Ryzen 7000 Testing

[Vorshim92]

This is a good plan. Getting some experience with SSDTs is an added bonus.

Can I ask if you ever had problems with IOREG like freezing or slowing down?
for me on Ventura I have to wait a lot of minute before I can see it and it's always slow, to take a save I spend like 5 minutes xD and my ioreg is very large, like 150MB

 

[CaseySJ]

Can I ask if you ever had problems with IOREG like freezing or slowing down?
for me on Ventura I have to wait a lot of minute before I can see it and it's always slow, to take a save I spend like 5 minutes xD and my ioreg is very large, like 150MB

I have not encountered this problem. My IOReg files are generally no larger than 35MB.

 

[CaseySJ]

tb4 WIP for hotplug and wake

for asus x670e :
for hotplug I think the method GPE - EE1B
for wake to method SB - IWAK

I ran some Thunderbolt hot-plug experiments yesterday, but all of them failed. This does not mean we should give up...if at first you don't succeed, try try again. Here are some notes from those experiments:

• I started by injecting some device properties that we believe are necessary for hot plug to work:
  • Root Port TBTC needs property PCIThunderbolt
  • Downstream Bridge at Address Zero TBDP needs property PCIHotplugCapable
  • Native Host Interface TBDC needs property linkDetails for enabling full speed (40 Gbps)
• I haven't looked at GPE hot plug events yet

With above properties I saw a glimmer of hope, but only a glimmer:

• I booted the system with my Belkin Thunderbolt 3 Dock connected
• Dock appears in IOReg and works fine
• Hot-unplugged the dock -> all IOReg devices turn red as expected
• Hot-plugged the dock -> IOThunderboltPort under TBDC flashed green -> this means a GPE event was triggered when dock was plugged in, and the event was seen by macOS
• But the dock did not reconnect

There are some unfortunate side effects from adding the above 3 properties:

• On reboot, the monitor cannot re-sync to the video signal; I just have to power off and power on the monitor
• Worse, however, is that BIOS seems to get reset on reboot
• Sleep will work when Thunderbolt device is connected, but wake will not work

 

[pbs888]

Have you checked in your debug log if it is fine?
try to disable not mandatory kext to boot and check the original post for my updated EFI
Also check your USB port mapping
with latest patch discovered by Casey all OS works perfectly and rock solid

if you have problems it could be also related to DDR5 and I would suggest to use AMD specific for them

hey @mariettosun!

// MMIO
I actually have similar concerns with MMIO — I've used the EFI provided by @CaseySJ from here as my starting point (I have Asus X670E Extreme) and modified last entry as I was getting different address on my motherboard. I also left only the second entry enabled and I was able to boot.

But, how do I know those MMIO entries are correct? I can see the MMIO in the logs, but AFAIK this is just telling me what's enabled/disabled. Is there anything else I can look at to verify MMIO is correctly set?

OCABC: MMIO devirt start
OCABC: MMIO devirt 0xF0000000 (0x8000 pages, 0x800000000000100D) skip 0
OCABC: MMIO devirt 0xFD000000 (0x1E00 pages, 0x800000000000100D) skip 1
OCABC: MMIO devirt 0xFEE00000 (0x1 pages, 0x8000000000000001) skip 0
OCABC: MMIO devirt 0xFEE01000 (0x11FF pages, 0x800000000000100D) skip 0
OCABC: MMIO devirt 0x1080000000 (0x20200 pages, 0x800000000000100D) skip 0
OCABC: MMIO devirt end, saved 675840 KB
OCABC: Only 168/256 slide values are usable!
OCABC: Valid slides - 88-255

Fun fact: with second ssd installed in the GEN-Z.2 slot (the new fancy easy access slot next to RAM slots) I was not able to boot at all. I had to remove it.

I am attaching my current EFI with logs just in case.

Any tips much appreciated!

 

[fabiosun]

Fun fact: with second ssd installed in the GEN-Z.2 slot (the new fancy easy access slot next to RAM slots) I was not able to boot at all. I had to remove it.

here is (probably) a MMIO problem)
try to see the same the log you have

your MMIO above are good
if you see on the second one 0 system does not boot
Every time you change your hardware configuration (about PCIE slot or NVME disks) you have to check if MMIO are good (the schema 0,1,0,0,0 is safer on this platform)

 

[CaseySJ]

...
Fun fact: with second ssd installed in the GEN-Z.2 slot (the new fancy easy access slot next to RAM slots) I was not able to boot at all. I had to remove it.

...

My Asus X670E Gene also has a GEN-Z.2 slot. In my case both sides of the GEN-Z are populated, and the system boots and works properly.

 

[pbs888]

here is (probably a MMIO problem)
try to see the same the log you have

your MMIO above are good
if you see on the second one 0 system does not boot
Every time you change your hardware configuration (about PCIE slot or NVME disks) you have to check if MMIO are good (the schema 0,1,0,0,0 is safer on this platform)

I've noticed that the MMIO in the logs always looks like I've specified it in my config.plist. So this MMIO in the logs is just telling us what are the addresses — not what the correct setting should be — is that correct?

