| Summary: | [KBL] [drm:intel_cpu_fifo_underrun_irq_handler] *ERROR* CPU pipe A FIFO underrun | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| Product: | DRI | Reporter: | Andrey <ryabinin.a.a> | ||||||||
| Component: | DRM/Intel | Assignee: | Intel GFX Bugs mailing list <intel-gfx-bugs> | ||||||||
| Status: | RESOLVED FIXED | QA Contact: | Intel GFX Bugs mailing list <intel-gfx-bugs> | ||||||||
| Severity: | normal | ||||||||||
| Priority: | medium | CC: | intel-gfx-bugs, nick.desaulniers | ||||||||
| Version: | unspecified | ||||||||||
| Hardware: | x86-64 (AMD64) | ||||||||||
| OS: | Linux (All) | ||||||||||
| Whiteboard: | |||||||||||
| i915 platform: | KBL | i915 features: | |||||||||
| Attachments: |
|
||||||||||
I see this as well even when booting a drm-tip tree kernel. 2016 Razer Blade Stealth (Kaby Lake) Ubuntu 16.10. Created attachment 129056 [details]
dmesg drm-tip
Created attachment 129057 [details]
an earlier dmesg with more info
If I add the kernel command line `i915.enable_rc6=0` this works around the issue. Note: this should still be fixed, a release candidate should not have this kind of bug. Oh, looks like rc6 stands for "Render C state 6": http://lxr.free-electrons.com/source/drivers/gpu/drm/i915/i915_params.c#L82 I just booted a drm-tip kernel with the lastest commit of: commit 4dbd7c0fbb78579ff491ef1184f78087055c5aa5 Author: Chris Wilson <chris@chris-wilson.co.uk> Date: Fri Feb 10 21:45:49 2017 +0000 drm-tip: 2017y-02m-10d-21h-45m-14s UTC integration manifest and removed the kernel command line parameter (i915.enable_rc6=0) and no longer seeing flicker at native resolutions. I assume that tree will be pulled from for the 4.10 kernel. Hello Andrey, please note the comment 7 and retest with the latest kernel (preferable drm-tip from git://anongit.freedesktop.org/git/drm-tip). If problem still persist, please mark status REOPENED, otherwise RESOLVED. Nick - can you also check this with the latest drm-tip? Reporters, please comment if issue is solved or not. Sory for delayed response. The issue has been fixed for me. Thanks for the feedback. Closing. Sorry to reopen, but I'm still seeing this issue. What can I provide to help debug it? Hi, Can you open new bnug not opening really old bug. Have you tested with latest drm-tip that is now 5.2rc4? Sure. Let's track in https://bugs.freedesktop.org/show_bug.cgi?id=108860. Note that there seems to be many bugs related to `drm:intel_cpu_fifo_underrun_irq_handler`. It might be good for someone to triage/dedup and better understand how pervasive the issue is. I'm doing build right now of drm-tip; will report in https://bugs.freedesktop.org/show_bug.cgi?id=108860. |
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.
Created attachment 128853 [details] dmesg This occasionally happens on resume, after display has been suspended. I.e. I can trigger this via following commands: xset dpms force suspend && xset dpms force on Sometimes screen remains black after that. There are two errors in dmesg: [ 850.328129] [drm:skl_set_cdclk] *ERROR* failed to inform PCU about cdclk change [ 850.533486] [drm:intel_cpu_fifo_underrun_irq_handler] *ERROR* CPU pipe A FIFO underrun Sometimes I see only the first error, but in that case display works, it doesn't remain black. So this might be a two separate issues, I'm going to open another bug for the first error. dmesg with drm.debug=1e attached. Kernel version - 4.9.2 Hardware - Dell XPS 13 9360