5.3.0-46-generic - i915 - frequent GPU hangs / resets rcs0

[Steve Murphy]

Hi,

Since update to HWE kernel 5.3.0-46-generic I am experiencing frequent (every couple of minutes) GPU hangs and reset manifesting as 2-3 seconds freezes of the GUI (other than the mouse pointer).

No particular triggers identified although have Chrome / Chromium running with Hardware Acceleration enabled does appear to increase the frequency.

I have seen incidences of these hangs in jounralctl output using previous kernels in the 5.3.0-xx series but they were very infrequent (one or twice in a week of daily usage)

System Info
steve@steve-Inspiron-5580:~$ inxi -SCGxxxz
System: Host: steve-Inspiron-5580 Kernel: 5.3.0-46-generic x86_64 bits: 64 compiler: gcc v: 7.5.0
           Desktop: Cinnamon 4.4.8 wm: muffin 4.4.2 dm: LightDM 1.26.0 Distro: Linux Mint 19.3 Tricia
           base: Ubuntu 18.04 bionic
CPU: Topology: Quad Core model: Intel Core i5-8265U bits: 64 type: MT MCP arch: Kaby Lake rev: B
           L2 cache: 6144 KiB
           flags: avx avx2 lm nx pae sse sse2 sse3 sse4_1 sse4_2 ssse3 vmx bogomips: 28800
           Speed: 1173 MHz min/max: 400/3900 MHz Core speeds (MHz): 1: 800 2: 800 3: 800 4: 800 5: 800 6: 800 7: 800
           8: 800
Graphics: Device-1: Intel vendor: Dell driver: i915 v: kernel bus ID: 00:02.0 chip ID: 8086:3ea0
           Display: x11 server: X.Org 1.20.5 driver: modesetting unloaded: fbdev,vesa resolution: 1920x1080~60Hz
           OpenGL: renderer: Mesa DRI Intel UHD Graphics (Whiskey Lake 3x8 GT2) v: 4.5 Mesa 19.2.8 compat-v: 3.0
           direct render: Yes

steve@steve-Inspiron-5580:~$ journalctl -b | grep i915
Apr 10 06:15:17 steve-Inspiron-5580 kernel: i915 0000:00:02.0: vgaarb: deactivate vga console
Apr 10 06:15:17 steve-Inspiron-5580 kernel: i915 0000:00:02.0: vgaarb: changed VGA decodes: olddecodes=io+mem,decodes=io+mem:owns=io+mem
Apr 10 06:15:17 steve-Inspiron-5580 kernel: [drm] Finished loading DMC firmware i915/kbl_dmc_ver1_04.bin (v1.4)
Apr 10 06:15:17 steve-Inspiron-5580 kernel: [drm] Initialized i915 1.6.0 20190619 for 0000:00:02.0 on minor 0
Apr 10 06:15:17 steve-Inspiron-5580 kernel: fbcon: i915drmfb (fb0) is primary device
Apr 10 06:15:17 steve-Inspiron-5580 kernel: i915 0000:00:02.0: fb0: i915drmfb frame buffer device
Apr 10 06:15:17 steve-Inspiron-5580 kernel: snd_hda_intel 0000:00:1f.3: bound 0000:00:02.0 (ops i915_audio_component_bind_ops [i915])
Apr 10 06:16:28 steve-Inspiron-5580 kernel: i915 0000:00:02.0: GPU HANG: ecode 9:0:0x00000000, hang on rcs0
Apr 10 06:16:28 steve-Inspiron-5580 kernel: i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
Apr 10 06:31:46 steve-Inspiron-5580 kernel: i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
Apr 10 06:37:48 steve-Inspiron-5580 kernel: i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
Apr 10 06:40:46 steve-Inspiron-5580 kernel: i915 0000:00:02.0: Resetting rcs0 for hang on rcs0

I note another user has reported similar issues on the same kernel at https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1861395/comments/52
---
ProblemType: Bug
ApportVersion: 2.20.9-0ubuntu7.14
Architecture: amd64
AudioDevicesInUse:
 USER PID ACCESS COMMAND
 /dev/snd/controlC0: steve 3920 F.... pulseaudio
 /dev/snd/pcmC0D0p: steve 3920 F...m pulseaudio
CurrentDesktop: X-Cinnamon
DistroRelease: Linux Mint 19.3
HibernationDevice: RESUME=none
InstallationDate: Installed on 2019-12-27 (104 days ago)
InstallationMedia: Linux Mint 19.3 "Tricia" - Release amd64 20191213
MachineType: Dell Inc. Inspiron 5580
Package: linux (not installed)
ProcFB: 0 i915drmfb
ProcKernelCmdLine: BOOT_IMAGE=/boot/vmlinuz-5.3.0-46-generic root=UUID=b0eaa5bb-0276-42d4-938f-ee6ce1627906 ro usb_storage.quirks=0bc2:2320: quiet splash vt.handoff=1
ProcVersionSignature: Ubuntu 5.3.0-46.38~18.04.1-generic 5.3.18
RelatedPackageVersions:
 linux-restricted-modules-5.3.0-46-generic N/A
 linux-backports-modules-5.3.0-46-generic N/A
 linux-firmware 1.173.17
Tags: tricia
Uname: Linux 5.3.0-46-generic x86_64
UpgradeStatus: No upgrade log present (probably fresh install)
UserGroups: adm cdrom dip input lpadmin plugdev sambashare sudo vboxusers
_MarkForUpload: True
dmi.bios.date: 07/02/2019
dmi.bios.vendor: Dell Inc.
dmi.bios.version: 2.4.0
dmi.board.name: 0K0DFT
dmi.board.vendor: Dell Inc.
dmi.board.version: A00
dmi.chassis.type: 10
dmi.chassis.vendor: Dell Inc.
dmi.modalias: dmi:bvnDellInc.:bvr2.4.0:bd07/02/2019:svnDellInc.:pnInspiron5580:pvr:rvnDellInc.:rn0K0DFT:rvrA00:cvnDellInc.:ct10:cvr:
dmi.product.family: Inspiron
dmi.product.name: Inspiron 5580
dmi.product.sku: 089D
dmi.sys.vendor: Dell Inc.

[Ubuntu Kernel Bot (ubuntu-kernel-bot)]

This bug is missing log files that will aid in diagnosing the problem. While running an Ubuntu kernel (not a mainline or third-party kernel) please enter the following command in a terminal window:

apport-collect 1872001

and then change the status of the bug to 'Confirmed'.

If, due to the nature of the issue you have encountered, you are unable to run this command, please add a comment stating that fact and change the bug status to 'Confirmed'.

This change has been made by an automated script, maintained by the Ubuntu Kernel Team.

[Steve Murphy (smurphos)]

apport information

[Steve Murphy (smurphos)]

apport information

[Steve Murphy (smurphos)]

apport information

[Steve Murphy (smurphos)]

apport information

[Steve Murphy (smurphos)]

apport information

[Steve Murphy (smurphos)]

apport information

[Steve Murphy (smurphos)]

apport information

[Steve Murphy (smurphos)]

apport information

[Steve Murphy (smurphos)]

apport information

[Steve Murphy (smurphos)]

apport information

[Steve Murphy (smurphos)]

apport information

[Steve Murphy (smurphos)]

apport information

[Steve Murphy (smurphos)]

apport information

[Steve Murphy (smurphos)]

apport information

[Steve Murphy (smurphos)]

apport information

[Steve Murphy (smurphos)]

The issue does not manifest on mainline kernel - 5.5.16-050516-generic

[Nikke (nmellegard)]

Experiencing the exact same problem on a Dell XPS 13 9630 since upgrade to kernel 5.3.0-46.

When problem is triggered, I can still move the mouse pointer for 2-3 seconds but can't interact with the desktop or any apps (clicking or typing does nothing). After that the mouse pointer stops for about a second. And then it all comes back; the clicks and keys that I types during the freeze comes flushing back...

Also noticed that when I'm on a telco, sound and video is completely unaffected throughout the freeze.

System info
--> inxi -SCGxxxz

System:
  Host: Dinky Kernel: 5.3.0-46-generic x86_64 bits: 64 compiler: gcc
  v: 9.2.1 Desktop: Gnome 3.34.3 wm: gnome-shell dm: GDM3 3.34.1
  Distro: Ubuntu 19.10 (Eoan Ermine)
CPU:
  Topology: Dual Core model: Intel Core i7-7500U bits: 64 type: MT MCP
  arch: Kaby Lake rev: 9 L2 cache: 4096 KiB
  flags: avx avx2 lm nx pae sse sse2 sse3 sse4_1 sse4_2 ssse3 vmx
  bogomips: 23199
  Speed: 600 MHz min/max: 400/3500 MHz Core speeds (MHz): 1: 600 2: 600
  3: 600 4: 600
Graphics:
  Device-1: Intel HD Graphics 620 vendor: Dell driver: i915 v: kernel
  bus ID: 00:02.0 chip ID: 8086:5916
  Display: x11 server: X.Org 1.20.5 driver: i915 compositor: gnome-shell
  resolution: 1920x1080~60Hz
  OpenGL: renderer: Mesa DRI Intel HD Graphics 620 (Kaby Lake GT2)
  v: 4.5 Mesa 19.2.8 compat-v: 3.0 direct render: Yes

From kernel.log:
Apr 10 09:21:50 Dinky kernel: [40438.893463] i915 0000:00:02.0: Resetting rcs0 for hang on rcs0

[Viktor (lifeisgoodmf)]

Same for me.

inxi -SCGxxxz
System: Host: viktorz Kernel: 5.3.0-46-generic x86_64 bits: 64 gcc: 7.5.0
           Desktop: Gnome 3.28.4 (Gtk 3.22.30-1ubuntu4) info: gnome-shell dm: gdm3 Distro: Ubuntu 18.04.4 LTS
CPU: Quad core Intel Core i5-9400H (-MT-MCP-) arch: Skylake rev.13 cache: 8192 KB
           flags: (lm nx sse sse2 sse3 sse4_1 sse4_2 ssse3 vmx) bmips: 19999
           clock speeds: min/max: 800/4300 MHz 1: 900 MHz 2: 900 MHz 3: 900 MHz 4: 900 MHz 5: 900 MHz 6: 900 MHz
           7: 900 MHz 8: 900 MHz
Graphics: Card: Intel Device 3e9b bus-ID: 00:02.0 chip-ID: 8086:3e9b
           Display Server: x11 (X.Org 1.20.5 ) driver: i915 Resolution: 1920x1080@60.00hz, 1920x1080@60.00hz
           OpenGL: renderer: Mesa DRI Intel UHD Graphics 630 (Coffeelake 3x8 GT2)
           version: 4.5 Mesa 19.2.8 (compat-v: 3.0) Direct Render: Yes

syslog:
Apr 10 14:56:25 viktorz kernel: [18006.588904] i915 0000:00:02.0: Resetting rcs0 for hang on rcs0

[DnsF (dfraipont-gmx)]

Same for me.

$ inxi -SCGxxxz
System:
  Host: xxx Kernel: 5.3.0-46-generic x86_64 bits: 64 compiler: gcc
  v: 9.2.1 Desktop: Gnome 3.34.3 wm: gnome-shell dm: GDM3 3.34.1
  Distro: Ubuntu 19.10 (Eoan Ermine)
CPU:
  Topology: Quad Core model: Intel Core i7-6700K bits: 64 type: MT MCP
  arch: Skylake-S rev: 3 L2 cache: 8192 KiB
  flags: avx avx2 lm nx pae sse sse2 sse3 sse4_1 sse4_2 ssse3 vmx
  bogomips: 63999
  Speed: 3896 MHz min/max: 800/4200 MHz Core speeds (MHz): 1: 3896 2: 3894
  3: 3896 4: 3896 5: 3896 6: 3896 7: 3896 8: 3896
Graphics:
  Device-1: Intel HD Graphics 530 vendor: ASUSTeK driver: i915 v: kernel
  bus ID: 00:02.0 chip ID: 8086:1912
  Display: x11 server: X.Org 1.20.5 driver: modesetting unloaded: fbdev,vesa
  compositor: gnome-shell resolution: 1920x1080~60Hz
  OpenGL: renderer: Mesa DRI Intel HD Graphics 530 (Skylake GT2)
  v: 4.5 Mesa 19.2.8 compat-v: 3.0 direct render: Yes

[Boštjan Žokš (bzoks)]

Same for me.
root@hpbo:~# inxi -SCGxxxz
System: Host: hpbo Kernel: 5.3.0-46-generic x86_64 bits: 64 compiler: gcc v: 9.2.1 Console: tty 1 wm: kwin_x11 dm: SDDM
           Distro: Ubuntu 19.10 (Eoan Ermine)
CPU: Topology: Quad Core model: Intel Core i7-8565U bits: 64 type: MT MCP arch: Kaby Lake rev: C L1 cache: 256 KiB
           L2 cache: 8192 KiB L3 cache: 8192 KiB
           flags: avx avx2 lm nx pae sse sse2 sse3 sse4_1 sse4_2 ssse3 vmx bogomips: 31999
           Speed: 969 MHz min/max: 400/4600 MHz Core speeds (MHz): 1: 969 2: 824 3: 825 4: 742 5: 743 6: 786 7: 822 8: 732
Graphics: Device-1: Intel UHD Graphics 620 vendor: Hewlett-Packard driver: i915 v: kernel bus ID: 00:02.0 chip ID: 8086:3ea0
           Display: server: X.Org 1.20.5 driver: modesetting unloaded: fbdev,vesa compositor: kwin_x11
           resolution: 1920x1080~60Hz, 2560x1440~60Hz
           OpenGL: renderer: Mesa DRI Intel UHD Graphics (Whiskey Lake 3x8 GT2) v: 4.5 Mesa 19.2.8 compat-v: 3.0
           direct render: Yes

root@hpbo:~# journalctl -b | grep i915
apr 10 10:57:00 hpbo kernel: i915 0000:00:02.0: vgaarb: deactivate vga console
apr 10 10:57:00 hpbo kernel: i915 0000:00:02.0: vgaarb: changed VGA decodes: olddecodes=io+mem,decodes=io+mem:owns=io+mem
apr 10 10:57:00 hpbo kernel: [drm] Finished loading DMC firmware i915/kbl_dmc_ver1_04.bin (v1.4)
apr 10 10:57:00 hpbo kernel: [drm] Initialized i915 1.6.0 20190619 for 0000:00:02.0 on minor 0
apr 10 10:57:00 hpbo kernel: fbcon: i915drmfb (fb0) is primary device
apr 10 10:57:00 hpbo kernel: i915 0000:00:02.0: fb0: i915drmfb frame buffer device
apr 10 10:57:00 hpbo kernel: snd_hda_intel 0000:00:1f.3: bound 0000:00:02.0 (ops i915_audio_component_bind_ops [i915])
apr 10 16:22:09 hpbo kernel: i915 0000:00:02.0: GPU HANG: ecode 9:0:0x00000000, hang on rcs0
apr 10 16:22:09 hpbo kernel: i915 0000:00:02.0: Resetting rcs0 for hang on rcs0

[Boštjan Žokš (bzoks)]

Got identical issue on other PC with same kernel version and onboard graphics (i915), also HP.

root@gromk:~# inxi -SCGxxxz
System: Host: gromk Kernel: 5.3.0-46-generic x86_64 bits: 64 compiler: gcc v: 9.2.1 Console: tty 4 wm: kwin_x11 dm: SDDM
           Distro: Ubuntu 19.10 (Eoan Ermine)
CPU: Topology: 6-Core model: Intel Core i7-8700 bits: 64 type: MT MCP arch: Kaby Lake rev: A L1 cache: 384 KiB
           L2 cache: 12.0 MiB L3 cache: 12.0 MiB
           flags: avx avx2 lm nx pae sse sse2 sse3 sse4_1 sse4_2 ssse3 vmx bogomips: 76799
           Speed: 1539 MHz min/max: 800/4600 MHz Core speeds (MHz): 1: 1539 2: 883 3: 2930 4: 1633 5: 1548 6: 1075 7: 1081
           8: 2129 9: 2623 10: 2214 11: 1257 12: 1777
Graphics: Device-1: Intel UHD Graphics 630 vendor: Hewlett-Packard driver: i915 v: kernel bus ID: 00:02.0 chip ID: 8086:3e92
           Display: server: X.Org 1.20.5 driver: modesetting unloaded: fbdev,vesa compositor: kwin_x11
           resolution: 2560x1440~60Hz, 1920x1080~60Hz
           OpenGL: renderer: Mesa DRI Intel UHD Graphics 630 (Coffeelake 3x8 GT2) v: 4.5 Mesa 19.2.8 compat-v: 3.0
           direct render: Yes
root@gromk:~# >

apr 10 23:03:45 gromk kernel: i915 0000:00:02.0: GPU HANG: ecode 9:0:0x00000000, hang on rcs0
apr 10 23:03:45 gromk kernel: i915 0000:00:02.0: Resetting rcs0 for hang on rcs0

[Nikke (nmellegard)]

Tried upgrading to kernel 5.3.0-47 which is in the Ubuntu proposed repository. Still get the freezes, but now it seems that the mouse pointer doesn't stop--only the interface stops responding for 2-3seconds and then it all jumps back in action again.

