Jump to content

DSDT fixes for Gigabyte boards

iSoprano

By iSoprano,
in Guides

 Share
1,909 posts in this topic

Recommended Posts

mm67
mm67
Posted
Posted
I tried to do it, but the problem remains: (

 

Try this one:

    Scope (_PR)
   {
       Processor (CPU0, 0x00, 0x00000410, 0x06)
       {
           Name (_CST, Package (0x04)
           {
               0x3,
               Package (0x4){ResourceTemplate (){Register (FFixedHW,0x01,0x02,0x000,0x0,)},One,One,0x3E8},
               Package (0x4){ResourceTemplate (){Register (FFixedHW,0x01,0x02,0x010,0x1,)},0x2,One,0x1F4},
               Package (0x4){ResourceTemplate (){Register (FFixedHW,0x01,0x02,0x030,0x3,)},0x4,0x39,0x064}
           })
           Name (_PSD, Package (0x05) {0x05, Zero, Zero, 0xFC, 0x04})
           Name (_PSS, Package (0x04)
           {
               Package (0x06) { Zero, Zero, 10, 10, 0x921, 0x921 },
               Package (0x06) { Zero, Zero, 10, 10, 0x81E, 0x81E },
               Package (0x06) { Zero, Zero, 10, 10, 0x71A, 0x71A },
               Package (0x06) { Zero, Zero, 10, 10, 0x613, 0x613 }
           })
       }
       Processor (CPU1, 0x01, 0x00000410, 0x06)
       {
           Alias (\_PR.CPU0._PSS, _PSS)
           Alias (\_PR.CPU0._CST, _CST)
           Alias (\_PR.CPU0._PSD, _PSD)
       }
       Processor (CPU2, 0x02, 0x00000410, 0x06)
       {
           Alias (\_PR.CPU0._PSS, _PSS)
           Alias (\_PR.CPU0._CST, _CST)
           Alias (\_PR.CPU0._PSD, _PSD)
       }
       Processor (CPU3, 0x03, 0x00000410, 0x06)
       {
           Alias (\_PR.CPU0._PSS, _PSS)
           Alias (\_PR.CPU0._CST, _CST)
           Alias (\_PR.CPU0._PSD, _PSD)
       }
   }

Link to comment
Share on other sites
SunKi
SunKi
Posted
Posted
Try this one:

    Scope (_PR)
   {
       Processor (CPU0, 0x00, 0x00000410, 0x06)
       {
           Name (_CST, Package (0x04)
           {
               0x3,
               Package (0x4){ResourceTemplate (){Register (FFixedHW,0x01,0x02,0x000,0x0,)},One,One,0x3E8},
               Package (0x4){ResourceTemplate (){Register (FFixedHW,0x01,0x02,0x010,0x1,)},0x2,One,0x1F4},
               Package (0x4){ResourceTemplate (){Register (FFixedHW,0x01,0x02,0x030,0x3,)},0x4,0x39,0x064}
           })
           Name (_PSD, Package (0x05) {0x05, Zero, Zero, 0xFC, 0x04})
           Name (_PSS, Package (0x04)
           {
               Package (0x06) { Zero, Zero, 10, 10, 0x921, 0x921 },
               Package (0x06) { Zero, Zero, 10, 10, 0x81E, 0x81E },
               Package (0x06) { Zero, Zero, 10, 10, 0x71A, 0x71A },
               Package (0x06) { Zero, Zero, 10, 10, 0x613, 0x613 }
           })
       }
       Processor (CPU1, 0x01, 0x00000410, 0x06)
       {
           Alias (\_PR.CPU0._PSS, _PSS)
           Alias (\_PR.CPU0._CST, _CST)
           Alias (\_PR.CPU0._PSD, _PSD)
       }
       Processor (CPU2, 0x02, 0x00000410, 0x06)
       {
           Alias (\_PR.CPU0._PSS, _PSS)
           Alias (\_PR.CPU0._CST, _CST)
           Alias (\_PR.CPU0._PSD, _PSD)
       }
       Processor (CPU3, 0x03, 0x00000410, 0x06)
       {
           Alias (\_PR.CPU0._PSS, _PSS)
           Alias (\_PR.CPU0._CST, _CST)
           Alias (\_PR.CPU0._PSD, _PSD)
       }
   }

 

No, it did not help. I tried already a lot of variations of PR but as soon as I get a C4 problem starts with the transition into sleep. I think that perhaps the problem should be sought in changing the method _PTS or something in that spirit.

