Jump to content

AnVAL (ACPI Loader)

valv

By valv,
in Chameleon

 Share
1,538 posts in this topic

Recommended Posts

Azimutz
Azimutz
Posted
Posted

Valv, i hope that my post wasn't somehow the catalyst for your decision!?

I was just trying to point that i don't think you're required to post sources of unfinished work?!

At least, that's what Apple does; they only post the sources when the work is finished to release. So, you shouldn't let this stuff get into you :D

Private testing, mate... or if the problem is with this site, there are others...

Link to comment
Share on other sites
Time2Retire
Time2Retire
Posted
Posted
Valv, i hope that my post wasn't somehow the catalyst for your decision!?

I was just trying to point that i don't think you're required to post sources of unfinished work?!

At least, that's what Apple does; they only post the sources when the work is finished to release. So, you shouldn't let this stuff get into you :(

Private testing, mate... or if the problem is with this site, there are others...

I'm afraid that it was my request for the source code – yeah yeah let's blame the messenger, thank you egberts – but you apparently don't seem to understand when developers must release their modifications, because a finished work is anything you release. That's your product there. This is what requires people to share the source code. The fact that a developer is not happy with its current state doesn't really matter.

 

What I found amusing was you naming Apple, as if they don't share the source (in time). That's just silly. Another weak defense was that it is dangerous to use. Why? Because there's no warning whatsoever that the boot file might fry your CPU. Ah I see. People should blindly trust someone they don't know, and when something goes bad. Well. Bad luck for you.

 

Also. Valv made it look like he is under pressure (can't work in this condition) but other people don't seem to feel any pressure whatsoever. Just look at the Chameleon trunk and the countless other open source projects.

 

p.s. I'm a girl and pretty productive myself, so thank you for your ignorant remarks. I guess you don't read much.

Link to comment
Share on other sites
The Real Deal
The Real Deal
Posted
Posted
Valv, can you implement support for two ati 5770 cards like Kabyl's bootloader? His lets you run two 5770's in crossfire (even though crossfire doesn't work under osx).

 

Basically his will show both cards in "slot-1" upon boot, and shows both in system profiler. Where as your bootloader boots fine, but when osx loads, any attempt to run finder or system preferences, just completely locks up the system.

 

 

I second that for HD 5870! :(

Link to comment
Share on other sites
cparm
cparm
Posted
Posted
Valv, i hope that my post wasn't somehow the catalyst for your decision!?

don't worry, it's not your fault,

 

At least, that's what Apple does; they only post the sources when the work is finished to release. So, you shouldn't let this stuff get into you :(

Private testing, mate... or if the problem is with this site, there are others...

 

First, you got to know that Apple have no obligation to release their sources,

 

second, please note also that even Apple do private testing, and each time they release the gold master, in the worst case, they also publish sources few days later,

 

for me, this rule is good because, nobody know if a developer who use chameleon's sources (or from an other open source project), and share his work in a binary form, will ever release his modifications,

 

in the present case, i'm sure that valv will publish his sources (one day :( ) because i know him, but as someone said before, the rules should be the same for everyone

Link to comment
Share on other sites
MadMungo
MadMungo
Posted
Posted

Hi valv,

 

I have a bit of a problem with 5.0.5 and loading of 5970 graphics driver under 10.6.5, what it dose is goes through the verbose boot and stops at the point where it normally draws/initialises the desktop. I can tell it has loaded, because if I ctrl+alt+delete I can hear the alert sound being played.

 

I have the following graphics options set in the Boot.plist and the video Bios in Extra

AutoResolution=Yes

UseAtiRom=Yes

VBIOS=Yes

 

Thanks.

Link to comment
Share on other sites
RayFlower
RayFlower
Posted
Posted

Using this with graphics enabler resulted in garbage on my screen with my ati 5750 hd with the new kexts.

 

 

I have to use ATI5000Controller.kext, ATIFramebuffer.kext & ATISupport.kext  from the macmini image to get my gpu working, anyone found a way around this?

 

 

 

^that was with version 4,version 5 doesn't work at all for me no display, but i guess that has everything to do with my outdated kext's

Link to comment
Share on other sites
artsin
artsin
Posted
Posted

GERMAN

 

Mit diesen Optionen läuft die HD5770. Seit dem Einbau der HD5770 geht bei mir NICHT mehr der Ruhezustand Modus! Als die alte GPU: NVIDIA-8800GTS 640MB noch Eingebaut war da ging der Sleep Modus noch!

 

AutoResolution=Yes

UseAtiRom=Yes

VBIOS=Yes

 

ENGLISH

 

These options run the HD5770. Since the installation of the 5770's with me no longer the sleep mode! When the old GPU: NVIDIA 8800GTS 640MB Built nor was there was the sleep mode yet!

