Right. So how does that explain the disappearance of onboard audio when a metal enabled gpu is used?
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
MP 1,1-5,1 OpenCore and the 2008 Mac Pro 3,1
- Thread starter amstel78
- WikiPost WikiPost
- Start date
- Sort by reaction score
- Status
That's strange. In this case their's an HDA for onboard audio and a HDA on the GPU. Both should be visible in ioreg. If they are working then a driver should be attached to each of the HDA PCIe devices.
OCLP for MacPro3,1 is using AppleALC for BigSur audio compatibility? I would start there. Test with and without AppleALC to see how the ioreg appears in each case.
i tried our OC package with an hd7870 and gtx770, both cards are flashed and got no audio issues...
but im not using any patcher, just the OpenCore: Project Legacy package. ?♂️
but im not using any patcher, just the OpenCore: Project Legacy package. ?♂️
Hi
Is there any tutorial of how to install-setup pciutils on big sur? I want to enable pice 2.0 speed on slot 2 but I can't install it.
I have found I think 3 different instances on GitHub about pciutils for big sur, but there is no guide of how to install it.
I tried make install but I get errors that I need directHW, even though I read in a previous post that is not needed.
I even tried copying files of directHW to some folders that pciutil install wants but then it errored on something else.
Is there any tutorial of how to install-setup pciutils on big sur? I want to enable pice 2.0 speed on slot 2 but I can't install it.
I have found I think 3 different instances on GitHub about pciutils for big sur, but there is no guide of how to install it.
I tried make install but I get errors that I need directHW, even though I read in a previous post that is not needed.
I even tried copying files of directHW to some folders that pciutil install wants but then it errored on something else.
- Yes, it does use a "soft" patch or more accurately, activate a feature in OpenCore that provides APFS support.
- Yes, by manually disabling the "EnableJumpstart" key in the OpenCore config file OCLP creates
- However, there is no need to disable it as it will not kick in once it determines APFS is already there.
You don't need directHW. Just add debug=0x144 to boot-args.
A bash script to produce more informative output than lspci -nntv
A bash script to produce more informative output than lspci -nntv - pcitree.sh
gist.github.com
I also have a MacPro3,1 that I obtained for small money, upgraded WIFI to BCM94360CD, GPU to RX480 and NVMe in slot 2. However pcitutils (here septic) only work with SIP disabled despite debug=0x144. I am trying to use a launchdaemon to set slot 2 to 5GT/s, but even sudo lspci fails with SIP enabled.You don't need directHW. Just add debug=0x144 to boot-args.
A bash script to produce more informative output than lspci -nntv
A bash script to produce more informative output than lspci -nntv - pcitree.sh
Since SIP is not mentioned in conjunction with pciutils I am suspecting some issues with access rights as the culprit on my 10.15.7 installation booting through OCLP 0.1.9. NVMe SSD works fine, but just at 2,5 GT/s x4.
I think you meant setpci instead of septic?
If you're using my pciutils, then you can use
sudo lspci -A detect to check what access methods you have:
Code:
Trying method darwin......using AppleACPIPlatformExpert...OK
Trying method intel-conf1...DirectHW.kext not loaded.
...no I/O permission...No.
Trying method intel-conf2......no I/O permission...No.If you don't include sudo (which runs the command as root), then you get this:
Code:
Trying method darwin......cannot open AppleACPIPlatformExpert (add boot arg debug=0x144 & run as root)...No.
Trying method intel-conf1......no I/O permission...No.
Trying method intel-conf2......no I/O permission...No.
lspci: Cannot find any working access method.I guess I'll write an EFI driver to do this so you don't need to use pciutils. If it doesn't survive sleep then a kext will be necessary.
Yeah sorry, auto spelling correction kicked in, I meant setpci.
I need to use sudo even with SIP disabled and I managed to change PCIe speed using the parameters shared by yourself and others in the other threads.
