External screen failure on Dell XPS 7390 2-in-1 (Ice Lake i7-1065G7 CPU) and Thunderbolt dock

[1.] One line summary of the problem:

After anywhere from 15min to two hours of use, external screens attached via the Thunderbolt dock go dark and have to be disabled/reenabled (or hard reboot) to get a signal again.

[2.] Full description of the problem/report:

After some time running 2 external monitors over the Thunderbolt 3 dock attached to the Dell XPS 7390 2-in-1, both external monitors suddenly lose their signal. The screens are still "seen" by the laptop, but display no picture (and the monitors go into power save thinking no device is attached).

The internal laptop screen continues to work with no issue.

Following messages can be seen in the journal:

[drm:intel_cpu_fifo_underrun_irq_handler [i915]] *ERROR* CPU pipe C FIFO underrun

or

[drm:intel_cpu_fifo_underrun_irq_handler [i915]] *ERROR* CPU pipe B FIFO underrun

Disabling and reenabling the screen in the GNOME display configuration sometimes fixes the issue, other times the entire device needs to be powered down and restarted for them to work again.

This issue persisted from kernel 5.4 all the way to the most recent version (5.5.9) on Arch Linux.

[3.] Keywords (i.e., modules, networking, kernel):

i915, firmware, intel UHD graphics

[4.] Kernel information

[4.1.] Kernel version (from /proc/version):

Linux version 5.5.9-arch1-2 (linux@archlinux) (gcc version 9.3.0 (Arch Linux 9.3.0-1)) #1 (moved) SMP PREEMPT Thu, 12 Mar 2020 23:01:33 +0000

[5.] Most recent kernel version which did not have the bug:

This issue persisted ever since I used the Thunderbolt dock with the device, from kernel 5.4 onwards.

[6.] Output of error message (found in dmesg):

[drm:intel_cpu_fifo_underrun_irq_handler [i915]] *ERROR* CPU pipe C FIFO underrun

[7.] A small shell script or example program which triggers the problem (if possible)

The problem cannot be reproduced easily, but seems to be triggered sometimes by a click or other activity in Chromium.

[8.] Environment

Arch Linux with GNOME desktop/Wayland

[8.1.] Software (add the output of the ver_linux script here)

[8.2.] Processor information (from /proc/cpuinfo):

cat /proc/cpuinfo

processor	: 0
vendor_id	: GenuineIntel
cpu family	: 6
model		: 126
model name	: Intel(R) Core(TM) i7-1065G7 CPU @ 1.30GHz
stepping	: 5
microcode	: 0x46
cpu MHz		: 825.986
cache size	: 8192 KB
physical id	: 0
siblings	: 8
core id		: 0
cpu cores	: 4
apicid		: 0
initial apicid	: 0
fpu		: yes
fpu_exception	: yes
cpuid level	: 27
wp		: yes
flags		: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx pdpe1gb rdtscp lm constant_tsc art arch_perfmon pebs bts rep_good nopl xtopology nonstop_tsc cpuid aperfmperf tsc_known_freq pni pclmulqdq dtes64 monitor ds_cpl vmx est tm2 ssse3 sdbg fma cx16 xtpr pdcm pcid sse4_1 sse4_2 x2apic movbe popcnt tsc_deadline_timer aes xsave avx f16c rdrand lahf_lm abm 3dnowprefetch cpuid_fault epb invpcid_single ssbd ibrs ibpb stibp ibrs_enhanced tpr_shadow vnmi flexpriority ept vpid ept_ad fsgsbase tsc_adjust bmi1 avx2 smep bmi2 erms invpcid avx512f avx512dq rdseed adx smap avx512ifma clflushopt intel_pt avx512cd sha_ni avx512bw avx512vl xsaveopt xsavec xgetbv1 xsaves dtherm ida arat pln pts hwp hwp_act_window hwp_epp hwp_pkg_req avx512vbmi umip pku ospke avx512_vbmi2 gfni vaes vpclmulqdq avx512_vnni avx512_bitalg avx512_vpopcntdq rdpid md_clear flush_l1d arch_capabilities
bugs		: spectre_v1 spectre_v2 spec_store_bypass swapgs itlb_multihit
bogomips	: 2996.00
clflush size	: 64
cache_alignment	: 64
address sizes	: 39 bits physical, 48 bits virtual
power management:

[8.3.] Module information (from /proc/modules):

i915 2449408 18 - Live 0x0000000000000000
intel_gtt 24576 1 i915, Live 0x0000000000000000
i2c_algo_bit 16384 1 i915, Live 0x0000000000000000
drm_kms_helper 237568 1 i915, Live 0x0000000000000000
drm 557056 10 i915,drm_kms_helper, Live 0x0000000000000000

[8.4.] Loaded driver and hardware information (/proc/ioports, /proc/iomem)

cat /proc/ioports

0000-0000 : PCI Bus 0000:00
  0000-0000 : dma1
  0000-0000 : pic1
  0000-0000 : iTCO_wdt
    0000-0000 : iTCO_wdt
  0000-0000 : timer0
  0000-0000 : timer1
  0000-0000 : keyboard
  0000-0000 : keyboard
  0000-0000 : rtc0
  0000-0000 : dma page reg
  0000-0000 : pic2
  0000-0000 : dma2
  0000-0000 : fpu
  0000-0000 : iTCO_wdt
    0000-0000 : iTCO_wdt
  0000-0000 : pnp 00:00
  0000-0000 : PNP0C09:00
    0000-0000 : EC data
  0000-0000 : PNP0C09:00
    0000-0000 : EC cmd
0000-0000 : PCI conf1
0000-0000 : PCI Bus 0000:00
  0000-0000 : pnp 00:00
  0000-0000 : ACPI PM1a_EVT_BLK
  0000-0000 : ACPI PM1a_CNT_BLK
  0000-0000 : ACPI PM_TMR
  0000-0000 : ACPI CPU throttle
  0000-0000 : ACPI PM2_CNT_BLK
  0000-0000 : pnp 00:02
  0000-0000 : ACPI GPE0_BLK
  0000-0000 : pnp 00:07
  0000-0000 : PCI Bus 0000:5a
  0000-0000 : 0000:00:02.0
  0000-0000 : PCI Bus 0000:01
    0000-0000 : PCI Bus 0000:02
      0000-0000 : PCI Bus 0000:04
  0000-0000 : PCI Bus 0000:2e
  0000-0000 : 0000:00:1f.4
    0000-0000 : i801_smbus

[8.5.] PCI information ('lspci -vvv' as root)

See lspci.txt

[8.6.] SCSI information (from /proc/scsi/scsi)

Empty

[8.7.] Other information that might be relevant to the problem (please look in /proc and include all information that you think to be relevant):

The driver generally has a problem with external monitors when the device is booted with the lid closed. The external monitors have no signal and the internal display turns on and off every second, with the journal printing:

Mär 09 11:05:03 notebook gnome-shell[2004]: Failed to post KMS update: Page flip of 91 failed, and no mode set available
Mär 09 11:05:04 notebook gnome-shell[2004]: Failed to post KMS update: Failed to set mode 2560x1440 on CRTC 91: Invalid argument

Although this seems to be a separate issue. Booting with the lid opened causes the issue not to appear.

[X.] Other notes, patches, fixes, workarounds:

None.