Clover General discussion

[maclinuxG4]

26 minutes ago, Jief_Machak said:

Hum. Send me your config.plist. You replace your serial number by "none" if you're worried (but they're free so nobody cares). But please change in place, don't delete anything. Same for your uuids.

i send you into a message.

 

could you chek also near (when use, why ) because with R5120.pkg, there no message

[Slice]

5 hours ago, Matgen84 said:

@Jief_Machak @Slice

 

Sorry for that silly question: have you planned for Clover to be compatible with Big Sur. It's complicated, I suspect. But I think this is expected by users.

I know your efforts are great, and deserve thanks. 

Yes, we planned.

[Jief_Machak]

My last commit compile well in my main folder, but not in my copy folder. Trying to understand...

 

Compiling with Xcode toolchain but not GCC. Still trying to understand 

Edited August 26, 2020 by Jief_Machak

[Badruzeus]

11 minutes ago, Jief_Machak said:

My last commit compile well in my main folder, but not in my copy folder. Trying to understand...

SHA: 5779130ea? Seems built fine here.

[Jief_Machak]

21 minutes ago, Badruzeus said:

SHA: 5779130ea? Seems built fine here.

Yes, with XCode. Seems to be an inlining function problem for GCC.

 

ok fixed.

Edited August 26, 2020 by Jief_Machak

[Badruzeus]

13 minutes ago, Jief_Machak said:

Yes, with XCode. Seems to be an inlining function problem for GCC.

Hmmb, and how to not print or hide these "missing local header" warnings? Thanks

Edited August 26, 2020 by Badruzeus

[iCanaro]

error

 

Spoiler

 

[Jief_Machak]

1 hour ago, Badruzeus said:

Hmmb, and how to not print or hide these "missing local header" warnings? Thanks

You have to add the file in the "[Sources]" section of the .inf of the module.

For example, "Local Header: /jiefland/5.devel/clover/clover-projects/clover--cloverhackycolor--compare_settings/include/library/ocapplechunklistlib.h not found in /JiefLand/5.Devel/Clover/Clover-projects/Clover--CloverHackyColor--compare_settings/Library/OcAppleChunklistLib/OcAppleChunklistLib.inf" means that OcAppleChunklistLib.h is missing from the [Sources] section of OcAppleChunklistLib.inf.
NOTE : in the message, ocapplechunklistlib.h is lower case even though the file is not.

1 hour ago, iCanaro said:

error

Yes, that was the error before I fixed it. Did you get the commit 74a5477c26563899a29e08e25ec1d72d759d3b1d ?

[Jief_Machak]

4 hours ago, maclinuxG4 said:

could you chek also near (when use, why ) because with R5120.pkg, there no message

@maclinuxG4 Here is a new version "5121.2-SC" (don't forgot to check it when you launch). This version try to resolve the ERROR CONFIG PLIST. If it's not booting but you have the debug.log, that's ok.

Thanks for support.

[Jief_Machak]

On 8/26/2020 at 2:27 PM, netgear said:

v5121. If you use an entries for recovery, the boot is blocked, as well as the hide option does not work, while, without the Entries, it starts.
AppleIntelCPUPM always remains active even with the "false" option.

 

 

To have everything working, I have to use 5114

Ok, no problem, I'll fix all that.

Could you try my special version 5121.3-SC ? Save preboot.log (or activate debug in boot section). And check there is no ERROR CONFIG PLIST.

That shouldn't be related to the problem you describe, but at least we'll be sure that my latest refactorings are not the culprit.

 

Tell me more about what you mean :

  - "If you use an entry for recovery" : is it if you create an entry, it doesn't work even if you boot normally ?

  - Hide option : yes, I modified that. What are you trying to hide ? HFS ? APFS ? Normal partition ? Preboot ? Recovery ?

  - AppleIntelCPUPM always remains active even with the "false" option : is it always patched, even when deactivated ?

Don't hesitate to do a video of your screen. That way, I'll be sure by which menu you went etc.

CloverX64.efi

 

EDIT : I can see in sources that if 0xE2 is locked, KPAppleIntelCPUPM will be always on. Do you have this message in your log "MSR 0xE2 is locked, PM patches will be turned on" ?

Edited August 27, 2020 by Jief_Machak

[netgear]

Yes. You understood everything.

 

If I use an Entries to start Recovery, it crashes; if I start from Scan Entries it works.

 

Config perfect

debug.rar

 

Hide Recovery:

Spoiler

<dict>
                    <key>Ignore</key>
                    <false/>
                    <key>Hidden</key>
                    <true/>
                    <key>Volume</key>
                    <string>0137013727BE7E4A97094640FD0B2C51</string>
                    <key>NoCaches</key>
                    <false/>
                    <key>FullTitle</key>
                    <string>Ripristino</string>
                    <key>Disabled</key>
                    <false/>
                    <key>Path</key>
                    <string>\12868A75-A750-4993-B4B7-2EE65A598E06\boot.efi</string>
                    <key>Type</key>
                    <string>OSXRecovery</string>
                    <key>Image</key>
                    <string>os_cata</string>
                    <key>VolumeType</key>
                    <string>Internal</string>
                    <key>InjectKexts</key>
                    <true/>
                </dict>

 

 

 

5121.3-SC, same problem:

 

Boot Recovery Entries:

 

No hide Entries Recovery.