 

Just in case, attaching my DSDT mini_dsdt.dsl.zip

Link to comment
Share on other sites
mm67
mm67
Posted
Posted
No, it did not help. I tried already a lot of variations of PR but as soon as I get a C4 problem starts with the transition into sleep. I think that perhaps the problem should be sought in changing the method _PTS or something in that spirit.

 

Just in case, attaching my DSDT mini_dsdt.dsl.zip

 

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. I moved the dongle to one of motherboard connectors and now C2:C4 usage levels running Windows look totally different:

post-375259-1262603297_thumb.png

 

This is a quote from ICH10 datasheet:

EHCI_BREAK_EN — R/W.

0 = EHCI traffic will not act as a break event.

1 = EHCI traffic acts as a break event, even if the BM_STS-ZERO_EN and POPUP_EN

bits are set. EHCI master activity will cause BM_STS to be set and will cause a break from C3/C4.

 

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.

Link to comment
Share on other sites
kdawg
kdawg
Posted
Posted
Thanks, mate! Very useful!

I think I can remove the floppy drive.

This is all just a learning exercise for you right? A lot of users have already removed things like Floppy, serial, printer ports etc. A lot of the things you're asking about have already been done or figured out.

 

If you're just trying to understand then have at it.

Link to comment
Share on other sites
SunKi
SunKi
Posted
Posted
Have you tried running at lower speeds ? Maybe there is a reason why some overclockers seem to use only C1E.

 

No, overclocking can cause unstable behavior and not to issue such a stable transition to sleep. Moreover, when properly made PR setpci and the voltage on the CPU (falls to 0.976) that C4E is working and the system behaves stably. Nevertheless, I tried to turn off acceleration, but nothing changed: (

Link to comment
Share on other sites
kdawg
kdawg
Posted
Posted
kdawg, not to throw a wrench in the gears of your theory, but my Gigabyte system with working sleep has iStat menus installed.

 

On my new p55 build, sleep problems are directly related to the optical drive. What a P.I.T.A this whole sleep thing is!

Small update on the auto-sleep issue. Dropbox has released a new version (v0.7.97) that has fixed auto-sleep for me. Still working on the iStat issue.
Link to comment
Share on other sites
rayap
rayap
Posted
Posted
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. I moved the dongle to one of motherboard connectors and now C2:C4 usage levels running Windows look totally different:

post-375259-1262603297_thumb.png

 

This is a quote from ICH10 datasheet:

EHCI_BREAK_EN �€” R/W.

0 = EHCI traffic will not act as a break event.

1 = EHCI traffic acts as a break event, even if the BM_STS-ZERO_EN and POPUP_EN

bits are set. EHCI master activity will cause BM_STS to be set and will cause a break from C3/C4.

 

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.

 

Great stuff mm67! In Ubuntu 9.10 with Bluetooth dongle out and all USB suspended, C3 reaches 100%, power usage down by more than 10 Watts.

In Vanilla SL, now with the CST method which gives "80", even the voltage switching is much improved and power usage similar to "00" The remaining USB devices esp. the Alum Keyboard appear to be hogging the system (reported in Powertop when in Ubuntu).

 

BTW, in Ubuntu Powertop was easy; what did you use in Windows 7? PowerInformer is stuck in Win7 and "perfmon" too tricky for me.

Link to comment
Share on other sites
mm67
mm67
Posted
Posted
Great stuff mm67! In Ubuntu 9.10 with Bluetooth dongle out and all USB suspended, C3 reaches 100%, power usage down by more than 10 Watts.

In Vanilla SL, now with the CST method which gives "80", even the voltage switching is much improved and power usage similar to "00" The remaining USB devices esp. the Alum Keyboard appear to be hogging the system (reported in Powertop when in Ubuntu).

 

BTW, in Ubuntu Powertop was easy; what did you use in Windows 7? PowerInformer is stuck in Win7 and "perfmon" too tricky for me.

 

Yes, voltages don't look as jumpy anymore when system is idling, even temps may have gone down 1-2 degrees. Hard to tell, could also be that room temperature is lower. For Windows readings I did use perfmon :censored2: It's not that hard, click Performance Monitor from left side of screen, monitor window opens. On top of monitor you have green + button, click that, select processor, click Add >> button and OK. Next use the view button on left side of that green + button and select report view. That's it.

