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

[startergo]

You can add refind to its own partition and add the appropriate core services folder and get refind showing in the system preferences prefpane. In theory that should tell refind to boot by default after OC. I had started down that path with my 3,1 but simply hadn’t the time to finish. I think the “connectdrivers” thing was killing me. As I could chainload sometimes I just didn’t realize why The behavior changed.
the install for refind calls this “ownhfs” but it does also work with apfs formatted drives. I didnt use the refind installer I simply manually placed the files in the right path and they showed up in the prefpane boot selector automatically.

Have you tried

GitHub - abdyfranco/next-loader: Next Generation EFI-Based Boot Manager designed for Intel Mac.

Next Generation EFI-Based Boot Manager designed for Intel Mac. - abdyfranco/next-loader

github.com

This is an enhanced refind version.

 

[h9826790]

You can add refind to its own partition and add the appropriate core services folder and get refind showing in the system preferences prefpane. In theory that should tell refind to boot by default after OC. I had started down that path with my 3,1 but simply hadn’t the time to finish. I think the “connectdrivers” thing was killing me. As I could chainload sometimes I just didn’t realize why The behavior changed.
the install for refind calls this “ownhfs” but it does also work with apfs formatted drives. I didnt use the refind installer I simply manually placed the files in the right path and they showed up in the prefpane boot selector automatically.

Thanks for the info. Sure I can try this path.

But I am a bit confused at this moment.

From my experience back in OC 0.5.2. If I select another start up disk in system preference, this will
1) disable OpenCore, if RequestBootVarRouting is disabled
2) no effect, if RequestBootVarRouting is enable in OpenCore

So, if that "start up disk" is rEFInd, I still can't get "OpenCore -> rEFInd".

However, from post #905, point 2, I should able to do it.

OpenCore on the Mac Pro

@startergo, I think we both may still have some level of misunderstanding. 1. Making your Mac boot OpenCore by default is one time operation during OpenCore setup. For this you somehow need to register it as the default in your firmware preferences. I can imagine the following ways: — From...

forums.macrumors.com

Most likely I did something wrong before. However, I am quite sure I did try select Windows as next boot drive in system preference with RequestBootVarRouting enabled. The result is the cMP reboot back to Mojave. I have to disable RequestBootVarRouting in other to make that start up disk selection work as expected.

I checked the OpenCore Changelog, it seems nothing changed for RequestBootVarRouting. Anyway, I will try again to enable RequestBootVarRouting and see if I can switch between OS. If yes, then I can go further to your suggested path. Which seems a good way to achieve what I want (a boot menu that can be switched On/Off via system preference -> start up disk)

 

[cdf]

I am quite sure I did try select Windows as next boot drive in system preference with RequestBootVarRouting enabled. The result is the cMP reboot back to Mojave. I have to disable RequestBootVarRouting in other to make that start up disk selection work as expected.

This is expected for CSM installations of Windows.

 

[h9826790]

Thanks for the info. Sure I can try this path.

But I am a bit confused at this moment.

From my experience back in OC 0.5.2. If I select another start up disk in system preference, this will
1) disable OpenCore, if RequestBootVarRouting is disabled
2) no effect, if RequestBootVarRouting is enable in OpenCore

So, if that "start up disk" is rEFInd, I still can't get "OpenCore -> rEFInd".

However, from post #905, point 2, I should able to do it.

OpenCore on the Mac Pro

@startergo, I think we both may still have some level of misunderstanding. 1. Making your Mac boot OpenCore by default is one time operation during OpenCore setup. For this you somehow need to register it as the default in your firmware preferences. I can imagine the following ways: — From...

forums.macrumors.com

Most likely I did something wrong before. However, I am quite sure I did try select Windows as next boot drive in system preference with RequestBootVarRouting enabled. The result is the cMP reboot back to Mojave. I have to disable RequestBootVarRouting in other to make that start up disk selection work as expected.

