Fix Unknown Synopsys component type issue in Comet Lake 2nd

[startpenghubingzhou]

This commit has fixed the issue #236 . This work is based on https://www.notion.so/Using-VoodooI2C-on-comet-lake-cpu-e-g-i5-10210u-142930887087445eaa533120455da5dc , which will force D0 in Comet Lake. And I completed it to check the platform to avoid this patch to be applied in other platforms. See code comments for more details.

[startpenghubingzhou]

@kprinssu As I have mentioned in #236 (comment), there 's two ways to check I have thought: by judging from IOMatchedID , or injecting a flag that marked 10th and judging this flag. But comparing with these two ways I think it's better to check it in code by judging from device name.

[startpenghubingzhou]

I also tried a compromise way: Set a flag in _DSM in the touchpad device DSDT code so it will be loaded as a property in the father device of VoodooPCIController. But it 's not easy to get a fatherdevice property for VoodooPCIController, and checking device_name flag is an easy and safe-enough way to do that.

Btw, the test last night showed that checking IONameMatched
is an UNBELIVIABLE way because it sometimes returns like 9dxx not 9d as required.

[ben9923]

Hey, thanks for the PR :)
I do think messing with the registers ourselves isn't necessary, as Apple's kexts can do it for us.

Why not doing something like this after IOPCIDevice::enablePCIPowerManagement?

physical_device.pci_device->setPowerState(kIOPCIDeviceOnState, physical_device.pci_device);

It will call IOPCIBridge::setDevicePowerState, that will call IOPCIDevice:: setPCIPowerState

Should be safe to do in all platforms. If device is already in D0 state the function will just skip it.

[startpenghubingzhou]

Hey, thanks for the PR :)
I do think messing with the registers ourselves isn't necessary, as Apple's kexts can do it for us.

Why not doing something like this after IOPCIDevice::enablePCIPowerManagement?

physical_device.pci_device->setPowerState(kIOPCIDeviceOnState, physical_device.pci_device);

It will call IOPCIBridge::setDevicePowerState, that will call IOPCIDevice:: setPCIPowerState

Should be safe to do in all platforms. If device is already in D0 state the function will just skip it.

Thx for your suggestions. As you know, I 'm new to iokit so I don' t really know all the functions. I ll edit it and then send a new commit again. :)

[startpenghubingzhou]

Hey, thanks for the PR :)
I do think messing with the registers ourselves isn't necessary, as Apple's kexts can do it for us.

Why not doing something like this after IOPCIDevice::enablePCIPowerManagement?

physical_device.pci_device->setPowerState(kIOPCIDeviceOnState, physical_device.pci_device);

It will call IOPCIBridge::setDevicePowerState, that will call IOPCIDevice:: setPCIPowerState

Should be safe to do in all platforms. If device is already in D0 state the function will just skip it.

Hey dude I have tried this one:

auto pci_device = physical_device.pci_device;
    pci_device->enablePCIPowerManagement(kPCIPMCSPowerStateD0);
    IOReturn ret = pci_device->setPowerState(kIOPCIDeviceOnState, physical_device.pci_device);

    if (ret != kIOReturnSuccess) {
        IOLog("%s::%s Set D0 state failed! Trying to patch it with Comet Lake patch...\n", getName(), physical_device.name);
        uint16_t oldPowerStateWord = pci_device->configRead16(0x80 + 0x4);
        uint16_t newPowerStateWord = (oldPowerStateWord & (~0x3)) | 0x0;
        // Modify 0x80 below to your findings.
        pci_device->configWrite16(0x80 + 0x4, newPowerStateWord);
        IOLog("%s::%s Successfully patched D0 state with Comet Lake patch!\n", getName(), physical_device.name);
    }

This can 't patch comet lake successfully because it seems that IOPCIBridge::setDevicePowerState always return kIOReturnSuccess. Tested in Xiaoxin Pro 13 with comet lake CPU.

[kprinssu]

I spoke with @ben9923 we can consider this a hack and then revert it in the future. We are thinking it's a macOS bug and hopefully it will be fixed in the future.

@startpenghubingzhou Do you have any issues with me merging it in?

[ben9923]

I spoke with @ben9923 we can consider this a hack and then revert it in the future. We are thinking it's a macOS bug and hopefully it will be fixed in the future.

@startpenghubingzhou Do you have any issues with me merging it in?

@startpenghubingzhou macOS 10.15.4 should be out together with the 10th Gen MacBook Air, which means full support for this hardware. Give it a try on it (whether latest beta or once it's released) if you can :)

[startpenghubingzhou]

I spoke with @ben9923 we can consider this a hack and then revert it in the future. We are thinking it's a macOS bug and hopefully it will be fixed in the future.

@startpenghubingzhou Do you have any issues with me merging it in?

Nothing if you think this can be a temporary solution, and also, I ll try to explore it deeper to try finding what caused that and find a perfect way to solve that but not just hacking it.:)

[startpenghubingzhou]

I spoke with @ben9923 we can consider this a hack and then revert it in the future. We are thinking it's a macOS bug and hopefully it will be fixed in the future.
@startpenghubingzhou Do you have any issues with me merging it in?

@startpenghubingzhou macOS 10.15.4 should be out together with the 10th Gen MacBook Air, which means full support for this hardware. Give it a try on it (whether latest beta or once it's released) if you can :)

Sure if I have time. For these days I m busy working on my homework so I might have enough time to test, but I ll do it as soon as I have the free time. :)

[startpenghubingzhou]

I spoke with @ben9923 we can consider this a hack and then revert it in the future. We are thinking it's a macOS bug and hopefully it will be fixed in the future.
@startpenghubingzhou Do you have any issues with me merging it in?

@startpenghubingzhou macOS 10.15.4 should be out together with the 10th Gen MacBook Air, which means full support for this hardware. Give it a try on it (whether latest beta or once it's released) if you can :)

@ben9923 also I asked somebody tested in 10.15.4 beta and it seems that it still need a patch of this.

[kprinssu]

Sounds good, we can revisit the issue at some later point in time. I am going to be merging this into master and get a new build going.

[marianopela]

@ben9923 also I asked somebody tested in 10.15.4 beta and it seems that it still need a patch of this.

I confirm it, not working on beta 6

[thefiredragon]

Can confirm on macOS 10.15.4 trackpad is not working at the moment.
Any updates here?

[thefiredragon]

on 10.15.3 Trackpad was working with the patch posted on the issues.
After upgrade to 10.15.4 trackpad works after sleep I noticed

[marianopela]

Can confirm on macOS 10.15.4 trackpad is not working at the moment.
Any updates here?

Working for me. Polling mode

Interrupts never worked for me

[kprinssu]

@thefiredragon @marianopela can you please carry the discussion over into #236? It makes it much easier to track everything in one location rather multiple.

[thefiredragon]

@thefiredragon @marianopela can you please carry the discussion over into #236? It makes it much easier to track everything in one location rather multiple.

We can,

for me it's working after second restart and clean kextcache. Also after an reboot.

[startpenghubingzhou]

@thefiredragon @marianopela can you please carry the discussion over into #236? It makes it much easier to track everything in one location rather multiple.

We can,

for me it's working after second restart and clean kextcache. Also after an reboot.

This may caused by a kextcache problem and have no connection with this patch.

[marianopela]

@Trilis29 this is not the right place to discuss. Just use latest release, which should be fine on your hardware. You’re probably missing XOSI patch or something

[kprinssu]

@Trilis29 stop commenting here. The changes are merged, you probably have another issue. Read https://voodooi2c.github.io/ and https://voodooi2c.github.io/#Troubleshooting/Troubleshooting.