 

/edit

The temperatures really do drop 2 degrees when C4 works properly. I can verify this by plugging the bluetooth dongle back to EHCI port, C4 is broken again. Then I open VoodooMonitor to see temps, run sudo setpci -s 0:1f.0 0xab.b=0 to disable EHCI and PCI activity from breaking C4 and after some seconds temps drop about 2 degrees. I'll test some time like this and if I don't see any side effects then I add that as a patch to dsdt.

Link to comment
Share on other sites
greythorne
greythorne
Posted
Posted

guys what do i need if i want to boot into 32bit kernel? You see at chameleon boot prompt i typed arch=i386 so that i could boot into 32bit kernel but when come to ACPI_SMC....plugin the system just hangs there. So do i need some 32bit kext inorder to boot into 32bit?

 

Thanks

Link to comment
Share on other sites
Krokus
Krokus
Posted
Posted
Small update on the auto-sleep issue. Dropbox has released a new version (v0.7.97) that has fixed auto-sleep for me. Still working on the iStat issue.

 

I too am having this problem. I only had sleep in the beginning and have been tweeking and starting over and over for months. I probably didn't have iStat installed at the time when it worked. I only installed it to check the temp. Is there any other way to check cpu temps and uninstall iStat?

Link to comment
Share on other sites
kdawg
kdawg
Posted
Posted
I too am having this problem. I only had sleep in the beginning and have been tweeking and starting over and over for months. I probably didn't have iStat installed at the time when it worked. I only installed it to check the temp. Is there any other way to check cpu temps and uninstall iStat?
VoodooMonitor. Look around you should be able to find it no problem.
Link to comment
Share on other sites
Johnny V
Johnny V
Posted
Posted
Thanks snackole!

 

Temps are lowered to around 50-52C... that's about where I started. PStateChanger is reporting all states values, mV and VID are 1275 and 24 respectively.

 

screenshot20100103at851.th.png

 

 

Not sure why the values are pegged all the same? Is it just my setup or is there a conflict somewhere?

 

Thanks!

 

Well I decided to reload NullCPUPowerManagement.kext, SleepEnabler.kext, VoodooPowerMini.kext and be done with it until I can understand DSDT better. Seems like temps are 3 degrees cooler in the 47C-50C range using snackole's DSDT with the above kexts installed. I'm sure this isn't the way to go but everything, including sleep, restart and shutdown are working.

Link to comment
Share on other sites
mm67
mm67
Posted
Posted
No, overclocking can cause unstable behavior and not to issue such a stable transition to sleep. Moreover, when properly made PR setpci and the voltage on the CPU (falls to 0.976) that C4E is working and the system behaves stably. Nevertheless, I tried to turn off acceleration, but nothing changed: (

 

On my system C4 seems to be directly connected to overclocking, if I have working C4 in dsdt then I can't even boot to OS X using anything higher than 8.5 * 375. Right now I have C4 commented out in dsdt and I am running at 8.5 * 450 with no problems. Nice Geekbench result : http://browse.geekbench.ca/geekbench2/view/202586

Link to comment
Share on other sites
SunKi
SunKi
Posted
Posted
On my system C4 seems to be directly connected to overclocking, if I have working C4 in dsdt then I can't even boot to OS X using anything higher than 8.5 * 375. Right now I have C4 commented out in dsdt and I am running at 8.5 * 450 with no problems. Nice Geekbench result : http://browse.geekbench.ca/geekbench2/view/202586

 

But in my case, all loaded and working steadily in both overclocked and without overclocking with C4. Also go to sleep, it stops halfway through and need an extra kick :) Tried a bunch of options DSDT. Changed the PR and _PTS. The last attempt - to add native DSDT _CST FFixedHW ... If not then my opinion is that the problem is in support of C4 in the BIOS that matherboard. Although .. I can check it at Win7 ;)

 

P.S.: Your result geekbench magnificent !

Link to comment
Share on other sites
cparm
cparm
Posted
Posted

As asked by theking, here is a version of pc_efi 10.5 (based on the AnV's work) with the restart fix 2.1 (but non auto, since i don't know how it work on acpi 2+), that can load a fadt.aml file,

 

plus it can also drop/replace or add, SSDT, HPET, SBST, ECDT, ASFT, DMAR, and of course DSDT tables

 

 

 

Link to comment
Share on other sites
jamonda
jamonda
Posted
Posted
No problems with Dropbox and iStat for me, auto-sleep works fine with those as long as I don't have that damn Samsung DVD connected

 

I don't have Dropbox or iStat menus either, and my EP45-UD3L doesn't sleep.

