Bios Versions / USB-C Thunderbolt Dongle Hotswap / Dell Docking Stations

[thiscodedbox]

Would you please add what version BIOS you've coded this for to the Readme? I see 1.4.4 as the most popular but I'm working on getting the WD15 dock working well and it recommends the 1.4.12 BIOS. If you're working with 1.4.4 still I'm happy to downgrade to be compatible with the project.

This is a great project.

[syscl]

@thiscodedbox I use 1.4.12 with default setting, the only thing I change is RAID -> AHCI.

You should keep using the latest BIOS, that's the reason why I create deploy.sh to auto patch ACPI tables when BIOS get updating.

Note: ACPI tables are sensitive to BIOS setting change as well. Once you change settings in BIOS, re dump your ACPI tables through pressing F4/Fn+F4 under Clover, then run deploy.sh. There's very few projects have such a convenient script to keep their laptops syncing to latest BIOS, latest ACPI patches. Thus deploy.sh is proven to be robust.

syscl

[thiscodedbox]

That's great to hear - I'm glad we can continue to get BIOS updates. Wow. Robust indeed!

[thiscodedbox]

Fantastic. Thanks @syscl !

FYI my 9350 is an i7-6560u with 16GB and a 1TB SM961 if you have any need for testing. I also use a Dell WD15 Dock and a generic USB-C Dongle that has 1xUSB3.0 1xHDMI 1xEthernet and USB-C Power Input (PD).

[syscl]

@thiscodedbox Thank you, I want to check if the USB-C work as expect.

Could you please give me the screenshot from DPCIManager once you plug in HDMI?

Thank you, syscl

[thiscodedbox]

This is with the WD15 Dell Dock connected. Everything works on it AFAIK.

[syscl]

@thiscodedbox I am really appreciate the detailed information you provide, thank you!

BTW, if you remove the USB-C dongle, what's the screenshot of the DPCIManager->Status?

Could you provideioreg+origin ACPI tables when HDMI plugin for me? Thank you.

Use ioreg from here

Thank you, syscl

[R1MnTnA]

@syscl If we do plug in a USB-C adapter of any kind, such as to HDMI, do we need to dump origin and rerun the script, to make it work correctly?

Also just as a side note.- Did you know you left your serial number that was generated for SMBIOS in the config.plist or is that just generic one? I read that if others use the same one that you could possibly get it blacklisted and then have problems with Apple sign in, message service or iCloud. (If you use those.) I generated a new one just in case.

[syscl]

@R1MnTnA, no you don't need to do that any time you plug in USB-C adapter.

Actually you should go to AppleStore and capture SMBIOS information in System Report->Hardware Overview. That's best for your laptop :)

syscl

[furrybanana]

Hi, hope its okay to hijack this post with different USB-C dongles (btw reinstalled and running fine on latest deploy). First of all: USB-C Network and USB-Hub do only work if they are plugged in and the device is booted. If I try to plug in this USB-C Hub from Anker when the system was booted without, nothing happens. My USB-C HDMI Adapter from Aukey is working fine when not plugged at boot.

I attached two pictures:

1. Anker USB-C Hub: 3x USB3.0 + Gigabit RJ45, model A8303 https://www.amazon.com/Anker-Ethernet-Including-ChromeBook-Aluminum/dp/B00XU6PM68

2. Aukey USB-C HDMI Adapter, model DE-CB-C40 https://www.amazon.com/AUKEY-Adapter-Supported-MacBook-Chromebook/dp/B00W98IC88/

[syscl]

@furrybanana Thank you, could you please provide screenshot of Thunderbolt?

Though I found on a real MacBook9,1, there's no device in Thunderbolt. But I am still curious:)

Thank you, syscl

[furrybanana]

Yeah its empty. 🥇

[syscl]

@furrybanana How about plug in thunderbolt then boot your laptop? Still blank in thunderbolt?

syscl

[furrybanana]

yes still like this:

[syscl]

@furrybanana Thank you.

syscl

[thiscodedbox]

I've reopened the issue and changed the title to include both things to keep us on topic. I will get the ioreg information to you @syscl soon. @furrybanana it would be awesome if we can solve the USB-C hotplug issue (usb not working when plugged in after startup) I move around with the notebook and plug into the Dell USB-C WD15 dock when I sit at my desk. it exhibits the same behavior that you describe.

