Jump to content

Clover Problems and Solutions

ErmaC

By ErmaC,
in Clover Development

 Share
3,206 posts in this topic

Recommended Posts

Sherlocks
Sherlocks
Posted
Posted

well that explains why you have some failures.

 

maybe mount or unmount issue in my case when reboot or terminal. "Resource busy -- try 'diskutil unmount'"

 

i did think ESP index problem like you said. but after test many times, exist fail case. its not good. 

 

FrodoKenny script always success make nvram by using -f. 100% success nvram file.

 

you said "i was not happy with a force unmount of ESP - better to leave it mounted." before.

 

but your script has fail case.

 

100% or has fail case. what is best?

 

i discussed this with syscl. and found better solution. not sure. but syscl and me have good result in test script.

 

sorry my english

 

thank you

Link to comment
Share on other sites
Sherlocks
Sherlocks
Posted
Posted

ok - i can i see this latest script? and better solution?

 

what is the fail case? do you mean failure to unmount? 

 

it's made by syscl. check. not sure include that you want case. it needs to confirm work.  

 

this script tested for week. and 100% success like FrodoKenny script when terminal and reboot.

@Slice

 

https://sourceforge.net/p/cloverefiboot/code/3978/tree//rEFIt_UEFI/Platform/kext_inject.c?diff=504c3833fd48f873331ebd05:3977

 

i want to tell r3978 commit

 

// Sierra debug kernel

UINT8   KBESieDebugSearchEXT[]   = { 0xE8, 0x47, 0x00, 0x00, 0x00, 0xE9, 0x09, 0x00, 0x00, 0x00, 0x48, 0x8B, 0x7D, 0xE8, 0xE8, 0xD9 };

UINT8   KBESieDebugReplaceEXT[]  = { 0xE8, 0x47, 0x00, 0x00, 0x00, 0x90, 0x90, 0x00, 0x00, 0x00, 0x48, 0x8B, 0x7D, 0xE8, 0xE8, 0xD9 };

UINT8   KBESieDebugSearchSIP[]   = { 0x31, 0xC9, 0x39, 0xC1, 0x0F, 0x85, 0x3C, 0x00, 0x00, 0x00, 0x48, 0x8B, 0x85, 0xF8, 0xFE, 0xFF };

UINT8   KBESieDebugReplaceSIP[]  = { 0x31, 0xC9, 0x39, 0xC1, 0xEB, 0x80, 0x90, 0x90, 0x90, 0x90, 0x48, 0x8B, 0x85, 0xF8, 0xFE, 0xFF };

 

// Sierra develop kernel

UINT8   KBESieDevelSearchEXT[]   = { 0xE8, 0x47, 0x00, 0x00, 0x00, 0xE9, 0x09, 0x00, 0x00, 0x00, 0x48, 0x8B, 0x7D, 0xE8, 0xE8, 0xD9 };

UINT8   KBESieDevelReplaceEXT[]  = { 0xE8, 0x47, 0x00, 0x00, 0x00, 0x90, 0x90, 0x00, 0x00, 0x00, 0x48, 0x8B, 0x7D, 0xE8, 0xE8, 0xD9 };

UINT8   KBESieDevelSearchSIP[]   = { 0x31, 0xC9, 0x39, 0xC1, 0x0F, 0x85, 0x3C, 0x00, 0x00, 0x00, 0x48, 0x8B, 0x85, 0xF8, 0xFE, 0xFF };

UINT8   KBESieDevelReplaceSIP[]  = { 0x31, 0xC9, 0x39, 0xC1, 0xEB, 0x80, 0x90, 0x90, 0x90, 0x90, 0x48, 0x8B, 0x85, 0xF8, 0xFE, 0xFF };

 

debug kernel patch and develop kernel patch is same value. no need both. only has need debug kernel or develop kernel

 

reason.

