OS9x86

[nightstalkerpoet]

Searching through old WinWorld downloads, I noticed that there is a Japanese ISO of System 7.1 Kanjitalk for x86 available. Since the goal of this site is keeping OS9 alive (as opposed to trying to stick only to PPC hardware), perhaps this image could be updated to OS 9? Also made me curious if in this version of Mac OS, the m68k emulator/converter could be replaced with one that converts PPC code to x86.

[nanopico]

there was a port of System 7 to x86.  It was an joint effort I believe with IBM.
All it could do was boot.  No software ran on it.
To make existing stuff work on it you would be writing a PPC emulator much like there is a 68k emulator currently.
So now you would have a 68k emulator on top of a PPC emulator on top of x86.
Since there is just enough in there to boot a very specific system, I am going to assume that
A. That actual system is not available anywhere
B. There is no emulator for either 68k or PPC
C. From a code point of view at the low levels involved, there is a very significant change between System 7 and OS 9.  You really would be just starting over. I don't think there is anything in there to use.

Still I would never say anything is impossible, but I'm not sure who would have the time and desire to actually accomplish this.

[darthnVader]

Searching through old WinWorld downloads, I noticed that there is a Japanese ISO of System 7.1 Kanjitalk for x86 available. Since the goal of this site is keeping OS9 alive (as opposed to trying to stick only to PPC hardware), perhaps this image could be updated to OS 9? Also made me curious if in this version of Mac OS, the m68k emulator/converter could be replaced with one that converts PPC code to x86.

If the goal is to run PPC apps, SheepShaver and Qemu-system-ppc work just fine on x86.

There is just no reason to reinvent the wheel.

Even if we did get our hands on systems 7 for x86, it's a far cry from being able to update that to OS 9. We'd really need the source code for OS 9, and I just don't see that happening, too many underlying copyrights and patens.

The only real reason to port and recompile OS 9 to x86 would be if you could get people to start writing apps for it again, and I just don't see that happening. I highly doubt OS 9 does address space randomization, it's not 64bit. So you have an OS that is insecure that is limited to 1.5 GB of Ram.

Now with the source code, you could fix these things, but in the end you would likely find what Apple found, it's just so much easer to dump it in favor of 'nix.

 

[nanopico]

Now with the source code, you could fix these things, but in the end you would likely find what Apple found, it's just so much easer to dump it in favor of 'nix.

This exactly.

[Naiw]

there was a port of System 7 to x86.  It was an joint effort I believe with IBM.
All it could do was boot.  No software ran on it.
To make existing stuff work on it you would be writing a PPC emulator much like there is a 68k emulator currently.
So now you would have a 68k emulator on top of a PPC emulator on top of x86.
Since there is just enough in there to boot a very specific system, I am going to assume that
A. That actual system is not available anywhere
B. There is no emulator for either 68k or PPC
C. From a code point of view at the low levels involved, there is a very significant change between System 7 and OS 9.  You really would be just starting over. I don't think there is anything in there to use.

Still I would never say anything is impossible, but I'm not sure who would have the time and desire to actually accomplish this.

Where from do you pull this information?

First of all we all know Mac OS 7 been running on at least 2 processors outside of 68k, PowerPC and AT&T Hobbit. This suggests that the 68k interpreter that was originally used was written in a high level language and was easily ported, if that's the case why would it been any more difficult get this going on x86?

We also know that classic Mac OS (and to not make any one confused here, classic Mac OS is the ROM) was written in 68k assembler for the most part, there was no way around that without compromising backwards compatibility.

We also know that System 7 does not implicitly mean PowerPC compatibility, System 7.0 had no ability to run PowerPC applications at all, infact I believe System 7.1.2 was the first to even have any PowerPC support. And even if it had if PowerPC compatibility was required for a particular platform that would not mean you would have the entire system on top of a PPC emulator that hosts a 68k emulator... Since the classic Mac OS to all it's knowledge runs on 68k it never had any knowledge the PowerPC at all, PowerPC emulation could be accomplished with a fairly simple user mode emulator- but I would argue it's unlikely this particular image would have any PowerPC compatibility.

And finally- if we overlook the fact that you claim this boots System 7.1 but didn't run any software (which by itself means there must be quite a bit of 68k emulation going on),
Kanjitalk was a system to be able to express multibyte characters within Mac OS- so I would reckon that unless this just isn't a mismarker and it's supposed to say 68k or PPC instead of x86, I would imagine it indeed runs software else it would be kind of pointless wouldn't it?

