WhatEverGreen Support Topic

[LockDown]

I have downloaded both (lilu & whatevergreen) sources. Compiled lilu in debug mode then compile whatevergreen as release (Profiling).

 

UPDATE:

Zero (0) warnings in xcode 9.2 

19 warning only occurs in xcode 8.3.3. 

Edited July 23, 2018 by ellaosx

[Ricky1994]

hi,

my hackintosh PC: i5 8500+RX580 8G(with igpu enabled necessary for hardware video decoding)+DisplayPort,WG 1.2.0 and Lilu 1.2.5. High Sierra 10.13.6. VBIOS is UEFI.just got a black screen in the second stage,but WG 1.1.7 and 1.1.8 is fine,anyboy know how to fix it ?

[wujek_bogdan]

@vit9696

I've just replaced the NvidiaGraphicsFixup kext with the new version of WhatEverGreen and I think I found a bug.

 

So far I've been using NvidiaGraphicsFixup + ngfxpatch=pikera boot arg that solved a High Sierra black screen after wake-up problem. Now, after replacing NvidiaGraphicsFixup with WhatEverGreen, the screen does not wake up properly after sleep. No matter is it auto-sleep or manual sleep (chosen from the Apple menu).

 

I see that ngfxpatch=pikera became agdpmod=pikera - I tried both, just in case, but it doesn't help.

 

// EDIT:

I've just found some weirdness. When I use the agdpmod=pikera flag (which, I assume, is a replacement for the ngfxpatch=pikera flag) then, on a log-in screen I have full HD resolution instead of 4K res. When I use the ngfxpatch=pikera (which is deprecated, I guess) then the log-in screen resolution is fine.

 

---

macOS 10.13.6, MSI GTX960 running 4k res, Nvidia Web Driver 387.10.10.10.40.105, Z87M-PLUS

Clover config: https://pastebin.com/ByDwhW8z

Kexts:

AppleALC.kext

FakeSMC_CPUSensors.kext

FakeSMC_LPCSensors.kext

RealtekRTL8111.kext

WhateverGreen.kext
FakeSMC.kext

FakeSMC_GPUSensors.kext Lilu.kext

Lilu.kext

SATA_Legacy.kext

 

All kexts up to date. Including Lilu.

Edited July 24, 2018 by wujek_bogdan

[xelanaiznac]

16 hours ago, wujek_bogdan said:

@vit9696

I've just replaced the NvidiaGraphicsFixup kext with the new version of WhatEverGreen and I think I found a bug.

 

So far I've been using NvidiaGraphicsFixup + ngfxpatch=pikera boot arg that solved a High Sierra black screen after wake-up problem. Now, after replacing NvidiaGraphicsFixup with WhatEverGreen, the screen does not wake up properly after sleep. No matter is it auto-sleep or manual sleep (chosen from the Apple menu).

 

I see that ngfxpatch=pikera became agdpmod=pikera - I tried both, just in case, but it doesn't help.

 

// EDIT:

I've just found some weirdness. When I use the agdpmod=pikera flag (which, I assume, is a replacement for the ngfxpatch=pikera flag) then, on a log-in screen I have full HD resolution instead of 4K res. When I use the ngfxpatch=pikera (which is deprecated, I guess) then the log-in screen resolution is fine.

 

---

macOS 10.13.6, MSI GTX960 running 4k res, Nvidia Web Driver 387.10.10.10.40.105, Z87M-PLUS

Clover config: https://pastebin.com/ByDwhW8z

Kexts:

AppleALC.kext

FakeSMC_CPUSensors.kext

FakeSMC_LPCSensors.kext

RealtekRTL8111.kext

WhateverGreen.kext
FakeSMC.kext

FakeSMC_GPUSensors.kext Lilu.kext

Lilu.kext

SATA_Legacy.kext

 

All kexts up to date. Including Lilu.

did you try this?

agdpmod=vit9696

[wujek_bogdan]

42 minutes ago, xelanaiznac said:

did you try this?

agdpmod=vit9696

 

No I didn't, I'll try. 

// Edit: I've just tried. It doesn't work either.

