MP 1,1-5,1 Manually Configured OpenCore on the Mac Pro

[cdf]

It isn't surprising you get a boot loop when using two different shikigva boot-args since that is not how it works. The value of shikigva is a sum and the result by adding parameters which each have a different value.

To clarify: Originally, @h9826790 had found 96 (32+64) to be a good choice for the cMP. With WEG 1.3.5, I tried 112 (16+32+64) and 240 (16+32+64+128), both causing a boot loop. It turns out that the presence of 32 with 16 causes the issue.

Another thing to test. Does adding -wegtree boot-arg trigger proper device renaming for your Mac Pros? GPU should be called GFX0 in IOReg if it does.

Indeed, it does:

 ...
 +-o IOU0@3  <class IOPCIDevice, id 0x1000001cb, registered, matched, active, busy 0 (960 ms), retain 12>
    | |   | +-o IOPP  <class IOPCI2PCIBridge, id 0x100000219, registered, matched, active, busy 0 (959 ms), retain 10>
    | |   |   +-o GFX0@0  <class IOPCIDevice, id 0x1000001cc, registered, matched, active, busy 0 (956 ms), retain 32>
    | |   |   | +-o AMDSupport  <class AMDSupport, id 0x100000412, registered, matched, active, busy 0 (0 ms), retain 5>
    ...

 

[CMMChris]

Great, so I would suggest that everyone goes ahead and reports the issues with WEG on the Insanelymac topic. Since -wegtree does trigger the appropriate renames, once developers realize there is an issue and fix the non-working shiki boot-args you guys will be able to use full video acceleration without the need for OpenCore or Clover. I personally can't do much regarding that on my own since nobody seems to believe me. LOL

 

[startergo]

Great, so I would suggest that everyone goes ahead and reports the issues with WEG on the Insanelymac topic. Since -wegtree does trigger the appropriate renames, once developers realize there is an issue and fix the non-working shiki boot-args you guys will be able to use full video acceleration without the need for OpenCore or Clover. I personally can't do much regarding that on my own since nobody seems to believe me. LOL

This is just 6 days ago:

@@ -133,8 +133,8 @@ Add `shikigva=16` argument to boot-args or to DeviceProperties in any GPU.

- _How can I play iTunes purchased videos in QuickTime on MacPro5,1 along with Apple TV+?_ 
For QuickTime movie playback along with TV+ on MacPro5,1 use one of the following: 
`shikigva=0` and OpenCore spoof to iMacPro1,1 (preferred). 
`shikigva=32 shiki-id=Mac-7BA5B2D9E42DDD94` without OpenCore.
OpenCore spoof to iMacPro1,1 (preferred). 
`shikigva=160 shiki-id=Mac-7BA5B2D9E42DDD94` without OpenCore.

- _How can I inject IOVARendererID/IOVARendererSubID in certain NVIDIA GPUs?_ 
NVIDIA drivers do not properly add these values necessary for VDA decoding for Maxwell and Pascal

 

[CMMChris]

@startergo What are you trying to say? None of this works, neither on a real Mac Pro nor on a Hackintosh. Shiki is broken right now. Some people in here tested it, a bunch of other people from the Hackintosh scene tested it, I personally tested it on my Hackintosh. I am trying to get the WEG developers on this issue for a couple of days now but as I said above, nobody does believe me.

 

[startergo]

None of this works, neither on a real Mac Pro nor on a Hackintosh

Can you please elaborate what exactly is not working? I am testing it now.
Playback of HEVC works. Does somebody know how to insert the onscreen debug messages from Lilu to a log file? I use -liludbgall argument and the messages scroll on the screen during boot.

