Jump to content

Clover General discussion


ErmaC
30,171 posts in this topic

Recommended Posts

Try remove clover folder or src folder.

Then try build again from clover build script.

 

I tested r4102. Build is nop

 

나의 LG-F800S 의 Tapatalk에서 보냄

 

Wait, "CloverGrowerPro" is still supported and it also has "src" folder compared to "Build_Clover.command"..?

Sorry, this one seems new to me.

  • Like 1
Link to comment
Share on other sites

Does not create one.   I'm guessing the kernel never starts.

 

This has nothing to do with it, F2 is in the clover GUI for preboot.log, debug.log is output no matter what. The log is created way before the kernel boots. I'm guessing you installed directly to the HFS+ partition? You can only write to VFAT volumes during boot, so you need to either have it on your ESP or use a USB.

  • Like 1
Link to comment
Share on other sites

This has nothing to do with it, F2 is in the clover GUI for preboot.log, debug.log is output no matter what. The log is created way before the kernel boots. I'm guessing you installed directly to the HFS+ partition? You can only write to VFAT volumes during boot, so you need to either have it on your ESP or use a USB.

 

 

Its true its on the HFS+ partition, I'll grab the SSD drive which has clover installed onto the ESP and try again.  Is the preboot.log, the debug.log you are referring to or do I need to enable the debug in clover config to generate an output log?

Link to comment
Share on other sites

This has nothing to do with it, F2 is in the clover GUI for preboot.log, debug.log is output no matter what. The log is created way before the kernel boots. I'm guessing you installed directly to the HFS+ partition? You can only write to VFAT volumes during boot, so you need to either have it on your ESP or use a USB.

 

 

Find attached both Preboot and Debug logs, Thanks!

 

I updated my signature I had swapped it for a Q8400 and forgot about that and the other Quad is as you said the Q9650 and both of these support SSE4.1.  Was the debug log helpful, I could not see any warning or error that stood out.

Debug and Preboot logs.zip

Link to comment
Share on other sites

Find attached both Preboot and Debug logs, Thanks!

 

I updated my signature I had swapped it for a Q8400 and forgot about that and the other Quad is as you said the Q9650 and both of these support SSE4.1.  Was the debug log helpful, I could not see any warning or error that stood out.

In your log

2:170  0:000  CPU Vendor = 756E6547 Model=6FB
2:170  0:000   The CPU not supported SSE4.1

  • Like 2
Link to comment
Share on other sites

@Slice @apianti

 

As you know, the recent versions of the Build_Clover.command script can add the proprietary EFI drivers (NTFS.efi, HFSPlus.efi and apfs.efi) to the package. Using such customized package however could create one potential problem: those drivers will be removed from the EFI folder, if a standard (non-customized) package is used right after the customized one.

The problem lies here: CloverPackage/package/Scripts.templates/EFIFolder/preinstall

Line 165 determines the list of the packages that should be removed, so the new version to be installed and according to that list, that should be all package identifiers, starting with "org.clover.". That means the proprietary drivers will be included in that list, because the customized package has already created package receipts for them, and therefore removed with the rest.

 

My proposition is that line to be changed from this:

pkgs=$(/usr/sbin/pkgutil --volume "$DEST_VOL" --pkgs | /usr/bin/grep -iE '@CLOVER_PACKAGE_IDENTITY@.')

to this:

pkgs=$(/usr/sbin/pkgutil --volume "$DEST_VOL" --pkgs | /usr/bin/grep -iE '@CLOVER_PACKAGE_IDENTITY@.' | /usr/bin/grep -Ev 'ntfs|apfs|hfsplus')

in order those identifiers to be excluded from the pkgs array, if they exist (already installed).

  • Like 1
Link to comment
Share on other sites

Pikeralpha wrote something disturbing in his blog: https://pikeralpha.wordpress.com/2017/07/04/do-i-use-apfs/

 

So is it better to use HFS+ even for the upcoming macOS and if so, is that possible?

 

APFS still is much better than HFS+. There's nothing disturbing or novel about his finding, those features have been under review since Apple announced the new FS last year. Unlike ZFS,  APFS just  lacks data checksum, end of the story, with or without Apple's firmware. All you have to do is to get a good and reliable data storage device, just like Apple does. That's it.

  • Like 4
Link to comment
Share on other sites

@Slice @apianti

 

As you know, the recent versions of the Build_Clover.command script can add the proprietary EFI drivers (NTFS.efi, HFSPlus.efi and apfs.efi) to the package. Using such customized package however could create one potential problem: those drivers will be removed from the EFI folder, if a standard (non-customized) package is used right after the customized one.

The problem lies here: CloverPackage/package/Scripts.templates/EFIFolder/preinstall

Line 165 determines the list of the packages that should be removed, so the new version to be installed and according to that list, that should be all package identifiers, starting with "org.clover.". That means the proprietary drivers will be included in that list, because the customized package has already created package receipts for them, and therefore removed with the rest.

 

My proposition is that line to be changed from this:

pkgs=$(/usr/sbin/pkgutil --volume "$DEST_VOL" --pkgs | /usr/bin/grep -iE '@CLOVER_PACKAGE_IDENTITY@.')

to this:

pkgs=$(/usr/sbin/pkgutil --volume "$DEST_VOL" --pkgs | /usr/bin/grep -iE '@CLOVER_PACKAGE_IDENTITY@.' | /usr/bin/grep -Ev 'ntfs|apfs|hfsplus')

in order those identifiers to be excluded from the pkgs array, if they exist (already installed).

I know nothing about this so I just accept your proposition.

