Jump to content

AMD High Sierra Kernel Release and Testing

Gigamaxx

By Gigamaxx,
in macOS 10.13 (High Sierra)

 Share
1,119 posts in this topic

Recommended Posts

bbmatias3
bbmatias3
Posted
Posted

Keep trying, this print thing is new to HS betas. I get that but if you keep trying it should get past that and load kernel.

Okay I’ll give it another try I get cpu hault then reboot I think I shouldn’t mess with prekernal. I did a copy of my working hard drive ryzen. Updates with intel machine to latest beta(worked) then disk edits made it not work

Link to comment
Share on other sites
Gigamaxx
Gigamaxx
Posted
  • Author
Posted

Okay I’ll give it another try I get cpu hault then reboot I think I shouldn’t mess with prekernal. I did a copy of my working hard drive ryzen. Updates with intel machine to latest beta(worked) then disk edits made it not work

 

Just swap out the kernel.

Link to comment
Share on other sites
Gigamaxx
Gigamaxx
Posted
  • Author
Posted

Has anyone made in progress on Vega compatibility with High Sierra? I am on 10.13.1, and anytime I run certain programs that demand graphics acceleration, my system freezes and either crashes or just stays frozen?

 

10.13.2 beta solved a lot of issues with the vega cards.

Link to comment
Share on other sites
Shaneee
Shaneee
Posted
Posted

In XLNC post-install I get a link failed error, see screenshot.

After reboot I have the gioscreenlockstate loop. Does the link error belongs to this?

Thanks in advanced.

 

 

This topic isn't for support. Please read the first post of this topic.

Link to comment
Share on other sites
bbmatias3
bbmatias3
Posted
Posted

Yes, it was fixed at 10.13.2 (17C79a) in both my FX-6300 & 1700X hacktosh with Bronya's Ryzen kernel (dated on Oct 4, 2017).

im still getting sleep wake issue. on latest beta it goes to sleep doesnt wake up im using latest lilu and what ever green 

Link to comment
Share on other sites
jalavoui
jalavoui
Posted
Posted

 

My display card is Intel internal display
but I want to know the reason of black screen.
 
Before OS X 10.11 There is no problem to enable CI
After OS X 10.12 Enabling CI causes a black screen (gIOScreenLockState 3 ... sm:0x0)
 
                     QE      CI (OPENGL/CL)
Before OS X 10.11    OK      OK
After OS X 10.12     OK      NO
If sm is 0x0 will enter the black screen.
 
sm is "systemMessage"

 

xnu Source: xnu-xxx/iokit/Kernel/IOService.cpp

 

function: IOService::updateConsoleUsers

#if HIBERNATION
	if (kIOMessageSystemHasPoweredOn == systemMessage)
	{
	    uint32_t lockState = IOHibernateWasScreenLocked();
	    switch (lockState)
	    {
		case 0:
		    break;
		case kIOScreenLockLocked:
		case kIOScreenLockFileVaultDialog:
		    gIOConsoleBooterLockState = kOSBooleanTrue;
		    break;
		case kIOScreenLockNoLock:
		    gIOConsoleBooterLockState = 0;
		    break;
		case kIOScreenLockUnlocked:
		default:
		    gIOConsoleBooterLockState = kOSBooleanFalse;
		    break;
	    }
	}
#endif /* HIBERNATION */


#if HIBERNATION
        if (!loginLocked) gIOConsoleBooterLockState = 0;
	IOLog("IOConsoleUsers: time(%d) %ld->%d, lin %d, llk %d, \n",
		(num != 0), gIOConsoleLockTime, (num ? num->unsigned32BitValue() : 0),
		gIOConsoleLoggedIn, loginLocked);
#endif /* HIBERNATION */


#if HIBERNATION
    if (gIOChosenEntry)
    {
	if (locked == kOSBooleanTrue) gIOScreenLockState = kIOScreenLockLocked;
	else if (gIOConsoleLockTime)  gIOScreenLockState = kIOScreenLockUnlocked;
	else                          gIOScreenLockState = kIOScreenLockNoLock;
	gIOChosenEntry->setProperty(kIOScreenLockStateKey, &gIOScreenLockState, sizeof(gIOScreenLockState));

	IOLog("IOConsoleUsers: gIOScreenLockState %d, hs %d, bs %d, now %ld, sm 0x%x\n",
		gIOScreenLockState, gIOHibernateState, (gIOConsoleBooterLockState != 0), now, systemMessage);
    }
#endif /* HIBERNATION */

I guess this is some sort of validation limit

 

But do not know the reason

 

 

 

this happened to me on 10.12

 

it's kind of random but very annoying

 

if u google some old versions of ioservice.cpp u can easily revert this trash code.