 

But anyway, If  ngfxpatch=pikera used to work fine before kext update, but no longer works after update then it seems to be a bug in the kext. I report it here, because I didn't find an another place where I could report bugs. Issues tracker is disabled on WhateverGreen Github.

Edited July 25, 2018 by wujek_bogdan

[FlymeDllVA]

Hi have graphics issues on xiaomi notebook pro 15'6 i7 8550u + hd 620. The method of incorporating graphics: https://github.com/daliansky/XiaoMi-Pro . The problem is the microblogs appearing in more complex animations. The videos from youtube and twitch work fine. But 3D graphics, photoshop, moving Windows, etc lag. Thanks for help!

Video with problem:

https://www.youtube.com/watch?v=f23DyPI4Z4U&feature=youtu.be

https://www.youtube.com/watch?v=_Q6VmusUvK4&feature=youtu.be

Config:

https://github.com/FlymeDllVa/XiaoMi-Pro/blob/master/EFI/CLOVER/config.plist

Full folder EFI:

https://github.com/FlymeDllVa/XiaoMi-Pro/tree/master/EFI

ioregSaved:

https://github.com/FlymeDllVa/XiaoMi-Pro/blob/master/ioregSaved.txt

Edited July 25, 2018 by FlymeDllVA

[Jingu]

On 7/23/2018 at 2:59 AM, vit9696 said:

ellaosx, I am not aware of any. Any log and Xcode/OS info?

 

For everyone else. Check your system logs. In most case it is clear from the debug log of Lilu and plugins what is wrong. The releases were thoroughly tested before publication. It is very likely that your configuration (e.g. ACPI) is done wrong.

 

Proper reference instructions for different GPUs (written by me and several colleagues) are only available in Russian at the moment, you could try to follow them if you have doubts. Translations are welcome.

Intel: https://www.applelife.ru/posts/729378

AMD: https://applelife.ru/threads/ati-hd4xx0-hd5xx0-amd-hd6xx0-hd7xx0-r5-r7-r9-rx-vega.28890/

NVIDIA: https://applelife.ru/threads/nvidia-gt-x-750-950-980-maxwell-1030-1080-pascal.1546195/

IQSV: https://applelife.ru/threads/zavod-intel-quick-sync-video.817923/

 

@vit9696thank you for the very detailed instructions for the Intel HD530. But the instructions don't seem to cover the particular black screen issues with iMac17,1 system definition.

 

I'm on High Sierra 10.13.16, Intel HD530, DisplayPort connector, no discrete GPU

 

Based on the instructions for Intel Graphics, I use:

 

Inject Intel = No

Clover ig-platform-id = 0x19120000

Removed FakePCIID_Intel_HD_Graphics.kext

Lilu 1.2.5, WhateverGreen 1.2.0

 

But, it only boots to a black screen

 

But, if I use Lilu 1.2.5, IntelGraphicsFixup 1.2.7, and

 

        <key>KextsToPatch</key>
        <array>
            <dict>
                <key>Comment</key>
                <string>10.11-SKL-1912000-4_displays</string>
                <key>Disabled</key>
                <true/>
                <key>Find</key>
                <data>AQMDAw==</data>
                <key>Name</key>
                <string>AppleIntelSKLGraphicsFramebuffer</string>
                <key>Replace</key>
                <data>AQMEAw==</data>
            </dict>

 

My iMac17,1 HD530 boots fine with full acceleration. Things are not quite ready for me to replace IntelGraphicsFixup with WhateverGreen.

 

Edited July 28, 2018 by Jingu

[Hackmac101]

8 hours ago, Jingu said:

 

@vit9696thank you for the very detailed instructions for the Intel HD530. But the instructions don't seem to cover the particular black screen issues with iMac17,1 system definition.

