DSDT fixes for Gigabyte boards

[FKA]

That page seems dead, I haven't seen any updates there for months.

 

looking for something similar on the english GB forum but can't see anything.

The tweaktown page blackosx has posted appears quite recent!

[mm67]

looking for something similar on the english GB forum but can't see anything.

The tweaktown page blackosx has posted appears quite recent!

 

Tweaktown should have the most recent versions. On german side JZ Board seems to have most recent versions.

http://www.jzelectronic.de/jz2/board.php?s...zeige&wo=28

[FKA]

Hi FormerlyKnownAs, did you find solution by any chance ?

anybody here know how to set up fixed disk order in OSX ?

 

NO - I believe this is native behaviour for Macs.

 

--- -- ----- ------ -

 

BTW - now awaiting delivery of an MSI P45 board

[kdawg]

NO - I believe this is native behaviour for Macs.

 

--- -- ----- ------ -

 

BTW - now awaiting delivery of an MSI P45 board

Nice. Saying goodbye to Gigabyte?

[FKA]

Nice. Saying goodbye to Gigabyte?

 

il est possible ! let you know how it goes ..

[Sull]

It would help if you posted the dsdt you are having trouble with.

 

I have trimmed down just like your dsdt and still get the compile error. I actually made it pretty much identical......

 

Here is it take a look. I really want to learn why this is happening..

 

Thanks

 

PS. I quickly tested your dsdt and everything worked fine. I noticed you cancelled out the internal SATA fix, I kept my graphics and my audio...... also may need some help doing my own c states..... I don't know much here.... I do know without disabler my temps were about 50C and they should be 30-32C idle, with your dsdt they were exactly that. Does Q9400 and Q9550 or whatever you have different states I assume...?

sull.dsl.zip

[keeza]

Here is it take a look. I really want to learn why this is happening..

Your problem was in your Lan addition. Compare your before and after dsdt and you'll see what I mean.

PS. I quickly tested your dsdt and everything worked fine. I noticed you cancelled out the internal SATA fix, I kept my graphics and my audio...... also may need some help doing my own c states..... I don't know much here.... I do know without disabler my temps were about 50C and they should be 30-32C idle, with your dsdt they were exactly that. Does Q9400 and Q9550 or whatever you have different states I assume...?

I'm not using the internal sata fix at the moment. P-states are processor specific. I used voodoomonitor to get my p-states, others have been using PState Changer which is supposed to be more accurate.

 

Quick Update on sleep timing:

Regarding my previous post here it turns out that the delay we seem to experience is not a result of sleep not working properly (as is suspected) as it appears that my imac experiences the same delay. Others also report the sleep delay before system sleep.

A google search also brought up possible reasons as to why the computer doesn't sleep when it should.

 

This article shed some light onto the subject for those that are interested.

 

What I would like to know:

Is there an app or utility that could monitor all computer activity prior to a full system sleep, that could capture all events, then I may be able to see what causes the sparodic variations in system sleep timings.

Sull_fixed.zip

[Sull]

@keeza, thank you! I really appreciate this. I tried finding the difference with my iPhone, but I found it a little hard. I'll check it out when I get home after work and I am no mobile

 

I am familiar with voodoo monitor. Though does it matter if I am running Disabler.kext or apples native powermanagent when checking p states ? P states are set by the processor so this should not matter right...

 

I also don't exactly know to to change p states in dsdt. Any tips ?

[keeza]

@keeza, thank you! I really appreciate this. I tried finding the difference with my iPhone, but I found it a little hard. I'll check it out when I get home after work and I am no mobile

 

I am familiar with voodoo monitor. Though does it matter if I am running Disabler.kext or apples native powermanagent when checking p states ? P states are set by the processor so this should not matter right...

 

I also don't exactly know to to change p states in dsdt. Any tips ?

You're welcome.

 

If you are using my dsdt then you can rip out disabler.kext as it has Device HPET fix applied. You need Appleintelpowermanagement kext to do its job if you want vanilla speedstepping.

P-states are hard coded into the processor, yes.

FormerlyKnownAs has an excellent guide on how to get p-state info over at the vanilla speedstepping thread here

- it contains all you need to know. You can then use your dsdt and replace the Name_PSS section with your processor P-States

 

BTW - now awaiting delivery of an MSI P45 board

 

don't care that USB 3 is here?

[leslieking]

Hi,

What does this scope do in Device (LPCB) ?

