| Summary: | Screen flickering till it goes blank (console login, no X) | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| Product: | DRI | Reporter: | Sten Heinze <sten.heinze> | ||||||
| Component: | DRM/Intel | Assignee: | Intel GFX Bugs mailing list <intel-gfx-bugs> | ||||||
| Status: | RESOLVED DUPLICATE | QA Contact: | Intel GFX Bugs mailing list <intel-gfx-bugs> | ||||||
| Severity: | normal | ||||||||
| Priority: | high | CC: | intel-gfx-bugs, russianneuromancer | ||||||
| Version: | DRI git | ||||||||
| Hardware: | x86-64 (AMD64) | ||||||||
| OS: | Linux (All) | ||||||||
| Whiteboard: | Triaged, ReadyForDev | ||||||||
| i915 platform: | SKL | i915 features: | display/Other | ||||||
| Bug Depends on: | |||||||||
| Bug Blocks: | 105908 | ||||||||
| Attachments: |
|
||||||||
|
Description
Sten Heinze
2018-09-24 19:07:57 UTC
Sten, With latest drmtip, can you attach full dmesg from boot with kernel parameters drm.debug=0x1e log_buf_len=4M. Latest drm-tip can be found from https://cgit.freedesktop.org/drm-tip How much impact this issue has for you? Does it happen every time when you reboot? This is important for us to prioritize bugs. (In reply to Lakshmi from comment #1) > Sten, With latest drmtip, can you attach full dmesg from boot with kernel > parameters drm.debug=0x1e log_buf_len=4M. I will do that when I get back to my laptop tomorrow. (In reply to Lakshmi from comment #2) > Latest drm-tip can be found from https://cgit.freedesktop.org/drm-tip That's what I cloned and compiled this past weekend. Two other important pieces of information are that 1) Adding nomodeset (at kernel boot/grub), the flickering is gone; 2) Using the 4.16.16-2 kernel from Debian, the flickering is not occurring (even in absence of nomodeset). (In reply to Lakshmi from comment #3) > How much impact this issue has for you? > Does it happen every time when you reboot? > This is important for us to prioritize bugs. It does happen on every boot, and within minutes the console screen is blank. It makes the laptop pretty much unusable. Created attachment 141744 [details]
dmesg
Over the last couple weeks I tried to bisect this problem (unfortunately a Core M is quite slow in compiling), but I could not reproduce the bug with any of the 14 kernels I compiled. In looking at every detail again, I noticed that I used Debian's 4.16 config for all kernels except the 4.17 to reproduce the bug. So I started the bisect again and compiled the first bisecting step twice, once with Debian's 4.16 and once with Debian's 4.17 config as starting point. (Then I removed Debian's signing key, and ran make olddefconfig.) And the different config causes the bug. I'll attach the diff. I'm not aware of an automated method to bisect configs. The most suspect config setting to me is CONFIG_SATA_MOBILE_LPM_POLICY, but I'm just guessing. I'd be happy to hear suggestions of how to make this more automated, and/or which config setting to prioritize in testing. (Also, even with a working kernel as far as this bug goes, I see "cpu pipe a fifo underrun" messages and flicker on shutdown - but maybe this should go in a different bug report?) Created attachment 142037 [details] [review] Diff in .config for 4.16+ kernel based on Debian's 4.16 and 4.17 kernels Setting CONFIG_SATA_MOBILE_LPM_POLICY=0 is not fixing the issue. It's looks similar to Bug 103229 - could you try this https://bugs.freedesktop.org/show_bug.cgi?id=103229#c26 workaround? (In reply to russianneuromancer from comment #9) > It's looks similar to Bug 103229 - could you try this > https://bugs.freedesktop.org/show_bug.cgi?id=103229#c26 workaround? I'll try when I get home. > https://bugs.freedesktop.org/show_bug.cgi?id=103229#c26 workaround?
With intel_idle.max_cstate=4, I see no flickering and the screen does not go blank.
So possibly your report if duplicate of bug 103229? Considering this as FIFO underrun issue on SKL. Closing this as a duplicate to Bug 103229. *** This bug has been marked as a duplicate of bug 103229 *** |
Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.