Edit:
Enabled debugging log with this argument:

• Add liludump=N to let Lilu DEBUG version dump log to /var/log/Lilu_VERSION_KERN_MAJOR.KERN_MINOR.txtafter N seconds

WhateverGreen:     weg @ (DBG) apple-fw proceeding with devprops due to arg
Lilu:   iokit @ (DBG) compatible property starts with pci1028,1701 and is 47 bytes
Lilu:   iokit @ (DBG) fixing compatible to have GFX0
WhateverGreen:   iokit @ (DBG) getOSData device-id has 67DF value
WhateverGreen:   iokit @ (DBG) getOSData revision-id has C7 value
WhateverGreen:   iokit @ (DBG) getOSData subsystem-vendor-id has 1028 value
WhateverGreen:   iokit @ (DBG) getOSData subsystem-id has 1701 value
WhateverGreen:     weg @ (DBG) fixing built-in
Lilu:    mach @ (DBG) found symbol __ZN15IORegistryEntry11setPropertyEPKcPvj at 0xffffff8002e203b0 (non-aslr 0xffffff8000a203b0)
Lilu:    mach @ (DBG) found symbol __ZNK15IORegistryEntry11getPropertyEPKc at 0xffffff8002e20600 (non-aslr 0xffffff8000a20600)
Lilu: patcher @ (DBG) from 0xFFFFFF8002E203B0 to 0xFFFFFF7F8775FD40 diff 0xFFFFFFFF8493F98B argument 8493F98B
Lilu: patcher @ (DBG) from 0xFFFFFF7F876D8296 to 0xFFFFFF8002E203B6 diff 0x000000007B74811B argument 7B74811B
Lilu: patcher @ (DBG) wrapped __ZN15IORegistryEntry11setPropertyEPKcPvj
Lilu: patcher @ (DBG) from 0xFFFFFF8002E20600 to 0xFFFFFF7F8775FF70 diff 0xFFFFFFFF8493F96B argument 8493F96B
Lilu: patcher @ (DBG) from 0xFFFFFF7F876D82AA to 0xFFFFFF8002E20606 diff 0x000000007B748357 argument 7B748357
Lilu: patcher @ (DBG) wrapped __ZNK15IORegistryEntry11getPropertyEPKc
WhateverGreen:   shiki @ (DBG) pre-config: online 0, bgra 0, compat 0, whitelist 0, id 1, stream 0, hwdrm 0
WhateverGreen:   shiki @ (DBG) requesting Mac-7BA5B2D9E42DDD94 board-id for gva
WhateverGreen:   shiki @ (DBG) changing shiki-id to Mac-7BA5B2D9E42DDD94
Lilu:    mach @ (DBG) MachInfo asKernel 0 object constructed
Lilu:    mach @ (DBG) readMachHeader for /System/Library/Extensions/AMDSupport.kext/Contents/MacOS/AMDSupport
Lilu:    mach @ (DBG) readMachHeader got magic FEEDFACF
Lilu:    mach @ (DBG) found executable at path: /System/Library/Extensions/AMDSupport.kext/Contents/MacOS/AMDSupport
Lilu:    mach @ (DBG) header processing found TEXT
Lilu:    mach @ (DBG) header processing found LINKEDIT
Lilu:    mach @ (DBG) header processing found SYMTAB
Lilu: patcher @ (DBG) loaded kinfo com.apple.kext.AMDSupport at 1 index
Lilu:    mach @ (DBG) found symbol _gLoadedKextSummaries at 0xffffff8003220868 (non-aslr 0xffffff8000e20868)
Lilu: patcher @ (DBG) _gLoadedKextSummaries address 0xffffff8003220868
Lilu:    mach @ (DBG) found symbol __ZN6OSKext25updateLoadedKextSummariesEv at 0xffffff8002ddd180 (non-aslr 0xffffff80009dd180)
Lilu: patcher @ (DBG) kext summaries (1) address FFFFFF8002DDD180 value 56415741E5894855
Lilu: patcher @ (DBG) from 0xFFFFFF8002DDD180 to 0xFFFFFF7F876D50E0 diff 0xFFFFFFFF848F7F5B argument 848F7F5B
Lilu: patcher @ (DBG) from 0xFFFFFF7F876D82C9 to 0xFFFFFF8002DDD191 diff 0x000000007B704EC3 argument 7B704EC3
Lilu:    mach @ (DBG) MachInfo asKernel 0 object constructed
Lilu:    mach @ (DBG) readMachHeader for /System/Library/Extensions/AMD9500Controller.kext/Contents/MacOS/AMD9500Controller