 

I'm on High Sierra 10.13.16, Intel HD530, DisplayPort connector, no discrete GPU

 

Based on the instructions for Intel Graphics, I use:

 

Inject Intel = No

Clover ig-platform-id = 0x19120000

Removed FakePCIID_Intel_HD_Graphics.kext

Lilu 1.2.5, WhateverGreen 1.2.0

 

But, it only boots to a black screen

 

But, if I use Lilu 1.2.5, IntelGraphicsFixup 1.2.7, and

 

        <key>KextsToPatch</key>
        <array>
            <dict>
                <key>Comment</key>
                <string>10.11-SKL-1912000-4_displays</string>
                <key>Disabled</key>
                <true/>
                <key>Find</key>
                <data>AQMDAw==</data>
                <key>Name</key>
                <string>AppleIntelSKLGraphicsFramebuffer</string>
                <key>Replace</key>
                <data>AQMEAw==</data>
            </dict>

 

My iMac17,1 HD530 boots fine with full acceleration. Things are not quite ready for me to replace IntelGraphicsFixup with WhateverGreen.

 

I have the same boot to black screen issue I have mentioned it in this thread, with iMac17,1 and platform ID 0x19120000. It is issue caused by disabling AGDP from WhateverGreen. 

Edited July 28, 2018 by Hackmac101

[Roger Smith]

I have a problem with CoreDisplay patches not being applied because of my VM configuration.

I have an HD Graphics passed through to Qemu via OVMF and Lilu says that my CPU is not supported. But I think that this is wrong, as I have an 8086:0412 device present.

 

If I change the CPU in Lilu to return haswell all is good.

 

There is also an OVMF rom that can init the IGPU too, here https://github.com/jam3st/edk2

[Jingu]

I understand that WhateverGreen 1.2.0 is only the initial merge of IntelGraphicsFixup into WhatverGreen. I had imagined that the code and functionality of IntelGraphicsFixup would be fully merged, but I'm discovering that is not entirely true.

 

IntelGraphicsFixup is designed to work with Inject Intel = Yes and FakePCIID_Intel_HD_Graphics. So, the graphics enabling is done with the help of Inject Intel / FakePCIID and IntelGraphicsFixup only adds the necessary corrections on top of that.

 

With WhateverGreen, we must use Inject Intel = No and remove FakePCIID_Intel_HD_Graphics. So, in this case, the graphics enabling and the corrections are all done by WhateverGreen. That's different from just adding the necessary corrections.

 

Well, something isn't right. My desktop core i3 6300 iMac17,1 HD530 (10.13.6) boots to a black screen. The poster from 2 posts above mentionned that it has do to with WhateverGreen disabling AGDP, while IntelGraphicsFixup wasn't.

[XantheFIN]

I needed to delete as i resulted not working acceleration for my intel hd4600 igpu if i removed FakePCIID_Intel_HD_Graphics on haswell refresh cpu.

[vit9696]

21 hours ago, Roger Smith said:

I have a problem with CoreDisplay patches not being applied because of my VM configuration.

I have an HD Graphics passed through to Qemu via OVMF and Lilu says that my CPU is not supported. But I think that this is wrong, as I have an 8086:0412 device present.

 

If I change the CPU in Lilu to return haswell all is good.

 

There is also an OVMF rom that can init the IGPU too, here https://github.com/jam3st/edk2

Could you give me the CPUID your QEMU returns? If you pass-through a Haswell IGPU, you are pretty much expected to pass a Haswell CPU via a -cpu Haswell option. I pretty much expect a legit CPUID from this… I understand that it might be less obvious, but having a Haswell IGPU on a non-Haswell CPU is pretty much insane.

 

