MP 1,1-5,1 RefindPlus|OpenCore by MyBootMgr

[JedNZ]

I have another question, and it relates to updating BigSur to 11.6.3.

I updated BigSur to 11.6.3 but encountered issues that I didn't experience with 11.6.1 and 11.6.2 (which both rebooted a couple of times, changing to the temporary Macintosh HD update drive (RAM disk?) then back again to my BigSur drive).

But not with 11.6.3, which didn't appear to be able to temporarily bless to the Macintosh HD drive, then switch back to my BigSur boot drive afterwards. Instead, I had to reset NVRAM, then boot into Mojave Recovery and disable SIP, then boot into Mojave and re-bless RefindPlus using Terminal, and then I was able to see the Macintosh HD option in the OC picklist which then carried on updating to 11.6.3, rebooting again and finally completing the process.

I'm trying to work out if this is a setting in RefindPlus (EFI/BOOT/config.conf) or if it's in EFI/OC/config.plist somewhere. I've had a look at Bless info in the OpenCore documentation (section 8.1) but it doesn't answer my questions, and in the RefindPlus (EFI/BOOT/) config.conf file it has a section relating to ignore_previous_boot, but I can't make head or tails of whether that's what I need to uncomment to activate it or not, and the same for normalise_csr. Happy to take this to the OpenCore thread if someone can confirm it's nothing to do with RefindPlus.

 

[Dayo]

A lot to catch up on but first thing is, I see a lot of "123" issue since "XYZ" but you need to remember that MyBootMgr covers a wide spectrum and I can't test every single one exhaustively. Hence, if there is no report about something, it doesn't get fixed. Secondly, reports have to provide enough information.

Anyway, here goes...

I noticed that my PCI drives have not been identifying as internal (since v075b?), which is what Innie should be doing. I noticed that Innie v1.2.1 is still being used in your package - so I copied the latest v1.3.0 into the EFI/OC/Kexts folder, and my system now shows PCI drives as internal. Can you please update this at some stage.

Innie is hit or miss and it is more likely that it will be removed altogether. Development appears to have stopped and it does not look like a long standing bug in v1.3.0 will ever be fixed. The situation v1.3.0 creates problems in other setups and ConfigFactory went back to v1.2.1 as it is stable for the scenarios where it works and does not result in greater problems than drives not being internal.

As mentioned earlier, ConfigFactory covers a wide spectrum but there are instances where you just have to do certain things yourself.

In the case of trying to make PCIe drives appear as internal:

1. It tries to get the DeviceProperty paths of drives via one method
   1. There there are situations where it cannot or where this does not work.
   2. It tries to test to see whether there is the situation where this will not work
2. It tries an alternative way to get the DeviceProperty paths if there are potential issues with the first way
   1. There there are situations where it cannot or where this does not work.
   2. It tries to test to see whether there is the situation where this will not work
3. It informs you and offers to use a kext and gives you two options
   1. If one doesn't work, try the other

If you still don't get the desired outcome after all this, you need to intervene manually.

However the next restart goes to a black screen and my cMP switches off after a minute or so.

This is an example of a report where there just isn't enough information to work with. Without log files, OC/RP, it might be possible to immediately know what is happening but in most cases not. No need here since you ultimately overcame the issue.

at least I know all I have to "adjust" in the config.plist that MyBootMgr creates is to delete the DeviceProperties entries for my GPU and it works, which is a great relief of sorts.

Well, blindly deleting the DeviceProperties is not a good idea. It was only recommended for a specific situation earlier.

Just made the change back to Innie from AHCI_3rdParty, and restart worked fine (no black screens of crash).

As mentioned, Innie v1.3.0 is hit or miss and AHCI_3rdParty covers some but not all setups. The ultimate way is to use DeviceProperties to make your drives appear as internal and if ConfigFactory cannot get it so, and the kext does not, you have to follow instructions to do it manually. Instructions are in the OpenCore for MacPro thread.

Happy to take this to the OpenCore thread if someone can confirm it's nothing to do with RefindPlus.

RefindPlus settings have nothing to do with what appears in OpenCore. There is often a mental block on this but as mentioned often, these are two separate utilities which Configfacgtory set up to co-exist. OpenCore is not influencing RefindPlus and the same applies in the other direction. To summarise, if something is observed in OpenCore, making a change in RefindPlus is not the answer.

For your specific case, did you use the OC_ALT instance for the update? Actually, your posts are framed as if you only have one OC instance. What happened to the others? Anyway, as I can see, you have run the update. I will leave that matter alone in your shoes.