Lilu:    mach @ (DBG) readMachHeader got magic FEEDFACF
Lilu:    mach @ (DBG) found executable at path: /System/Library/Extensions/AMD9500Controller.kext/Contents/MacOS/AMD9500Controller
Lilu:    mach @ (DBG) header processing found TEXT
Lilu:    mach @ (DBG) header processing found LINKEDIT
Lilu:    mach @ (DBG) header processing found SYMTAB
Lilu: patcher @ (DBG) loaded kinfo com.apple.kext.AMD9500Controller at 2 index
Lilu:    mach @ (DBG) MachInfo asKernel 0 object constructed
Lilu:    mach @ (DBG) readMachHeader for /System/Library/Extensions/IOPlatformPluginFamily.kext/Contents/PlugIns/ACPI_SMC_PlatformPlugin.kext/Contents/MacOS/ACPI_SMC_PlatformPlugin
Lilu:    mach @ (DBG) readMachHeader got magic FEEDFACF
Lilu:    mach @ (DBG) found executable at path: /System/Library/Extensions/IOPlatformPluginFamily.kext/Contents/PlugIns/ACPI_SMC_PlatformPlugin.kext/Contents/MacOS/ACPI_SMC_PlatformPlugin
Lilu:    mach @ (DBG) header processing found TEXT
Lilu:    mach @ (DBG) header processing found LINKEDIT
Lilu:    mach @ (DBG) header processing found SYMTAB
Lilu: patcher @ (DBG) loaded kinfo com.apple.driver.ACPI_SMC_PlatformPlugin at 3 index
Lilu:    mach @ (DBG) MachInfo asKernel 0 object constructed
Lilu:    mach @ (DBG) readMachHeader for /System/Library/Extensions/IOPlatformPluginFamily.kext/Contents/PlugIns/X86PlatformPlugin.kext/Contents/MacOS/X86PlatformPlugin
Lilu:    mach @ (DBG) readMachHeader got magic FEEDFACF
Lilu:    mach @ (DBG) found executable at path: /System/Library/Extensions/IOPlatformPluginFamily.kext/Contents/PlugIns/X86PlatformPlugin.kext/Contents/MacOS/X86PlatformPlugin
Lilu:    mach @ (DBG) header processing found TEXT
Lilu:    mach @ (DBG) header processing found LINKEDIT
Lilu:    mach @ (DBG) header processing found SYMTAB
Lilu: patcher @ (DBG) loaded kinfo com.apple.driver.X86PlatformPlugin at 4 index
Lilu:    mach @ (DBG) MachInfo asKernel 0 object constructed
Lilu:    mach @ (DBG) readMachHeader for /System/Library/Extensions/AppleIntelMCEReporter.kext/Contents/MacOS/AppleIntelMCEReporter
Lilu:    mach @ (DBG) readMachHeader got magic FEEDFACF
Lilu:    mach @ (DBG) found executable at path: /System/Library/Extensions/AppleIntelMCEReporter.kext/Contents/MacOS/AppleIntelMCEReporter
Lilu:    mach @ (DBG) header processing found TEXT
Lilu:    mach @ (DBG) header processing found LINKEDIT
Lilu:    mach @ (DBG) header processing found SYMTAB
Lilu: patcher @ (DBG) loaded kinfo com.apple.driver.AppleIntelMCEReporter at 5 index
Lilu:    user @ (DBG) loadFilesForPatching 5
Lilu:    user @ (DBG) requesting file /System/Library/Frameworks/VideoToolbox.framework/Versions/A/VideoToolbox at 0
Lilu:    user @ (DBG) have 6 mods for /System/Library/Frameworks/VideoToolbox.framework/Versions/A/VideoToolbox (read as 7861200)
Lilu:    mach @ (DBG) found section __text size 0 in segment 0
Lilu:    user @ (DBG) findSectionBounds returned vmsegment 0 vmsection 1960 sectionptr 0xffffff8e19616960 size 3245695
Lilu:    user @ (DBG) this patch will start from 0 entry and will replace 1 findings
Lilu:    user @ (DBG) found entry of BF 69 patch
Lilu:    user @ (DBG) using it off 51916 pageOff 516A0 new 53000 segOff 53276