I even managed to create a launch daemon and the change in PCIe speed seems to survive sleep (see attached).
Not sure if the “keep alive” in the plist accomplished that or if it is just a coincidence.
However it only works when SIP is disabled, else pcilib complains that AppleACPIPlatformExpert cannot be opened and DirectHW.kext is not installed like in you second example.
Would your version help with additional insight about why that method is not available beyond “no I/O permission” as I would like to understand what blocks the permission: generic SIP limits or wrong permission setting kicking in with SIP activation (if that even exists).
I am going to check if allowing device configuration in SIP (csr-active-config=0x80) would do the trick but a kext or efi driver specific for modifying negotiated link speed would be a great thing to have.
Unfortunately I have no clue how to do it and I suspect it has a steep learning curve attached to it.
Hence your help would be highly appreciated.
Update: I tested various SIP settings and it is sufficient to just ALLOW_KERNEL_DEBUGGER, which in hindsight is quite obvious as we added debug=0x144 to the boot-args. Hence csr-active-config needs to be set to <08000000> in the OpenCore config and the launch daemon switches PCIe speed to 2.0 resulting in NVMe speed of around 1350 MB/s.
I need to use sudo even with SIP disabled and I managed to change PCIe speed using the parameters shared by yourself and others in the other threads.
I even managed to create a launch daemon and the change in PCIe speed seems to survive sleep (see attached).
Not sure if the “keep alive” in the plist accomplished that or if it is just a coincidence.
However it only works when SIP is disabled, else pcilib complains that AppleACPIPlatformExpert cannot be opened and DirectHW.kext is not installed like in you second example.
Would your version help with additional insight about why that method is not available beyond “no I/O permission” as I would like to understand what blocks the permission: generic SIP limits or wrong permission setting kicking in with SIP activation (if that even exists).
I am going to check if allowing device configuration in SIP (csr-active-config=0x80) would do the trick but a kext or efi driver specific for modifying negotiated link speed would be a great thing to have.
Unfortunately I have no clue how to do it and I suspect it has a steep learning curve attached to it.
Hence your help would be highly appreciated.
Update: I tested various SIP settings and it is sufficient to just ALLOW_KERNEL_DEBUGGER, which in hindsight is quite obvious as we added debug=0x144 to the boot-args. Hence csr-active-config needs to be set to <08000000> in the OpenCore config and the launch daemon switches PCIe speed to 2.0 resulting in NVMe speed of around 1350 MB/s.
Attachments
Last edited:
Now that I am in the final stage of fine-tuning I discovered another small issue:
The SuperDrive stops responding after sleep.
Background is the use of AppleIntelPIIXata.kext 2.5.1 from Mojave to re-enable access to the IDE devices in the MacPro3,1. All works great after cold boot or restart, however after system sleep the Superdrive disappears from the system info -> disc burning section and the eject button in the upper right of the keyboard stops working.
It still shows in the system info -> ATA section though.
One can also see in IORegistryExplorer that the IOSCSIPeripheralsDevice and lower objects attached to the IOSCSIPeripheralDeviceNub are missing. No specific entry regarding SuperDrive in the pm-log though.
I have attached system report and screenshots with further details.(some entries like serial number are replaced by zeros though)
This happens both on 10.15.7 as well as 11.4 using the same EFI generated with OCLP 0.1.9.
This is just a minor annoyance as I rarely use the Superdrive these days and the easy solution would be to attached a modern SATA drive to one on the extra ports on the main board and put it in the drive compartment. This will likely be less than €30.
However I am keen on understanding why this happens. Any ideas to further explore this symptom are highly welcome.
Update: Issue solved. I found a dormant ATAPI DVD-Drive in my storage and thought I revive that. Now with two drives connected to the IDE controller and jumped as master and slave rather than cable select, drives remain fully visible and functioning after sleep. Maybe there is still some issues with the Mojave driver that go all the way back to this post, but I don't have time to do more tests.
