Jump to content

GA-EX58 and GA-X58A DSDT native power management modifications


  • Please log in to reply
1476 replies to this topic

#1241
Dajjal

Dajjal

    InsanelyMac Protégé

  • Members
  • PipPip
  • 72 posts
  • Gender:Male
  • Location:WI
Before you do twaektown has F7 while Gigabyte has F7J .......
i have also solved my issue so a note to others if this happens to you.The Realtek RTG 2.0.6 driver is what broke my Sleep/Shutdown! Removing this brings my system back to it's former Glory...........Bless

#1242
Rankrotten

Rankrotten

    InsanelyMac Protégé

  • Members
  • PipPip
  • 52 posts
  • Location:UK

Before you do twaektown has F7 while Gigabyte has F7J .......
i have also solved my issue so a note to others if this happens to you.The Realtek RTG 2.0.6 driver is what broke my Sleep/Shutdown! Removing this brings my system back to it's former Glory...........Bless


I used the @bios tool to update the motherboard and it grabbed version F7 from Gigabyte as the latest, no indication of a beta (F7J) so I'll presume it's the same as the version on tweaktown?

On startup it updated the on-chip SATA3 firmware.

Tried the eSATA and it was still showing Unknown AHCI Standard Controller at 1.5 Gigabit negotiated speed.

Installed the JMicron 36xx package and it now shows thus:

Attached File  Screen_shot_2011_06_29_at_01.01.03.png   47KB   78 downloads

But a hot swap has the same effect on my machine as on yours and the negotiated link falls back to 1.5 Gigabits. Very strange.

Attached File  Screen_shot_2011_06_29_at_01.13.25.png   45.66KB   32 downloads

A final thought, is the eSATA chip not a JMB362 as opposed to the JMB363 as listed in the profiler?

#1243
Dajjal

Dajjal

    InsanelyMac Protégé

  • Members
  • PipPip
  • 72 posts
  • Gender:Male
  • Location:WI
I dont know what's the difference between the 2 Bios updates to be honest, all i found was that It added 3Tb support.....
Well according to the Board Spec's it's JMB362 so i i'm not sure why it says JMB363.......

I did some Test which i posted over at TonyM's but i could not really say if the speed was drastically Decreased cause the test's varied a bit before and after.......
http://www.tonymacx8...php?f=7&t=21444

#1244
FUT1L1TY

FUT1L1TY

    InsanelyMac Geek

  • Members
  • PipPipPip
  • 110 posts
Hello Rankrotten and Dajjal,

In my experience trying to get hot swap to work, I would get 3.0 Gigabit negotiated speed only if the drive was mounted at startup. Any drive the was mounted or remounted at a later point would connect at 1.5 Gigabit. This drove me crazy so I stopped using it. I was however able to get true hotswap (except for the full speed part) using this thread and JMB36X chipset.

I inserted this code at the end of Device (PEX0)

[codebox] Device (EATA)
{
Name (_ADR, Zero)
OperationRegion (BAR0, PCI_Config, 0x40, 0x04)
Field (BAR0, WordAcc, NoLock, Preserve)
{
Offset (0x00),
M1, 8,
Offset (0x01),
M2, 8,
Offset (0x02),
M3, 8
}

Method (_DSM, 4, NotSerialized)
{
Store (0xb3, M1)
Store (0xa1, M2)
Store (0xc2, M3)
Return (0x00)
}
}

Device (EPTA)
{
Name (_ADR, One)
OperationRegion (BAR0, PCI_Config, 0x40, 0x04)
Field (BAR0, WordAcc, NoLock, Preserve)
{
Offset (0x00),
M1, 8,
Offset (0x01),
M2, 8,
Offset (0x02),
M3, 8
}

Method (_DSM, 4, NotSerialized)
{
Store (Zero, M1)
Store (Zero, M2)
Store (Zero, M3)
Return (Zero)
}
}
[/codebox]

Unfortunately, I can't remember if I was using a PCIe card with a Jmicron chip or the onboard chip. The above code will work with either, it's just a matter of inserting it into to the proper location/address space within the DSDT.

EDIT: The above DSDT edit is for JMB 36x chipsets only where true hot swap is not working. True hot swap is defined by the ability to mount a drive, physically eject it, and insert a different drive on the same port and have it automatically mount. If hot swap is already working for you on your JMB 36x ports then you do not need this edit. Drives mounted in this fashion will connect at 1.5 Gigabits regardless of their interface speed. This is not fixable as far as I can tell. Finally, the location of the above edit varies based on your motherboard. Do not insert it without first checking the address of your Jmicron chipset(s).

