would somehow restrict the emulated CPU type.

You can get some impossible system + CPU combinations but as long as you boot the correct MorphOS version (and the CPU is supported in the first place) it ought to just work. Naturally if the kvm-pr can't emulate a MorphOS compatible CPU it won't work. My point here is that Pegasos II machine emulation doesn't give you any better options regarding this.

Granted, I don't use kvm-pr myself so I have no intimate knowledge about the limitations it poses. Yet, I see no way how Pegasos II machine emulation would improve the situation there.

[ Edited by Piru 28.08.2021 - 08:07 ]

Quote:

There would not be any emulation of CPU as same (PowerPC) CPU which runs Qemu natively would run MorphOS natively (through Qemu). Like VirtualBox + co. in x86 world. Or AROS/hosted on Linux.