Clover General discussion

[vector sigma]

1 hour ago, zkingtut said:

@vector sigma

Please my laptop keyboard and trackpad stopped to work once I've updated clover from 5018 to 5033 

What would be the reason ? If I compare between the two EFI are exactly the same drivers and kexts 

So I rolled back to 5018 

Thank you in advance 

 

Between 5018 and 5033 there are a lot of commits regarding translations (which does not affect Clover any how) and other minor fixes. What can affect keyboards/trackpad maybe r5028, but not sure.

[joevt]

1 hour ago, vector sigma said:

so why the root of the target volume is a perfect place....

Ok, but the selected target volume is only part of the install. The actual target is the EFI partition on the same disk as the target, correct? Backups are not put in EFI because space is limited in the EFI partition. The same could be said for partitions on a USB stick - in that case a backup location related to the current OS might be preferable.

 

1 hour ago, vector sigma said:

Additionally you will don't know if your backup is the backup of your internal HD/SSD or your pen drive (in case you want update Clover on a usb stick)

A backup could contain a text log file that describes the installation options, the version of clover to be installed, the target, source of the backup, and destination EFI (I suppose the last two are the same thing). "diskutil list" for the entire setup, "diskutil info /dev/disk?" for the affected disks, "diskutil info /dev/disk?s?" for the EFI partition, the target, and current OS if it differs). Actually, this log file could be a copy of the new "Clover_Install_Log.txt" file that is placed in the EFI partition.

 

2 hours ago, vector sigma said:

The package also actually takes care of not making a lot of of backups and only takes latest

I have multiple macOS partitions (different versions of macOS). Each has a EFI-Backups folder. I have 8 backup folders in the EFI-Backups folder on the root of my HighSierra disk. The folders are named after the version number (r4077 to r4910; I am currently running r4945). Each version folder has one or more date/time folders (date/time of backup). Each date/time folder has an EFI folder. The new log file would go in the date/time folder, outside the EFI folder. The EFI folder contains an old log file for when that EFI folder's Clover was last installed. Hard drive space is cheap, so I don't mind having all these folders (nearly 1000 MB on the HighSierra partition) and I can delete old backups manually.

[Slice]

5 hours ago, zkingtut said:

@vector sigma

Please my laptop keyboard and trackpad stopped to work once I've updated clover from 5018 to 5033 

What would be the reason ? If I compare between the two EFI are exactly the same drivers and kexts 

So I rolled back to 5018 

Thank you in advance 

 

Please provide preboot.log. I see no reason Clover 5018 will be better then 5033.

[STLVNUB]

I'm running 5022 at the moment, when I upgrade to 5033 my computer doesn't boot, just sits at Apple logo, no progress bar.

 edit: fixed, doesn't like the AudioDxe.efi

Edited August 5, 2019 by STLVNUB

[vector sigma]

12 hours ago, zkingtut said:

Your guessing is right r5028 start problem .... but what is the reason or which driver affect that behavior ? 

In short, try to disable PatchAPIC in config or at Clover's GUI at boot...

[vector sigma]

12 hours ago, joevt said:

Ok, but the selected target volume is only part of the install. The actual target is the EFI partition on the same disk as the target, correct? Backups are not put in EFI because space is limited in the EFI partition. The same could be said for partitions on a USB stick - in that case a backup location related to the current OS might be preferable.

 

A backup could contain a text log file that describes the installation options, the version of clover to be installed, the target, source of the backup, and destination EFI (I suppose the last two are the same thing). "diskutil list" for the entire setup, "diskutil info /dev/disk?" for the affected disks, "diskutil info /dev/disk?s?" for the EFI partition, the target, and current OS if it differs). Actually, this log file could be a copy of the new "Clover_Install_Log.txt" file that is placed in the EFI partition.

 

I have multiple macOS partitions (different versions of macOS). Each has a EFI-Backups folder. I have 8 backup folders in the EFI-Backups folder on the root of my HighSierra disk. The folders are named after the version number (r4077 to r4910; I am currently running r4945). Each version folder has one or more date/time folders (date/time of backup). Each date/time folder has an EFI folder. The new log file would go in the date/time folder, outside the EFI folder. The EFI folder contains an old log file for when that EFI folder's Clover was last installed. Hard drive space is cheap, so I don't mind having all these folders (nearly 1000 MB on the HighSierra partition) and I can delete old backups manually.

By your words I can only understand that you agreed with me? (only one backup directory at the root of the target Volume which its ESP always belong to... i.e. just target the same volume)

[chris1111]

I have never seen any other application or program in macOS that creates a bacup on the root disk; including Apple

Edited August 5, 2019 by chris1111

[Matgen84]

Hi @vector sigma @chris1111

 

Just my opinion: EFI-Backups folder seems to be in root of my target volume. Maybe I made a mistake in my comment.

 

 

[Dajjal]

On my Z77 system anything past 4988 I get no output via hdmi on my Sapphire Radeon Nitro+ RX 590.I have no display port so that's not an option.I can however Remote Desktop and see that everything is working.Can someone point out the obvious that I'm missing plz?