My question is how do I know this is correct or does it just work by accident? In the open core docs it says to randomly disabling/enabling them and checking if system boots up? With 5 record we have quite a lot of possibilities to check

Can I for example add just a single MMIO and try to boot with just a single MMIO set? Once I know what the value for the first one should be check I can add the second to check for enabled/disabled..and so on and on until I have all 5.

I am still experiencing some instability issues and I am looking for potencial EFI errors.

 

[pbs888]

My Asus X670E Gene also has a GEN-Z.2 slot. In my case both sides of the GEN-Z are populated, and the system boots and works properly.

That's very interesting. I was unable to find the correct MMIO values to boot with the GEN-Z.2 slot installed.

How did you come up with your MMIO list? Is there any log-driven method for it or do I have to check every single combination available? 5 x MMIO entires with 2 possible value for each give us quite a lot of possible combinations

 

[fabiosun]

I've noticed that the MMIO in the logs always looks like I've specified it in my config.plist. So this MMIO in the logs is just telling us what are the addresses — not what the correct setting should be — is that correct?

the correct settings are due to the experience you get on the platform in this case AM5
The first people who had these systems did the testing for us and found the correct pattern to start
on AM5 as I wrote to you having SKIP 1 on the second MMIO area is currently the best solution
You can also activate the first, the second, the third and the fourth and your system would still start
Since I have this platform I have experienced that the best combination is the one you have in your debug (numbers can change from manufacturer to manufacturer and from bios to bios as well as from installed hardware)
But 0,1,0,0,0 schema for MMIO works for everyone correctly

My question is how do I know this is correct or does it just work by accident?

With the experience of the AMD community

With 5 record we have quite a lot of possibilities to check

my previous platform have add 19 MMIO areas

Can I for example add just a single MMIO and try to boot with just a single MMIO set?

it is what you did in your log
if they are 0 you can also exclude them from your config and they will be skipped to 0 by the system anyway

 

[fabiosun]

@pbs888 maybe you are using in a wrong way all patches sets did by Casey for Aquantia ethernet
And also you do not use new patch useful in Ventura OS if you have wifi and TB enabled in your bios

 

[pbs888]

@pbs888 maybe you are using in a wrong way all patches sets did by Casey for Aquantia ethernet
And also you do not use new patch useful in Ventura OS if you have wifi and TB enabled in your bios

I am actully running Monterey — I've tried Ventura but I preffer more stable Monterey (if it is actually more stable)

My BIOS settings are:

• onboard USB controller enabled
• onboard WIFI/BT disabled
• onboard Intel i255-V 2.5Gbe enabled
• onboard Aquantia AWC113CS 10Gbe enabled
• onboard USB4/Thunderbolt enabled

I've added all patches (except the Thunderbolt one as I have not tried it yet) and everything works as expected.

My first though was that there must be an issue with MMIO entires, but if those are fined maybe I've messup something in the USB configuration?

I've done my USB Mapping with Hackintool with the onboard USB4/Thunderbolt enabled and I made sure to change Fenvi to 255 and I've delited the ports I was not planning on using. I kept all USB C but only kept 3 of the USB2/3.

Please see image below. Do I have too many USB controllers?






Just looking for potencial issues..

 

[fabiosun]

@pbs888 have you disabled your USB audio? or extreme doesn't have Realtek USB audio?

 

[agent2]

Hey guys, sorry if this is a bit unrelated to the current discussion, but about a week ago there was a brief discussion on this thread about USB DVD/CD-players not working. I stumbled across this article and thought that it might clear up a little of the confusion that arose here.
Cheers!

macOS 13.2 Breaks Compatibility With Pioneer's CD/DVD Drives [Updated]

The latest version of macOS Ventura breaks compatibility with Pioneer's lineup of USB-connected CD/DVD/Blu-ray drives for the Mac. Apple released macOS 13.2 in late January with support for a new Security Keys for Apple ID feature and various bug fixes.

www.macrumors.com

 

[CaseySJ]

Hey guys, sorry if this is a bit unrelated to the current discussion, but about a week ago there was a brief discussion on this thread about USB DVD/CD-players not working. I stumbled across this article and thought that it might clear up a little of the confusion that arose here.
Cheers!

macOS 13.2 Breaks Compatibility With Pioneer's CD/DVD Drives [Updated]

The latest version of macOS Ventura breaks compatibility with Pioneer's lineup of USB-connected CD/DVD/Blu-ray drives for the Mac. Apple released macOS 13.2 in late January with support for a new Security Keys for Apple ID feature and various bug fixes.

www.macrumors.com

I saw this a few moments ago. It even affects Pioneer USB CD/DVD drives. There’s also a problem with SMB file sharing.

Both of these problems are in Ventura. I haven’t personally encountered them, but I would expect Apple to fix them in a 13.2.1 or similar minor update.

