X99 Motherboards with Socket 2011-3 Now Available

[Manic Harmon1c]

Great work!

As far as I can tell manic harmon1c was the first to figure out what needed to change in the kernel and tonymac is the first person to get OS X running on the X99 platform.

Actually, I would like to clarify something. I actually didn't "figure it out," another user mentioned it on a certain hackintosh blog.
Unfortunately the forum wouldn't let me post a link to it.

The user's name is stinga11, he mentioned that cpuid.c was missing something. So I compared cpuid.c to cpuid.h and saw what he was talking about.

I really didn't mean to take credit for that, I just wanted to share the information because I wanted to see these builds working.

 

[tonymacx86]

Actually, I would like to clarify something. I actually didn't "figure it out," another user mentioned it on a certain hackintosh blog.
Unfortunately the forum wouldn't let me post a link to it.

The user's name is stinga11, he mentioned that cpuid.c was missing something. So I compared cpuid.c to cpuid.h and saw what he was talking about.

I really didn't mean to take credit for that, I just wanted to share the information because I wanted to see these builds working.

Thanks for the clarification. The necessary fix had already been discovered years ago- it was one line of code. It's exactly the same as Ivy Bridge and Sandy Bridge early kernels and adding cpuid to the source. The key this time was using an older version of Xcode (5.0.1) and Command Line Tools from March 2014.

 

[Manic Harmon1c]

tonymac, do you want to do a source build without any modifications? I can see if I can compare the 2 and come up with a binary patch. unless of course I can finally get it to build. I'll try 5.0.1 and the March command line tools.

 

[stinga11]

tonymac, do you want to do a source build without any modifications? I can see if I can compare the 2 and come up with a binary patch. unless of course I can finally get it to build. I'll try 5.0.1 and the March command line tools.

Is you use clover you can use the following for add support for Haswell-E.
<key>KernelAndKextPatches</key>
<dict>
<key>FakeCPUID</key>
<string>0x0306C0</string>
<key>KernelCpu</key>
<true/>
</dict>

 

[shilohh]

Stinga11 is brilliant. He provided the patched kernel binary patch that most used for the 10.8-10.9.1 IvyBridgeE installs and helped alot with getting speedstep going in x79. @tonymacx86, thanks for clarifying. It seems that we're all recycling good info but I guess the key is knowing when and where to apply it all. I know that the kernel's code changes from build to build but it would be great to have a bin patch for the vanilla kernel using perl or a hex editor so we can use XCPM. Thanks again for the great work and the great site!

 

[tonymacx86]

Is you use clover you can use the following for add support for Haswell-E.
<key>KernelAndKextPatches</key>
<dict>
<key>FakeCPUID</key>
<string>0x0306C0</string>
<key>KernelCpu</key>
<true/>
</dict>

Thanks I will try this.

 

[Manic Harmon1c]

We already tried that in the other thread, I think the FakeCPUID is not complete yet (it's meant to replace the kernelCPU patch, I remember Slice saying something about it). It may help, though along with the custom kernel. For example, TimeWalker has created a custom built Clover package that autopatches the kernel for atom but it still needs a FakeCPUID to work properly. So if the intention of that is just to get power management working I think it could be helpful (I'm not sure if Haswell will use AICPM at all though).

 

[Manic Harmon1c]

5.0.1 and March command line tools still didn't work for me. I just get all kinds of file not found errors.

 

[stinga11]

Stinga11 is brilliant. He provided the patched kernel binary patch that most used for the 10.8-10.9.1 IvyBridgeE installs and helped alot with getting speedstep going in x79.

Hey thanks for that. I try to learn from all community experts.

I think that the vanilla kernel with bin. patch is not good idea because before when we were testing a unsupported cpu the first thing we were doing was delete AICPM or use a null kext, but now the AICPM do not does anything because haswell uses the XCPM and the only way for disable is using a compiled kernel that do not has XCPM

 

[Manic Harmon1c]

Yeah the problem is that source built kernels don't have support for icloud, app store etc. Last time I checked.

I figured out my problem thanks to you stinga11. I didn't realize I was supposed to build in /usr/local/src.