Bug 111250 - i915 external monitor not working in usb-c/thunderbolt dock on X1 gen6
Summary: i915 external monitor not working in usb-c/thunderbolt dock on X1 gen6
Status: NEW
Alias: None
Product: DRI
Classification: Unclassified
Component: DRM/Intel (show other bugs)
Version: XOrg git
Hardware: x86-64 (AMD64) Linux (All)
: high major
Assignee: Stanislav Lisovskiy
QA Contact: Intel GFX Bugs mailing list
URL:
Whiteboard: Triaged, ReadyForDev
Keywords:
Depends on:
Blocks:
 
Reported: 2019-07-29 14:50 UTC by tmai
Modified: 2019-09-17 14:19 UTC (History)
2 users (show)

See Also:
i915 platform: KBL
i915 features: display/Other


Attachments
dmesg while reproducing the error twice (312.44 KB, text/plain)
2019-07-29 14:50 UTC, tmai
no flags Details
dmesg with drm-tip kernel from 06-aug-2019 (310.12 KB, text/plain)
2019-08-06 12:24 UTC, tmai
no flags Details
dmesg with drm-tip kernel from 29-aug-2019 (7.87 MB, text/plain)
2019-08-29 11:04 UTC, tmai
no flags Details
dmesg with drm-tip kernel from 29-aug-2019 shorter version (4.53 MB, text/plain)
2019-08-29 11:17 UTC, tmai
no flags Details
dmesg dump after booting to console (1.45 MB, text/plain)
2019-09-06 14:58 UTC, tmai
no flags Details
Experimental patch to check if this is a race condition when USB dock is plugged (996 bytes, patch)
2019-09-16 11:55 UTC, Stanislav Lisovskiy
no flags Details | Splinter Review
dmesg with drm-tip kernel from 16-sep-2019 with patch, booted to console, noname usb-c dock (301.75 KB, text/plain)
2019-09-17 14:00 UTC, tmai
no flags Details
dmesg with drm-tip kernel from 16-sep-2019 with patch, booted to console, lenovo dock (2.80 MB, text/plain)
2019-09-17 14:02 UTC, tmai
no flags Details
dmesg with drm-tip kernel from 16-sep-2019 with patch, booted to gui (kde), lenovo dock (6.37 MB, text/plain)
2019-09-17 14:03 UTC, tmai
no flags Details

Note You need to log in before you can comment on or make changes to this bug.
Description tmai 2019-07-29 14:50:55 UTC
Created attachment 144909 [details]
dmesg while reproducing the error twice

Hi,

I cannot get external monitors to work on my X1 gen6 on Linux (Ubuntu) when they are connected over a dock. I tried the official dock from lenovo as well as an third-party usb-c adapter. The systems immediately freezes if I hotplug an external monitor or it freezes when I try to configure the external monitor (which is not enabled in the settings up to this point). It unfreezes when the cable is disconnected. In some cases booting with the external monitor ended up at a log-in screen with reduced screen resolution (below full HD).

Kernel: 5.3.0-994-generic
(from here: https://kernel.ubuntu.com/~kernel-ppa/mainline/drm-tip/2019-07-27/)
However, it affects various other kernel versions as well.

OS: Kubuntu 19.04

Machine: Lenovo X1 (Gen6); 20KGS03800

Display connector: HDMI and DP

dmesg is attached (external monitor was connected twice after boot up)

Additional info:

External monitors were working on Kubuntu 18.04 earlier this year. The other day, I tried the livesystem 18.04.2 which is already affected by the bug.

Please, let me know if I can provide any further information from my end or point me to any other place where I should report this bug instead.

Thanks to anyone who looks into this.
Comment 1 Lakshmi 2019-08-01 12:00:05 UTC
IS the attached log is from drmtip? If not, can you please verity this issue with drmtip (https://cgit.freedesktop.org/drm-tip)?
Comment 2 tmai 2019-08-05 11:06:02 UTC
I used the drm-tip kernel from the Ubuntu team.