The SuperDrive stops responding after sleep.
Background is the use of AppleIntelPIIXata.kext 2.5.1 from Mojave to re-enable access to the IDE devices in the MacPro3,1. All works great after cold boot or restart, however after system sleep the Superdrive disappears from the system info -> disc burning section and the eject button in the upper right of the keyboard stops working.
It still shows in the system info -> ATA section though.
One can also see in IORegistryExplorer that the IOSCSIPeripheralsDevice and lower objects attached to the IOSCSIPeripheralDeviceNub are missing. No specific entry regarding SuperDrive in the pm-log though.
I have attached system report and screenshots with further details.(some entries like serial number are replaced by zeros though)
This happens both on 10.15.7 as well as 11.4 using the same EFI generated with OCLP 0.1.9.
This is just a minor annoyance as I rarely use the Superdrive these days and the easy solution would be to attached a modern SATA drive to one on the extra ports on the main board and put it in the drive compartment. This will likely be less than €30.
However I am keen on understanding why this happens. Any ideas to further explore this symptom are highly welcome.
Update: Issue solved. I found a dormant ATAPI DVD-Drive in my storage and thought I revive that. Now with two drives connected to the IDE controller and jumped as master and slave rather than cable select, drives remain fully visible and functioning after sleep. Maybe there is still some issues with the Mojave driver that go all the way back to this post, but I don't have time to do more tests.
Last edited:
I have been using OCLP 0.2.2 from dortania to run a cMP 3,1 (see signature) on both Big Sur and Monterey successfully. I just added an eeprom flashed Titan-Ridge Ver 2 TB3 AIC to the cMP3,1 and although it shows up correctly in System Information-Hardware-PCI and Thunderbolt/USB4 for both TB and USB 3.1 with loaded drivers, and I can also see it in IOReg PCI (see attached files) Thunderbolt and USB disks, as well as TB-to-Ethernet adapters, etc. don't mount/setup correctly. It seems I need to create a SSDT file and edited config.plist to add this functionality. I have such files from a cMP5,1 installation that are said to work for the Titan-Ridge but they clearly they need mods for the 3,1. I've tried several times using using Open Core Configurator to make the mods but just end with a frozen boot each time!
Is there someone here familiar with how to mod either the existing booting config.plist or the 5,1 version and creating the correct SSDT for installing in the EFI folder generated by OCLP?
I'll attach the booting config.plist now working on the 3,1 as well as a copy of the above-mentioned EFI folder for the cMP5,1 said to add support for the Titan-Ridge Ver 2 TB3 AIC, and the IOReg file generated from the 3,1 booting with non-functional Titan-Ridge card.
Thank you for any advice and help you can offer. I'm not asking for someone to fix it but rather to teach me how to fix it.
Is there someone here familiar with how to mod either the existing booting config.plist or the 5,1 version and creating the correct SSDT for installing in the EFI folder generated by OCLP?
I'll attach the booting config.plist now working on the 3,1 as well as a copy of the above-mentioned EFI folder for the cMP5,1 said to add support for the Titan-Ridge Ver 2 TB3 AIC, and the IOReg file generated from the 3,1 booting with non-functional Titan-Ridge card.
Thank you for any advice and help you can offer. I'm not asking for someone to fix it but rather to teach me how to fix it.
Last edited:
I don’t own any Thunderbolt hardware, but there are hackintosh builts that have working Thunderbolt.
This is in German, but the config files may help.
For full functionality you need a mac specific firmware flash.
www.hackintosh-forum.de
www.hackintosh-forum.de
www.hackintosh-forum.de
This is in German, but the config files may help.
For full functionality you need a mac specific firmware flash.
i9 10900k / Gigabyte Z490 Vision D - Er läuft - Hackintosh-Forum - Deine Anlaufstelle für Hackintosh & mehr...
