Jump to content

Chameleon RC5 mode with mem detection enabled and automatic P-States & C-States generation for native power managment


kozlek
 Share

1,214 posts in this topic

Recommended Posts

The only thing I have right now is that the OS sounds are cracky (Latest VooDooHDA). Playing music with iTunes is fine. The AudioChip is an Analog Devices AD 1981

 

My FSB is detected fine and the MP is corrected. But still have a problem with the audio.

 

Thx for replying.

 

With the VoodooHDA.kext v. 2.7.2 along with the VoodooHDA v1.1 preference pane, the SS2 noise reduction feature must be enabled to stop the crackling noise. However, this must be done after every reboot as the buggy preference pane refuses to permanently save and resets to defaults after each reboot.

 

Reference: http://www.projectosx.com/forum/index.php?showtopic=355

Link to comment
Share on other sites

*** SOLVED **************************[/size]

 

With the VoodooHDA.kext v. 2.7.2 along with the VoodooHDA v1.1 preference pane, the SS2 noise reduction feature must be enabled to stop the crackling noise. However, this must be done after every reboot as the buggy preference pane refuses to permanently save and resets to defaults after each reboot.

 

Reference: http://www.projectosx.com/forum/index.php?showtopic=355

 

 

mmm thats indeed changing:

 

FSB corrected to 200MHz + SSE2 + Noise reduction better OS sounds | Bad itunes sound

 

FSB corrected to 200MHz + SSE2 off + Noise reduction off bad OS sounds | Good itunes sound

 

FSB @ 185Mhz good sounds in OS and iTunes

 

My Memory is DD2 666Mhz and my CPU is 200MHz FSB. So I think it has to do with the FSB devider for Memory and CPU. If FSB is set to 200 the AD1981 gets maybe a FSB which is to high. If I lower it to something in the middle like 185 (between 166 and 200) everything is fine.

 

Also a MP of 15 actualy works. Geekbench at 210Mhz MP15 give 4950 points ! Amzing but with crappy sound.

 

 

*** The VooDooBattery is causing the Crackling sounds. Found out that removing and replacing the AC adapter plug was changing the crackling noise. So I removed the Voodoobattery and now mem / cpu detected correctly and sounds works fine ! ***

Link to comment
Share on other sites

Win,partially..with MemDetection turned the system boots up but it is just unbearably slow, it feels like there's a ton of interrupts happening while I use the system, slowing everything down.

Hi, sorry for the huge delay :mellow: and thanks for reading the instructions.

There are problems with some laptops and memdetect, i just didn't knew about the hang and it seems you're not alone on that.

Can't tell much about the rest; try to check around... maybe a power management problem. You do have some wrong keys on the boot.plist:

- arch should be used like this:

	   <key>arch</key>
   <string>i386</string>

 

- "Graphics Mode" like this:

	   <key>Kernel Flags</key>
   <string>"Graphics Mode"=1024x768x32</string>

and by the way, 1024x768x32 is the default resolution.

 

- on RestartFix the default is true, no need to specify it unless you're using my branch :P or testing.

 

But none of this is really a problem and it's certainly not relevant.

On the other hand, i'd be careful with the smbios.plist.. for instance,

don't know if that's the most appropriate Mac for you?

Did you try Chameleon's defaults? without smbios.plist...

Link to comment
Share on other sites

I thought I would share some findings about my setup.

 

I finally got the main trunk of Chameleon to work with my ATI5850 with a quite dirty workaround. It's not pretty, but I get into the desktop. First of all let's see how I made it, and then the tradeoff.

 

- Dumped the GFX ROM in Windows Z with GPU Z.

 

- Modified the official RC5 r.630:

 

        { 0x100268B8,  "Motmot"},

       { 0x100268BE,  "Motmot"},

       { 0x10026898,  "Motmot"},

       { 0x10026899,  "Motmot"}

 

Seriously guys, WTF??? 6899 is Uakari, why no love at all for us???

 

- Added in apple.boot,plist:

 

<key>UseAtiROM</key>

<string>Yes</string>

 

- Booted 10.6.5

 

Now this produced a functional desktop, but I think the power management for the GFX card does not load. I think so because of these 2 very inconvenient glitches:

 

- Sleep works, the machine goes to sleep like a baby, but resuming, the CPU and all works again, drives come back to life and so do the fans, but video remains black. The framebuffer cannot restart.