Also there's a new line in the journal about an

--> journalctl -b | grep i915

apr 11 11:06:13 Dinky kernel: mei_hdcp 0000:00:16.0-b638ab7e-94e2-4ea2-a552-d1c54b627f04: bound 0000:00:02.0 (ops i915_hdcp_component_ops [i915])
apr 11 11:10:26 Dinky kernel: i915 0000:00:02.0: Resetting rcs0 for hang on rcs0

--> inxi -SCGxxxz
System: Host: Dinky Kernel: 5.3.0-47-generic x86_64 bits: 64 compiler: gcc v: 9.2.1 Desktop: Gnome 3.34.3 wm: gnome-shell
           dm: GDM3 3.34.1 Distro: Ubuntu 19.10 (Eoan Ermine)
CPU: Topology: Dual Core model: Intel Core i7-7500U bits: 64 type: MT MCP arch: Kaby Lake rev: 9 L1 cache: 128 KiB
           L2 cache: 4096 KiB L3 cache: 4096 KiB
           flags: avx avx2 lm nx pae sse sse2 sse3 sse4_1 sse4_2 ssse3 vmx bogomips: 23199
           Speed: 600 MHz min/max: 400/3500 MHz Core speeds (MHz): 1: 642 2: 695 3: 646 4: 639
Graphics: Device-1: Intel HD Graphics 620 vendor: Dell driver: i915 v: kernel bus ID: 00:02.0 chip ID: 8086:5916
           Display: server: X.Org 1.20.5 driver: i915 compositor: gnome-shell resolution: 1920x1080~60Hz
           OpenGL: renderer: Mesa DRI Intel HD Graphics 620 (Kaby Lake GT2) v: 4.5 Mesa 19.2.8 compat-v: 3.0
           direct render: Yes

[pcworld (pcworld)]

This is probably the same as bug #1871867, can you confirm?

I've had this bug ever since upgrading to Ubuntu 18.10 (i.e. the Linux 5.3 kernel series). However before the 5.3.0-46-generic update (when I was on 5.3.0-45-generic and earlier versions), I had this issue (freezes of around ~10s) only a few times per day. After the 5.3.0-46-generic update, this bug happened several times per hour, sometimes repeating in the timeframe of minutes.
Though I only rarely get the "GPU HANG: ecode 9:0:0x00000000, hang on rcs0" message prior to the "Resetting rcs0 for hang on rcs0" message, for what's it worth.

To put this into numbers, from April 4 to April 9 (daily usage), I only got 6 freezes, whereas since the update to 5.3.0-46-generic (which I performed on April 10), I've got 122 hangs already. This makes using the whole system really annoying.

There are some interesting discussions about this bug on https://bbs.archlinux.org/viewtopic.php?id=250765, though note that there seem to be multiple bugs and it got worse with later kernel versions, including unrecoverable freezes on kernel 5.4 (which were reported in Ubuntu as bug #1861395).

I've reverted to Linux 4.19 LTS mainline kernel from https://kernel.ubuntu.com/~kernel-ppa/mainline/ for now, and haven't had any more freezes yet.

[sanette (sanette-linux)]

I agree, it seems exactly the same as bug #1872009

I have uploaded a test program (needs SDL1.2) that kind of immediately triggers the hang
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1872009/+attachment/5352362/+files/test.exe

[sanette (sanette-linux)]

I'm sorry I realized you need more dependencies to run my test program, it's going to be complicated.

But I think any program that intensively sends data to the GPU should trigger the problem as well

[sanette (sanette-linux)]

OK here is a new version, I think now you just need to do

sudp apt install libsdl-ttf2.0-0
./test.exe

it should work

[sanette (sanette-linux)]

interestingly, my demo runs also 25% faster on 5.3.0-45 than on 5.3.0-46...

[sanette (sanette-linux)]

and 50% faster on 4.4.0 :(

[ekinox09 (ekinox09)]

Hello all, i think i have the same issue. Encountered on Ubuntu 18.04 and kernels 5.3.0, 5.4.31, 5.5.16 and 5.5.6 (ppa mainline build). I'm in a virtualized VM with Proxmox
I can EASILY reproduce it within a minute (start Kodi for example).

inxi -SCGxxxz:

System: Host: hystUbuTest Kernel: 5.5.6-050506-generic x86_64 bits: 64 gcc: 9.2.1
           Desktop: Gnome 3.28.4 (Gtk 3.22.30-1ubuntu4) info: gnome-shell dm: gdm3 Distro: Ubuntu 18.04.4 LTS
CPU: Quad core Common KVM (-MCP-) arch: Netburst Prescott rev.1 cache: 16384 KB
           flags: (lm nx sse sse2 sse3) bmips: 21697
           clock speeds: max: 2712 MHz 1: 2712 MHz 2: 2712 MHz 3: 2712 MHz 4: 2712 MHz
Graphics: Card: Intel HD Graphics 620 bus-ID: 00:10.0 chip-ID: 8086:5916
           Display Server: x11 (X.Org 1.20.5 ) drivers: vmware (unloaded: modesetting,fbdev,vesa)
           Resolution: 1920x1080@60.00hz
           OpenGL: renderer: Mesa DRI Intel HD Graphics 620 (Kaby Lake GT2)
           version: 4.5 Mesa 19.2.8 (compat-v: 3.0) Direct Render: Yes

journalctl -b | grep i915:

avr 12 13:08:54 hystUbuTest kernel: i915 0000:00:10.0: vgaarb: deactivate vga console
avr 12 13:08:54 hystUbuTest kernel: i915 0000:00:10.0: Invalid PCI ROM header signature: expecting 0xaa55, got 0x0000
avr 12 13:08:54 hystUbuTest kernel: i915 0000:00:10.0: vgaarb: changed VGA decodes: olddecodes=io+mem,decodes=io+mem:owns=io+mem
avr 12 13:08:54 hystUbuTest kernel: [drm] Finished loading DMC firmware i915/kbl_dmc_ver1_04.bin (v1.4)
avr 12 13:08:55 hystUbuTest kernel: [drm] Initialized i915 1.6.0 20191101 for 0000:00:10.0 on minor 0
avr 12 13:08:55 hystUbuTest kernel: snd_hda_intel 0000:00:11.0: bound 0000:00:10.0 (ops i915_audio_component_bind_ops [i915])
avr 12 13:08:55 hystUbuTest kernel: fbcon: i915drmfb (fb0) is primary device
avr 12 13:08:55 hystUbuTest kernel: i915 0000:00:10.0: fb0: i915drmfb frame buffer device
avr 12 13:08:58 hystUbuTest /usr/lib/gdm3/gdm-x-session[973]: 915G, E7221 (i915), 915GM, 945G, 945GM, 945GME, Pineview GM,
avr 12 13:08:58 hystUbuTest /usr/lib/gdm3/gdm-x-session[973]: (II) intel(0): Using Kernel Mode Setting driver: i915, version 1.6.0 20191101
avr 12 13:14:33 hystUbuTest kernel: i915 0000:00:10.0: GPU HANG: ecode 9:2:0x00000000, stopped heartbeat on bcs0
avr 12 13:14:33 hystUbuTest kernel: drm/i915 developers can then reassign to the right component if it's not a kernel issue.
avr 12 13:14:33 hystUbuTest kernel: i915 0000:00:10.0: Resetting bcs0 for stopped heartbeat on bcs0

Afterwards, i have such issues that will block the computer and need a reboot to solve:

Apr 12 13:05:45 hystUbuTest kernel: [ 1025.728158] i915 0000:00:10.0: Resetting bcs0 for preemption time out
Apr 12 13:05:50 hystUbuTest kernel: [ 1031.617233] i915 0000:00:10.0: Resetting bcs0 for preemption time out
Apr 12 13:05:57 hystUbuTest kernel: [ 1037.761206] i915 0000:00:10.0: Resetting bcs0 for preemption time out
Apr 12 13:06:03 hystUbuTest kernel: [ 1043.649152] i915 0000:00:10.0: Resetting bcs0 for preemption time out
Apr 12 13:06:08 hystUbuTest kernel: [ 1049.537229] i915 0000:00:10.0: Resetting bcs0 for preemption time out

Let me know how to contribute with testing some patchs ...

[ekinox09 (ekinox09)]

I've seen that going back to Kernel 4.19 has improved the situation.
This is not the case for me (4.19.0-041900-generic from kernel.ubuntu.com/~kernel-ppa/mainline/):

journalctl -b | grep i915:

avr 12 14:07:03 hystUbuTest kernel: i915 0000:00:10.0: Invalid PCI ROM header signature: expecting 0xaa55, got 0x0000
avr 12 14:07:03 hystUbuTest kernel: i915 0000:00:10.0: vgaarb: changed VGA decodes: olddecodes=io+mem,decodes=io+mem:owns=io+mem
avr 12 14:07:03 hystUbuTest kernel: [drm] Finished loading DMC firmware i915/kbl_dmc_ver1_04.bin (v1.4)
avr 12 14:07:05 hystUbuTest kernel: [drm] Initialized i915 1.6.0 20180719 for 0000:00:10.0 on minor 0
avr 12 14:07:05 hystUbuTest kernel: snd_hda_intel 0000:00:11.0: bound 0000:00:10.0 (ops i915_audio_component_bind_ops [i915])
avr 12 14:07:05 hystUbuTest kernel: i915 0000:00:10.0: fb0: inteldrmfb frame buffer device
avr 12 14:07:07 hystUbuTest /usr/lib/gdm3/gdm-x-session[935]: 915G, E7221 (i915), 915GM, 945G, 945GM, 945GME, Pineview GM,
avr 12 14:07:07 hystUbuTest /usr/lib/gdm3/gdm-x-session[935]: (II) intel(0): Using Kernel Mode Setting driver: i915, version 1.6.0 20180719
avr 12 14:09:38 hystUbuTest kernel: i915 0000:00:10.0: Resetting rcs0 for hang on rcs0
avr 12 14:10:16 hystUbuTest kernel: i915 0000:00:10.0: Resetting rcs0 for hang on rcs0
avr 12 14:10:24 hystUbuTest kernel: i915 0000:00:10.0: Resetting rcs0 for hang on rcs0

[Nikke (nmellegard)]

Seems that a fix is backported to kernel 5.4.0-16.19 for Focal:
  * system hang: i915 Resetting rcs0 for hang on rcs0 (LP: #1861395)
    - drm/i915/execlists: Always force a context reload when rewinding RING_TAIL

https://launchpad.net/ubuntu/+source/linux/5.4.0-16.19

Has anyone tried that kernel?

[ekinox09 (ekinox09)]

Hello, for your information, i've tested:
- cod/tip/drm-tip/2020-04-11 mainline build (from https://kernel.ubuntu.com/~kernel-ppa/mainline/drm-tip/2020-04-11/). No improvement:
       [drm] GPU HANG
       [drm] *ERROR* vcs0 reset request timed out
- cod/tip/drm-tip/2020-04-13 mainline build (from https://kernel.ubuntu.com/~kernel-ppa/mainline/drm-intel-nightly/2020-04-13/). No improvement:
       [drm] GPU HANG
       [drm] Resetting rcs0 for CS error
- cod/tip/drm-intel-next/2020-03-14 mainline build (from https://kernel.ubuntu.com/~kernel-ppa/mainline/drm-intel-next/2020-03-14/). Better !
       No more CPU HANG
       [drm] *ERROR* Fault errors on pipe A: 0x00000080
       Note: Lot of bugs linked to this not stable kernel BUT no more CPU HANG for the moment...
Hope it helps.

[Ohad Lutzky (lutzky)]

FWIW, I'm experiencing very similar symptoms on the same kernel (5.3.0-46-generic), but on different hardware: This is a Lenovo ThinkPad E490. One thing that makes it *much* more frequent is running jackd with realtime priority. Running with ardour and jack, during the freezes audio proceeds normally (even MIDI events and synths work fine), but keyboard, mouse and monitor freeze. This occurs about once a minute!
I've also experienced longer crashes (gave up and rebooted after a minute) when performing skype calls or just watching youtube, but those are far less frequent.

[Geir Isene (qc-e-9h)]

On my Dell XPS15, the frequency of the freezes depends on what I am running. While browsing, doing e-mail, terminal work and the like, it freezes maybe once every 10 minutes. But while playing Wesnoth, it freezes at least once per minute(!). It doesn't matter if I run on the i915 or the Nvidia gpu. The freezes are funny, because everything BUT the mouse freezes. I can move the mouse pointer around, but everything else is stuck for 5-8 seconds or so. I run i3-wm straight on X with no desktop environment on Ubuntu 19.10 with the 5.3.0-46-generic kernel.

[Geir Isene (qc-e-9h)]

...and no freezes before the update to the 5.3.0-46-generic kernel.

[Grzegorz Dabrowski (grzegorz-dabrowski)]

I can reproduce this bug by starting a call via zoom.us. The bug exists on 5.3.0-46-generic, previous kernel 5.3.0-45-generic works fine.

