Mavericks kernel testing on AMD (formerly Mountain Lion kernel testing on AMD)

[theconnactic]

We made history here! Now Mountain Lion works on AMD Machines. The support is still limited to Phenom and FX-Series CPUs, but soon we'll have fixes for other CPU models ALL CPUS, but nVidia GPUs won't boot without either safe boot (-x) or disabling GeforceGA.plugin and GeforceGLDriver.bundle (thus no QE/CI - hopefully a fix will come soon).

 

amd_kernel.zip -----> latest kernel: credits to Andy Vandjick! (Pag 74, post #1468)

 

This kext is obligatory: you must install it before starting your machine: http://cl.ly/0o3s0d2c001A -----> AnVsCoreCrypto! No more need for any patched kext.

 

Thanks to all that developed, tested, commented, helped to spread the news, to all who believed in what was deemed impossible. Success was our reward, and i'm very proud of it.

 

Happy new year!

 

 

 

 

EDIT: We have an IRC room for talking about development of a Mountain Lion working kernel for AMD and other unsupported CPUs. Join us! Server chat.osx86.hu, room #LegacyKernel

 

 

Well, i suppose we're lucky for Apple's early release of its Mountain Lion kernel as open source. Thanks to this, a patched Atom kernel is already done, although it doesn't work on my little Atom netbook. Hopefuly, an AMD patched kernel will be around soon but, as everybody should know, this one will be a troublesome build. The AMD kernel for the now superseded Mac OSX Lion is very limited, to be kind. In most AMD machines, basic functionalities doen't even run, like - yes, believe it! - the Finder. It could be quite useless, but it's not, thanks to many alternatives and workarounds to be found at the community.

 

The problem for AMD is that any patched kernel since Lion is strictly i386. To have access to 64 bit apps in an AMD Lion install, it's necessary a Buldozer CPU, since it has support for a key instruction set in OSX Lion, the ssse3 (please, don't confuse it with sse3). All other AMD processors should run strict 32-bit, forbidding crucial apps like the newer xcodes. A solution for this would be a sss3 emulator. Anyone willing or able to write?

 

(silence)

 

As i already knew. But even when some good-willing and capable-minded individual(s) do write the emulator, there's another issue: maybe Lion would be good to go with AMD, but Mountain Lion needs a 64-bit-only legacy kernel. And 64-bit legacy kernels equals instant reboots in AMD rigs in Lion, even on ssse3-capable Buldozer machines. You can try it for yourself, with a 64-bit patch of RAW x86 (boot with arch=x64_86): http://osx86.co/f100...687/page23.html

 

The question: is it really needed to figure out why these problems plague Lion running on AMD to make AMD work under Mountain Lion? Or the behavior of 64-bit patches would be completely different in ML? Anyway, it seems obvious to me that a sss3 emulator is crucial, since the vast majority of AMD systems are non-Buldozer.

 

So, there's an anxious waiting at AMD realm, to finally get an early update of AMD systems, to be shoulder to shoulder with Intel users.

[Huckleberry Pie]

That's why I stuck to Intel. AMD seems good on Windows and other OSes, but for something that didn't support a non-Intel CPU from the start, it would be too much work to do.

[theconnactic]

Snow Leopard and earlier iterations of OsX used to run flawlessly on AMD systems. Maybe it comes to my mind only because my AMD computer isn't invited to the party, but seems to me that, while apple is trying to use the power of intel CPUs to the maximum, which is good for performance, it makes its system more and more dependent on Intel's capability of innovate and improve. Intel has the edge now and i'm pretty sure that it will have it in the near future. But then, who knows?

 

OsX isn't anymore a straight x86-based system (or, more accurately, x86_64), but an Intel EM64T system. As it goes, Microsoft is making improvements in multithreading for Windows 8, so it could use the full eight cores of, say, AMD Piledriver upcoming CPUS. Maybe that gives an Intel-high-end-level of performance to this AMD CPUs, which will surely cost less. Is the newest releases of OsX optimized for multithreading? If so, a Piledriver hackintosh is not bad a prospect for power users, and could well worth the effort of writing emulators etc.