I checked the OpenCore Changelog, it seems nothing changed for RequestBootVarRouting. Anyway, I will try again to enable RequestBootVarRouting and see if I can switch between OS. If yes, then I can go further to your suggested path. Which seems a good way to achieve what I want (a boot menu that can be switched On/Off via system preference -> start up disk)

Just tested, RequestBootVarRouting work as expected now.

I can stay boot via OpenCore, and switch between Mojave / Catalina in system preference -> start up disk.

So now, I can go next step and see how to make rEFInd shows up in system preference, and bridge boot that via OpenCore in order to provide boot selector for Radeon VII.
[automerge]1580748625[/automerge]

Have you tried

GitHub - abdyfranco/next-loader: Next Generation EFI-Based Boot Manager designed for Intel Mac.

Next Generation EFI-Based Boot Manager designed for Intel Mac. - abdyfranco/next-loader

github.com

This is an enhanced refind version.

Thanks for this suggestion. I will keep this in mind. But at this moment, I want something that can show up in system preference - start up disk. Since Ludacrisvp has experience on that. I will start with rEFInd first.

 

[Ludacrisvp]

Have you tried

GitHub - abdyfranco/next-loader: Next Generation EFI-Based Boot Manager designed for Intel Mac.

Next Generation EFI-Based Boot Manager designed for Intel Mac. - abdyfranco/next-loader

github.com

This is an enhanced refind version.

I’ve not tried next-loader.

after perusing the github for this it seems to be a dead / abandoned project.

 

[startergo]

I’ve not tried next-loader.

after perusing the github for this it seems to be a dead / abandoned project.

It worked last time I tested it but yes it is abandoned.

 

[h9826790]

I remember something now. When I first boot with the Radeon VII on my cMP, I released that has a little different behaviour than the RX580.

With non flashed RX580, my monitor has no sign to wake up until OS GPU driver is loaded.

With Radeon VII, my monitor will try to wake up as soon as I power up my cMP. But then get no signal and sleep again until OS GPU driver is loaded.

I think this is a clear sign that the Radeon VII's GOP is somehow different than the RX580. Even it seems "more compatible to cMP" (because able to wake up the screen once power on), but this actually make it can't display the OC boot picker.

 

[vit9696]

@h9826790 it may be that VII GOP is simply broken, but there still are some tests one can do to check this.

@df3475, as for broken colours, this looks like your screen is operating in 30-bit mode yet GOP believes it is 24 or something like that. I think it might still be doable to resolve this.

I will try to find some time this week or a bit later to have a look at this and PM you. Do not expect much, as we are moving on to other stuff though.

 

[h9826790]

@h9826790 it may be that VII GOP is simply broken, but there still are some tests one can do to check this.

@df3475, as for broken colours, this looks like your screen is operating in 30-bit mode yet GOP believes it is 24 or something like that. I think it might still be doable to resolve this.

I will try to find some time this week or a bit later to have a look at this and PM you. Do not expect much, as we are moving on to other stuff though.

Totally understand that, thanks for all your help and contribution anyway.

For info, there are two members here with Radeon VII get exactly same symptom. Therefore, most like not a particular faulty Radeon VII's problem.

Also, in both case, the Radon VII can display the verbose mode / Apple logo with loading bad perfectly once go beyond the OS boot picker. This is another sign that the GOP itself is working, just somehow can't display the boot picker yet.

 

[rroumen]

@h9826790 it may be that VII GOP is simply broken, but there still are some tests one can do to check this.

First of all thanks a lot vit for your support. I think I speak for many cMP users who already love OpenCore and it is getting better with every rev update. About the radeon VII I have the feeling we are closer to solution than you think. Without the:

• Added Mac GOP support to ProvideConsoleGop quirk

I would get NO first graphics boot screen with apple logo. That now works flawlessly and also the monitor sleeps directly after boot but turns on again at the point where the boot picker is displayed. I will try a bit more with different settings to see if I can get anything.