[vector sigma]

7 hours ago, Matgen84 said:

Just my opinion: EFI-Backups folder seems to be in root of my target volume. Maybe I made a mistake in my comment.

just as expected 

[joevt]

9 hours ago, vector sigma said:

By your words I can only understand that you agreed with me? (only one backup directory at the root of the target Volume which its ESP always belong to... i.e. just target the same volume)

I think so. I like the way it has worked for me up to now. Create the version folder in the EFI-Backups folder, create the date/time folder, create the backup, don't delete old backups, etc.

 

My one suggestion is to have a copy of the install log in the date/time folder containing the backup and to add any useful missing info to the install log. But maybe those changes aren't useful if the backup is always on the same disk as the EFI partition.

 

I don't care too much about where the EFI-Backups folder is. I take back my reservation about placing backups on a USB stick - If a USB stick is large enough to contain a macOS installer then it should be large enough to contain a few backups.

 

For Catalina, I don't like that you have to make it writable to do the backup. Does it make it readonly again after the backup is finished?

 

[vector sigma]

51 minutes ago, joevt said:

For Catalina, I don't like that you have to make it writable to do the backup. Does it make it readonly again after the backup is finished?

Yeah, this is a good point, I'll try to secure the root file system some how depending if was or wasn't read only before the installation, I think I can.. and so no, to answer your question, actually the change is persistent only until the next reboot (not a permanent change to the end).

An easy and easy solution is to make the installer ask for a reboot...  achievable by just put the following in the Distribution file: 

'onConclusion="RequireRestart"'

the above will make a prominent alert window which prevent Installer.app to not exit until you accept to reboot

Anyway I found singular the story of 'killal Finder' to fix a thing that is not a bug but a declared safety improvement, until now... with your remark, which make me feel better Lol.

 

59 minutes ago, joevt said:

I don't care too much about where the EFI-Backups folder is. I take back my reservation about placing backups on a USB stick - If a USB stick is large enough to contain a macOS installer then it should be large enough to contain a few backups.

No one apparently knows we can decide the backup limit in nvram:

Clover.KeepBackupLimit = 1 (or another number... even 0 which mean no backup)

(10 by default that maybe is too high.. 2 is perfect imho)

and/or backup directly into the *ESP:

Clover.BackupDirOnDestVol = false # (ESP), true (target volume)

so I don't even know why we ended up in this fancy discussion which I guess was about  "default behaviour", I hope. The original creator of this package was (is) a very smart person...

 

good night

 

* this is intended if you selected the ESP in the package.. otherwise will be the target volume anyway

[LockDown]

@vector sigma

 

buildme:

buildme Beta, Clover v2.5k r5035
EDKII in use: vUDK2018
Patches in use: Patches_for_UDK2018

1) build Clover		   5) update Clover	     9) show diff
2) make pkg		   6) update edk2	    10) quit
3) make iso		   7) status
4) build all		   8) check remote changes
Please enter your choice: 

does 4) build all also update clover & edk2 on the fly?

[Matgen84]

2 hours ago, ellaosx said:

@vector sigma

 

buildme:

buildme Beta, Clover v2.5k r5035
EDKII in use: vUDK2018
Patches in use: Patches_for_UDK2018

1) build Clover		   5) update Clover	     9) show diff
2) make pkg		   6) update edk2	    10) quit
3) make iso		   7) status
4) build all		   8) check remote changes
Please enter your choice: 

does 4) build all also update clover & edk2 on the fly?

 

The answer is the script itself.

 

Build all= Build Clover, Build PKG, Build Iso

 

 

Edited August 6, 2019 by Matgen84

[Slice]

2 hours ago, ellaosx said:

@vector sigma

 

buildme:

buildme Beta, Clover v2.5k r5035
EDKII in use: vUDK2018
Patches in use: Patches_for_UDK2018

1) build Clover		   5) update Clover	     9) show diff
2) make pkg		   6) update edk2	    10) quit
3) make iso		   7) status
4) build all		   8) check remote changes
Please enter your choice: 

does 4) build all also update clover & edk2 on the fly?

If UDK2018 is used then no update for EDK2.

[vector sigma]

5 hours ago, Slice said:

If UDK2018 is used then no update for EDK2.

yes infact is always the same vUDK2018, maybe 'renew edk2' would be better.

Slice the script is ready for the stable release 2019:

declare -r PATCHES=Patches_for_UDK2018

if [[ $PATCHES == Patches_for_UDK2018 ]]; then
  declare -r edk2Rel=vUDK2018
  edk2Link=https://codeload.github.com/tianocore/edk2/zip/vUDK2018 # UDK2018
else
  declare -r edk2Rel=edk2-stable201905
  edk2Link=https://github.com/tianocore/edk2/archive/${edk2Rel}.zip # edk2-stable201905
fi

any plan to update Patches_for_EDK2? Looks like edk2-stable201905 is almost compatible, only a reference to a removed pkg need to be changed (which now I forgot).

