Closed MarioG-X closed 3 years ago
Please amend your report to follow the issue template including providing debug logs at a level specified in the template. Also, since your issue needs quite a bit of explanation, provide screenshots annotated or commented.
In addition, please restructure to bear in mind that I do not know about your system and something like " Linux partitions on the HD disk1" have no meaning to someone that does not know your system.
What is "HD disk1"? Do you mean Linux Partitions on "HD disk2" show up or do non-Linux disks on "HD disk1" show up and only Linux Partitions do not?
Point is that do you need to mention "HD disk1" to convey the issue you have at all?
I have pasted the template back into your post so you can edit. Thanks
EDIT Thanks for the update. Will take a look and get back
I would leave the title alone since the "APFS volume FX 10.15.7" is also missing. Not only Linux, I clarified. Was in the original post.
The Mac OS display issues are down to the Multi Instance APFS Container item. This needs to focus on the Linux issue.
That puzzles me since the missing FX 10.15.7 volume is in its own container all by itself. It's clear in the diskutil list. If it's related to what you say, why would that one be missing but not Monterey Beta which is also in it's own container? Hopefully your next improvements will get it all. Thanks!
I see. Will take a closer look but let's still split the issues by OS type as most likely any relationships will be along those lines.
BTW, simple test would be just to run with SyncAPFS disabled and see whether the Mac Items return.
I commented syncapfs and your're right, the missing FX 10.15.7 volume did show up, but no Linux's. Unfortunately, so did all the Preboot partition which I had to boot from. Log attached in case it helps: 21r14c3647.log
Please replace your current RP file with the attached (rename as needed) and share the log. X159-BOOTx64.efi.zip
Note that some config settings have been hardcoded in the file and things like volume badges disabled as they can clutter log files.
Also, please run the standard rEFind and share the log from that at Level 4.
Finally, use the exactly the same refind.conf file for both RP and rEFInd runs. You don't need to rename to config.conf (there will be a message in RP but it will go away). The required additional tokens for the test have been set in the efi so no need to add anything extra to the rEFInd config.
Thanks
Here is the log from Refind boot, took a little longer but was normal. Loglevel 4 in refind.conf, I removed RP config for test: refind.log
Here is the log from the RP boot you sent. Unfortunately, RP displayed its title at the top but nothing else. After 4 minutes I had to power down/up. 21r14r4136.log
Now looks like a variant of this issue: https://github.com/dakanji/RefindPlus/issues/45 Will have to go back over that.
FYI, retested and left for 25 minutes, log stopped at same line.
... left for 25 minutes
No need for that. If RP has not loaded within 25 seconds of the header showing, it's not going to be loading.
Anyway, please run the attached and share the log. Just contains additional debug prints to try to see where it is hanging (issue is a memory management issue related to the one linked earlier)
No need for a rEFInd run this time obviously.
Here you go: 21r14x4908.log
Thanks ... please repeat with X161. Added more prints.
161 log, many more lines than 160: 21r15q4441.log
Thanks ... unfortunately not yet enough to be definite but hopefully 162 would do X162-BOOTx64.efi.zip
Hoping, got the same issue here, only see my Catalina Drive and working with APFS, but not seeing my Linux boot manager(s)
162 is MUCH bigger but there was an error trying to open the file. I copied it to another partition and it looks ok: 21r15s5737.log
So I get what is basically happening, a memory addressing conflict in a part of the code dealing with certain Linux setups, but haven't yet gone through searching for this ... likely a typo somewhere.
Anyway, did some general tweaks which may impact on this in X163. If it solves it, great. Otherwise it will have to go to the dreaded line by line check.
Progress! It did not freeze, it displayed a menu, However, there is a major and minor issue:
Linux can be selected and boot started, but all 4 linux's fail quickly at the same place, too soon to create logs. Pic is attached. All macOSs boot fine. Strange but true: when I first tested I booted suse153bum and it booted, but only one time! After that it got the same error as the other three linux's. Power down/up made no difference. The attached log is for the failed suse153bu boot. All partitions boot using refind. 21r16q5701.log Note that all the failed linux logs look the same. Unfortunately, I don't have the log to the one boot that worked.
Menu icon issue: The icons for the two Velementary and Open1 partitions are not correct. In refind, they display the correct icons named .VolumeIcon.icns and .VolumeIcon.png located in the root (/) partition. RP displays a generic penguin for Velementary and dimmed elementary icon for Open1. Refind displays the correct icon for both elementary os partitions, blue and white E, see previous post pics.
Here is where all the linux boots stop:
Great. Should be possible to get across the line then.
Three pics overlapped to fit entire screen. Booting opensuse, note the previous post pic was booting elementary (ubuntu) but both stop at the same place, here is the log to the pics: 21r16s4302.log
Looks like a secure boot issue. Can you try without this activated?
This is a 2014 Mac mini, there is no secure boot. I ran the recovery/security option and it only offers firmware password, no secure boot options.
@MarioG-X isn't it as it states '... SUSE Linux Enterprise Secure Boot Signkey: ...'
Dunno about SUSE and secure boot feature, but seem like it might be, check here
Anyhow, been trying the X163 efi on my Macmini Late 2009, it seem to fail to find my extra drivers as they are located in EFI\refind\drivers_x64\ and not as where it seems to search in EFI\refind\drivers_x64\drivers[_x64]
Will attempt to relocate them there and get back...
This is not a SUSE issue, been booting fine for years with refind. Also booted elementary, fedora, etc. with refind in the past. BTW, my drivers are in EFI\refind\drivers_x64\ and RP finds them fine.
