VoodooPower 1.2.3

[sdsdysj]

Work great on C2D 5450 !!!

[pasi]

Same here: Intel C2D , Vanialla Kernel, OpenRestart.kext used - without Voodoo speedstep shuzdowm works.

All other is OK, much better stepping than other 1.1.X versions before

.

 

killed shutown to on hp530, everything else ok

[cvad]

The same result - killed shutdown. On previous version GenericCPUPowerManagement.kext all OK.

[Superhai]

There was a small mistype in the shutdown hack routine so it didnt shut down the highest cpu. I have uploaded a new release (same version number) which should work now. For other fixes, I will see if I can figure out. But without a full bug report don't expect it to happen.

 

It loaded up fine (older versions always KP'd) on my Pentium M. But typing sysctl -a | grep freq resulted in a KP. I don't have a camera right now so can't post a pic, and it didn't contain any useful info either (no backtrace). It was a page fault but without a backtrace or at least a list of the loaded module ranges, I couldn't make sense of the EIP.

 

Thats an odd one. It only modifes the frequency and multiplier field, so I guess one of those got corrupted or have a value that breaks the sysctl routine in the kernel. Does it happen every time?

[Konami®]

There was a small mistype in the shutdown hack routine so it didnt shut down the highest cpu. I have uploaded a new release (same version number) which should work now. For other fixes, I will see if I can figure out. But without a full bug report don't expect it to happen.

Thats an odd one. It only modifes the frequency and multiplier field, so I guess one of those got corrupted or have a value that breaks the sysctl routine in the kernel. Does it happen every time?

 

I downloaded and still not working here the shutdown and happen all the time, restart works.

[stellarola]

I downloaded and still not working here the shutdown and happen all the time, restart works.

 

 

Hey, same here. I tried the newest version and shutdown is still not functioning. I'm using openhaltrestart.kext if you think that matters at all.

 

-Stell

[pasi]

shutdown still broken

[mitch_de]

There was a small mistype in the shutdown hack routine so it didnt shut down the highest cpu. I have uploaded a new release (same version number) which should work now.

 

Hi Superhai !

I also tried to look into that part of code for some fixing.

 

#pragma mark -

#pragma mark Shutdown/Restart Hack

int

FallbackHaltRestart(unsigned int type)

{

UInt32 count = GlobalIOCpuCount - 1 ;

switch (type) {

case kPERestartCPU:

......

break;

case kPEHaltCPU:

while (count--) {

if (GlobalIOCpuArray[count]) GlobalIOCpuArray[count]->haltCPU();

}

break;

}

return OriginalHaltRestart(type);

}

 

Can it be that the while (count--) does minus the count "too much" so that highest cpu didnt get haltet ?

I changed UInt32 count = GlobalIOCpuCount - 1; to count = GlobalIOCpuCount (count is one higher) but shutdown

also not work compiling it new.

 

Can it be that our (most use that?) OpenHaltRestart.kext can be a problem ?

 

THANKS

[Pietruszka]

Works great for me.... I've got Intel Dual-Core E2160.....it's 1,8GHz but it's overclocked to 2,7GHz.....

 

and now it works like under windows mean speedstep from 1,8 to 2,7 (x6 - 1,8 x7 - 2,1 x8 - 2,4 x9 2,7) so it's great. CPU-X shows this results

 

don't know about changing voltage because CPU-X shows nothing...hope it works

 

big thanks for this kext . waiting for GUI

[Aranius]

Samsung R560 -> Kills shutdown, downloaded about 20min ago... I never needed Openhaltrestart.kext or anything like that, worked right out of the box... until now^^

GenericCPUPowerManagement never did that^^

[Superhai]

Gosh, sorry, fixed it halfway. Try now

 

I also tried to look into that part of code for some fixing.

Can it be that our (most use that?) OpenHaltRestart.kext can be a problem ?

 

#pragma mark -
#pragma mark Shutdown/Restart Hack

int
FallbackHaltRestart(unsigned int type)
{
UInt32 count = GlobalIOCpuCount;
switch (type) {
	case kPERestartCPU:
		__asm__ volatile("outb %b0,%w1" : : "a" (KeyboardReset), "Nd" (KeyboardPort));
		//Not working? Force reboot via division by zero
		__asm__ volatile("lidt no_inter\n"
						 "xor %%eax, %%eax\n"
						 "xor %%ecx, %%ecx\n"
						 "div %%ecx, %%eax\n"
						 "ret\n"
						 "no_inter:\n"
						 ".word 0\n"
						 ".long 0\n"
						  : : );
		break;
	case kPEHaltCPU:
		while (--count) {
			if (GlobalIOCpuArray[count]) GlobalIOCpuArray[count]->haltCPU();
		}
		break;
}
return OriginalHaltRestart(type);
}

 

This is how the fixed version looks like. And there is no need to use the OpenHaltRestart.kext together with this.