[thiscodedbox]

@syscl - hopefully this is enough data. Let me know if you want more. I'm going to go digging in the DSDT to see if there's a method that applies to USB3.0 but not to USB3.1, but I'm still learning about ACPI. thiscodedbox_ioreg_1_26_2017_XPS_13_9350_10_12_3.zip

[thiscodedbox]

on the XPS 15 9550 the USB-C port in windows (much like the TB3 port) supposedly turns (fully) on when something is plugged into it. It stood to reason the XPS 13 9350 is the same way. The guys in this thread claim to have hotplug working: http://www.insanelymac.com/forum/topic/319764-guide-dell-xps-15-9550-sierra-10122-quick-installation/page-2

They found the the USB-C port creates a bridge device on RP15 on the XPS 15. Same on the XPS 13, but it's RP01 instead.

I took their SSDT-TB, compiled it, and replaced RP15 with RP01. I also tossed in their three relevant patches from their Clover config.plist

I had to remove SSDT-4 as a lot of the code was repeated from there and it was doing weird stuff. After I rebooted I saw the USB 3.1 controller as a removable PCI device - like an ExpressCard used to be in OSX (I used one for firewire on my old x230 so I recognized it). When I 'powered it off' from the menu it killed USB-C, and even the 3.1 controller disappeared from the system entirely!

Alas - still no hot swap - but it seems like this is at least addressing it. Might be able to finagle their SSDT-TB work into our different XPS 13 ecosystem.

Also, I noticed a difference in IOreg when the USB-C is plugged back in after being removed (attempted hotswap.) IOreg has different properties for AppleUSBXHCIPCI. This is even without the SSDT-TB.aml I was using.

before (working) / after (not working):

CapabilityFlags 0x8000 / 0x10000 CurrentPowerState 0x3 / 0x2 DesiredPowerState n/a / 0x3 MachineState n/a / 0x8

Just FYI. I'm still understanding ACPI etc. So I'm not sure how useful this is. attached is my modified SSDT-TB-RP01.aml

SSDT-TB-RP01.aml.zip

Here are the patches I tried (Credit all those in the thread above):

<dict>
<key>Comment</key>
<string>rename _RMV to XRMV</string>                                        
<key>Disabled</key>
<false/>
<key>Find</key>
<data>
X1JNVg==
</data>
<key>Replace</key>
<data>
WFJNVg==
</data>
</dict>

<dict>
<key>Signature</key>
<string>SSDT</string>
<key>TableId</key>
<string>xh_rvp10</string>
</dict>

dict>
<key>Comment</key>
<string>10.12.2 fudge USB 3.1 device ID</string>
<key>Disabled</key>
<false/>
<key>Find</key>
<string>30783135623638303836</string>
<key>InfoPlistPatch</key>
<true/>
<key>Name</key>
<string>AppleUSBXHCIPCI</string>
<key>Replace</key>
<string>30783135623538303836</string>
</dict>

[syscl]

@thiscodedbox The problem is _RMV -> XRMV, it is very possible this method will affect other devices as well(_RMV lies in other devices as well). Thus not a good solution. I will try to find better solution.

syscl

[thiscodedbox]

Are you suggesting we should patch the specific _RMV entities in the ssdt directly via maciasl rather than hotpatching all of ACPI?

[syscl]

@thiscodedbox try this SSDT SSDT-4.aml.zip

• Drop xh_rvp07

• Apply 10.12.2 fudge USB 3.1 device ID patch you provided above.

Report if this work, syscl

[thiscodedbox]

@syscl Kernel Panic on AppleUSBXHCIPCI with and without 10.12.2 fudge usb patch.

I noticed that the fudge patch patches 0x15b68086 to 0x15b58086.

the USB 3.1 device as detected on my system (and @furrybanana 's system from his screen shots above) is 0x15b58086 already - do you know if the fudge patch is meant to patch the a plist for the driver to attach to our device, or is it supposed to patch the system's detection (Inject a FakePCIid?)

If the latter, should the patch be reversed? Or, should we make a custom fakePCIID.kext to patch our system?

[syscl]

@thiscodedbox What cause the panic? The SSDT-4 or the kext patch?

OK, thank you for the information, I will refine it later today. I have due tomorrow.

Indeed, the kext patch need to be refined to fit our USB 3.1 device.

syscl

[thiscodedbox]

@syscl SSDT-4 - with just the patch boot is OK.