Kernel: 5.3.0-994-generic
(from here: https://kernel.ubuntu.com/~kernel-ppa/mainline/drm-tip/2019-07-27/)

I just tried to compile the kernel from https://cgit.freedesktop.org/drm-tip but could not boot it afterwards. Not sure what exactly went wrong there.

Is it necessary to compile the linked kernel or is the version I used sufficient/close enough?
Comment 3 Lakshmi 2019-08-06 08:32:52 UTC
(In reply to tmai from comment #2)
> I used the drm-tip kernel from the Ubuntu team.
> 
> Kernel: 5.3.0-994-generic
> (from here:
> https://kernel.ubuntu.com/~kernel-ppa/mainline/drm-tip/2019-07-27/)
> 
> I just tried to compile the kernel from https://cgit.freedesktop.org/drm-tip
> but could not boot it afterwards. Not sure what exactly went wrong there.
> 
> Is it necessary to compile the linked kernel or is the version I used
> sufficient/close enough?

It is recommended to verify the issue with https://cgit.freedesktop.org/drm-tip.
In general here are the instructions to build kernel, check if that helps you. 
https://kernelnewbies.org/KernelBuild
Comment 4 tmai 2019-08-06 12:24:05 UTC
Created attachment 144956 [details]
dmesg with drm-tip kernel from 06-aug-2019

I finally managed to compile and boot a new kernel. I used 'make INSTALL_MOD_STRIP=1 modules_install' to reduce the initram-img and get it to boot.

Before I copied the attached dmesg I booted my system, plugged in a display-port cable from my monitor and unplugged it a few seconds later after the system froze. This got it to unfreeze and let me prepare the dmesg dump.
Comment 5 tmai 2019-08-29 07:04:03 UTC
This bug also affects fedora 30 booted from a live stick (sry, I did not compile a drm-tip kernel for that OS, too).

It also is reproducible on an official Lenovo docking station and no-name usb-c dock as well.

This error description on a non-lenovo machine could be related to my problem: https://bugs.freedesktop.org/show_bug.cgi?id=110629
Comment 6 tmai 2019-08-29 11:04:19 UTC
Created attachment 145204 [details]
dmesg with drm-tip kernel from 29-aug-2019

reproduced problem with hdmi and displayport cable
Comment 7 tmai 2019-08-29 11:17:24 UTC
Created attachment 145205 [details]
dmesg with drm-tip kernel from 29-aug-2019 shorter version
Comment 8 tmai 2019-08-29 11:19:33 UTC
I created a new dmesg-dump with the drm-tip kernel 5.3.0-rc6+

Sine the last dump is now a few weeks old.

Please let me know, if I can provide anything else to support your work or if I should post this bug somewhere else, additionally.
Comment 9 Stanislav Lisovskiy 2019-09-06 08:35:23 UTC
In dmesg it looks like whenever it tries to detect DP the dpcd read times out:

[   11.576987] [drm:intel_dp_detect [i915]] [CONNECTOR:92:DP-1]
[   11.585363] [drm:intel_dp_aux_xfer [i915]] dp_aux_ch timeout status 0x7d4003ff
[   11.593726] [drm:intel_dp_aux_xfer [i915]] dp_aux_ch timeout status 0x7d4003ff
[   11.602089] [drm:intel_dp_aux_xfer [i915]] dp_aux_ch timeout status 0x7d4003ff
...
[   11.819474] [drm:intel_dp_aux_xfer [i915]] dp_aux_ch timeout status 0x7d4003ff
[   11.827834] [drm:intel_dp_aux_xfer [i915]] dp_aux_ch timeout status 0x7d4003ff
[   11.836194] [drm:intel_dp_aux_xfer [i915]] dp_aux_ch timeout status 0x7d4003ff
[   11.844556] [drm:intel_dp_aux_xfer [i915]] dp_aux_ch timeout status 0x7d4003ff
[   11.844561] [drm:drm_dp_dpcd_access [drm_kms_helper]] Too many retries, giving up. First error: -110


Can you boot in console mode and check how it works then? (add "3" after "drm.debug" in the kernel command line).
Comment 10 tmai 2019-09-06 14:58:47 UTC
Created attachment 145279 [details]
dmesg dump after booting to console

I added the parameter 3 to boot to console which resulted in the attached dmesg_dump (I re-used the last drm-tip kernel I compiled).

