[Guide] HP EliteBook 820 840 G2 I5-5300U HD5500 El Capitan [ENG]

[stealthdave]

GOOD NEWS, EVERYONE: SLEEP IS NOW WORKING!!!

The problem? Video RAM was set to 512Mb in BIOS. I changed the value to 128Mb, and BAM! Sleep started working!

That was a long way to go, but thank you SO much for your help, @RehabMan!

- Dave

 

[RehabMan]

GOOD NEWS, EVERYONE: SLEEP IS NOW WORKING!!!

The problem? Video RAM was set to 512Mb in BIOS. I changed the value to 128Mb, and BAM! Sleep started working!

That was a long way to go, but thank you SO much for your help, @RehabMan!

- Dave

So 64mb DVMT-prealloc and 128mb DVMT-size should be the recommendation. What is the BIOS default?

 

[stealthdave]

The only setting in the BIOS available is Video memory size under Advanced -> Device Configurations, and the default value is 32Mb. The OP recommended setting it to 512Mb, which is what I did. Setting it to 128Mb allowed sleep to work. I didn't think to try and change it until I started Googling the crap out of sleep even more than I already had and started down even more rabbit holes.

One thing that is still not working is closing the lid to sleep. Opening the lid wakes the computer up, but closing the lid turns off the backlight and changes brightness to 0, but never actually sleeps. This is something I can potentially live with (not having sleep was not), but it's certainly a "very nice to have".

 

[RehabMan]

The only setting in the BIOS available is Video memory size under Advanced -> Device Configurations, and the default value is 32Mb. The OP recommended setting it to 512Mb, which is what I did. Setting it to 128Mb allowed sleep to work. I didn't think to try and change it until I started Googling the crap out of sleep even more than I already had and started down even more rabbit holes.

I'm not familiar with your BIOS options, but it sounds like Video memory size corresponds to DVMT-prealloc... and it should usually be set to 64mb for most Broadwell laptops. And there are reports that it is best at 128mb if your laptop has a UHD screen.

One thing that is still not working is closing the lid to sleep. Opening the lid wakes the computer up, but closing the lid turns off the backlight and changes brightness to 0, but never actually sleeps. This is something I can potentially live with (not having sleep was not), but it's certainly a "very nice to have".

Keep in mind that the 15 sec delay still applies to lid initiated sleep. Also, all sleep assertions have to clear. When testing it is best to use 'pmset -g' to see that all sleep assertions have cleared before attempting sleep... then wait the 15-20 seconds before giving up on it.

If still not working, let me know and we'll explore other options (has to do with _PRW on the LID device in ACPI).

 

[stealthdave]

Keep in mind that the 15 sec delay still applies to lid initiated sleep. Also, all sleep assertions have to clear. When testing it is best to use 'pmset -g' to see that all sleep assertions have cleared before attempting sleep... then wait the 15-20 seconds before giving up on it.

If still not working, let me know and we'll explore other options (has to do with _PRW on the LID device in ACPI).

After additional testing, it appears that it is going to sleep after closing the lid, but it's taking an extraordinarily long time the first time the lid is closed after a reboot: upwards of 5-10 minutes. I didn't get an exact amount but way longer than the usual 15 seconds. This first "sleep" doesn't have the slow flashing power light in the front of the unit like sleep normally should. I'm not quite certain if this is genuine sleep or just the screen turning off.

Subsequent closings of the lid gives me sleep as expected with the slow blinking indicator light. In both cases, opening the lid wakes the unit... mostly.

Unfortunately, waking from sleep seems to not fully wake all of the USB devices. All of the external ports are fine, but the internal port used for the web cam gets disabled. IOReg also shows a duplicate Bluetooth entry after waking from sleep, making two or more entries for the Bluetooth device on the XHCI hub where one is disabled. It seems like this duplicate entry might actually be the normal restart process for devices after waking, and the webcam is just not getting the signal, but that's just supposition.

I've attached before and after IOreg dumps along with the relevant portion of the pmset log.

Thank you!

Update - Additional testing shows that lid sleep is intermittent, not always the "second" lid close or necessarily any lid close. I'm not sure what the trigger is.

 

[RehabMan]

After additional testing, it appears that it is going to sleep after closing the lid, but it's taking an extraordinarily long time the first time the lid is closed after a reboot: upwards of 5-10 minutes. I didn't get an exact amount but way longer than the usual 15 seconds. This first "sleep" doesn't have the slow flashing power light in the front of the unit like sleep normally should. I'm not quite certain if this is genuine sleep or just the screen turning off.

Sleep is not active until the LED is blinking. You should check for sleep assertions to be clear before testing.

Unfortunately, waking from sleep seems to not fully wake all of the USB devices. All of the external ports are fine, but the internal port used for the web cam gets disabled. IOReg also shows a duplicate Bluetooth entry after waking from sleep, making two or more entries for the Bluetooth device on the XHCI hub where one is disabled. It seems like this duplicate entry might actually be the normal restart process for devices after waking, and the webcam is just not getting the signal, but that's just supposition.

I've attached before and after IOreg dumps along with the relevant portion of the pmset log.

Post EFI/Clover folder as ZIP (omit 'themes' folder).

 

[stealthdave]

Sleep is not active until the LED is blinking. You should check for sleep assertions to be clear before testing.

One consistent assertion for keeping the system awake is ApplePushServiceTask (com.apple.apsd). Turning this off resolved most of the assertions that prevented the system from going to sleep through timeouts or lid closing.

Post EFI/Clover folder as ZIP (omit 'themes' folder).

Attached.

 

[RehabMan]

Attached.

Try disabling XWAK. To do it:
- add patch in config.plist to change Method(XWAK,0,NotSerialized) to ZWAK
in config.plist/ACPI/Patches/DSDT
Comment: change Method(XWAK,0,S) to ZWAK
Find: <5857414b 08>
Replace: <5a57414b 08>

Then add an SSDT-XWAK.aml. Here is source (SSDT-XWAK.dsl):

// Disabling XWAK

DefinitionBlock("", "SSDT", 2, "hack", "XWAK", 0)
{
    External(_SB.PCI0.XHC, DeviceObj)

    // In DSDT, native XWAK is renamed ZWAK
    // As a result, calls to it land here.
    Method(_SB.PCI0.LPBC.XHC.XWAK)
    {
        // do nothing
    }
}
//EOF

 

[stealthdave]

Try disabling XWAK. To do it:
- add patch in config.plist to change Method(XWAK,0,NotSerialized) to ZWAK
in config.plist/ACPI/Patches/DSDT
Comment: change Method(XWAK,0,S) to ZWAK
Find: <5857414b 08>
Replace: <5a57414b 08>

Then add an SSDT-XWAK.aml. Here is source (SSDT-XWAK.dsl):

No change; the Webcam is still disabled after the computer wakes up. I added the new patch to the end of the array, in case order of operations is important. I did notice that earlier in the Patches that there is already a change _WAK to ZWAK entry: find = <5f5741b 01> / replace = <5a57414b 01>. I presume that these changes are intentionally distinct?

 

[RehabMan]

No change; the Webcam is still disabled after the computer wakes up. I added the new patch to the end of the array, in case order of operations is important. I did notice that earlier in the Patches that there is already a change _WAK to ZWAK entry: find = <5f5741b 01> / replace = <5a57414b 01>. I presume that these changes are intentionally distinct?

Need ioreg and EFI/Clover (as zip, no themes).

XWAK not related to _WAK.