Jump to content

Mavericks kernel testing on AMD (formerly Mountain Lion kernel testing on AMD)

theconnactic

By theconnactic,
in OS X 10.9 (Mavericks)

 Share
6,414 posts in this topic

Recommended Posts

wjywbs
wjywbs
Posted
Posted

I found the reason of flooding OPEMU errors. I add a panic and halt the machine when it detects the first error. Then I found that IOATAFamily crashed in the stack trace. I used a patched IOATAFamily. I restored the original one and the error disappeared,

 

@sinetek: Should we change the handling of OPEMU errors? I'm thinking about adding saved_state->isf.rip += bytes_skip; before return 0. It keeps running the current opcode. Or is it better to exit the process that caused problems?

opemu.c, in opemu_ktrap
bad:
	{
		/* Well, now go in and get the asm text at least.. */
		const char *instruction_asm;
		instruction_asm = ud_insn_asm(&ud_obj);

                // This change is for debug only, not for production.
		printf ( "kOPEMU, %d:  %s\n", bytes_skip, instruction_asm) ;
		panic();
		pmCPUHalt(PM_HALT_PANIC);
		return 0;
	}

@tragediana150, I think I haven't set up my graphics driver correctly. I only added GE=Yes. Maybe that's not enough. I replaced my stock HD6850 and put in a engineering sample HD6970. The boot stops at PCI configuration end with 6970 and sinetek's kernel. With Савва's kernel, it boots but shows a white screen without mouse cursor or log in screen.

Link to comment
Share on other sites
tragediana150
tragediana150
Posted
Posted

If you do not have full QE/CI the system is freezing, you are right. I had the same problem, system is running maybe 3-7minutes and then freeze, I could only moving the cursor with the coloured circle spinning. But if you enable QE/CI then you've got no freezing at all ;)

Link to comment
Share on other sites
tragediana150
tragediana150
Posted
Posted (edited)

Hey guys! been gone for a while whats the latest stable kernel release for mavericks, cheers!

Has much progress been made?

zip.gif  mach_kernel.zip - from Савва

zip.gif  BSA_kernel.zip - R9 by Andy

This is two latest kernels. Bronya said that the only Sinetek's kernel for old cpus (v6_kernel) :)

Edited by tragediana150
  • Like 1
Link to comment
Share on other sites
ADHDMedia
ADHDMedia
Posted
Posted

Screenflow, Wirecast, iMovie and many other professional apps tested with BSA V9 kernel and Final Cut looks working too

Very stable kernel and the system only freeze when I plug in my Windows Phone into an USB2 port (or maybe just the Apple trolling)

USB3 works fine

 

 

 

Thank you, right now Im stuck using 10.0.4 because anything newer than that gives me graphics corruption in my preview window.  I'll check out the R9 kernel.  

Link to comment
Share on other sites
ADHDMedia
ADHDMedia
Posted
Posted

So I tried to boot with the r9 kernel, and it gets stuck here: 

 

                   1/17/14 4:50:10.000 PM kernel[0]: hfs: mounted ADHD Mavericks on device root_device

 

…and doesn't go any further.  Any suggestions?

 

On another note, I just noticed with the Kabba kernel I get this message repeatedly in the console: 

 

                   1/17/14 4:45:57.352 PM Dock[702]: ImageIO: ConvertBytes vImageConvert_AnyToAny error: 'kvImageInvalidParameter'
                   1/17/14 4:45:57.352 PM Dock[702]: ImageIO: PNG No IDATs written into file
                   1/17/14 4:45:57.366 PM Dock[702]: ImageIO: ConvertBytes vImageConvert_AnyToAny error: 'kvImageInvalidParameter'
                   1/17/14 4:45:57.366 PM Dock[702]: ImageIO: PNG No IDATs written into file
                   1/17/14 4:45:57.367 PM Dock[702]: ImageIO: ConvertBytes vImageConvert_AnyToAny error: 'kvImageInvalidParameter'
                   1/17/14 4:45:57.367 PM Dock[702]: ImageIO: PNG No IDATs written into file
                   1/17/14 4:45:57.376 PM Dock[702]: ImageIO: ConvertBytes vImageConvert_AnyToAny error: 'kvImageInvalidParameter'
                   1/17/14 4:45:57.376 PM Dock[702]: ImageIO: PNG No IDATs written into file
                   1/17/14 4:45:57.376 PM Dock[702]: ImageIO: ConvertBytes vImageConvert_AnyToAny error: 'kvImageInvalidParameter'
                   1/17/14 4:45:57.377 PM Dock[702]: ImageIO: PNG No IDATs written into file
 
…does anyone know what this is?  Or how to fix it?  
 
Thanks guys. 
Link to comment
Share on other sites
ADHDMedia
ADHDMedia
Posted
Posted

 

So I tried to boot with the r9 kernel, and it gets stuck here: 

 

                   1/17/14 4:50:10.000 PM kernel[0]: hfs: mounted ADHD Mavericks on device root_device

 

…and doesn't go any further.  Any suggestions?

