Bug 102997 - [BAT][CTG only] igt@* - Incomplete
Summary: [BAT][CTG only] igt@* - Incomplete
Alias: None
Product: DRI
Classification: Unclassified
Component: DRM/Intel (show other bugs)
Version: DRI git
Hardware: Other All
: lowest normal
Assignee: Intel GFX Bugs mailing list
QA Contact: Intel GFX Bugs mailing list
Whiteboard: ReadyForDev
: 103061 (view as bug list)
Depends on:
Blocks: 105984
  Show dependency treegraph
Reported: 2017-09-26 10:10 UTC by Marta Löfstedt
Modified: 2019-11-29 17:26 UTC (History)
1 user (show)

See Also:
i915 platform: GM45
i915 features: display/Other


Description Marta Löfstedt 2017-09-26 10:10:17 UTC
Last dmesg:
[  476.001352] [IGT] kms_pipe_crc_basic: starting subtest nonblocking-crc-pipe-A-frame-sequence
[  476.001519] [drm:drm_mode_addfb2] [FB:76]
[  476.007618] [drm:drm_mode_setcrtc] [CRTC:36:pipe A]
[  476.007644] [drm:drm_mode_setcrtc] [CONNECTOR:47:LVDS-1]
[  486.368118] [drm:drm_atomic_helper_wait_for_dependencies] *ERROR* [CRTC:36:pipe A] flip_done timed out
[  496.608114] [drm:drm_atomic_helper_wait_for_dependencies] *ERROR* [PLANE:27:primary A] flip_done timed out
[  506.848132] [drm:drm_atomic_helper_wait_for_flip_done] *ERROR* [CRTC:36:pipe A] flip_done timed out
[  506.848350] [drm:drm_mode_setcrtc] [CRTC:46:pipe B]
[  509.853442] [IGT] kms_pipe_crc_basic: exiting, ret=99
[  509.853878] [drm:status_store] [CONNECTOR:54:VGA-1] force updated from 2 to 0 or reprobing
[  509.853896] [drm:drm_helper_probe_single_connector_modes] [CONNECTOR:54:VGA-1]
[  509.853961] [drm:intel_crt_detect [i915]] [CONNECTOR:54:VGA-1] force=1
[  509.863527] [drm:intel_crt_detect [i915]] CRT not detected via hotplug
[  509.865643] [drm:gmbus_xfer [i915]] GMBUS [i915 gmbus vga] NAK for addr: 0050 w(1)
[  509.865712] [drm:gmbus_xfer [i915]] GMBUS [i915 gmbus vga] NAK on first message, retry
[  509.867876] [drm:gmbus_xfer [i915]] GMBUS [i915 gmbus vga] NAK for addr: 0050 w(1)
[  509.867899] [drm:drm_do_probe_ddc_edid] drm: skipping non-existent adapter i915 gmbus vga
[  509.867966] [drm:intel_crt_get_edid [i915]] CRT GMBUS EDID read failed, retry using GPIO bit-banging
[  509.868031] [drm:intel_gmbus_force_bit [i915]] enabling bit-banging on i915 gmbus vga. force bit now 1
[  509.868686] [drm:drm_do_probe_ddc_edid] drm: skipping non-existent adapter i915 gmbus vga
[  509.868733] [drm:intel_gmbus_force_bit [i915]] disabling bit-banging on i915 gmbus vga. force bit now 0
[  509.868782] [drm:intel_crt_detect_ddc [i915]] CRT not detected via DDC:0x50 [no valid EDID found]
[  509.868797] [drm:drm_helper_probe_single_connector_modes] [CONNECTOR:54:VGA-1] status updated from connected to disconnected
[  509.868813] [drm:drm_helper_probe_single_connector_modes] [CONNECTOR:54:VGA-1] disconnected

Comment 1 Marta Löfstedt 2017-09-26 10:11:12 UTC
Probably related to BUG 102996.
Comment 2 Chris Wilson 2017-09-26 10:45:27 UTC
Note the end of the dmesg shows signs of fs corruption; an uncaptured panic?
Comment 4 Marta Löfstedt 2017-10-03 12:52:08 UTC
Now also incomplete on:


This also has the: "[drm:drm_atomic_helper_wait_for_flip_done] *ERROR* [CRTC:36:pipe A] flip_done timed out" print I have seen this pattern on some KBL incompletes, but there pstore is working and we will see softdog triggered.

