Chameleon with DSDT override

[stellarola]

I tried this method of fixing the DSDT on the MSI Wind. When trying to load up without a disabler kext for CPUPowermanagement. It read that the CPU was unsupported. I assume it's because its an ATOM CPU?

 

Also to note. This CPU works with vanilla kernel and has similar instructions to that of a Core based CPU. SSE3, SSSE3.

 

 

Any ideas?

 

Thanks,

 

-Stell

 

EDIT: Attached dsdt.dsl. I corrected the errors on this one. I have not patched this specific one to fix IntelCpupowermanagement, etc.

MSIWind_ATOM_dsdt.dsl.zip

[mackerintel]

I tried this method of fixing the DSDT on the MSI Wind. When trying to load up without a disabler kext for CPUPowermanagement. It read that the CPU was unsupported. I assume it's because its an ATOM CPU?

 

Also to note. This CPU works with vanilla kernel and has similar instructions to that of a Core based CPU. SSE3, SSSE3.

Supported features is not only difference that can exist between processors. While it's true that Atom is similar to cut-down Core with Hyper-threading instead of Dual-Core it probably uses Speedstepping more similar to Pentium M with its own additions. Also the CPUID version (family,model,stepping) are probably different. And I think that AppleCPUPowermanagement executes CPUID and compares the output with built-in table. In this case there is nothing that DSDT override can do about it. At least until Apple makes their own netbook (a kind "Macbook Mini"). This is also the reason why AppleCPUPowermanagement can't be used with AMDs. But I can also be wrong and perhaps your DSDT is screwed in another way (decompiled DSDT is about 15000 lines of codes it means 15000 ways it can be screwed up)

[stellarola]

Supported features is not only difference that can exist between processors. While it's true that Atom is similar to cut-down Core with Hyper-threading instead of Dual-Core it probably uses Speedstepping more similar to Pentium M with its own additions. Also the CPUID version (family,model,stepping) are probably different. And I think that AppleCPUPowermanagement executes CPUID and compares the output with built-in table. In this case there is nothing that DSDT override can do about it. At least until Apple makes their own netbook (a kind "Macbook Mini"). This is also the reason why AppleCPUPowermanagement can't be used with AMDs. But I can also be wrong and perhaps your DSDT is screwed in another way (decompiled DSDT is about 15000 lines of codes it means 15000 ways it can be screwed up)

 

 

I appreciate the input, brother. Well, hopefully something will come out of it. I got this working on my desktop mobo/cpu just fine.

 

You think they'll be any chance of a DFE 132 Chameleon with this DSDT override in the future?

 

Thanks mackerintel!

 

-Stell

[mackerintel]

I appreciate the input, brother. Well, hopefully something will come out of it.

CPUID has to be patched. Either on the disk or in the memory. The second method is used by voodoo-kernel( http://code.google.com/p/xnu-dev/ ) to make OSX work on AMD. But in case of AppleCPUP-M- it's probably wouldn't bring anything since powermanagement is probably (just a wild guess) not implemented the same way on Atom that on supported CPUs.

I got this working on my desktop mobo/cpu just fine.

 

You think they'll be any chance of a DFE 132 Chameleon with this DSDT override in the future?

Tamás Kosárzky recently posted a message on his site about big unified bootloader .

 

Thanks mackerintel!

You are welcome

[stellarola]

CPUID has to be patched. Either on the disk or in the memory. The second method is used by voodoo-kernel( http://code.google.com/p/xnu-dev/ ) to make OSX work on AMD. But in case of AppleCPUP-M- it's probably wouldn't bring anything since powermanagement is probably (just a wild guess) not implemented the same way on Atom that on supported CPUs.

 

Tamás Kosárzky recently posted a message on his site about big unified bootloader .

You are welcome

 

Perfecto.

 

Link to Tamás Kosárzky's site?

 

-Stell

[mackerintel]

Perfecto.

 

Link to Tamás Kosárzky's site?