After attaching the monitor to the docking station I got a couple of FIFO underrun messages directly on the console. I took a photo, however, they are also reported in the dmesg dump.
Comment 11 Stanislav Lisovskiy 2019-09-09 12:51:09 UTC
(In reply to tmai from comment #10)
> Created attachment 145279 [details]
> dmesg dump after booting to console
> 
> I added the parameter 3 to boot to console which resulted in the attached
> dmesg_dump (I re-used the last drm-tip kernel I compiled).
> 
> After attaching the monitor to the docking station I got a couple of FIFO
> underrun messages directly on the console. I took a photo, however, they are
> also reported in the dmesg dump.

Did the external display show anything in console mode or the system freeze? FIFO underruns are irrelevant to the problem and can happen for short time during modeset.
Comment 12 tmai 2019-09-09 13:08:35 UTC
No, the monitor did not show anything. More precisely, I tried two different monitors and two different cables. None worked.
Comment 13 Stanislav Lisovskiy 2019-09-10 07:43:45 UTC
(In reply to tmai from comment #12)
> No, the monitor did not show anything. More precisely, I tried two different
> monitors and two different cables. None worked.

There might be a race condition between fbdev polling the outputs and then executing probing and usb hub discovery, because what I see in the log is that we get a hotplug event and start probing, before the usb hub is even completely discovered:

  2.788336] kernel: [drm:drm_fb_helper_hotplug_event.part.0 [drm_kms_helper]] 

here goes fbdev hotplug event and we start probing:

[    2.788346] kernel: [drm:drm_client_modeset_probe [drm]] 
[    2.788353] kernel: [drm:drm_helper_probe_single_connector_modes [drm_kms_helper]] [CONNECTOR:86:eDP-1]
[    2.788378] kernel: [drm:intel_dp_detect [i915]] [CONNECTOR:86:eDP-1]
[    2.788434] kernel: [drm:intel_dp_print_rates [i915]] source rates: 162000, 216000, 270000, 324000, 432000, 540000
[    2.788473] kernel: [drm:intel_dp_print_rates [i915]] sink rates: 162000, 270000
[    2.788497] kernel: [drm:intel_dp_print_rates [i915]] common rates: 162000, 270000
[    2.788545] kernel: [drm:intel_dp_detect [i915]] MST support? port A: no, sink: no, modparam: yes
[    2.789155] kernel: [drm:drm_add_display_info [drm]] non_desktop set to 0
[    2.789168] kernel: [drm:drm_add_edid_modes [drm]] ELD: no CEA Extension found
[    2.789176] kernel: [drm:drm_add_display_info [drm]] non_desktop set to 0
[    2.789195] kernel: [drm:drm_helper_probe_single_connector_modes [drm_kms_helper]] [CONNECTOR:86:eDP-1] probed modes :
[    2.789205] kernel: [drm:drm_mode_debug_printmodeline [drm]] Modeline "1920x1080": 60 141000 1920 1936 1952 2104 1080 1083 1097 1116 0x48 0xa

here we start detecting DP-1, despite usb device discovery hasn't completed yet

[    2.789210] kernel: [drm:drm_helper_probe_single_connector_modes [drm_kms_helper]] [CONNECTOR:92:DP-1]
[    2.789241] kernel: [drm:intel_dp_detect [i915]] [CONNECTOR:92:DP-1]
[    2.797582] kernel: [drm:intel_dp_aux_xfer [i915]] dp_aux_ch timeout status 0x7d4003ff
[    2.805917] kernel: [drm:intel_dp_aux_xfer [i915]] dp_aux_ch timeout status 0x7d4003ff
[    2.814248] kernel: [drm:intel_dp_aux_xfer [i915]] dp_aux_ch timeout status 0x7d4003ff
[    2.822575] kernel: [drm:intel_dp_aux_xfer [i915]] dp_aux_ch timeout status 0x7d4003ff
[    2.830902] kernel: [drm:intel_dp_aux_xfer [i915]] dp_aux_ch timeout status 0x7d4003ff
[    2.839228] kernel: [drm:intel_dp_aux_xfer [i915]] dp_aux_ch timeout status 0x7d4003ff
[    2.847554] kernel: [drm:intel_dp_aux_xfer [i915]] dp_aux_ch timeout status 0x7d4003ff
[    2.855880] kernel: [drm:intel_dp_aux_xfer [i915]] dp_aux_ch timeout status 0x7d4003ff
[    2.864206] kernel: [drm:intel_dp_aux_xfer [i915]] dp_aux_ch timeout status 0x7d4003ff
[    2.872531] kernel: [drm:intel_dp_aux_xfer [i915]] dp_aux_ch timeout status 0x7d4003ff
[    2.880856] kernel: [drm:intel_dp_aux_xfer [i915]] dp_aux_ch timeout status 0x7d4003ff
[    2.889183] kernel: [drm:intel_dp_aux_xfer [i915]] dp_aux_ch timeout status 0x7d4003ff