#1245
Rankrotten

Rankrotten

    InsanelyMac Protégé

  • Members
  • PipPip
  • 52 posts
  • Location:UK
Hi FUT1L1TY

Tried your DSDT edits but it changed my main SATA ports from ICH10 to ESB2 AHCI which is not correct for my board. It's probably only cosmetic as the data transfer rate was the same as before.

Hotswap was never a problem for the eSATA on my machine and it still works fine.

Thanks for the info and I'll have a further look at the edits.

#1246
aikidoka25

aikidoka25

    InsanelyMac Sage

  • Members
  • PipPipPipPipPip
  • 358 posts
UPDATE: Problem solved.
I created another account and copied dock preferences from this new account into the current one.
I guess it was nothing to do with 10.6.8.
Thanks guys.


After upgrading to 10.6.8, now I noticed in Activity Monitor the Dock process hovers around 100% of CPU.
Anyone has the same issue?
Any pointers for me on what to check?

Attached File  Screen_shot_2011_06_30_at_12.13.30_PM.png   164.78KB   40 downloads

#1247
FUT1L1TY

FUT1L1TY

    InsanelyMac Geek

  • Members
  • PipPipPip
  • 110 posts

Hi FUT1L1TY

Tried your DSDT edits but it changed my main SATA ports from ICH10 to ESB2 AHCI which is not correct for my board. It's probably only cosmetic as the data transfer rate was the same as before.

Hotswap was never a problem for the eSATA on my machine and it still works fine.

Thanks for the info and I'll have a further look at the edits.


Please read my edit to my original post.

#1248
Dajjal

Dajjal

    InsanelyMac Protégé

  • Members
  • PipPip
  • 72 posts
  • Gender:Male
  • Location:WI
I just thought i should Post this here instead of the Lion thread since it's kinda more DSDt i guess :blink: .Can anyone else confirm that this is what their Logs say?
10.6.8

[ PCI configuration begin ]
localhost kernel[0]: console relocated to 0xf10000000
localhost kernel[0]: PCI configuration changed (bridge=4 device=3 cardbus=0)
localhost kernel[0]: [ PCI configuration end, bridges 10 devices 33 ]
localhost kernel[0]: AppleIntelCPUPowerManagement: Turbo Ratios 0001
localhost kernel[0]: AppleIntelCPUPowerManagement: (built 16:44:42 Jun 7 2011) initialization complete

10.7 GM

[ PCI configuration begin ]
localhost kernel[0]: console relocated to 0xf10000000
localhost kernel[0]: PCI configuration changed (bridge=4 device=3 cardbus=0)
localhost kernel[0]: [ PCI configuration end, bridges 10 devices 33 ]
localhost kernel[0]: AppleIntelCPUPowerManagement: Turbo Ratios 1112
localhost kernel[0]: AppleIntelCPUPowerManagement: (built 13:08:12 Jun 18 2011) initialization complete



#1249
JamesLittler

JamesLittler

    InsanelyMac Legend

  • Members
  • PipPipPipPipPipPipPipPip
  • 1,022 posts
  • Gender:Male
  • Location:Portsmouth, England
I've just noticed that after updating to 10.6.7 back in March I'm not getting turbo ratios.
This is my kernel log when searching the string 'local kernel[0]: AppleIntel':
Mar  1 23:47:23 localhost kernel[0]: AppleIntelCPUPowerManagementClient: ready
Mar  8 01:05:27 localhost kernel[0]: AppleIntelCPUPowerManagementClient: ready
Mar  9 00:30:54 localhost kernel[0]: AppleIntelCPUPowerManagementClient: ready
Mar  9 18:15:48 localhost kernel[0]: AppleIntelCPUPowerManagementClient: ready
Mar  9 18:40:18 localhost kernel[0]: AppleIntelCPUPowerManagementClient: ready
Mar 28 22:26:52 localhost kernel[0]: AppleIntelCPUPowerManagement: Turbo Ratios 1112
Mar 28 22:26:52 localhost kernel[0]: AppleIntelCPUPowerManagement: initialization complete
Mar 28 22:26:52 localhost kernel[0]: AppleIntelCPUPowerManagementClient: ready
Mar 28 22:37:01 localhost kernel[0]: AppleIntelCPUPowerManagement: Turbo Ratios 1112
Mar 28 22:37:01 localhost kernel[0]: AppleIntelCPUPowerManagement: initialization complete
Mar 28 22:37:01 localhost kernel[0]: AppleIntelCPUPowerManagementClient: ready
Jun 25 01:09:43 localhost kernel[0]: AppleIntelCPUPowerManagementClient: ready
Jun 29 11:06:18 localhost kernel[0]: AppleIntelCPUPowerManagementClient: ready
Jun 29 14:55:08 localhost kernel[0]: AppleIntelCPUPowerManagementClient: ready

As you can see I'm getting Turbo ratios back in March before I updated to 10.6.7, since then I'm just getting 'ready', no 'Turbo ratios 1112'.

