So the word gives that there are porting morphos to x64.does anybody know how fare there are? I have Steen a video were morphos is booking.but what i wanna know is had there been any Work done onbit since then and if so how much and is there any new plans like smp og virtuel memory and kernel protection or is IT as i sucspect and rough portover???

@retro

Quote:

retro wrote: So the word gives that there are porting morphos to x64.does anybody know how fare there are? I have Steen a video were morphos is booking.but what i wanna know is had there been any Work done onbit since then and if so how much and is there any new plans like smp og virtuel memory and kernel protection or is IT as i sucspect and rough portover???

Yes, work continues to be done on the port of MorphOS to x64, but the MorphOS Dev. Team does not discuss how far their progress has gone, or when they anticipate a first release will be ready. You probably saw the "Proof of Concept" video that BigFoot showed at one of the shows in Europe somewhere (I can't remember which show exactly), but other than that one video, I have not seen any other examples of the port working. Anyone who knows the MorphOS Dev. Team members, knows how capable they are at creating an OS and adding features and software to it over time, will know that it takes time and a very lot of effort, but the MorphOS Dev. Team continues to amaze me with their progress, all for just a one time license registration fee. The team is really great IMHO!

Moving to the x64 platform is a huge task though, so I don't think any of the current users are expecting a release of the port (or it may be more of a rewrite, than a port) in the near future. Probably more like another year or so, but I might be pleasantly surprised. I did not even expect the "Proof of Concept" video to be possible for another 6 months, and that video is already many months old, so they did something much faster than I thought possible.

I do wish that they would publish a road map for which features the new x64 version of MorphOS will have in it, but again, that is not their usual way, as they do not like to announce anything, until it is ready for release, or nearly ready for release.

_________________
Amiga! The computer that inspired so many, to accomplish so much, but has ended up in the hands of . . . . . . . . . .

@amigadave

Thanks for your input

@amigadave

I guess the biggest problem is to implement the transparent emulation of PPC and 68k on X64. It is certainly a huge and demanding project

@OlafS25

They already have Morphos PPC running under emulation perhaps not integrated under Morphos X64 though.

@kamelito

as I wrote "transparent"

@OlafS25

Let's hope the Q-Box springs to life. About time really

_________________

@OlafS25

They can use QEMU for both 68k and ppc.

_________________
Pegasos II G4@1GHz 2GB Radeon 9250 256MB
AmigaOS4.1 fe - MorphOS - Debian 9 Jessie

@OlafS25

I can’t see transparent nor invisible :)
They shouldn’t care. Maintained PPC apps could be recompiled. Dead PPC well who cares.

Last edited by kamelito on 14-Nov-2020 at 06:19 PM.

@OlafS25

Quote:

OlafS25 wrote: @amigadave I guess the biggest problem is to implement the transparent emulation of PPC and 68k on X64. It is certainly a huge and demanding project

I'm not sure if "transparent emulation of PPC and 68k" is a goal of the MorphOS Dev. Team for the x64 version of MorphOS. There was some talk long ago about completely breaking backward compatibility with 68k Amiga software, in order to instead provide many modern features, such as multi-core support, memory protection, etc., but the "Proof of Concept" seemed to me to be more of just a port of our existing PPC version of MorphOS, to the x64 platform, and no mention of advanced new features have been confirmed (or at least I have not looked closely enough to know for sure what is being planned, or promised).

Personally, I was hoping that the x64 version of MorphOS would have the ability to easily run Linux programs, without even needing to recompile them, sort of like what you get by using a hosted OS and the Rabbit Hole feature, but implemented differently, without hosting the new MorphOS, but I don't think that is planned, and probably it is not even possible. My desire is to have tons of software, or more accurately, my concern, is not having enough great software to run, when the x64 version of MorphOS is first released, but I'll just have to wait, and trust that the MorphOS Dev. Team has thought about this concern as well, and they have a solution. They haven't let me down so far, so it is easier for me to sit back and wait to see what amazing work they come up with.

As you point out, including emulation of PPC that is as transparent as the current version of MorphOS3.14 is at running 68k Amiga software, will probably be a huge task. I'm more interested in having the MorphOS Dev. Team come up with a really great new OS that is very similar to our existing PPC version of MorphOS, but I don't care about it being able to run PPC programs through emulation, because hopefully most good PPC MorphOS software can be recompiled for the x64 platform, and as long as there is a version of UAE that gets ported to the x64 version of MorphOS, that will run all of the 68k Amiga software sufficiently well for me.