Lilu:    user @ (DBG) first page bytes are 86 77 2C 0 48 83 C4 10
Lilu:    user @ (DBG) ref find 0
Lilu:    user @ (DBG) pushing off 276 to patch
Lilu:    mach @ (DBG) found section __text size 0 in segment 0
Lilu:    user @ (DBG) findSectionBounds returned vmsegment 0 vmsection 1960 sectionptr 0xffffff8e19616960 size 3245695
Lilu:    user @ (DBG) this patch will start from 0 entry and will replace 1 findings
Lilu:    user @ (DBG) found entry of BF 69 patch
Lilu:    user @ (DBG) using it off 51980 pageOff 516A0 new 53000 segOff 532E0
Lilu:    user @ (DBG) ref find 0
Lilu:    user @ (DBG) pushing off 2E0 to patch
Lilu:    user @ (DBG) skipping not requested patch /System/Library/Frameworks/VideoToolbox.framework/Versions/A/VideoToolbox for 2
Lilu:    user @ (DBG) skipping not requested patch /System/Library/Frameworks/VideoToolbox.framework/Versions/A/VideoToolbox for 3
Lilu:    user @ (DBG) skipping not requested patch /System/Library/Frameworks/VideoToolbox.framework/Versions/A/VideoToolbox for 4
Lilu:    user @ (DBG) skipping not requested patch /System/Library/Frameworks/VideoToolbox.framework/Versions/A/VideoToolbox for 5
Lilu:    user @ (DBG) requesting file /System/Library/PrivateFrameworks/CoreFP.framework/Versions/A/CoreFP at 1
Lilu:    user @ (DBG) have 1 mods for /System/Library/PrivateFrameworks/CoreFP.framework/Versions/A/CoreFP (read as 44294016)
Lilu:    mach @ (DBG) found section __text size 0 in segment 0
Lilu:    user @ (DBG) findSectionBounds returned vmsegment 0 vmsection 17D0 sectionptr 0xffffff8e1ab2d7d0 size 21423781
Lilu:    user @ (DBG) this patch will start from 0 entry and will replace 1 findings
Lilu:    user @ (DBG) found entry of 1 0 patch
Lilu:    user @ (DBG) using it off 5FF5A8 pageOff 5FE830 new 600000 segOff 600D78
Lilu:    user @ (DBG) first page bytes are E0 43 33 14 BA 48 8B 8D
Lilu:    user @ (DBG) ref find 0
Lilu:    user @ (DBG) pushing off D78 to patch
Lilu:    user @ (DBG) requesting file /System/Library/PrivateFrameworks/AppleGVA.framework/Versions/A/AppleGVA at 2
Lilu:    user @ (DBG) have 5 mods for /System/Library/PrivateFrameworks/AppleGVA.framework/Versions/A/AppleGVA (read as 2087888)
Lilu:    user @ (DBG) skipping not requested patch /System/Library/PrivateFrameworks/AppleGVA.framework/Versions/A/AppleGVA for 0
Lilu:    user @ (DBG) skipping not requested patch /System/Library/PrivateFrameworks/AppleGVA.framework/Versions/A/AppleGVA for 1
Lilu:    user @ (DBG) skipping not requested patch /System/Library/PrivateFrameworks/AppleGVA.framework/Versions/A/AppleGVA for 2
Lilu:    mach @ (DBG) found section __cstring size 5 in segment 0
Lilu:    user @ (DBG) findSectionBounds returned vmsegment 0 vmsection C9900 sectionptr 0xffffff8e196de900 size 100603
Lilu:    user @ (DBG) this patch will start from 0 entry and will replace 1 findings
Lilu:    user @ (DBG) found entry of 62 6F patch
Lilu:    user @ (DBG) using it off 16242 pageOff 15700 new DF000 segOff DFB42
Lilu:    user @ (DBG) first page bytes are 4F 62 74 61 69 6E 65 64
Lilu:    user @ (DBG) ref find 0
Lilu:    user @ (DBG) pushing off B42 to patch