 

  else if (NumSieDebug == 1) {

      Num = SearchAndReplace(Kernel, KERNEL_MAX_SIZE, KBESieDebugSearchEXT, sizeof(KBESieDebugSearchEXT), KBESieDebugReplaceEXT, 1) +

            SearchAndReplace(Kernel, KERNEL_MAX_SIZE, KBESieDebugSearchSIP, sizeof(KBESieDebugSearchSIP), KBESieDebugReplaceSIP, 1);

      DBG_RT(Entry, "==> kernel Sierra Debug: %d replaces done.\n", Num);

  }

  else if (NumSieDevel == 1) {

      Num = SearchAndReplace(Kernel, KERNEL_MAX_SIZE, KBESieDevelSearchEXT, sizeof(KBESieDevelSearchEXT), KBESieDevelReplaceEXT, 1) +

            SearchAndReplace(Kernel, KERNEL_MAX_SIZE, KBESieDevelSearchSIP, sizeof(KBESieDevelSearchSIP), KBESieDevelReplaceSIP, 1);

      DBG_RT(Entry, "==> kernel Sierra Development: %d replaces done.\n", Num);

  }

 

clover alway enter "  else if (NumSieDebug == 1) {" if we use sierra.

never enter  " else if (NumSieDevel == 1) {". we don't need "  else if (NumSieDevel == 1) {"

 

because it's same hex value like i said above.

 

as result. we have to remain debug or develop kernel hex code. it's just name difference. choose debug or develop. i don't know better name.

 

thank you.

Archive (1).zip

  • Like 1
Link to comment
Share on other sites
tluck
tluck
Posted
Posted

well my script and syscl script is very similar.

 

2 things i changed

 

1) rm -r of the mountpoint is very dangerous

no need to remove the mount point

and if you want to that better to use rmdir  

- so you dont remove the entire folder if it does not get mounted

2) unmount -f  is not ideal - can leave the filesystem in corrupt mode.

Link to comment
Share on other sites
Sherlocks
Sherlocks
Posted
Posted

well my script and syscl script is very similar.

 

2 things i changed

 

1) rm -r of the mountpoint is very dangerous

no need to remove the mount point

and if you want to that better to use rmdir  

- so you dont remove the entire folder if it does not get mounted

2) unmount -f  is not ideal - can leave the filesystem in corrupt mode.

 

okay. i will discuss this with syscl

 

i want to work script that make nvram 100% all cases. if you can, give me improved file.

 

thank you

Link to comment
Share on other sites
tluck
tluck
Posted
Posted

compare closely my 1.6 and syscl v1.6

 

removed some minor unnecessary calls and streamlined a few things

 

no need read nvram multiple times - stuffed in variable

 

also added logic to write to write to / of OS only if no separate ESP partition

 

added check for:

if [[ "${gEmuVariableLegacy}" == "true" || "${gEmuVariablePresent}" != "" ]];

 

fixed issues with legacy _toLower Case etc 

 

my logoutHook is CloverDaemon-stopservice

 

see attached for 4 files - replaced

nvram-save.zip

Link to comment
Share on other sites
Sherlocks
Sherlocks
Posted
Posted

compare closely my 1.6 and syscl v1.6

 

removed some minor unnecessary calls and streamlined a few things

 

no need read nvram multiple times - stuffed in variable

 

also added logic to write to write to / of OS only if no separate ESP partition

 

added check for:

if [[ "${gEmuVariableLegacy}" == "true" || "${gEmuVariablePresent}" != "" ]];

 

fixed issues with legacy _toLower Case etc 

 

my logoutHook is CloverDaemon-stopservice

 

see attached for 4 files - replaced

 

confirmed

 

Last login: Tue Jan 24 10:53:35 on console

Supreme-MBP:~ supreme$ sudo /etc/rc.shutdown.d/80.save_nvram_plist.local

Password:

umount(/Volumes/ESP-01): Resource busy -- try 'diskutil unmount'

Supreme-MBP:~ supreme$ sudo cat /Library/Logs/CloverEFI/rc.shutdown.log 

v1.6 © 2017 syscl/lighting/Yating Zhou

Found ESP on disk0s1 and found Clover in /Volumes/ESP-01/EFI/CLOVER

Target path: /Volumes/ESP-01 saved new values to NVRAM

Supreme-MBP:~ supreme$ 

 

i didn't change brightness or any nvram value. nvram always refresh when reboot or terminal in your script.

