netkas.org forum
September 25, 2017, 02:31:19 PM *
Welcome, Guest. Please login or register.
Did you miss your activation email?

Login with username, password and session length
News: Information for registering users http://forum.netkas.org/index.php/topic,2246.0.html
 
   Home   Help Search Login Register  
Pages: [1]
  Print  
Author Topic: Apple's High Sierra Driver Bake-Off  (Read 2205 times)
Rominator
Hero Member
*****
Offline Offline

Posts: 2131



« on: June 09, 2017, 06:13:14 AM »

So, Apple threw us some curve balls in High Sierra

I dug out my RX460 and RX480 and discovered that my bar sink has a leak. The 480 had been marinating in some stinky water for 5 months.

Amazingly it still works.

Tossed them both in a 5,1 (a real one) and was surprised to discover that the 460 worked great in Sierra and the 480 worked great in High Sierra (dumb name, up there with Nvidia calling 2nd Titan Pascal the "Xp")

But 460 not liking HS and 480 not liking S (neither is patched)

Preliminary tests, I didn't try any tricks or patching. Just chucked them in there with an EFI GT630 for disk selection.

Guess I'll try some eGPU fun on a 6,1 next.
Logged

Before asking a question, check your "Personal Settings" and be sure that you have "Brain Services" set to "On".
nekton1
Jr. Member
**
Offline Offline

Posts: 59


« Reply #1 on: July 16, 2017, 11:56:49 PM »

Same here in 3,1 hacked to run Sierra and High Sierra.  RX460 works fine in Sierra but not with High Sierra (up to beta 3) and not even with a helper card (ancient HD2500 XT).
Logged
Fl0r!an
Sr. Member
****
Offline Offline

Posts: 348


« Reply #2 on: July 17, 2017, 10:15:39 AM »

Vega support seems to be there, too. Have seen a positive report @tmx86 with Frontier Edition. Might even work when flashed with Mac EFI, since the old "crash when the card has been init'ed before OS"-bug seems to be solved on Vega.
Logged
nekton1
Jr. Member
**
Offline Offline

Posts: 59


« Reply #3 on: July 19, 2017, 09:09:19 AM »

What could be stopping the RX460 in HS but not in S?
The AMD9510Controller.kext has the correct 460 ID in the right location as does the AMDRadeonX4100.kext.
I'd have expected 460 support to be native in HS as in Sierra because the late 2016 and 2017 15-inch MBPs use 460 discrete GPUs as an option.
Any ideas where to look next?
Logged
nekton1
Jr. Member
**
Offline Offline

Posts: 59


« Reply #4 on: July 19, 2017, 08:33:32 PM »

HS 10.13 beta 3 installs successfully on late 2016 MBP with Pro 460 option so the 460 drivers in the OS are correct for that model. It must be a specific problem for RX460 boards and cMP but I can't spot the cause. Grabbing at straws but could it be the Acre framebuffer?
Logged
nekton1
Jr. Member
**
Offline Offline

Posts: 59


« Reply #5 on: July 31, 2017, 05:41:37 AM »

According to Mork vom Ork, the reason why the RX460 is no longer working in 10.13 is because:

"There seems to be a new checkroutine within AMD9500Controller: it checks for the Part# of the installed AMD card.
If it finds this value for Part#: 113-4E3531U or 113-4E353BU , then it sets "Radeon RX 580" as modelname and chooses "Orinoco" (which will match perfectly for RX480!) instead of "AMD,RadeonFramebuffer" as its related framebuffer.
 
That's because Apple's eGPU kit contains Sapphire AMD RX 580 PULSE as their bundled cards."

I've looked in the AMD9500Controller but have yet to spot where this check is coded.  Anyone have a pointer?
Logged
nekton1
Jr. Member
**
Offline Offline

Posts: 59


« Reply #6 on: August 02, 2017, 10:11:49 PM »

Rominator,

Did you ever discover why HS does not like the RX460?  I've been through the kexts and framebuffers with a fine-tooth comb and cannot spot the problem.
Logged
nekton1
Jr. Member
**
Offline Offline

Posts: 59


« Reply #7 on: August 08, 2017, 06:02:14 AM »

Here is the first part of the typical WindowServer crash in the Console of 10.13 beta 4 and 5 when the RX 460 is installed—it seems to involve the AMDRadeonX4000GLdriver.  Is there any way somebody other than apple can fix whatever is wrong with the driver?  Or do I just wait for Apple to respond to the crash feedback.