Do you know anything about Asus DVD drives preventing hacks from auto-sleeping?

Link to comment
Share on other sites
mm67
mm67
Posted
Posted
I don't have Dropbox or iStat menus either, and my EP45-UD3L doesn't sleep.

Do you know anything about Asus DVD drives preventing hacks from auto-sleeping?

 

Never heard anything about those, but it is easy to test if that is causing problems. Just disconnect your drive and check if your system has auto-sleep without it.

Link to comment
Share on other sites
jamonda
jamonda
Posted
Posted
Never heard anything about those, but it is easy to test if that is causing problems. Just disconnect your drive and check if your system has auto-sleep without it.

 

I've just tested. It's not the Asus drive. My bluetooth dongle, maybe?

And can you please help me with C4? This is my current output for reggie_se and setpci commands:

 

Last login: Wed Dec 30 18:24:05 on console

Mac-Pro-P45-2:~ jamonda$ sudo reggie_se -D PhysAddr -r -a 0xe0000000 -B 32

Password:

. 4.7.2 (263)

Memory Address - 0x00000000e0000000 (local)

Address 0xe0000000 0xffffffff |1 1 1 1|1 1 1 1|1 1 1 1|1 1 1 1|1 1 1 1|1 1 1 1|1 1 1 1|1 1 1 1|

Mac-Pro-P45-2:~ jamonda$ setpci -s 0:1f.0 0xa6.b

00

Mac-Pro-P45-2:~ jamonda$

 

An this is what I get with the code you posted do Blackosx:

 

Last login: Mon Jan 4 18:36:20 on console

Mac-Pro-P45-2:~ jamonda$ sudo reggie_se -D PhysAddr -r -a 0xe0000000 -B 32

Password:

. 4.7.2 (263)

Memory Address - 0x00000000e0000000 (local)

Address 0xe0000000 0xffffffff |1 1 1 1|1 1 1 1|1 1 1 1|1 1 1 1|1 1 1 1|1 1 1 1|1 1 1 1|1 1 1 1|

Mac-Pro-P45-2:~ jamonda$ setpci -s 0:1f.0 0xa6.b

80

Link to comment
Share on other sites
mm67
mm67
Posted
Posted
I've just tested. It's not the Asus drive. My bluetooth dongle, maybe?

And can you please help me with C4? This is my current output for reggie_se and setpci commands:

 

Last login: Wed Dec 30 18:24:05 on console

Mac-Pro-P45-2:~ jamonda$ sudo reggie_se -D PhysAddr -r -a 0xe0000000 -B 32

Password:

. 4.7.2 (263)

Memory Address - 0x00000000e0000000 (local)

Address 0xe0000000 0xffffffff |1 1 1 1|1 1 1 1|1 1 1 1|1 1 1 1|1 1 1 1|1 1 1 1|1 1 1 1|1 1 1 1|

Mac-Pro-P45-2:~ jamonda$ setpci -s 0:1f.0 0xa6.b

00

Mac-Pro-P45-2:~ jamonda$

 

An this is what I get with the code you posted do Blackosx:

 

Last login: Mon Jan 4 18:36:20 on console

Mac-Pro-P45-2:~ jamonda$ sudo reggie_se -D PhysAddr -r -a 0xe0000000 -B 32

Password:

. 4.7.2 (263)

Memory Address - 0x00000000e0000000 (local)

Address 0xe0000000 0xffffffff |1 1 1 1|1 1 1 1|1 1 1 1|1 1 1 1|1 1 1 1|1 1 1 1|1 1 1 1|1 1 1 1|

Mac-Pro-P45-2:~ jamonda$ setpci -s 0:1f.0 0xa6.b

80

 

There seems to be a million possible reasons for auto-sleep not working, just ask kdawg :)

 

That 80 tells that you have C4 working.

 

 

No, overclocking can cause unstable behavior and not to issue such a stable transition to sleep. Moreover, when properly made PR setpci and the voltage on the CPU (falls to 0.976) that C4E is working and the system behaves stably. Nevertheless, I tried to turn off acceleration, but nothing changed: (

 

Where do you see this 0.976 value ? You wouldn't happen to have Gigabyte's DES turned on ?

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