[Egbert van der Wal (eggie)]

Update: I just tried the recently released linux-image-5.3.0-48-generic that was made available through eoan-proposed, but I still get the `Resetting rcs0 for hang on rcs0` error, so it's back to mainline 5.4.28-050428-generic for me.

[Maurizio (xthephreakx)]

Experiencing the same issues on linux mint 19.3.

System:
  Host: maurizio-LM Kernel: 5.3.0-46-generic x86_64 bits: 64 compiler: gcc
  v: 7.5.0 Desktop: Cinnamon 4.4.8 info: plank wm: muffin 4.4.2
  dm: LightDM 1.26.0 Distro: Linux Mint 19.3 Tricia
  base: Ubuntu 18.04 bionic
CPU:
  Topology: Quad Core model: Intel Core i5-6500 bits: 64 type: MCP
  arch: Skylake-S rev: 3 L2 cache: 6144 KiB
  flags: lm nx pae sse sse2 sse3 sse4_1 sse4_2 ssse3 vmx bogomips: 25599
  Speed: 800 MHz min/max: 800/3600 MHz Core speeds (MHz): 1: 800 2: 800
  3: 800 4: 800
Graphics:
  Device-1: Intel HD Graphics 530 vendor: Hewlett-Packard driver: i915
  v: kernel bus ID: 00:02.0 chip ID: 8086:1912
  Display: x11 server: X.Org 1.20.5 driver: modesetting unloaded: fbdev,vesa
  resolution: 1920x1200~60Hz, 1920x1200~60Hz
  OpenGL: renderer: Mesa DRI Intel HD Graphics 530 (Skylake GT2)
  v: 4.5 Mesa 19.2.8 compat-v: 3.0 direct render: Yes

Running to dual screen and also noticed that the size settings(icon size) on my second panel get reset to default... nut sure when it happens tough

[Rimas Kudelis (rq)]

I have this issue as well. Rebooted my Thinkpad X1 Carbon 6th gen today, and now I'm getting these UI lockups for a few seconds every few minutes, even though the mouse cursor still moves and Spotify keeps playing what it is playing.

In my case, this seems related to PhpStorm 2020.1. It's enough to just scroll a file back and forth for like 30 seconds in PhpStorm for the bug to manifest. Meanwhile scrolling in Firefox or Gnome Terminal doesn't seem to trigger this.

PhpStorm writes a thread dump each time it happens. I'm attaching one here, in case it might be useful.

[trickv (trickv)]

I think I've been having the same issue as well, but I'm also not sure that this is 100% "new". The symptoms I'm having are consistent (mouse cursor continues to move, video freezes for ~2 seconds at a time leading me to think that the machine has crashed, GPU HANG and Resetting rcs0 log events, but machine is running underneath which is obvious because audio keeps working.) I can kill the offending video process (Zoom) and the issue subsides.

I can reliably reproduce the problem on 5.3.0-46-generic (Ubuntu 19.10 latest) and downgrading to 5.3.0-45-generic seemed at first to fix the problem. But I think that it's just less frequent on the previous kernels. I use my machine 5 days a week for 2-4 hours a day on Zoom. With the -46 kernel, simply launching a Zoom meeting kicks off the problem. However today after downgrading to the -45 kernel, while the problem doesn't immediately happen, it has "glitched" once where the machine had the same behavior for ~2 seconds and went back to working as expected.

Looking back over the last 3 months, these events - assuming that the "Resetting rcs0 for hang on rcs0" log event indicates the 2 second video glitch, it's been rather common. It's just *much* more common on the -46 kernel. Attached is a grep of journalctl since 2020-01-01 for the relevant log messages:
- kernel boot message to show the version
- i915 GPU HANG
- i915 Resetting rcs0 for hang on rcs0

Command: journalctl -o short-iso --since=2020-01-01 | egrep "Command line: |Resetting rcs0 for hang on rcs0|GPU HANG"

[Eduard Biceri-Matei (cubusbacau)]

Hi,

I'm having the same issue, i can consistently trigger it using PyCharm. Initially i wanted to report it there, but the issue is already reported and sent me here (https://youtrack.jetbrains.com/issue/JBR-2269)

Dmesg logs:
i915 0000:00:02.0: Resetting rcs0 for hang on rcs0

PyCharm logs:
WARN - .diagnostic.PerformanceWatcher - UI freezed for 8041ms, details saved to /home/***

Trigger:
- scroll a larger file in the IDE

Thread dumps shows thread in TIMED_WAITING ("waiting on condition")

Behavior:
- Almost everything (clock, keyboard, UI) freezes for up to 8 seconds;
- mouse still moves, audio still works

System:
PopOS! 19.10 with 5.3.0-7648-generic on Lenovo ThinkPad E590

[Maurizio (xthephreakx)]

i downgraded for the time beeing to 4.15.0.96 were the issue doesnt exist didnt want to but i have to work on this machine...

[Robert Hardy (rhardy)]

I just started having this issue on Ubuntu 19.10 w/ 5.3.0-46-generic. What a mess...
My system is an Asrock Z370M-ITXax with an Intel i3-8100.

sudo cat /sys/class/drm/card0/error
GPU HANG: ecode 9:0:0x00000000, hang on rcs0
Kernel: 5.3.0-46-generic x86_64

As if all my Intel NIC problems weren't bad enough. What happened to Intel....

[Robert Hardy (rhardy)]

For what it is worth I just downgraded my kernel to 5.3.0-45.37 and I haven't had a hang since:
https://launchpad.net/ubuntu/+source/linux/5.3.0-45.37

Not clear at all from the change log as to what brought this on in 5.3.0-46.38:
https://launchpad.net/ubuntu/+source/linux/5.3.0-46.38

[Robert Hardy (rhardy)]

I wanted the security patches in 5.3.0-46.38 so I tried updating to the kernel from proposed i.e. 5.3.0-47.39. Sadly hangs returned within a couple minutes of trying to use the kernel.

# cat /sys/class/drm/card0/error | head -2
GPU HANG: ecode 9:0:0x00000000, hang on rcs0
Kernel: 5.3.0-47-generic x86_64

[htrex (hantarex)]

Same issue here, consistently triggered by PhpStorm (it's an IDE similar to PyCharm).

journalctl -o short-iso --since=2020-01-01 | egrep "Command line: |Resetting rcs0 for hang on rcs0|GPU HANG"
2020-04-18T12:43:21+0200 OrionXPS kernel: Command line: BOOT_IMAGE=/boot/vmlinuz-5.3.0-46-generic root=UUID=34223b98-7d1b-48e3-a62d-6c7373fa8c21 ro quiet splash acpi_rev_override=5 i915.modeset=1 nouveau.modeset=0 vt.handoff=1
2020-04-18T12:48:29+0200 OrionXPS kernel: i915 0000:00:02.0: GPU HANG: ecode 9:0:0x00000000, hang on rcs0
2020-04-18T12:48:29+0200 OrionXPS kernel: i915 0000:00:02.0: Resetting rcs0 for hang on rcs0

[Rimas Kudelis (rq)]

@htrex:
I'm using PHPStorm as well, and the issue seems gone after I rebooted to previous kernel (5.3.0-45-generic). I suggest you do the same until this is resolved.

[Jon V (mumonkan)]

i am also seeing this same issue on Mint 19.2 Tina on a dell XPS13.
only after a recent (today) apt update.

i can ssh in and tail the kern.log and watch it happen, same kind of sporadic usage, like switching between slack-desktop and chrome etc.

# Apr 18 22:09:59 fig kernel: [ 1138.173506] i915 0000:00:02.0: GPU HANG: ecode 9:1:0x00000000, hang on rcs0
Apr 18 22:09:59 fig kernel: [ 1138.173517] i915 0000:00:02.0: Resetting rcs0 for hang on rcs0

uname -a
Linux fig 5.4.3-050403-generic

# cat /sys/class/drm/card0/error |head
GPU HANG: ecode 9:1:0x00000000, hang on rcs0
Kernel: 5.4.3-050403-generic x86_64
Driver: 20190822
Time: 1587272999 s 864492 us
Boottime: 1137 s 921564 us
Uptime: 1136 s 132425 us

[Sledge HaMMeR (kh-ubuntuone)]

Same here for a while now, really annoying

$ inxi -SCGxxxz
System: Host: Slenvy Kernel: 5.3.0-46-generic x86_64 bits: 64 gcc: 7.5.0
           Desktop: Gnome 3.28.4 (Gtk 3.22.30-1ubuntu4) info: gnome-shell dm: gdm3 Distro: Ubuntu 18.04.4 LTS
CPU: Quad core Intel Core i7-8550U (-MT-MCP-) arch: Kaby Lake rev.10 cache: 8192 KB
           flags: (lm nx sse sse2 sse3 sse4_1 sse4_2 ssse3 vmx) bmips: 15999
           clock speeds: min/max: 400/4000 MHz 1: 796 MHz 2: 799 MHz 3: 799 MHz 4: 799 MHz 5: 796 MHz 6: 799 MHz
           7: 798 MHz 8: 795 MHz
Graphics: Card-1: Intel UHD Graphics 620 bus-ID: 00:02.0 chip-ID: 8086:5917
           Card-2: NVIDIA GP108M [GeForce MX150] bus-ID: 01:00.0 chip-ID: 10de:1d10
           Display Server: x11 (X.Org 1.20.5 ) drivers: modesetting,nvidia (unloaded: fbdev,vesa,nouveau)
           Resolution: 1920x1080@60.00hz
           OpenGL: renderer: Mesa DRI Intel UHD Graphics 620 (Kabylake GT2)
           version: 4.5 Mesa 19.2.8 (compat-v: 3.0) Direct Render: Yes

[Viacheslav Semykrasov (7krasov)]

The same for Phpstorm

slava@slava-Latitude-5500:~$ inxi -SCGxxxz
System: Host: slava-Latitude-5500 Kernel: 5.3.0-46-generic x86_64 bits: 64 compiler: gcc v: 9.2.1 Desktop: Gnome 3.34.3
           wm: gnome-shell dm: GDM3 3.34.1 Distro: Ubuntu 19.10 (Eoan Ermine)
CPU: Topology: Quad Core model: Intel Core i5-8365U bits: 64 type: MT MCP arch: Kaby Lake rev: C L2 cache: 6144 KiB
           flags: avx avx2 lm nx pae sse sse2 sse3 sse4_1 sse4_2 ssse3 vmx bogomips: 30399
           Speed: 800 MHz min/max: 400/4100 MHz Core speeds (MHz): 1: 800 2: 801 3: 800 4: 800 5: 800 6: 800 7: 800 8: 800
Graphics: Device-1: Intel UHD Graphics 620 vendor: Dell driver: i915 v: kernel bus ID: 00:02.0 chip ID: 8086:3ea0
           Display: x11 server: X.Org 1.20.5 driver: i915 compositor: gnome-shell resolution: 1920x1080~60Hz
           OpenGL: renderer: Mesa DRI Intel UHD Graphics (Whiskey Lake 3x8 GT2) v: 4.5 Mesa 19.2.8 compat-v: 3.0
           direct render: Yes

[Eugene86 (eugene86)]

Affected me with 5.3.0-46-lowlatency (Ubuntu 18.04.4 LTS)
VGA compatible controller: Intel Corporation UHD Graphics 620
Hardware: Dell Latitude 7490

[Benjamin Schmid (benbuntu)]

Same as Grzegorz Dabrowski: I can reproduce this bug by starting a call via zoom.us. It occurs very quickly and then sometimes calms down to an acceptable level.

Attached /sys/class/drm/card0/error

[Chris Carpenter (chriscarpenter12)]

Saw the same hanging issues constantly with IntelliJ Idea using 5.3.0-46. Rolling back to 5.3.0-45 resolves the freezing for now.

Came here from: https://youtrack.jetbrains.com/issue/JBR-2269

[sanette (sanette-linux)]

So, this problem is serious.
I am surprised that no one else has mentioned the coil noise issue.
I can repeatedly and consistently hear it on 5.3.0-46, while it's absent in 5.3.0-45.
I think this is the sign of something quite bad happening. (maybe too much data transiting on the bus between cpu and gpu?)

[Andrea Righi (arighi)]

This looks similar to LP: #1861395

I've prepared a test kernel (5.3.0-48.41+lp1861395v1), backporting the following fixes that seems to have fixed the problem in 5.4:

 b1339ecac661 drm/i915/execlists: Always force a context reload when rewinding RING_TAIL
 f26a9e959a7b drm/i915/gt: Detect if we miss WaIdleLiteRestore
 22b7a426bbe1 drm/i915/execlists: Preempt-to-busy

Test kernel is available here:
https://kernel.ubuntu.com/~arighi/LP-1861395/5.3/

It'd be great if you could give it a try, thanks.

[Jozef Lammers (jlammrs)]

@Andrea, can I use this kernel with UEFI/SecureBoot?

[Sultan Alsawaf (kerneltoast)]

@jlammrs You can't use that kernel with secure boot because it's unsigned.

[sanette (sanette-linux)]

@Andrea, for me this seems to solve the problem. Thanks!

(except that my demo is still slightly slower than with 4.4.0 (13fps instead of 15fps) but this is probably unrelated)

[Egbert van der Wal (eggie)]

@arighi - it helps, but introduces different problems on my Dell XPS 15 7590. I get much less frequent lockups but they still do occur. It seems to introduce similar issues with Nouveau.

My dmesg gives me lots of messages like this:

[ 67.714761] nouveau 0000:01:00.0: fifo: SCHED_ERROR 08 []
[ 67.855754] nouveau 0000:01:00.0: DRM: failed to idle channel 0 [DRM]

[ 5396.978950] nouveau 0000:01:00.0: tmr: stalled at ffffffffffffffff

[ 5412.086427] nouveau 0000:01:00.0: timeout
[ 5412.086464] WARNING: CPU: 11 PID: 11576 at drivers/gpu/drm/nouveau/nvkm/subdev/mmu/vmmtu102.c:44 tu102_vmm_flush+0x130/0x140 [nouveau]

and lots of more information. I'll add the full output of dmesg as an attachment.

Not sure if the new problems are introduced by your patch but I didn't experience them before on 5.3 and are also not occuring on the 5.4.28-050428-generic kernel from `mainline`. So back to 5.4.28 for me, again.

[Sultan Alsawaf (kerneltoast)]

@eggie You shouldn't be using nouveau on your hardware; it has almost no support for the Turing architecture from Nvidia. You should blacklist nouveau and then reboot. If you want to make use of your Nvidia card, you'll have to use the binary Nvidia driver sadly.

[Egbert van der Wal (eggie)]

@kerneltoast I don't want to make use of my nvidia card. I'm attempting to use i915 exclusively to conserve battery and prime doesn't work properly at all anyway. I'm saving my nvidia GPU for my dualboot Windows partition.

[Egbert van der Wal (eggie)]

Oh, wait, are you implying that I should actually blacklist nouveau in order to disable it and use Intel exclusively? I'll give that a try.