Spoiler

$ qemu-system-x86_64 -cpu help
Available CPUs:
x86              486                                                  
x86   Broadwell-IBRS  Intel Core Processor (Broadwell, IBRS)          
x86 Broadwell-noTSX-IBRS  Intel Core Processor (Broadwell, no TSX, IBRS)  
x86  Broadwell-noTSX  Intel Core Processor (Broadwell, no TSX)        
x86        Broadwell  Intel Core Processor (Broadwell)                
x86           Conroe  Intel Celeron_4x0 (Conroe/Merom Class Core 2)   
x86        EPYC-IBPB  AMD EPYC Processor (with IBPB)                  
x86             EPYC  AMD EPYC Processor                              
x86     Haswell-IBRS  Intel Core Processor (Haswell, IBRS)            
x86 Haswell-noTSX-IBRS  Intel Core Processor (Haswell, no TSX, IBRS)    
x86    Haswell-noTSX  Intel Core Processor (Haswell, no TSX)          
x86          Haswell  Intel Core Processor (Haswell)                  
x86   IvyBridge-IBRS  Intel Xeon E3-12xx v2 (Ivy Bridge, IBRS)        
x86        IvyBridge  Intel Xeon E3-12xx v2 (Ivy Bridge)              
x86     Nehalem-IBRS  Intel Core i7 9xx (Nehalem Core i7, IBRS update)
x86          Nehalem  Intel Core i7 9xx (Nehalem Class Core i7)       
x86       Opteron_G1  AMD Opteron 240 (Gen 1 Class Opteron)           
x86       Opteron_G2  AMD Opteron 22xx (Gen 2 Class Opteron)          
x86       Opteron_G3  AMD Opteron 23xx (Gen 3 Class Opteron)          
x86       Opteron_G4  AMD Opteron 62xx class CPU                      
x86       Opteron_G5  AMD Opteron 63xx class CPU                      
x86           Penryn  Intel Core 2 Duo P9xxx (Penryn Class Core 2)    
x86 SandyBridge-IBRS  Intel Xeon E312xx (Sandy Bridge, IBRS update)   
x86      SandyBridge  Intel Xeon E312xx (Sandy Bridge)                
x86 Skylake-Client-IBRS  Intel Core Processor (Skylake, IBRS)            
x86   Skylake-Client  Intel Core Processor (Skylake)                  
x86 Skylake-Server-IBRS  Intel Xeon Processor (Skylake, IBRS)            
x86   Skylake-Server  Intel Xeon Processor (Skylake)                  
x86    Westmere-IBRS  Westmere E56xx/L56xx/X56xx (IBRS update)        
x86         Westmere  Westmere E56xx/L56xx/X56xx (Nehalem-C)          
x86           athlon  QEMU Virtual CPU version 2.5+                   
x86         core2duo  Intel(R) Core(TM)2 Duo CPU     T7700  @ 2.40GHz 
x86          coreduo  Genuine Intel(R) CPU           T2600  @ 2.16GHz 
x86            kvm32  Common 32-bit KVM processor                     
x86            kvm64  Common KVM processor                            
x86             n270  Intel(R) Atom(TM) CPU N270   @ 1.60GHz          
x86          pentium                                                  
x86         pentium2                                                  
x86         pentium3                                                  
x86           phenom  AMD Phenom(tm) 9550 Quad-Core Processor         
x86           qemu32  QEMU Virtual CPU version 2.5+                   
x86           qemu64  QEMU Virtual CPU version 2.5+                   
x86             base  base CPU model type with no features enabled    
x86             host  Enables all features supported by the accelerator in the current host
x86              max  Enables all features supported by the accelerator in the current host

Recognized CPUID flags:
  fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 pn clflush ds acpi mmx fxsr sse sse2 ss ht tm ia64 pbe
  pni pclmulqdq dtes64 monitor ds-cpl vmx smx est tm2 ssse3 cid fma cx16 xtpr pdcm pcid dca sse4.1 sse4.2 x2apic movbe popcnt tsc-deadline aes xsave osxsave avx f16c rdrand hypervisor
  fsgsbase tsc-adjust bmi1 hle avx2 smep bmi2 erms invpcid rtm mpx avx512f avx512dq rdseed adx smap avx512ifma pcommit clflushopt clwb intel-pt avx512pf avx512er avx512cd sha-ni avx512bw avx512vl
  avx512vbmi umip pku ospke avx512vbmi2 gfni vaes vpclmulqdq avx512vnni avx512bitalg avx512-vpopcntdq la57 rdpid
  avx512-4vnniw avx512-4fmaps spec-ctrl
  syscall nx mmxext fxsr-opt pdpe1gb rdtscp lm 3dnowext 3dnow
  lahf-lm cmp-legacy svm extapic cr8legacy abm sse4a misalignsse 3dnowprefetch osvw ibs xop skinit wdt lwp fma4 tce nodeid-msr tbm topoext perfctr-core perfctr-nb
  invtsc
  ibpb
  xstore xstore-en xcrypt xcrypt-en ace2 ace2-en phe phe-en pmm pmm-en
  kvmclock kvm-nopiodelay kvm-mmu kvmclock kvm-asyncpf kvm-steal-time kvm-pv-eoi kvm-pv-unhalt kvm-pv-tlb-flush kvmclock-stable-bit
  kvm-hint-dedicated
  
