i915 - after return from suspend display is flickering / *ERROR* CPU pipe A FIFO underrun

[arno]

Test this with Ubuntu 20.04 (Mate) beta installer. Also happen with current official version.
Kernel is 5.4.0-21 #25 (if I remember right).
All works fine, but after returning from suspend the display is flickering.
For the installer / USB stick usage that might not be a problem, but the same behaviour in a normal environment would be a no go (for using suspend)

Graphics is driven by an Intel Skylake core-m5 processor.
---
ProblemType: Bug
ApportVersion: 2.20.11-0ubuntu27
Architecture: amd64
AudioDevicesInUse:
 USER PID ACCESS COMMAND
 /dev/snd/controlC0: user 1124 F.... pulseaudio
CasperMD5CheckResult: skip
CurrentDesktop: MATE
DistroRelease: Ubuntu 20.04
InstallationDate: Installed on 2020-04-17 (0 days ago)
InstallationMedia: Ubuntu-MATE 20.04 LTS "Focal Fossa" - Beta amd64 (20200402)
Lsusb:
 Bus 002 Device 001: ID 1d6b:0003 Linux Foundation 3.0 root hub
 Bus 001 Device 003: ID 8087:0a2b Intel Corp.
 Bus 001 Device 002: ID 04f2:b560 Chicony Electronics Co., Ltd HP HD Camera
 Bus 001 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub
MachineType: HP HP EliteBook Folio G1
Package: linux (not installed)
ProcFB: 0 i915drmfb
ProcKernelCmdLine: BOOT_IMAGE=/vmlinuz-5.4.0-24-generic root=UUID=8fb86658-3799-4204-94df-7cf37441baa3 ro quiet splash vt.handoff=7
ProcVersionSignature: Ubuntu 5.4.0-24.28-generic 5.4.30
RelatedPackageVersions:
 linux-restricted-modules-5.4.0-24-generic N/A
 linux-backports-modules-5.4.0-24-generic N/A
 linux-firmware 1.187
Tags: focal
Uname: Linux 5.4.0-24-generic x86_64
UpgradeStatus: No upgrade log present (probably fresh install)
UserGroups: adm cdrom dip lpadmin lxd plugdev sambashare sudo
_MarkForUpload: True
dmi.bios.date: 01/05/2020
dmi.bios.vendor: HP
dmi.bios.version: N91 Ver. 01.45
dmi.board.name: 8170
dmi.board.vendor: HP
dmi.board.version: KBC Version 29.73
dmi.chassis.type: 10
dmi.chassis.vendor: HP
dmi.modalias: dmi:bvnHP:bvrN91Ver.01.45:bd01/05/2020:svnHP:pnHPEliteBookFolioG1:pvr:rvnHP:rn8170:rvrKBCVersion29.73:cvnHP:ct10:cvr:
dmi.product.family: 103C_5336AN
dmi.product.name: HP EliteBook Folio G1
dmi.product.sku: V1C39EA#ABD
dmi.sys.vendor: HP
---
ProblemType: Bug
ApportVersion: 2.20.11-0ubuntu27
Architecture: amd64
AudioDevicesInUse:
 USER PID ACCESS COMMAND
 /dev/snd/controlC0: user 1124 F.... pulseaudio
CasperMD5CheckResult: skip
CurrentDesktop: MATE
DistroRelease: Ubuntu 20.04
InstallationDate: Installed on 2020-04-17 (0 days ago)
InstallationMedia: Ubuntu-MATE 20.04 LTS "Focal Fossa" - Beta amd64 (20200402)
Lsusb:
 Bus 002 Device 001: ID 1d6b:0003 Linux Foundation 3.0 root hub
 Bus 001 Device 003: ID 8087:0a2b Intel Corp.
 Bus 001 Device 002: ID 04f2:b560 Chicony Electronics Co., Ltd HP HD Camera
 Bus 001 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub
MachineType: HP HP EliteBook Folio G1
Package: linux (not installed)
ProcFB: 0 i915drmfb
ProcKernelCmdLine: BOOT_IMAGE=/vmlinuz-5.4.0-24-generic root=UUID=8fb86658-3799-4204-94df-7cf37441baa3 ro quiet splash vt.handoff=7
ProcVersionSignature: Ubuntu 5.4.0-24.28-generic 5.4.30
RelatedPackageVersions:
 linux-restricted-modules-5.4.0-24-generic N/A
 linux-backports-modules-5.4.0-24-generic N/A
 linux-firmware 1.187
Tags: focal
Uname: Linux 5.4.0-24-generic x86_64
UpgradeStatus: No upgrade log present (probably fresh install)
UserGroups: adm cdrom dip lpadmin lxd plugdev sambashare sudo
_MarkForUpload: True
dmi.bios.date: 01/05/2020
dmi.bios.vendor: HP
dmi.bios.version: N91 Ver. 01.45
dmi.board.name: 8170
dmi.board.vendor: HP
dmi.board.version: KBC Version 29.73
dmi.chassis.type: 10
dmi.chassis.vendor: HP
dmi.modalias: dmi:bvnHP:bvrN91Ver.01.45:bd01/05/2020:svnHP:pnHPEliteBookFolioG1:pvr:rvnHP:rn8170:rvrKBCVersion29.73:cvnHP:ct10:cvr:
dmi.product.family: 103C_5336AN
dmi.product.name: HP EliteBook Folio G1
dmi.product.sku: V1C39EA#ABD
dmi.sys.vendor: HP

[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 1872760

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.