 

-Stell

 

Sorry I made a typo . Correct spelling is Tamás Kosárszky. Linking to his site gives me an error. Just search for "Tamás Kosárszky" or "Chameleon EFI" in any search engine

[Shadow_2k]

Works like a charm. Tested with a downloaded AppleRTC 1.2.2 and a AppleCPUPowermanagement.kext that I copied over from my real mac.

[xtreme_xtc]

Thanks mackerintel for adding DSDT patching.

Fixed my acer 5610z it can SLEEP, however it will not resume, such that it will attempt to wake, and it will initialize the system and get no where.

 

Anyway i have uploaded the corrected dsl and aml.

 

Also another note if you guys get cpu thermal threshold signal must sleep in verbose remove all thermal kexts it will prevent your system from sleeping on boot.

corrected_5610z_dsdt.zip

[xhala_debru]

Hi,

 

doesn't work for me.

I got the dsdt.aml created by fass's patch, placed in the root of my partition and then replaced chameleon's boot file by yours.

 

At boot time, I see something about dsdt, boot volume is found, kext begin loading.

Then a message about power management time out.

After that boot process resumes but never completes.

I never reach the login window and booting in safe mode doesn't help neither.

[mackerintel]

Hi,

 

doesn't work for me.

I got the dsdt.aml created by fass's patch, placed in the root of my partition and then replaced chameleon's boot file by yours.

 

At boot time, I see something about dsdt, boot volume is found, kext begin loading.

Then a message about power management time out.

After that boot process resumes but never completes.

I never reach the login window and booting in safe mode doesn't help neither.

1) Try putting unmodified dump to /DSDT.aml (to know if it's a problem in Chameleon or DSDT patcher)

2) Post your DSDT

3) Try disabling modified and non-apple kexts

[xhala_debru]

1) Try putting unmodified dump to /DSDT.aml (to know if it's a problem in Chameleon or DSDT patcher)

2) Post your DSDT

3) Try disabling modified and non-apple kexts

Hi,

 

1) Unfortunately I can't because iasl complains about 2 errors of uninitialized Local0 variables.

It's in the Method (\_WAK, 1, NotSerialized) , around line 287

 

2) Here's my dsdt.dsl renamed in .txt to be able to upload it.

Don't know how to do it other way

 

 

 

3) Apart from Timemachine Fix (IONetworkingfalily.kext) and AppleHDA, I don't think I have anything else.

In /Extra/Extensions I've removed IntelCPUPMDisabler.kext but left SMBIOSEnabler.kext

 

Thanks for everything,

 

Xhala

dsdt.txt

[mitch_de]

Hi, i also get an warnimg (recompile) with GA-Ep35-DS3, but no errors so my dsdt.aml has beed generated.

./dsdt_fixed.dsl 222: Method (_WAK, 1, NotSerialized)

Warning 1080 - ^ Reserved method must return a value (_WAK)

 

Question:

How can users, which not have the cpu=1 problem, see that it worked (changed something) ?

I have the problem (so i tried this) that my Gigabyte BIOS didnt serve the PSTATES (ACPI i think) of the CPU for superhais speedstep.

I cant see any debug outputs (.logs) or any new text on bootscreen using the boot (12) and my dsdt.alm.

I installes boot (from the acpi12 (page1)) in my root and also the dsdt.alm (must be big letters?) . Before i tried an other boot (was called V2 but that was older than 12) which gave coloured ASCI screen - so i am sure that i did installed all right.

 

THANKS a lot for that very interesting way of BIOS fixing!

dsdt_GA_EP35_DS3.dsl.zip

[fassl]

1) Unfortunately I can't because iasl complains about 2 errors of uninitialized Local0 variables.

 

The new Patcher should fix this, check it out: http://forum.insanelymac.com/index.php?sho...t=0#entry947456

 

 

greetz

 

@mackerintel: I tried version acpi12 and it behaves the same like the first version you posted, the PC freezes short after the Kernel is loaded, either with unmodified dsdt.aml or modified.

 

Specs.