I've noticed talk of adding 'generate c and p states - yes' to com.apple.boot.plist, but have not added them.
This is the contents of my com.apple.boot.plist:
<key>EthernetBuiltIn</key>
	<string>Yes</string>
	<key>GraphicsEnabler</key>
	<string>Yes</string>
	<key>Kernel</key>
	<string>mach_kernel</string>
	<key>Kernel Flags</key>
	<string>GraphicsEnabler=Yes</string>
	<key>Timeout</key>
	<string>1</string>

I have attached my dsdt but this has been edited to allow for native power management.

ex58-ud5 rev.1

Thanks in advance

Attached Files



#1250
JamesLittler

JamesLittler

    InsanelyMac Legend

  • Members
  • PipPipPipPipPipPipPipPip
  • 1,022 posts
  • Gender:Male
  • Location:Portsmouth, England
James-Littlers-Mac:~ James$ ioreg -lw0 | grep PerformanceStateArray
	| |   | |   "PerformanceStateArray" = (<e90a0000d0fb01000a0000000a0000000015000000150000>,<640a0000d0fb01000a0000000a0000000014000000140000>,<df090000c8a901000a0000000a0000000013000000130000>,<5a090000a08601000a0000000a0000000012000000120000>,<d5080000384401000a0000000a0000000011000000110000>,<50080000e02801000a0000000a0000000010000000100000>,<cb07000018f600000a0000000a000000000f0000000f0000>,<46070000a8de00000a0000000a000000000e0000000e0000>,<c106000098b700000a0000000a000000000d0000000d0000>,<3c060000f8a700000a0000000a000000000c0000000c0000>)
James-Littlers-Mac:~ James$ ioreg -lw0 | grep CSTInfo
James-Littlers-Mac:~ James$

Does this mean C states are not working?

#1251
Dajjal

Dajjal

    InsanelyMac Protégé

  • Members
  • PipPip
  • 72 posts
  • Gender:Male
  • Location:WI
@ JamesLittler

I wish i could answer your question but i myself am looking for answers.One note though, i noticed your boot.plist has 2 entries for GraphicsEnabler,i believe you can remove the second one '' <string>GraphicsEnabler=Yes</string> ''

#1252
JamesLittler

JamesLittler

    InsanelyMac Legend

  • Members
  • PipPipPipPipPipPipPipPip
  • 1,022 posts
  • Gender:Male
  • Location:Portsmouth, England
Booting with 'GeneratePStates=Yes GenerateCStates=Yes' has no effect on the output of ioreg.
If I run cpu-x it looks about right though the version I have is in russian so I don't know exactly what's happening but values change as I do things, so I guess that's a good sign, right?

This is making me think that C&P states are working but 10.6.7 may have changed the way the output is logged in console.

#1253
d00d

d00d

    InsanelyMac Sage

  • Members
  • PipPipPipPipPipPip
  • 477 posts
  • Gender:Male
  • Location:41.648324, -70.635260

James-Littlers-Mac:~ James$ ioreg -lw0 | grep PerformanceStateArray
	 | |   | |   "PerformanceStateArray" = (<e90a0000d0fb01000a0000000a0000000015000000150000>,<640a0000d0fb01000a0000000a0000000014000000140000>,<df090000c8a901000a0000000a0000000013000000130000>,<5a090000a08601000a0000000a0000000012000000120000>,<d5080000384401000a0000000a0000000011000000110000>,<50080000e02801000a0000000a0000000010000000100000>,<cb07000018f600000a0000000a000000000f0000000f0000>,<46070000a8de00000a0000000a000000000e0000000e0000>,<c106000098b700000a0000000a000000000d0000000d0000>,<3c060000f8a700000a0000000a000000000c0000000c0000>)
 James-Littlers-Mac:~ James$ ioreg -lw0 | grep CSTInfo
 James-Littlers-Mac:~ James$

Does this mean C states are not working?

Read my last note for how to check CStates as of 10.6.8.

#1254
JamesLittler

JamesLittler

    InsanelyMac Legend

  • Members
  • PipPipPipPipPipPipPipPip
  • 1,022 posts
  • Gender:Male
  • Location:Portsmouth, England
So:

21. (added 20100726) Add the following to com.apple.Boot.plist if you're using the C2RC5 (trunk revision 192 and newer) bootloader;

<key>GenerateCStates</key>
<string>yes</string>


To confirm, use ioreg to look at IOService:/AppleACPIPlatformExpert/CPU0@0/AppleACPICPU/ACPI_SMC_PlatformPlugin.
You will see CPUPLimit with a value of 0x0, a different value if it's not activated, or not there at all if not enabled.