[vector sigma]

6 hours ago, Matgen84 said:

The answer is the script itself.

 

Build all= Build Clover, Build PKG, Build Iso

8 hours ago, ellaosx said:

does 4) build all also update clover & edk2 on the fly?

Only builds and no update of any kind. buildme is done to work with stable releases of edk2 so once you got it there's no need to update or search for update of it. Just update Clover when a need be (you can see remote changes) and your are done! the option to update edk2 serves only if you are playing with Patches_for_UDK2018/EDK2 and you want a clean workspace, otherwise you can ignore it (until Patches_for_* get an update). I made this wrapper for my self but I hope is helpfull for you all.

Edited August 6, 2019 by vector sigma

[onemanOSX]

On 8/5/2019 at 11:51 AM, Slice said:

Please provide preboot.log

I am having similar unbootable issue with r5033 (r5018 is fine). Attached are both logs from both versions.

 

Thanks.

 

r5033 - preboot.log

 

r5018 - preboot.log.zip

Edited August 6, 2019 by onemanOSX

[bronxteck]

how are we supposed to implement patch APIC after commit 5028. most people that have it enabled in config are having issues booting to a KP. and dont know why the machine that always booted fine now suddenly KP's at boot after rev 5028 and higher. is it a bug or a feature? 

for the guys above try disabling patch APIC if you have it enabled and see if now it works.

Edited August 6, 2019 by bronxteck

[Badruzeus]

57 minutes ago, bronxteck said:

how are we supposed to implement patch APIC after commit 5028. most people that have it enabled in config are having issues booting to a KP. and dont know why the machine that always booted fine now suddenly KP's at boot after rev 5028 and higher. is it a bug or a feature? 

My problem after r5028++ was machine with 2nd Gen CPU unable to HALT (reboot-shutdown); already tested leaving only mandatory *.efi on \drivers\UEFI (DataHub, FSInject, OsxAptio*, VBoxHfs, SMCHelper), and just minumum 3rd party kexts (HWSensor's FakeSMC v3.5.x & PS2Controller). Also tried clear NVRAM several times with F11 key, as well.. using Clover's on the fly ACPI Patches instead of my custom DSDT/SSDT, booting 10.13.6 upto 10.15 Beta.. etc.

BUT all didn't solve the issue, TILL I uncommited r5028 AcpiPatcher.c (rolling back to r5027) and now my mach is back to normal again.

 

#EDIT: Thanks to Florin9doi, anyway.

APIC_unPatch.diff.zip

r5036_APIC_unPatched.zip

Edited August 6, 2019 by Badruzeus

[onemanOSX]

26 minutes ago, Badruzeus said:

My problem after r5028++ was machine with 2nd Gen CPU unable to HALT (reboot-shutdown); already tested leaving only mandatory *.efi on \drivers\UEFI (DataHub, FSInject, OsxAptio*, VBoxHfs, SMCHelper), and just minumum 3rd party kexts (HWSensor's FakeSMC v3.5.x & PS2Controller). Also tried clear NVRAM several times with F11 key, as well.. using Clover's on the fly ACPI Patches instead of my custom DSDT/SSDT, booting 10.13.6 upto 10.15 Beta.. etc.

BUT all didn't solve the issue, TILL I uncommited r5028 AcpiPatcher.c (rolling back to r5027) and now my mach is back to normal again.

 

#EDIT: Thanks to Florin9doi, anyway.

APIC_unPatch.diff.zip

r5036_APIC_unPatched.zip

This solved for my build as well! Thank you @Badruzeus and @Florin9doi

[Slice]

1 hour ago, Badruzeus said:

My problem after r5028++ was machine with 2nd Gen CPU unable to HALT (reboot-shutdown); already tested leaving only mandatory *.efi on \drivers\UEFI (DataHub, FSInject, OsxAptio*, VBoxHfs, SMCHelper), and just minumum 3rd party kexts (HWSensor's FakeSMC v3.5.x & PS2Controller). Also tried clear NVRAM several times with F11 key, as well.. using Clover's on the fly ACPI Patches instead of my custom DSDT/SSDT, booting 10.13.6 upto 10.15 Beta.. etc.

BUT all didn't solve the issue, TILL I uncommited r5028 AcpiPatcher.c (rolling back to r5027) and now my mach is back to normal again.

 

#EDIT: Thanks to Florin9doi, anyway.

APIC_unPatch.diff.zip

r5036_APIC_unPatched.zip

Is it the same result if just set PatchAPIC=NO?

 

[Badruzeus]

35 minutes ago, Slice said:

Is it the same result if just set PatchAPIC=NO?

Sorry, is this an option to add on config.plist, or..?

[vector sigma]

Just now, Badruzeus said:

Sorry, is this an option to add on config.plist, or..?

Yes, and you can do that at boot time as well in the GUI

[Badruzeus]

1 minute ago, vector sigma said:

Yes, and you can do that at boot time as well in the GUI

Thanks. So, I need re-using r5028++ AcpiPatcher.c then boot with this `PatchAPIC=NO` right?

Will test it ASAP.