 

syscl script reflect smart nvram value if change ex. brightness, nvram refresh. but if don't change brightness, nvram not refresh.

 

in your script, although shown "Resource busy -- try 'diskutil unmount'", nvram file is refresh.

 

maybe need to test more times. i don't know that touch CloverDaemon service better.

 

thank you

Link to comment
Share on other sites
tluck
tluck
Posted
Posted

" in your script, although shown "Resource busy -- try 'diskutil unmount'", nvram file is refresh."

 

this message is not a real problem - just means umount failed to unmount the filesystem - as i prefer not to force unmount if busy

 

yes it will write the nvram.plist only if the file contents changed from the last version on the filesystem. same logic.

it checks the entire nvram.plist for changes - not just brightness

 

run the script 2 or 3 times from command line to see if the file changes

$ cat /Library/Logs/CloverEFI/rc.shutdown.log
v1.6 (c) 2017 syscl/lighting/Yating Zhou
Found ESP on disk0s1 and found Clover in /Volumes/ESP-01/EFI/CLOVER
Found ESP on disk1s1 and found Clover in /Volumes/ESP-11/EFI/CLOVER
Found ESP on disk2s1
Target path: /Volumes/ESP-01 no change in NVRAM
Target path: /Volumes/ESP-11 no change in NVRAM
Link to comment
Share on other sites
Sherlocks
Sherlocks
Posted
Posted

" in your script, although shown "Resource busy -- try 'diskutil unmount'", nvram file is refresh."

 

this message is not a real problem - just means umount failed to unmount the filesystem - as i prefer not to force unmount if busy

 

yes it will write the nvram.plist only if the file contents changed from the last version on the filesystem. same logic.

it checks the entire nvram.plist for changes - not just brightness

 

run the script 2 or 3 times from command line to see if the file changes

$ cat /Library/Logs/CloverEFI/rc.shutdown.log
v1.6 (c) 2017 syscl/lighting/Yating Zhou
Found ESP on disk0s1 and found Clover in /Volumes/ESP-01/EFI/CLOVER
Found ESP on disk1s1 and found Clover in /Volumes/ESP-11/EFI/CLOVER
Found ESP on disk2s1
Target path: /Volumes/ESP-01 no change in NVRAM
Target path: /Volumes/ESP-11 no change in NVRAM
Okay. I want to revert CloverDaemon orig. Can i use it?

 

나의 LG-F410S 의 Tapatalk에서 보냄

Link to comment
Share on other sites
tluck
tluck
Posted
Posted

well sure.

CloverDeamon (orig) just waits in sleep loop to run the /etc/rc.shutdown.d/*.local scripts upon shutdown - when it is killed it runs StopService part of the daemon.

so it is redundant with logoutHook (CloverDaemon-stopservice)

my approach is not to worry about whether these shutdown scripts will complete - but run via logoutHook earlier.

Link to comment
Share on other sites
PMheart
PMheart
Posted
Posted

it's made by syscl. check. not sure include that you want case. it needs to confirm work.

 

this script tested for week. and 100% success like FrodoKenny script when terminal and reboot.

 

@Slice

 

https://sourceforge.net/p/cloverefiboot/code/3978/tree//rEFIt_UEFI/Platform/kext_inject.c?diff=504c3833fd48f873331ebd05:3977

 

i want to tell r3978 commit

 

debug kernel patch and develop kernel patch is same value. no need both. only has need debug kernel or develop kernel

 

reason.

