Bug 103287

Summary: [CI] - igt@* - dmesg-warn - *ERROR* Link Training Unsuccessful
Product: DRI Reporter: Marta Löfstedt <marta.lofstedt>
Component: DRM/IntelAssignee: shashank.sharma <shashank.sharma>
Status: CLOSED WORKSFORME QA Contact: Intel GFX Bugs mailing list <intel-gfx-bugs>
Severity: critical    
Priority: highest CC: intel-gfx-bugs
Version: DRI git   
Hardware: Other   
OS: All   
Whiteboard: ReadyForDev
i915 platform: KBL i915 features: display/Other

Description Marta Löfstedt 2017-10-16 06:39:24 UTC
CI_DRM_3229 KBL-shards igt@gem_caching@write 	

[  224.177526] [drm:intel_dp_get_link_train_fallback_values [i915]] *ERROR* Link Training Unsuccessful
[  224.734245] [drm:intel_dp_get_link_train_fallback_values [i915]] *ERROR* Link Training Unsuccessful

NOTE: LSPCON fix see BUG 102295 and fix for linktraining BUG 102359 was integrated to CI_DRM_3228. I file a new bug for above linktraining dmesg warning, to avoid confusion with the previous bugs.  

Comment 1 Marta Löfstedt 2017-10-16 06:43:03 UTC


[  283.838139] [drm:intel_dp_get_link_train_fallback_values [i915]] *ERROR* Link Training Unsuccessful
[  284.271625] [drm] Atomic update on pipe (A) took 284 us, max time under evasion is 250 us
[  284.701592] [drm] Atomic update on pipe (A) took 306 us, max time under evasion is 250 us
[  285.263078] [drm:intel_dp_get_link_train_fallback_values [i915]] *ERROR* Link Training Unsuccessful
[  285.569340] [drm:intel_dp_get_link_train_fallback_values [i915]] *ERROR* Link Training Unsuccessful
[  286.139792] [drm:intel_dp_get_link_train_fallback_values [i915]] *ERROR* Link Training Unsuccessful
[  286.411885] [drm] Atomic update on pipe (B) took 307 us, max time under evasion is 250 us
[  286.849299] [drm:intel_dp_get_link_train_fallback_values [i915]] *ERROR* Link Training Unsuccessful
[  287.159871] [drm:intel_dp_get_link_train_fallback_values [i915]] *ERROR* Link Training Unsuccessful
[  287.452562] [drm] Atomic update on pipe (C) took 278 us, max time under evasion is 250 us
[  287.729101] [drm:intel_dp_get_link_train_fallback_values [i915]] *ERROR* Link Training Unsuccessful
[  288.644852] [drm:intel_dp_get_link_train_fallback_values [i915]] *ERROR* Link Training Unsuccessful
Comment 13 Marta Löfstedt 2017-10-18 12:17:48 UTC
Here is maybe an interesting observation.

This show all the times we have hit the issue:

for CI_DRM_3256 all occurrences are from shards-kbl2.
I am not saying that there is something particularly wrong with that machine. Which is clearly visible by looking at other runs.
Comment 14 Marta Löfstedt 2017-11-01 07:45:42 UTC
There has been no linktraining issues on the KBL-shards since CI_DRM_3291 until CI_DRM_3304.

So, it appear as if: 

d0582552491e drm-tip: 2017y-10m-27d-19h-16m-21s UTC integration manifest
be33be5df2a0 drm/i915: Clean up the mess around hdmi_12bpc_possible()
857c416e58dc drm/i915: Improve DP downstream HPD handling
5161d058dff4 drm/i915: Fix BXT lane latency optimal setting with MST
742745f1ee4f drm/i915: Stop using encoder->type in intel_ddi_enable_transcoder_func()
f49b44ab8403 drm/i915: Start using output_types for DPLL selection
3a6d84e677d4 drm/i915: Pass crtc state to intel_prepare_dp_ddi_buffers()
5448f53ffe62 drm/i915: Don't use encoder->type in intel_ddi_set_pipe_settings()
4efa63085536 drm-tip: 2017y-10m-27d-15h-25m-04s UTC integration manifest

unintentionally fixed something.
IGT-Version: 1.20-g1fc4de1c for both CI_DRM_3291 and CI_DRM_3292.

I will close this bug.

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.