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

thefinalprophecy · November 10, 2013

So many different versions of the modified kernel posted--which is the best to test with at this time?

Diazz72 · November 10, 2013

So many different versions of the modified kernel posted--which is the best to test with at this time?

I am using RC7 and its best so far. . .

theconnactic · November 10, 2013

Support-related stuff deleted...

Sinetek · November 10, 2013

I'm unable to extract a RAR file, either with Unarchiver or unrarx. plus, it's very slow..

there might be an error still in the opemu, around RIP section

Andy Vandijck · November 10, 2013

@Bronzovka: I made a diff containing the PM implementation for the kernel.

You might want to add this.

It will enable the new rtc_clock_stepped and rtc_clock_stepping algorithms and generate a System.kext for using VoodooPState.kext (the custom version).

I tried building the kernel with only this patch done and it builds without any problems...

PM_Patch_for_Bronya.diff.zip

Bronya · November 10, 2013

@Bronzovka: I made a diff containing the PM implementation for the kernel.

You might want to add this.

It will enable the new rtc_clock_stepped and rtc_clock_stepping algorithms and generate a System.kext for using VoodooPState.kext (the custom version).

I tried building the kernel with only this patch done and it builds without any problems...

Yes ! It is rtclock )) .

Thank ! Maybe tomorrow .

+ira · November 10, 2013

@Bronzovka: I made a diff containing the PM implementation for the kernel.

You might want to add this.

It will enable the new rtc_clock_stepped and rtc_clock_stepping algorithms and generate a System.kext for using VoodooPState.kext (the custom version).

I tried building the kernel with only this patch done and it builds without any problems...

Tested your latest kernel/kexts etc and Launchpad pink icons gone away, i noticed some lack of performance compared to first Shane build but system is working good anyway my hardware is {censored} i think it should work better in early amd rigs. Thanks for the work Andy!

Cheers!

Bronya · November 10, 2013

+ira, my kernel don't work for your cpu ? =)))

+ira · November 10, 2013

+ira, my kernel don't work for your cpu ? =)))

Will try it soon Bronya, i'm currently making room in my HDD's for another clean install and will report back.. Also, maybe later tonight i would put that P4 rig to work, i'm at -3GMT time...

MrApocalypse · November 10, 2013

Something good and weird happened. Yesterday I couldn't login to App Store, game me the unknown error message. I woke up this morning and I could login into App Store. :thumbsup_anim:

Now Mavericks is 99% functional on my rig. The broken 1% would be iCloud. Not complaining, just a report.

Andy Vandijck · November 10, 2013

Can somebody with an FX cpu (K15) boot with Bronzovka's kernel and post me a report of

sysctl -a | grep machdep.cpu

?

I'm going to mod the PM kext for FX cpus.

Thanks in advance...

EDIT: I would prefer it to be a Bulldozer CPU so I can implement it right.

AMD kernel dev manual helps in part but need cpu signature for accurate ID...

thefinalprophecy · November 10, 2013

What's the key to getting this working in VMware with AMD CPU? I have a home lab ESXi 5.5 host that I want to run a Mavericks Server on. I followed a guide with [url="http://www.insanelymac.com/forum/topic/279450-why-insanelymac-does-not-support-tonymacx86/"]#####[/url] to make a bootable USB drive, replaced the mach_kernel with the RC7 version from this thread but it does not boot. Just hangs on Apple logo, no apparent CPU usage at all. AMD FX-8320 CPU.

Thanks!

granik · November 10, 2013

Can somebody with an FX cpu (K15) boot with Bronzovka's kernel and post me a report of

sysctl -a | grep machdep.cpu

?

I'm going to mod the PM kext for FX cpus.

Thanks in advance...

EDIT: I would prefer it to be a Bulldozer CPU so I can implement it right.

AMD kernel dev manual helps in part but need cpu signature for accurate ID...