Servus Leute,nachdem ich eine Weile mit zwei Hackintoshs Blut geleckt habe, habe ich mich jetzt entschieden bei meinem Desktop doch auf Intel umzusteigen (vorher AMD 1950x).Habe mir gestern das Gigabyte Z490 Vision D und einen i9 10900k gekauft.Nachdem…
www.hackintosh-forum.de
Gigabyte Z390 DESIGNARE: OpenCore-EFI-Ordner und Anleitung - Hackintosh-Forum - Deine Anlaufstelle für Hackintosh & mehr...
hackintosh-forum.de/attachment/139405/Für Hackintosh-Konfigurationen mit Gigabyte Z390 DESIGNARE mit Intel Core i9, i7, i5 oder i3 und Radeon-RX-GPU (RX 580, Vega 56, Vega 64, RX 5700 XT, Radeon VII etc.) oder auch für den Betrieb ohne externe GPU (nur…
www.hackintosh-forum.de
My last Hackintosh?! Gigabyte Z590i Vision D + 11900k + Radeon Pro W5500 - Hackintosh-Forum - Deine Anlaufstelle für Hackintosh & mehr...
So Leute,es ist mal wieder Zeit für einen neuen Build. Der Alte ist ganz schön in die Jahre gekommen. Spaß beiseite. Den 11900k empfinde ich von der Performance her keine wirkliche Verbesserung zum 10900k. Aber ich hätte gerne PCIE 4.0, so kommt aus…
www.hackintosh-forum.de
Thanks Inspector42. I used those German sites to do the original flash of the eeprom on the card. The card is seen correctly in the IOReg Device Tree and if I connect another Mac in target mode using Thunderbolt cable it sees the Mac (see screenshot) but not the SSD which does not mount on the cMP3,1 as an external disk. Similarly, if I connect a Thunderbolt-to-Ethernet adapter or a Thunderbolt to Firewire adapter, it shows up in System Reports-Hardware under Thunderbolt/USB4 but the adapter cannot be set up in Sys Pref. I am fairly sure the issue is to do with PCI ports and requires an SSDT file in ACPI and a modification to the open core config.plist in the OC folder in the EFI folder created by OCLP. Last screenshot shows Mac connected in Target mode but Mac SSD does not show in list or on cMP desktop.
The problem is I don't know how to make the modification.
The problem is I don't know how to make the modification.
Attachments
Last edited:
It has been a while that I looked into DSDT/SSDT and I am far from being savvy in the iASL language that they are built of.
It is my understanding that most SSDT modifications started by comparing the ACPI file from the Mac with those of a recent UEFI PC.
There is a repository for the ACPI files extracted from basically every Mac Model but I can’t find the link anymore. If you could get the ACPI files from one of the systems I linked before it might give you an idea what is missing. As the devices show up in the registry and even have drivers attached to it, it likely goes beyond just adding the device or enabling it through a SSDT.
Let’s hope someone with more ACPI knowledge can chime in here.
Edit: I just re-read the original post and it seems you are already beyond my suggestion.
It is my understanding that most SSDT modifications started by comparing the ACPI file from the Mac with those of a recent UEFI PC.
There is a repository for the ACPI files extracted from basically every Mac Model but I can’t find the link anymore. If you could get the ACPI files from one of the systems I linked before it might give you an idea what is missing. As the devices show up in the registry and even have drivers attached to it, it likely goes beyond just adding the device or enabling it through a SSDT.
Let’s hope someone with more ACPI knowledge can chime in here.
Edit: I just re-read the original post and it seems you are already beyond my suggestion.
Last edited:
Could someone here help me out with a request? If you have a cMP 5,1 with the eeprom-flashed Titan-Ridge TB3 AIC installed and working, could you capture the IOReg output only for Thunderbolt and save and post it here. I need it for comparison with the cMP IOReg output to Thunderbolt to try to modify the cMP5,1 SSDT file to work with the cMP3,1.