 

- The GFX fan spins at slowest settings all the time, which is nice for the ears, but it produces lots of unwanted visual glitches, intermittent white pixels etc. The card clearly overheats since the fan cannot adjust on the load.

 

So guys, can we at least set the proper framebuffer names in the main trunk? We are talking about cards officially supported on OSX, they should be given as much importance as the NVidia.

 

I hope my report was of any use for the devs. Apart of that the booter works very well, and I got P-states on my i7 :(

 

Cheers

Link to comment
Share on other sites

What does this actualy mean ?

 

It seems P states and C states are not injected,

 

 

FADT: Restart Fix applied !
Invalid character found in ProcessorOP 0x5c!
Invalid character found in ProcessorOP 0x5c!
ACPI CPUs not found: C-States not generated !!!
Invalid character found in ProcessorOP 0x5c!
Invalid character found in ProcessorOP 0x5c!
ACPI CPUs not found: P-States not generated !!!
FADT: Restart Fix applied !
Invalid character found in ProcessorOP 0x5c!
Invalid character found in ProcessorOP 0x5c!
ACPI CPUs not found: C-States not generated !!!
Invalid character found in ProcessorOP 0x5c!
Invalid character found in ProcessorOP 0x5c!
ACPI CPUs not found: P-States not generated !!!
Patched ACPI version 2 DSDT

 

How can I correct this ?

 

TIA

Link to comment
Share on other sites

Here's the latest trunk (r635) compiled just incase someone needs it.

 

Chameleon2_RC5_Trunk_r635.zip

 

 

I also want to report that it works pretty well on my Gigabyte GA-EP35-DS3L motherboard, E8400 CPU, Reference Sapphire 4890.

Everything works fine: C & P states generated and working properly, CPU and Mem. identified correctly including their speeds..etc

 

Thanks a bunch for the hard work guys, keep it up!

Link to comment
Share on other sites

As far as I can tell, vanilla speedstepping isn't working for me as P and C-States are not being generated in bdmesg. AppleLPC is loading via DSDT fix. CPU temps are 10 degrees hotter than with NullCPUPowerManahement. Any ideas?

 

ToBeMac: Our rigs are pretty similar... would you mind posting the bdmesg from your desktop?

 

Using PCI-Root-UID value: 1
No SMBIOS replacement found.
Intel P35/G33 DRAM Controller [8086:29c0] at 00:00.0
Frequency detected: 533 MHz (1066) Dual Channel 
CAS:5 tRC:7 tRP:7 RAS:24 (5-7-7-24)
Slot: 0 Type 19 2048MB (DDR2 SDRAM) 1066MHz Vendor=Mushkin
     PartNo=991599996599 SerialNo=00000000
Slot: 2 Type 19 2048MB (DDR2 SDRAM) 1066MHz Vendor=Mushkin
     PartNo=991599996599 SerialNo=00000000
CPU is Intel® Core(tm)2 Duo CPU     E8400  @ 3.00GHz, family 0x6, model 0x17
Patched DMI Table
Found SMBIOS System Information Table 1
Customizing SystemID with : 00000000-0000-0000-0000-001a4d5f28cc
efi_inject_get_devprop_string NULL trying stringdata
Read HFS+ file: [hd(0,2)/Extra/DSDT.aml] 18681 bytes.
ACPI table not found: SSDT.aml
FADT: Restart Fix applied!
FADT: Using custom DSDT!
Starting Darwin x86_64
Read HFS+ file: [hd(0,2)/Extra/Themes/Default/boot.png] 10950 bytes.

