trx40 [Discussion] - TRX40 Bare Metal - Vanilla Patches

[fabiosun]

a bit OFF Topic (I will ban my self later)

we are happy to say that also with clover boot loader is possible to install from scratch and also update from a previous OSX version

I do not advise Clover boot loader because many problem could happen during installation steps and not solved by now:

1)GetOS Detection

2)MatchOS

but with patches you can find in thread below I have had success!

 

[Jaidy]

hello you wonderful people,

I am facing an interesting problem. My EFI, while boots just fine, does not show me the macOS installer, (the external one) to boot from to do an installation from scratch at the OC's picker menu. Can some one please take a look at my EFI and fix the config.plist? I would be eternally grateful.

 

Here is the link to the EFI.

 

[iGPU]

4 hours ago, Jaidy said:

hello you wonderful people,

I am facing an interesting problem. My EFI, while boots just fine, does not show me the macOS installer, (the external one) to boot from to do an installation from scratch at the OC's picker menu. Can some one please take a look at my EFI and fix the config.plist? I would be eternally grateful.

 

Here is the link to the EFI.

 

 

I answered this already, here. Change your scan policy, yours is 3481 (try to carefully read the thread). And "SecureBootModel" is best left "Disabled".

 

Your should also look at the Quirk settings I posted on same page here.

 

AppleMCEReporterDisabler.kext is not necessary. I still use AppleALC.kext and WEG (but limit with "agdpmod=pikera"on the 6900XT). I've not seen anyone use "RtWlanU.kext" or "RtWlanU1827.kext" on the TRX40 builds, but perhaps you have add-on hardware of which I'm unaware.

[Jaidy]

2 hours ago, iGPU said:

 

I answered this already, here. Change your scan policy, yours is 3481 (try to carefully read the thread). And "SecureBootModel" is best left "Disabled".

 

Your should also look at the Quirk settings I posted on same page here.

 

AppleMCEReporterDisabler.kext is not necessary. I still use AppleALC.kext and WEG (but limit with "agdpmod=pikera"on the 6900XT). I've not seen anyone use "RtWlanU.kext" or "RtWlanU1827.kext" on the TRX40 builds, but perhaps you have add-on hardware of which I'm unaware.

Thank you very much. I have a usb wifi that gives me faster speeds.

[Shaneee]

On 6/26/2021 at 5:39 PM, fabiosun said:

@Ploddles

 

We are both using this patch, but it is not an official one I think

do you remember where it was taken from or have you taken it in the past from my config.plist?

I can't find it and strangely it works the same way as the official one ...

@iGPU and @Shaneee can you take a look when you have time too?

 

thank you very much

 

It's XLNC's patch,

 

XLNC - Remove version check and panic - 10.15/11.0/12.0

 

[fabiosun]

23 minutes ago, Shaneee said:

 

It's XLNC's patch,

 

XLNC - Remove version check and panic - 10.15/11.0/12.0

 

thank you to clarify this

my doubt was about the two patches below (system works well we both or one of them)

 

different pattern same job?

 

[fabiosun]

@all

I know, you do not love clover bootloader...but to try patches job..is a fast way 🙂

 

 

 

[valmeida]

@ fabiosun I installed your patches and I can boot BS beta 4  but not Monterey . Here is my config.PLIST and the error message I get when trying to install Monterey .  

config.plist.zip

[fabiosun]

@valmeida

maybe you have to update opencore and all your kext 

then if you use bt or Wi-Fi kext maybe you have to do some work on those

 

[iGPU]

2 hours ago, valmeida said:

@ fabiosun I installed your patches and I can boot BS beta 4  but not Monterey . Here is my config.PLIST and the error message I get when trying to install Monterey .  

config.plist.zip 7.19 kB · 0 downloads

 

Some of the BT stuff I commented upon here.

 

I211 is not working in Monterey (the SmallTreeIntel82576.kext is broken) and if used will crash the system. This is true for both Intel and AMD. To avoid crash, either disable the kext or don't plug anything into the I211 ports. Aquantia ports are working just fine, but do need a different kernel patch that I posted earlier.

 

And no patches are yet working for Monterey beta 2 on AMD.

 

 

Edited July 1, 2021 by iGPU

[fabiosun]

@valmeidaAre you sure then you are installing Beta 1...

[Shaneee]

8 hours ago, fabiosun said:

thank you to clarify this

my doubt was about the two patches below (system works well we both or one of them)

 

different pattern same job?