Lilu:    user @ (DBG) skipping not requested patch /System/Library/PrivateFrameworks/AppleGVA.framework/Versions/A/AppleGVA for 4
Lilu:    user @ (DBG) requesting file /System/Library/PrivateFrameworks/AppleVPA.framework/Versions/A/AppleVPA at 3
Lilu:    user @ (DBG) have 1 mods for /System/Library/PrivateFrameworks/AppleVPA.framework/Versions/A/AppleVPA (read as 999312)
Lilu:    mach @ (DBG) found section __cstring size 5 in segment 0
Lilu:    user @ (DBG) findSectionBounds returned vmsegment 0 vmsection 39E60 sectionptr 0xffffff8e1964de60 size 21216
Lilu:    user @ (DBG) this patch will start from 0 entry and will replace 1 findings
Lilu:    user @ (DBG) found entry of 62 6F patch
Lilu:    user @ (DBG) using it off 387A pageOff 31A0 new 3D000 segOff 3D6DA
Lilu:    user @ (DBG) first page bytes are 20 45 72 72 6F 72 0 55
Lilu:    user @ (DBG) ref find 0
Lilu:    user @ (DBG) pushing off 6DA to patch
Lilu:    user @ (DBG) ignoring file /System/Library/Frameworks/CoreDisplay.framework/Versions/A/CoreDisplay at 4, no mods out of 1 apply
Lilu:    user @ (DBG) loadDyldSharedCacheMapping 5
Lilu:    user @ (DBG) mapped 4 entries out of 5
Lilu:    user @ (DBG) loading lookup 0 current off is 0
Lilu:    user @ (DBG) successful finding at 0
Lilu:    user @ (DBG) loading lookup 1 current off is 0
Lilu:    user @ (DBG) matched off 0 with 0th
Lilu:    user @ (DBG) chose 8
Lilu:    user @ (DBG) loading lookup 2 current off is 10
Lilu:    user @ (DBG) chose 10
Lilu:    user @ (DBG) loading lookup 3 current off is 18
Lilu:    user @ (DBG) chose 18
Lilu:    mach @ (DBG) found symbol _cs_validate_range at 0xffffff8002d08730 (non-aslr 0xffffff8000908730)
Lilu: patcher @ (DBG) from 0xFFFFFF8002D08730 to 0xFFFFFF7F876F2790 diff 0xFFFFFFFF849EA05B argument 849EA05B
Lilu: patcher @ (DBG) from 0xFFFFFF7F876D82EB to 0xFFFFFF8002D08744 diff 0x000000007B630454 argument 7B630454
Lilu:    mach @ (DBG) found symbol _current_map at 0xffffff800285a570 (non-aslr 0xffffff800045a570)
Lilu:    mach @ (DBG) found symbol _get_map_min at 0xffffff800280b4a0 (non-aslr 0xffffff800040b4a0)
Lilu:    mach @ (DBG) found symbol _get_task_map at 0xffffff800280ab40 (non-aslr 0xffffff800040ab40)
Lilu:    mach @ (DBG) found symbol _vm_map_switch_protect at 0xffffff800285a820 (non-aslr 0xffffff800045a820)
Lilu:    mach @ (DBG) found symbol _vm_map_check_protection at 0xffffff8002859680 (non-aslr 0xffffff8000459680)
Lilu:    mach @ (DBG) found symbol _vm_map_read_user at 0xffffff8002859590 (non-aslr 0xffffff8000459590)
Lilu:    mach @ (DBG) found symbol _vm_map_write_user at 0xffffff80028594a0 (non-aslr 0xffffff80004594a0)
Lilu:    mach @ (DBG) found symbol _task_set_main_thread_qos at 0xffffff80027f5ce0 (non-aslr 0xffffff80003f5ce0)
Lilu: patcher @ (DBG) from 0xFFFFFF80027F5CE0 to 0xFFFFFF7F876F4C90 diff 0xFFFFFFFF84EFEFAB argument 84EFEFAB
Lilu: patcher @ (DBG) from 0xFFFFFF7F876D82FF to 0xFFFFFF80027F5CE6 diff 0x000000007B11D9E2 argument 7B11D9E2
Lilu: patcher @ (DBG) wrapped _task_set_main_thread_qos
Lilu:    mach @ (DBG) found symbol _vm_shared_region_map_file at 0xffffff8002888460 (non-aslr 0xffffff8000488460)
Lilu: patcher @ (DBG) from 0xFFFFFF8002888460 to 0xFFFFFF7F876F4930 diff 0xFFFFFFFF84E6C4CB argument 84E6C4CB