K8 (oldest non-FX) does not fit?

aziiz3 · November 10, 2013

here you go @Andy, mine is Fx-6100 which is bulldozer architecture.

using bronya's latest kernel :

Darwin Abdul-Azizs-iMac.local 13.0.0 Darwin Kernel Version 13.0.0: суббота, 9 ноября 2013 г. 02:42:04 (MSK); root:xnu-2422.1.72_by_bronya_sinetek_anv_rc7/BUILD/obj/RELEASE_X86_64 x86_64

machdep.cpu.max_basic: 13
machdep.cpu.max_ext: 2147483678
machdep.cpu.vendor: AuthenticAMD
machdep.cpu.brand_string: AMD FX(tm)-6100 Six-Core Processor             
machdep.cpu.family: 21
machdep.cpu.model: 1
machdep.cpu.extmodel: 0
machdep.cpu.extfamily: 6
machdep.cpu.stepping: 2
machdep.cpu.feature_bits: 395049983 513286667
machdep.cpu.extfeature_bits: 802421759 29999103
machdep.cpu.signature: 6295314
machdep.cpu.brand: 0
machdep.cpu.features: FPU VME DE PSE TSC MSR PAE MCE CX8 APIC SEP MTRR PGE MCA CMOV PAT PSE36 CLFSH MMX FXSR SSE SSE2 HTT SSE3 PCLMULQDQ MON SSSE3 CX16 SSE4.1 SSE4.2 POPCNT AES XSAVE OSXSAVE AVX1.0
machdep.cpu.extfeatures: SYSCALL XD 1GBPAGE EM64T LAHF RDTSCP TSCI
machdep.cpu.logical_per_package: 6
machdep.cpu.cores_per_package: 6
machdep.cpu.microcode_version: 21
machdep.cpu.processor_flag: 1
machdep.cpu.mwait.linesize_min: 64
machdep.cpu.mwait.linesize_max: 64
machdep.cpu.mwait.extensions: 3
machdep.cpu.mwait.sub_Cstates: 0
machdep.cpu.thermal.sensor: 0
machdep.cpu.thermal.dynamic_acceleration: 0
machdep.cpu.thermal.invariant_APIC_timer: 0
machdep.cpu.thermal.thresholds: 0
machdep.cpu.thermal.ACNT_MCNT: 1
machdep.cpu.thermal.core_power_limits: 0
machdep.cpu.thermal.fine_grain_clock_mod: 0
machdep.cpu.thermal.package_thermal_intr: 0
machdep.cpu.thermal.hardware_feedback: 0
machdep.cpu.thermal.energy_policy: 0
machdep.cpu.xsave.extended_state: 7 832 960 1073741824
machdep.cpu.arch_perf.version: 0
machdep.cpu.arch_perf.number: 0
machdep.cpu.arch_perf.width: 0
machdep.cpu.arch_perf.events_number: 0
machdep.cpu.arch_perf.events: 0
machdep.cpu.arch_perf.fixed_number: 0
machdep.cpu.arch_perf.fixed_width: 0
machdep.cpu.cache.linesize: 64
machdep.cpu.cache.L2_associativity: 16
machdep.cpu.cache.size: 2097152
machdep.cpu.tlb.inst.small: 48
machdep.cpu.tlb.inst.large: 24
machdep.cpu.tlb.data.small: 32
machdep.cpu.tlb.data.small_level1: 1024
machdep.cpu.tlb.data.large: 32
machdep.cpu.tlb.data.large_level1: 1024
machdep.cpu.address_bits.physical: 48
machdep.cpu.address_bits.virtual: 48
machdep.cpu.core_count: 6
machdep.cpu.thread_count: 6

Andy Vandijck · November 10, 2013

here you go @Andy, mine is Fx-6100 which is bulldozer architecture.

using bronya's latest kernel :

