joe82

ok tried it on a d430... pro's you can activate wake from usb in bios and it sleeps now also will wake from usb mouse.... but for some reason it drives the trackpad crazy it jumps all over the place had to diable it in bios. i have not tried vga, bluetooth. or dock. so your definitely on the right track in getting the little gremlins worked out. have yet to try your original d620 dsdt edits on my d620 intel.... waiting for a new hdd. will report back on that one when i get it back up.

 

update after a couple reboots trackpad is working again but it will not wake from usb mouse anymore but it can still sleep wile having usb wake enabled in bios wich it would not do previously it had to be off in bios for sleep to work properly before,

 

I usually do not use the track-pad, or the "eraser" in the keyboard much.

I prefer a mouse, either wireless, or wired. I have seen signs of what you

describe happening in the track-pad. I have not had time to look into this.

 

The only thing that comes to mind is usually there are some parameters

associated with a mouse, acceleration, for example. Perhaps they are

not being restored.

 

Joe

WOW.. thats amazing work dude .. its stuff like that we need onboard our team.

 

Joe, if you dont mind, could you add the same changes to the D4x0 DSDT so i could give it a test ?

 

 

Here it is.

 

First some notes:

 

For the D4X0, I used the DSDT from EDP 1.9 as a starting point.

 

I do not know what chipset is used in the D4X0 machines,

but I do know the D620 uses the ICH7 chipset. So the updates

to the D4X0 DSDT are based on the D620.

 

1) The DSDT USB0-3 devices are UHCI devices.

Using lspci and looking at the data sheet, I found these devices

DO NOT have a power management capability in the PCI configuration

space. I believe you can still have a machine that supports power

management, but these devices are on or off. This is why I did not

include the _DSM() methods in these devices. They were present in

the D4X0 DSDT, so I commented them out. I figured it would be best

to start with what worked for me. I also added the other pieces.

 

2) The DSDT EHCI device is an EHCI device.

Using lspci and looking at the data sheet, I found this device

DOES have a power management capability in the PCI configuration

space. I did include the _DSM() methods in the EHCI device. The version

that is in the D4X0 ECHI device was largely a super set of

my version. Since I am not clear what side effects the additional

items will cause, I commented out that version, and added my version.

This may not matter. I also added the other pieces.

 

3) I changed the AAPL,Haslib to AAPL,HasLid in the _DSM method

in the VID device.

 

Lastly I suggest updating the BIOS parameters to match the values I used.

 

Let me know how it goes.

 

Joe

dsdt_d4x0.tar.gz

Hi All,

 

I have a D620 with a DELL TrueMobile 360 Bluetooth card. Overall it has worked well, but

I loose the connection to my Bluetooth mouse after a wake from sleep. I have noted others seem to

have the same problem from the notes in their signatures. I decided to look into this problem.

 

After reading various forums on other websites, I decided to try some changes to the DSDT,

specifically the USB* and EHCI devices. I compared the same devices in the DSDT from a

Macbook1,1. The two machines are largely the same.

 

I added the following to the USB0-3 devices.

 

Method (_S3D, 0, NotSerialized)

{

Return(0x03)

}

 

Method (_S4D, 0, NotSerialized)

{

Return(0x03)

}

 

and the following to the EHCI device.

 

Method (_DSM, 4, NotSerialized)

{

Store (Package (0x04)

{

"AAPL,clock-id",

Buffer (One)

{

One

},

 

"device_type",

Buffer (0x05)

{

"EHCI"

}

 

}, Local0)

DTGP (Arg0, Arg1, Arg2, Arg3, RefOf (Local0))

Return (Local0)

}

 

Method (_S3D, 0, NotSerialized)

{

Return(0x03)

}

 

Method (_S4D, 0, NotSerialized)

{

Return(0x03)

}

 

With these changes, I am able to get the USB-based Bluetooth adaptor to reconnect with the mouse after a wake from sleep.

I am also able to change the hibernatemode parameter in pmset back to the default value of 3.

Here are my pmset values.

 

 

Active Profiles:

AC Power -1*

Battery Power -1

Currently in use:

hibernatefile /var/vm/sleepimage

hibernatemode 3

disksleep 10

sleep 10

displaysleep 3

lidwake 1

ttyskeepawake 1

womp 0

acwake 0

halfdim 1

autorestart 0

 

To be complete, here are the BIOS parameters I have updated that are related to the USB.

 

Onboard Devices

External USB ports: Enabled

Integrated USB Hub: High Speed

Power Management

USB Wake Support: Enabled

POST Behavior

USB Emulation: Enabled

 

 

I am also running 10.6.8 with EDP 1.9. I have not had time to try EDP 1.9.1 or 1.9.2

When the EDP is configured, I choose all the Apple versions of the drivers.

 

Although these changes seem to solve my original problem, I believe the DSDT

is removing the power from the Bluetooth adaptor when going to sleep. This

makes it hard to get a wake-on-bluetooth-mouse-event to work.

 

I decided to start with a wired USB-mouse. Using lspci, I was able to verify the

power management registers were getting set when this mouse was used. I was

able to get the machine to wake on a usb-mouse button hit.

Looking at the "Advanced..." tab from the "Open Bluetooth Preferences..." tab,

wake-on-bluetooth-mouse-event seems to be supported, but the

"Allow Bluetooth devices to wake this computer" is grayed-out. It seems I do not

have a supported Bluetooth adaptor. I have been able to "trick" the OS into thinking

I have a supported adaptor, but with the power being turned-off, it does not work. I

have seen other people comment on this problem, and they have even found the

line that disables the power, but their DSDTs are different. I will put this on my list of

things to look into. If I ever find the way to keep the power connected, I will most

likely try to get the wake-on-bluetooth-mouse-event to work.