[houdini68 (houdini68)]

I have experienced the same problem!
Excerpt from the log of this morning:
Apr 23 07:13:18 rudy-ThinkPad-T590 kernel: [ 5282.177837] i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
Apr 23 07:13:26 rudy-ThinkPad-T590 kernel: [ 5290.177889] i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
Apr 23 07:13:44 rudy-ThinkPad-T590 kernel: [ 5308.225930] i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
Apr 23 07:20:50 rudy-ThinkPad-T590 kernel: [ 5734.242384] i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
Apr 23 07:33:39 rudy-ThinkPad-T590 kernel: [ 6503.203480] i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
Apr 23 07:33:47 rudy-ThinkPad-T590 kernel: [ 6511.203533] i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
Apr 23 07:37:38 rudy-ThinkPad-T590 kernel: [ 6742.247806] i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
Apr 23 07:37:46 rudy-ThinkPad-T590 kernel: [ 6750.243886] i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
Apr 23 07:39:55 rudy-ThinkPad-T590 kernel: [ 6879.176068] i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
Apr 23 07:40:31 rudy-ThinkPad-T590 kernel: [ 6915.240116] i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
As soons as SoapUI was started the application freezes continually...

Since I downgraded to 5.3.0-45-generic, the hangings disappeared.
As it was said previously until the problem is fixed let's continue to work with the version 5.3.0-45.

[houdini68 (houdini68)]

I must say it is a serious problem... Intellij, SoapUI, ... hanging continually for a few seconds... For a professional environment, it is impossible to work...

[houdini68 (houdini68)]

Same for me. (I have switched to 45).

$ inxi
CPU~Quad core Intel Core i7-8665U (-MT-MCP-) speed/max~800/4800 MHz Kernel~5.3.0-45-generic x86_64 Up~3:08 Mem~5845.4/39821.6MB HDD~5025.0GB(53.9% used) Procs~389 Client~Shell inxi~2.3.56
rudy@rudy-ThinkPad-T590:~/bin/MLOZ/INTEG$ inxi -SCGxxxz
System: Host: rudy-ThinkPad-T590 Kernel: 5.3.0-45-generic x86_64 bits: 64 gcc: 7.5.0
           Desktop: Gnome 3.28.4 (Gtk 3.22.30-1ubuntu4) info: gnome-shell dm: gdm3 Distro: Ubuntu 18.04.4 LTS
CPU: Quad core Intel Core i7-8665U (-MT-MCP-) arch: Kaby Lake rev.12 cache: 8192 KB
           flags: (lm nx sse sse2 sse3 sse4_1 sse4_2 ssse3 vmx) bmips: 16799
           clock speeds: min/max: 400/4800 MHz 1: 800 MHz 2: 800 MHz 3: 800 MHz 4: 800 MHz 5: 800 MHz 6: 800 MHz
           7: 800 MHz 8: 800 MHz
Graphics: Card: Intel Device 3ea0 bus-ID: 00:02.0 chip-ID: 8086:3ea0
           Display Server: x11 (X.Org 1.20.5 ) drivers: fbdev (unloaded: modesetting,vesa)
           Resolution: 1440x810@60.00hz, 1920x1080@60.00hz
           OpenGL: renderer: Mesa DRI Intel UHD Graphics (Whiskey Lake 3x8 GT2)
           version: 4.5 Mesa 19.2.8 (compat-v: 3.0) Direct Render: Yes
rudy@rudy-ThinkPad-T590:~/bin/MLOZ/INTEG$ inxi -SCGxxxz
System: Host: rudy-ThinkPad-T590 Kernel: 5.3.0-45-generic x86_64 bits: 64 gcc: 7.5.0
           Desktop: Gnome 3.28.4 (Gtk 3.22.30-1ubuntu4) info: gnome-shell dm: gdm3 Distro: Ubuntu 18.04.4 LTS
CPU: Quad core Intel Core i7-8665U (-MT-MCP-) arch: Kaby Lake rev.12 cache: 8192 KB
           flags: (lm nx sse sse2 sse3 sse4_1 sse4_2 ssse3 vmx) bmips: 16799
           clock speeds: min/max: 400/4800 MHz 1: 800 MHz 2: 800 MHz 3: 800 MHz 4: 800 MHz 5: 800 MHz 6: 800 MHz
           7: 799 MHz 8: 800 MHz
Graphics: Card: Intel Device 3ea0 bus-ID: 00:02.0 chip-ID: 8086:3ea0
           Display Server: x11 (X.Org 1.20.5 ) drivers: fbdev (unloaded: modesetting,vesa)
           Resolution: 1440x810@60.00hz, 1920x1080@60.00hz
           OpenGL: renderer: Mesa DRI Intel UHD Graphics (Whiskey Lake 3x8 GT2)
           version: 4.5 Mesa 19.2.8 (compat-v: 3.0) Direct Render: Yes

[Andrea Righi (arighi)]

@sanette-linux thanks for testing! Can you provide some details about the demo that seems to be slower? Do you notice the same slowness also with a simple glxgears for example?

[Hans van den Bogert (hbogert)]

#57 this works.

Hard to install on 18.04 though, you need the headers package (linux-headers-5.3.0-48_5.3.0-48.41_all.deb) from eoan

[pcworld (pcworld)]

Right, comment #57 is missing the non-generic headers package (which the -generic package depends on). The kernel is installable but makes dpkg sad.

[Sultan Alsawaf (kerneltoast)]

@eggie Yes, you should actually blacklist the nouveau module. Or instead, if you don't want your Nvidia card to drain power while you're using exclusively Intel graphics, run the following commands:

sudo -i
cat << EOF | sudo tee /etc/udev/rules.d/00-pcidevices.rules
ACTION=="add", KERNEL=="0000:01:00.0", SUBSYSTEM=="pci", RUN+="/bin/sh -c 'echo 1 > /sys/bus/pci/devices/0000:01:00.0/remove'"
EOF

This adds a permanent rule to eject the Nvidia card from your system so that it won't use any power. Reboot for it to take effect. On some laptops this doesn't stop the Nvidia card from draining power, but on Dell laptops it does. This is what I use on my Precision 5540, which is essentially the same laptop as yours.

[Andrea Righi (arighi)]

@hbogert I've uploaded all the required deb's, it should be easier to install the test kernel now:

https://kernel.ubuntu.com/~arighi/LP-1861395/5.3/

Thanks!

[Benjamin Gemmill (benjamin-gemmill)]

@arighi I'm testing your kernel on an affected dell xps 13 9300, and it looks like there are momentary hangs but no crashing any more with your patches.

In dmesg just after a hang, I see:
[ 338.955106] i915 0000:00:02.0: GPU HANG: ecode 11:1:0xfffffffe, in Xorg [1315], hang on rcs0
[ 338.956204] i915 0000:00:02.0: Resetting rcs0 for hang on rcs0

...and then things resume.

In 5.3.0-46 without your patches, this would have crashed my session.

If it helps with more debugging, I've attached the log from
/sys/class/drm/card0/error

[Robert Hardy (rhardy)]

The test kernel 5.3.0-48.41+lp1861395v1 seems to be a major improvement!
On 5.3.0-45.37 pauses still happened rarely i.e. a single 2 second pause every 15 hrs and I hadn't noticed them until I was watching the logs for them.
5.3.0-46 and later were pausing several times a minute and the system was unusable.
This test kernel 5.3.0-48.41+lp1861395v1 has yet to pause at all. I will keep testing it but I suspect we have a winner here.

[Egbert van der Wal (eggie)]

I've been running the 5.3.0-48-generic_5.3.0-48.41+lp1861395v1 for 2 days now. It definitely did get better but this afternoon I did get another "Resetting rcs0 for hang on rcs0" so it doesn't seem to be solved completely.

@kerneltoast: thanks for the suggestion, I did that. No more nouveau errors in dmesg and battery life has improved significantly. Awesome.

[tomatomat (mkatzer)]

The test kernel of #57 works well for me. Before I could trigger the hangs immediately and frequently (every few seconds) when running the linuxcnc simulator (heavily using mesa GL I guess) on an i9-9900K desktop and kernel 5.3.0-46, that is gone with the test kernel. I never saw the problem on my Dell XPS13 9343 (i7-5500U, running same 18.04 OS and 5.3.0-46 kernel), by the way.

[sanette (sanette-linux)]

@arighi
It's graphical demo using openGL via SDL1.2 where 1000 small discs move around and all collisions are detected, and print FPS. It uses both CPU and GPU quite heavily. (On my XPS 13 9350, intel_gpu_top reports about 55% GPU usage, and top reports 46% CPU). I was developing this for my research and this is why the hanging bug stroke me strongly.
I made quick tests with the kernels that are currently on my laptop:

5.3.0-48-generic #41+lp1861395v1: 12 FPS
5.3.0-46 9 FPS + Noise (inbetween hangs, of course)
5.3.0-45 12 FPS
5.3.0-42 12 FPS
4.4.0 15 FPS

(aside remark: on my antique desktop PC with Athlon 64 5600 CPU and Radeon HD6450 GPU, the performance is better: 16FPS, which is quite humiliating ;) )

[sanette (sanette-linux)]

@arighi
I cannot test with glxgears because it syncs with screen refresh, so always 60FPS.

[pcworld (pcworld)]

@sanette: You can run glxgears with `vblank_mode=0 glxgears` to turn off vsync.

[Rob N (robn)]

Another +1 for test kernel from #57 fixing it (in my case, Zoom killing the system for a couple of minutes after it starts as described in https://forums.linuxmint.com/viewtopic.php?t=316392). Lenovo X1 G7, Intel GPU, 18.04.

Thanks!

[Benjamin Gemmill (benjamin-gemmill)]

@arighi
Unfortunately hard crashes still occur with the test kernel you posed in #57.

dmesg:

[ 1955.974958] i915 0000:00:02.0: GPU HANG: ecode 11:1:0xfffffffe, in Xorg [1073], hang on rcs0
[ 1955.976105] i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
[ 1991.965817] i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
[ 1999.965828] i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
[ 2007.965803] i915 0000:00:02.0: Resetting rcs0 for hang on rcs0

i915 error log attached.

[Mathieu Lavigne (mlavigne)]

My comment may be irrelevant but as an IntelliJ user on Ubuntu 18.04.4 LTS, I experience the same problem.

I've found a workaround by adding the following line in IntelliJ VM Options :

-Dsun.java2d.opengl=true

Hope this may help someone !

[ekinox09 (ekinox09)]

Hi, using the test kernel proposed on message #57, i have a problem. May someone help me ?
Syslog: "modprobe: FATAL: Module i915 not found in directory /lib/modules/5.3.0-48-generic"
(i've tried to copy the i915 directory found in /drm in a previous kernel version without success)

[pcworld (pcworld)]

@ekinox09: Have you installed the linux-modules-extra-5.3.0-48-generic package? Should be included in there. Though at least for Eoan the updated kernels seem to be in the official repos now anyway.

On topic: The proposed -48 kernel works really well for me, haven't had a hang ever since updating (5 days ago), I grepped the logs for hangs[0]. Even before the -46 update I had an occasional hang about once a day, they seem to be gone so far. Thanks for backporting the fixes.

[0]: zgrep -ah 'Resetting rcs0 for hang on rcs0' /var/log/syslog*

[ekinox09 (ekinox09)]

Extra modules package installation solved my problem. Thank you. Starting test from now.

[ekinox09 (ekinox09)]

Still have the issue with the kernel proposed in message #57 (5.3.0-48-generic_5.3.0-48.41+lp1861395v1_amd64.deb).

[ 107.771197] i915 0000:00:10.0: Resetting rcs0 for stuck wait on rcs0
[ 121.775946] i915 0000:00:10.0: GPU HANG: ecode 9:0:0x00000000, hang on rcs0
[ 121.776956] i915 0000:00:10.0: Resetting rcs0 for hang on rcs0
[ 225.752786] i915 0000:00:10.0: Resetting vcs0 for hang on vcs0, vecs0
[ 225.753827] i915 0000:00:10.0: Resetting vecs0 for hang on vcs0, vecs0

Note: I reproduce the issue with the same program (Kodi media player)
Note: The problem occured regularly before; here, i observed 1 HANG within 10 minutes (was observed many times per minute before)
Note: The only kernel without this error i've tested is this one "https://kernel.ubuntu.com/~kernel-ppa/mainline/drm-intel-next/2020-03-14/". But it is not stable; so, the best result is clearly with 5.3.0-48.

[Chris Glass (tribaal)]

For the record, this still happens with the latest Eoan kernel (5.3.0-51-generic).

[ 68.285097] i915 0000:00:02.0: GPU HANG: ecode 9:0:0x00000000, hang on rcs0
[ 68.286107] i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
[ 76.301387] i915 0000:00:02.0: Resetting rcs0 for hang on rcs0

[Uwe Terborg (madmax1403)]

Hi,
where can i get the fix? i'm on Ubuntu 18.04. with the kernel 5.3.0-46

BR
Uwe

[Eduard Biceri-Matei (cubusbacau)]

@mlavigne #82
That fixed it for me... PyCharm works now, no hangs anymore.
Kernel 5.3.0-7648 PopOS 19.10
Pycharm 2019.3.4

[Sledge HaMMeR (kh-ubuntuone)]

@Uwe Terborg (madmax1403) #88

I'm passively reading along, and AFAIK there is no official/proper fix yet.
In the meantime, boot from kernel 5.3.0-45-generic

[Benjamin Gemmill (benjamin-gemmill)]

To save others time, I tested a bunch of kernels yesterday.

5.3.0-51.44 from here:
https://launchpad.net/ubuntu/+source/linux/
still shows the issue.

dmesg:
[43374.714454] [drm:intel_pipe_update_end [i915]] *ERROR* Atomic update failure on pipe A (start=12240 end=12241) time 150 us, min 2385, max 2399, scanline start 2379, end 2401
[44804.600121] i915 0000:00:02.0: GPU HANG: ecode 11:1:0xfffffffe, in Xorg [1075], hang on rcs0
[44804.601152] i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
[44807.725143] Asynchronous wait on fence i915:cinnamon[1445]:9138 timed out (hint:intel_atomic_commit_ready+0x0/0x54 [i915])
[44812.590311] i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
[44820.590611] i915 0000:00:02.0: Resetting rcs0 for hang on rcs0

The latest mainline kernels 5.6.7, 5.4.35, and drm-tip as of 2020-04-27 from here:
https://kernel.ubuntu.com/~kernel-ppa/mainline/
Also show the issue.

@kleber-souza: if there was a fix committed, can you please point to where? Glad to test it out.

[Chris Glass (tribaal)]

Reverting the status of that bug to confirmed was a bit premature :)

My previous comment already mentions 5.3.0-51-generic not being fixed (before it being marked fix commited).

From IRC:

"we should get a eoan/linux kernel in -proposed at the latest by tomorrow, when that happens an automated comment will be added to the bug report"

So, when a package with the comited fix is available ("tomorrow") you can test it again. 5.3.0-51.44 is not the version containing the patch, however.