 

AutoResolution=Yes

UseAtiRom=Yes

VBIOS=Yes

 

 

SYSTEM:

 

 

OS: SnowLeopard 10.6.5 32/64bit

Motherboard: Gigabyte GA-P35-DS4 Rev. 1.0

CPU: Intel Q6600 @ 2.44GHz

GPU: XFX (Retail) HD5770 XXX 1024MB HDMI/DVI/DP

post-228956-1290591993_thumb.png

Link to comment
Share on other sites
mitch_de
mitch_de
Posted
Posted

"Also, as Valv himself pointed out to me, messing with some of the sources - "maxvolt" for instance if unlocked - could result in fried CPUs so a stable version that is then commented accordingly is safer for all."

Not really - CPU doesnt accept values over(under) an cpu specific max(min). That also is guilty for Pstate VID and FID - if to high (or too low to) , MSR writes are ignored by the cpu itself.

BUt for sure: Such things better be pretested by knowledged people here before let it get public for all .

Link to comment
Share on other sites
Breathless
Breathless
Posted
Posted
Hi all, do you have this phenomenon, before the Chameleon boot theme appears, there is a log show like :
ntfs_fixup:magic doesn't match xxxx

, I have search in the google without answer. I do know how this cause, even the unofficial version exist this problem, but when I use RC4 boot version ,it does not appears.

 

After test archive 3 : CStates and PState works great, the PState is five, now it is nine: CPU=Pentium dual core .

 

Here is the log: dmesg.rtf

 

Regards.

 

 

I have the same exact issue. I didn't get this error in RC4.

 

NTFS_Fixup: magic doesn't match "blah blah blah"....

 

When I boot up all my NTFS drives are there no problem. Everything works as before.

 

It doesn't seem that this is a very common issue since there is virtually no info on this in google. Anyone have any ideas?

Link to comment
Share on other sites
Breathless
Breathless
Posted
Posted
@Breathless, did this happened with anval ? if yes, which version ?

 

Which version would you like me to try? Give me a link to one of em and i'll give a test.

 

This is the first i've ever encountered it. I'm not sure if i've tried Anval before.... probably have in all my messing around, but not sure.

Link to comment
Share on other sites
Time2Retire
Time2Retire
Posted
Posted

Some people here have TM1 and/or TM2 disabled in MSR Tools. On the HP G72 notebook I am using TM2 is off so I had a look in the new Intel CPU Monitoring with DTS/PECI PDF and noticed that it should be enabled on the HP, and possibly for other people here as well. This is what I found out:

 

BIOS can enable or disable TM1 and TM2 (software should not disable any of the two).

TM1 is enabled by BIOS setting bit 3 of IA32_MISC_ENABLE.

TM1 was introduced with Intel® Pentium® 4.

TM1 controls the processor core temperature by modulating the duty cycle of processor clock.

TM2 is enabled by BIOS setting bit 13 of IA32_MISC_ENABLE.

TM2 was introduced with Intel® Pentium® M.

TM2 controls the processor temperature by reducing the operating frequency and voltage of the processor.

TMn are two automatic thermal monitoring mechanisms calibrated in the factory, and software cannot alter the trigger conditions.

Chameleon doesn't seem to check and/or enable any of the two.

Chameleon doesn't seem to read IA32_MISC_ENABLE, IA32_THERM_INTERRUPT, IA32_THERM_CONTROL nor IA32_THERM_STATUS.

 

The question now is: Should Chameleon do this?

 

p.s. We need a new / more complete Open Source version of MSR Tools (command line tool). One that shows all cores and uses PECI (when available).

Link to comment
Share on other sites
valv
valv
Posted
  • Author
Posted
Chameleon doesn't seem to read IA32_MISC_ENABLE, IA32_THERM_INTERRUPT, IA32_THERM_CONTROL nor IA32_THERM_STATUS.
IA32_MISC_ENABLE & IA32_THERM_STATUS are already used here!

Also on the local version am using, I've managed to make the boot-loader probe for TM & TM2 presence.

btw,

TM1 is enabled by BIOS setting bit 3 of IA32_MISC_ENABLE.

TM2 is enabled by BIOS setting bit 13 of IA32_MISC_ENABLE.

thank you for the valuable info. To be honest I was looking after those.

 

p.s. We need a new / more complete Open Source version of MSR Tools (command line tool). One that shows all cores and uses PECI (when available).
Once, I've managed to make one working for leopard (32bit). but I remember having bad times with dyld becoming crippled, and the whole system became unstable. The hardest part was to mess with IOKIT. Because, as u may know, rdmsr/wrmsr are privilege level 0 aware.

A good idea would be to get something that 'd also work on snow (32/64).

It'd be great if someone could point us to the source code of MSR Tools (even incomplete) if any.

Link to comment
Share on other sites
mitch_de
mitch_de
Posted
Posted

Yep - some kext Hero which could make new MSRTool would be great. As known , even the 32 Bit Version (running in general) gets instable / stalls the system after some time. Too bad that the kernel istself(or Apple .kext) havent any sysctl -a values listed which shows the changing MHZ - only fixed MHz values are shown (for my knowledge/testings).