[arno (star-gmx)]

apport information

[arno (star-gmx)]

apport information

[arno (star-gmx)]

apport information

[arno (star-gmx)]

apport information

[arno (star-gmx)]

apport information

[arno (star-gmx)]

apport information

[arno (star-gmx)]

apport information

[arno (star-gmx)]

apport information

[arno (star-gmx)]

apport information

[arno (star-gmx)]

apport information

[arno (star-gmx)]

apport information

[arno (star-gmx)]

apport information

[arno (star-gmx)]

apport information

[arno (star-gmx)]

apport information

[arno (star-gmx)]

apport information

[arno (star-gmx)]

apport information

[arno (star-gmx)]

apport information

[arno (star-gmx)]

apport information

[arno (star-gmx)]

apport information

[arno (star-gmx)]

apport information

[arno (star-gmx)]

apport information

[arno (star-gmx)]

apport information

[arno (star-gmx)]

apport information

[arno (star-gmx)]

apport information

[arno (star-gmx)]

apport information

[arno (star-gmx)]

apport information

[arno (star-gmx)]

apport information

[arno (star-gmx)]

apport information

[arno (star-gmx)]

apport information

[arno (star-gmx)]

apport information

[arno (star-gmx)]

apport information

[arno (star-gmx)]

apport information

[arno (star-gmx)]

apport information

[arno (star-gmx)]

apport information

[arno (star-gmx)]

apport-collect returns message that dpkg-query doesn't find a packet matching linux

[arno (star-gmx)]

Seems I can't change status to complete ... !?

[Seth Forshee (sforshee)]

The attached logs are from the current boot, which doesn't show anything from when the corruption actually happened. Can you get logs from a previous boot where you did see the corruption? For example, if you know this happened the last time you booted, you can use 'journalctl -b -1', or 'jornalctl -b -2' for two boots ago, etc.

[arno (star-gmx)]

Mh, I bootet (all looks fine), go to standby, recover from standby - screen shutters. Than I did the apport-collect. That is wrong timing? What do you expect to see in logs, so that I can take care?

[Seth Forshee (sforshee)]

Oh, no that is fine then. Sorry, on the other bug you said the system was unusable after suspend so I assumed that you could not run apport-collect in that state.

Unfortunately the logs don't show anything strange happening.

[Seth Forshee (sforshee)]

Actually there is something in the logs that I missed the first time:

[ 69.108687] [drm:intel_cpu_fifo_underrun_irq_handler [i915]] *ERROR* CPU pipe A FIFO underrun

Do you know whether earlier 5.4 kernels in focal also had this same issue? If not, do you know the last version you used which did not flicker after suspend?

[arno (star-gmx)]

This is first attempt with focal/5.4 kernel, the beta. Not sure but I think the installer has an even earlier kernel version but the same problem.
Before I was fighting with ubuntu 19.10 (see
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1863489
https://bugs.launchpad.net/bugs/1853044)
These kernel has powerdrain, but not this kind of issue. So to answer your question: 5.3.0-23 has no flicker.
I didn't use that laptop for a while, so I did not tried that fix. Instead I just jumped to the 20.04 beta. Hope that helps.

[Seth Forshee (sforshee)]

Can you try a vanilla 5.4.30 kernel from here, and see if it has the same problem? That will let us know if any of the fixes we've applied on top of 5.4 is responsible.

https://kernel.ubuntu.com/~kernel-ppa/mainline/v5.4.30/

Note that you will need to disable secure boot to boot into this kernel. Thanks!

[arno (star-gmx)]

Not sure that I know how to install that kernel.
I checked the previous
5.4.0-21-generic #25
kernel, but that flickers too after resuming from suspend.

[arno (star-gmx)]

As far as I can see there is no build for amd64 (failed) for the 5.4.30 mainline.

[arno (star-gmx)]

I checked for the latest update (5.3.0-25-generic) and (as there is no 5.4 build) the latest 5.5.18 mainline kernel.
Both flicker after suspend!

[arno (star-gmx)]

Why status is incomplete - I did everything that has been requested...

[Seth Forshee (sforshee)]

Because someone has to manually flip it once information has been provided. Setting it to confirmed, buy you should also be able to do that for your own bug.

