Clover General discussion

[Slice]

10 hours ago, jsl2000 said:

That's all right for me because I can wait for your special version until you're available.

Take this version. It has no console switching

CLOVERX64.efi.zip

 

1 hour ago, kushwavez said:

How can I transform OpenCore ACPI patches to Clover?

In OC I have these (and more like these) for my Thunderbolt 3 controller on X1C6.

(all patches are here: https://github.com/tylernguyen/x1c6-hackintosh/blob/master/EFI-OpenCore/EFI/OC/config.plist, and the TB3*.amls for them are in EFI/OC/ACPI)

TableSignature: SSDT
OemTableId: 52565037 52746433
TableLength: 7453
Find: 5F505330
Replace: 58505330
Comment: TB RP09:_PS0 to RP09:XPS0
Count: 1

Find: 5F494E49
Replace: 58494E49
Comment: RP09._INI to RP09.XINI for ICM disable
Count: 1
Skip: 11

TableSignature: SSDT
OemTableId: 50726F6A 53736474
TableLength: 6103
Find: 5F555043
Replace: 58555043
Comment: XHC1: _UPC to XUPC

 

If I simply just put the Find & Replace value in Clover, Thunderbolt 3 does not work unfortunately.

First one

            <key>RenameDevices</key>
            <dict>
                <key>_SB.PCI0.RP09._PS0</key>
                <string>XPS0</string>
            </dict>

Same for others but look your DSDT for exact names/pathes.

Count and Skip are not correct because we must rename all occurence of the name.

Long path in Clover guaranties that rename will be only where needed.

[jsl2000]

9 hours ago, Slice said:

Before I make you new version, test please correction in config.plist:

GUI->

<key>ProvideConsoleGop</key>

<true/>

<key>ConsoleMode</key>

<string>0</string>

Unfortunately this new CLOVERX64.efi still got crashed on "O" selection and unable to save debug log.

Only successful booting log and new config.plist are attached here.

Is it possible related to OpenCore integration codes because older than Clover 5116 seems no such a crash.

2021-3-4_7-40-21_CLOVERX64.EFI.log config.plist

Edited March 4, 2021 by jsl2000

[Slice]

Impossible.

[jsl2000]

40 minutes ago, Slice said:

Impossible.

Impossible related to OC integration or unable to debug log ?

Thanks.

[Slice]

Impossible that it is related to opencore integration.

Can you confirm that 5116 works and 5117 no?

[jsl2000]

32 minutes ago, Slice said:

Impossible that it is related to opencore integration.

Can you confirm that 5116 works and 5117 no?

Yes, 5116 can boot 10.15.7 without such a crash but it can not boot Big Sur without integration of OpenCore.

Edited March 4, 2021 by jsl2000

[Slice]

But the crash is before system loading so doesn't matter what system you want to boot.

I want to refine my question.

Is it true that "O" in Clover GUI causes crash on version 5117 and doesn't on 5116?

13 hours ago, jsl2000 said:

Unfortunately this new CLOVERX64.efi still got crashed on "O" selection and unable to save debug log.

Only successful booting log and new config.plist are attached here.

Is it possible related to OpenCore integration codes because older than Clover 5116 seems no such a crash.

2021-3-4_7-40-21_CLOVERX64.EFI.log 59.23 kB · 3 downloads config.plist 17.95 kB · 2 downloads

There is a question. What is the crash with this my new version and why there is no debug.log. It must be!

[chris1111]

@Slice I test Legacy bootx64 on ESP on my OptiPlex 790 Intel I5 2500k / Geforce GT 710 Legacy mode Red screen here also a user notice the same on my forum on X58 P6TSE

R5129 no issue red screen

Edited March 4, 2021 by chris1111

[jsl2000]

11 hours ago, Slice said:

But the crash is before system loading so doesn't matter what system you want to boot.

I want to refine my question.

Is it true that "O" in Clover GUI causes crash on version 5117 and doesn't on 5116?

There is a question. What is the crash with this my new version and why there is no debug.log. It must be!

I have double checked and quite sure that there is no debug log produced on crashed with red screen.

No such issue at Clover 5116 which can not boot/install Big Sur yet due to lack of OpenCore integration.

Successful boot log of 10.11.6 was attached for your reference.

I found the other users (e.g. chris1111) who use legacy booting also have the same problem as mine !

debug.log

Edited March 5, 2021 by jsl2000

[jsl2000]

7 hours ago, chris1111 said:

@Slice I test Legacy bootx64 on ESP on my OptiPlex 790 Intel I5 2500k / Geforce GT 710 Legacy mode Red screen here also a user notice the same on my forum on X58 P6TSE

R5129 no issue red screen

Can you share your legacy booting Clover 5129 (CLOVERX64.efi) & config.plist without such a crash for me to test ? If the CLOVERX64.efi is release version please let me know its commit code.