 

[rroumen]

Well, after messing with config.plist settings for 1.5 hour and multiple pram resets to recover from failing boots here my conclusions:
* ProvideConsoleGop quirk is required, no display until final apple boot logo display without this.
* If I set builtin text renderer I HAVE to set Display Resolution to something e.g. "Max" or I get an "OCC: Cannot setup ASCII output" error and there is no display until final boot.
* I tried lower resolution but no success
* I tried 80x25 console setting, no success
* I tried multiple combinations of resolution, console setting and builtin text rendered false/true but all without success.
* I tried ForceText and ForceGraphics on Os and UI but this doesn't seem to make any difference, behaviour stays the same: No boot text and full graphics boot logo.

* Something to note, during the many reboots I did I have seen several times a white block in the top left corner of an otherwise black screen, this stayed there for few seconds until display output was reset and it looked a lot like a cursor. When I tried to take a picture of course I didn't see it.

So far my best config.plist attached below.

 

[astonius86]

Still looking for some insight into my Windows booting issue if anyone can help. Log attached.

 

[cdf]

When booting with GOP support into Mojave or earlier you may see some text on the Apple logo screen. You can hide this text by enabling IgnoreTextInGraphics.

 

[ashman70]

I have read the install instructions for Opencore and they seem rather convoluted. Do I just replace my config.plist file with the one provided or is there more to it?

 

[h9826790]

When booting with GOP support into Mojave or earlier you may see some text on the Apple logo screen. You can hide this text by enabling IgnoreTextInGraphics.

Thanks for this info. I am also studying if anything in UEFI Quirks can help the Radeon VII to show boot picker.

e.g. ReplaceTabWithSpace, or ReconnectOnResChange, etc

 

[rroumen]

Thanks for this info. I am also studying if anything in UEFI Quirks can help the Radeon VII to show boot picker.

e.g. ReplaceTabWithSpace, or ReconnectOnResChange, etc

Was just thinking the same thing and haven’t tried these yet. I think I tried one of the quirks without the ForceGraphics setting which lead to a non booting system and PRAM reset...

 

[h9826790]

Was just thinking the same thing and haven’t tried these yet. I think I tried one of the quirks without the ForceGraphics setting which lead to a non booting system and PRAM reset...

I turn those options on one at at time now. So far, tested

ReplaceTabWithSpace
ReconnectOnResChange
SanitiseClearScreen
AppleUserInterfaceTheme
AppleImageConversion

with ForceGraphics and ForceText, no luck so far.

Always the same situation, no OC boot picker, but then the Apple boot logo show up perfectly before loading bar start.

 

[cdf]

Always the same situation, no OC boot picker, but then the Apple boot logo show up perfectly before loading bar start.

Perhaps you could try disabling ShowPicker and hold Escape during boot to bring up the menu (Option should also work, but the timing has to be right).

 

[h9826790]

Perhaps you could try disabling ShowPicker and hold Escape during boot to bring up the menu (Option should also work, but the timing has to be right).

Just read that part 3 minutes ago, and about to test it now ?

Anyway, anyone interested to test if can boot via USB 3.0?
XhciDxe — XHCI USB controller support driver from MdeModulePkg. This driver is included in most firmwares starting with Sandy Bridge generation. For earlier firmwares or legacy systems it may be used to support external USB 3.0 PCI cards.
[automerge]1580831235[/automerge]

Perhaps you could try disabling ShowPicker and hold Escape during boot to bring up the menu (Option should also work, but the timing has to be right).

Tested, no change.

100% sure the boot picker is working in blind. My screen stay black (but awake) until I press Enter. Then the Apple logo shows up in no time.

 

[startergo]

I finally got my hibernation working using this kext (Lilu dependent). There are no more crashes, sutdowns or KP. That will extend the life of the internal components (especially the PSU) and reduce the heat output. I also managed to do WOL using this tool from another PC on the local network so I can do remote login to my cMP. For some reason WOL from the open-wrt router Luci does not work.