 

[pbs888]

@pbs888 have you disabled your USB audio? or extreme doesn't have Realtek USB audio?
View attachment 9684

It is Realtek ALC4082. I was unable to map it - it just would not show up at all in the Hackintool USB section. Did you have to do anything specific for the USB to show up?

 

[pbs888]

@pbs888 have you disabled your USB audio? or extreme doesn't have Realtek USB audio?
View attachment 9684

I can also see you have AURA LED here - does it mean you can control RGB? I have a Fractal Torrent Compact with RGB but I never though RGB will work via MacOS How did you get it to work?

 

[CaseySJ]

I can also see you have AURA LED here - does it mean you can control RGB? I have a Fractal Torrent Compact with RGB but I never though RGB will work via MacOS How did you get it to work?

I use Jonas Malaco's liquidctl to control RGB via macOS on my X670E Gene:

GitHub - liquidctl/liquidctl: Cross-platform CLI and Python drivers for AIO liquid coolers and other devices

Cross-platform CLI and Python drivers for AIO liquid coolers and other devices - liquidctl/liquidctl

github.com

Asus Z690 ProArt Creator WiFi (Thunderbolt 4) + i7-12700K + AMD RX 6800 XT

** Automating Liquidctl at Start Up ** Purpose: To install a script that will run automatically at startup to configure Aura LED color channels. Procedure: There are two parts to the procedure: Creating a script to configure color channels Creating and registering a LaunchAgent Part 1...

www.tonymacx86.com

An earlier post in this thread:

Ryzen 7000 Testing

Aura LED controller via liquidctl: % liquidctl list Device #0: AsusTek Aura LED Controller (experimental) % liquidctl status AsusTek Aura LED Controller (experimental) ├── ARGB channels: 2 └── RGB channels: 1 # set all channels to green % liquidctl -m Aura set sync color static...

forum.amd-osx.com

 

[Galve2000]

Are you using the RX 6600 XT device-id for your RX 6650 XT, with this being injected via the DeviceProperties section of your config.plist. As you aren't using an SSDT for the dGPU.

Your DeviceProperties for the RX 6650 XT should look something like this.

View attachment 9664

The Device Path would be different as this screenshot is taken from an Intel system's config.plist.

So I am missing some of the 0,name 1,name etc entries.. but I do have the PCi root path Directory and device ID Data in my config.plist:

Do you have -radcodec boot arg? You need it to get hardware decoding/
encoding on spoofed GPUs.

I added -radcodec to my boot args, it did not make a difference, unfortunately


here's my GFX util output, if you could help me make an SSDT, I would be most grateful.



its not the end of the world that GFX Bench Metal does't work

Heaven, Geekbench 5, Blender and SotTR benchmark all work and I know Hardware Acceleration (Metal) is working. but if we could fix this, it would be great.

if you need my GFX0 IORegExplore screenshot, I would be happy to provide it.

 

[Edhawk]

Just add the other five DeviceProperties, as shown in the screenshot below. You might not need an SSDT if the framebuffer patches work with your AMD system.



I don't think you need the -radcodec boot argument either, but that is another matter.

The device path for your discrete GPU is (_SB.PCI0.GPP0.SWUS.SWDS.GFX0) according to the gfxutil screenshot.

 

[ExtremeXT]

Just add the other five DeviceProperties, as shown in the screenshot below. You might not need an SSDT if the framebuffer patches work with your AMD system.

View attachment 9706

I don't think you need the -radcodec boot argument either, but that is another matter.

The device path for your discrete GPU is (_SB.PCI0.GPP0.SWUS.SWDS.GFX0) according to the gfxutil screenshot.

Why the device_type property? That doesn't do anything, and you also need only @0 name, and the other 3 will take it from the first one. Your device-id is wrong, it's 32 bit little endian, so the correct value for spoofing to 6600XT is FF730000.

-radcodec is needed on any spoofed GPU (along with WEG) to fix hardware decoding/encoding support and DRM.

WhateverGreen automatically renames the GPU to GFX0, the actual name is usually VGA_ on Navi.

So I am missing some of the 0,name 1,name etc entries.. but I do have the PCi root path Directory and device ID Data in my config.plist:

View attachment 9703


I added -radcodec to my boot args, it did not make a difference, unfortunately

View attachment 9704
here's my GFX util output, if you could help me make an SSDT, I would be most grateful.

View attachment 9705

its not the end of the world that GFX Bench Metal does't work

Heaven, Geekbench 5, Blender and SotTR benchmark all work and I know Hardware Acceleration (Metal) is working. but if we could fix this, it would be great.

if you need my GFX0 IORegExplore screenshot, I would be happy to provide it.

You only need an SSDT if you need to define PCI Bridges, in your case devprops are enough. You should have:
device-id | Data | FF730000
model | String | AMD Radeon RX 6650 XT
@0,name | String | ATY,Henbury