 

clover alway enter " else if (NumSieDebug == 1) {" if we use sierra.

never enter " else if (NumSieDevel == 1) {". we don't need " else if (NumSieDevel == 1) {"

 

because it's same hex value like i said above.

 

as result. we have to remain debug or develop kernel hex code. it's just name difference. choose debug or develop. i don't know better name.

 

thank you.

Hi. Sherlocks.

Seems the new patch hex code don't exist. Both the two patches.

Check http://www.insanelymac.com/forum/topic/284656-clover-general-discussion/page-609#entry235588

Link to comment
Share on other sites
Sherlocks
Sherlocks
Posted
Posted

Hi. Sherlocks.

Seems the new patch hex code don't exist. Both the two patches.

Hello.

In kernel? I didnt add this hex. Slice added it. I just report wrong code. I didnt check hex code in kernel.

 

나의 LG-F410S 의 Tapatalk에서 보냄

Link to comment
Share on other sites
PMheart
PMheart
Posted
Posted

Hello.

In kernel? I didnt add this hex. Slice added it. I just report wrong code. I didnt check hex code in kernel.

 

나의 LG-F410S 의 Tapatalk에서 보냄

Yes. In kernel.

Have some looks and I'm almost sure the new patches may not work properly cuz no data can be found.

  • Like 1
Link to comment
Share on other sites
Sherlocks
Sherlocks
Posted
Posted

Yes. In kernel.

Have some looks and I'm almost sure the new patches may not work properly cuz no data can be found.

 

you're right. i checked hex code in kernel in 10.12.3 official.

debug EXT and SIP hex code are not exist. i tried to find detail hex code. but not exist similar code for disable any features in kernel.

 

as result, debug sierra kernel patch is not working now.

 

or other new macos sierra for developer? cool information

 

thank you:)

  • Like 1
Link to comment
Share on other sites
PMheart
PMheart
Posted
Posted

you're right. i checked hex code in kernel in 10.12.3 official.

debug EXT and SIP hex code are not exist. i tried to find detail hex code. but not exist similar code for disable any features in kernel.

 

as result, debug sierra kernel patch is not working now.

 

or other new macos sierra for developer? cool information

 

thank you:)

Yes. Both of the new patches will not work. And no similar code too. So just comment them out.  :wink_anim:

I think bypassing the loading path check and SIP is enough. Nothing more we should do.

 

BTW. Normally how did you check a similar code based on the existing patch quickly? I wanna know. Thanks. :)

Link to comment
Share on other sites
Sherlocks
Sherlocks
Posted
Posted

@Slice

 

i have boot problem. my laptop can't never boot. i tested r3984. i think edk2 issue when building clover. edk2 Checked out revision 23704.

 

shown shortly

osxaptiofix

+++++++++++++++++++++++++++++++++++++

 

and force reboot. i never see boot process.

 

here is preboot.log

Link to comment
Share on other sites
Slice
Slice
Posted
Posted

Since 23702, clover has problem? I will check work or not work in detail from any code.

 

나의 LG-F410S 의 Tapatalk에서 보냄

Yes, up to KP after OS started. Nothing interesting in boot.log.

Anyway, how many DSDT/kext/kernel patches do you use? Some of them crashed with EDK2 >= 23702

@Slice

 

i have boot problem. my laptop can't never boot. i tested r3984. i think edk2 issue when building clover. edk2 Checked out revision 23704.

 

shown shortly

osxaptiofix

+++++++++++++++++++++++++++++++++++++

 

and force reboot. i never see boot process.

 

here is preboot.log

What computer?

Link to comment
Share on other sites
Sherlocks
Sherlocks
Posted
Posted

Yes, up to KP after OS started. Nothing interesting in boot.log.

Anyway, how many DSDT/kext/kernel patches do you use? Some of them crashed with EDK2 >= 23702

 

What computer?

Good news. I found something.

I will post here.

 

나의 LG-F410S 의 Tapatalk에서 보냄

Link to comment
Share on other sites
 Share
Go to topic listing
×
×
  • Create New...