 

You'd need to ask XLNC about that one.

 

[fabiosun]

Thanks @iGPUfor a tip given..

It is possible to use his sets of minimal patches to boot 11.4 and 11.5 b4, in my case I have had problems due AMDRyzenCPUPowerManagement.kext which caused without those two patches a reboot in early stage

now patches are 11 to boot in 11.4 and 11.5 b4 with trx40 system

[iGPU]

I'm working on Monterey ß1. While the patch list is not yet reduced, I did, despite reading complaints about BT not working, have Airdrop working just fine as shown in Spoiler below. (See this previous post about how to get Airdrop to work.) All of this is under OC v071.

 

[So far, patches list for Big Sur + Monterey (ß1) is reduced to 24 22 17 + 2 for Aquantia, from the initial 46+2. See next post for file.]

 

When booting, Monterey is slow with the Apple progress stalling about 1/2 way through, requiring some patience.

 

AirDrop is working:

 

Spoiler

 

 

First, below are kexts used (note that some are limited to Big Sur by MaxKernel 20.99.99). The AirportBrcmFixup.kext does not need to be loaded for the 4360 as this work OOTB under Monterey, unlike in Big Sur, which requires this kext to be injected for Airdrop to work.

 

 

While the driver for I211 is loaded:

 

 

The port is not active (an ethernet cable is plugged in and works under Big Sur) as shown in Spoiler below (along with the active Aquantia port).

 

Spoiler

 

 

 

The Aquantia port is working well, but uses a new Patch just for Monterey (this was uploaded with previous BS-minimum patch plist file):

 

 

While the SmallTree kext is apparently loading, it must not be attaching to the correct location. Attached is the log for all kext files loaded.

 

 

loaded kext log.zip

Edited July 8, 2021 by iGPU

[fabiosun]

grazie @iGPU i can confirm all system booting fine from 11.4 release to Monterey B1 (tested also 11.5b4) with above limited set of patches

I am investigating some stuff about temperatures

[iGPU]

On 7/4/2021 at 3:13 AM, fabiosun said:

grazie @iGPU i can confirm all system booting fine from 11.4 release to Monterey B1 (tested also 11.5b4) with above limited set of patches

I am investigating some stuff about temperatures

 

@fabiosun

 

I've updated my above post to show that AMD-SMC kext files run, at least in my setup, with the posted minimal patches in Monterey ß1. Attached is the debug version of a recent OC v071 that I've been using for these tests. No config, ACPI, Kext (in previous post), Utility or Resources folders are included to keep size small.

 

As for temps, if the computer is not touched, the baseline temp is ~46 C. Moving windows about, etc, cause the temp to spike to ~63 C and then rapidly fall.

 

 

Edited July 10, 2021 by iGPU

[fabiosun]

@iGPUi have 10 ° in all 3 system I can boot, my results in temperature are similar to yours

With previous EFI posted in this thread  I have had 10° less

I am investigating this  (not happy for this I mean)

 

I didn't do an accurate testing (sorry) because I am trying to have patches working also for Clover Bootloader and a X570 system 

confirmed your update about AMD PowerGadget

 

On 7/2/2021 at 11:33 PM, fabiosun said:

Thanks @iGPUfor a tip given..

It is possible to use his sets of minimal patches to boot 11.4 and 11.5 b4, in my case I have had problems due AMDRyzenCPUPowerManagement.kext which caused without those two patches a reboot in early stage

now patches are 11 to boot in 11.4 and 11.5 b4 with trx40 system

 

to explain better this assertion 🙂

I usually do not use RebuildAppleMemoryMap quirk

and without two patches I have a pretty instant reboot

Bios and hardware settings are not involved in my opinion

[iGPU]

21 minutes ago, fabiosun said:

@iGPUi have 10 ° in all 3 system I can boot, my results in temperature are similar to yours

With previous EFI posted in this thread  I have had 10° less

I am investigating this  (not happy for this I mean)

 

I didn't do an accurate testing (sorry) because I am trying to have patches working also for Clover Bootloader and a X570 system 

confirmed your update about AMD PowerGadget

 

to explain better this assertion 🙂

I usually do not use RebuildAppleMemoryMap quirk

and without two patches I have a pretty instant reboot

Bios and hardware settings are not involved in my opinion

 

 

@fabiosun

 

I just finished test booting your EFI v070 from 1st post on this thread. I only substituted my patches (your have 15; mine 18) and my mono-specific MmioWhitelist; otherwise all is from your EFI folder (no updated kexts, etc). When I did this, I could not boot into either Big Sur or Monterey.