here USB hub is found to be connected, however we are already trying to communicate with DP-1:

[    2.889485] kernel: usb 1-3: New USB device found, idVendor=17ef, idProduct=3074, bcdDevice= 0.00
[    2.889486] kernel: usb 1-3: New USB device strings: Mfr=1, Product=2, SerialNumber=3
[    2.889487] kernel: usb 1-3: Product: USB Billboard
[    2.889488] kernel: usb 1-3: Manufacturer: Cypress Semiconductor
[    2.889488] kernel: usb 1-3: SerialNumber: 11AD1D00B625480C19290B00
[    2.897508] kernel: [drm:intel_dp_aux_xfer [i915]] dp_aux_ch timeout status 0x7d4003ff
[    2.905834] kernel: [drm:intel_dp_aux_xfer [i915]] dp_aux_ch timeout status 0x7d4003ff
[    2.914159] kernel: [drm:intel_dp_aux_xfer [i915]] dp_aux_ch timeout status 0x7d4003ff
[    2.916129] kernel: tsc: Refined TSC clocksource calibration: 1991.999 MHz
[    2.916134] kernel: clocksource: tsc: mask: 0xffffffffffffffff max_cycles: 0x396d4ffc055, max_idle_ns: 881590662783 ns
[    2.916144] kernel: clocksource: Switched to clocksource tsc
[    2.922510] kernel: [drm:intel_dp_aux_xfer [i915]] dp_aux_ch timeout status 0x7d4003ff
[    2.930869] kernel: [drm:intel_dp_aux_xfer [i915]] dp_aux_ch timeout status 0x7d4003ff
[    2.939228] kernel: [drm:intel_dp_aux_xfer [i915]] dp_aux_ch timeout status 0x7d4003ff
[    2.947586] kernel: [drm:intel_dp_aux_xfer [i915]] dp_aux_ch timeout status 0x7d4003ff
[    2.955945] kernel: [drm:intel_dp_aux_xfer [i915]] dp_aux_ch timeout status 0x7d4003ff
[    2.964303] kernel: [drm:intel_dp_aux_xfer [i915]] dp_aux_ch timeout status 0x7d4003ff
[    2.972661] kernel: [drm:intel_dp_aux_xfer [i915]] dp_aux_ch timeout status 0x7d4003ff
[    2.981020] kernel: [drm:intel_dp_aux_xfer [i915]] dp_aux_ch timeout status 0x7d4003ff
[    2.989379] kernel: [drm:intel_dp_aux_xfer [i915]] dp_aux_ch timeout status 0x7d4003ff
[    2.997737] kernel: [drm:intel_dp_aux_xfer [i915]] dp_aux_ch timeout status 0x7d4003ff
[    3.006096] kernel: [drm:intel_dp_aux_xfer [i915]] dp_aux_ch timeout status 0x7d4003ff
[    3.014454] kernel: [drm:intel_dp_aux_xfer [i915]] dp_aux_ch timeout status 0x7d4003ff
[    3.016119] kernel: usb 1-4: new high-speed USB device number 4 using xhci_hcd
[    3.016287] kernel: usb 4-2.2: new SuperSpeedPlus Gen 2 USB device number 3 using xhci_hcd
[    3.022812] kernel: [drm:intel_dp_aux_xfer [i915]] dp_aux_ch timeout status 0x7d4003ff
[    3.031171] kernel: [drm:intel_dp_aux_xfer [i915]] dp_aux_ch timeout status 0x7d4003ff
[    3.039011] kernel: usb 4-2.2: New USB device found, idVendor=17ef, idProduct=3070, bcdDevice= a.74
[    3.039012] kernel: usb 4-2.2: New USB device strings: Mfr=1, Product=2, SerialNumber=0
[    3.039013] kernel: usb 4-2.2: Product: USB3.1 Hub             
[    3.039014] kernel: usb 4-2.2: Manufacturer: VIA Labs, Inc.         
[    3.039531] kernel: [drm:intel_dp_aux_xfer [i915]] dp_aux_ch timeout status 0x7d4003ff
[    3.041652] kernel: hub 4-2.2:1.0: USB hub found
[    3.041953] kernel: hub 4-2.2:1.0: 4 ports detected
[    3.047897] kernel: [drm:intel_dp_aux_xfer [i915]] dp_aux_ch timeout status 0x7d4003ff
[    3.056260] kernel: [drm:intel_dp_aux_xfer [i915]] dp_aux_ch timeout status 0x7d4003ff
[    3.056266] kernel: [drm:drm_dp_dpcd_access [drm_kms_helper]] Too many retries, giving up. First error: -110
[    3.056270] kernel: [drm:drm_helper_probe_single_connector_modes [drm_kms_helper]] [CONNECTOR:92:DP-1] disconnected