 

[Dayo]

delete the DeviceProperties entries for my GPU

Update:

I checked the vanilla config file with the GPU DeviceProperties with the OC Author and he can't see any reason why the error should be generated from that file.

Are you sure you did not have another OC setup config, not created by ConfigFactory or manually modified, present in another ESP on your machine when you were running?

To check things further, can you try this amendment to the config file and let me know if it works?

    <key>DeviceProperties</key>
    <dict>
        <key>Add</key>
        <dict>
            <key>PciRoot(0x0)/Pci(0x3,0x0)/Pci(0x0,0x0)</key>
            <dict>
                <key>model</key>
                <string>AMD Radeon RX 590 - TEST</string>
            </dict>
        </dict>
        <key>Delete</key>
        <dict/>
    </dict>

This only disables the Delete section

If it works, please load VideoProc and share a screenshot of the screen that shows the GPU Acceleration status.
I want to see whether you get a AMD Radeon RX 590 - TEST string there.

Thanks

 

[Dayo]

Update 2:

I have used the same GPU DeviceProperties setup you have issues with on my machine (with the PCI path amended accordingly) and it works as expected. So the OC Dev is correct that there is something else going on with your machine.

Suspected situations are:

• You have another OC setup with a corrupted config file, perhaps wrongly manually amended, on another ESP that is what is actually being loaded when you think this is being loaded.
  • This can happen after an NVRAM reset and you have multiple ESPs setup with OC
• You have some form of FileSystem Corruption
  • This could be a manifestation of trying to run OC from an ESP on a drive connected to a PCIe Slot.
• Some other unknown/unusual issue

 

[JedNZ]

Thanks for that Dayo.

I’ve checked all my drives and there are no other RefindPlus and/or OC ESP setups anywhere.

My ESP is located on a HFS+ formatted hard drive in one of my SATA II bays that only contains backed up files (I.e. no OS therefore not bootable), so not in a PCI slot and meets the recommendations of it being HFS+ formatted.

I’ll try the DeviceProperties you suggested when I get home after work and let you know the results. Thanks for looking into it.

 

[JedNZ]

For your specific case, did you use the OC_ALT instance for the update? Actually, your posts are framed as if you only have one OC instance. What happened to the others? Anyway, as I can see, you have run the update. I will leave that matter alone in your shoes.

OC_ALT failed for me, hence why I just tried updating to 11.6.3 with OC. I will explore why OC_ALT didn't work once OC is sorted.

To check things further, can you try this amendment to the config file and let me know if it works?
<key>DeviceProperties</key>
<dict>
<key>Add</key>
<dict>
<key>PciRoot(0x0)/Pci(0x3,0x0)/Pci(0x0,0x0)</key>
<dict>
<key>model</key>
<string>AMD Radeon RX 590 - TEST</string>
</dict>
</dict>
<key>Delete</key>
<dict/>
</dict>

That worked fine - no issues with OC and OC pick list and the GPU shows correctly "AMD Radeon RX 590 - TEST" - screenshots below.

 

[Dayo]

I wonder whether it was just a temporary glitch, something in the NVRAM perhaps, and whether it will now work if you also restore the Delete section. Can you try this to close out?

Speaking of NVRAM, I wonder why you needed all this:

I had to reset NVRAM, then boot into Mojave Recovery and disable SIP, then boot into Mojave and re-bless RefindPlus using Terminal

Why not just toggle SIP off in RefindPlus then boot directly into Mojave from there to do what you wanted to do?

 

[Macschrauber]

As this was a 4.1 the nvram could be messed up. If there is some oddity in it considder a firmware reconstruction.

or at least force a nvram garbage collection by triple nvram reset.

 

[JedNZ]

I wonder whether it was just a temporary glitch, something in the NVRAM perhaps, and whether it will now work if you also restore the Delete section. Can you try this to close out?

I’ll try that. If it works, all good. If it doesn’t, is your temporary fix okay to stay in place (I.e. will it harm anything having nothing in the delete section?)

Why not just toggle SIP off in RefindPlus then boot directly into Mojave from there to do what you wanted to do?

I had to go through the convoluted NVRAM reset because I couldn’t get to the RefindPlus screen - my cMP just blank screened and would shut off.

As this was a 4.1 the nvram could be messed up. If there is some oddity in it considder a firmware reconstruction.

or at least force a nvram garbage collection by triple nvram reset.

I did the 5x NVRAM reset already but maybe it is an issue with my firmware - I was thinking of getting the reconstruction (already have the info from Tsialex).

 

[Dayo]

The Delete Section will likely be removed in the next release. It is only there as an artefact of the option to use the Orinoco framebuffer. That is, it went in because of that option and a few other things were added just because they could be added. However, the Orinoco framebuffer option was recently removed from ConfigFactory and there is no actual need for that section at this time.