 

My c.a.boot.plist:

 

<plist version="1.0">
<dict>
<key>Graphics Mode</key>
<string>1680x1050x32</string>
<key>Kernel</key>
<string>mach_kernel</string>
<key>Kernel Flags</key>
<string>-v</string>
<key>arch</key>
<string>i386</string>
<key>Timeout</key>
<string>3</string>	
<key>Default Partition</key>
<string>hd(0,3)</string>
<key>Instant Menu</key>
<string>No</string>
<key>GraphicsEnabler</key>
<string>Yes</string>
<key>PciRoot</key>
<string>1</string>
<key>SMUUID</key> 
<string>40FE6F5D-84AE-3F61-950E-674FB1BD36CF</string>
	<key>GeneratePStates</key>
<string>Yes</string>
	<key>GenerateCStates</key>
<string>Yes</string>
<key>DropSSDT</key>
<string>Yes</string>
	<key>Rescan</key>
<string>Yes</string>
</dict>
</plist>

SMC.tiff

LPC.tiff

Link to comment
Share on other sites

I was even able to get things running optimally by picking the correct macbook pro model that had my CPU back in 2007 in smbios.plist - as soon as I stepped down from a MacBookPro4,1 to a MacBookPro3,1 things started humming. All this by simply reading and doing a little thinking.

Hi, does anyone know which MacBook model I should put in my SMBIOS.plist file if I'm using an Atom based Netbook or will this make no difference as no MacBooks exist with an Atom CPU?

Link to comment
Share on other sites

As far as I can tell, vanilla speedstepping isn't working for me as P and C-States are not being generated in bdmesg. AppleLPC is loading via DSDT fix. CPU temps are 10 degrees hotter than with NullCPUPowerManahement. Any ideas?

 

ToBeMac: Our rigs are pretty similar... would you mind posting the bdmesg from your desktop?

 

Using PCI-Root-UID value: 1
No SMBIOS replacement found.
Intel P35/G33 DRAM Controller [8086:29c0] at 00:00.0
Frequency detected: 533 MHz (1066) Dual Channel 
CAS:5 tRC:7 tRP:7 RAS:24 (5-7-7-24)
Slot: 0 Type 19 2048MB (DDR2 SDRAM) 1066MHz Vendor=Mushkin
     PartNo=991599996599 SerialNo=00000000
Slot: 2 Type 19 2048MB (DDR2 SDRAM) 1066MHz Vendor=Mushkin
     PartNo=991599996599 SerialNo=00000000
CPU is Intel® Core(tm)2 Duo CPU     E8400  @ 3.00GHz, family 0x6, model 0x17
Patched DMI Table
Found SMBIOS System Information Table 1
Customizing SystemID with : 00000000-0000-0000-0000-001a4d5f28cc
efi_inject_get_devprop_string NULL trying stringdata
Read HFS+ file: [hd(0,2)/Extra/DSDT.aml] 18681 bytes.
ACPI table not found: SSDT.aml
FADT: Restart Fix applied!
FADT: Using custom DSDT!
Starting Darwin x86_64
Read HFS+ file: [hd(0,2)/Extra/Themes/Default/boot.png] 10950 bytes.

 

My c.a.boot.plist:

 

<plist version="1.0">
<dict>
<key>Graphics Mode</key>
<string>1680x1050x32</string>
<key>Kernel</key>
<string>mach_kernel</string>
<key>Kernel Flags</key>
<string>-v</string>
<key>arch</key>
<string>i386</string>
<key>Timeout</key>
<string>3</string>	
<key>Default Partition</key>
<string>hd(0,3)</string>
<key>Instant Menu</key>
<string>No</string>
<key>GraphicsEnabler</key>
<string>Yes</string>
<key>PciRoot</key>
<string>1</string>
<key>SMUUID</key> 
<string>40FE6F5D-84AE-3F61-950E-674FB1BD36CF</string>
	<key>GeneratePStates</key>
<string>Yes</string>
	<key>GenerateCStates</key>
<string>Yes</string>
<key>DropSSDT</key>
<string>Yes</string>
	<key>Rescan</key>
<string>Yes</string>
</dict>
</plist>

 

Same to me , speedstep stopped working with 10.6.5.

Doenst matter if i use handmade DSDT or the autogen by bootloader.

There must be some changes in 10.6.5.

I believe a lot of people have that problem but didnt see it (not easy to look if real MHz change).

Link to comment
Share on other sites

Can someone please explain this error:

Invalid character found in ProcessorOP 0x2e!