Fujitsu Siemens Amilo PI 1556

T2500 @2ghz

1,5Gb 667 Ram

Geforce 7600 Go

Leopard 10.5.5

[mackerintel]

@mackerintel: I tried version acpi12 and it behaves the same like the first version you posted, the PC freezes short after the Kernel is loaded, either with unmodified dsdt.aml or modified.

 

Specs.

Fujitsu Siemens Amilo PI 1556

T2500 @2ghz

1,5Gb 667 Ram

Geforce 7600 Go

Leopard 10.5.5

 

Can you try the debug version and post the result?

http://forum.insanelymac.com/index.php?s=&...st&p=944496

[fassl]

Oh sorry, it works, must have been me mixing up the bootloaders :/ -> human fail

 

greetz

[xhala_debru]

@fassl

I tried your aptcher and even if it does take care of the compiling errors and creates an dsdt.aml

But still, I have the same problem.

 

The computer hangs just before launching the loginwindow.app

 

What bothers me a bit is that after the forced reboot (push the button), my OC is messed a bit and I have to manually reapply the parameters in Bios.

And that normally, my computer is rock stable.

[mackerintel]

Oh sorry, it works, must have been me mixing up the bootloaders :/ -> human fail :censored2:

 

Never mind. Now I'm trying to mess around further with my DSDT to solve some secondary issues. If I find something will inform you for addition in your patcher. Surfing through the net I found a ioreg dump from macbook and macbook air. Extracted DSDTs from theese dumps. Might be useful so I attach it

macbook.zip

[fassl]

You sure your Graphicscard is set up properly? And also do you have any decrypt kext like dsmos.kext or AppleDecrypt.kext loaded?

 

greetz

[xhala_debru]

You sure your Graphicscard is set up properly? And also do you have any decrypt kext like dsmos.kext or AppleDecrypt.kext loaded?

 

greetz

Sure, the graphic card is correctly setup.

I use it with Extended display on two screens, QE/CI supported.

And the way I do is I clone my working install on a spare disk, reinstall chameleon with patch and then remove the blablaintelblacpu.kext

I'd have to check again but dsmos.kext is still there but don't know if it's loaded (I'm 99% sure it is)

 

But I saw that relikwie had exactly the same problem as me.

And he's having roughly the same conguration as I do, which is:

- Gigabyte MB

- Nvidia GC

- GFX strings for GC (I don't remeber if I've EFI strings for sound and net)

 

So I think we may have a non working config

[riws]

As fassl knows my aml File won't compile, too many Errors. So smarties?

[fassl]

As fassl knows my aml File won't compile, too many Errors. So smarties?

 

Removed everything that error'ed, maybe it will work riws.zip

[BlackCH]

My dsdt.dsl wont compile even thoug is unmodified. I dont know if the decomp is wrong or what.

Using iasl from the pacher I get:

ACPI Error (dmutils-0261): Argument count mismatch for method \_TZ_.C39F 1 0 [20080926]

in the first line of the .dsl file..

I attached my dsdt.dat dump made on linux. maybe any of you can try to decompile it and see whats wrong with it...

HP6720s_dsdt.dat.zip

[mackerintel]

Released a new version. New features are dropping SSDT and specifiing DSDT filename (very useful for testing of different DSDTs)

[Superhai]

I tried this, but was not very fond of my mb i think. I recompiled the original DSDT and same problem. Error is roughly

 

ACPI: RSDP.....

AcpiInitializeTables: AE_NO_MEMORY

 

and then panic Unable to load "ACPI"

 

tried this with the newest and also the acpiv12 version. This is on the Gigabyte GA-P35-DQ6

[mackerintel]

I tried this, but was not very fond of my mb i think. I recompiled the original DSDT and same problem. Error is roughly

 

ACPI: RSDP.....

AcpiInitializeTables: AE_NO_MEMORY

 

and then panic Unable to load "ACPI"

 

tried this with the newest and also the acpiv12 version. This is on the Gigabyte GA-P35-DQ6

 

So you have this problem both with and without DSDT override? Or only when overriding DSDT with its recompilation? Can you also try the debug version with "-v" parameter? (but be patient it contains long sleeps)