Darwin Abdul-Azizs-iMac.local 13.0.0 Darwin Kernel Version 13.0.0: суббота, 9 ноября 2013 г. 02:42:04 (MSK); root:xnu-2422.1.72_by_bronya_sinetek_anv_rc7/BUILD/obj/RELEASE_X86_64 x86_64

machdep.cpu.max_basic: 13
machdep.cpu.max_ext: 2147483678
machdep.cpu.vendor: AuthenticAMD
machdep.cpu.brand_string: AMD FX(tm)-6100 Six-Core Processor             
machdep.cpu.family: 21
machdep.cpu.model: 1
machdep.cpu.extmodel: 0
machdep.cpu.extfamily: 6
machdep.cpu.stepping: 2
machdep.cpu.feature_bits: 395049983 513286667
machdep.cpu.extfeature_bits: 802421759 29999103
machdep.cpu.signature: 6295314
machdep.cpu.brand: 0
machdep.cpu.features: FPU VME DE PSE TSC MSR PAE MCE CX8 APIC SEP MTRR PGE MCA CMOV PAT PSE36 CLFSH MMX FXSR SSE SSE2 HTT SSE3 PCLMULQDQ MON SSSE3 CX16 SSE4.1 SSE4.2 POPCNT AES XSAVE OSXSAVE AVX1.0
machdep.cpu.extfeatures: SYSCALL XD 1GBPAGE EM64T LAHF RDTSCP TSCI
machdep.cpu.logical_per_package: 6
machdep.cpu.cores_per_package: 6
machdep.cpu.microcode_version: 21
machdep.cpu.processor_flag: 1
machdep.cpu.mwait.linesize_min: 64
machdep.cpu.mwait.linesize_max: 64
machdep.cpu.mwait.extensions: 3
machdep.cpu.mwait.sub_Cstates: 0
machdep.cpu.thermal.sensor: 0
machdep.cpu.thermal.dynamic_acceleration: 0
machdep.cpu.thermal.invariant_APIC_timer: 0
machdep.cpu.thermal.thresholds: 0
machdep.cpu.thermal.ACNT_MCNT: 1
machdep.cpu.thermal.core_power_limits: 0
machdep.cpu.thermal.fine_grain_clock_mod: 0
machdep.cpu.thermal.package_thermal_intr: 0
machdep.cpu.thermal.hardware_feedback: 0
machdep.cpu.thermal.energy_policy: 0
machdep.cpu.xsave.extended_state: 7 832 960 1073741824
machdep.cpu.arch_perf.version: 0
machdep.cpu.arch_perf.number: 0
machdep.cpu.arch_perf.width: 0
machdep.cpu.arch_perf.events_number: 0
machdep.cpu.arch_perf.events: 0
machdep.cpu.arch_perf.fixed_number: 0
machdep.cpu.arch_perf.fixed_width: 0
machdep.cpu.cache.linesize: 64
machdep.cpu.cache.L2_associativity: 16
machdep.cpu.cache.size: 2097152
machdep.cpu.tlb.inst.small: 48
machdep.cpu.tlb.inst.large: 24
machdep.cpu.tlb.data.small: 32
machdep.cpu.tlb.data.small_level1: 1024
machdep.cpu.tlb.data.large: 32
machdep.cpu.tlb.data.large_level1: 1024
machdep.cpu.address_bits.physical: 48
machdep.cpu.address_bits.virtual: 48
machdep.cpu.core_count: 6
machdep.cpu.thread_count: 6

Ok, family is the same as 0x00600000 which happens to be Bulldozer and Piledriver.

I noticed that it uses the same mechanism as K10 (kext defaults for families above 0x00200000 to the following:

inline UInt32 AMDGetK11Frequency(UInt8 fid, UInt8 did) { return (100 * (fid + 0x08) >> did);}

Which is bad for K15.

K15 is the same as K10 namely:

inline UInt32 AMDGetK10Frequency(UInt8 fid, UInt8 did) { return (100 * (fid + 0x10) >> did);}

According to the AMD CPU docs.

I modified the kext, could you guys (gils + aziiz) test the new PM pack?

Normally the frequency should be correct now...