[theconnactic]

Given the progress made in Lion towards 64-bit support (running on a 32-bit kernel, we don't have a working arch x86_64 kernel yet) on ssse3 capable CPUs (FX series), i think that running ML on AMD will be soon a reality.

[tle88]

EM64T is in Pentium4 6xx models, Pentium4 5x1 models and Celeron 3x1 and 3x6 models.

They are still quite common and widely used processors.

 

Many of us still have Pentium D series 64-bit processor, which has Intel EM64T too but only SSE3.

SSSE3 in Core2Duo has only 4 multimedia orders more. It would be really nice thing to give the

working legacy_kernel for this useful Pentium D processor.

 

Fortunately 10.6.8 works well with some 10.6.7 kexts and voodooHDA 2.7.4 , but working Lion

or Mountain Lion would be definitely nice.

[applelight]

As i already knew. But even when some good-willing and capable-minded individual(s) do write the emulator, there's another issue: maybe Lion would be good to go with AMD, but Mountain Lion needs a 64-bit-only legacy kernel. And 64-bit legacy kernels equals instant reboots in AMD rigs in Lion, even on ssse3-capable Buldozer machines. You can try it for yourself, with a 64-bit patch of RAW x86 (boot with arch=x64_86): http://osx86.co/f100...687/page23.html

 

The question: is it really needed to figure out why these problems plague Lion running on AMD to make AMD work under Mountain Lion? Or the behavior of 64-bit patches would be completely different in ML? Anyway, it seems obvious to me that a sss3 emulator is crucial, since the vast majority of AMD systems are non-Buldozer.

 

So, there's an anxious waiting at AMD realm, to finally get an early update of AMD systems, to be shoulder to shoulder with Intel users.

 

Thanks for the low down Connatic. For political reasons (I don't like monopolies), I wanted to build a loaded AMD hackintosh to run Premiere Pro 6. I can see that it's a nearly impossible task. Much time saved. Much appreciated.

[theconnactic]

You're welcome! Subscribe to this topic, any changes in the scenario i'll post here.

[theconnactic]

I've been taking a look at the XNU source of Mountain Lion. If i would make changes so it boots on AMD machines, which of the source files i should change, and what should i look for to change?

 

There are the following folders, plus some documents, and the folders names are bsd, config, external headers, iokit, libkern, libsa, libsyscalls, makedefs, osfmk, pexpert, security, setup, tools. The documents are kgmacros, lidbmacros.py, makefile, apple_license and a readme with instructions (but not the kind of instructions i'm looking for).

 

So where to start? I'm not the most capable (with slightly above basic knowledge of C, and in the process of studying assembly by myself) but i'm more than willing. Any help would be so much appreciated. Even if all this ends up in an epic fail.

 

P.S.: i now the ssse3 calls are a particularly difficult issue, especially for non-Bulldozer AMD CPUs, but i intend to do a basic experiment, following a suggestion from Pooky, of changing the ssse3 calls to sse3 calls, whenever they appear. To do it, though, i have to build myself a lot more assembly language proficiency.

[atlee]

I've been taking a look at the XNU source of Mountain Lion. If i would make changes so it boots on AMD machines, which of the source files i should change, and what should i look for to change?

 

I haven't been around for a very long time but i'm deciding to come back for a little while it's just a hell of a lot of work to even re-master editing the XNU source for AMD to work. Best option would be just to wait for Apple to support AMD and Intel in the future so it will them make your job and life so much easier.

[theconnactic]

What you're telling me is to sit and wait to Apple to support AMD CPUs somewhere in the future, right? In other words (since Apple has never shown a sign that it would do that), to give up and either stick to the achievements we've already had for AMD machines (fully working Snow Leopard) or change to an Intel set-up.