Process:               WindowServer [140]
Path:                  /System/Library/PrivateFrameworks/SkyLight.framework/Versions/A/Resources/WindowServer
Identifier:            WindowServer
Version:               600.00 (306)
Code Type:             X86-64 (Native)
Parent Process:        launchd [1]
Responsible:           WindowServer [140]
User ID:               88

Date/Time:             2017-08-08 09:04:12.823 +0900
OS Version:            Mac OS X 10.13 (17A330h)
Report Version:        12
Anonymous UUID:        Deleted for privacy


Time Awake Since Boot: 33 seconds

System Integrity Protection: disabled

Crashed Thread:        0  Dispatch queue: com.apple.main-thread

Exception Type:        EXC_BAD_INSTRUCTION (SIGILL)
Exception Codes:       0x0000000000000001, 0x0000000000000000
Exception Note:        EXC_CORPSE_NOTIFY

Termination Signal:    Illegal instruction: 4
Termination Reason:    Namespace SIGNAL, Code 0x4
Terminating Process:   exc handler

Application Specific Information:
StartTime:2017-08-08 09:03:59
GPU:AMD

Thread 0 Crashed:: Dispatch queue: com.apple.main-thread
0   com.apple.AMDRadeonX4000GLDriver   0x000000010e9aaf18 glrATI_SI_InitMiscTextureInfo + 180
1   com.apple.AMDRadeonX4000GLDriver   0x000000010e9e4f5c glrATI_SI_InitHardwareContext + 268
2   com.apple.AMDRadeonX4000GLDriver   0x000000010e995625 glrInitializeContext + 26
3   libGPUSupportMercury.dylib       0x00007fff5c7dc430 gldCreateContext + 1170
4   GLEngine                         0x00007fff4e2ffd69 gliCreateContextWithShared + 750
5   com.apple.SkyLight               0x00007fff656dc8b6 slglsCreateContext + 249
6   com.apple.SkyLight               0x00007fff657be5d0 initialize_display_context + 2632
7   com.apple.SkyLight               0x00007fff657bda25 CGXInitializeGL + 18
8   com.apple.SkyLight               0x00007fff657c07b0 WSInitialize + 3170
9   com.apple.SkyLight               0x00007fff6585b825 SLXServer + 732
10  WindowServer                     0x00000001007bddde 0x1007bd000 + 3550
11  libdyld.dylib                    0x00007fff6b61817d start + 1

Thread 1:
0   libsystem_kernel.dylib           0x00007fff6b75ee76 mach_msg_trap + 10
1   libsystem_kernel.dylib           0x00007fff6b75e390 mach_msg + 60
2   com.apple.CoreDisplay            0x00007fff43babaf9 0x7fff43b0e000 + 645881
3   com.apple.CoreDisplay            0x00007fff43bacc66 0x7fff43b0e000 + 650342
4   libsystem_pthread.dylib          0x00007fff6b8a26c1 _pthread_body + 340
5   libsystem_pthread.dylib          0x00007fff6b8a256d _pthread_start + 377
6   libsystem_pthread.dylib          0x00007fff6b8a1c5d thread_start + 13

Thread 2:
0   libsystem_kernel.dylib           0x00007fff6b7686da __workq_kernreturn + 10
1   libsystem_pthread.dylib          0x00007fff6b8a206a _pthread_wqthread + 1035
2   libsystem_pthread.dylib          0x00007fff6b8a1c4d start_wqthread + 13

Thread 3:
0   libsystem_kernel.dylib           0x00007fff6b7686da __workq_kernreturn + 10
1   libsystem_pthread.dylib          0x00007fff6b8a226f _pthread_wqthread + 1552
2   libsystem_pthread.dylib          0x00007fff6b8a1c4d start_wqthread + 13

Thread 4:
0   libsystem_kernel.dylib           0x00007fff6b7686da __workq_kernreturn + 10
1   libsystem_pthread.dylib          0x00007fff6b8a206a _pthread_wqthread + 1035
2   libsystem_pthread.dylib          0x00007fff6b8a1c4d start_wqthread + 13

Thread 0 crashed with X86 Thread State (64-bit):
  rax: 0x00007f9674010960  rbx: 0x00007f967400f400  rcx: 0x0000000000000000  rdx: 0x00007f9674010960
  rdi: 0x00007f9674010948  rsi: 0x0000000000000050  rbp: 0x00007ffeef441db0  rsp: 0x00007ffeef441db0
   r8: 0x00007f96740109b0   r9: 0x0000000000000002  r10: 0x00000000fffc0000  r11: 0x00007f9670e00000
  r12: 0x00007f967680b280  r13: 0x00007f9671800600  r14: 0x00007f9671800600  r15: 0x00007f9672a002d0
  rip: 0x000000010e9aaf18  rfl: 0x0000000000010246  cr2: 0x000000010ea347e0
  
Logical CPU:     0
Error Code:      0x00000000
Trap Number:     6


Logged
nekton1
Jr. Member
**
Offline Offline

Posts: 59


« Reply #8 on: August 14, 2017, 09:36:11 PM »

Does anyone have a suggestion why the presence of the RX460 in a MacPro 5,1 with a stock Apple HD2600XT (as a helper card) prevents the machine booting with a spinner freeze about 1 cm from the right end of the boot progress bar (point where graphics driver loads)?  Without the RX460, the boot is normal.
Should the cards be in specific slots? I'd guess not but perhaps there is some voodoo....
My gut feeling is that Apple screwed the pooch with the AMDX4000driver bundle in some way.
Logged
Pages: [1]
  Print  
 
Jump to:  

Powered by MySQL Powered by PHP Powered by SMF 1.1.21 | SMF © 2015, Simple Machines
SMFAds for Free Forums
Valid XHTML 1.0! Valid CSS!