Is that correct?

#1255
JamesLittler

JamesLittler

    InsanelyMac Legend

  • Members
  • PipPipPipPipPipPipPipPip
  • 1,022 posts
  • Gender:Male
  • Location:Portsmouth, England
This is the line I think refers to IOService:/AppleACPIPlatformExpert/CPU0@0/AppleACPICPU/ACPI_SMC_PlatformPlugin

CPU0@0 <class IOACPIPlatformDevice, id 0x100000111, registered, matc$

Does that mean C states aren't activated?

EDIT:
Booting with GenerateCStates=Yes makes no difference to this line.

EDIT 2:
Mark-i shows the cpu going from 12-21 but does that mean C states are working?

#1256
d00d

d00d

    InsanelyMac Sage

  • Members
  • PipPipPipPipPipPip
  • 477 posts
  • Gender:Male
  • Location:41.648324, -70.635260

So:




Is that correct?

That's my first note, read my last note one page back.

#1257
JamesLittler

JamesLittler

    InsanelyMac Legend

  • Members
  • PipPipPipPipPipPipPipPip
  • 1,022 posts
  • Gender:Male
  • Location:Portsmouth, England
Great, that all checks out and booting with GenerateCStates=yes seems to lower my cpu temps by about 5C at idle.

Thanks

#1258
xbb

xbb

    InsanelyMac Protégé

  • Members
  • PipPip
  • 65 posts

Great, that all checks out and booting with GenerateCStates=yes seems to lower my cpu temps by about 5C at idle.

Thanks


I think I have your same problem. I'm using GenerateCStates and GeneratePStates. X58A-UD7 - i7 930
I'm using 10.6.7 with custom DSDT with the fixed listed in the first post, except CPU states.
CPUPLimit shows 0
PerformanceStateArray is filled with values.
No turbo ratios in logs.
temp is okay but frequency seems only switching between 1200 and 2200, looking at SMC Monitor.
Maybe it's a problem with chameleon?

#1259
xbb

xbb

    InsanelyMac Protégé

  • Members
  • PipPip
  • 65 posts
Last time turbo ratios was working and then not:

Jul 25 13:50:31 localhost kernel[0]: AppleIntelCPUPowerManagement: Turbo Ratios 1112
Jul 25 13:50:31 localhost kernel[0]: AppleIntelCPUPowerManagement: initialization completeRTC: Only single RAM bank (128 bytes)
Jul 25 13:50:31 localhost kernel[0]: AppleIntelCPUPowerManagementClient: ready
Jul 25 14:20:25 localhost kernel[0]: AppleIntelCPUPowerManagement: initialization complete
Jul 25 14:20:25 localhost kernel[0]: AppleIntelCPUPowerManagementClient: ready

I was updating fakesmc.

drwxr-xr-x 3 root wheel 102 Jul 25 14:23 IntelCPUMonitor.kext/
drwxr-xr-x 3 root wheel 102 Jul 25 14:16 SuperIOFamily.kext/
drwxr-xr-x 3 root wheel 102 Jul 25 14:15 RadeonMonitor.kext/
drwxr-xr-x 3 root wheel 102 Jul 25 13:44 FakeSMC.kext/

Edit:

Removing SuperIOFamily.kext brings back Turbo Ratios, but no change in SMC Monitor, switching only between 1200-2200

Jul 26 01:18:32 localhost kernel[0]: AppleIntelCPUPowerManagement: Turbo Ratios 1112

So it's really only estetical? Is there a way to see the real frequency? Tried mark-i also shows vcore ranging from 0.9 to almost 1.2 and frequency around 1700-2000

#1260
SuperHack

SuperHack

    InsanelyMac Protégé

  • Members
  • Pip
  • 39 posts
Much thanks for helping put all this together!

The only issue I have is idle sleep. I figured out my system idle sleeps correctly as long as Safari is closed. When I use RIP3IanNet auto-sleep works even when Safari is open, but about 1/3 times a try to resume the monitor doesn't come back on (no signal from GPU). I could try the RIP script again (changed a couple things since I last tried), but does anyone have a solution to the black screen issue or is there something simple that can be made to allow idle sleep to work when Safari is open?





0 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users

© 2014 InsanelyMac  |   News  |   Forum  |   Downloads  |   OSx86 Wiki  |   Mac Netbook  |   PHP hosting by CatN  |   Designed by Ed Gain  |   Logo by irfan  |   Privacy Policy