 

I think this is due to other settings within OC and has nothing to do with v070 vs v071. My OC settings also include various ACPI and DevProp injections, which, while I understand many here don't like them, I think they help the functionality of Hackintoshes.

 

I'll work on a sampled EFI and attach to this post in a few minutes...

 

[fabiosun]

@iGPUEFI you have downloade is only till Big Sur beta 4 so you have to leave "my" patches and try with your BS 11.5 beta 4

45 minutes ago, iGPU said:

 

 

@fabiosun

 

I just finished test booting your EFI v070 from 1st post on this thread. I only substituted my patches (your have 15; mine 18) and my mono-specific MmioWhitelist; otherwise all is from your EFI folder (no updated kexts, etc). When I did this, I could not boot into either Big Sur or Monterey.

 

I think this is due to other settings within OC and has nothing to do with v070 vs v071. My OC settings also include various ACPI and DevProp injections, which, while I understand many here don't like them, I think they help the functionality of Hackintoshes.

 

I'll work on a sampled EFI and attach to this post in a few minutes...

 

.

I think temperatures are too elevate and for now I do not know why

I can use your EFi without touching nothing

obviously I miss a couple of MMIO if I do in that way but system starts and works..

For now I can't have a system as I like if I include new Monterey b1 patches...

if you use your patches in my 070 EFi couldn't work because you disable two patches as we said and you discovered  yesterday and without them I need of RebuilAppleMemoryMap quirk ..to see if your system boots you should only change your MMIO and maybe something with WEG I do not use as kext and either as bootlargs..but it is only a my curiosity to see also for you some degrees less happen

 

to summarise with a 13 patches combination my system needs RebuilAppleMemoryMap that usually I do not use)

if you put your complete set of patches in my 070 EFi you have to enable that quirk..and to gain some degrees also DisableIOMapper..

but my goal is always to reduce stuff 🙂

 

Edited July 4, 2021 by fabiosun
formatting

[iGPU]

On 7/4/2021 at 10:42 AM, fabiosun said:

@iGPUEFI you have downloade is only till Big Sur beta 4 so you have to leave "my" patches and try with your BS 11.5 beta 4

.

I think temperatures are too elevate and for now I do not know why

I can use your EFi without touching nothing

obviously I miss a couple of MMIO if I do in that way but system starts and works..

For now I can't have a system as I like if I include new Monterey b1 patches...

if you use your patches in my 070 EFi couldn't work because you disable two patches as we said and you discovered  yesterday and without them I need of RebuilAppleMemoryMap quirk ..to see if your system boots you should only change your MMIO and maybe something with WEG I do not use as kext and either as bootlargs..but it is only a my curiosity to see also for you some degrees less happen

 

to summarise with a 13 patches combination my system needs RebuilAppleMemoryMap that usually I do not use)

if you put your complete set of patches in my 070 EFi you have to enable that quirk..and to gain some degrees also DisableIOMapper..

but my goal is always to reduce stuff 🙂

 

 

I'm not understanding why not using RebuilAppleMemoryMap is reducing anything; it just seems like we use one Quirk or another.

 

The OC docs discuss MAT:

Spoiler

 

 

In the error log created when booting OC Debug, search for "MAT" and you'll find that MAT = 1, meaning our mobos support MAT:

Spoiler

 

 

Therefore, it would seem that Booter/Quirks, based on MAT, can be set as:

Spoiler

 

 

*******

 

UPDATE:

 

With multiple, very slow boots into Monterey ß1, I remembered on InsanelyMac some users on the Intel side commenting that turning on 'auto login' reduces boot time. It does. (This is most likely just early beta hiccups.)

 

Enable it by going into Users & Groups as shown below:

 

 

Edited July 5, 2021 by iGPU

[fabiosun]

A spike of 63 degrees when system is doing nothing for me it is no good

the sense of the test I was asking is to see if also for other users with that Efi is possible to have a temperatures reduction as I see of 10 degrees and more.

then we can also  discute about mat or what it is correct for an Apple system with an unsupported thread ripper Cpu
 

[iGPU]

Attached to this post is an EFI that seems to boot into Big Sur and Monterey ß1 on the TRX40 without too much difficulty, using the minimal patch list from recent post. Comments are made on most problematic sections of OC in several Spoilers below. This is the Debug version of OC v071 (be2d9fe) from 5 July.