[RickyUnix (riccardo-noc)]

Same issue with my DELL XPS 15 7590 and 5.3.0-46-generic

The only way was reverting to the previous 5.3.0-45-generic

By the way I'm also experiencing some

[280608.468810] mce: CPU13: Package temperature above threshold, cpu clock throttled (total events = 1027143)
[280608.468811] mce: CPU5: Package temperature above threshold, cpu clock throttled (total events = 1027144)
[280608.468812] mce: CPU3: Package temperature above threshold, cpu clock throttled (total events = 1027144)
[280608.468813] mce: CPU11: Package temperature above threshold, cpu clock throttled (total events = 1027143)
[280608.468815] mce: CPU15: Package temperature above threshold, cpu clock throttled (total events = 1027144)
[280608.468815] mce: CPU7: Package temperature above threshold, cpu clock throttled (total events = 1027143)
[280608.468817] mce: CPU6: Package temperature above threshold, cpu clock throttled (total events = 1027144)
[280608.468817] mce: CPU14: Package temperature above threshold, cpu clock throttled (total events = 1027144)

after the revert, very often. (dmesg logs)

[Matthieu Vion (magentix)]

Same experience with 5.3.0-51-generic on Bionic. Randomly freeze for 5 seconds.

Apr 30 10:30:45 xxx kernel: [ 7955.957968] i915 0000:00:02.0: GPU HANG: ecode 9:0:0x00000000, hang on rcs0
Apr 30 10:30:45 xxx kernel: [ 7955.958978] i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
Apr 30 10:41:41 xxx kernel: [ 8611.960926] i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
Apr 30 10:44:05 xxx kernel: [ 8755.986655] i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
Apr 30 10:50:34 xxx kernel: [ 9144.005272] i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
Apr 30 11:18:01 xxx kernel: [10791.905314] i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
Apr 30 11:40:08 xxx kernel: [12118.887548] i915 0000:00:02.0: Resetting rcs0 for hang on rcs0

[Sledge HaMMeR (kh-ubuntuone)]

Can confirm the issue ("Resetting rcs0 for hang on rcs0") is still around using kernel 5.3.0-51-generic (ubuntu 18.04 LTS, HP i7 Envy laptop). However, it seems to me a lot less frequent, only happened once over the last couple hours.

With that being said, I can see the "Package temperature above threshold, cpu clock throttled" now as well and quite often too :(

[Rich (richintheusa)]

I can also confirm that this problem continues in 5.3.0-51-generic

However, I see it very frequently (not just for 5 seconds... happening all the time) with two different use cases.

1) When i have a headset (with a mic) plugged into the front audio port. One of my cores also goes to 100% utilization.

or

2) When I run Zoom conferencing software.

Here is the journalctl output with 5.3.0-51-generic:

journalctl -b | grep i915
Apr 30 08:02:17 breath-mint kernel: i915 0000:00:02.0: vgaarb: deactivate vga console
Apr 30 08:02:17 breath-mint kernel: i915 0000:00:02.0: vgaarb: changed VGA decodes: olddecodes=io+mem,decodes=io+mem:owns=io+mem
Apr 30 08:02:17 breath-mint kernel: [drm] Finished loading DMC firmware i915/kbl_dmc_ver1_04.bin (v1.4)
Apr 30 08:02:17 breath-mint kernel: [drm] Initialized i915 1.6.0 20190619 for 0000:00:02.0 on minor 1
Apr 30 08:02:17 breath-mint kernel: fbcon: i915drmfb (fb0) is primary device
Apr 30 08:02:17 breath-mint kernel: i915 0000:00:02.0: fb0: i915drmfb frame buffer device
Apr 30 08:02:26 breath-mint kernel: snd_hda_intel 0000:00:1f.3: bound 0000:00:02.0 (ops i915_audio_component_bind_ops [i915])
Apr 30 08:03:11 breath-mint kernel: [drm:intel_pipe_update_end [i915]] *ERROR* Atomic update failure on pipe B (start=4635 end=4636) time 115 us, min 1073, max 1079, scanline start 1072, end 1080
Apr 30 08:04:15 breath-mint kernel: [drm:intel_pipe_update_end [i915]] *ERROR* Atomic update failure on pipe B (start=8521 end=8522) time 130 us, min 1073, max 1079, scanline start 1072, end 1081
Apr 30 08:04:40 breath-mint kernel: i915 0000:00:02.0: GPU HANG: ecode 9:0:0x00000000, hang on rcs0
Apr 30 08:04:40 breath-mint kernel: i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
Apr 30 08:04:48 breath-mint kernel: i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
Apr 30 08:04:56 breath-mint kernel: i915 0000:00:02.0: Resetting rcs0 for hang on rcs0

My temporary work around is to downgrade my kernel to 5.0. I do not experience the problem there.

Thanks

[Zuccster (uhdoch06y)]

Still there with 5.3.0-51-generic

Apr 30 15:59:53 hobbes kernel: [13727.891384] i915 0000:00:02.0: GPU HANG: ecode 9:0:0x00000000, hang on rcs0
Apr 30 15:59:53 hobbes kernel: [13727.892392] i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
Apr 30 16:00:01 hobbes kernel: [13735.890719] i915 0000:00:02.0: Resetting rcs0 for hang on rcs0

[Ubuntu Kernel Bot (ubuntu-kernel-bot)]

This bug is awaiting verification that the kernel in -proposed solves the problem. Please test the kernel and update this bug with the results. If the problem is solved, change the tag 'verification-needed-eoan' to 'verification-done-eoan'. If the problem still exists, change the tag 'verification-needed-eoan' to 'verification-failed-eoan'.

If verification is not done by 5 working days from today, this fix will be dropped from the source code, and this bug will be closed.

See https://wiki.ubuntu.com/Testing/EnableProposed for documentation how to enable and use -proposed. Thank you!

[Chris Glass (tribaal)]

I can confirm that the problem is fix in the -proposed kernel (5.3.0-52-generic).

Marking verification-done.

[jraby (raby-jean)]

What would be a good way to test the proposed 5.3.0-52-generic kernel for eoan on 18.04?
add the eoan proposed repo and install only the kernel pkgs?

[Chris Glass (tribaal)]

For all other subscribers to this bug, now is the time to help :)

Enable -proposed on affected systems:

$ echo "deb http://archive.ubuntu.com/ubuntu/ eoan-proposed main" >> /etc/apt/source.list

Update apt sources:

$ apt update

Install the linux kernel (and friends) from -proposed:

$ sudo apt install linux-headers-5.3.0-52 linux-headers-5.3.0-52-generic linux-image-5.3.0-52-generic linux-modules-5.3.0-52-generic linux-modules-extra-5.3.0-52-generic

Remove or comment the -proposed line in /etc/apt

Reboot.

Verify that the bug if fixed (and that nothing else broke), then report here :)

[Rich (richintheusa)]

I followed the steps above on Linux Mint, and the output of "journalctl -b | grep i915" no longer had the "GPU HANG or Resetting rcs0 for hang on rcs0" errors.

I also confirmed that running Zoom no longer has any lags or hangs. It works as expected.

So confirming this fix worked!

(Sorry, I'm new to this forum and not sure if I'm supposed to put tags on my message like Chris did above).

tags: added: verification-done-eoan
removed: verification-needed-eoan

Thanks.

[jraby (raby-jean)]

Initial testing suggests that the problem is resolved in 5.3.0-52.

I saw hangs every time I'd try to crop a picture in gThumb, now it works like it should.
(haven't / cannot test related Zoom video hangs, but will report back if not fixed.)

[Chris Glass (tribaal)]

Perfect, no need to change tags anymore.

Now that at least 2 persons verified the bug is fixed with the -proposed kernel, the SRU process will follow its course now and the fixed linux will land in the archives in due time: there is a mandatory 7 days wait time, so it should hit the archives next week, unless another fix rolled in this SRU breaks something else (I hope not :) ).

Thanks for your help!

[Egbert van der Wal (eggie)]

I've done some testing with 5.3.0-52 over the past few hours and it seems to work properly. However, the kernel in #57 also worked significantly better than the one from the repositories but I did still experience the GPU HANG a couple of times over the course of 2 days so I am not completely certain yet. Until today, 5.4.28-050428 has been the only one I tested that worked without any issues over several weeks. I'll keep running this one to see if the issue still pops up or not.

Is 5.3.0-52 in proposed different from the patches applied in #57?

[Benjamin Gemmill (benjamin-gemmill)]

Unfortunately this problem still persists. i915 error log attached.

$ uname -a
Linux spore 5.3.0-52-generic #46-Ubuntu SMP Wed Apr 29 20:35:46 UTC 2020 x86_64 x86_64 x86_64 GNU/Linux

# dmesg
[ 662.818303] i915 0000:00:02.0: GPU HANG: ecode 11:1:0xfffffffe, in Xorg [1113], hang on rcs0
[ 662.819457] i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
[ 670.809758] i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
[ 678.809853] i915 0000:00:02.0: Resetting rcs0 for hang on rcs0

[Robert Hardy (rhardy)]

I thought 5.3.0-48.41+lp1861395v1 fixed it for me. I haven't logged a hang for a very long time.
Now out of the blue my audio stopped working altogether. I tried rebooting into 5.3.0-51.44 and I still don't have audio but I'm not hanging with that either.

[Robert Hardy (rhardy)]

Just an FYI for those trying to figure out what the heck is going on.

The fix related to this thread wasn't in 5.3.0-51.44.
From the changelog, which is hard to find, the fix was committed in 5.3.0-52.46 which is currently in eoan-proposed:

5.3.0-46-generic - i915 - frequent GPU hangs / resets rcs0 (LP: #1872001)
    - drm/i915/execlists: Preempt-to-busy
    - drm/i915/gt: Detect if we miss WaIdleLiteRestore
    - drm/i915/execlists: Always force a context reload when rewinding RING_TAIL

It is not recommended to run proposed kernels for production systems but it is possible to add them to systems which would otherwise be broken (like mine right now) using this repository:
https://launchpad.net/~canonical-kernel-team/+archive/ubuntu/proposed

I had hoped since that also contained a bunch of audio fixes it might have fixed my inability to play sound but I haven't solved that one yet. I still have yet to see a hang on 5.3.0-52.46 either but again video playback with sound and heavy office document work seemed to trigger hangs for me so with no sound I could just be avoiding the problem...

[jraby (raby-jean)]

Ran all day yesterday with 5.3.0-52.46 (on 18.04), no GPU hang problem at all. So that's great.

However, I'm having an unrelated problem (only) with this kernel and I'm not sure where I should report it. Could someone direct me to the right place?

---

With 5.3.0-52.46 , I'm getting the following every time docker starts up (and any docker operation just hang). This doesn't happen with the latest 5.3.0 kernel for 18.04 (5.3.0-46.38~18.04.1):

May 2 08:33:57 x1 kernel: [ 13.231646] general protection fault: 0000 [#1] SMP PTI
May 2 08:33:57 x1 kernel: [ 13.231650] CPU: 3 PID: 1524 Comm: dockerd Tainted: G U 5.3.0-52-generic #46-Ubuntu
May 2 08:33:57 x1 kernel: [ 13.231651] Hardware name: LENOVO 20KHCTO1WW/20KHCTO1WW, BIOS N23ET71W (1.46 ) 02/20/2020
May 2 08:33:57 x1 kernel: [ 13.231656] RIP: 0010:ovl_open_realfile+0x4a/0x150 [overlay]
May 2 08:33:57 x1 kernel: [ 13.231658] Code: 44 8b 6f 40 65 48 8b 04 25 28 00 00 00 48 89 45 d0 31 c0 49 8b 7e 28 41 81 cd 00 00 04 04 e8 cd d4 ff ff 49 89 c7 48 8b 45 c8 <48> 8b 40 68 48 81 78 60 62 6a 65 6a 0f 84 a6 00 00 00 65 48 8b 04
May 2 08:33:57 x1 kernel: [ 13.231659] RSP: 0018:ffffc32f011abc08 EFLAGS: 00010202
May 2 08:33:57 x1 kernel: [ 13.231660] RAX: afb362bc303d5400 RBX: ffff9ee1f2ae1a68 RCX: 0000000000000001
May 2 08:33:57 x1 kernel: [ 13.231661] RDX: 0000000000008000 RSI: ffff9ee1f2ae1a68 RDI: ffff9ee23bf18780
May 2 08:33:57 x1 kernel: [ 13.231661] RBP: ffffc32f011abc50 R08: ffff9ee238f0ccc0 R09: 0000000000000000
May 2 08:33:57 x1 kernel: [ 13.231662] R10: ffff9ee23cdf8600 R11: 0000000000000000 R12: ffff9ee1ebc24b00
May 2 08:33:57 x1 kernel: [ 13.231663] R13: 0000000004048001 R14: ffff9ee1f2ac2d48 R15: ffff9ee238f0ccc0
May 2 08:33:57 x1 kernel: [ 13.231664] FS: 00007fc82cff9700(0000) GS:ffff9ee2424c0000(0000) knlGS:0000000000000000
May 2 08:33:57 x1 kernel: [ 13.231665] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
May 2 08:33:57 x1 kernel: [ 13.231665] CR2: 000000c000193000 CR3: 0000000431906005 CR4: 00000000003606e0
May 2 08:33:57 x1 kernel: [ 13.231666] Call Trace:
May 2 08:33:57 x1 kernel: [ 13.231670] ? ovl_llseek+0x120/0x120 [overlay]
May 2 08:33:57 x1 kernel: [ 13.231673] ovl_open+0x57/0x90 [overlay]
May 2 08:33:57 x1 kernel: [ 13.231675] do_dentry_open+0x143/0x3a0
May 2 08:33:57 x1 kernel: [ 13.231677] vfs_open+0x2d/0x30
May 2 08:33:57 x1 kernel: [ 13.231678] do_last+0x194/0x8f0
May 2 08:33:57 x1 kernel: [ 13.231680] path_openat+0x8d/0x270
May 2 08:33:57 x1 kernel: [ 13.231682] ? __slab_free+0x217/0x330
May 2 08:33:57 x1 kernel: [ 13.231683] do_filp_open+0x91/0x100
May 2 08:33:57 x1 kernel: [ 13.231685] ? __alloc_fd+0x46/0x150
May 2 08:33:57 x1 kernel: [ 13.231687] do_sys_open+0x17e/0x290
May 2 08:33:57 x1 kernel: [ 13.231688] __x64_sys_openat+0x20/0x30
May 2 08:33:57 x1 kernel: [ 13.231690] do_syscall_64+0x5a/0x130
May 2 08:33:57 x1 kernel: [ 13.231692] entry_SYSCALL_64_after_hwframe+0x44/0xa9
May 2 08:33:57 x1 kernel: [ 13.231694] RIP: 0033:0x56200aa6b4fa
May 2 08:33:57 x1 kernel: [ 13.231695] Code: e8 8b 3d f8 ff 48 8b 7c 24 10 48 8b 74 24 18 48 8b 54 24 20 4c 8b ...