Please test and post the following table (see screenshot)

EDIT: Oh yeah, and also the CPU max frequency please... so I can compare...

AMD_PM_REL3.zip

+ira · November 10, 2013

Awesome, cool Sinetek ! ))

Interested !

Thank for mtrr.c

I created mach_kernel_rc7 !!! Soon full.diff !

It is working with chameleon , i tested !

Soon for ML 10.8 and Lion !

Updated , i CASUAL DELETE "HASWEL" =)))

mach_rc7_10_9.zip

Congrats Bronya, it seems that you made a reliable kernel, performance seems better and system is running smoothly anyway the other problems still persist like the small artifacts etc...

Shaneee · November 10, 2013

@Andy, Is this info useful for sorting some issues? http://www.insanelymac.com/forum/topic/281450-mavericks-kernel-testing-on-amd-formerly-mountain-lion-kernel-testing-on-amd/?p=1966096

Shaneee · November 10, 2013

This is what I get when I try to boot in VMware (verbose mode). Any suggestion? Running RC7.

This is not a help topic...

aziiz3 · November 10, 2013

Ok, family is the same as 0x00600000 which happens to be Bulldozer and Piledriver.

I noticed that it uses the same mechanism as K10 (kext defaults for families above 0x00200000 to the following:

inline UInt32 AMDGetK11Frequency(UInt8 fid, UInt8 did) { return (100 * (fid + 0x08) >> did);}

Which is bad for K15.

K15 is the same as K10 namely:

inline UInt32 AMDGetK10Frequency(UInt8 fid, UInt8 did) { return (100 * (fid + 0x10) >> did);}

According to the AMD CPU docs.

I modified the kext, could you guys (gils + aziiz) test the new PM pack?

Normally the frequency should be correct now...

Please test and post the following table (see screenshot)

EDIT: Oh yeah, and also the CPU max frequency please... so I can compare...

my cpu max frequency should be : 3300Mhz/3.3Ghz

and this is the table:

Andy Vandijck · November 10, 2013

my cpu max frequency should be : 3300Mhz/3.3Ghz

and this is the table:

ps copy.jpg

Damn... still not good...

Going to check it tomorrow evening

I'm tired so I'm gonna get some sleep...

aziiz3 · November 10, 2013

Damn... still not good...

Going to check it tomorrow evening

I'm tired so I'm gonna get some sleep...

OK, Sleep tight andy :angel:

ameris_cyning · November 10, 2013

I probably should have said something earlier

I think I might have an idea of how to conquer the iCloud problem

I am using a kernel compiled from the sources with a Power Management fix, yet I have Messages.app working without a hitch.

So, maybe that means that our AMD brethren could boot up on an Intel machine with their AMD hackintosh hard drive, sign into Messages, then place it back into your machine. Can anyone test this theory?

Konner · November 11, 2013

Can somebody with an FX cpu (K15) boot with Bronzovka's kernel and post me a report of

sysctl -a | grep machdep.cpu

?

I'm going to mod the PM kext for FX cpus.

Thanks in advance...

EDIT: I would prefer it to be a Bulldozer CPU so I can implement it right.

AMD kernel dev manual helps in part but need cpu signature for accurate ID...