 

Well, not giving contempt to your advice by any means, since it's quite reasonable, i would really want to know how to do it, how to patch a kernel for it to working with AMD CPUs, or at least where to start, where and what to look for on it to change, what kind of programming resources must i have etc. Do you know about any useful information on these subjects? Would be much appreciated.

 

I'm willing to try, as i said, even being a beginner with just above basics of C, no assembly, no ruby etc. I'm willing to do it because no one else is doing it, so i think someone has to start. Did the hackintosh community become like uni/###### users, who just use what is natively supported? If that's the point, why not simply buy a Mac? It "just works" and it has a better price tag than ever, and saves people the time (and time is money, you know) they spend tweaking things, even on a natively supported hardware.

 

For me, is kind of a science experiment. Of course, i love OSX and would love if it simply worked on my AMD computer with little to no effort. Of course, a minimum level of hardware compatibility is necessary to even start to think about running OSX on an average PC computer. But, if i have to spend lots of money (hey, i'm not rulling out spending some in supported and cool peripherals, and i myself did it more than once) assembling a whole new computer with natively supported parts, i would be better switching to a Mac.

 

Thank you very much for your answer! And the AMD quest goes on!

[bcobco]

for me programming is like black magic... but i offer myself for beta testing

[theconnactic]

Thank you very much. Even it being a long term goal, bcobco, you know. Perhaps the beta test phase never comes, because i don't know yet how and where to start and, as for programing, i'm still learning - know C, but no assembly etc - and i'm slow to learn, and this is a hobby for me - i'm not a professional programmer and don't want to be one, since i'm doing well in my own professional activity. But i'm commited to it and i'm gathering info about it. If someone more skilled than me (that means, every dev. in Insanelymac, lol) decides to join efforts, this will go way faster.

[bcobco]

dont worry. im following this thread.

also, this seems to be the first and official OSX-10.8 on AMD thread. so please keep updated the first topic as header with information (and perhaps links) for all osx86 community. maybe make sticky thread.

best regards

[atlee]

What you're telling me is to sit and wait to Apple to support AMD CPUs somewhere in the future, right? In other words (since Apple has never shown a sign that it would do that), to give up and either stick to the achievements we've already had for AMD machines (fully working Snow Leopard) or change to an Intel set-up.

 

When you say noone is doing it you are completly wrong, Go talk to andyvand (AnV) to see what he has done in terms of the code he has written, I remember back when he said he made it boot to a certain point then it crashed but atleast he has done a lot of the work.

 

Some reading for you man which will do most of the homework for you:

http://osx86.co/f100/lion-with-amd-t7106/page100.html

http://www.hack-my-mac.fr/index.php/forum/48-pc-bureau/28494-tutoriel-amd-fx-buldo#28896

[theconnactic]

Atlee, i'm afraid you simply didn't understand what i said, or didn't read carefully, so i'll take the liberty of quote myself:

 

I'm willing to try, as i said, even being a beginner with just above basics of C, no assembly, no ruby etc. I'm willing to do it because no one else is doing it.

 

I'm obviously not intending to remind you basic grammar, but i didn't use the present continuous tense without reason. [ ]

 

I never said that no one else ever done anything for us AMD users - and that would have been plain stupid, since i used Andy's and Nawcom's work when running Snow Leo, and Bronzovka and Raw's when in Lion, installing it with Devout's method by the way. I said, as you can read, as much as an AMD Mountain Lion kernel is concerned, that no one else is working on it (to be fair, no one else appears to), so it's up to me and anyone else who really wants to stick to AMD, or does not have another option. This is not matter of opinion, but of fact.

 

I would really appreciate if Andy, Nawcom, Bronzovka, RAW etc show up with something for us, since they have experience and are far far far more skilled than i am. But it's not their obligation, and it's totally understandable if they can't or simply don't want to do it. They already done much, as you yourself pointed. Therefore, as Gringo Vermelho usually answers to lazy hackintoshers, why not try to do it ourselves?

 

Thank you!