 

Scope (\)
{
OperationRegion (SCPP, SystemIO, 0xB2, One)
Field (SCPP, ByteAcc, NoLock, Preserve)
{
	SMIP,   8
}
}

 

Removing it cause a kernel panic. I just want to know what does it do..

[kdawg]

il est possible ! let you know how it goes ..

I was pouring over someone elses DSDT for their MSI P55 mobo and it seem to be very bloated with a lot of unused stuff. For example serial and GFX devices that don't exist. We'll have to start a MSI DSDT thread.

[waldo_pepperjack]

Have you tried running at lower speeds ? Maybe there is a reason why some overclockers seem to use only C1E.

 

I managed find to reason that broke C4 on my Gigabyte board. I had a bluetooth dongle connected to a USB 2.0 hub and I had already earlier noticed that dongle's led seems to be flashing all the time. ...

 

As that bit seems to be set it is easy to see why that constantly flashing bluetooth dongle breaks C4 when it is connected to EHCI port.

 

mm67, can you tell us what brand of BT dongle that was? (we can avoid it ...)

[leslieking]

I solved the SMBus kernel panic issue, after removing Device (PSMR) from Device (LPCB)..

I think, the problem was Device PSMR.

 

This is my SMBus code:

 

Device (SBUS) //<-- Renamed from PX43
{
   Name (_ADR, 0x001F0003)
   Device (BUS0)
   {
       Name (_CID, "smbus")
       Name (_ADR, Zero)
       Device (DVL0)
       {
           Name (_ADR, 0x57)
           Name (_CID, "diagsvault")
       }
   }
}

 

Ok, but $kextstat | grep SMBus doesn't show any kext loaded. Any idea why does this happen?

 

86:~ zoliky$ kextstat | grep SMBus
86:~ zoliky$

[waldo_pepperjack]

Two questions to throw out for you all... I've read what I could find, including the APIC specs and the PNP devids ... seen some references to this here and in the P5K PRO install thread also...

 

1. I'm asking for a sanity check as I remove devices, while I try to cull back my dsdt to a more minimal subset.

Does this make sense, on a GA-EP45-UD3L board? :

==========================================

 

 

I have removed devices and any calls/call structures to:

...............

• LPT*
  
• UAR*
  
• ECP*
  
• FDC*
  
• PS2* (I have no PS2 periphs)
  

 

 

 

Can these be removed similarly, with relative impunity? I'm leaning towards yes, but still wondering:

................

• SPKR (BIOS speaker) - used by Parallels Desktop, but I use VirtualBox
  
• COPR (FPU/math coprocessor)
  
• PIC (legacy IRQ)
  

 

 

Any more (very) easily removable ones that I've forgotten?

 

 

 

 

==========================================

2. Has anyone experimented with either of these?

A. Sensing the fan speed using Device (FAN)

 

 

B. Getting the Power Button to work correctly (maybe bringing up the Shutdown Menu) by changing the adding a _UID entry ... ... or something...

I've googled quite a bit, read what I've found, and it seems like no one has found a way to make PWRB behave in a truly Apple-esque manner, presenting the user w/ the Shutdown Menu. Insta-sleep (_HID -> _CID in PWRB) is not bad, but not stock behavior.

[kdawg]

Anyone have any ideas how to access bit 18 of this register:

PxCMD—Port [5:0] Command Register (D31:F2)

 

On my MSI board value of that bit is 0 and it doesn't need IOAHCIBlockStorageInjector.kext. On my Gigabyte board that bit is set and the kext file is necessary. Only problem is that the bit is marked as R/WO, maybe there is no other way to modify it but to modify bios ?

Would it be bit 18 (Hot Plug Capable Port) or 21 (External SATA Port) of D31:F2 that would be they key? I see they're both R/WO however.

 

Would it look something like this if it were writable? I guess there would be 6 of them one for each offset or SATA port?

OperationRegion (SAT0, PCI_Config, 0x118, 0x04)
Field (SAT0, ByteAcc, NoLock, Preserve) 
{
		,   18,
	HPCP,   0 //<-- Hot Plug Capable Port  p.553
}

[keeza]

B. Getting the Power Button to work correctly (maybe bringing up the Shutdown Menu) by changing the adding a _UID entry ... ... or something...

I've googled quite a bit, read what I've found, and it seems like no one has found a way to make PWRB behave in a truly Apple-esque manner, presenting the user w/ the Shutdown Menu. Insta-sleep (_HID -> _CID in PWRB) is not bad, but not stock behavior.

Control + eject brings up that menu on an apple keyboard. As the apple keyboard doesn't have a power button there isnt a way to map that menu to the keyboard power button with the apple keyboard.