Ishamis-Mac-Pro:~ ishami$ sysctl -a | grep machdep.cpu
machdep.cpu.max_basic: 13
machdep.cpu.max_ext: 2147483678
machdep.cpu.vendor: AuthenticAMD
machdep.cpu.brand_string: AMD FX(tm)-8350 Eight-Core Processor           
machdep.cpu.family: 21
machdep.cpu.model: 2
machdep.cpu.extmodel: 0
machdep.cpu.extfamily: 6
machdep.cpu.stepping: 0
machdep.cpu.feature_bits: 395049983 1050161675
machdep.cpu.extfeature_bits: 802421759 32227327
machdep.cpu.signature: 6295328
machdep.cpu.brand: 0
machdep.cpu.features: FPU VME DE PSE TSC MSR PAE MCE CX8 APIC SEP MTRR PGE MCA CMOV PAT PSE36 CLFSH MMX FXSR SSE SSE2 HTT SSE3 PCLMULQDQ MON SSSE3 FMA CX16 SSE4.1 SSE4.2 POPCNT AES XSAVE OSXSAVE AVX1.0 F16C
machdep.cpu.extfeatures: SYSCALL XD 1GBPAGE EM64T LAHF RDTSCP TSCI
machdep.cpu.logical_per_package: 8
machdep.cpu.cores_per_package: 8
machdep.cpu.microcode_version: 21
machdep.cpu.processor_flag: 1
machdep.cpu.mwait.linesize_min: 64
machdep.cpu.mwait.linesize_max: 64
machdep.cpu.mwait.extensions: 3
machdep.cpu.mwait.sub_Cstates: 0
machdep.cpu.thermal.sensor: 0
machdep.cpu.thermal.dynamic_acceleration: 0
machdep.cpu.thermal.invariant_APIC_timer: 0
machdep.cpu.thermal.thresholds: 0
machdep.cpu.thermal.ACNT_MCNT: 1
machdep.cpu.thermal.core_power_limits: 0
machdep.cpu.thermal.fine_grain_clock_mod: 0
machdep.cpu.thermal.package_thermal_intr: 0
machdep.cpu.thermal.hardware_feedback: 0
machdep.cpu.thermal.energy_policy: 0
machdep.cpu.xsave.extended_state: 7 832 960 1073741824
machdep.cpu.arch_perf.version: 0
machdep.cpu.arch_perf.number: 0
machdep.cpu.arch_perf.width: 0
machdep.cpu.arch_perf.events_number: 0
machdep.cpu.arch_perf.events: 0
machdep.cpu.arch_perf.fixed_number: 0
machdep.cpu.arch_perf.fixed_width: 0
machdep.cpu.cache.linesize: 64
machdep.cpu.cache.L2_associativity: 16
machdep.cpu.cache.size: 2097152
machdep.cpu.tlb.inst.small: 48
machdep.cpu.tlb.inst.large: 24
machdep.cpu.tlb.data.small: 64
machdep.cpu.tlb.data.small_level1: 1024
machdep.cpu.tlb.data.large: 64
machdep.cpu.tlb.data.large_level1: 1024
machdep.cpu.address_bits.physical: 48
machdep.cpu.address_bits.virtual: 48
machdep.cpu.core_count: 8
machdep.cpu.thread_count: 8

This is on an FX-8350 with bronya's kernel.

MariusT · November 11, 2013

Ok, family is the same as 0x00600000 which happens to be Bulldozer and Piledriver.

I noticed that it uses the same mechanism as K10 (kext defaults for families above 0x00200000 to the following:

inline UInt32 AMDGetK11Frequency(UInt8 fid, UInt8 did) { return (100 * (fid + 0x08) >> did);}

Which is bad for K15.

K15 is the same as K10 namely:

inline UInt32 AMDGetK10Frequency(UInt8 fid, UInt8 did) { return (100 * (fid + 0x10) >> did);}

According to the AMD CPU docs.

I modified the kext, could you guys (gils + aziiz) test the new PM pack?

Normally the frequency should be correct now...

Please test and post the following table (see screenshot)

EDIT: Oh yeah, and also the CPU max frequency please... so I can compare...