 

System specs:

HP dc7600 cmt

MB = intel 945 ICH7

Processor Intel Pentium D 945 (3.40GHz dual core)

VGA = NVidia GeForce 7600 GT 256MB

RAM 2.5GB DDR2 533MHz

BIOS - latest

OS 10.6.5 with legacy kernel

 

DSDT.aml yes

 

error message at the bottom of code:

bdmesg
LoadConfig: [hd(0,2)/Extra/com.apple.Boot.plist] 674 bytes.
CPU: Intel® Pentium® D CPU 3.40GHz
max: 17 current: 0
CPU: Vendor/Family/ExtFamily: 0x756e6547/0xf/0x0
CPU: Model/ExtModel/Stepping: 0x6/0x0/0x5
CPU: Multipliers x10:         max=170, min=0
CPU: MaxDiv/CurrDiv:          0x0/0x0
CPU: TSCFreq:                 3389MHz
CPU: CPUFreq:                 3389MHz
CPU: FSBFreq:                 199MHz
CPU: Enhanced SpeedStep:      1
CPU: NoCores/NoThreads:       2/2
CPU: Features:                0x268002a1
Chipset is 945G (pci id 0x27708086)
We have an NVIDIA Card
Creating 2 Mode Tables
New table with id : 0
New table with id : 1
First Standard VESA Table at offset 0x8b33
Second Standard VESA Table at offset 0x2d4e
Table #0 has 16 modes
Table #1 has 30 modes

Darwin/x86 boot v5.0.132 - Chameleon v5.0.1 r164
Build date: 2010-11-19 22:48:22
2559MB memory
VESA v3.0 256MB (NVIDIA)
Scanning device 80...Press any key to enter startup options.(10) (9) (8) Use  keys to select the startup volume.    hd(0,2) Snow Leopard HD                 
   hd(1,2) Leopard                         
   hd(2,1) DC7600-C                        
Press Enter to start up Darwin/x86 with no options, or you can:
 Type -v and press Enter to start up with diagnostic messages
 Type ? and press Enter to learn about advanced startup options

boot: LoadConfig: [hd(0,2)/Extra/com.apple.Boot.plist] 674 bytes.
Setting boot-uuid to: 73286EBC-B4F7-3BCF-8C6E-4688824E5682
Closing VBios
Table #0 : Freeing backup	[OK]
Table #0 : Freeing		[OK]
Table #1 : Freeing backup	[OK]
Table #1 : Freeing		[OK]
Freeing map			[OK]
LoadConfig: [hd(0,2)/System/Library/CoreServices/SystemVersion.plist] 479 bytes.
Loading Darwin 10.6
Loading kernel legacy_kernel
LoadDrivers: /Extra/Extensions.mkext
LoadDrivers: /System/Library/Caches/com.apple.kext.caches/Startup/Extensions.mkext
LoadACPI: [hd(0,2)/Extra/DSDT.aml] 35508 bytes.
Using PCI-Root-UID value: 0
nVidia GeForce 7600 GT 256MB NV4b [10de:0391] :: PciRoot(0x0)/Pci(0x1,0x0)/Pci(0x0,0x0)
NVCAP: 04000000-00000300-1c000000-0000000a-00000000
@0,display-cfg: ffffffff
@1,display-cfg: ffffffff
No SMBIOS replacement found.
Slot: 0 Type 19 1024MB (DDR2 SDRAM) 533MHz Vendor=Kingston
     PartNo= SerialNo=6820B583
Slot: 1 Type 19 512MB (DDR2 SDRAM) 533MHz Vendor=Samsung
     PartNo=M378T6553CZ3-CD53C SerialNo=031013E7
Slot: 2 Type 19 1024MB (DDR2 SDRAM) 533MHz Vendor=Kingston
     PartNo= SerialNo=98209983
Patched DMI Table
Found SMBIOS System Information Table 1
Customizing SystemID with : f6c51d2e-2bcd-da11-bbda-35aa68ec0016
LoadACPI: [hd(0,2)/Extra/DSDT.aml] 35508 bytes.
FADT: Restart Fix applied !
[b]Invalid character found in ProcessorOP 0x2e![/b]
[b]Invalid character found in ProcessorOP 0x2e![/b]
[b]ACPI CPUs not found: P-States not generated !!![/b]
Patched ACPI version 1 DSDT
Starting Darwin x86
Press any key to continue...