Udpating MaciASL from Ver 1.4 to 1.6.1 solved this problem, but still no cigar; I tried recompiling the SSDT for the cMP5,1 using the IOReg Thunderbolt output for the cMP3,1 — there were no compile errors and the cMP still boots but no change in the TB hardware status. All devices (Target-mode MBA, TB to Ethernet and TB to FW adapters) connected to the TB3 ports are seen in System Information – Hardware, but are not functional.
Driving me nuts; has nobody got a working OC folder for the EFI folder for this card in the cMP3,1?
Last edited:
Success—in addition to recompiling the SSDT-TBOLT3.aml from D2M2 for my cMP3.1 IOReg TB path, I also added some lines to the config.plist generated from OCLP for the cMP3,1 and it came good.
If people here want to do this for the cMP3,1 I'll post a blow-by-blow description of how I did it.
If people here want to do this for the cMP3,1 I'll post a blow-by-blow description of how I did it.
Attachments
Congratulation ! I have considered Thunderbolt for a short while but could not see the value it adds. Considering MacPro3,1 is limited to PCIe 2.0 (x16 slots) and PCIe 1.0 (x4 slots) what is the actual throughput you achieve ? My NVMe SSD in slot 2 has appr. 1500 MB/s, which would be around 10 GT/s.
I also need to force PCIe 2.0 mode through a Lauch daemon. It the flashed Titan Ridge natively negotiating 2.0 speed ?
I also need to force PCIe 2.0 mode through a Lauch daemon. It the flashed Titan Ridge natively negotiating 2.0 speed ?
It is showing 20 Gb/s on the connected TB port and 40 Gb/s on the open port.
I haven't clocked the true speed yet.
The boot NVMe Samsung SSD in the HighPoint 7101 uses the AIC's own PCIe Gen 3.0 chip so it gets much better speeds.
But this Titan-Ridge V2 TB card was a gift from a friend breaking an old Windows PC so no cost and a bit of fun/time hacking and coding to get it running.
Last edited:
I'm working on an EFI driver to fix the PCIe 2.0 link rate issue. It's gotten a little out of hand - I added the pci ids database from pciutils to recreate the output of pcitree.sh in EFI Shell. I'm currently working on adding code to search for hidden PCIe devices of Thunderbolt add-in cards. I would like the code to at least enable the USB of my GC-ALPINE RIDGE (to replace the startup.nsh method).
i'd be curious about your config settings... i recently tried to install Catalina to a cMP 3,1 (E5462 x2, 64GB RAM, PCIe SSD, unflashed Nvidia GTX 780 Ti, OS X El Capitan) using OpenCore (and the directions provided in the "OpenCore on the Mac Pro" thread) and couldn't get the Catalina install to proceed (after the first reboot, the screen stayed black). tried again using OCLP (and the directions provided at dortania.github.io) and had the same result: trying to install High Sierra first resulted in an error (something about a missing firmware partition), then trying to install Catalina resulted in another error (something about the machine not being able to run Catalina). finally had to use Dosdude1's Catalina patcher to get it up and running, but it is not as elegant a solution as using OpenCore.
i can't help but think it has something to do with the Nvidia GPU, since i have two other cMPs (one 4,1 > 5,1, one 5,1) with AMD Radeon RX 580s that i had no problems using OpenCore with...
has anyone else here had these problems?
Not specifically and the massage about a missing firmware partition might point at other issues.
You can try using MyBootMgr (link in my sig) to set things up and let me know how it goes. Should work with unflashed Nvidia GTX 780 Ti just fine.
Will suggest you do try dropping to 32 GB for the installation as was suggested. You then need to properly implement the 64 GB mode when done (sudo nvram boot-args="maxmem=63488" in Terminal). This will limit the actual used RAM to 62 GB which is apparently the real maximum amount that can be used without a speed hit.
You can add this boot arg to OpenCore and/or RefindPlus so that it is always available (after NVRAM reset etc)