[Wiwi]

Hi all, and thanks superhai for your work, as always !!!

 

Just some questions :

 

What are CStates ? Which is the difference with P and T State, not clear...

 

Edit (thanks tom hardware):

The closest AMD-based tech related to this [C State] is Cool and Quiet. Basically, if the processor is idle for a period of time, the computer shuts down the core to a specific degree. The longer the system is idle, the more the core is shut off. This saves power and in turn reduces noise.

 

The second paragraph basically answers whether you should use C-State or not. C-State is a personal choice. If the computer is only going to be idle for short periods of time, the paper states that more power can be saved by letting the core idle without entering C-State.

 

Basically, use C-State if it makes sense and you want to. If you use Cool and Quiet on your AMD machines, then you may want to use this too (Although this is a more intensive form of Cool and Quiet, because it actually shuts down parts of the core).

Isn't it amazing what questions Google can answer?

 

 

What's "LoopTimerFactor" option ???

 

That's all.

[Superhai]

I modified the old GenericCPUControl Application to work with this kext also, but it will be deprecated and will not be updated anymore. But you can use it until I get time to create prefpane and statusbartool.

[cvad]

OK.

Thank`s

[Superhai]

Just some questions :

What are CStates ? Which is the difference with P and T State, not clear...

What's "LoopTimerFactor" option ???

 

LoopTimerFactor is how often it check load and change state. It is changed to a factor do avoid dividing values (which is kind of "heavy" on the cpu compared to a shift). To find the ms value you can use this formula 2^x * 10. So 2 = 40ms and, 4 = 160 ms, 6 = 640 ms.

 

A short explanation of the different states.

 

G-States

Global power states:

G0 - Operational state

G1/G2 - Sleeping state

G3 Mechanical Off

D-States

Device power states:

D0 - Full power

D1/D2 - Low power

D3 - Off

S-States

Sleeping states: CPU is non-operational.

S0 - No sleep

S1 - No context lost

S2 - CPU context lost

S3 - All context except memory lost

S4 - All devices off

S5 - All context off (Soft off)

C-States

Idle states (CPU in non-execution mode)

C0 - Normal execution

C1 - "No latency" nap

C2 - Latency, but no software sideeffects

C3 - OS must watch for CPU cache states

C4/C5/C6 - CPU specific states

P-States

Performance states

P0 - Highest performance state

Px - Lower performance

T-States

Throttle states

T0 - No throttle

Tx - Throttle

 

S-States are user controllable, and according to users need. C-States are OS controllable, and (based on performance profile) changed whenever the CPU is not doing anything (basically waiting for things to do, but still in a working state. P-States are OS controllable, (based as well on peformance profile). While C-States only change when CPU is idle, P-States changes during execution. T-States are "a sort of" P-State, as it changes during execution, they can be OS controllable but often also HW-controllable. They are mostly used during thermal events on the CPU (overheat etc.). P-States change voltage and frequency parameters, but T-States "stops" the CPU for brief moments. The uses and when to go to a certain state is based on (user or powerdevice controlled) power profiles, latency and device interaction etc.

[mitch_de]

I modified the old GenericCPUControl Application to work with this kext also, but it will be deprecated and will not be updated anymore. But you can use it until I get time to create prefpane and statusbartool.

Thanks !

Working for me now !

[Konami®]

Working great now! shutdown and restart works, thanks a lot man!

[adrian mak]

should voodoo kernel be used in order to use voodoopower ?

[Konami®]

should voodoo kernel be used in order to use voodoopower ?

 

I am using vanilla kernel 9.6 from 10.5.6 and works great!

[Superhai]

should voodoo kernel be used in order to use voodoopower ?

 

Why is it so hard to read the first post?

[Maniac10]

My athlon 64 3000+ could not boot, I get an instant kp when the kext is loaded. I'll try to get a camera to take a picture later.

[weaksauce12]

Just tested the v1.2.0 update on my Gigabyte GA-EP45-UD3P with BOOT-132 and Vanilla, works great! No more OpenHaltRestart for me, whoohoo! Thanks Superhai!

[Aranius]

Thanks, works like a charm now

[go2mp]

THx just works fine @ msi wind

[GsRrrr]

On T7200 c2d sony vgn-sz330p/b(laptop)vanilla ipc 10.5.6 installed generic (didnt shut down, no batt info), Openhalt (same result) Voodoo v1.2 No result either.

 

 

I havn't messed with my osx setup for awhile b/c of school. did a fresh install w/ IPC. Am i doing something wrong? Am i suppose to remove the Generic and Openhalt Kext? then install Voodoo?

 

If so how do i remove the generic and openhalt kexts?

 

thanks.