@MarioG-X Right I agree then it ought still to work one would assume...
Funny when booting from my internal ESP volume it finds the drivers from drivers_x64. The screen dump image was an attempt to boot of a REFIND MGR bootable USB Key. But even when finding my drivers incl. the iso9xxx.efi it doesn't find the bootable DVD in my optical drive :/
Also it fails me to make it create a second line entry various tools like shell_x64.efi etc.
@stefws ... please create a separate item to deal with those issues.
@MarioG-X ... Are you still able to boot via rEFInd? Please check. If it goes well, then try X164. Doesn;t include fixes but additional fingerprinting around likely problem area X164-BOOTx64.efi.zip
Original Refind boots all linux and macOS partitions fine. RP 164 is the same as 163, finds all partitions, macOS's boot but all linux's crash as in previous posts. Here is the log to 164: 21r16y2117.log
In case it helps, here is that same partition successfully booting from rEFInd: refind-suse.log
Could it be related to RP passing null load options? It could be why linux can't find things. See this line towards end of RP log: Starting 'vmlinuz-5.3.18-59.19-default' ... Load Options:- '' <– Empty load options
Original rEFInd log shows the load options line as: Using load options 'ro root=LABEL=Vsuse153bum initrd=boot\initrd-5.3.18-59.19-default'
Could it be related to RP passing null load options?
Yes It is. Spotted that in the last log but just got round to attempted fix in X165 X165-BOOTx64.efi.zip
Unfortunately back to hanging displaying only title at top, no menu displayed. log: 21r17t5536.log
BTW, I noticed that for the second disk in the RP log (disk0 in diskutil, SSD 128GB) shown in the log it has the line: Found MBR Partition Table on Volume Below But that is a GUID disk. Also: The volumes FS 10.15.7 and Monterey Beta are not on the second disk, they are on the first disk in the RP log (disk1 in diskutil HHD). Could be I am confused if the second disk for RP is emulated, I noticed it has all the macOS volumes no matter what physical disk they are on. If this sounds suspicious, look at the diskutil lists in the previous posts to match them up.
Try X166 ... Will consider the MBR thing later. It's not critical.
Sorry tp take long, had an errand, back now for the day. 166 still freezes, log: 21r17w1607.log
The two previous might have essentially been the same file. Try X167 X167-BOOTx64.efi.zip
167 same freeze same way: 21r17w4057.log
Okay ... have some things to deal with. Will take a closer look tomorrow.
Okay. The crash is at a very innocuous point and not with something that immediately jumps out as being a potential issue. Does point to memory management issues so I did some tidying up on the preceding lines in X168. X168-BOOTx64.efi.zip
Hang again but a different ending line (looks earlier than 167) in the log: 21r18q3653.log
OK. Looks like the actual issue is not actually in the function where the crash has been happening.
I have replaced the code with that from rEFInd verbatim in X169.
I expect that this will also crash in which case, I know I definitely need to look for the problem elsewhere. If it doesn't, that will actually be great as I know I just need to keep looking. I expect it will hang though. X169-BOOTx64.efi.zip
169 hangs at different place: 21r18r4604.log
Thanks ... As expected. Have to sift through to find the issue.
Please try X170 which would hopefully help with a start point X170-BOOTx64.efi.zip
170 hang looks same place, log: 21r18s2150.log
OK ... Just trying to narrow down potential point of interest.
Please give X171 a go X171-BOOTx64.efi.zip
No problem I get it. I was an IBM mainframe systems programmer for 40 years coding assembly language. Had to do this with millions of lines of code. I am checking for your response every few minutes so you don't have to wait too long for the logs. 21r18s5322.log
Cheers. Please try X172 X172-BOOTx64.efi.zip
Major movement! I booted the first linux partition and Big Sur (from Preboot) successfully. Do you want me to test booting all other partitions? Only cosmetic issue is the icon for the elementary OS partitions. Here are logs: 21r18t5408.log 21r18t5553.log
test other partitions?
Might as well to be sure.
All 4 linux and 5 macOSs boot fine!
Describe the Bug: RP only displays 4 out of the 9 bootable systems on its menu. See details below.
To Reproduce: Install RefindPlus Modify config to support APFS Boot machine
Expected Behaviour: All 9 bootable volumes should display.
Environment: RefindPlus Version : v0.13.2.AG Hardware Computer : Mac mini 2014, Macmini7,1 with internal an SSD and hard drive (originally fusion drive) Motherboard : Dual-Core Intel Core i7 GPU Model : Intel Iris Affected OS : All macOS, Opensuse Linux, Elementary OS Linux
Log Files I ran the debug version of RefindPlus and attached the RP log (level 3): 21r13t2138.log
Screenshots Here are the 4 macOS systems RP displays:
Here are the 9 systems Refind displays (requires 2 screen files since many Preboots take space):
Additional Context:
See the two attached diskutil lists for locations and volume names of the 5 macOS systems: diskutil list.txt diskutil apfs list.txt
This system has 9 bootable systems that display in refind: Linux BTRFS partition Vsuse153bum Linux BTRFS partition Vsuse153bu Linux EXT4 partition Velementary Linux EXT4 partition Open1 APFS volume FX 10.15.7 APFS volume Monterey Beta APFS volume FS 10.14.6 APFS volume FS 10.15.7 APFS volume FS 11.5.
Additional information for the 4 Linux partitions:
The missing systems are all 4 linux and APFS volume FX 10.15.7. Note they are on the HD not SSD.
I noticed the log shows “Found MBR Partition Table” which is looks wrong since it’s GPT.
Here is the config: config.conf.txt