With FX-8350 default clock 200MHz x 20

Mariuss-Mac-Pro:AMD_PM_REL3 marius$ uname -a
Darwin Mariuss-Mac-Pro.local 13.0.0 Darwin Kernel Version 13.0.0: za  9 nov 2013 21:28:59 CET; SinetekBronzovkAnVoodoo v2.0.0:xnu-2422.1.72-AMD/BUILD/obj/RELEASE_X86_64 x86_64
Mariuss-Mac-Pro:AMD_PM_REL3 marius$ sysctl machdep.cpu
machdep.cpu.max_basic: 13
machdep.cpu.max_ext: 2147483678
machdep.cpu.vendor: GenuineIntel
machdep.cpu.brand_string: Intel(R) Xeon(R) CPU           E
machdep.cpu.family: 21
machdep.cpu.model: 2
machdep.cpu.extmodel: 1
machdep.cpu.extfamily: 0
machdep.cpu.stepping: 6
machdep.cpu.feature_bits: 395049983 1050161675
machdep.cpu.extfeature_bits: 801373183 32227327
machdep.cpu.signature: 67190
machdep.cpu.brand: 0
machdep.cpu.features: FPU VME DE PSE TSC MSR PAE MCE CX8 APIC SEP MTRR PGE MCA CMOV PAT PSE36 CLFSH MMX FXSR SSE SSE2 HTT SSE3 PCLMULQDQ MON SSSE3 FMA CX16 SSE4.1 SSE4.2 POPCNT AES XSAVE OSXSAVE AVX1.0 F16C
machdep.cpu.extfeatures: SYSCALL 1GBPAGE EM64T LAHF RDTSCP TSCI
machdep.cpu.logical_per_package: 8
machdep.cpu.cores_per_package: 8
machdep.cpu.microcode_version: 21
machdep.cpu.processor_flag: 1
machdep.cpu.mwait.linesize_min: 64
machdep.cpu.mwait.linesize_max: 64
machdep.cpu.mwait.extensions: 3
machdep.cpu.mwait.sub_Cstates: 0
machdep.cpu.thermal.sensor: 0
machdep.cpu.thermal.dynamic_acceleration: 0
machdep.cpu.thermal.invariant_APIC_timer: 0
machdep.cpu.thermal.thresholds: 0
machdep.cpu.thermal.ACNT_MCNT: 1
machdep.cpu.thermal.core_power_limits: 0
machdep.cpu.thermal.fine_grain_clock_mod: 0
machdep.cpu.thermal.package_thermal_intr: 0
machdep.cpu.thermal.hardware_feedback: 0
machdep.cpu.thermal.energy_policy: 0
machdep.cpu.xsave.extended_state: 7 832 960 1073741824
machdep.cpu.arch_perf.version: 0
machdep.cpu.arch_perf.number: 0
machdep.cpu.arch_perf.width: 0
machdep.cpu.arch_perf.events_number: 0
machdep.cpu.arch_perf.events: 0
machdep.cpu.arch_perf.fixed_number: 0
machdep.cpu.arch_perf.fixed_width: 0
machdep.cpu.cache.linesize: 64
machdep.cpu.cache.L2_associativity: 16
machdep.cpu.cache.size: 2097152
machdep.cpu.address_bits.physical: 48
machdep.cpu.address_bits.virtual: 48
machdep.cpu.core_count: 8
machdep.cpu.thread_count: 8
Mariuss-Mac-Pro:AMD_PM_REL3 marius$

Attached PStateChanger

Screen+Shot+2013-11-11+at+8.55.56.png

hope it helps you ~

Duran Keeley · November 11, 2013

I probably should have said something earlier

I think I might have an idea of how to conquer the iCloud problem

I am using a kernel compiled from the sources with a Power Management fix, yet I have Messages.app working without a hitch.

So, maybe that means that our AMD brethren could boot up on an Intel machine with their AMD hackintosh hard drive, sign into Messages, then place it back into your machine. Can anyone test this theory?

Sadly no go.... Changed the HDD to my Intel machine (Haswell i5-4570) cloned my extra folder to my AMD HDD (with nvram's & nvram module,smbios and so on), booted with the first kernel compiled by Shanee. Activation fails.

It wouldn't be a true test if I didn't boot to the original kernel. Tried activation and it worked perfectly.

Reboot onto my AMD machine and boom there goes iMessage again inactive, iCloud still works.

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

Recommended Posts

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites

Link to comment

Share on other sites