  
  
  npt lbrv svm-lock nrip-save tsc-scale vmcb-clean flushbyasid decodeassists pause-filter pfthreshold
  xsaveopt xsavec xgetbv1 xsaves
  arat
  
  
 

 

Regarding black screen issues, it looks like CDF code was causing issues, and it is now optional in the latest master.

Also, if your IGPU device-id is not present in Apple kexts, you must pass a legit device-id via Clover properties.

[WinstonAce]

32 minutes ago, vit9696 said:

Regarding black screen issues, it looks like CDF code was causing issues, and it is now optional in the latest master.

Also, if your IGPU device-id is not present in Apple kexts, you must pass a legit device-id via Clover properties.

 

Thanks, that did the trick for me.

one question

do you recommend using clover to rename:

GFX0 to IGPU

PEGP to GFX0

or let WEG do the renaming?

 

TIA 

[vit9696]

Either way is fine, but I suppose using Green is commonly more reliable. The less you do, the more stable it works.

Edited July 29, 2018 by vit9696

[LockDown]

who is responsible for ngfxcompat=1 flag

Is it WEG, Nvidia web driver or the kernel?

Edited July 30, 2018 by ellaosx

[TheRacerMaster]

It's parsed by WhateverGreen (it overrides NVDARequiredOS with the current OS version when set).

[LockDown]

8 minutes ago, TheRacerMaster said:

It's parsed by WhateverGreen (it overrides NVDARequiredOS with the current OS version when set).

ah ok. so it means that the flag is useless w/o installing WhateverGreen. 

Thanks

Edited July 30, 2018 by ellaosx

[123456789323]

Guys anyone of you tried to use WEG with the newest Mojave Beta ? 

With Beta 4 everything worked with my UHD 630 set as Connecterless, with Beta 5 it doesn't work anymore.

 

EDIT: Problem fixed!

Edited July 30, 2018 by DSM2

[LockDown]

2 hours ago, DSM2 said:

Guys anyone of you tried to use WEG with the newest Mojave Beta ? 

With Beta 4 everything worked with my UHD 630 set as Connecterless, with Beta 5 it doesn't work anymore.

 

EDIT: Problem fixed!

what was your fix?

[xelanaiznac]

hi guys. is there someone who can pass vdadecoder checker with x99-x299?

is because igpu is mandatory?

no issue with my z370, I can pass vdadecoder checker with last version of WEG.

thank you for any help

[Pavo]

1 hour ago, xelanaiznac said:

hi guys. is there someone who can pass vdadecoder checker with x99-x299?

is because igpu is mandatory?

no issue with my z370, I can pass vdadecoder checker with last version of WEG.

thank you for any help

I can and I have no iGPU because I have Xeon CPUs. No issue with vdadecoder checker

Henrys-iMac-Pro:~ henrybrock$ ./Dropbox/VDADecoderChecker
Hardware acceleration is fully supported

 

VDADecoderChecker.zip

Edited July 31, 2018 by Pavo

[Roger Smith]

On 7/29/2018 at 8:26 PM, vit9696 said:

Could you give me the CPUID your QEMU returns? If you pass-through a Haswell IGPU, you are pretty much expected to pass a Haswell CPU via a -cpu Haswell option. I pretty much expect a legit CPUID from this… I understand that it might be less obvious, but having a Haswell IGPU on a non-Haswell CPU is pretty much insane.