In normal situation, at least on my desktop it looks completely different - first we get USB hub discovered, then hotplug event and then we start probing.

Have no clue currently how to fix that, however may be adding some artificial delay, before doing actual probing might help and confirm that idea.

I will attach some patch soon, once I figure out what can be done.
Comment 14 Stanislav Lisovskiy 2019-09-16 11:55:04 UTC
Created attachment 145373 [details] [review]
Experimental patch to check if this is a race condition when USB dock is plugged
Comment 15 Stanislav Lisovskiy 2019-09-16 11:56:18 UTC
(In reply to tmai from comment #12)
> No, the monitor did not show anything. More precisely, I tried two different
> monitors and two different cables. None worked.

Please test if possible with the patch above, I see that intel_dp_detect is called simultaneously with usb hub being set up, which might result in some kind of race condition.
Comment 16 tmai 2019-09-17 14:00:51 UTC
Created attachment 145392 [details]
dmesg with drm-tip kernel from 16-sep-2019 with patch, booted to console, noname usb-c dock
Comment 17 tmai 2019-09-17 14:02:00 UTC
Created attachment 145393 [details]
dmesg with drm-tip kernel from 16-sep-2019 with patch, booted to console, lenovo dock
Comment 18 tmai 2019-09-17 14:03:45 UTC
Created attachment 145394 [details]
dmesg with drm-tip kernel from 16-sep-2019 with patch, booted to gui (kde), lenovo dock
Comment 19 tmai 2019-09-17 14:19:41 UTC
First of all, thanks for the patch and all the explaining. I am learning a few things here. :-)

I attached 3 new dmesg files. All were created with a fresh drm-tip kernel + patch.

A brief description of the results:

1. Test with noname-dock booted to console: The Monitor (via hdmi cable) was connected at a 1024xsomething resolution. Besides the low resolution everything seemed to work fine. Afterwards I rebooted to gui (kde). Same result, monitor was connected, but with a low resolution. I could not change the resolution via the kde settings menu.

2. Test with lenovo-dock booted to console: No change here. Still no connection to the monitor(s). I tried to monitors with two cables (hdmi and dp). The hdmi-monitor works fine when connected directly to the notebook. After playing around a little the system eventually crashed. I had to reboot with a hard shutdown.

3. Test with lenovo-dock booted to gui (kde): Same thing as with the console: no monitor worked via the dock. However, I noticed that the hdmi monitor connected to the laptop directly still showed a screen, when I connected the other monitor as via dp at the dock. In that scenario the laptop screen itself went dark, but still showed the mouse cursor.


Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.