Lilu: patcher @ (DBG) from 0xFFFFFF7F876D8321 to 0xFFFFFF8002888474 diff 0x000000007B1B014E argument 7B1B014E
Lilu:    mach @ (DBG) found symbol _vm_shared_region_slide at 0xffffff8002889000 (non-aslr 0xffffff8000489000)
Lilu: patcher @ (DBG) from 0xFFFFFF8002889000 to 0xFFFFFF7F876F4B50 diff 0xFFFFFFFF84E6BB4B argument 84E6BB4B
Lilu: patcher @ (DBG) from 0xFFFFFF7F876D8340 to 0xFFFFFF8002889011 diff 0x000000007B1B0CCC argument 7B1B0CCC
Lilu:    user @ (DBG) params are B562000 37E4E000 E0DF000 7FFFCB562000 1D000
Lilu:    user @ (DBG) patch for /System/Library/Frameworks/VideoToolbox.framework/Versions/A/VideoToolbox in 7FFF37B1D000 7FFF37E7F000
Lilu:    user @ (DBG) found a possible match for 0 of 10C48348002C7786
Lilu:    user @ (DBG) found a possible match for 1 of 874F6854DC78941
Lilu:    user @ (DBG) found a possible match for 2 of 2C71A8E8F7894C
Lilu:    user @ (DBG) found a possible match for 3 of E7894C0874E4854D
Lilu:    user @ (DBG) failed to match a complete page with 0
Lilu:    user @ (DBG) 1/1 found BF 69 6D 72
Lilu:    user @ (DBG) found request for W^X switch-off 7
Lilu:    user @ (DBG) found p_csflags offset 308
Lilu:    user @ (DBG) obtained write permssions
Lilu:    user @ (DBG) restored write permssions
Lilu:    user @ (DBG) done reading patches for 53276
Lilu:    user @ (DBG) patch for /System/Library/Frameworks/VideoToolbox.framework/Versions/A/VideoToolbox in 7FFF37B1D000 7FFF37E7F000
Lilu:    user @ (DBG) 1/1 found BF 69 6D 72
Lilu:    user @ (DBG) found request for W^X switch-off 7
Lilu:    user @ (DBG) obtained write permssions
Lilu:    user @ (DBG) restored write permssions
Lilu:    user @ (DBG) done reading patches for 532E0
Lilu:    user @ (DBG) patch for /System/Library/PrivateFrameworks/AppleGVA.framework/Versions/A/AppleGVA in 7FFF3BD1A000 7FFF3BE01000
Lilu:    user @ (DBG) found a possible match for 0 of 64656E696174624F
Lilu:    user @ (DBG) found a possible match for 1 of 2F0054566D6F7246
Lilu:    user @ (DBG) found a possible match for 2 of 6F6F52646C697542
Lilu:    user @ (DBG) found a possible match for 3 of 72617262694C2F74
Lilu:    user @ (DBG) found what we are looking for 62 6F 61 72 64 2D 69 64
Lilu:    user @ (DBG) obtained write permssions
Lilu:    user @ (DBG) restored write permssions
Lilu:    user @ (DBG) 1/0 found 73 68 69 6B
Lilu:    user @ (DBG) miss at 0: 73 vs 62
Lilu:    user @ (DBG) done reading patches for DFB42
Lilu:    user @ (DBG) patch for /System/Library/PrivateFrameworks/AppleVPA.framework/Versions/A/AppleVPA in 7FFF3C235000 7FFF3C275000
Lilu:    user @ (DBG) found a possible match for 0 of 5500726F72724520
Lilu:    user @ (DBG) found a possible match for 1 of 74726F707075736E
Lilu:    user @ (DBG) found a possible match for 2 of 5600495041206465
Lilu:    user @ (DBG) found a possible match for 3 of 3A6F666E69204150
Lilu:    user @ (DBG) found what we are looking for 62 6F 61 72 64 2D 69 64
Lilu:    user @ (DBG) obtained write permssions
Lilu:    user @ (DBG) restored write permssions
Lilu:    user @ (DBG) 1/0 found 73 68 69 6B
Lilu:    user @ (DBG) miss at 0: 73 vs 62
Lilu:    user @ (DBG) done reading patches for 3D6DA