<3>[  516.576177] [drm:drm_atomic_helper_wait_for_flip_done] *ERROR* [CRTC:36:pipe A] flip_done timed out
<7>[  516.576265] [drm:_intel_set_memory_cxsr [i915]] memory self-refresh is enabled (was disabled)
<7>[  516.576325] [drm:verify_connector_state.isra.72 [i915]] [CONNECTOR:54:VGA-1]
<7>[  516.576370] [drm:intel_atomic_commit_tail [i915]] [CRTC:36:pipe A]
Comment 5 Marta Löfstedt 2017-10-03 12:53:56 UTC
*** Bug 103061 has been marked as a duplicate of this bug. ***
Comment 6 Marta Löfstedt 2017-10-04 09:41:06 UTC
reproduced again:

last dmesg:
<7>[  468.433621] [drm:intel_panel_actually_set_backlight [i915]] set backlight PWM = 453348
<3>[  478.688171] [drm:drm_atomic_helper_wait_for_flip_done] *ERROR* [CRTC:46:pipe B] flip_done timed out
<7>[  478.688263] [drm:_intel_set_memory_cxsr [i915]] memory self-refresh is enabled (was disabled)
<7>[  478.688313] [drm:verify_connector_state.isra.72 [i915]] [CONNECTOR:47:LVDS-1]
<7>[  478.688359] [drm:intel_atomic_commit_tail [i915]] [CRTC:46:pipe B]
Comment 7 Marta Löfstedt 2017-10-05 07:08:25 UTC
reproduced again with "flip_done timed out" last in dmesg.

<7>[  455.639593] [drm:intel_panel_actually_set_backlight [i915]] set backlight PWM = 453348
<3>[  465.888183] [drm:drm_atomic_helper_wait_for_flip_done] *ERROR* [CRTC:46:pipe B] flip_done timed out
<7>[  465.888273] [drm:_intel_set_memory_cxsr [i915]] memory self-refresh is enabled (was disabled)
<7>[  465.888339] [drm:verify_connector_state.isra.72 [i915]] [CONNECTOR:47:LVDS-1]
<7>[  465.888396] [drm:intel_atomic_commit_tail [i915]] [CRTC:46:pipe B]
Comment 9 Marta Löfstedt 2017-10-19 08:34:19 UTC
(In reply to Marta Löfstedt from comment #8)
> starting at CI_DRM_3244 there is continous incompletes on
> igt@gem_exec_suspend@basic-s3
> https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_3244/fi-ctg-p8600/
> igt@gem_exec_suspend@basic-s3.html
> https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_3257/fi-ctg-p8600/
> igt@gem_exec_suspend@basic-s3.html

This trend is continuing still solid line of incompletes up to CI_DRM_3261.
Theory is that it started when we wnet up to 4.14.0-rc5+,
so someone should do a bisect.
Comment 10 Elizabeth 2017-11-14 20:06:19 UTC
Rising priority since it is BAT.
Comment 11 Marta Löfstedt 2017-11-16 06:38:41 UTC
The incompletes on igt@gem_exec_suspend@basic-s3 has continues since 4.14.0-rc5+. However, from CI_DRM_3350 we added kernel config CONFIG_ORC_UNWINDER to improve stacktraces and from that run CTG no longer incompletes on the igt@gem_exec_suspend@basic-s3. on 2 runs it has incompletes on  igt@kms_pipe_crc_basic@suspend-read-crc-pipe-b instead, but the coverage is significantly improved.
Comment 12 Marta Löfstedt 2017-12-07 07:23:52 UTC

<3>[  453.087273] [drm:drm_atomic_helper_wait_for_dependencies] *ERROR* [CRTC:47:pipe B] flip_done timed out
<3>[  463.327266] [drm:drm_atomic_helper_wait_for_dependencies] *ERROR* [PLANE:38:primary B] flip_done timed out
<7>[  463.327420] [drm:_intel_set_memory_cxsr [i915]] memory self-refresh is disabled (was disabled)
<3>[  473.567271] [drm:drm_atomic_helper_wait_for_flip_done] *ERROR* [CRTC:47:pipe B] flip_done timed out
<7>[  473.567416] [drm:_intel_set_memory_cxsr [i915]] memory self-refresh is enabled (was disabled)

running: igt/kms_pipe_crc_basic/read-crc-pipe-b-frame-sequence

[240/288] skip: 50, pass: 190 |                               
FATAL: command execution failed
Completed CI_IGT_test CI_DRM_3467/fi-ctg-p8600/0 : FAILURE
CI_IGT_test runtime 574 seconds
Rebooting fi-ctg-p8600
Comment 13 Marta Löfstedt 2017-12-07 11:11:24 UTC
Softdog on:

last dmesg:
<7>[   47.417493] [drm:_intel_set_memory_cxsr [i915]] memory self-refresh is disabled (was enabled)
<3>[   57.823271] [drm:drm_atomic_helper_wait_for_flip_done] *ERROR* [CRTC:37:pipe A] flip_done timed out
<7>[   57.824605] [IGT] debugfs_test: exiting, ret=0
<7>[   57.838127] [drm:g4x_compute_pipe_wm [i915]] primary A watermarks: normal=24, SR=54, HPLL=154
<7>[   57.838170] [drm:g4x_compute_pipe_wm [i915]] FBC watermarks: SR=3, HPLL=4
<3>[   68.063282] [drm:drm_atomic_helper_wait_for_dependencies] *ERROR* [CRTC:37:pipe A] flip_done timed out
<3>[   78.303234] [drm:drm_atomic_helper_wait_for_dependencies] *ERROR* [PLANE:28:primary A] flip_done timed out
<7>[   78.303563] [drm:_intel_set_memory_cxsr [i915]] memory self-refresh is disabled (was disabled)
<3>[   88.543234] [drm:drm_atomic_helper_wait_for_flip_done] *ERROR* [CRTC:37:pipe A] flip_done timed out
<7>[   88.543499] [drm:_intel_set_memory_cxsr [i915]] memory self-refresh is enabled (was disabled)

running: igt/debugfs_test/read_all_entries

[002/288] pass: 2 -                       
owatch: TIMEOUT!
owatch: timeout for /dev/watchdog0 set to 10 (requested 10)
FATAL: command execution failed
Completed CI_IGT_test CI_DRM_3470/fi-ctg-p8600/0 : FAILURE
CI_IGT_test runtime 177 seconds
Rebooting fi-ctg-p8600
Comment 14 Marta Löfstedt 2017-12-11 07:18:22 UTC

<3>[  489.440225] [drm:drm_atomic_helper_wait_for_dependencies] *ERROR* [PLANE:28:primary A] flip_done timed out
<7>[  489.440383] [drm:_intel_set_memory_cxsr [i915]] memory self-refresh is disabled (was disabled)
<3>[  499.680276] [drm:drm_atomic_helper_wait_for_flip_done] *ERROR* [CRTC:37:pipe A] flip_done timed out
<7>[  499.680429] [drm:_intel_set_memory_cxsr [i915]] memory self-refresh is enabled (was disabled)

running: igt/kms_pipe_crc_basic/read-crc-pipe-a-frame-sequence

[238/288] skip: 50, pass: 188 -                               
owatch: TIMEOUT!
owatch: timeout for /dev/watchdog0 set to 10 (requested 10)
FATAL: command execution failed

NOTE: run.log starts with a lot of:
Unpacking IGT_4051
tar: ./bin/intel-gpu-overlay: time stamp 2017-12-08 20:11:29 is 258.670573436 s in the future
Comment 15 Marta Löfstedt 2018-01-02 07:27:05 UTC

[  325.087233] [drm:drm_atomic_helper_wait_for_flip_done] *ERROR* [CRTC:37:pipe A] flip_done timed out

then incomplete timeout on:


running: igt/kms_chamelium/vga-hpd-fast       
[204/288] skip: 40, pass: 163, dmesg-warn: 1 |
Build timed out (after 17 minutes). Marking the build as aborted.
Comment 17 Marta Löfstedt 2018-01-25 12:41:51 UTC
This is a meta bug for Incompletes on CTG.
Comment 18 Marta Löfstedt 2018-01-25 12:49:20 UTC
This is a Meta bug to capture all unexplained incompletes on CTG
Comment 20 Marta Löfstedt 2018-04-10 12:38:35 UTC
Why has this high priority?
Comment 21 Jani Saarinen 2018-04-10 12:44:49 UTC
This should not as CTG. Re-prioritizing.
Comment 22 Francesco Balestrieri 2018-07-03 14:58:57 UTC
This is a metabug with no issues blocking it, no point keeping it around. Closing.
Comment 23 Martin Peres 2018-07-27 13:20:30 UTC
(In reply to Francesco Balestrieri from comment #22)
> This is a metabug with no issues blocking it, no point keeping it around.
> Closing.

Well, we have plenty of meta bugs like this, and they are here because the issues keep on happening... So yeah, we need to keep it open.
Comment 24 Francesco Balestrieri 2018-08-14 07:47:29 UTC
> Well, we have plenty of meta bugs like this, and they are here because the issues 
> keep on happening... So yeah, we need to keep it open.

Then perhaps we need to track this differently, e.g. with a specific "incomplete" label + platform field, if your purpose is to have a list of incompletes on this platform. I don't see the point in keeping a bug open for things that may happen in the future.
Comment 25 Francesco Balestrieri 2018-08-14 07:49:04 UTC
I'm at least moving this to lowest priority since there is no action required at the moment.
Comment 26 Martin Peres 2019-11-29 17:26:20 UTC
-- GitLab Migration Automatic Message --

This bug has been migrated to freedesktop.org's GitLab instance and has been closed from further activity.

You can subscribe and participate further through the new bug through this link to our GitLab instance: https://gitlab.freedesktop.org/drm/intel/issues/47.

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.