Nothwithstanding this, it should not be causing the observed issue (and may be needed for something else later) hence trying to reach a conclusion. In summary, even if it works, you can take it out as this is likely to be the default as said.

Getting round to some oddities in your ConfigFactory debug logs

1. You say you have PCIe drives that you want to appear as internal but ConfigFactory returns a PCIe Count of Zero (0). That is, it didn't detect any PCIe Drive.
   • Can you run this command in Terminal and post the result? It might be longish and need to go under spoiler tags: system_profiler SPPCIDataType | grep -B 2 'AHCI Controller\|NVM Express Controller'
2. ConfigFactory only detects One (1) Mac OS instance, presumably the Mojave one, but you are apparently also running Big Sur as well (as a minimum).
   • Did you disconnect the disk(s) with other Mac OS (Big Sur or other) while running ConfigFactory or if connected, is it on a SATA port or on a PCIe Slot?
3. Can you share a screenshot of the RefindPlus Screen?
   • Press F10 to get this. The file will be saved in the ESP.

You might want to activate the firmware rebuild as it is bascially an essential item for cMP by now.
Either one of Macschrauber and tsialex can do the deed for you.

I couldn’t get to the RefindPlus screen - my cMP just blank screened and would shut off.

More so in light of this.

 

[amstel78]

OC_ALT failed for me, hence why I just tried updating to 11.6.3 with OC. I will explore why OC_ALT didn't work once OC is sorted.

FWIW, I did not have to use OC's alternate instance when updating from 11.2 to 11.6.3. I just let it do its thing under the standard OC instance. After downloading, it rebooted, selected the update volume automatically in OC boot picker, installed, rebooted again, and I was up and running without issue.

 

[Dayo]

No Response on Requests for Info ... Moving on

---

Updated MyBootMgr to v078
See change log in Post 1 for details

 

[Dayo]

Updated MyBootMgr to v079
See change log in Post 1 for details

 

[amstel78]

Updated MyBootMgr to v079
See change log in Post 1 for details

v079 installed without any issues on my end. Thanks, Dayo.

 

[JedNZ]

v079 didn’t copy Innie 1.30 into the Kext folder for me. Otherwise works fine.

 

[Dayo]

For the "crazies" running Monterey and wondering about reported 12.3 issues, note that the required Power Management Patch, ASPP-Override.kext, was added in the MyBootMgr v078 release:



I don't run Monterey myself, but should be good to go.

 

[startergo]

ASPP-Override.kext

Isn't this AGPM-override?

 

[Dayo]

The kext is called ASPP-Override

 

[freqrider]

For some reason I can’t boot windows Legacy anymore via RefindPlus? After updating to 0.7.9 package? Shows up in bootpicker? I will try older version to see if it an issue with windows or the latest package.

Update: Can confirm, .0.7.8 boots Win Legacy. I will try to rebuild 0.7.9 and see? The 0.7.8 config in 0.7.9 did not work either? I noticed I could no longer see my Linux drive in 0.7.9 even when selecting ‘yes’ to ‘are you running linux?’ In Config.

 

[Dayo]

Seems you are not actually asking questions on the sentences where you use question marks and that these are in fact statements and not questions.

This odd writing style makes it really difficult to follow the logic of your post.

Anyway, there is no reason why RP in v079 should behave differently to that in v078 in terms of Legacy Windows but you can always just swap the older one in if you prefer. Will check this later.

In any case, a report about RP without a log file, preferably on GitHub, is of little use. Especially one like not seeing Linux.

 

[freqrider]

Seems you are not actually asking questions on the sentences where you use question marks and that these are in fact statements and not questions.

This odd writing style makes it really difficult to follow the logic of your post.

Anyway, there is no reason why RP in v079 should behave differently to that in v078 in terms of Legacy Windows but you can always just swap the older one in if you prefer. Will check this later.

In any case, a report about RP without a log file, preferably on GitHub, is of little use. Especially one like not seeing Linux.

While I can understand it being a bit confusing to some, using a question mark at the end of a statement is justifiable and perfectly acceptable in some instances and considered to be proper English. It can imply shock, disbelief or sarcasm, to give some examples, as if to say “really?”

When Is a Question Not a Question? - DAILY WRITING TIPS

Inclusion or exclusion of a question mark is usually a straightforward matter. However, there are instances in which what are framed as questions should end

www.dailywritingtips.com

As for latest RP not working, I swapped my 0.7.8 build and it now boots windows/linux fine. I have also swapped out the OC instances for my custom config as your builds haven't been able to boot my Monterey installs for some time. Fresh, 0.7.9 build did not boot 12.3 for me.