 

TIA

Link to comment
Share on other sites

Terente:

 

I'm guessing that those error messages (one for each CPU core?) have to do with Chameleon 2.0 RC5s native CPU Power Management, which will not work with a Pentium D CPU.

 

If you've added GenerateCStates/PStates=yes to your com.apple.Boot.plist, remove them and see if the error message goes away.

 

If not, then maybe you can get rid of them by explicitly setting "GeneratePStates" and "GenerateCStates"="No".

 

If that doesn't help, then just ignore those messages. You cannot use this feature with your CPU because it is not compatible with AppleIntelCPUPowerManagement.kext. You must use something like VoodooPower.kext instead.

 

mitchde:

 

P- and C-states are still working here with Chameleon 2.0 RC5 r647 and 10.6.5. CPU is C2D E8500, ICH10R/P45 chipset, MacPro3,1 model identifier.

Hell, even S3 sleep/wake started working. Hackintosh Paradise. ;)

Link to comment
Share on other sites

"mitchde: P- and C-states are still working here with Chameleon 2.0 RC5 r647 and 10.6.5. CPU is C2D E8500, ICH10R/P45 chipset, MacPro3,1 model identifier."

There were major changes in the .plists of the ACPI_SMC_Plotformplugin of 10.6.5.

My LegancyPM (injects some other PM data into that plugin) didnt work anymore for my iMac Modell.

SOLVED by deleting LegancyPM (used in /Extra) and edited iMac11,1 for my usage. Speedstep works now again.

(More details in the valv bootloader thread).

PS: The problem doesnt belongs to RC or valv or handmade dsdt Pstates - only iMac11,1 (modded) worked.

Link to comment
Share on other sites

Hi, does anyone know which MacBook model I should put in my SMBIOS.plist file if I'm using an Atom based Netbook or will this make no difference as no MacBooks exist with an Atom CPU?

I think you just answered your own question. :)

 

You can try MacPro3,1 as that model has no CPUPLimit by default. Note that none of this works if you don't get the c-states and the p-states generating properly, though. :)

Link to comment
Share on other sites

Do I still need the nullcpupowermanagement kext, with this bootloader (with C and P states enabled) and without a DSDT on a Core 2 duo PC?

 

You're the only one who can answer that question. Why did you need nullcpupowermanagement in the first place?

 

If you have nullcpupowermanagement loaded, Chameleon 2.0 RC5 P- and C-states will not work.

 

DSDT or no DSDT doesn't change that.

Link to comment
Share on other sites

I think you just answered your own question. :)

 

You can try MacPro3,1 as that model has no CPUPLimit by default. Note that none of this works if you don't get the c-states and the p-states generating properly, though. ;)

Thanks, I've give MacPro3,1 a go. :)

 

Another user with the same Netbook as mine (Acer Aspire One) has managed to get Sleep working with this new version of Chameleon, so I think C-states are generating Ok.

 

So far no one seems to has managed to get Speedstep on the CPU to work (yet!) without extra extensions like VoodooPower so P-states are going to need some more work though...

Link to comment
Share on other sites

I have an EVGA X58 SLI3. It's known that EVGA X58 bioses newer than sz2z don't work with native speedstep. I've found no documented cases, and all attempts of mine and other people's results in a kernel panic without using a disabler. What should I be looking for if I want to get it working?

 

I have C2RC5 r635, DSDT with minimal mods per the gigabyte 1366 guide on tonymacx86, iMac11,1, tried DropSSDT=yes and no, GenerateCStates=yes, PStates=no, and ssdt.aml ssdt-1.aml with the p states in them.

 

sz2z is older than the SLI3 and isn't available for it.

 

Also my 1600 ram shows as 1066.

Link to comment
Share on other sites

Seeing a strange problem since getting all this stuff working - and since everyone on this thread is probably a native OSX speedstep user, perhaps this is the best audience, so bear with me...

 

As I mentioned a few posts ago, I got all the native speedstepping working by following instructions on this thread and by creating a legacy custom kext to eliminate the CPUPLimit for my MacBookPro3,1 - so all is well and the machine steps through all 10 stages and I'm running around on average 6 to 10 degrees cooler.