[Robert Hardy (rhardy)]

The lack of kernel stability here is really getting frustrating....
With recent kernels audio playback over HDMI has stopped working for me altogether.
I'm getting zero traction resolving it.
I get audio playback when I plug in a headset but short cables are preventing use of that. Longer ones ordered but with pandemic ETA in weeks.
I was able to get audio over Bluetooth working as a temporary work around but that was so unstable that it crashed my computer so badly I could not reboot.
With the new 5.3.0-52.46 kernel I've seen both multiple crashes of blueman and kernel stack traces with kernels failing to the point of being unable to reboot once.
No i915 hangs though ugh...

[Benjamin Gemmill (benjamin-gemmill)]

i915 does still hang with the new 5.3.0-52.46 kernel, see the log in #106.

I can also confirm that docker is not happy with it.

While the changes may have fixed some things, and I'm glad that it's showing improvement for others, there are still hangs and X crashes there that are fairly easy for me to reproduce.

I'm happy testing any proposed kernel that comes along.

[Robert Hardy (rhardy)]

Here is the syslog content just before the hard crash on proposed 5.3.0-52 from which I could not reboot and finally had to hit my reset button. Please note this syslog extract contains a long string of UTF8 characters I had never seen before in a syslog just before the crash where I had to hit the reset button. I did not remove those.
For lack of a better option, I'm still running it and BT audio has worked so far today..

[Robert Hardy (rhardy)]

I got about another 4 hrs before I had a 2nd hard crash with 5.3.0-52. I once again could not shutdown or reboot and was forced to hit my reset button. What a mess....

Basically BT audio works until the connection stops. Pausing an audio stream encourages the BT audio to unpair. The Audio sink still shows connected in Linux even though my DVD player shows it has unpaired.

After that happens does it doesn't do that cleanly and causes a kernel panic when you attempt to reconnect the audio. Afterwards you cannot cleanup the broken audio sink or power off Bluetooth.
Any kind of authentication seems to freeze the session. Things are very broken.

[Robert Hardy (rhardy)]

After it was pretty clear a stable kernel wasn't coming any time soon for 19.10 I gave up and took a dangerous leap. The only way I was going to get a 5.4 kernel was to move to 20.04.
It was dangerous and I had to get there through a do-release-upgrade -d. Highly not recommended if you don't know what you are doing. I basically backed everything up and planned on things failing.
I did it from the command line from multiple root shells launched ahead of time using screen in case X died.

I had Virtualbox 6.1.6 installed before I went through my upgrade. It was not a clean upgrade as something as part of the upgrade process forced the uninstall of Virtualbox 6.1.6.

I suspect it was because some of the dependencies of Virtualbox 6.1.6 only existed in Universe on focal Virtualbox 6.1.6 was uninstalled as part of the upgrade to 20.04.

The upgrade actually blew up but I knew enough to recover. That blew a lot of things out of the water and was not listed on the upgrade screen or I would not have proceeded. I shutdown virtualbox and let it be temporarily uninstalled. Then I did a apt-get dist-upgrade and once that completed I tried to install Virtualbox 6.1.6 again but failed due to missing dependencies. I regretted doing an apt-get autoremove --purge just before hand.
I had to manually download and install libqt5opengl5:amd64 for focal and then an apt-get install -f was able to fix python and in turn Virtualbox 6.1.6. Then once Virtualbox installed things worked again.

After all the hell on 19.10 suddenly all my problems are gone. My NIC drivers still work thanks to the updated Intel e1000e 3.8.4 release DKMS package which Intel engineering kindly provided. That badly needs to be native to Ubuntu 20.04 as the kernel driver at least back on 5.3.0 didn't work for me.

I have seen no i915 hangs with the kernel 5.4.0-28-generic. I haven't seen any flaky Bluetooth behavior that led to kernel panics but to be fair I didn't need to use Bluetooth audio sink on 20.04 as my HDMI audio pass-through works properly again!

[Egbert van der Wal (eggie)]

@rhardy good that it worked all out for you.

However, just for your information: upgrade to 20.04 wasn't necessary to run a newer kernel. I've been running a 5.4 kernel for quote some time on 19.10 using the mainline utitlity from http://ppa.launchpad.net/cappelikan/ppa/ubuntu which works perfectly.

And after a couple of days of testing on 5.3.0-52 from eoan-proposed, I now feel confident enough to say that it fixes the i915 issues I experienced before, so it definitely improves the situation.

[Michael Unger (unger123)]

Had massive problems with GUI freezes on 18.04 (especially when running CERN's ROOT). Installing the 5.3.0-52-generic kernel fixed it for me, thanks a lot Chris!

[Rob Cork (rcork)]

I haven't had a hang since installing 5.3.0-52-generic on April 30th.

[Guillaume FRANCOIS (guillaume-francois55)]

Hello,

I installed today "5.3.0-52-generic #46-Ubuntu SMP" following comment https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1872001/comments/101 (and fixing the typo for /etc/sourceS.list)

It removed all visible hang issued I have before especially with IntelliJ and the whole desktop.

Since I boot at 15h36 I got only 3 lines in journalctl, which is way less then before.

mai 04 15:44:08 francogu-linux-7540 kernel: [drm:intel_pipe_update_end [i915]] *ERROR* Atomic update failure on pipe C (start=62403 end=62404) time 145 us, min 1073, max 1079, scanline start 1070, end 1079
mai 04 15:57:58 francogu-linux-7540 kernel: [drm:intel_pipe_update_end [i915]] *ERROR* Atomic update failure on pipe C (start=112170 end=112171) time 162 us, min 1073, max 1079, scanline start 1069, end 1079
mai 04 17:18:25 francogu-linux-7540 kernel: [drm:intel_pipe_update_end [i915]] *ERROR* Atomic update failure on pipe B (start=401843 end=401844) time 210 us, min 1073, max 1079, scanline start 1072, end 1086

[Uwe Terborg (madmax1403)]

Hi,

i had the problems with 5.3.0-28, 5.3.0-45, 5.3.0-46 and 5.3.0-51 but not with 4.18.0-25!

BR
Uwe

[Chris Glass (tribaal)]

Indeed, it seems like the kernel linked above breaks dockerd.

dmesg with "general protection fault" and stacktrace. all docker commands hang.

https://pastebin.ubuntu.com/p/VRbFSmPkPs/

:(

[Kleber Sacilotto de Souza (kleber-souza)]

Hi,

Thank you all for your testing. We will investigate further the issues reported by @benjamin-gemmill, but it seems to be a improvement already.

Regarding the issues with docker, it seems to be the same root cause as the issue reported on bug 1876645. We are investigating it and we will probably release a new kernel in -proposed soon to fix this issue.

[Ohad Lutzky (lutzky)]

Can confirm that 5.3.0-52-lowlatency does not appear to exhibit the i915 issue on my Lenovo E490, with the same testing from #34. Thank you!

[Andrea Righi (arighi)]

According to comment #106 it looks like the bug is still happening also in 5.3.0-52-generic, so maybe we have just reduced the probability to hit the bug.

Let's try a different approach.

Assuming that the bug wasn't happening in -45 and it started to happen in -46, the bug is likely to be introduced by:

 411b24761d28 "drm/i915: Schedue request retirement in execlists_user_end()"

that was supposed to fix a power consumption issue.

The alternative approach that I would like to try is to revert all the commits that I backported and revert also 411b24761d28. This would re-introduce the power consumption issue, but if we are not hitting the GPU HANG anymore, it would tell us the exact commit that introduced the bug and we can figure out an alternative way of fixing the power consumption problem.

So, I've upload a new test kernel (5.3.0-52.46+lp1861395v1) with the changes mentioned above:
https://kernel.ubuntu.com/~arighi/LP-1861395/5.3.0-52.46+lp1861395v1/

It would be great if someone could give it a try and report if they are still experiencing the GPU HANG bug also with this new test kernel. Thanks!

[pcworld (pcworld)]

Andrea, sorry to disappoint you, but I've definitely had this bug even in earlier 5.3 kernel versions (i.e. also on -45), just with a lower frequency. See my comment 24 (https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1872001/comments/24) for details.

Interestingly contrary to others, I haven't had it happen once since I upgraded to your 5.3.0-48-generic kernel.
I use the following command to check syslog for freezes:
 zgrep -ah 'Resetting rcs0 for hang on rcs0' /var/log/syslog*

[Robert Hardy (rhardy)]

I also had the issue on the 5.3.0-45 kernel but it happened once for 2 seconds every two days.
It wasn't even noticeable unless you really looked for it. The 5.3 kernel series seems to be a total write off at this point.

While it likely isn't allowed as a solution it might be less effort to back port the focal 5.4.0-28-generic kernel to eoan. I've been on that for several days. Since I upgraded to focal, I've seen no BT kernel panics, no issues with HDMI passthrough audio not working, no i915 hangs or any of the other issues I was seeing on the 5.3.0-x kernels.

[Benjamin Gemmill (benjamin-gemmill)]

@arighi and @kleber-souza the issue persists in the test kernel.

i915 error log attached.

#uname -a
Linux spore 5.3.0-52-generic #46+lp1861395v1 SMP Tue May 5 16:19:02 UTC 2020 x86_64 x86_64 x86_64 GNU/Linux

#dmesg
[ 166.187536] i915 0000:00:02.0: GPU HANG: ecode 11:1:0xfffffffe, in Xorg [1081], hang on rcs0
[ 166.188564] i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
[ 169.281006] Asynchronous wait on fence i915:cinnamon[1449]:d4e timed out (hint:intel_atomic_commit_ready+0x0/0x58 [i915])
[ 174.178202] i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
[ 182.178161] i915 0000:00:02.0: Resetting rcs0 for hang on rcs0

If it helps rule out anything, I've observed the issue as far back as 5.3.0-19, and the mesa folks have been chasing it since December 2019:
https://gitlab.freedesktop.org/mesa/mesa/issues/2026

I'm able to reproduce this pretty reliably by generating a lot of text in either gnome terminal or a Jetbrains IDE and scrolling up and down through it.

The workaround in #82 helps with Jetbrains, and oibaf's ppa's mesa with KMS seems to make gnome terminal happy.

Removing either lets me test this out. Still happy to test any proposed kernels :)

[Andrea Righi (arighi)]

@pcworld @rhardy @benjamin-gemmill thank you very much for the feedback! We will keep investigating about this bug.

[Nicolas (n-herve)]

Same bug for me, kernel 5.3.0-51-generic, ubuntu 19.10. It only happens when working on PyCharm (2 sec freeze, mouse moving, getting back what I typed after the freeze)

[RickyUnix (riccardo-noc)]

I did some investigations about the history of kernels run in my system before the issue through journalctl:

----------------------

root=UUID=.... ro quiet splash mem_sleep_default=deep vt.handoff=1
2020-03-31T10:33:14+0200 kernel: Command line: BOOT_IMAGE=/boot/vmlinuz-5.3.0-42-generic root=UUID=.... ro quiet splash mem_sleep_default=deep vt.handoff=1
2020-03-31T10:55:30+0200 kernel: Command line: BOOT_IMAGE=/boot/vmlinuz-5.3.0-42-generic root=UUID=.... ro quiet splash mem_sleep_default=deep vt.handoff=1
2020-04-21T09:13:43+0200 kernel: Command line: BOOT_IMAGE=/boot/vmlinuz-5.3.0-46-generic root=UUID=.... ro quiet splash mem_sleep_default=deep vt.handoff=1
2020-04-21T10:16:47+0200 kernel: i915 0000:00:02.0: GPU HANG: ecode 9:0:0x00000000, hang on rcs0
2020-04-21T10:16:47+0200 kernel: i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
2020-04-21T10:19:25+0200 kernel: i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
2020-04-21T10:19:58+0200 kernel: i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
2020-04-21T10:21:23+0200 kernel: i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
2020-04-21T10:21:51+0200 kernel: i915 0000:00:02.0: Resetting rcs0 for hang on rcs0

[...]

2020-04-21T15:02:56+0200 kernel: i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
2020-04-21T15:08:37+0200 kernel: i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
2020-04-21T15:08:53+0200 kernel: i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
2020-04-21T15:20:57+0200 kernel: i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
2020-04-21T15:21:05+0200 kernel: i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
2020-04-21T15:43:26+0200 kernel: i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
2020-04-21T15:45:21+0200 kernel: Command line: BOOT_IMAGE=/boot/vmlinuz-5.3.0-45-generic root=UUID=.... ro quiet splash mem_sleep_default=deep vt.handoff=1

----------------------

I can say for sure that it started with 5.3.0-46, but when i reverted to the 5.3.0-45, that i never run before the issue (because i did a jump to 45) the hangs disappeared but the cpu now overheats very easily... the fan are always on when I use something like IDEs.

----------------------

[280608.468815] mce: CPU7: Package temperature above threshold, cpu clock throttled (total events = 1027143)
[280608.468817] mce: CPU6: Package temperature above threshold, cpu clock throttled (total events = 1027144)
[280608.468817] mce: CPU14: Package temperature above threshold, cpu clock throttled (total events = 1027144)

----------------------

I am considering going back to 5.3.0-42

[htrex (hantarex)]

I'm using a Dell XPS 9560 and found that booting with kernel 5.3.0-45 avoids IntelliJ's IDE (PhpStorm) to trigger the GPU hang, while the workaround on comment #82 to doesn't make any difference when using kernel 5.3.0-46 or 5.3.0-51.

[jraby (raby-jean)]

Another datapoint:

I just tested again on 5.3.0-43 since I thought I remember having seen the issue there too.

It does have the problem:

May 7 09:03:25 x1 kernel: [ 0.000000] Linux version 5.3.0-43-generic (buildd@lcy01-amd64-024) (gcc version 7.5.0 (Ubuntu 7.5.0-3ubuntu1~18.04)) #36~18.04.2-Ubuntu SMP Thu Mar 19 16:03:35 UTC 2020 (Ubuntu 5.3.0-43.36~18.04.2-generic 5.3.18)
May 7 09:04:27 x1 kernel: [ 69.099086] i915 0000:00:02.0: GPU HANG: ecode 9:0:0x00000000, hang on rcs0
May 7 09:04:27 x1 kernel: [ 69.100094] i915 0000:00:02.0: Resetting rcs0 for hang on rcs0

(I've gone back to 4.15 to avoid this problem and the docker issue mentioned above)

[houdini68 (houdini68)]

May 7 09:02:11 rudy-ThinkPad-T590 kernel: [16362.901051] i915 0000:00:02.0: GPU HANG: ecode 9:0:0x00000000, hang on rcs0
May 7 09:02:11 rudy-ThinkPad-T590 kernel: [16362.902070] i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
May 7 09:50:21 rudy-ThinkPad-T590 kernel: [19252.894099] i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
May 7 14:35:29 rudy-ThinkPad-T590 kernel: [32445.859961] i915 0000:00:02.0: Resetting rcs0 for hang on rcs0

On 14:35 it was very noticeable...

rudy@rudy-ThinkPad-T590:~$ inxi -SCGxxxz to report
System: Host: rudy-ThinkPad-T590 Kernel: 5.3.0-51-generic x86_64 bits: 64 gcc: 7.5.0
           Desktop: Gnome 3.28.4 (Gtk 3.22.30-1ubuntu4) info: gnome-shell dm: gdm3 Distro: Ubuntu 18.04.4 LTS
CPU: Quad core Intel Core i7-8665U (-MT-MCP-) arch: Kaby Lake rev.12 cache: 8192 KB
           flags: (lm nx sse sse2 sse3 sse4_1 sse4_2 ssse3 vmx) bmips: 16799
           clock speeds: min/max: 400/4800 MHz 1: 800 MHz 2: 800 MHz 3: 800 MHz 4: 800 MHz 5: 800 MHz 6: 800 MHz
           7: 800 MHz 8: 800 MHz
Graphics: Card: Intel Device 3ea0 bus-ID: 00:02.0 chip-ID: 8086:3ea0
           Display Server: x11 (X.Org 1.20.5 ) drivers: fbdev (unloaded: modesetting,vesa)
           Resolution: 1600x900@59.99hz, 1920x1080@60.00hz
           OpenGL: renderer: Mesa DRI Intel UHD Graphics (Whiskey Lake 3x8 GT2)
           version: 4.5 Mesa 19.2.8 (compat-v: 3.0) Direct Render: Yes
rudy@rudy-ThinkPad-T590:~$

[Marcelo Vanzin (vanzin)]

I was also running into pretty horrendous graphics perf (with bionic), and fixed it after I did the following (did all at once so not sure which one did it):

- updated to the 5.3.0-51 kernel
- set the java2d option to opengl (from #82) - although I had issues in other apps too
- updated the Xorg server to xserver-xorg-hwe-18.04 (+ other hwe drivers)

My guess is that the last one is what did the trick. Much better for the last couple of days.