[Naiw]

And as a follow up, it's mislabeled. it's a 68k system.

[nanopico]

there was a port of System 7 to x86.  It was an joint effort I believe with IBM.
All it could do was boot.  No software ran on it.
To make existing stuff work on it you would be writing a PPC emulator much like there is a 68k emulator currently.
So now you would have a 68k emulator on top of a PPC emulator on top of x86.
Since there is just enough in there to boot a very specific system, I am going to assume that
A. That actual system is not available anywhere
B. There is no emulator for either 68k or PPC
C. From a code point of view at the low levels involved, there is a very significant change between System 7 and OS 9.  You really would be just starting over. I don't think there is anything in there to use.

Still I would never say anything is impossible, but I'm not sure who would have the time and desire to actually accomplish this.

Where from do you pull this information?

First of all we all know Mac OS 7 been running on at least 2 processors outside of 68k, PowerPC and AT&T Hobbit. This suggests that the 68k interpreter that was originally used was written in a high level language and was easily ported, if that's the case why would it been any more difficult get this going on x86?

We also know that classic Mac OS (and to not make any one confused here, classic Mac OS is the ROM) was written in 68k assembler for the most part, there was no way around that without compromising backwards compatibility.

We also know that System 7 does not implicitly mean PowerPC compatibility, System 7.0 had no ability to run PowerPC applications at all, infact I believe System 7.1.2 was the first to even have any PowerPC support. And even if it had if PowerPC compatibility was required for a particular platform that would not mean you would have the entire system on top of a PPC emulator that hosts a 68k emulator... Since the classic Mac OS to all it's knowledge runs on 68k it never had any knowledge the PowerPC at all, PowerPC emulation could be accomplished with a fairly simple user mode emulator- but I would argue it's unlikely this particular image would have any PowerPC compatibility.

And finally- if we overlook the fact that you claim this boots System 7.1 but didn't run any software (which by itself means there must be quite a bit of 68k emulation going on),
Kanjitalk was a system to be able to express multibyte characters within Mac OS- so I would reckon that unless this just isn't a mismarker and it's supposed to say 68k or PPC instead of x86, I would imagine it indeed runs software else it would be kind of pointless wouldn't it?

Here's the best explination.
http://lowendmac.com/2014/star-trek-apples-first-mac-os-on-intel-project/

The 68k was all ported.  It was all running x86 code.  This is why no applications ran.
The 68k emulator was never ported.

The 68k emulator runs on top of the PPC system.  A lot of the system is running PPC code.  The emulator is for backwards compatability and the remaining pieces of the OS in 68k.
Eln can elaborate more on that.  Thing is backwards compatability was compromised and not all 68k applications would run.

The thing is, the emulator wouldn't have to run ontop of a PPC emulator in the end, but you would need both a PPC emulator and 68k emulator.  An x86 kernel would have to marshal between them and between x86 code.

If it's as trivial and easy as you imply, then by all means give it a go.

[DieHard]

If it's as trivial and easy as you imply, then by all means give it a go.

Diehard was relaxing and sipping the morning java before opening his retail store to the masses... then he read the last line... surely a mistake as coffee blasted all over his LED screen with muffled snorted laugh.

[nanopico]

If it's as trivial and easy as you imply, then by all means give it a go.

Diehard was relaxing and sipping the morning java before opening his retail store to the masses... then he read the last line... surely a mistake as coffee blasted all over his LED screen with muffled snorted laugh.

I don't think there is a mistake.

[Naiw]

there was a port of System 7 to x86.  It was an joint effort I believe with IBM.
All it could do was boot.  No software ran on it.
To make existing stuff work on it you would be writing a PPC emulator much like there is a 68k emulator currently.
So now you would have a 68k emulator on top of a PPC emulator on top of x86.
Since there is just enough in there to boot a very specific system, I am going to assume that
A. That actual system is not available anywhere
B. There is no emulator for either 68k or PPC
C. From a code point of view at the low levels involved, there is a very significant change between System 7 and OS 9.  You really would be just starting over. I don't think there is anything in there to use.

Still I would never say anything is impossible, but I'm not sure who would have the time and desire to actually accomplish this.

