1109 replies to this topic

[Master Chief]

scrax, on Mar 9 2010, 04:25 AM, said:

Can be also the quad core CPU instead of mine dualcore, maybe in part, or not?

Must be.

scrax, on Mar 9 2010, 04:25 AM, said:

I'll try to reduce mine /S/L/E folder...

Good.  It helped here so...

scrax, on Mar 9 2010, 04:25 AM, said:

Yes it's the screen resolution at boot but with Chameleon adding the right resolution worked,  with Revolution v0.6.20 at boot it's the apple that is not in center but the resolution looks right, desktop resolution is ok.

Please check function showBootLog in bootlogo.c and open boot.h where you should find:

#define APPLE_LOGO_X	736 // ((1600 - 128) / 2)
#define APPLE_LOGO_Y	536 // ((DEFAULT_SCREEN_HEIGHT - 128) / 2)

#define APPLE_LOGO_WIDTH  128
#define APPLE_LOGO_HEIGHT 128

Set for my boot screen resolution of 1600 x 1200.  You do have this in your version of Revolution, do you?

[scrax]

in bootlogo.c my function showBootLogo is this:

void showBootLogo()
{
	uint8_t *bootImageData = NULL;
	uint8_t *appleBootLogo = (uint8_t *) decodeRLE(appleLogoRLE, 686, 16384); 
	convertImage(APPLE_LOGO_WIDTH, APPLE_LOGO_HEIGHT, appleBootLogo, &bootImageData);

	uint16_t x = APPLE_LOGO_X; // (VIDEO(width) - 128) / 2; 
	uint16_t y = APPLE_LOGO_Y; // (VIDEO(height) - 128) / 2; 

	drawDataRectangle(x, y, APPLE_LOGO_WIDTH, APPLE_LOGO_HEIGHT, bootImageData);

	free(bootImageData);
	free(appleBootLogo); 
}

and in bootlogo.h I have only this:

#define MIN(x, y) ((x) < (y) ? (x) : (y))
#define MAX(x, y) ((x) > (y) ? (x) : (y))

#define VIDEO(x) (bootArgs->Video.v_ ## x)

const unsigned char appleLogoRLE[] =
{
	0xff, 0x01, 0xff, 0x01, 0xff, 0x01, 0x50, 0x01, 0x01, 0x02, 0x01, 0x17,
	....
	0xff, 0x01, 0xb3, 0x01
};

[Master Chief]

scrax, on Mar 10 2010, 12:34 AM, said:

in bootlogo.c my function showBootLogo is this... and in bootlogo.h I have only this:

Please check boot.h and fix the values, or reinstate the old code from a previous version.

[scrax]

Ok, now it's perfect in the middle.
Thank's

[Beerkex'd]

Master Chief, on Feb 26 2010, 11:02 AM, said:

While this works, it is far too complicated.  Period.  Thus what we need is a simple tool to dump and convert our SMBIOS data.  I would opt for an OS X specific tool.  A tool that let's us work with the SMBIOS data [add/change/remove] would be a plus to me.  That would be cool.  Any takers?

Here's something that might be useful: http://www.insanelym...howtopic=153582

[Master Chief]

Beerkex'd, on Mar 18 2010, 03:01 PM, said:

Here's something that might be useful: http://www.insanelym...howtopic=153582

Nice.

Thank you for the link

[Beerkex'd]

You're welcome.

For those who prefer the point-and-click way, EvOToolsX from our friends in Spain can also dump DMI info:

http://www.osx86.es/?p=755

[bossob]

Hi all. is it still in advance user state only.

[Beerkex'd]

Read the first post and tell us what you think.  

[fc bayern]

After 10.6.3 update...

I have a slower boot time.(earlier was 12 spins, now is 21 spins after the 10.6.3)

When I remove AppleHDA.kext - no matter which version (10.6.2 or 10.6.3, the same rezult) , the boot time is 12 spins again.
As soon as I put AppleHDA.kext in S/L/E, the boot time get slower , and it is 21 spins
Strange, a?

[bossob]

Beerkex'd, on Mar 30 2010, 04:53 AM, said:

Read the first post and tell us what you think.  

It was just a question.because it was dead in here for a long time

[scrax]

fc bayern, on Mar 30 2010, 05:44 PM, said:

After 10.6.3 update...

I have a slower boot time.(earlier was 12 spins, now is 21 spins after the 10.6.3)

When I remove AppleHDA.kext - no matter which version (10.6.2 or 10.6.3, the same rezult) , the boot time is 12 spins again.
As soon as I put AppleHDA.kext in S/L/E, the boot time get slower , and it is 21 spins
Strange, a?