My version of 5129 still got crash on booting if "O" is entered.

Edited March 5, 2021 by jsl2000

[Slice]

This debuglog has no interest because I have to see a moment of crash.

[jsl2000]

1 hour ago, Slice said:

This debuglog has no interest because I have to see a moment of crash.

Thanks.

I found Clover 5127 can boot Big Sur without such a crash on "Option" selection.

[kushwavez]

On 3/3/2021 at 7:13 PM, Slice said:

First one

            <key>RenameDevices</key>
            <dict>
                <key>_SB.PCI0.RP09._PS0</key>
                <string>XPS0</string>
            </dict>

Same for others but look your DSDT for exact names/pathes.

Count and Skip are not correct because we must rename all occurence of the name.

Long path in Clover guaranties that rename will be only where needed.

Thanks! that did the trick!

My eGPU over Thunderbolt 3 is working now finally  

 

I still need to tweak it tho, now the eGPU only showing up as an internal PCIe device, and it says Thunderbolt 3 driver not found.

This isn't a big problem, but because of that I can't select "Prefer eGPU" option on apps and I have to connect a monitor to the card for apps to utilise the card. 

[Slice]

11 hours ago, jsl2000 said:

Thanks.

I found Clover 5127 can boot Big Sur without such a crash on "Option" selection.

OK.

Test please this version. I want a log if a crash

CLOVERX64.efi.zip

[jsl2000]

3 hours ago, Slice said:

OK.

Test please this version. I want a log if a crash

CLOVERX64.efi.zip 646.87 kB · 1 download

Thank you very much for your patience and kind help.

No more crash with this 5131 version now !

Edited March 5, 2021 by jsl2000

[Slice]

Thank, now I know the problem.

[Slice]

We celebrate 10th anniversary of Clover project!

http://web.archive.org/web/20140317061109/http://www.projectosx.com/forum/lofiversion/index.php/t2008-0.html

 

[Jief_Machak]

On 2/28/2021 at 8:37 AM, Slice said:

Install OpenRuntime.efi.

 

@Jief_Machak

See. For legacy boot we lived without AptioMemoryFix and all was good.

Then we switch to OpenRuntime that does almost same functions switched on-off by Quirks.

Then we switch to Clover+OC and now legacy boot can't works without OpenRuntime. Red Screen panic. Why?

The problem seems related to ConsoleGop?

Can we reproduce this in QEMU ? 

[blackosx]

2 hours ago, Slice said:

We celebrate 10th anniversary of Clover project!

Well done Slice. Amazing really!

[Sherlocks]

We celebrate 10th anniversary of Clover project!
http://web.archive.org/web/20140317061109/http://www.projectosx.com/forum/lofiversion/index.php/t2008-0.html
 

thank you for support and hard work!!

Sent from my SM-N960N using Tapatalk

[Slice]

2 hours ago, Jief_Machak said:

Can we reproduce this in QEMU ? 

Yes, but I already fix this. 

correct use of virtual pointer

I don't know why OpenRuntime cures it.

[Jief_Machak]

I can see you replaced a memset by ZeroMem. That means there is a bug in memset, right ? So should we fix that bug instead ?

 

I also don't understand why you test if self.getCloverDirFullPath() is empty because it cannot be. Else, you'll get a panic at Self object initialization "if ( m_CloverDirFullPath.isEmpty() ) panic("m_CloverDirFullPath.isEmpty()");"

[Slice]

8 hours ago, Jief_Machak said:

I can see you replaced a memset by ZeroMem. That means there is a bug in memset, right ? So should we fix that bug instead ?

 

I also don't understand why you test if self.getCloverDirFullPath() is empty because it cannot be. Else, you'll get a panic at Self object initialization "if ( m_CloverDirFullPath.isEmpty() ) panic("m_CloverDirFullPath.isEmpty()");"

The key change is I move pointer calculation from args to local variable. 

ZeroMem was not the key. It had no influence.

[Slice]

This line

FPath = XFP.wc_str();

was the key feature that prevented the panics. 

So I must calculate the pointer before use it as arg.

Status = OcStorageInitFromFs(&mOpenCoreStorage, FileSystem, FPath, NULL);

It means that when I do

Status = OcStorageInitFromFs(&mOpenCoreStorage, FileSystem, XFP.wc_str(), NULL);

then it will crash.

When the driver OpenRuntime.efi is installed there is no crash at this place.

[Slice]

Name "Clover" started at 14.03.2011

http://web.archive.org/web/20140317060435/http://www.projectosx.com/forum/lofiversion/index.php/t2008-50.html