 

[Ludacrisvp]

I finally got my hibernation working using this kext (Lilu dependent). There are no more crashes, sutdowns or KP. That will extend the life of the internal components (especially the PSU) and reduce the heat output. I also managed to do WOL using this tool from another PC on the local network so I can do remote login to my cMP. For some reason WOL from the open-wrt router Luci does not work.

What configuration was hibernate not working with (curious if it was a "just you" or an "anyone" type issue)?
Also, was sleep impacted or was it only hibernate (which I see as 'write to disk and full power down')?

 

[startergo]

Also, was sleep impacted or was it only hibernate

Sleep. Most of the time it would not re-initialize the monitor from sleep (black screen), crash (KP) or shut down.

 

[startergo]

@h9826790 The USB3 driver XhciDxe.efi does not let the machine to boot:

02:271 00:073 OC: Driver XhciDxe.efi at 1 is being loaded...
02:356 00:085 OC: Driver XhciDxe.efi at 1 is successfully loaded!
02:415 00:059 OC: Driver XhciDxe.efi at 1 needs connection.
02:474 00:059 OC: Connecting drivers...
03:595 01:120 XhcCreateUsb3Hc: Capability length 0x20
04:422 00:826 XhcCreateUsb3Hc: HcSParams1 0x400087F
04:481 00:058 XhcCreateUsb3Hc: HcSParams2 0xFC0000FA
04:541 00:059 XhcCreateUsb3Hc: HcCParams 0x200EF81
04:613 00:072 XhcCreateUsb3Hc: DBOff 0x1800
04:674 00:061 XhcCreateUsb3Hc: RTSOff 0x1000
04:734 00:059 XhcCreateUsb3Hc: UsbLegSupOffset 0x0
04:794 00:060 XhcCreateUsb3Hc: DebugCapSupOffset 0x80
04:868 00:073 XhcDriverBindingStart: failed to enable 64-bit DMA on 64-bit capable controller @ 75A38F18 (Unsupported)
04:928 00:060 XhcSetBiosOwnership: called to set BIOS ownership
04:987 00:058 XhcResetHC!
05:114 00:127 XhcInitSched:DCBAA=0x7281C000
05:952 00:838 XhcInitSched: Created CMD ring [7281C400~7281D400) EVENT ring [7281D400~7281F400)

 

[Ludacrisvp]

@h9826790 The USB3 driver XhciDxe.efi does not let the machine to boot:

02:271 00:073 OC: Driver XhciDxe.efi at 1 is being loaded...
02:356 00:085 OC: Driver XhciDxe.efi at 1 is successfully loaded!
02:415 00:059 OC: Driver XhciDxe.efi at 1 needs connection.
02:474 00:059 OC: Connecting drivers...
03:595 01:120 XhcCreateUsb3Hc: Capability length 0x20
04:422 00:826 XhcCreateUsb3Hc: HcSParams1 0x400087F
04:481 00:058 XhcCreateUsb3Hc: HcSParams2 0xFC0000FA
04:541 00:059 XhcCreateUsb3Hc: HcCParams 0x200EF81
04:613 00:072 XhcCreateUsb3Hc: DBOff 0x1800
04:674 00:061 XhcCreateUsb3Hc: RTSOff 0x1000
04:734 00:059 XhcCreateUsb3Hc: UsbLegSupOffset 0x0
04:794 00:060 XhcCreateUsb3Hc: DebugCapSupOffset 0x80
04:868 00:073 XhcDriverBindingStart: failed to enable 64-bit DMA on 64-bit capable controller @ 75A38F18 (Unsupported)
04:928 00:060 XhcSetBiosOwnership: called to set BIOS ownership
04:987 00:058 XhcResetHC!
05:114 00:127 XhcInitSched:DCBAA=0x7281C000
05:952 00:838 XhcInitSched: Created CMD ring [7281C400~7281D400) EVENT ring [7281D400~7281F400)

Perhaps it’s just an unsupported card by this driver?

 

[startergo]

Perhaps it’s just an unsupported card by this driver?

The only way to find out is if someone else tests it.