FS#55562 - [linux] Graphics glitches out when laptop wakes from suspend

Attached to Project: Arch Linux
Opened by Harish (sitwano) - Saturday, 09 September 2017, 08:53 GMT
Last edited by Jan de Groot (JGC) - Friday, 31 May 2019, 06:43 GMT
Task Type Bug Report
Category Kernel
Status Closed
Assigned To Tobias Powalowski (tpowa)
Architecture x86_64
Severity High
Priority Normal
Reported Version
Due in Version Undecided
Due Date Undecided
Percent Complete 100%
Votes 1
Private No

Details

Description:

Graphics glitches out when laptop resumes from sleep/suspend. After timeout, the display goes blank as expected for power saving. After a few more minutes the laptop goes to suspend to ram as expected. However, soemtimes upon waking from suspend, the contents on the screen shakes violently on the monitor indicating something wrong with the graphics driver.

This seems to happen in the 4.12.x series kernel. I have not yet seen this happen on 4.11 series kernel and LTS-4.9 series kernel.

Relevant Laptop details:

Thinkpad X250
Architecture: x86_64
Model name: Intel(R) Core(TM) i5-5200U CPU @ 2.20GHz
Graphics: Intel Integrated HD 5500

Steps to reproduce:
- wait for laptop to dim screen and then go to suspend.
- Wake computer by opening lid or pressing power button.

Steps to temporarily fix:
- Force laptop to suspend by closing lid.
- Reopen laptop lid to stop glitching.

A bug report has been opened in bugzilla.freedesktop.org --> https://bugs.freedesktop.org/show_bug.cgi?id=102587

The following error is shown in dmesg and journal when the graphics glitch happens:
kernel: [drm:intel_cpu_fifo_underrun_irq_handler [i915]] *ERROR* CPU pipe A FIFO underrun

Please note that I have currently reverter to 4.9 LTS kernel to avoid this problem. If anyone is experiencing this bug, please post the relevent log messages as indicated by Elizabeth in the freedesktop bugzilla report.
This task depends upon

Closed by  Jan de Groot (JGC)
Friday, 31 May 2019, 06:43 GMT
Reason for closing:  Fixed
Additional comments about closing:  cannot reproduce + closed upstream
Comment by Harish (sitwano) - Saturday, 09 September 2017, 11:08 GMT
Oh and I forgot to mention I'm using Gnome DE 3.24.2.
Comment by loqs (loqs) - Friday, 15 September 2017, 20:08 GMT
Seems same issue or related to  FS#55629 
Comment by Benjamin (BrauerSuzuki) - Sunday, 17 September 2017, 08:02 GMT
I'm having exactly the same issue! Details: x86_64, i7-5500U, kernel 4.12.12-1, fluxbox (system up to date).
Comment by Harish (sitwano) - Sunday, 17 September 2017, 20:12 GMT
@Benjamin, glad I'm not the only one! Please follow or engage with the bug report here https://bugzilla.freedesktop.org/show_bug.cgi?id=102587
Please note that this seems to be unrelated to https://bugs.archlinux.org/task/55629 as turning the IOMMU off in grub does not resolve this issue.
Comment by Benjamin (BrauerSuzuki) - Wednesday, 20 September 2017, 13:56 GMT
Now I haven't seen the bug in some days. Maybe the upgrade to 4.12.12-1 fixed the problem (when I wrote my last comment, I forgot to reboot after the update). Can you confirm, Harish?
Comment by Harish (sitwano) - Wednesday, 20 September 2017, 15:28 GMT
Hi Benjamin, I switched to latest 4.12 kernel. I was on zen 4.13.2 kernel to see if issue was fixed, but it turns out that its not. Will report back my experience after a few days with 4.12.13 kernel to eliminate possible regression. This issue isn't really predictable, but it seems to happens after displays blanks out (but not always) suggesting some power-saving issue. I also removed all kernel parameters such as IOMMU.
Comment by Harish (sitwano) - Wednesday, 20 September 2017, 20:14 GMT
i915.enable_psr=0 as a kernel parameter might help since it causes problems, according to arch wiki. 4.9.x kernel appears to turn it off by default, and it may be that some kernel version after that turned it back on as default again. I have currently added it as grub kernel parameter but I am yet to test this out and verify.

check out arch wiki page https://wiki.archlinux.org/index.php/intel_graphics#Screen_flickering for possible fixes! I also included it in the original bug report for the devs at freedesktop.org

Edit: it seems that psr is already disabled by default since kernel 4.9 according to https://bugs.freedesktop.org/show_bug.cgi?id=101111 so might as well try disable rc6 instead.
Comment by Benjamin (BrauerSuzuki) - Saturday, 23 September 2017, 19:33 GMT
I was wrong. It happend again with 4.12.12-1. Thanks for the link! I will try it soon.
Comment by Benjamin (BrauerSuzuki) - Tuesday, 03 October 2017, 09:22 GMT
with rc6 disabled my fan doesn't stop spinning which is even more annoying... :-(
Comment by Harish (sitwano) - Tuesday, 03 October 2017, 15:09 GMT
But has it stopped the graphics glitch though? ;) If you want peace of mind just shift to LTS kernel with hibernation patch (if you use hibernation). The devs seem to have begun working on this bug over at bugzilla.freedesktop.org
Comment by Benjamin (BrauerSuzuki) - Monday, 09 October 2017, 21:04 GMT
I was not patient enough to see if rc6 would have helped. Now I can report that kernel 4.13.4 is still affected. I have finally downgraded to 4.11.9. This was the latest "safe" kernel I have cached (at least I hope so).
Comment by Harish (sitwano) - Sunday, 22 October 2017, 16:16 GMT
I can also confirm that that 4.11.9 is the last kernel that does not exhibit this problem. I strongly suspect that kernel 4.12 introduced this regression.
Comment by Benjamin (BrauerSuzuki) - Wednesday, 07 March 2018, 10:25 GMT
Has it been fixed in the latest kernel? I like to know before I update.

Loading...