 

On another note, I just noticed with the Kabba kernel I get this message repeatedly in the console: 

 

                   1/17/14 4:45:57.352 PM Dock[702]: ImageIO: ConvertBytes vImageConvert_AnyToAny error: 'kvImageInvalidParameter'
                   1/17/14 4:45:57.352 PM Dock[702]: ImageIO: PNG No IDATs written into file
                   1/17/14 4:45:57.366 PM Dock[702]: ImageIO: ConvertBytes vImageConvert_AnyToAny error: 'kvImageInvalidParameter'
                   1/17/14 4:45:57.366 PM Dock[702]: ImageIO: PNG No IDATs written into file
                   1/17/14 4:45:57.367 PM Dock[702]: ImageIO: ConvertBytes vImageConvert_AnyToAny error: 'kvImageInvalidParameter'
                   1/17/14 4:45:57.367 PM Dock[702]: ImageIO: PNG No IDATs written into file
                   1/17/14 4:45:57.376 PM Dock[702]: ImageIO: ConvertBytes vImageConvert_AnyToAny error: 'kvImageInvalidParameter'
                   1/17/14 4:45:57.376 PM Dock[702]: ImageIO: PNG No IDATs written into file
                   1/17/14 4:45:57.376 PM Dock[702]: ImageIO: ConvertBytes vImageConvert_AnyToAny error: 'kvImageInvalidParameter'
                   1/17/14 4:45:57.377 PM Dock[702]: ImageIO: PNG No IDATs written into file
 
…does anyone know what this is?  Or how to fix it?  
 
Thanks guys. 

 

 

No one else gets this message with this kernel?

Link to comment
Share on other sites
ADHDMedia
ADHDMedia
Posted
Posted

Screenflow, Wirecast, iMovie and many other professional apps tested with BSA V9 kernel and Final Cut looks working too

Very stable kernel and the system only freeze when I plug in my Windows Phone into an USB2 port (or maybe just the Apple trolling)

USB3 works fine

 

 

So I got my rig to boot with the r9 kernel and I still have corruption in the newest version of FCPX.  Anyone else know what's causing this?  Kernel issue?  Not a kernel issue? 

Link to comment
Share on other sites
ADHDMedia
ADHDMedia
Posted
Posted
Final Cut Pro X Problem with OS X Mavericks ???

 

All versions after 10.0.4 have graphics corruption on all videos in the preview window. I'm wondering if this has to do with the kernels or if this is a problem that will need to be fixed elsewhere? I have the same problem with the newest version of iMovie.

Link to comment
Share on other sites
Sinetek
Sinetek
Posted
Posted

I found the reason of flooding OPEMU errors. I add a panic and halt the machine when it detects the first error. Then I found that IOATAFamily crashed in the stack trace. I used a patched IOATAFamily. I restored the original one and the error disappeared,

 

@sinetek: Should we change the handling of OPEMU errors? I'm thinking about adding saved_state->isf.rip += bytes_skip; before return 0. It keeps running the current opcode. Or is it better to exit the process that caused problems?

opemu.c, in opemu_ktrap
bad:
	{
		/* Well, now go in and get the asm text at least.. */
		const char *instruction_asm;
		instruction_asm = ud_insn_asm(&ud_obj);

                // This change is for debug only, not for production.
		printf ( "kOPEMU, %d:  %s\n", bytes_skip, instruction_asm) ;
		panic();
		pmCPUHalt(PM_HALT_PANIC);
		return 0;
	}

@tragediana150, I think I haven't set up my graphics driver correctly. I only added GE=Yes. Maybe that's not enough. I replaced my stock HD6850 and put in a engineering sample HD6970. The boot stops at PCI configuration end with 6970 and sinetek's kernel. With Савва's kernel, it boots but shows a white screen without mouse cursor or log in screen.

no, it would be better to get a full stack trace instead of merely halting. plus panic never returns, so that halt is never reached.

 

it would make sense though that the thread gets confused if the sse functions return a traditional error code with negative value. if the kernel thread fails, there is nothing we can do i'm afraid. So the panic on mov is caused by the ATA driver? even then, it shouldn't panic on a valid mov instruction.

 

Here's how I imagine it goes, tell me what you think:    instruction stream -> bad opcode ==> opcode emu, returns invalid value of -n, pc -= n; ==> invalid instruction stream ==> opemu hangs and confused.

 

in any case, error handling should work as is, so long as no emulator function returns a non-zero value (ie instruction length of caught instruction).

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

 

I appreciate your help, but these issues are completely unrelated to the problem I'm talking about.  I said "video corruption" meaning artifacts and whatnot, here's a screen: 

 

6pzgJ.png

Link to comment
Share on other sites
spakk
spakk
Posted
Posted

I appreciate your help, but these issues are completely unrelated to the problem I'm talking about. I said "video corruption" meaning artifacts and whatnot, here's a screen:

 

6pzgJ.png

 

I suspect that the graphics misfire caused Comes by your graphics card or by lack of compatibility of the graphics card drivers.