[Egbert van der Wal (eggie)]

Another occurrence of this bug for me too, unfortunately:

[23965.090982] i915 0000:00:02.0: GPU HANG: ecode 9:0:0x00000000, hang on rcs0
[23965.091992] i915 0000:00:02.0: Resetting rcs0 for hang on rcs0

This is the first for me in a week of using 5.3.0-52-generic so it's definitely less common.

[ekinox09 (ekinox09)]

Hello, it took me few seconds running Kodi to encounter the problem with 5.3.0-52.46 from https://kernel.ubuntu.com/~arighi/LP-1861395/5.3.0-52.46+lp1861395v1/

[ 177.891537] i915 0000:00:10.0: Resetting bcs0 for hang on bcs0
[ 179.875568] i915 0000:00:10.0: Resetting bcs0 for hang on bcs0
[ 181.890600] i915 0000:00:10.0: GPU recovery timed out, cancelling all in-flight rendering.
[ 181.891827] i915 0000:00:10.0: Resetting chip for hang on bcs0

In my case, the problem occurs more with 5.3.0-52 than 5.3.0-51.

[Robert Hardy (rhardy)]

To be clear 5.3.0-51 did NOT include the backported patches.
Proposed 5.3.0-52.46 had some and 5.3.0-52.46+lp1861395v1 had more.
Out of curiosity could the people who are still having hangs on the 5.3.0-52.46 proposed or 5.3.0-52.46+lp1861395v1 kernel have different generations of CPUs?

I've got a Coffee Lake CPU and the patches seemed to do a better job of fixing issues for me.
Perhaps we really have two different issues here.

@pcworld we saw similar behaviour. What is your CPU model?
Here is mine:
# grep "model name" /proc/cpuinfo
CPU model name : Intel(R) Core(TM) i3-8100 CPU @ 3.60GHz

[pcworld (pcworld)]

@rhardy I have an i5-8250U and hangs have stopped ever since updating to the proposed 5.3.0-48-generic kernel (didn't try newer kernels yet).

[Guillaume FRANCOIS (guillaume-francois55)]

As written in https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1872001/comments/118,

I had no hangs with "5.3.0-52-generic" or the extended one "lp1861395v1"

I did not check the logs with "5.3.0-52.46+lp1861395v1" but with the previous version the number of entries where very limited.

Unfortunatly I had to rollback to 5.3.0-51 as docker integration was a totally broken and preventing daily work.

In the meantime, I tried to upgrade to Ubuntu 20.04 with kernel 5.4.0x but the upgrade broke docker and even using docker.io instead of docker-ce did not fix the issue, as docker-ce is not yet available on 20.04.

I did not have the time to check for the "hangs" on 20.04 with kernel 5.4.0 but I can give a try if needed.

Later on, I'll probably do a format to perform a clean install on 20.04 to check is docker can work.

My laptop is a Dell Precision 7450 with an Intel i9-9880H.

[Robert Hardy (rhardy)]

Hmm so patches were effective for:
i5-8250U Kaby Lake R i.e. Gen 8
i9-9880H and i3-8100 both Coffee Lake Gen 9

It's hard enough debugging this directly on hardware. Doing it in a VM isn't going to be helpful.
Which CPU models are still failing with 5.3.0-52.46 or 5.3.0-52.46+lp1861395v1?

[Benjamin Gemmill (benjamin-gemmill)]

i7-1065G7 Ice Lake Gen 10 is still showing issues with 5.3.0-52

[Rob Cork (rcork)]

Strange, i have an i7-1065G7 and haven't had the issue since upgrading to 5.3.0-52. @benjamin-gemmill, what are you doing to test the issue. Maybe i can replicate on my system to see if i get the error to. So far i've been error free since April 30th on this kernel.

System: Host: XPS-13-9300 Kernel: 5.3.0-52-generic x86_64 bits: 64 Desktop: Cinnamon 4.4.8 Distro: Linux Mint 19.3 Tricia
CPU: Topology: Quad Core model: Intel Core i7-1065G7 bits: 64 type: MT MCP L2 cache: 8192 KiB
           Speed: 1805 MHz min/max: 400/3900 MHz Core speeds (MHz): 1: 1187 2: 1158 3: 1196 4: 1200 5: 1200 6: 1200 7: 1202
           8: 1142
Graphics: Device-1: Intel driver: i915 v: kernel
           Display: x11 server: X.Org 1.20.5 driver: modesetting unloaded: fbdev,vesa
           resolution: 2560x1440~60Hz, 2560x1440~60Hz
           OpenGL: renderer: Mesa DRI Intel Iris Plus Graphics (Ice Lake 8x8 GT2) v: 4.5 Mesa 19.2.8

[Robert Hardy (rhardy)]

For me initially I saw it when working with large office documents and when attempting to do certain 1080p A/V playback. When it really got bad I couldn't use a browser as it was happening all the time.

[Dave Kosmal (p-dave-x)]

I had the issue with Mint cinnamon desktop running kernel 5.3.0-46-generic.

Reverted back to 5.3.0-45-generic and it was resolved.

I noticed it mostly when using Webstorm IDE, although I don't think it was specific to this application. Intermittent but very frequent.

[houdini68 (houdini68)]

Again
May 11 15:49:59 rudy-ThinkPad-T590 kernel: [ 4727.312939] i915 0000:00:02.0: GPU HANG: ecode 9:0:0x00000000, hang on rcs0
May 11 15:49:59 rudy-ThinkPad-T590 kernel: [ 4727.313946] i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
May 12 05:48:08 rudy-ThinkPad-T590 kernel: [11927.393097] i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
May 12 05:50:08 rudy-ThinkPad-T590 kernel: [12047.397104] i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
May 12 06:04:00 rudy-ThinkPad-T590 kernel: [12879.363685] i915 0000:00:02.0: Resetting rcs0 for hang on rcs0

[Mika Joenperä (mika-joenpera)]

>Which CPU models are still failing with
>5.3.0-52.46 or 5.3.0-52.46+lp1861395v1?
i7-9700k is failing..

[Rusted (steko-segmentas)]

I had a problem for a while, with PHPStorm being laggy on ubuntu (had this problem for over a year)
Recently lag became unusable, after I tweaked a few things in ubuntu config (i think I increased swappiness)
In Intellij discussion they linked to this linux kernel ticket.

Lags in a way of:
Phpstorm window freezes for 5 seconds
sometimes phpstorm does not unfreeze, need to kill the process
laggy scrolling inside Phpstorm file window
sometimes freezes inside settings menus
laptop fan spins up audibly more than usual
google chrome can become laggy while running phpstorm

Ubuntu 18.04, kernel 5.3.0-51
pc: lenovo t480,
graphics: intel 620
CPU: Intel® Core™ i5-8250U CPU @ 1.60GHz × 8

Could not figure out how to upgrade kernel version, because I already had stable one

Upgrade to 19.10 did not help, kernel stayed the same version

After I did upgrade to 20.04, it seems so far that problems disappeared. Now kernel is 5.4.0-29-generic

[Boštjan Žokš (bzoks)]

I can also confirm that after upgrade to Ubuntu 20.04 with kernel 5.4.0-28-generic (12 days ago) there was not a single occurrence of this GPU HANG - on both machines I reported earlier.

[Chris Higgins (chris-higgins)]

Thinkpad T490
 - Ubuntu 19.10
 - Intel Graphics UHD 620 (Whiskey Lake)
 - Intel i7-8665U
 - uptime since kernel upgrade : 7days

Upgrade to 5.3.0-52-generic (5.3.0-52.46) last week has fixed the problem completely.
No more reset messages in dmesg with daily use

[Egbert van der Wal (eggie)]

I've been running a couple of days not on the new release from eoan-proposed: 5.3.0-53-generic and this solved the docker issues introduced by earlier versions. So far no GPU hangs either. Not sure if anything related to this bug was changed in -53 in comparison to -52, but it's looking good.

[ekinox09 (ekinox09)]

My CPU is an Intel i5-7200 (with Intel HD 620)(Kaby lake). I can reproduce the issue running Kodi with 5.3.0-52.46+lp1861395v1

[sanette (sanette-linux)]

running 5.3.0-52-generic for a couple of weeks I still have occasional hangs, maybe not triggered by the same things. It seems for instance to happen when I open a PDF from within chromium-browser.
It lasts 3-4 seconds.

[12467.199744] perf: interrupt took too long (2520 > 2500), lowering kernel.perf_event_max_sample_rate to 79250
[12872.912144] i915 0000:00:02.0: GPU HANG: ecode 9:1:0xfffffffe, in chromium-browse [6961], hang on rcs0
[12872.913153] i915 0000:00:02.0: Resetting rcs0 for hang on rcs0

[Noe (noecordmart)]

Running Ubuntu 18.04.4 LTS with 5.3.0-51-generic in Intel i5-9600K (with Intel UHD Graphics 630).
- The system freezes every 5-10 minutes when running libreoffice (6.0.7.3).
- Running Gimp (2.10.18 from snap) it freezes every few senconds:
[ 2960.739364] i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
[ 2968.739685] i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
[ 2976.740123] i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
[ 2984.740477] i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
[ 2992.740852] i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
[ 3002.757318] i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
[ 3010.757689] i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
[ 3018.758094] i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
[ 3028.742571] i915 0000:00:02.0: Resetting rcs0 for hang on rcs0

[Launchpad Janitor (janitor)]

This bug was fixed in the package linux - 5.3.0-53.47

---------------
linux (5.3.0-53.47) eoan; urgency=medium

  * eoan/linux: 5.3.0-53.47 -proposed tracker (LP: #1877257)

  * Intermittent display blackouts on event (LP: #1875254)
    - drm/i915: Limit audio CDCLK>=2*BCLK constraint back to GLK only

  * Unable to handle kernel pointer dereference in virtual kernel address space
    on Eoan (LP: #1876645)
    - SAUCE: overlayfs: fix shitfs special-casing

linux (5.3.0-52.46) eoan; urgency=medium

  * eoan/linux: 5.3.0-52.46 -proposed tracker (LP: #1874752)

  * alsa: make the dmic detection align to the mainline kernel-5.6
    (LP: #1871284)
    - ALSA: hda: add Intel DSP configuration / probe code
    - ALSA: hda: fix intel DSP config
    - ALSA: hda: Allow non-Intel device probe gracefully
    - ALSA: hda: More constifications
    - ALSA: hda: Rename back to dmic_detect option
    - [Config] SND_INTEL_DSP_CONFIG=m
    - [packaging] Remove snd-intel-nhlt from modules

  * built-using constraints preventing uploads (LP: #1875601)
    - temporarily drop Built-Using data

  * ubuntu/focal64 fails to mount Vagrant shared folders (LP: #1873506)
    - [Packaging] Move virtualbox modules to linux-modules
    - [Packaging] Remove vbox and zfs modules from generic.inclusion-list

  * linux-image-5.0.0-35-generic breaks checkpointing of container
    (LP: #1857257)
    - SAUCE: overlayfs: use shiftfs hacks only with shiftfs as underlay

  * shiftfs: broken shiftfs nesting (LP: #1872094)
    - SAUCE: shiftfs: record correct creator credentials

  * Add debian/rules targets to compile/run kernel selftests (LP: #1874286)
    - [Packaging] add support to compile/run selftests

  * shiftfs: O_TMPFILE reports ESTALE (LP: #1872757)
    - SAUCE: shiftfs: fix dentry revalidation

  * getitimer returns it_value=0 erroneously (LP: #1349028)
    - [Config] CONTEXT_TRACKING_FORCE policy should be unset

  * 5.3.0-46-generic - i915 - frequent GPU hangs / resets rcs0 (LP: #1872001)
    - drm/i915/execlists: Preempt-to-busy
    - drm/i915/gt: Detect if we miss WaIdleLiteRestore
    - drm/i915/execlists: Always force a context reload when rewinding RING_TAIL

  * alsa/sof: external mic can't be deteced on Lenovo and HP laptops
    (LP: #1872569)
    - SAUCE: ASoC: intel/skl/hda - set autosuspend timeout for hda codecs

  * Eoan update: upstream stable patchset 2020-04-22 (LP: #1874325)
    - ARM: dts: sun8i-a83t-tbs-a711: HM5065 doesn't like such a high voltage
    - bus: sunxi-rsb: Return correct data when mixing 16-bit and 8-bit reads
    - net: vxge: fix wrong __VA_ARGS__ usage
    - hinic: fix a bug of waitting for IO stopped
    - hinic: fix wrong para of wait_for_completion_timeout
    - cxgb4/ptp: pass the sign of offset delta in FW CMD
    - qlcnic: Fix bad kzalloc null test
    - i2c: st: fix missing struct parameter description
    - cpufreq: imx6q: Fixes unwanted cpu overclocking on i.MX6ULL
    - media: venus: hfi_parser: Ignore HEVC encoding for V1
    - firmware: arm_sdei: fix double-lock on hibernate with shared events
    - null_blk: Fix the null_add_dev() error path
    - null_blk: Handle null_add_dev() failures properly
    - null_blk: fix spuri...

[Benjamin Gemmill (benjamin-gemmill)]

Unfortunately, the errors persist in 5.3.0-53.47

i915 error logs attached.