  • Like 4
Link to comment
Share on other sites

Well, that makes very little sense. There are two problems I see in both, there appears to be some weird page count computation error, and looks like an allocation is being made inside the region that needs allocated. Seems that printing out the memmap when this error happens would be more useful since maybe also other stuff changed.... Damn. Anyway here's the current issue:

 

As you can see for AptioFix, the relocation block appears to be cut off by a driver, or maybe even clover itself. What's really weird is that the region it wants is completely available according to your memmap, I have a feeling that there must be another allocation here between these two not seen when your memmap was captured.

Requested mem: 5D9C000 - 17701FFF, Pages: 11966, Size: 11966000
Reloc block: 17020000 - 2C97AFFF, Pages: 1595B, Size: 1595B000
Available 000000000EDE4000-000000002C941FFF 000000000001DB5E 000000000000000F
LoaderCode 000000002C942000-000000002CADCFFF 000000000000019B 000000000000000F

Then AptioFix2,

Error allocating 0x275 pages at 0x0000000000ed28000 alloc type 2
Available  0000000000100000-000000000EDA3FFF 000000000000ECA4 000000000000000F
BS_Data    000000000EDA4000-000000000EDE3FFF 0000000000000040 000000000000000F
Available  000000000EDE4000-000000002C941FFF 000000000001DB5E 000000000000000F

So, first, I think that there needs to be a better method of determining if the allocation will fail, I think it has to do with the page computation method(s). There's something weird going on there.

Second, and I think this is probably the big thing, need to probably prevent boot time allocations from happening in lower memory where the kernel/device-tree/runtime/etc is going to try to be allocated. This can be easily achieved by hooking the memory allocation functions and modifying boot time allocations to be made above 4GB. Although I'm not sure if this is entirely possible/would not break stuff, but I also don't see why it would since all code should, in theory, be 64bit. Hmmmmmm....

 

Hey everyone, not sure if this is still relevant, but I my have stumbled over the same problem:

 

http://www.insanelymac.com/forum/topic/319764-guide-dell-xps-15-9550-sierra-10122-quick-installation/?p=2457877

I calculate my slide parameter and add it to clover config, but during boot clover tries to allocate memory in a different location OR the slide parameter now points to a different page than my calculation.

 

I cannot boot up without free200 variant, but I currently suspect this variant of AF of interfering with my iGPU (if that is possible) since it has become unstable under load, for example in FCPX or compressor using iGPU.

Link to comment
Share on other sites

Hey everyone, not sure if this is still relevant, but I my have stumbled over the same problem:

 

http://www.insanelymac.com/forum/topic/319764-guide-dell-xps-15-9550-sierra-10122-quick-installation/?p=2457877

I calculate my slide parameter and add it to clover config, but during boot clover tries to allocate memory in a different location OR the slide parameter now points to a different page than my calculation.

 

I cannot boot up without free200 variant, but I currently suspect this variant of AF of interfering with my iGPU (if that is possible) since it has become unstable under load, for example in FCPX or compressor using iGPU.

 

AptioFix changes to slide=0. Don't use any boot fix except the ones that are with clover, those other fixes literally just destroy memory to make the space needed. Did you try both AptioFix and AptioFix2? If so, please show the image of screen where it fails to boot for both. Memory maps would be good too.

Link to comment
Share on other sites

This is the memmap:
post-14795-0-56034100-1499333841_thumb.jpg

 

This is the error with slide=138 when booting with AptioFix264 on latest release of Clover:

post-14795-0-12837200-1499335916_thumb.jpg

 

Without any configured slide value it tries to allocate 0x14495 pages and fails

I had to switch to AptioFix2 due to a new BIOS for my Notebook, which apparently requires the use of v2

Link to comment
Share on other sites

Oh wait you set slide so it will use that, also that is the correct place where the slide is, slide=138 is 0x100000 + (0x200000 * 138) = 0x11500000. Then kernel is loaded address + 0x100000. I think you want slide=139. That being said you have plenty of space at slide=0. I can't tell you where on earth is the best place to allocate, I would probably just remove slide argument and let it pick.

Link to comment
Share on other sites

That is what I tried initially and didn't work, also why I decided to post this in this dev oriented thread - because I do hope I don't come across like just seeking support.

 

Here's the result with no slide argument and having Clover pick a page by itself:

post-14795-0-76817100-1499336968_thumb.jpg

 

Here's the result with slide=0 for trial:

post-14795-0-03357300-1499336984_thumb.jpg

 

Link to comment
Share on other sites

That is what I tried initially and didn't work, also why I decided to post this in this dev oriented thread - because I do hope I don't come across like just seeking support.

 

Here's the result with no slide argument and having Clover pick a page by itself:

attachicon.gifIMG_20170706_123759.jpg

 

Here's the result with slide=0 for trial:

attachicon.gifIMG_20170706_123923.jpg

Did you try dart=0 ?

Link to comment
Share on other sites

Really, I didn't use private drivers. VBosHFS.efi is enough for me. NTFS.efi is not needed (yes, bootmgr.efi located on FAT32).

apfs.efi is not ready to use.

Link to comment
Share on other sites

Yeah, dart=0 is set in my clover.config

I'll try to zip up my clover folder and upload it here

 

Nothing to do with dart=0. You don't have enough free pages at slide=0 for that big of allocation. You may not have them at this point in the boot at all below 4GB barrier. It's hard to say what is the result with AptioFix instead of AF2?

Link to comment
Share on other sites

Really, I didn't use private drivers. VBosHFS.efi is enough for me. NTFS.efi is not needed (yes, bootmgr.efi located on FAT32).

apfs.efi is not ready to use.

Yeah, you don't need them, I just thought your response was great.

Link to comment
Share on other sites

×
×
  • Create New...