 

  Reveal hidden contents

$ qemu-system-x86_64 -cpu help
Available CPUs:
x86              486                                                  
x86   Broadwell-IBRS  Intel Core Processor (Broadwell, IBRS)          
x86 Broadwell-noTSX-IBRS  Intel Core Processor (Broadwell, no TSX, IBRS)  
x86  Broadwell-noTSX  Intel Core Processor (Broadwell, no TSX)        
x86        Broadwell  Intel Core Processor (Broadwell)                
x86           Conroe  Intel Celeron_4x0 (Conroe/Merom Class Core 2)   
x86        EPYC-IBPB  AMD EPYC Processor (with IBPB)                  
x86             EPYC  AMD EPYC Processor                              
x86     Haswell-IBRS  Intel Core Processor (Haswell, IBRS)            
x86 Haswell-noTSX-IBRS  Intel Core Processor (Haswell, no TSX, IBRS)    
x86    Haswell-noTSX  Intel Core Processor (Haswell, no TSX)          
x86          Haswell  Intel Core Processor (Haswell)                  
x86   IvyBridge-IBRS  Intel Xeon E3-12xx v2 (Ivy Bridge, IBRS)        
x86        IvyBridge  Intel Xeon E3-12xx v2 (Ivy Bridge)              
x86     Nehalem-IBRS  Intel Core i7 9xx (Nehalem Core i7, IBRS update)
x86          Nehalem  Intel Core i7 9xx (Nehalem Class Core i7)       
x86       Opteron_G1  AMD Opteron 240 (Gen 1 Class Opteron)           
x86       Opteron_G2  AMD Opteron 22xx (Gen 2 Class Opteron)          
x86       Opteron_G3  AMD Opteron 23xx (Gen 3 Class Opteron)          
x86       Opteron_G4  AMD Opteron 62xx class CPU                      
x86       Opteron_G5  AMD Opteron 63xx class CPU                      
x86           Penryn  Intel Core 2 Duo P9xxx (Penryn Class Core 2)    
x86 SandyBridge-IBRS  Intel Xeon E312xx (Sandy Bridge, IBRS update)   
x86      SandyBridge  Intel Xeon E312xx (Sandy Bridge)                
x86 Skylake-Client-IBRS  Intel Core Processor (Skylake, IBRS)            
x86   Skylake-Client  Intel Core Processor (Skylake)                  
x86 Skylake-Server-IBRS  Intel Xeon Processor (Skylake, IBRS)            
x86   Skylake-Server  Intel Xeon Processor (Skylake)                  
x86    Westmere-IBRS  Westmere E56xx/L56xx/X56xx (IBRS update)        
x86         Westmere  Westmere E56xx/L56xx/X56xx (Nehalem-C)          
x86           athlon  QEMU Virtual CPU version 2.5+                   
x86         core2duo  Intel(R) Core(TM)2 Duo CPU     T7700  @ 2.40GHz 
x86          coreduo  Genuine Intel(R) CPU           T2600  @ 2.16GHz 
x86            kvm32  Common 32-bit KVM processor                     
x86            kvm64  Common KVM processor                            
x86             n270  Intel(R) Atom(TM) CPU N270   @ 1.60GHz          
x86          pentium                                                  
x86         pentium2                                                  
x86         pentium3                                                  
x86           phenom  AMD Phenom(tm) 9550 Quad-Core Processor         
x86           qemu32  QEMU Virtual CPU version 2.5+                   
x86           qemu64  QEMU Virtual CPU version 2.5+                   
x86             base  base CPU model type with no features enabled    
x86             host  Enables all features supported by the accelerator in the current host
x86              max  Enables all features supported by the accelerator in the current host