# uname -a
Linux spore 5.3.0-53-generic #47-Ubuntu SMP Thu May 7 12:18:16 UTC 2020 x86_64 x86_64 x86_64 GNU/Linux

# dmesg
[ 1537.362289] mce: CPU2: Core temperature above threshold, cpu clock throttled (total events = 2)
[ 1537.362290] mce: CPU6: Core temperature above threshold, cpu clock throttled (total events = 2)
[ 1537.362291] mce: CPU0: Package temperature above threshold, cpu clock throttled (total events = 2)
[ 1537.362292] mce: CPU4: Package temperature above threshold, cpu clock throttled (total events = 2)
[ 1537.362292] mce: CPU6: Package temperature above threshold, cpu clock throttled (total events = 2)
[ 1537.362294] mce: CPU2: Package temperature above threshold, cpu clock throttled (total events = 2)
[ 1537.362348] mce: CPU1: Package temperature above threshold, cpu clock throttled (total events = 2)
[ 1537.362349] mce: CPU7: Package temperature above threshold, cpu clock throttled (total events = 2)
[ 1537.362350] mce: CPU3: Package temperature above threshold, cpu clock throttled (total events = 2)
[ 1537.362352] mce: CPU5: Package temperature above threshold, cpu clock throttled (total events = 2)
[ 1537.364048] mce: CPU6: Core temperature/speed normal
[ 1537.364049] mce: CPU2: Core temperature/speed normal
[ 1537.364049] mce: CPU0: Package temperature/speed normal
[ 1537.364050] mce: CPU4: Package temperature/speed normal
[ 1537.364052] mce: CPU6: Package temperature/speed normal
[ 1537.364053] mce: CPU2: Package temperature/speed normal
[ 1537.364108] mce: CPU1: Package temperature/speed normal
[ 1537.364109] mce: CPU5: Package temperature/speed normal
[ 1537.364110] mce: CPU3: Package temperature/speed normal
[ 1537.364111] mce: CPU7: Package temperature/speed normal
[ 1990.917267] i915 0000:00:02.0: GPU HANG: ecode 11:1:0xfffffffe, in Xorg [1097], hang on rcs0
[ 1990.918390] i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
[ 1998.907399] i915 0000:00:02.0: Resetting rcs0 for hang on rcs0
[ 2006.907435] i915 0000:00:02.0: Resetting rcs0 for hang on rcs0

[jt (johannes-thraen)]

Since Linux X 5.3.0-53-generic it seems that I don't experience the hangs any more, but idle power consumption severely degraded: 5.3.0-51 uses 2/3 of the power that ..-53 uses. It's also worse than on 4.15.
Powertop showing GPU being powered on 100% of the time

[litjens (jhcl)]

5.3.0-53-generic solved the problem for me (xubuntu 19.10). No problems or rcs0 messages for 2 days now.
Thank you for your help.

[Noe (noecordmart)]

Solved in 5.3.0-53-generic for me.
In may case GPU is not always at 100%. It is scaling frequency well from 350Mhz to max

[RickyUnix (riccardo-noc)]

Dell XPS 15 7590 - 5.3.0-53-generic #47~18.04.1-Ubuntu

I can say that the problem is solved now

it no longer has hangs and the CPU has returned to work without overloads.
The fans work normally and the PC is no longer overheated

Thank you

[Viacheslav Semykrasov (7krasov)]

Dell Latitude 5500
Ubuntu 19.10, 5.3.0-53-generic - no freezing so far
Ubuntu 20.04, 5.4.0-31-generic - works perfectly too

[ekinox09 (ekinox09)]

Hi, with 5.3.0-53 generic, i've encountered only 1 error (i915 0000:00:10.0: Resetting vcs0 for hang on vcs0) but no impact on performance and no freeze until now. Will continue to test but seems promizing. Thank you.

[ekinox09 (ekinox09)]

Hi, with 5.3.0-53 generic, i've encountered 2 errors that have frozen the system during 2 or 3 minutes. After this error, the system came back to normal. These 2 errors has been observed with Kodi at the "beginning" of the session (in the firts minutes after a boot), and only once per boot. I'm still testing.
Here is the syslog when an error occurs:

May 30 23:30:57 hystKodi kernel: [ 150.872582] i915 0000:00:10.0: GPU HANG: ecode 9:0:0x00000000, hang on rcs0
May 30 23:30:57 hystKodi kernel: [ 150.873591] i915 0000:00:10.0: Resetting rcs0 for hang on rcs0
May 30 23:31:05 hystKodi kernel: [ 158.851005] i915 0000:00:10.0: Resetting rcs0 for hang on rcs0
May 30 23:31:13 hystKodi kernel: [ 166.850875] i915 0000:00:10.0: Resetting rcs0 for hang on rcs0
May 30 23:31:21 hystKodi kernel: [ 174.850907] i915 0000:00:10.0: Resetting rcs0 for hang on rcs0
May 30 23:31:23 hystKodi kernel: [ 176.866911] i915 0000:00:10.0: Resetting rcs0 for hang on rcs0
May 30 23:31:31 hystKodi kernel: [ 184.866958] i915 0000:00:10.0: Resetting rcs0 for hang on rcs0
May 30 23:31:39 hystKodi kernel: [ 192.866879] i915 0000:00:10.0: Resetting rcs0 for hang on rcs0
May 30 23:31:41 hystKodi kernel: [ 194.850880] i915 0000:00:10.0: Resetting rcs0 for hang on rcs0
May 30 23:31:43 hystKodi kernel: [ 196.866960] i915 0000:00:10.0: Resetting rcs0 for hang on rcs0
May 30 23:31:45 hystKodi kernel: [ 198.850929] i915 0000:00:10.0: Resetting rcs0 for hang on rcs0
May 30 23:31:53 hystKodi kernel: [ 206.850890] i915 0000:00:10.0: Resetting rcs0 for hang on rcs0
May 30 23:31:55 hystKodi kernel: [ 208.866949] i915 0000:00:10.0: Resetting rcs0 for hang on rcs0
May 30 23:31:57 hystKodi kernel: [ 210.850916] i915 0000:00:10.0: Resetting rcs0 for hang on rcs0
May 30 23:31:59 hystKodi kernel: [ 212.866937] i915 0000:00:10.0: Resetting rcs0 for hang on rcs0
May 30 23:32:01 hystKodi kernel: [ 214.850973] i915 0000:00:10.0: Resetting rcs0 for hang on rcs0
May 30 23:32:03 hystKodi kernel: [ 216.866919] i915 0000:00:10.0: Resetting rcs0 for hang on rcs0
May 30 23:32:05 hystKodi kernel: [ 218.850958] i915 0000:00:10.0: Resetting rcs0 for hang on rcs0
May 30 23:32:07 hystKodi kernel: [ 220.866897] i915 0000:00:10.0: Resetting rcs0 for hang on rcs0
May 30 23:32:09 hystKodi kernel: [ 222.850968] i915 0000:00:10.0: Resetting rcs0 for hang on rcs0
May 30 23:32:11 hystKodi kernel: [ 224.866970] i915 0000:00:10.0: Resetting rcs0 for hang on rcs0
May 30 23:32:13 hystKodi kernel: [ 226.850920] i915 0000:00:10.0: Resetting rcs0 for hang on rcs0
May 30 23:32:15 hystKodi kernel: [ 228.866923] i915 0000:00:10.0: Resetting rcs0 for hang on rcs0
May 30 23:32:17 hystKodi kernel: [ 230.850905] i915 0000:00:10.0: Resetting rcs0 for hang on rcs0
May 30 23:32:19 hystKodi kernel: [ 232.866958] i915 0000:00:10.0: Resetting rcs0 for hang on rcs0
May 30 23:32:21 hystKodi kernel: [ 234.850911] i915 0000:00:10.0: Resetting rcs0 for hang on rcs0
May 30 23:32:23 hystKodi kernel: [ 236.866940] i915 0000:00:10.0: Resetting rcs0 for hang on rcs0
May 30 23:32:25 hystKodi kernel: [ 238.850895] i915 0000:00:10.0: Resetting rcs0 for hang on rcs0
May 30 23:32:27 hystKodi kernel: [ 240.866943] i915 0000:00:1...

[jt (johannes-thraen)]

I looked at this again. The GPU does scale down sometimes. But still the power consumption has without any doubt degraded a lot between 51 and 53 minor versions. The difference is immediately seen by either by looking at powertop or the power consumption or expected battery time in freshly booted kernels respectively. 53 only last 2/3 of the time that 51 lasts.

[ekinox09 (ekinox09)]

I've tried Ubuntu 20.04 (5.4.0-33-generic) hoping it will be better. But it's not the case; i've encoutered the problem quickly, even without using Kodi for stressing the GPU. Here the problem froze completly the system; hard reboot is the only way out. Will come back to Ubuntu 18.04 and 5.3.0-53 generic.
For info, here is the syslog with the problem:

Jun 5 18:28:41 hystKodi3 systemd-udevd[343]: Worker [2067] terminated by signal 11 (SEGV)
Jun 5 18:28:44 hystKodi3 PackageKit: daemon quit
Jun 5 18:28:44 hystKodi3 systemd[1]: packagekit.service: Succeeded.
Jun 5 18:28:45 hystKodi3 kernel: [ 317.504521] i915 0000:00:10.0: GPU HANG: ecode 9:1:0x00000000, hang on rcs0
Jun 5 18:28:45 hystKodi3 kernel: [ 317.505529] i915 0000:00:10.0: Resetting rcs0 for hang on rcs0
Jun 5 18:28:48 hystKodi3 systemd-udevd[2100]: free(): invalid next size (fast)
Jun 5 18:28:48 hystKodi3 systemd-udevd[343]: Worker [2100] terminated by signal 6 (ABRT)
Jun 5 18:28:53 hystKodi3 kernel: [ 325.505416] i915 0000:00:10.0: Resetting rcs0 for hang on rcs0
Jun 5 18:30:01 hystKodi3 CRON[2110]: (root) CMD ([ -x /etc/init.d/anacron ] && if [ ! -d /run/systemd/system ]; then /usr/sbin/invoke-rc.d anacron start >/dev/null; fi

[ekinox09 (ekinox09)]

Hello, it seems that there is no more activity on this bug. But i still face i915 CPU hangs and resetting. How can i have it solved ? Should i create another bug ? Thanks for your help.

[Hans van den Bogert (hbogert)]

It was solved, yet I am now on kernel 5.4.0-43 and I am hitting this now again since a week when using Jetbrains products. Clearly a regression, how could this 've happened. The linux-image* package changelog is pretty much worthless to investigate this.

[Hans van den Bogert (hbogert)]

You still can't edit comments on LP :S?

I am running kernel 5.4.0-42 instead of the earlier mentioned 5.4.0-43

[Serhiy (xintx-ua)]

Confirming, only Blender is affected in my case. Kernel 5.4.0-42, Intel UHD Graphics 620 (rev 07).

[Vsevolod Velichko (torkvemada)]

Same story here on Ubuntu 20.10.
Kernel version 5.8.0-25-generic on Dell XPS 15 9570, with Intel UHD Graphics 630 (Mobile).

[Luis Vega (lrvz)]

I am seeing a very similar issue on my Lenovo T480s while using the Intel UHD Graphics 620 (rev 7).
A simple example is that YouTube videos in any browser (Chrome/Chromium/Firefox/etc.) while in the browser window play at 60fps but as soon as you use full screen it falls back to 30fps on the latest 5.4 kernels (-54,-53,-52). I did a quick test and booted a 4.15.0-123 kernel and that seemed to resolve the problem. Seems like a regression.

[philicorda (philicorda)]

This doesn't seem to be fixed with recent Ubuntu kernels.

I have constant GPU hangs too.

kernel: Linux luto 5.8.0-41-generic #46~20.04.1-Ubuntu SMP

[ 20.503897] i915 0000:00:02.0: vgaarb: deactivate vga console
[ 20.505641] i915 0000:00:02.0: vgaarb: changed VGA decodes: olddecodes=io+mem,decodes=io+mem:owns=io+mem
[ 20.525280] [drm] Initialized i915 1.6.0 20200515 for 0000:00:02.0 on minor 0
[ 20.570726] fbcon: i915drmfb (fb0) is primary device
[ 20.570858] i915 0000:00:02.0: fb0: i915drmfb frame buffer device
[ 2252.688165] unhandled error in i915_error_to_vmf_fault: -22
[ 2252.688324] WARNING: CPU: 1 PID: 1455 at drivers/gpu/drm/i915/gem/i915_gem_mman.c:211 i915_error_to_vmf_fault+0x7a/0x90 [i915]
[ 2252.688325] Modules linked in: nls_iso8859_1 snd_hda_codec_realtek snd_hda_codec_generic ledtrig_audio snd_hda_intel snd_intel_dspcfg snd_ice1712 snd_hda_codec snd_cs8427 snd_i2c snd_ice17xx_ak4xxx snd_ak4xxx_adda i915 snd_ac97_codec snd_hda_core snd_mpu401_uart snd_hwdep snd_seq_midi drm_kms_helper input_leds cec snd_seq_midi_event rc_core i2c_algo_bit snd_rawmidi fb_sys_fops snd_seq syscopyarea sysfillrect ac97_bus snd_pcm snd_seq_device snd_timer sysimgblt at24 snd soundcore serio_raw mac_hid sch_fq_codel w83627ehf hwmon_vid coretemp parport_pc ppdev lp drm parport ip_tables x_tables autofs4 uas usb_storage hid_generic usbhid hid gpio_ich i2c_i801 psmouse i2c_smbus pata_acpi lpc_ich r8169 realtek video
[ 2252.688398] RIP: 0010:i915_error_to_vmf_fault+0x7a/0x90 [i915]
[ 2252.688450] vm_fault_gtt+0x1d7/0x540 [i915]

[ilove my cat (ilovemycat-deactivatedaccount)]

I facing some sort of similar Error. I reported a Bug, but nobody answering till now. Don't now wahts going on now.

This was the first time i Report a bug, did I do something wrong?

Anyway, i facing this problem only when using vmware. The Problems appers inside the vm, i get some heavy screen disortion:

https://bugs.launchpad.net/ubuntu/+source/linux-signed-hwe-5.8/+bug/1914850

[Jeff Van Epps (lordbah)]

Seeing i915 GPU HANG in 5.11.0-7620-generic.

Aug 14 23:46:50 topper kernel: [88920.160585] i915 0000:00:02.0: [drm] Resetting rcs0 for preemption time out
Aug 14 23:46:50 topper kernel: [88920.161300] i915 0000:00:02.0: [drm] *ERROR* rcs0 reset request timed out: {request: 00000001, RESET_CTL: 00000001}
Aug 14 23:46:50 topper kernel: [88920.170516] i915 0000:00:02.0: [drm] GPU HANG: ecode 9:1:85dffffa, in chrome [59885]

Aug 16 19:38:07 topper kernel: [ 5901.552794] i915 0000:00:02.0: [drm] Resetting rcs0 for preemption time out
Aug 16 19:38:07 topper kernel: [ 5901.553541] i915 0000:00:02.0: [drm] *ERROR* rcs0 reset request timed out: {request: 00000001, RESET_CTL: 00000001}
Aug 16 19:38:07 topper kernel: [ 5901.560089] i915 0000:00:02.0: [drm] GPU HANG: ecode 9:1:85dffffa, in Xorg [2179]