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

[theconnactic]

Building Andy's kernel with Xcode.

[Science]

I'm back and have redesigned the kernel some more.

It needs testing A.S.A.P. (+feedback )

Source diff is included.

I've also made the diff ready for Lion (needs merging).

Pentium D and AMD users should like it

It shouldn't panic anymore

Don't ask which version it is please (it's 10.8.2 lol)

Thanks! Takes longer to boot, but stuff like the dictation thingy won't crash anymore!

[Andy Vandijck]

Soon I'll post a new one.

This one has also a sse3 opemu for first gen Athlon64 and Pentium 4...

[theconnactic]

Great job, Andy! I still didn't have the opportunity to test your latest kernel, but i already compiled it here on my laptop, since i'm not home.

 

Do you have any idea why all the kernels seem to have a high degree of incompatibility with nVidia GPUs? Maybe, perhaps, some dyld thing...?

[byransays]

I'm back and have redesigned the kernel some more.

It needs testing A.S.A.P. (+feedback )

Source diff is included.

I've also made the diff ready for Lion (needs merging).

Pentium D and AMD users should like it

It shouldn't panic anymore

Don't ask which version it is please (it's 10.8.2 lol)

 

This one seems to be working very well, it doesnt crash with the trackpad prefpane.

I do notice I cant access my NTFS partition anymore though...is this kernel related? The system doesnt even see it.

 

I am still running safeboot because of Nvidia, however maybe its due to the kernel or because of me changing smbios to imac12,2 but even as buggy as they are with no qe/ci it runs a bit faster with things like launchpad.

 

edit: scratch that, foot in my mouth. I put in an SMBIOS but it didnt load it, it still sees it as a macpro 3,1

 

edit 2: scratch that, I remembered I chose to do a macpro 5,1 smbios instead...funny thing is it took the serial I chose for the SMBIOS but didnt take the SMproductname, it still sees it as a 3,1 not a 5,1

[Andy Vandijck]

Here is the new kernel, opemu enhanced for SSE3 emulation (some P4's with 64bit and early Athlon64 cpu's require SSE3 emulation).

Test please

 

Use RAID mode, jtsanborn. It usually implies AHCI.

 

P.S.: Great job, Andy! I still didn't have the opportunity to test your latest kernel, but i already compiled it here on my laptop, since i'm not home.

 

Do you have any idea why all the kernels seem to have a high degree of incompatibility with nVidia GPUs? Maybe, perhaps, some dyld thing...?

I wish I knew...

I don't have a clue why the NVidia GPU's don't work.

(I have a NVidia Geforce 8300 GS and I also use this kernel... it seems to work fine, but I have Intel :X)

amd_kernel.zip

SinetekBronzovkAnVoodoo_rc3.diff.zip

[instant idiot]

Test please

AMD Athlon II x4 640:

32 and 64-bit apps working!

Stock corecrypto working!

It even wash--oh, I guess it didn't wash the dishes after all. Must've been the cat.

No more QuickTime panics.

 

Intel Pentium D 3.4 GHz:

64-bit apps working!

32-bit apps, well, uTorrent launches but TextWrangler doesn't. I don't think I have any more 32-bit only apps on my system to test, but if anyone wants to point out a few, I'll test them.

Stock corecrypto working!

No more QuickTime panics.

 

enhanced for SSE3 emulation

I haven't any 64-bit non-SSE3 CPUs.

 

Just curious, with all this emulation going on, will there be any significant performance impacts?

[spakk]

its stucks on jnl disk0s4 : journal replay done

[Deltac0]

Test please

 

Phenom II x4 965

 

Boots slowly, but is otherwise awesome work!

Only few graphic glitches exist in my system (ATI).

32 and 64 seems both to be working 100%!

 

Now only the graphics issues need a fix! (Both ATI & NVidia)

 

 

 

EDIT: Forgot this: I used vanilla corecrypto and it passed the FIPS test fine.

[instant idiot]

Slowly Is how many time? 5 minutes?

Come to think of it, mine boots a bit slowly too; it boots in about 1:40 with the Athlon II (even longer for the Pentium D) from Chameleon to the login screen.

[Lacedaemon]

Try mach_sse42_only from here. I think it's the best working kernel at the moment.

Thanks for the suggestion, instant idiot! I'll try AnV's latest amd_kernel first, then if that doesn't work, I'll go for the mach_sse42_only.

[instant idiot]

As far as I can tell, AnV's latest kernel does work as well as mach_sse42_only.

[Lacedaemon]

As far as I can tell, AnV's latest kernel does work as well as mach_sse42_only.

Not for me. :glare: The installer still crashes with AnV's kernel, now moving on to mach_sse42_only.

 

Update: Booted with mach_sse42_only, got "Waiting for root device". Looks like I'll need that AppleNForceATA after all!

Edited January 7, 2013 by Lacedaemon

[mac_carol]

I'm back and have redesigned the kernel some more.

It needs testing A.S.A.P. (+feedback )

 

 

Found 1 more issue

 

is it necessary to perform the FIPS TEST twice ?