 

By "too much difficulty", I mean that occasionally, the AMD-SMC kext prevents a boot. As I've mentioned, unless absolutely necessary, I keep it disabled. The developer has not specifically written it for the TRX40 platform, nor updated it recently.

 

 

1. ACPI (Internal HS05 power for USB/BT; also see Update at bottom)

 

Spoiler

The ACPI section contains several files. The top red highlighted section is good for all TRX40 mobos. 

 

The pink section will probably enable either item 6 or 7 but not both. Each of these files also adjusts the USB ports

and will probably require some editing to fine-tune for your mobo (seems to be manufacturer-dependent, so what's 

here is okay for MSI but will need some changes for Gigabyte or other brands. You can try booting with both disabled

if you're uncertain. Basically, item 6 does not define the USB port that supplies power to the internal AX200 module,

which is "XHC/HS05" (see 2nd image below from the HackCheck app), while item 7 does supply power to this port. Except for this difference, the two ACPI files (6/7) are identical. So, if you need to change for your mobo, look for which port powers the AX200 module when studying IORE. (Also, see update added to end of this post on 22 July.)

 

The green section was not necessary for Big Sur, but Monterey ß1 would not boot unless I had the TB SSDT enabled

(I have a TB AIC present). Disable/enable as you see fit.

 

 

 

If HS05 is not defined, this removes the USB power from the internal AX200 BT device:

 

 

The Delete section removes a common error, I'd leave enabled.

 

 

 

 

 

2. Booter - DevProperties Sections

 

Spoiler

This section was somewhat discussed in another post (above).

 

Edit the MmioWhitelist section for your mobo since there is variation between manufacturers. Also, as shown in example,

even BIOS settings can influence the list. See this post I wrote on how to create a Mmiowhitelist.

 

The DevProp injection turned out to be important for booting into Monterey ß1 (all could be off when booting into Big Sur; odd).

I've left 2 sets of GPUs (both inactivated with "#" symbols) as examples. One is for the RX580, the other, for the 6900XT.

 

Other entries are for NVMe drives, I211 and Aquantia ports, audio, and Wifi/BT entries. Using Hackintool, or the new HackCheck available

in the download section on this forum, to definitively verify the PCI addresses for your mobo.

 

 

 

 

3. Kernel

 

Spoiler

I would recommend leaving all enabled, even WEG as there is a boot-arg present (discussed below) for NAVI GPUs.

 

I211 kext is disabled for Monterey since it is presently not working (the drivers are loaded, but the port is inactive). Not fixed as of mid-July 2021.

 

DummyPowerManagement is left disabled. The initial thought was this Quirk should not be necessary for TRX40 platform. , However, some users seem to need it, so this behavior will need to be investigated by each user. (It seems to be a requirement for Gigabyte TRX40 mobos.)

 

 

 

 

4. MIsc and NVRAM

 

Spoiler

The Misc/Boot section is fairly unremarkable but is set up to limit lots of boot entries (scan policy) in the menu system

as well as set up the OC graphic menu system. Unless you have a really good reason to change, I'd leave all these values

as is.

 

The Debug section (not shown), can have the DisplayDelay set to 0, SysReport to NO, and the Target to 3 to stop debug 

reporting.

 

The NVRAM section has inactivated (through the "#" symbol) many boot-arg entries. It is there to save

time if you need to recall some. Most are unnecessary for typical boots. If your boots are stable remove the 

"-v" to stop verbose boots.

 

 

 

5. PlatformInfo

 

Spoiler

The descriptions below are probably sufficient. Do adjust/edit the Generic and Memory sections as to your mobo

configuration.

 

I'd leave the drivers as is.

 

 

 

 

 

 

 

6. Validation a config plist file:

 

Spoiler

This is a complicated image. The upper left is a Terminal window; the right, the EFI/OC boot folder.

 

The enclosed description states how to run 'ocvalidate' to see if your config.plist file is correct

for the current version (commit, actually). Each new commit updates ocvalidate and is only

accurate for that build. ocvalidate is present in the OC/Utilities/ocvalidate folder.

 

When a new commit comes out, I run this on my current config.plist file to see if any changes

were made. If it passes, then what ever was changed in that commit, is not affecting the config.plist

file structure.

 

In the example, an error as found. The errors are due to the "fill-in" entries I made in the PlatformInfo section discussed above.

Once these have proper entries, if you run again, the errors will be gone.