Recognized CPUID flags:
  fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 pn clflush ds acpi mmx fxsr sse sse2 ss ht tm ia64 pbe
  pni pclmulqdq dtes64 monitor ds-cpl vmx smx est tm2 ssse3 cid fma cx16 xtpr pdcm pcid dca sse4.1 sse4.2 x2apic movbe popcnt tsc-deadline aes xsave osxsave avx f16c rdrand hypervisor
  fsgsbase tsc-adjust bmi1 hle avx2 smep bmi2 erms invpcid rtm mpx avx512f avx512dq rdseed adx smap avx512ifma pcommit clflushopt clwb intel-pt avx512pf avx512er avx512cd sha-ni avx512bw avx512vl
  avx512vbmi umip pku ospke avx512vbmi2 gfni vaes vpclmulqdq avx512vnni avx512bitalg avx512-vpopcntdq la57 rdpid
  avx512-4vnniw avx512-4fmaps spec-ctrl
  syscall nx mmxext fxsr-opt pdpe1gb rdtscp lm 3dnowext 3dnow
  lahf-lm cmp-legacy svm extapic cr8legacy abm sse4a misalignsse 3dnowprefetch osvw ibs xop skinit wdt lwp fma4 tce nodeid-msr tbm topoext perfctr-core perfctr-nb
  invtsc
  ibpb
  xstore xstore-en xcrypt xcrypt-en ace2 ace2-en phe phe-en pmm pmm-en
  kvmclock kvm-nopiodelay kvm-mmu kvmclock kvm-asyncpf kvm-steal-time kvm-pv-eoi kvm-pv-unhalt kvm-pv-tlb-flush kvmclock-stable-bit
  kvm-hint-dedicated
  
  
  
  npt lbrv svm-lock nrip-save tsc-scale vmcb-clean flushbyasid decodeassists pause-filter pfthreshold
  xsaveopt xsavec xgetbv1 xsaves
  arat
  
  
 

 

Regarding black screen issues, it looks like CDF code was causing issues, and it is now optional in the latest master.

Also, if your IGPU device-id is not present in Apple kexts, you must pass a legit device-id via Clover properties.

 

Hi

 

Qemu supports the Penryn CPU. I think that there are some MSRs that cause CPU host not to be supported by MacOS.  So WhateverGreen says you don't have an HD Graphics 4600 because you are using Penryn. I can work around that. If I manually enable the IGFX patches I get unacelerated output.

 

To enable acceleration I also need to remove the MSR 0x35 from the AppleIntelHD5000Graphics kext. I think that is used to determine the number of cores, so I just return 4.

 

ie patch  from:

 

hexdump AppleIntelHD5000Graphics.orig | diff - =(hexdump AppleIntelHD5000Graphics.patch )
8111c8111
< 00200b0 3100 31db 31c9 0fd2 b9a2 0035 0000 320f
---
> 00200b0 3100 31db 31c9 0fd2 b8a2 0008 0004 9090

 

Without this patch I don't get acceleration on the VM.

 

[vit9696]

So, do I understand the problem correctly, that you cannot use -cpu Haswell because of QEMU/XNU compatibility issues, and are forced to use Penryn?

 

If this is the case, I added a workaround with the use of a boot-arg — lilucpu, allowing one to specify the emulated CPU generation. lilucpu=6 will do for Haswell. It may be imperfect, but there are several reasons for it:

— Detecting the GPU by its own device-id is an unfortunate maintenance burden, as the amount of IGPU device-ids per each generation is extremely large, Intel continues to add them, and even now we do not know them all.

— Being ready to patch every kext for every IGPU generation slows the boot process down and increases the memory footprint, so WhateverGreen tries to do best to optimise it.

 

Since the existence of a Haswell IGPU on Penryn is not really possible on real hardware, it feels fair for me. Another possible use for an option could be some newly announced CPU, which is faked to use the drivers from a previous generation, so it is probably good to have anyway.

Edited August 1, 2018 by vit9696

[Roger Smith]

Thats awesome. More than what I was expecting. Yeah I was thinking about the case when you have different IP on the same die in the future (eg core x gpu y etc). Much appreciated and much more than I expected. Thanks again.

[xelanaiznac]

10 hours ago, Pavo said:

I can and I have no iGPU because I have Xeon CPUs. No issue with vdadecoder checker

Henrys-iMac-Pro:~ henrybrock$ ./Dropbox/VDADecoderChecker
Hardware acceleration is fully supported

 

VDADecoderChecker.zip

 

thank you Pavo. I found out that if I use vega64, my hardware acceleration is fully supported.

as soon as I put my Nvidia gtx 1080, I have the error as picture

this happens on my x99 and on my x299 config (tried in both systems).

 

no issue with my z370 with both amd or nvidia