 

[CMMChris]

Can you please elaborate what exactly is not working? I am testing it now.

Spoofing iMacPro1,1 for AppleGVA doesn't work which is necessary to get full H.264 and H.265 acceleration on supported AMD GPUs.

Playback of HEVC works.

But CPU decoded and thus non-accelerated.

 

[startergo]

But CPU decoded and thus non-accelerated.

No, GPU decoded.

 

[CMMChris]

In this case please tell us:
- How did you check if your AMD does decode and encode HEVC
- How did you get it to work
- Share some kind of proof

 

[startergo]

Try these.

 

[CMMChris]

You are the first one who claims that the old way of spoofing iMacPro1,1 with shikigva=32 still works (it used to in the past, now it doesn't). Already had some guys here in this forum try it and it didn't work on Catalina. Same applies to a ton of Hackintosh users. WEG developers came up with new shikigva values now which are supposed to work but they don't work as well. See here for reference: https://www.insanelymac.com/forum/t...support-topic/?do=findComment&comment=2699100

So as requested please provide a proof for your words. If you are on Mojave this might be the reason it works for you.

 

[startergo]

If you are on Mojave this might be the reason it works for you.

Indeed I am testing on Mojave. 160 works too.

 

[Pressure]

Try these.

Wouldn't that require the GPU to be in slot-1?

 

[startergo]

Wouldn't that require the GPU to be in slot-1?

Please elaborate. What requires the GPU to be in slot 1?

 

[CMMChris]

Slot is not relevant, just the rename to GFX0 is, otherwise no HEVC. @Pressure

 

[startergo]

-wegtree renames slot 1 to GFX0 and slot 2 to GFX1.

 

[Pressure]

Please elaborate. What requires the GPU to be in slot 1?

Slot is not relevant, just the rename to GFX0 is, otherwise no HEVC. @Pressure

-wegtree renames slot 1 to GFX0 and slot 2 to GFX1.

It's just my GPU is in slot-2 for better airflow around it. The config.plist file says PSX1 to GFX0 in the files @startergo attached.

 

[startergo]

The config.plist file says PSX1 to GFX0 in the files

That patch is not enabled. -wegtree takes care of everything.


Catalina HEVC playback.