 

 

TEST #1: After corecrypto.kext started!

 

 

 

 

TEST #2: After Kernel is LP64

 

[happel]

Here is the new kernel, opemu enhanced for SSE3 emulation (some P4's with 64bit and early Athlon64 cpu's require SSE3 emulation).

Test please

 

This one is great! Boot takes about twice the time of sinetek's mach_sse42_only, but hey, everything just works .

[Mohamed Khairy]

@andyvand

 

on some PC there is a lag on screen , videos and audios --> overcomes by this command under terminal to make cpu usage over 20 % this fix this lags

yes > /dev/null

 

can this fixed on kernel like making a flag on kernel allow users having this problem to make cpu usage like 10 / 15 % on idle state

 

thanks

[Andy Vandijck]

Yes.

I will be recoding it later today

[AkimoA]

Could it be that the window server launching the gui interfaces doesnt get full attention of the kernel yet ?

 

just brainstorming !

 

 

pfff

 

i just compared the logs off my intel system with a 5750 and the amd system with the 5770

 

seems like there is no diffrence at all ....

 

bollox !

 

also booted the intel system with latest anvkernel also POSTs Fips twice and macx swapon gab takes round 30 sec (SSD though)

 

no GUI trouble either offcourse

 

@mohamed

 

 

yes > /dev/null

 

that trick didnt work here

[Andy Vandijck]

Few... the work is about done.

I've optimized the opemu so that it runs ssse3 emu pass first, then sse3 emu, then syscall, then monitor skip and last fisttp replacement to fistp (just added that)

Needless to say this is much faster code...

Post soon...

[theconnactic]

Way to go, Andy!

[Andy Vandijck]

Here it is... the new kernel.

It should be more efficient in performing instruction emulation (First SSSE3 pass, then SSE3 pass in 3 steps as this is more efficient, next sysenter, then monitor/mwait skipp for those without SSE3 and last fisttp to fistp replacement for those without SSE3).

It also has fisttp to fistp on-the-fly in memory replacement (added it in opemu) for those without SSE3.

Source diff included.

Test please and report back too

SinetekBronzovkAnVoodoo_rc4.diff.zip

amd_kernel.zip

[Rasmus Rasmussen]

Hey Anv Worked here, and were faster compared to your previous kernel

 

but for some reason my 6850 acts up where my 4870 is acting as it should, without garbling in osx.

 

Thx here take a 

 

Best Regards Racermussen

[theconnactic]

Hi, Andy!

 

Just tested your latest and greatest. Boots fine with -x, faster than earlier iterations. But the best thing, that convinces me that we're going to the right direction, is that, without -x, it boots smoothly... to a white screen with mouse pointer cursor (that i can control at will). Okay, this is not working, but it's far far better than the previous heavy fractal fragments without any mouse cursor or even the white screen with spinning pinwheel and frozen mouse cursor of the previous kernel. This evolution hints that the problem with nVidia (and some ATI) graphics is perhaps related to encryption, not a library.

 

Best regards and keep on the good work!

 

EDIT: Maybe it's not encryption after all: booting with -f gave five consecutive KPs with kextd.

 

EDIT 2: Finally the boot completed successful with -f. Unfortunately, bad news: the heavy fractal fragments are back.

[Mohamed Khairy]

the problem of lag not fixed ?

[byransays]

Hi, Andy!

 

Just tested your latest and greatest. Boots fine with -x, faster than earlier iterations. But the best thing, that convinces me that we're going to the right direction, is that, without -x, it boots smoothly... to a white screen with mouse pointer cursor (that i can control at will). Okay, this is not working, but it's far far better than the previous heavy fractal fragments without any mouse cursor or even the white screen with spinning pinwheel and frozen mouse cursor of the previous kernel. This evolution hints that the problem with nVidia (and some ATI) graphics is perhaps related to encryption, not a library.

 

Best regards and keep on the good work!

 

EDIT: Maybe it's not encryption after all: booting with -f gave five consecutive KPs with kextd.

 

EDIT 2: Finally the boot completed successful with -f. Unfortunately, bad news: the heavy fractal fragments are back.

 

Probably not encryption as safeboot wouldnt give us native 1080p if the video card was recognized with encrypted kexts.

 

You're farther along than me, with the new kernel I still get grey screen, then cut to black terminal screen where I can type in the top left.

 

Here it is... the new kernel.

It should be more efficient in performing instruction emulation (First SSSE3 pass, then SSE3 pass in 3 steps as this is more efficient, next sysenter, then monitor/mwait skipp for those without SSE3 and last fisttp to fistp replacement for those without SSE3).

It also has fisttp to fistp on-the-fly in memory replacement (added it in opemu) for those without SSE3.

Source diff included.

Test please and report back too

 

It seems a bit faster yes, it still gets caught up somewhere around USBF and mbinit and pthread(sorry I just missed it)

Edit: pictures are better for you probably. The red is where it hangs for a bit.