 

So obviously, I now no longer need Sleepenabler.kext and NULLCPUPowerManagement.kext - hooray. The computer sleeps, and I even got video mirroring working now, which never worked before (surprised that it's related, but I'll take it!)

 

HOWEVER - I can no longer sleep the machine with my iPhone plugged in - it just crashes immediately and eventually returns a sleep failure code of 0x12000770. It's very clearly a crash in the USB driver and this is indicated in the dump. Has anyone else experienced this behavior? Other simple USB devices (mouse, keyboard, mass storage, etc) are fine and cause no problem - so far it's just the iPhone. When I switched back to NullCPUManagement and SleepEnabler, the computer sleeps fine with the iPhone plugged in. Not a tradeoff I'm willing to make and I can live with unplugging the phone before sleep, but this is bizarre.

 

If any kind soul is interesting in helping, I'll be happy to upload the dump details.

 

Thanks!

Link to comment
Share on other sites

Maybe there's a problem with the sleep/wake DSDT code for your USB ports.

There are many posts about patching the EHCI/UHCI USB devices.

 

Here's zhell's oldie but goodie, lots of useful tips later in the thread as well:

http://www.insanelymac.com/forum/index.php?showtopic=168014

More:

http://www.insanelymac.com/forum/index.php...t&p=1240686

http://www.insanelymac.com/forum/index.php?showtopic=212815

 

Try different USB controller settings in your BIOS.

 

Also take a look at the official documentation for Chameleon (VoodooProjects forum, General board) there are three USB fix flags you can add to /Extra/com.apple.Boot.plist.

Link to comment
Share on other sites

Maybe there's a problem with the sleep/wake DSDT code for your USB ports.

There are many posts about patching the EHCI/UHCI USB devices.

...

Hey, thanks for the quick response! I've done some DSDT patching for EHCI/UHCI but as you said maybe there's something more so I'll look through your links. :P Funny thing is that it's just the iPhone that causes problems (other USB devices are ok), and this sleep problem only arises when AppleIntelCPUManagement is active. Anyway I'll let you know how it goes!

Link to comment
Share on other sites

Hey, thanks for the quick response! I've done some DSDT patching for EHCI/UHCI but as you said maybe there's something more so I'll look through your links. ;) Funny thing is that it's just the iPhone that causes problems (other USB devices are ok), and this sleep problem only arises when AppleIntelCPUManagement is active. Anyway I'll let you know how it goes!

Just a update, I went back through these various EHCI/UHCI DSDT modding tutorials linked above and I did find two mistakes in my DSDT, one was using the same value for the 'AAPL,clock-id' parameter across all my USB devices (prolly not a big deal) and the other was injecting an ICH10's device IDs into my UHCI methods when it was not needed (I have an ICH8 which is natively supported, doh!) - sometimes cutting and pasting is bad. :)

 

Anyway the problem with AppleIntelCPUPowerManagement was not fixed and still remains but it was nice to fix those mistakes. I will do some more testing to see if devices other than iPhones cause this - I also want to narrow down to see if it's EHCI only, which I suspect.

 

If anyone is interested, I'll post the crash dump pointing to IOUSBFamily.

 

Thanks!

Link to comment
Share on other sites

Alright so let me get this straight:

 

If I have NullCPUPM, C2RC5's C and P states won't work right, but what if I pass the -allowAppleCPUPM flag at boot? Will they still not work unless I delete NullCPUPM?

 

If C2RC5 says that it has generated the C states correctly before I start booting, does that absolutely mean that it's true? Is there a chance it could be doing it wrong?

 

Let's assume that I've done everything right to get this working, but I still can't boot without NullCPUPM or else I get a panic, what else should I be looking for?

 

Thanks

Link to comment
Share on other sites

If C2RC5 says that it has generated the C states correctly before I start booting, does that absolutely mean that it's true?

 

Chameleon does that independently of AppleIntelCPUPowerManagement.kext, which is not even loaded at that point.

 

That message is of no consequence to you because the Chameleon generated C- and P-states are meaningless on a system where AppleIntelCPUPowerManagement is not loaded.

Link to comment
Share on other sites

 Share

×
×
  • Create New...