Last edited by amigadave on 15-Nov-2020 at 03:01 AM.

_________________
Amiga! The computer that inspired so many, to accomplish so much, but has ended up in the hands of . . . . . . . . . .

@OlafS25 Quote:

OlafS25 wrote: @amigadave I guess the biggest problem is to implement the transparent emulation of PPC and 68k on X64. It is certainly a huge and demanding project

It's enough to use UAE for the old 68K apps/games.

PowerPC apps can be recompiled.

Both 68K and PowerPC are legacy architectures, and it's better to (finally) focus on some modern stuff.

@Fl@sh



Quote:

They can use QEMU for both 68k and ppc.

Yes, if they GPL MorphOS, they can do that :)

_________________
B5D6A1D019D5D45BCC56F4782AC220D8B3E2A6CC

@amigadave

I do not know how current MorphOS users think (more a question to you ;) ) but one of the main reason why people not used Aros because it not integrated 68k like MorphOS or AmigaOS. So for me to keep the feeling people currently have when using MorphOS on PPC transparent integration of 68k and PPC would be important. But that is only my view ;) we will see

And regarding directly running Linux software I cannot imagine that this is possible without using Linux as base. Of course porting from linux could be made much easier by porting required components from linux to MorphOS.

Last edited by OlafS25 on 15-Nov-2020 at 09:03 AM.

@retro

Quote:

retro wrote: So the word gives that there are porting morphos to x64.does anybody know how fare there are?

No.

Quote:

I have Steen a video were morphos is booking.

The key feature in that video was the new kernel, and while that was very impressive in itself, it didn't really offer any clues whatsoever to the rest of the questions you have asked here. Sorry...

Quote:

but what i wanna know is had there been any Work done onbit since then and if so how much and is there any new plans like smp og virtuel memory and kernel protection or is IT as i sucspect and rough portover???

There has been snippets of info communicated here and there during the years. But the AMD64 migration was announced quite exactly nine years(!) ago. That's so long that it's difficult to know if plans back then still applies or if it has been something of a moving target. Some of the developers we saw back then even seems to have faded out of the picture, but the core still remains. Is the plan they had nine years ago still the same or has it changed? It could have changed several times. Impossible for an outsider to know. Possibly the work on the X5000's firmware by MorphOS developers and its MorphOS port could be seen as the first concrete bricks being laid down in this path? The thread (the first post) over at Morph.zone that warns about X5000 firmware upgrades suggest it was.

But anyway. Even a "rough portover" would remove one of the key features of MorphOS, the [binary] Amiga compatibility. This because of the endianness difference. Which makes it no point in trying to maintain that aspect of the OS, so it would in practice become similar to AROS in that aspect, i.e. source compatible where the OS and applications could be ported/recompiled to other architectures. But one snippet of info that one of the developers shared on Morph.zone at one point (was it Piru?) is that the new MorphOS will not even be source compatible; that changes will be needed to be done to source code for applications in order for them to run, albeit changes would be kept to a minimum. This indeed suggests new Amiga-incompatible features. Like real SMP, real 64-bit computing/addressing, etc.

The new kernel supports both real SMP and real 64-bit. The rest of the OS is a different matter. But IMHO it wouldn't make sense to go all this way without going all the way. A clean slate restart, no "a-box", instead all OS components and all applications running directly on the new kernel. In fact, I would be disappointed if it would turn out to be anything different than this. I would want MorphOS to move forward, on its own merits. Not staying in limbo of not really being retro and not really being "Next Generation" either.

@TRIPOS

that is all nice. I only see the problem who will do the new software for the OS. The OS developers alone cannot do it. And if existing software no longer works you have at best a new OS with modern features but no software.

@TRIPOS

Quote:

But anyway. Even a "rough portover" would remove one of the key features of MorphOS, the [binary] Amiga compatibility. This because of the endianness difference. Which makes it no point in trying to maintain that aspect of the OS, so it would in practice become similar to AROS in that aspect, i.e. source compatible where the OS and applications could be ported/recompiled to other architectures. But one snippet of info that one of the developers shared on Morph.zone at one point (was it Piru?) is that the new MorphOS will not even be source compatible; that changes will be needed to be done to source code for applications in order for them to run, albeit changes would be kept to a minimum. This indeed suggests new Amiga-incompatible features. Like real SMP, real 64-bit computing/addressing, etc.