[automerge]1575056384[/automerge]

FCPX HEVC playback. Of course 10.15.1. I believe @h9826790 said things are broken in later Catalina development release. No official release yet. Maybe the things will get better.

 

[CMMChris]

HEVC acceleration on the AMD is not working on your machine as for anyone else. VideoProc shows red light, CPU load is high. The GPU load displayed in Activity Monitor is something else. Decoder / Encoder load isn't displayed there.

 

[startergo]

CPU load is high.

Where did you see that?

 

[startergo]

So this is what was recommended to us by the developers:
"No good, I guess… In this case you can try, but it is hard to say what will the results be. Basically you need shikigva=160, and in addition to that you also need to get other parts patched, not just the AppleGVA decoder. At least that would include AppleGVACore, AppleVPA, but most likely something else. The potential candidates for patching can be found by grepping board-id in Frameworks, PrivateFrameworks, /usr.


The patch you need is here: https://github.com/acidanthera/WhateverGreen/blob/10bea0d/Resources/Patches.plist#L242, just rework Patches.plist and rebuild with Clean (Cleaning before Build is important, as Xcode 11 broke dependency solving).





On the other side, your power management is broken because of /System/Library/Extensions/IOPlatformPluginFamily.kext/Contents/PlugIns/ACPI_SMC_PlatformPlugin.kext. It does not have a plist file in Resources for iMacPro1,1. For this you can use our https://github.com/acidanthera/CPUFriend."

 

[CMMChris]

Yeah, but the thing is, what vit wrote there is ********. I honestly am kinda of shocked this guy doesn't even know the extension he is developing for. If you go back in the topic a bit you can see that he doesn't really know what he is talking about. He isn't even aware which AMD cards are natively supported for encoding / decoding and thus DRM in macOS. How is he supposed to help with such issues then?

Regarding his last statements:
- AppleGVA patching is fully implemented in Whatevergreen (that is what the shiki boot-arg is mainly for)
- AppleVPA (GPU based picture rendering) patching has been added to Whatevergreen earlier this year after I suggested it to them and explained the mechanism behind it
- No other frameworks or components need to be patched as spoofing iMacPro1,1 board-id for those components works absolutely fine in Mojave, resulting in full utilization of the AMD encoder / decoder and DRM support even when running SMBIOSes without native support for it or SMBIOSes which are designed to utilize an IGPU

The only issue here is that Whatevergreen or more specifically the Shiki part isn't compatible with Catalina anymore. They simply need to fix this, but if they don't even acknowledge there is an issue that of course won't happen.

 

[cdf]

- No other frameworks or components need to be patched as spoofing iMacPro1,1 board-id for those components works absolutely fine in Mojave, resulting in full utilization of the AMD encoder / decoder and DRM support even when running SMBIOSes without native support for it or SMBIOSes which are designed to utilize an IGPU

Perhaps I’m not recalling the situation with Mojave correctly, but hardware acceleration, let alone DRM support, has never fully worked by just patching these components in a cMP. Only when we changed the board-id in the SMBIOS using OpenCore did we finally get HEVC encoding, and I’m not sure we even tested DRM. Therefore, looking for occurrences of board-id in other components seems appropriate, at least for the cMP.

 

[CMMChris]

The reason HEVC did not work was due to the missing GFX0 rename. Without it only H.264 will work. This can now be fixed with the new -wegtree boot arg.

 

[h9826790]

The reason HEVC did not work was due to the missing GFX0 rename. Without it only H.264 will work. This can now be fixed with the new -wegtree boot arg.

It's more than that.

Even with GFX0 renaming, and iMac Pro board-id injection by using WEG, still no HEVC hardware encoding.

And HEVC hardware decoding can work regardless GFX0 or not (by using WEG board-id injection). However, in this case, VideoProc cannot detect it properly.

 

[startergo]

The reason HEVC did not work was due to the missing GFX0 rename. Without it only H.264 will work. This can now be fixed with the new -wegtree boot arg.

I am using -wegtree argument. Is it working now for you in the Catalina development release?