 

if u wanna keep this code try force the status, just change code to one of this:

 

-a

kIOScreenLockNoLock:

         gIOConsoleBooterLockState = 0;

-b

kIOScreenLockUnlocked:

         gIOConsoleBooterLockState = kOSBooleanFalse;

 

i would test option a) first ;)

Link to comment
Share on other sites
jalavoui
jalavoui
Posted
Posted

for the dam screen lock issue i've got it fixed on 10.12 by reverting code and adding a extra

gIOScreenLockState = kIOScreenLockNoLock;

 

check updateConsoleUsers()

 

this might bring some undesired side effects - try it if u wish to.

void IOService::updateConsoleUsers(OSArray * consoleUsers, IOMessage systemMessage)
{
    IORegistryEntry * regEntry;
    OSObject *        locked = kOSBooleanFalse;
    uint32_t          idx;
    bool              publish;
    OSDictionary *    user;
    static IOMessage  sSystemPower;
    
    regEntry = IORegistryEntry::getRegistryRoot();
    
    if (!gIOChosenEntry)
        gIOChosenEntry = IORegistryEntry::fromPath("/chosen", gIODTPlane);
    
    IOLockLock(gIOConsoleUsersLock);
    
    if (systemMessage)
    {
        sSystemPower = systemMessage;
#if HIBERNATION
        if ((kIOMessageSystemHasPoweredOn == systemMessage) && IOHibernateWasScreenLocked())
        {
            locked = kOSBooleanTrue;
        }
#endif /* HIBERNATION */
    }
    
    if (consoleUsers)
    {
        OSNumber * num = 0;
        gIOConsoleLoggedIn = false;
        for (idx = 0;
             (user = OSDynamicCast(OSDictionary, consoleUsers->getObject(idx)));
             idx++)
        {
            gIOConsoleLoggedIn |= ((kOSBooleanTrue == user->getObject(gIOConsoleSessionOnConsoleKey))
                                   && (kOSBooleanTrue == user->getObject(gIOConsoleSessionLoginDoneKey)));
            if (!num)
            {
                num = OSDynamicCast(OSNumber, user->getObject(gIOConsoleSessionScreenLockedTimeKey));
            }
        }
        gIOConsoleLockTime = num ? num->unsigned32BitValue() : 0;
    }
    
    if (!gIOConsoleLoggedIn
        || (kIOMessageSystemWillSleep == sSystemPower)
        || (kIOMessageSystemPagingOff == sSystemPower))
    {
        locked = kOSBooleanTrue;
    }
    else if (gIOConsoleLockTime)
    {
        clock_sec_t  now;
        clock_usec_t microsecs;
        
        clock_get_calendar_microtime(&now, &microsecs);
        if (gIOConsoleLockTime > now)
        {
            AbsoluteTime deadline;
            clock_interval_to_deadline(gIOConsoleLockTime - now, kSecondScale, &deadline);
            thread_call_enter_delayed(gIOConsoleLockCallout, deadline);
        }
        else
        {
            locked = kOSBooleanTrue;
        }
    }
    
    publish = (consoleUsers || (locked != regEntry->getProperty(gIOConsoleLockedKey)));
    if (publish)
    {
        regEntry->setProperty(gIOConsoleLockedKey, locked);
        if (consoleUsers)
        {
            regEntry->setProperty(gIOConsoleUsersKey, consoleUsers);
        }
        OSIncrementAtomic( &gIOConsoleUsersSeed );
    }
    
#if HIBERNATION
    if (gIOChosenEntry)
    {
        if (locked == kOSBooleanTrue) gIOScreenLockState = kIOScreenLockLocked;
        else if (gIOConsoleLockTime)  gIOScreenLockState = kIOScreenLockUnlocked;
        else                          gIOScreenLockState = kIOScreenLockNoLock;
        
        gIOScreenLockState = kIOScreenLockNoLock;
        gIOChosenEntry->setProperty(kIOScreenLockStateKey, &gIOScreenLockState, sizeof(gIOScreenLockState));
    }
#endif /* HIBERNATION */
    
    IOLockUnlock(gIOConsoleUsersLock);
    
    if (publish)
    {
        publishResource( gIOConsoleUsersSeedKey, gIOConsoleUsersSeedValue );
        
        MessageClientsContext context;
        
        context.service  = getServiceRoot();
        context.type     = kIOMessageConsoleSecurityChange;
        context.argument = (void *) regEntry;
        context.argSize  = 0;
        
        applyToInterestNotifiers(getServiceRoot(), gIOConsoleSecurityInterest, 
                                 &messageClientsApplier, &context );
    }
}
  • Like 1
Link to comment
Share on other sites
  • 2 weeks later...
jsl
jsl
Posted
Posted

im still getting sleep wake issue. on latest beta it goes to sleep doesnt wake up im using latest lilu and what ever green 

In my 1700X hackintosh only CLOVER r.4259 with Lilu(1.1.7), AppleALC(1.1.4), & WhateverGreen (1.0.4) all dated on 2017-08-27 can fix sleep/wakeup issue at 10.13.1 (17B48). Please try it like mine.

Link to comment
Share on other sites
XLNC
XLNC
Posted
Posted

In my 1700X hackintosh only CLOVER r.4259 with Lilu(1.1.7), AppleALC(1.1.4), & WhateverGreen (1.0.4) all dated on 2017-08-27 can fix sleep/wakeup issue at 10.13.1 (17B48). Please try it like mine.