I agree. One of the features of MorphOS was being able to run 68K. But I don't think the endian difference can be used as an excuse now. Because x86 has emulated 68K for years. Look at Amithlon which spawned the UAE JIT engine. That was a highly optimsed AmigaOS/68K emulator. To the point some thought it was actually AmigaOS running on x86. And it had x86 native apps that hooked into the emulated 68K OS. Back then there was only BSWAP. Now there has been MOVEBE for years. x86 has had direct big endian support for over a decade. All they need is an intelligent compiler to compile code as big endian for x86.

MOS is also split into the Q/Box and A/Box. They just need an x86 Q/Box and an endian agnostic A/Box. Depending on how it works they could do it the OpenFirmware way. OpenFirmware always works in big endian no matter what the native CPU is. Any pointers passed to and fro must be in big endian. They could work the same. Like network programming. Or BCPL.

PPC shouldn't be that hard as it's simpler then 68K. All codes are the same size. So it should be easy enough to emulate and it's been done before. Though I think emulation is a step too far when they could make use of the simpler coding and do a load time translation of PPC codes. An advanced Elf loader that can read in a PPC Elf and rewrite it into memory as an x86 conversion on the fly.

@Hypex

The MorphOS X64 demo shown at Amiga34 in Neuss was able to "play" PPC and 68k code transparently.

@OlafS25

Quote:

OlafS25 wrote: @amigadave I do not know how current MorphOS users think (more a question to you ;) ) but one of the main reason why people not used Aros because it not integrated 68k like MorphOS or AmigaOS.

The importance of transparent 68k emulation is going down and down. If I look though the programs that I use at least semi-regularly there s probably only one 68k program (tvpaint) the rest is MorphOS native.
Clipboard sharing and the ability to launch direct from Amient that's enoght for legacy stuff. 68 is legacy, ppc, too.
Time to move on and leave the pulled brakes behind.

_________________
My programs: via.bckrs.de
MorphOS user since V0.4 (2001)

@OlafS25

Quote:

OlafS25 wrote: @amigadave I do not know how current MorphOS users think (more a question to you ;) ) but one of the main reason why people not used Aros because it not integrated 68k like MorphOS or AmigaOS. So for me to keep the feeling people currently have when using MorphOS on PPC transparent integration of 68k and PPC would be important. But that is only my view ;) we will see

I can't answer for other MorphOS users, but I have a feeling that many if not most of them would be like me, and would be okay with using some version of UAE to run all of their old Amiga 68k software, if that is the price we have to pay to get true smp multicore support, full 64bit architecture support, and real memory protection. Those three modern features that are impossible to achieve while keeping transparent backward compatibility with Amiga 68k software, and it's 31bit address space. It should (IMO) be more important to implement those three features in any new OS, specially when UAE works so well these days, and launching it can be done so painlessly and easily. I am afraid that the MorphOS Dev. Team is not being so bold to do what I have described, as they are afraid they will fracture the MorphOS user base yet again, if they deliver something that does not have some type of backward compatibility with the current PPC versions of MorphOS. Yes, we shall see, when they finally release MorphOS for the x64 architecture. I will be excited to see what magic they can achieve.

Quote:

And regarding directly running Linux software I cannot imagine that this is possible without using Linux as base. Of course porting from linux could be made much easier by porting required components from linux to MorphOS.

That part was just a "pipe dream" of mine, to allow us to access the huge library of Linux software from day one. I know it is not going to be included, and I only mentioned it because it was something that I had hoped for years ago, when the idea of a new x64 version of MorphOS was first mentioned, and that it "might" be a complete break from backward compatibility with the AmigaOS 68k API. (or should that be ABI?)

_________________
Amiga! The computer that inspired so many, to accomplish so much, but has ended up in the hands of . . . . . . . . . .

It shouldn’t be rocket science to port the linux kernel to run ontop of Quark through the usual virtualization methods.

_________________
B5D6A1D019D5D45BCC56F4782AC220D8B3E2A6CC