With 10.6.2 I was using the AppleHDA.kext submitted by MC in the P5K Pro thread, it was edited to have a faster boot time, with 10.6.3 I've reinstalled it 'cause the AppleHDA.kext used in 10.6.3 wasn't working (I have integrate also the legacy ket needed to make it work)

[fc bayern]

scrax, on Apr 4 2010, 07:36 PM, said:

With 10.6.2 I was using the AppleHDA.kext submitted by MC in the P5K Pro thread, it was edited to have a faster boot time, with 10.6.3 I've reinstalled it 'cause the AppleHDA.kext used in 10.6.3 wasn't working (I have integrate also the legacy ket needed to make it work)

I copy all the 10.6.2 AppleHDA.kext into 10.6.3 AppleHDA.kext , exept the info and versio.plist, and get back to 12 spins

[scrax]

fc bayern, on Apr 4 2010, 07:39 PM, said:

I copy all the 10.6.2 AppleHDA.kext into 10.6.3 AppleHDA.kext , exept the info and versio.plist, and get back to 12 spins

Nice idea, i'll try this method  

[ntsmkfob]

I've been following this thread with interest. My system is a Badaxe2 with Q6600 overclocked to 3GHz. SL10.6.3 is on an SSD and the configuration I had (AseseBLN1.19 bootloader with nullcpu/fakesmc/sleepenabler/AHCIPortInjector in E/E and VoodooPowerMini in S/L/E was taking between 5.20 and 6.05 revs to boot (each segment as 5 minute chunk). I've been playing with my DSDT fixed a few issues but never got speedstep working until now, so kept reverting back to the above setup.

There are quite a few samples of working PR DSDT sections for Q6600 and I think I've tried most of them. Now I just have fakesmc/AHCI... in E/E and no VoodooPowerMini in S/L/E. AppleIntelCPUManagement is starting up, AppleLPC is loading and now I'm down to less than 1 rev - usually 0.50-0.55 - superfast. Sleep and shutdown/restart work fine. The problem is LPC loads but isn't initialised, so the CST isn't set up. CPU temps have risen about 8-10 deg - not to horrendous levels from around 36-42 to 47-52 in normal use, so it seems that VoodooPowerMini was enabling CST, as the P state values are pretty much identical. DropSSDT=Yes doesn't make any difference to the problem - I still get the 'ACPI_SMC_PlatformPlugin::pushCPU_CSTData - _CST evaluation failed' message in the kernel log.

In the hope that someone can cast an eye over my DSDT and tell me where I've gone wrong, I've attached it below. Obviously, the problem may not be in the DSDT, but at least it will steer me to looking elsewhere.

Something else is that CPU0 is consistently 5/6 deg hotter than the rest. Is this normal?

Attached Files

•  dsdt.zip   14.01K   5 downloads

[mastershredder]

ntsmkfob

I have a X48 based board with Q6600 also overclocked to 3GHz i have looked at your DSDT and added what i changed to make the SpeedStep work on my board a few things to note are that i believe you  should have smbios plist set to mac pro 3.1 otherwise the speedstep wont work properly also since we are running different intel based chipsets and also different q6600's the voltages that mine are set to may make yours unstable as i actually have set mine to custom  undervolted voltages as my chip turned out to be a good-un so if any instability occurs just substitute the voltages and you should be set i believe the VID for the q6600 G0 revision is set at 1.2875 according to my motherboard however i can overclock it to 3GHz and undervolt it to 1.21675 and it is absolutely stable yours may not be also yours may be the B3 stepping of the Q6600 which ran hotter and probably takes more voltage to reach 3GHz however give this a go and report back there is no harm in trying as the voltages are lower than your board would set at stock to be safe i have the DSDT voltages set to 1.292 which should be very compatible with all but the worst batches of Q6600's

Attached Files

•  DSDTQ66003GHZ.zip   13.89K   16 downloads

[ntsmkfob]

Great! Thanks for your help mastershredder. I'll start playing over the weekend. I have MacPro3,1 set already.

[ntsmkfob]

Weird! I put a long reply in here that seems to have disappeared. Anyway, I am pretty sure the _PR section of the DSDT is being ignored, as if I add DropSSDT=Yes into cabp, speedstep stops working - speed and voltages are maxed out. This is with any of the several _PR sections I've tried. Otherwise, speedstep works, but the voltages didn't change when I used your DSDT. Other stuff from the DSDT is working fine, and I can easily screw the boot process by getting something wrong in the  DSDT.

Thanks for your help.

[Master Chief]

ntsmkfob, on Apr 19 2010, 12:21 PM, said:

Weird! I put a long reply in here that seems to have disappeared. Anyway, I am pretty sure the _PR section of the DSDT is being ignored, as if I add DropSSDT=Yes into cabp, speedstep stops working - speed and voltages are maxed out. This is with any of the several _PR sections I've tried. Otherwise, speedstep works, but the voltages didn't change when I used your DSDT. Other stuff from the DSDT is working fine, and I can easily screw the boot process by getting something wrong in the  DSDT.

Thanks for your help.

Is that [probably] because this isn't about Intel's SpeedStep nor DSDT hacking?

[petopay]

How sure are you this works?