Link to comment
Share on other sites
carlo_67
carlo_67
Posted
Posted

Thank you sir, that's all I wanted to know ?

 

I have a phenom 2 x6 1075 with 2 ati 5770 in crossfire and no problems from year 10.6/10.7/10.8/10.9 
uses chameleon chameleon wizard and set with your framebuffer 
Aticonfig = Vervet or Hoolock

Anyway guys used Pandora installation is perfect

Thank you sir, that's all I wanted to know

no cpu overclck 
ASUS M4A78LT-M | AMD Phenom II X6 1075T@3.67GHz OC 
You need to set the bios manual 
CPU multiplier x15

Schermata 2014-01-19 alle 10.34.01.png

Schermata 2014-01-19 alle 10.38.01.png

  • Like 2
Link to comment
Share on other sites
ADHDMedia
ADHDMedia
Posted
Posted

 

I have a phenom 2 x6 1075 with 2 ati 5770 in crossfire and no problems from year 10.6/10.7/10.8/10.9 
uses chameleon chameleon wizard and set with your framebuffer 
Aticonfig = Vervet or Hoolock

Anyway guys used Pandora installation is perfect

no cpu overclck 
ASUS M4A78LT-M | AMD Phenom II X6 1075T@3.67GHz OC 
You need to set the bios manual 
CPU multiplier x15

 

 

Same problem for me, only with the latest versions of iMove and FCPX though.  He said it was an SSSE3 problem, so I guess I'll just have to wait for a fix. 

 

Davids-Mac-Pro-2:~ DMeister08$ ioreg | grep ATY

    | |   |   | +-o ATY,Vervet@0  <class AtiFbStub, id 0x100000468, registered, matched, active, busy 0 (6 ms), retain 32>

    | |   |   | | +-o ATY_Vervet  <class AMDFramebuffer, id 0x10000046c, registered, matched, active, busy 0 (5 ms), retain 41>

    | |   |   | +-o ATY,Vervet@1  <class AtiFbStub, id 0x100000469, registered, matched, active, busy 0 (1 ms), retain 32>

    | |   |   | | +-o ATY_Vervet  <class AMDFramebuffer, id 0x100000473, registered, matched, active, busy 0 (0 ms), retain 41>

    | |   |   | +-o ATY,Vervet@2  <class AtiFbStub, id 0x10000046a, registered, matched, active, busy 0 (0 ms), retain 32>

    | |   |   | | +-o ATY_Vervet  <class AMDFramebuffer, id 0x100000478, registered, matched, active, busy 0 (0 ms), retain 38>

    | |   |   | +-o ATY,Vervet@3  <class AtiFbStub, id 0x10000046b, registered, matched, active, busy 0 (0 ms), retain 32>

    | |   |   | | +-o ATY_Vervet  <class AMDFramebuffer, id 0x10000047d, registered, matched, active, busy 0 (0 ms), retain 38>

Link to comment
Share on other sites
Duran Keeley
Duran Keeley
Posted
Posted

Hi Guys

 

Just a heads up

 

In the last couple of minutes I have spent my time cleaning up this topic, please everyone post to the correct topics thank you.

 

For Questions:

AMD - Questions & Answers

 

For Pre-release Testing:

AMD - Pre-release Testing

 

For Help:

HELP TOPIC - further help in chat.osx86.hu/#legacykernel (IRC)

  • Like 2
Link to comment
Share on other sites
Duran Keeley
Duran Keeley
Posted
Posted

Hi Durran, I appreciate you to keep this forum clear, but I think the new release/build OSX is related with kernel.. to see worked or not worked,... except we got some problem so we can go to [AMD] Pre-release Testing ..

Sorry if I'm wrong ..

 

best regards

pico

Hi pico joe

 

Not quite true, Beta kernel sources do not get released to the public. The only time we get the sources is when apple releases the source of the public non beta update for everyone. Hence right now we are using old kernel source (10.9) on a new update (10.9.2).

 

So even if you are having problems with the kernel released on this topic, it is technically proven to work for 10.9. Since there was no kernel change in 10.9.1 that seems to work fine as well.

  • Like 1
Link to comment
Share on other sites
latvian88
latvian88
Posted
Posted

Had a random KP this morning. 

 

qd2VZvk.jpg

 

OS: Mavericks 10.9.1

CPU: AMD A8-5600K APU 

Kernel: BSA Kernel v9 ( andyvand:xnu-2422.1.72-AMD )

 

btw, not asking for help. just mentioning that this happened in case it helps with future kernel development.

Link to comment
Share on other sites
spakk
spakk
Posted
Posted (edited)

current version 13C44 works reliably on AMD Hackintosh :)

more info see here http://www.insanelymac.com/forum/topic/294712-amd-pre-release-testing/page-5?do=findComment&comment=1988625

Edited by spakk
  • Like 1
Link to comment
Share on other sites
 Share
Go to topic listing
×
×
  • Create New...