 

v5114:

[Jief_Machak]

3 hours ago, netgear said:

Config perfect

I fixed the hide custom entry.

There is "ERROR CONFIG PLIST" in your log. These are bugs I've introduced with my refactors. Let's find that first.

Boot with this attached EFI (version 5121.5-SC) and send me your debug.log

NOTE : I've dumped the config.plist in the debug.log. If you are worried about serial number or uuid, remove them from the debug.log. Send me that in a private message.

You've got "MSR 0xE2 is locked, PM patches will be turned on", so that's why AppleIntelCPUPM always remains ON.

 

Issue on custom entry that point to a recovery is not fixed. I works on my debug platform, but I've tried with HFS. I think yours is APFS. I'll have a look into that in few days (I'm away till Monday). I'll also improve so that the recovery entry is always next to the "normal" boot entry.

 

CloverX64.efi

[netgear]

5121.5-SC Entries hide Recovery OK!

 

System Catalina no Boot...

[Matgen84]

33 minutes ago, mifjpn said:

Hello.
I compiled build-e9f0f8c and tried it.
Scanning the EFI and displaying the entries went well.
But when I chose Catalina,
"__String::char32At(size_t i) :i>= length(). System halted A fatal error happened. System halted" was displayed on the screen and stopped.
Is there anything I can do?

 

Hi
Unfortunately, the only thing to do: waiting a new commit 

[Guest]

@Sliceis it planned from you and your staff to add a support for TRX40 platform?

It seem it is possible but some problem with OCquirks /Devirtualize MMIO seems do not work well here)

 

Edited August 28, 2020 by Guest

[D-an-W]

2 hours ago, mifjpn said:

Thank you,Matgen84 .

I wait.

 

 

Same happened to me too, the Clover.app installer was also corrupt.

 

In fact, you got further than me, mine failed with a CPU Halted error before the GUI was loaded.

Edited August 28, 2020 by D-an-W

[Slice]

7 hours ago, fabiosun said:

@Sliceis it planned from you and your staff to add a support for TRX40 platform?

It seem it is possible but some problem with OCquirks /Devirtualize MMIO seems do not work well here)

 

No idea what is it.

[Guest]

New Amd CPU (3960x, 3970x, 3990x Threadripper)

Edited August 28, 2020 by Guest

[Slice]

14 hours ago, mifjpn said:

Hello.
I compiled build-e9f0f8c and tried it.
Scanning the EFI and displaying the entries went well.
But when I chose Catalina,
"__String::char32At(size_t i) :i>= length(). System halted A fatal error happened. System halted" was displayed on the screen and stopped.
Is there anything I can do?

Same result booting Mojave.

Debug.log

62:325  0:074  Patching DSDT:
62:400  0:075   - [Rename oem _DSM to ZDSM]:ATTENTION : CmpDev called with a name with length() != 4
62:600  0:199  ATTENTION : CmpDev called with a name with length() != 4
62:725  0:125  ATTENTION : CmpDev called with a name with length() != 4
62:800  0:074  ATTENTION : CmpDev called with a name with length() != 4
62:891  0:091  ATTENTION : CmpDev called with a name with length() != 4
62:966  0:075  ATTENTION : CmpDev called with a name with length() != 4

and so on, infinite loop.

 

[Slice]

CmpDev called with empty name.

And this ATTENTION... set messages is finite. Ended by

46:477  0:000  ATTENTION : CmpDev called with a  with length() != 4 0
46:478  0:000   - [Rename my XDSM to _DSM]: pattern 5844534D, bin not found / already patched!
46:481  0:003   - [Rename GFX0 to IGPU]:

followed by screen message __String::char32At(size_t i) :i>= length(). 

[Slice]

This condition

          if ( gSettings.DSDTPatchArray[i].PatchDsdtTgt.notEmpty() ) {
            Len = FixAny((UINT8*)NewTable, Len,
                         gSettings.DSDTPatchArray[i].PatchDsdtFind,
                         gSettings.DSDTPatchArray[i].PatchDsdtReplace);
            //DBG(" OK\n");
          }else{
            //DBG("Patching: renaming in bridge\n");
            Len = FixRenameByBridge2((UINT8*)NewTable, Len, gSettings.DSDTPatchArray[i].PatchDsdtTgt, gSettings.DSDTPatchArray[i].PatchDsdtFind, gSettings.DSDTPatchArray[i].PatchDsdtReplace);
          }

Means "else" will be executed when  PatchDsdtTgt isEmpty. This is not the logic being before.

[Slice]

I think logic will be if (...isEmpty())

[Slice]

New problem

It happen at kext/Info.plist parse for CFBundleVersion. 

[Slice]

It appears at parsing <data></data> which is possible value.

[Slice]

Committed.

 

Next problem: stop after

0:710  0:001  Hiding entries with string Preboot

next line should be

0:744  0:002  === [ LoadDrivers ] =============================

but absent.