 

By the way, i already did my first experience: i take RAWx86's patch for Lion 10.7.4 (http://osx86.co/f100...d-kernel-t7687/) and applied it to the Mountain Lion XNU. As expected, a total and complete fiasco, lol.

[Devout]

Good on you for this project, I know we all have doubts, it's a good project to keep to yourself. I'm not familiar with how to add support for AMD on the older mach kernels (Lion/Snow Leopard etc.), but those who added support had an experience background, and to an extent only made so much work over a period of time, and it took them a while. Keep going, as it's better to be determined then just sit here!

 

But, maybe another hobby like xcode coding is more enticing

[theconnactic]

Thank you for the moral support, Devout.

 

Anyway, if you have a good idea, don't be shy: if your hobby is something like Xcode coding, you're with all probability much more experienced in programming languages than i am (remember that my skills in C are just above basics, and i'm just beginning to study assembly language by myself), so you can possibly give some good advice.

[atlee]

Try not to be to tight on correcting my English or grammar because I do whatever without following certain order, I don't even bother to correct before I post most times, Don't really care as forums are just quick reply message systems. Anyway Go talk to the bigger devs as they already have all the starting gear and code and ask them if you can use it and work upon it. Saves you starting from scratch and going through the failing periods they have already done.

[Devout]

Thank you for the moral support, Devout.

 

Anyway, if you have a good idea, don't be shy: if your hobby is something like Xcode coding, you're with all probability much more experienced in programming languages than i am (remember that my skills in C are just above basics, and i'm just beginning to study assembly language by myself), so you can possibly give some good advice.

 

hehe. Nope I'm not programming, but I've started learning C-Objective Programming. I'm more of a designer then a programmer, but I want to implement my design into my own systems, so I want to save time and effort of finding someone else to implement these and just do it myself . But hobbies can get really fun, and I'm glad you've found something you're interested in.

[theconnactic]

Atlee, man, i was not correcting your english: you missed the whole point. But thanks for your advice, it's valuable!

 

Devout, just like me: i'm not a programmer, less even than you are, but it would be really nice if i could make apps to address my own needs. Then my hobby can reveal itself as an useful thing. Good look with your studies, and thank you very much for your guides, without which i'd never start making a hackintosh at the first place.

 

About my work with the ML kernel, no progress yet: so many new things to learn and understand, so few available material or people to help me do it, and time's another issue. I confess i'm taking a licking... but i'm still kicking!

[bcobco]

maybe this helps...

there is a patch for modifying cpu instructions to adapt dyld and libSystem.dylib binaries to non intel cpus.

AvN made a kernel (Darwin compu.local 10.8.0 Darwin Kernel Version 10.8.0: ma 22 aug 2011 22:12:25 CEST; legacy kernel v8 :xnu-1504.15.3/BUILD/obj/RELEASE_I386 i386) with that patch built in the kernel. maybe he can drop some ideas

[theconnactic]

That could really help, bcobco. Where is this patch available?

 

How can i reach AnV out to get some advice? He's not being posting here for a long time; maybe he's too busy to invest time and effort on this kernel subject for now.

 

Thank you very much!

[bcobco]

look at this thread. with comments from aVn about this patch

http://osx86.co/f36/...-be-nice-t6845/

 

EDIT

i found a version here

http://www.osx86.net...sn_patcher.html

the other version i found somewhere i dont remember... but i have a copy.

PM me your email and i send you both versions

 

also, look for the darwin kernel modified by avn (or nawcom) against the original darwin source code to see what things change.

[theconnactic]

Thank you, bcobco: i'll check this out ASAP.

[SS01]

This is something that's been in the back of my head for a while... I'm taking a look at the XNU sources, given how much I know about programming (basic C++) I doubt I'll get anywhere, but if I find/do anything interesting I'll be sure to let you know. Also, I will offer myself for beta testing too - doubt a Deneb Phenom is really what you're after, but it can't hurt to volunteer can it?