Link to comment
Share on other sites
JBraddock
JBraddock
Posted
Posted
IA32_MISC_ENABLE & IA32_THERM_STATUS are already used here!

Also on the local version am using, I've managed to make the boot-loader probe for TM & TM2 presence.

btw, thank you for the valuable info. To be honest I was looking after those.

@Valv, I've been using the boot10 for a while and it seems ok. But, sometimes, during the boot, OSX gets stuck. No spinning ball, no kernel panic. I removed auto resolution and vbios keys from the boot.plist but it still happens randomly.

 

When I boot with -v, it shows that right before power management it gets stuck.

 

Any idea what may cause this? Can anyone confirm this?

Link to comment
Share on other sites
valv
valv
Posted
  • Author
Posted
@Valv, I've been using the boot10 for a while and it seems ok. But, sometimes, during the boot, OSX gets stuck. No spinning ball, no kernel panic. I removed auto resolution and vbios keys from the boot.plist but it still happens randomly.

 

When I boot with -v, it shows that right before power management it gets stuck.

 

Any idea what may cause this? Can anyone confirm this?

This seems to be related to usb, and more specifically to hard bios reset. You can fix it by adding the following key to your boot.plist:

	<key>USBBusFix</key>
<string>Yes</string>

btw, don't forget to remove any other uhci/ehci related keys.

Link to comment
Share on other sites
JBraddock
JBraddock
Posted
Posted
This seems to be related to usb, and more specifically to hard bios reset. You can fix it by adding the following key to your boot.plist:
	<key>USBBusFix</key>
<string>Yes</string>

btw, don't forget to remove any other uhci/ehci related keys.

I have these keys in the boot.plist which are related to usb. Should I remove the other two or leave them as they are. Since they are set to No, do they still need to be removed?

		<key>UHCIreset</key>
<string>No</string>
<key>USBBusFix</key>
<string>No</string>
<key>USBLegacyOff</key>
<string>No</string>

 

In my previous boot.plist which I started to use upon your suggestions, we had the followings.

<key>USBBusFix</key>
<string>Yes</string>
<key>UHCIreset</key>
<string>Yes</string>
<key>USBLegacyOff</key>
<string>Yes</string>
<key>EHCIacquire</>
<string>Yes</string>

 

So, what should I use?

Link to comment
Share on other sites
valv
valv
Posted
  • Author
Posted
I have these keys in the boot.plist which are related to usb. Should I remove the other two or leave them as they are. Since they are set to No, do they still need to be removed?

		<key>UHCIreset</key>
<string>No</string>
<key>USBBusFix</key>
<string>No</string>
<key>USBLegacyOff</key>
<string>No</string>

yes, remove the first and last keys, and replace USBBusFix value with a Yes
In my previous boot.plist which I started to use upon your suggestions, we had the followings.

<key>USBBusFix</key>
<string>Yes</string>
<key>UHCIreset</key>
<string>Yes</string>
<key>USBLegacyOff</key>
<string>Yes</string>
<key>EHCIacquire</>
<string>Yes</string>

 

So, what should I use?

USBBusFix takes care (so to say) of all those keys (now). That is, when it is enabled: all the others get a Yes!
Link to comment
Share on other sites
JBraddock
JBraddock
Posted
Posted
yes, remove the first and last keys, and replace USBBusFix value with a Yes

USBBusFix takes care (so to say) of all those keys (now). That is, when it is enabled: all the others get a Yes!

Ok I did as you suggested. I'll let you know if the problem is solved. Need some testing.

Link to comment
Share on other sites
Time2Retire
Time2Retire
Posted
Posted
IA32_MISC_ENABLE & IA32_THERM_STATUS are already used here! Also on the local version am using, I've managed to make the boot-loader probe for TM & TM2 presence. btw, thank you for the valuable info. To be honest I was looking after those.

You're welcome. I found them in xnu/osfmk.cpuid.h as 

#define CPUID_FEATURE_TM	  _Bit(29)	/* Thermal Monitor (TM1) */
#define CPUID_FEATURE_TM2	 _HBit(8)	/* Thermal Monitor 2 */

 

Once, I've managed to make one working for leopard (32bit). but I remember having bad times with dyld becoming crippled, and the whole system became unstable. The hardest part was to mess with IOKIT. Because, as u may know, rdmsr/wrmsr are privilege level 0 aware...

Ah so that might explain why my command line tool fails with a segmentation fault. Need to dig into this matter at a later stage. Thanks.

 

BTW: Can this info be found with sysctl: 

p->CPU.MaxCoef, p->CPU.MaxDiv, p->CPU.CurrCoef and p->CPU.CurrDiv

TIA

Link to comment
Share on other sites
 Share
Go to topic listing
×
×
  • Create New...