I will see if I can enable logging again and post them on GitHub but with my custom setup, I’m afraid it will only add to the confusion. (Yes, I’m one of the ‘crazies’ running latest Monterey on my 3,1!)

Here’s another example for you Spock (logical) brained individuals.

I went for a walk last night and I think I saw a ghost? 👻🤪

 

[freqrider]

While I can understand it being a bit confusing to some, using a question mark at the end of a statement is justifiable and perfectly acceptable in some instances and considered to be proper English. It can imply shock, disbelief or sarcasm, to give some examples, as if to say “really?”

When Is a Question Not a Question? - DAILY WRITING TIPS

Inclusion or exclusion of a question mark is usually a straightforward matter. However, there are instances in which what are framed as questions should end

www.dailywritingtips.com

As for latest RP not working, I swapped my 0.7.8 build and it now boots windows/linux fine. I have also swapped out the OC instances for my custom config as your builds haven't been able to boot my Monterey installs for some time. Fresh, 0.7.9 build did not boot 12.3 for me.

I will see if I can enable logging again and post them on GitHub but with my custom setup, I’m afraid it will only add to the confusion. (Yes, I’m one of the ‘crazies’ running latest Monterey on my 3,1!)

Here’s another example for you Spock (logical) brained individuals.

I went for a walk last night and I think I saw a ghost? 👻🤪

Admittedly, I did go a bit overboard with the question marks on my post! 😐

 

[Dayo]

It can imply shock, disbelief or sarcasm

It is just the fast spreading scourge of the High Rising Terminal AKA Australian Question Intonation AKA Uptalk.
It's one thing to listen to it but trying to read it takes things to another level. Anyway, best we move on.

As for latest RP not working, I swapped my 0.7.8 build and it now boots windows/linux fine.

This is the one I need the RP log from the 079 build file. Note that the log request was not OpenCore related, so your customisations do not impact this.

You said the RP in v079 does not boot Legacy Windows and does not detect Linux (Doesn't say which Linux distro) ... I need to see your RP log output. Thanks.

Fresh, 0.7.9 build did not boot 12.3 for me.

Which one of the three instances did you select?
Not all can boot Monterey but at least one will do ... that's the point of having more than one.

I’m one of the ‘crazies’ running latest Monterey on my 3,1!

The "crazies" thing was tongue in cheek and not specifically directed at 3,1 users. MyBootMgr caters for 3,1 to 5,1 as the great thing so far is that anything a 5,1 can run on OpenCore can also be run on a 3,1 (with tweaks at times).

 

[ThetaTauMan]

I have a couple of errors in ConfigFactory, and a not really working OC result. I somehow doubt that they are related, but what to I know.

The two issues in configfactory are (1) following the Wifi question and (2) ... I can't seem to remember ... but it happens repeatably and in both cases terminal is open with a failed action. I am attaching a picture of each, and I assume you will recognize when the second one is happening based on the command it is trying to execute.

One interesting thing (to me) is that I was able to continue past; when I gave up and tried to force quit configfactory, due to it being nonresponsive, it responded by going on to what I assume is the next question.

The OC problem occurs when I actually try to boot. Refind+ works fine, but when I go into either of the OC selections all that OC presents are the reboot, shutdown, reset NVRAM options. No bootable selections. Right now I have only two drives in my cMP 3,1. Both happen to have new installations of ElCapitan, specifically for the purposes of trying to move past that. (My real install drives have all been pulled for the duration of the exercise, until I get things to work.) I have been setting the drive that I have been running configfactory on as the backup boot, and the other drive as the target (with the MyBootManager install of refind/OC in EFI). Both drives show up for selection in Refind, but as mentioned, neither show up in OC, once it boots.

Maybe I am misunderstanding something, but it seems like I ought to be able OC boot something in order to then install High Sierra, etc. (the whole point afterall)

Anyhow, if you need config files or logs, let me know where to find them.

 

[Dayo]

Hi,

The errors in ConfigFactory are because you do not have Python 3 installed.
Actually quite fortuitous as while ConfigFactory tries to work around this, it was set up a long time ago and I am unable to test that condition any longer since I do have Python 3 installed.

The errors can be overcome by installing Python 3 but if you don't mind, I'll like to fix the workaround instead. So hold off installing this.

Initially, can you share the ConfigFactory debug log? You will find it in the /Users/Shared/MyBootMgr folder. If not there, rerun ConfigFactory and select the option to run in debug mode.

Thanks