 

 

I am going to attach my modified dsdt.dsl and the compiled dsdt.aml files. Note this

DSDT also includes a fix to get the VGA port to work correctly on a docking station.

dsdt_d620_gma.tar.gz

It has been a while since I updated this, but I have some additional information.

 

First, using EDP 1.7, I went back to see if I could add back the device-properties

line in com.apple.boot.plist. After adding it back it still worked correctly when in

the docking station. It also worked out of the docking station. The only problem

I observed is then the screen entered display sleep. It does exit display sleep, but

the screen seems to be at a lower resolution. It is still 1440x900, but it is as if half

the pixels are not being used. I was able to track the problem to the option

 

AAPL01,DualLink

 

It seems with displays of this size, 1440x900, this needs to be set. So I updated

and added the following

 

"AAPL01,DualLink",

Buffer (0x04)

{

0x01, 0x00, 0x00, 0x00

},

 

"AAPL01,Height",

Buffer (0x04)

{

0x84, 0x03, 0x00, 0x00

},

 

"AAPL01,Width",

Buffer (0x04)

{

0xA0, 0x05, 0x00, 0x00

},

 

This fixes the display sleep problem, and it still works in the docking station.

 

I have also moved from 10.6.7 and EDP 1.7 to 10.6.8 and EDP 1.9. I have

not had time to update to EDP 1.9.1 or 1.9.2. Things have changed in Chameleon

since EDP 1.7. The com.apple.boot.plist has been changed to org.chameleon,Boot.plist.

The device-properties is no longer present in the new file. The good news is these

changes have not been a problem for this docking station fix.

 

At this point I decided to look back at the fix to this parameter.

 

"AAPL,HasLid",

Buffer (0x04)

{

0x01, 0x00, 0x00, 0x00

},

 

This parameter had been updated from AAPL,Haslid to AAPL,HasLid.

 

I did a strings on /System/Library/Extensions/AppleIntelIntegratedFramebuffer.kext/AppleIntelIntegratedFramebuffer

 

After pruning back some of the junk I got.

 

AAPL,HasPanel

AAPL,HasLid

AAPL,NumFramebuffers

AAPL,NumDisplays

AAPL,DisplayConfig

AAPL,SelfRefreshSupported

AAPL,display-alias

AAPL,boot-display

saved-config

saved-timing0

saved-timing1

AAPL00,CurrentDisplay

AAPL00,IODisplayMode

AAPL00,Pipe

AAPL00,BootDisplay

AAPL00,Depth

AAPL00,Width

AAPL00,Height

AAPL00,Interlace

AAPL00,Refresh

AAPL00,EDID

AAPL00,display-alias

AAPL00,IgnoreConnection

AAPL00,no-hotplug-interrupt

AAPL01,DualLink

AAPL01,T1

AAPL01,T2

AAPL01,T3

AAPL01,T4

AAPL01,T5

AAPL01,T6

AAPL01,T7

AAPL01,PixelFormat

AAPL01,BacklightIntensity

AAPL01,InverterFrequency

AAPL01,CurrentDisplay

AAPL01,IODisplayMode

AAPL01,Pipe

AAPL01,BootDisplay

AAPL01,Depth

AAPL01,Width

AAPL01,Height

AAPL01,Interlace

AAPL01,Refresh

AAPL01,EDID

AAPL01,display-alias

AAPL01,IgnoreConnection

AAPL01,no-hotplug-interrupt

 

So this driver does reference AAPL,HasLid.

 

I have also found the same string in

 

/System/Library/Extensions/AppleIntelGMAX3100FB.kext/AppleIntelGMAX3100FB

 

I am not clear what systems use this driver, but the same fix might work in that case.

 

Anyway this fix has been stable for me for some time now.

 

I still have not found a way to get the DVI port to work, but it is still on the list.

No, the DVI port is still a problem.

I would like to look into this as soon as I get time.

Hi all,

 

First nice work by all. I have not been an Apple user, but now that I have access to a machine running the

Apple OS, it has given me something to think about.

 

Since I have an interest in using a D620 in a docking station, I decided to try the solution posted by "Fireman Sam".

I was able to implement the documented changes, and found it worked at advertised.

 

I started looking at why the DVI port did now work. I looked over "dumps" from real Apple machines, and found an

interesting different between the Apple version of a variable, and the one in the D620 DSDT. The variable, from the

EDP 1.7 DSDT is:

 

"AAPL,Haslid"

 

Apple uses

 

"AAPL,HasLid"

 

Note the "Capital-L".

 

I updated the DSDT, and gave it a try. It still worked.

 

I backed-out the DSDT change (detection of lid/dock), and gave it a try.

It still worked. Using "system profiler" I was able to note the internal display

was not "configured" when docked, but was when not docked. This should

help save on power, I guess.

 

So my modified fix is;

1) Remove the device-properties from com.apple.boot.plist (As documented by Fireman Sam).

2) Change "AAPL,Haslid" to "AAPL,HasLid" in DSDT.

 

To be accurate, I did update two of the other variables, "subsystem-vendor-id", and "subsystem-id".

This should not matter. I also noted that there are various other variables, in the Apple version. Most

I do not know what they do, and I am not clear which should be in DSDT, and which should be in

com.apple.boot.plist.

 

This version of the fix, seems to move the decision making from the DSDT, to the OS level. This fix should

be more portable across laptops, and other VGA chipsets. "Wake from sleep" still works. I did not

try mirror, and there are still other things to test.