Where from do you pull this information?

First of all we all know Mac OS 7 been running on at least 2 processors outside of 68k, PowerPC and AT&T Hobbit. This suggests that the 68k interpreter that was originally used was written in a high level language and was easily ported, if that's the case why would it been any more difficult get this going on x86?

We also know that classic Mac OS (and to not make any one confused here, classic Mac OS is the ROM) was written in 68k assembler for the most part, there was no way around that without compromising backwards compatibility.

We also know that System 7 does not implicitly mean PowerPC compatibility, System 7.0 had no ability to run PowerPC applications at all, infact I believe System 7.1.2 was the first to even have any PowerPC support. And even if it had if PowerPC compatibility was required for a particular platform that would not mean you would have the entire system on top of a PPC emulator that hosts a 68k emulator... Since the classic Mac OS to all it's knowledge runs on 68k it never had any knowledge the PowerPC at all, PowerPC emulation could be accomplished with a fairly simple user mode emulator- but I would argue it's unlikely this particular image would have any PowerPC compatibility.

And finally- if we overlook the fact that you claim this boots System 7.1 but didn't run any software (which by itself means there must be quite a bit of 68k emulation going on),
Kanjitalk was a system to be able to express multibyte characters within Mac OS- so I would reckon that unless this just isn't a mismarker and it's supposed to say 68k or PPC instead of x86, I would imagine it indeed runs software else it would be kind of pointless wouldn't it?

Here's the best explination.
http://lowendmac.com/2014/star-trek-apples-first-mac-os-on-intel-project/

The 68k was all ported.  It was all running x86 code.  This is why no applications ran.
The 68k emulator was never ported.

The 68k emulator runs on top of the PPC system.  A lot of the system is running PPC code.  The emulator is for backwards compatability and the remaining pieces of the OS in 68k.
Eln can elaborate more on that.  Thing is backwards compatability was compromised and not all 68k applications would run.

The thing is, the emulator wouldn't have to run ontop of a PPC emulator in the end, but you would need both a PPC emulator and 68k emulator.  An x86 kernel would have to marshal between them and between x86 code.

If it's as trivial and easy as you imply, then by all means give it a go.

Project Star Trek was never commercialized, even if this wasn’t a mislabeled 68k iso (which it is, I checked) there is no way it would been the base of such system.

I’m very well aware of how the 68k emulator fits into the system and I’m saying once again the 68k emulator is what runs the entire system- it’s true it’s there for compatibility reason; and the major reason is there wouldn’t have been a Mac OS for PowerPC without it, The entire system runs on top of it.

If you don’t need supervisor compatibility the emulation is a lot more trivial and often way way faster too.

If you hand me Mac OS 7 for x86; I can sure make a PoC.

---

Edit:

I obviously missed your last comment about marshalling, no it doesn't have to marshall between anything.
This is exactly what Rosetta did on Mac OS X and also the reason it didn't support kexts- the (PowerPC) emulation was performed entirely in userspace.
In addition to that "the 68k emulator was never ported", Mac OS did NOT have a 68k emulator before Gary Davidians nanokernel, he made both of them to get Mac OS running on PowerPC (well originally Motorola 88000, Not AT&T Hobbit as I wrote in the previous post- unfortunately mixed it with BeOS), This was done in parallel with Project Star Trek.
https://en.wikipedia.org/wiki/Mac_68k_emulator

But besides that Mac OS doesn't have a kernel- unless we're talking PowerPC systems that has the nanokernel (which as I repeatedly said is more of a "hypervisor"; especially the first version than anything else)

[dr bu]

dr bu for sure enjoy die hard/somebody! speak about him/herself in 3rd person

[Jubadub]

there was a port of System 7 to x86.  It was an joint effort I believe with IBM.

Wow. This just blew my mind. I had heard of the project, but never knew they ever had compiled it!!

I just felt the same kinda shock as the 2005 one.

Honestly, though, I'd be more interested in porting OS 9 over to the Talos II than anything Intel or AMD, modern (with their "glorious" IME and PSP) or old. Regardless of how possible or unviable the task would be, porting OS 9 to x86, to me, sound as interesting as "how to replace my eyes with two Nintendo 3DS cameras" or anything else similarly random. Possible? Maybe. A good idea? I don't think so.