[syscl]

@thiscodedbox So, when AppleUSBXHCIPCI causes panic? Could you please give me some more details?

Thank you, syscl

[thiscodedbox]

Sorry - I think I had a partial script run issue, I redid my config.plist and no KP anymore, but, unfortunately, no hotplug either.

[syscl]

@thiscodedbox Thank you, due to the id you mention above, I think we need to change the patch code.

syscl

[furrybanana]

I did read some wikis on arch and they had the same issues mention.

https://wiki.archlinux.org/index.php/Dell_XPS_13_(9350)

a bug report for linux kernel: https://bugzilla.kernel.org/show_bug.cgi?id=115121

a patch/ fix: http://marc.info/?l=linux-acpi&m=145929159015853&w=2 confirming its working: https://bugzilla.kernel.org/show_bug.cgi?id=115121#c27

it's above my skill but what is possible with the information? can somebody port it? thank you both @syscl and @thiscodedbox

[furrybanana]

Tested my Fantec ALU-25U31 USB 3.1 USB-C Case. It only works when booting plugged in. After ejecting the drive it won't be recognized again. I didn't do any benchmark because I just had a spare 250gb SATA-II HDD installed. I will benchmark it with a 500gb Samsung Evo SSD later when I find the time. It is connected with a Belkin F2CU030BT1M-BLK 3.1 USB C to USB C cable (0,9m).

http://www.belkin.com/us/F2CU030-Belkin/p/P-F2CU030/

http://www.fantec.de/en/products/storage-devices/hard-drive-cases/25-inch-hard-drive-cases/produkt/details/artikel/1795_fantec_alu_25u31/

[R1MnTnA]

Hi @syscl, not sure if this is any help to this topic, but I found this on the forums. - https://www.tonymacx86.com/threads/usb-c-hotplug-questions.211313/

[kultigspritzig]

If it's of any help, a USB to HDMI dongle I have seems to be working fine for the most part, except for sometimes the external display will not wake up at all after booting.

I do not see a trace of it anywhere in the System Information and the Graphscs/Displays section reports as if it was connected via Thunderbolt.

Can get more information if needed.

[R1MnTnA]

@kultigspritzig do you have audio also working with that adapter? If so which one do you have? Does anyone else have there USB-C to HDMI adapters working with video and audio?

[thiscodedbox]

@R1MnTnA - I have a Dell WD15 dock which has Realtek USB2.0 Audio working out of the box - no kext needed. . My other dongle/adapter doesn't have audio - only ethernet/hdmi/usb3 - which also work fine. (no hot-plugging yet though)

(ignore my too-many audio devices - I have a usb audio interface and virtual interfaces for remote access)

[R1MnTnA]

@thiscodedbox ok thanks for this info. Although what I was hoping or looking for was video and audio straight out of HDMI with an adapter, but it looks as though this is still not possible. So are you saying that the 2nd adapter you mentioned doesn't have a separate audio plug but HDMI only has video output no audio and video to a TV? Ultimately I would like to watch and hear movies directly from USB-C/TB3 to HDMI.

[kultigspritzig]

@R1MnTnA Yeah, sound works via USB-C.

[R1MnTnA]

@syscl Strange behavior with my USB-C to HDMI adapter. Keeps disconnecting. Possibly incompatible? Video works audio does not.- 000281.621945 IOUSBHostHIDDevice: IOUSBHostHIDDevice::interruptRetry: resetting device due to IO failures 000281.696371 AppleUSBHostResources@: AppleUSBHostResources::allocateDownstreamBusCurrentGated: assuming successful wakeUnits 150 sleepUnits 0 000281.701769 AppleUSBHostResources@: AppleUSBHostResources::allocateDownstreamBusCurrentGated: assuming successful wakeUnits 150 sleepUnits 0 Report bytes: input:1 output:5 feature:129Collections : 2Buttons : input:0 output:0 feature:0Values : input:0 output:1 feature:2Element value capacity 3832Report count: 4Report ID: 0 input:0 output:0 feature:0Report ID: 238 input:8 output:40 feature:8Report ID: 239 input:8 output:8 feature:136Report ID: 237 input:8 output:8 feature:1032ApplePS2Keyboard: sending key e=33 down

[kultigspritzig]

BTW Anyone tried with 1.4.13 BIOS released in Feb? All changes Dell lists for this release are related to USB Type-C and Dell docking stations.

