Microphone led state does not persist after sleep

[igorkulman]

Steps reproduce:

• press fn+f4 to disable the microphone, the led on f4 turns on
• put the Thinkpad to sleep
• wake up the Thinkpad
• observe the led on f4 being off although the microphone is still muted

[MSzturc]

It's a problem in ACPI. We have to restore LED State after wakeup. We have to patch _WAK and inside you have to restore LED state ( we created an Variable which should be still present after wakeup ). Have a look here for reference : https://github.com/zhtengw/EFI-for-X1C6-hackintosh/blob/e73ddb21e1192990e3e2156f3613b735bf71de75/ACPI/SSDT-SLPWAK-X1C6.dsl

[xma]

I've modified T480 configuration with this idea
xma/T480-Clover@7320f08

It's within a method WKBD, called from _WAK -> EXT4
next to required call for switching led state on thinkpad case
idea is to avoid putting more logic on upper level, and keeping everything related to KBD within same dsl file

[MSzturc]

It's within a method WKBD, called from _WAK -> EXT4
next to required call for switching led state on thinkpad case
idea is to avoid putting more logic on upper level, and keeping everything related to KBD withing same dsl file

I experiment with a similar solution currently. Im not a fan of mixing two separate functionalities. The current solution with the ssdt & the acpi patch overhelmed many users. Patching 2 ssdt's will make their experience worse ( and it will drive my supporttime up :-) )

A better solution would be to write an kext that synchronizes the states of the led's. It could solve this problem and another bug i noticed: When you mute your sound by reducing volume to 0 the sound muted led will stay off.

My goal (for Thinkpad Assistant 2.0) would be to ship it with a kext that does 3 things:

• Patches the SSDT's dynamically on load. This would fix pitfalls like LPC & LPCB mismatch
• Synchronize the state of LEDs ( to get rid of the 2/3 button bindings thats only need because Thinkpad Assistant doesn't know what the current LED state is ).
• Introduce a Communication Layer between Thinkpad Assistant (which still will do the rendering of the HUD) and Assistant Kext. With that we get rid of the Keybinding problematic (almost all F-Key are bound right now and using an Apple Bluetooth Keyboard that has F13-F19 on it causes the problem that this Buttons are not usable for us since the are already used by Thinkpad Assistant) With that we could have as many binding as we want

@xma @EETagent @tylernguyen @Sniki
Do you have experience in writing a kext? For me it would be the first time and i could need some help and advise

[tylernguyen]

@MSzturc
I do not think there's a need to write one from scratch There's already such a kext. See https://github.com/qwerty12/ThinkPadMuteLEDSetter

We would only need to clean it up and make it compatible with Catalina and ThinkpadAssistant.

[junaedahmed]

Hi,
I found a simple solution to this problem without modifying other SSDT.
According to ACPI spec _TTS (transition to sleep) is a reserved method which is run once before sleep and once again after wake, but as it is not mandatory in all thinkpad's DSDT this method is not implemented, we can exploit this inconsistency to restore any method calling after wake with only just this snippet:

    Scope (\)
    {

        // This ACPI reserved method is run once before sleep and once after awakened
        Method (_TTS, 1, NotSerialized)
        {
            If (_OSI ("Darwin"))
            {
                // Arg0 contains the system state of transition
                // for wake state it is Zero.
                If (Arg0 == Zero & \_SB.PCI0.LPC.EC.LED1 == One)
                {
                    \_SB.PCI0.LPC.EC.HKEY.MMTS (0x02)
                }
            }
        }
    }

I tested it it works.... Pls test yourselves and report.

best wishes,
Junaed

[Sniki]

Hi,
I found a simple solution to this problem without modifying other SSDT.
According to ACPI spec _TTS (transition to sleep) is a reserved method which is run once before sleep and once again after wake, but as it is not mandatory in all thinkpad's DSDT this method is not implemented, we can exploit this inconsistency to restore any method calling after wake with only just this snippet:

    Scope (\)
    {

        // This ACPI reserved method is run once before sleep and once after awakened
        Method (_TTS, 1, NotSerialized)
        {
            If (_OSI ("Darwin"))
            {
                // Arg0 contains the system state of transition
                // for wake state it is Zero.
                If (Arg0 == Zero & \_SB.PCI0.LPC.EC.LED1 == One)
                {
                    \_SB.PCI0.LPC.EC.HKEY.MMTS (0x02)
                }
            }
        }
    }

I tested it it works.... Pls test yourselves and report.

best wishes,
Junaed

Thanks for this, we need this for LED Blink fix as well, im learning ACPI slowly and this is what vit9696 suggested me but I wasn’t able to formulate that time and which is the right way.
No need to patch _WAK and add extra complications

[Sniki]

@MSzturc @junaedahmed @tylernguyen @tluck
Thanks to @junaedahmed post above, with the SSDT providing a fix for saving Mic Mute LED toggle state after wake from sleep, i also included the cleanest possible Power and ThinkPad LED breathing light after wake from sleep. Now we don't need to patch WAK by renaming it to ZWAK and providing all those methods, here is my final SSDT that i think should be used in all the ThinkPads line-up as it's the cleanest one as far as i have researched:

// _TTS Method (TransitionToState) to fix LED issues like:
// Power Button LED and Red LED blinking after Wake from Sleep
// Save Microphone Mute F4 toggle LED State after Wake from Sleep
// Credits: @junaedahmed (Mic Mute LED) @Sniki (Power LED)

DefinitionBlock ("", "SSDT", 1, "ThinkPad", "LED", 0)
{
    External (_SB.PCI0.LPC.EC.HKEY.MMTS, MethodObj)
    External (_SB.PCI0.LPC.EC.LED1, IntObj)
    External (_SI._SST, MethodObj)
    
    Method (_TTS, 1, NotSerialized)
    {
        If (_OSI ("Darwin"))
        {
            // Arg0 contains the system state of transition
            // for wake state it is Zero.
            If (Arg0 == Zero & \_SB.PCI0.LPC.EC.LED1 == One)
            {
                \_SB.PCI0.LPC.EC.HKEY.MMTS (0x02)
            }
            If (Arg0 == Zero)
            {
                \_SI._SST(One)
            }
        }
    }
}

[junaedahmed]

Hi @Sniki,
Nice work, one correction,
External (_SB.PCI0.LPC.EC.LED1, MethodObj) here MethodObj should be IntObj.

[Sniki]

Hi @Sniki,
Nice work, one correction,
External (_SB.PCI0.LPC.EC.LED1, MethodObj) here MethodObj should be IntObj.

Thanks, will do the correction.
Update: Done !

[MSzturc]

@Sniki @junaedahmed
What if SB.PCI0.LPC.EC.LED1 is not present? Will this code work or crash?

[junaedahmed]

@MSzturc
I was thinking using it on SSDT-KBRD.dsl. If we use on other SSDT like Sniki's one, then we could try conditionally referencing it like this:

Method (_TTS, 1, NotSerialized)
    {
        If (_OSI ("Darwin"))
        {
            // Arg0 contains the system state of transition
            // for wake state it is Zero.
            // Here conditional reference will yield false if LED1 not found and the other functionality 
            will not break.
            If (Arg0 == Zero & CondRefOf (\_SB.PCI0.LPC.EC.LED1 == One))
            {
                \_SB.PCI0.LPC.EC.HKEY.MMTS (0x02)
            }
            If (Arg0 == Zero)
            {
                \_SI._SST(One)
            }
        }
    }

I think you should add it on KBRD and those who want power led fix too should make their own custom ssdt. it will be simpler to maintain for you. Most hackintosher already use power led fix on _wak method so they don't have to change any thing new.

[Sniki]

@MSzturc
I was thinking using it on SSDT-KBRD.dsl. If we use on other SSDT like Sniki's one, then we could try conditionally referencing it like this:

Method (_TTS, 1, NotSerialized)
    {
        If (_OSI ("Darwin"))
        {
            // Arg0 contains the system state of transition
            // for wake state it is Zero.
            // Here conditional reference will yield false if LED1 not found and the other functionality 
            will not break.
            If (Arg0 == Zero & CondRefOf (\_SB.PCI0.LPC.EC.LED1 == One))
            {
                \_SB.PCI0.LPC.EC.HKEY.MMTS (0x02)
            }
            If (Arg0 == Zero)
            {
                \_SI._SST(One)
            }
        }
    }

Agree, this should be a must as if any ThinkPad model or newer model doesn't have the method at least the LED Blink Fix will work.
It may also be a good idea to add an extra comment as note: you may want to adapt the LED1 device according to your ACPI if it is different than LED1

Note: there is an error with your change, please check it on maciASL

[junaedahmed]

@Sniki
Sorry posted without testing. Use it like this:

Method (_TTS, 1, NotSerialized)
    {
        If (_OSI ("Darwin"))
        {
            // Arg0 contains the system state of transition
            // for wake state it is Zero.
            If CondRefOf (\_SB.PCI0.LPC.EC.LED1)
            {
                If (Arg0 == Zero & \_SB.PCI0.LPC.EC.LED1 == One)
                {
                    \_SB.PCI0.LPC.EC.HKEY.MMTS (0x02)
                }
            }
            
            If (Arg0 == Zero)
            {
                \_SI._SST(One)
            }
        }
    }

[Sniki]

@Sniki
Sorry posted without testing. Use it like this:

Method (_TTS, 1, NotSerialized)
    {
        If (_OSI ("Darwin"))
        {
            // Arg0 contains the system state of transition
            // for wake state it is Zero.
            If CondRefOf (\_SB.PCI0.LPC.EC.LED1)
            {
                If (Arg0 == Zero & \_SB.PCI0.LPC.EC.LED1 == One)
                {
                    \_SB.PCI0.LPC.EC.HKEY.MMTS (0x02)
                }
            }
            
            If (Arg0 == Zero)
            {
                \_SI._SST(One)
            }
        }
    }

There is still an issue,
I believe it should be:

If (CondRefOf (\_SB.PCI0.LPC.EC.LED1))

Check it one more time and correct me if im wrong, as there were still errors.

Thanks !

Update: also ThinkPad exceeds 6 characters for OEM ID, so we want something shorter, either "hack" short word for hackintosh or whatever you guys prefer.

[junaedahmed]

Yes there should be one more pair of parentheses.