Open Core NVRAM Emulation - Initiating

[Nemfrod]

Hi all!

[Apologies; I am naturally verbose; I just want to get this posted.] 

I want to get this question posted, so -- with apologies -- I am just going to start with the simple question, and work up from there if necessary (to writing down error messages for booting-variations).  [Also, my config.plist et al are on another machine (being the machine I am trying to set up).] 

I had a working Clover/Sierra installation on my triple-boot GA-Z170X-UD5, except I can not put it to sleep [I leave it running continuously], so -- when some of my HDDs failed -- I thought I would do an Open Core [OpenCore] installation. 

The problem is that... it turns out that the Clover and the Open Core installations use different NVRAM settings, and thus can not be both run on the same hardware. 

My solution was, of course, to use emulated NVRAM for both.  That is... the idea would be to switch across to Open Core [precious sleep!!], but I do not want to commit myself until it is working properly.  (The incumbent Clover setup is now working [again] that way.)

 

The question is... is this a catch-22?  [Presumably not, of course.]

I *can* boot, starting initially with either the installation USB or the installed EFI (noting that the screen is obviously slow, the About This Mac does nothing and it thinks the NVidia driver is incompatible, and sleep is odd)... without any attempt at using emulated NVRAM. 

I *can not* boot the installed OS with it configured for emulated NVRAM.  [From memory, it outright refuses to recognise it as a bootable volume... although that might have been because I selected the non-UEFI boot alternative -- in which case the problem is "not a properly-configured OC installation"-ish.]

I *can* boot from a plain USB into an installed OS set up for emulated NVRAM. 

I can not remember what happens if I try to boot into the installed OS via the installation USB, with the latter set up for emulated NVRAM.

 

I take it that I am supposed to boot into the installed OS -- with or without the Emulated NVRAM set up in the config.plist, I do not know -- presumably from/using my non-Emulated_NVRAM set-up USB... and then run the thing that dumps the NVRAM to NVRAM.plist [LogoutHook/launchd].

With Emulated NVRAM set up, I can not get that far [from memory]. 

[Conversely] starting with a config.plist set up for normal NVRAM... I get an error as follows. 

 

nvram: Error getting variable - '4D1FDA02-38C7-4A6A-9CC6-4BCCA8B30102:boot-path': (iokit/common) data was not found

 

This happens with or without an existing NVRAM.plist file.  (I tried borrowing the one from Clover.)  I tried manually editing the pertinent entry into said NVRAM file, and it fails (differently) -- with the following error essage.

 

Fatal error: Cannot access device node!

 

It does not make sense to me; it has booted successfully, and it is trying to *create* the NVRAM.plist file; it should work if there is no such file present.  [I have tried that.]

Incidentally,  " sudo nvram -p "  does return some NVRAM settings.

I would have thought it would work if I...  * had an NVRAM.plist file,  * booted without/with?? Emulated_NVRAM,  * reset the NVRAM,  * launched the target OS set up for emulated NVRAM, and  * ran the dump program.  (Again... I *have* tried it *without* an NVRAM.plist file.)

 

Again, apologies, but I would have to spend an hour or so switching things around and rebooting to get all the right error messages listed verbatim.  I can/shall do that if it is more complicated than just doing things in the right order.

 

Thank you in advance.

 

p.s.  Is there an easier way to have Open Core and Clover on the same hardware?

 

____  Update  ____

 

I have added (in config.plist), under NVRAM/Add/  4D1FDA02-38C7-4A6A-9CC6-4BCCABB30102,  an entry as follows.

    <key>boot-path</key>

    <string>PciRoot(0x0)/Pci(0x1C,0x5)/Pci(0x0,0x0)/Sata(0x1,0xFFFF,0x0)/HD(1,GPT,{UUID})</string>

(This is listed by the MAP command in the EFI shell.  [Always append " -b", for a page-by-page view.]  [E.g. "FS2:" then "LS" to see the contents.])

(I also had this address plus "/\EFI\OC\OpenCore.efi" (and separately {that} but without the starting "/") but that makes no difference.)

I no longer get the "Error getting variable..." error noted above.  Noting that I have created a folder "NVRAM" (from the root of the EFI folder)...now, it says as follows.  [Later:  that [folder creation] was on the EFI volume, and the LogoutHook program is using my MacOS HDD, so it is irrelevant.]

    "

Found boot device at /dev/disk4s2

Already mounted at /

cp: //NVRAM/nvram.plist: Permission denied

Fatal error: failed to copy nvram.plist!

    "

Note the extra "/" -- at least I take it that there is an extra "/".

!  Note that disk4s2 is the actual MacOS partition, not the EFI partition.

The interesting thing is that it also makes no difference whether I use the address of the MacOS partition or the EFI partition.  It seems that  (although it did not work without a "boot-path" entry)  it gets the path itself... but stuffs up by adding an extra "/".  Possibly this is a bug?

 

... Okay...  I just got it to write the file.  It *is* a permissions error (and apparently the "/" is *not* extra).  I had read-only access to the MacOS HDD [which strikes me as very odd], but it successfully created the "NVRAM" folder anyway (!?)... but could not write to it.  I changed the permissions of "Everyone" for the new folder, from read-only to read-write, and Voila! -- it worked.  (I also changed it so that I, and Administrator, have read-write access to the drive.)

 

Now the question is...  (a) do I change it back so that I have read-only access to my own HDD (!?!), and  (b) who needs permission to the NVRAM folder for it to work?  ..And  (c) shouldn't it be using the EFI volume?

(Of course, the question remains of why I have to do this stuffing around when (apparently/presumably) it works OOB for everyone else.)

Edited March 13, 2023 by Nemfrod
Updated efforts... and I have fixed it.