AppleALC ?

which one are you using ?

Link to comment
Share on other sites
bbmatias3
bbmatias3
Posted
Posted

In my 1700X hackintosh only CLOVER r.4259 with Lilu(1.1.7), AppleALC(1.1.4), & WhateverGreen (1.0.4) all dated on 2017-08-27 can fix sleep/wakeup issue at 10.13.1 (17B48). Please try it like mine.

 

Okay thank you!

Sleep still not working in 10.13.2

Link to comment
Share on other sites
Gigamaxx
Gigamaxx
Posted
  • Author
Posted

In my 1700X hackintosh only CLOVER r.4259 with Lilu(1.1.7), AppleALC(1.1.4), & WhateverGreen (1.0.4) all dated on 2017-08-27 can fix sleep/wakeup issue at 10.13.1 (17B48). Please try it like mine.

Does it work in 13.2, I think that’s where the bug started. It may be a kernel issue since the 13.2 kernel is 17.3 and our kernel is 17.0.

Link to comment
Share on other sites
jsl
jsl
Posted
Posted

AppleALC ?

which one are you using ?

Post #84 at http://www.insanelymac.com/forum/topic/325987-whatevergreen-support-topic/page-5

Does it work in 13.2, I think that’s where the bug started. It may be a kernel issue since the 13.2 kernel is 17.3 and our kernel is 17.0.

I'm pretty sure that is due to GPU compatibility instead of kernel issue.

No such sleep/wake-up blank screen issue at 10.13.2 in my 1700X & FX-6300 hackintoshs with AMD 7950 & 5770/6870 GPU respectively.

But if replaced its GPU with RX-560 all got sleep/wake-up problems immediately at 10.13.2, but no issue at 10.13.1 in both AMD hackintoshs.

Okay thank you!

Sleep still not working in 10.13.2

It could be fixed by GPU replacement with compatible ones such as AMD 7950, 6870, or 5770, as well as GT-730, GTX-680, etc.

No such a sleep/wake-up issue in my hackintoshs if one of the above GPUs was used.

Link to comment
Share on other sites
cyrhex
cyrhex
Posted
Posted

In my 1700X hackintosh only CLOVER r.4259 with Lilu(1.1.7), AppleALC(1.1.4), & WhateverGreen (1.0.4) all dated on 2017-08-27 can fix sleep/wakeup issue at 10.13.1 (17B48). Please try it like mine.

Did bronya's kernel for 10.13 compatible to 10.13.2?

Link to comment
Share on other sites
Gigamaxx
Gigamaxx
Posted
  • Author
Posted

How about the sleep/resume issue?

 

In Sierra with my RX 480 sleep only worked in version 12.6 after the last Bronya kernel release for 12.6. When using the older kernel in Sierra 12.6 sleep didn’t work but it did in the previous versions 12.4, 12.5 etc. using the original older kernel(s). Sierra 12.6 had some major changes to the AMD graphics kexts so it probably needed a kernel to match for it to operate correctly. I know once that last Sierra kernel came out I was real happy with the improved performance and functionality.

 

High Sierra 13.2 beta 6 works with Bronya’s HS 17.0 kernel but no sleep wake function for my RX 480 and RX Vega 64. Nvidia gt210 sleep function works in High Sierra for me, I will test my R7 265 to see if it works. Bronya’s HS kernel was released very early so it’s based on the HS 17.0 kernel which has evolved into 17.3 now. If you use older kexts which some people do for their cards to work like Fury X there is a limit as to how far back you can go before they stop being compatible with the newer versions.

 

bbmatias3 didn’t you have sleep working in 13.1? But with the same configuration once you were in 13.2 it stopped working, is that correct?

Link to comment
Share on other sites
bbmatias3
bbmatias3
Posted
Posted

In Sierra with my RX 480 sleep only worked in version 12.6 after the last Bronya kernel release for 12.6. When using the older kernel in Sierra 12.6 sleep didn’t work but it did in the previous versions 12.4, 12.5 etc. using the original older kernel(s). Sierra 12.6 had some major changes to the AMD graphics kexts so it probably needed a kernel to match for it to operate correctly. I know once that last Sierra kernel came out I was real happy with the improved performance and functionality.

 

High Sierra 13.2 beta 6 works with Bronya’s HS 17.0 kernel but no sleep wake function for my RX 480 and RX Vega 64. Nvidia gt210 sleep function works in High Sierra for me, I will test my R7 265 to see if it works. Bronya’s HS kernel was released very early so it’s based on the HS 17.0 kernel which has evolved into 17.3 now. If you use older kexts which some people do for their cards to work like Fury X there is a limit as to how far back you can go before they stop being compatible with the newer versions.

yes the method i used no longer worked on 10.13.1

bbmatias3 didn’t you have sleep working in 13.1? But with the same configuration once you were in 13.2 it stopped working, is that correct?

yes that correct that method didnt work on 10.13.2

i dont mind staying on 10.13.1 but there is some gpu issues(performance) that are fixed on 13.2

will see im sure a solution will show up 

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