[thiscodedbox]

Yep - I'm on it for the past week or two - I haven't noticed anything better/worse regarding my WD15 dock or a generic HDMI/gigabit/USB dongle in either Windows 7 x64 or 10.12.3

On Mon, Feb 27, 2017 at 8:40 PM, kultigspritzig notifications@github.com wrote:

BTW Anyone tried with 1.4.13 BIOS released in Feb? The only changes Dell lists for this release are related to USB-Type C an Dell docking stations.

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/syscl/XPS9350-macOS/issues/33#issuecomment-282940379, or mute the thread https://github.com/notifications/unsubscribe-auth/AOfzCNaCDbSaq3vrtYBjisi_ybiXWZ5Jks5rg6UogaJpZM4Lt9rk .

[dpassmor]

Hey guys, I wrote the original SSDT-TB and accompanying patch for my Dell XPS 15 9550. I hope it's working for most. It's not perfect, since macOS doesn't like a USB controller disappearing without being powered down properly, but I don't think that's something that can be solved via ACPI. Still looking into it though...

My question is, has anyone tried this with an actual Thunderbolt (not USB-C) device? I haven't gotten my hands on one yet to test.

[thiscodedbox]

Hi @dpassmor - I was just thinking of reaching out to you but I haven't done all my due diligence to know what I'm talking about. I haven't gotten my hands on a thunderbolt 3 device to test either. AFAIK USB-C hotswap is not working for anyone on the 9350 yet - I can plug in to the 3rd party dongle and/or the dell WD15 dock (both USB-C 3.1 devices, not TB) and get USB-PD charging and HDMI/DP video hotswap (these seem to be passive features, pass-through and not software/ACPI dependent), but the USB ports, USB audio, and USB LAN on the Dock/dongle are only detected if plugged in at boot, and when removed, will not come back without a reboot.

[dpassmor]

The only USB-C device I have is a flash drive, which works with hot plugging on my 9550 with 10.12.3. I'm not sure how the dock is implemented; can you attach an ioreg dump when it is plugged in?

[syscl]

Hello @dpassmor ,

How are you?

I want to leave a big THANK YOU to you for the nice SMBIOS workaround and tiny SSDT for HWP, thank you!

BTW, thanks for the SSDT-TB, I took a look in it this afternoon, and refined it SSDT-XHC.dsl.zip

The refined version works pretty well on XPS 13 9350(may work well on XPS with Skylake/Kabylake models as well). What I changed in the new SSDT-XHC(based on your work):

• Remove Device(HS12), actually, we don't need that way to inject camera device. We have more general solution to fix webcam problem
• Append absent devices with _UPC and _PLD such that XPS 13 and XPS 15 can share the same SSDT-XHC
• Add MBSD method such that our system support SuperDrive
• Some cleanup...

As a side note, I still can't test thunderbolt because I don't have any thunderbolt devices. And I don't know if this new SSDT-XHC fixes the hot plug issue or not? Could somebody do me a favor to show a working thunderbolt ioreg?

Have a nice day, syscl

[maz-1]

Hello @dpassmor ,

How are you?

I want to leave a big THANK YOU to you for the nice SMBIOS workaround and tiny SSDT for HWP, thank you!

BTW, thanks for the SSDT-TB, I took a look in it this afternoon, and refined it SSDT-XHC.dsl.zip

The refined version works pretty well on XPS 13 9350(may work well on XPS with Skylake/Kabylake models as well). What I changed in the new SSDT-XHC(based on your work):

* Remove Device(HS12), actually, we don't need that way to inject camera device. We have more general solution to fix webcam problem

* Append absent devices with _UPC and _PLD such that XPS 13 and XPS 15 can share the same SSDT-XHC

* Add MBSD method such that our system support SuperDrive

* Some cleanup...

As a side note, I still can't test thunderbolt because I don't have any thunderbolt devices. And I don't know if this new SSDT-XHC fixes the hot plug issue or not? Could somebody do me a favor to show a working thunderbolt ioreg?

Have a nice day, syscl

Why defining _RMV and _DSM method under RP15? AFAIK the ACPI path of type-c controller is SB.PCI0.RP01.PXSX

[maz-1]

With this ssdt I get type-c hotplug, but still have to plug in a type-c device before boot, and will lost type-c availability after wake up from sleep. SSDT-TYPC.zip