 

As far as that menu working on my imac - I cant get it to work via my power button

[mm67]

Would it be bit 18 (Hot Plug Capable Port) or 21 (External SATA Port) of D31:F2 that would be they key? I see they're both R/WO however.

 

Would it look something like this if it were writable?

OperationRegion (SAT0, PCI_Config, 0x1F, 0x04)
Field (SAT0, ByteAcc, NoLock, Preserve) 
{
		,   18,
	HPCP,   0 //<-- Hot Plug Capable Port  p.553
}

 

Bit 18 is the key, bit 21 is not even set. And you can't access those bits quite like that, they are on memory mapped registers.

[kdawg]

Bit 18 is the key, bit 21 is not even set. And you can't access those bits quite like that, they are on memory mapped registers.

So We'd need to know the memory locations in order to set them?

[mm67]

So We'd need to know the memory locations in order to set them?

 

Yes, but that's not the problem. Problem is that those bits can only be written once and bios does that one time already.

[LocusOfControl]

Hi

 

Just for info, if anyone is having sleep problems

 

check for

 

Jan 13 10:08:00 LocusOfControl configd[13]: PowerManagement configd: System Sleep prevented by active remote login session (600 second threshold).
Jan 13 10:08:00 LocusOfControl configd[13]: PowerManagement configd: tty sleep preventer: /dev/ttys000
Jan 13 10:08:00 LocusOfControl configd[13]: PowerManagement configd: System Sleep prevented by active remote login session (600 second threshold).
Jan 13 10:08:00 LocusOfControl configd[13]: PowerManagement configd: tty sleep preventer: /dev/ttys000

 

 

in system.log. I got this following an update of remote desktop (may only happen if you have another computer on

the same network as I just booted a windows pc with bonjour installed)

 

The solution is go to energy saving and turn off wake for ethernet access (I agree with kdawg that sometimes dropbox

seems to cause issues as well)

[FKA]

The solution is go to energy saving and turn off wake for ethernet access (I agree with kdawg that sometimes dropbox

seems to cause issues as well)

 

unless of course you want your machine to WOL !

[kdawg]

unless of course you want your machine to WOL !

I reverted back to the vanilla networking kext and started using the "Wake for network access" option, or whatever it's called. Right now idle-sleep, Bonjour and WOL all work well.

 

@FKA I know you have problems with your networking. Sorry man. BUT, you got that shiny new MSI mobo on the way!

[FKA]

I reverted back to the vanilla networking kext and started using the "Wake for network access" option, or whatever it's called. Right now idle-sleep, Bonjour and WOL all work well.

 

@FKA I know you have problems with your networking. Sorry man. BUT, you got that shiny new MSI mobo on the way!

 

I'm actually good .. using vanilla kext, WOL OS X option and wake on ring option in BIOS!

The latter seemed to do the trick, also set Gigi and full duplex manualy in network prefs'.

 

still the old girl will most likely be on ebay soon

[sk1nhead]

I reverted back to the vanilla networking kext and started using the "Wake for network access" option, or whatever it's called. Right now idle-sleep, Bonjour and WOL all work well.

 

@FKA I know you have problems with your networking. Sorry man. BUT, you got that shiny new MSI mobo on the way!

 

regarding WOL, what do you use for bonjour? ifconfig? i might try using the vanilla kext but cant live w/o bonjour because of my TimeCapsule. Right now i'm using RealtekR1000SL but we all know WOL option will be gone because of this kext.

 

BTW on my system idle sleep is working 50/50 percent of the time - sometimes it'll work and sometimes it's out of sync (eg set to 5 min and it will sleep more than that), using an LG GH22LS50 and even updated the firmware to TL02 but still the same.. so scour around and found an old solution here which is insert a media disc (dvd/Cd) and voilla! idle sleep works %100 percent of time..

[FKA]

regarding WOL, what do you use for bonjour? ifconfig? i might try using the vanilla kext but cant live w/o bonjour because of my TimeCapsule. Right now i'm using RealtekR1000SL but we all know WOL option will be gone because of this kext.

 

BTW on my system idle sleep is working 50/50 percent of the time - sometimes it'll work and sometimes it's out of sync (eg set to 5 min and it will sleep more than that), using an LG GH22LS50 and even updated the firmware to TL02 but still the same.. so scour around and found an old solution here which is insert a media disc (dvd/Cd) and voilla! idle sleep works %100 percent of time..

 

Try it vanilla .. the last bonjour SW update may have solved that problem for you.