 

 

 

 

*****

 

UPDATE (22 July 2021):

 

This is an expansion on how to cancel the internal AX200 WiFi/BT module discussed above in the ACPI section. For more discussion on the WiFi/BT issue and AirDrop, see this post (p 93).

 

After removing HS05, while the internal AX200 has its BT disabled (after removing its USB power), the WiFi portion at BYS4 is still active (although somewhat broken) as shown below:

 

 

To completely remove the device, another SSDT is required, 9-SSDT-TRX40-OFF-AX200.aml (the code is shown in Spoiler below and the file added as another attachment).

 

Spoiler

DefinitionBlock ("", "SSDT", 1, "APPLE ", "NoWLAN", 0x00001000)
{
    External (_SB_.S0D2.D2A0.BYUP.BYD4, DeviceObj)
    External (_SB_.S0D2.D2A0.BYUP.BYD4.BYS4, DeviceObj)

    Scope (_SB.S0D2.D2A0.BYUP.BYD4.BYS4)
    {
        Method (_DSM, 4, NotSerialized)  // _DSM: Device-Specific Method
        {
            If (!Arg2)
            {
                Return (Buffer (One)
                {
                     0x03                                             // .
                })
            }

            Return (Package (0x06)
            {
                "class-code", 
                Buffer (0x04)
                {
                     0xFF, 0xFF, 0xFF, 0xFF                           // ....
                }, 

                "vendor-id", 
                Buffer (0x04)
                {
                     0xFF, 0xFF, 0x00, 0x00                           // ....
                }, 

                "device-id", 
                Buffer (0x04)
                {
                     0xFF, 0xFF, 0x00, 0x00                           // ....
                }
            })
        }
    }
}

 

 

 

After rebooting with this SSDT enabled, the following is seen. Now there is no device at BYS4 and the power is removed at HS05.

 

 

Note that this discussion is based on the internal AX200 being located at BYS4. Check IORE to see if your device is at a different location. If it is, you'll need to accordingly modify the SSDT files.

 

 

EFI.zip

 

9-SSDT-TRX40-OFF-AX200.aml.zip

 

Edited July 22, 2021 by iGPU
Update 22 July to add how to cancel BYS4

[iGPU]

59 minutes ago, fabiosun said:

A spike of 63 degrees when system is doing nothing for me it is no good

the sense of the test I was asking is to see if also for other users with that Efi is possible to have a temperatures reduction as I see of 10 degrees and more.

then we can also  discute about mat or what it is correct for an Apple system with an unsupported thread ripper Cpu
 

 

Small spike seems of no concern to me. I do not have AIC water cooler for CPU, so maybe worse transients in my case.

 

But even when I tested on other Intel systems in past with AIC water coolers, temps were always jumping up and down based on what's happening. I'm not convinced of overall accuracy and steady states of these values. I take them as relative values looking for significant over-heating problems, mainly during the initial build or for build changes. Aside from that I don't use or run the kexts (more overhead, no?).

 

In BIOS, I have fans set to kick up based on temps. So if the fans are quiet, I know the temps are not bad and I get on with using the computer. If the fans were to stay at full throttle, then temps are elevated and I'd need to investigate. (I've never had this happen to date.) Fans average out the small blips in the temp readings. I do not hear fan surges, accompanying those spikes I posted earlier.

 

 

[fabiosun]

@iGPU

I just asked to test that EFI in the conditions that are necessary to make it go and see if on yours or on other trx40 systems there is the same decrease in temperature

at the moment it was not possible.

To explain me better, with the new patch system, converting a 30-minute 4k video with Davinci resolve to DCP format I usually reach temperatures of 70 degrees with the efi in OP

With the new patches, quirks and so on about 84 degrees 

t's too much for me

 

EFI can also be completely adhering to the principles of configuration.pdf, but that's not good for me and I wanted to understand why

At the moment it was not possible but I remind you if we had followed the specifications and indications of the manual or the opencore devs, perhaps today we would still be in ProxMox 🙂 I

don't want to remind you how they define an amd system that uses" junk kernel patches to boot", also called "borked"

Fortunately, we tried other things .. and it's going well.

thanks anyway for your detailed messages 🙂

[fabiosun]

added in OP Opencore 071 EFI for my rig (for 11.4.x and 11.5.x) I am using by now Clover 5137 to boot in Monterey B1

It uses @iGPUreduced set of Kernel patches and some quirk adjustment for it (RebuildAppleMemoryMap and others)

 

UPDATE 17-07-2021