[arno (star-gmx)]

Thanks Seth, but (don't know why) this status "confirmed" was not an option to select for me. That is something that can just another one than the reporter.
Do you think it deals with re-introduction of power saving ?
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1853044

Here I have reported the same and I tested that with same machine, no flickering at that moment.
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1863489

[Sjoerd Nijboer (cobalt-ethon)]

There seems to be a known issue of i915 related screen flickering with relation to power states, refresh rates and eDP for intel i915 in recent kernels. (If i understand the issue correctly)

Can you confirm that the flickering is thesame as the videos in the description of: https://gitlab.freedesktop.org/drm/intel/-/issues/272 ?

[arno (star-gmx)]

It seems to be more in the video. BUT - screen is ok BEFORE suspend.
It happens here only after suspend or - what I observed yesterday - while calibrating powertop.

[Sjoerd Nijboer (cobalt-ethon)]

So it looks thesame (maybe a little less) but happens under different circumstances?
I'm experiencing a similar issue and maybe got a temporary fix.
All I know is that with my laptop (Lenovo Yoga C940 14IL) i've got thesame flickering as in that issue on Kubuntu 19.10, Kubuntu 20.04 and Fedora KDE 31.
Those are all 5.3.x and 5.4.x kernels if I got my kernel numbers right.
I've tried upgrading Kubuntu 20.04 to mainline kernel 5.6.0-050600 and that didn't work.
Upgrading to 5.7-rc2 seems to work, but network card doesn't work out of the box.

I've tried this issue with OpenGL 2.0, 3.1 and XRender with various settings for vsync, that does have major effect to occurrence, and that's why I think that how often it occurs isn't that important.

Could you try to install 5.6.0-050600 and 5.7-rc2 to confirm that the underlying issue is thesame and fixed in 5.7?

[arno (star-gmx)]

Mh reading the report in your mentioned bug I see that bug should be fixed already in 5.4.0-42 and 5.4.x
This is not what I observe. The newest I have been testing was 5.5.18 mainline from ubuntu.
Also I didn't have this issue in older 19.10/18.04 - but there I did have a much higher power consuption due to missing rc6 states. Do they fixed that by reverting this power saving? Hope that is not a "either or" thing.
Before doing a test let me do a backup first ....

[arno (star-gmx)]

Things get worse with 5.7.rc2.
It is not just a missed module, also Display is flickering right from start (not onyl after suspend). No - this is not the way to go .... Sorry

[arno (star-gmx)]

Help please, that flickering drives me crazy ....

[Seth Forshee (sforshee)]

I think the upstream i915 developers might be able to best help you, since the issue still exists in 5.7. I checked and this report looks to be the same problem:

https://gitlab.freedesktop.org/drm/intel/-/issues/347

I'd suggest trying to help the developers debug the problem there.

Looks like intel_idle.max_cstate=4 might be a workaround, of course then you are back to higher power consumption.

[arno (star-gmx)]

I can confirm that the workaround helps, but I need the power saving states too.

[arno (star-gmx)]

The bug in that mentioned list is also old (2month). Can you recommend a list where this upstream things are handled best... there seems to be so many (confused).

[Seth Forshee (sforshee)]

The mailing list for i915 is <email address hidden>.

[arno (star-gmx)]

As far as I see : no response from mailing-list. :( So all the effort here is wasted?

[arno (star-gmx)]

As for me there are 2 ways to get rid of flickering, but both are not make me happy:
Either in grub add:
- GRUB_CMDLINE_LINUX="intel_idle.max_cstate=4"
or (that is new to me)
- change screen resolution(!?!) Anything else than 1920x1080 works! If I can add custom resolutions I would play with that. But only 1600x900 matches my ration.

Maybe this 2nd workaround gives the developers a clue what is the original issue here. Hope to get some feedback. Maybe this is the reason there are not to many that complains about that issue, although nowaday this resolution is standard.

Beside that I have testet a lot of other things, kernel options as
GRUB_CMDLINE_LINUX="i915.enable_psr=0"
or in i915.conf
options i915 enable_guc=2
options i915 fastboot=1

[arno (star-gmx)]

The error message from Kernel about this issue is:

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

[Alejandro Santoyo Gonzalez (al3jandrosg)]

The same issue on my Lenovo T14s, frequently hanging after suspend. I'm using Ubuntu 20.04, with HWE kernel version 5.8.0-40-generic. I have two external monitors with 1980x1080 res, connected through a 2xHDMI to USB-C connector using DisplayLink.

This is really annoying! Are there any efforts being made to provide definitive root cause analysis and ship the solution into an upcoming kernel version?

[arno (star-gmx)]

I would be great to see any progress in it.
Just tested Manjaro with newer 5.9 kernel, but the issue keeps same.

[arno (star-gmx)]

Would like to push it again.