Bug 103186 - [BAT] igt@* - dmesg-warn: *ERROR* Link Training Unsuccessful | *ERROR* LSPCON mode hasn't settled
Summary: [BAT] igt@* - dmesg-warn: *ERROR* Link Training Unsuccessful | *ERROR* LSPCON...
Status: CLOSED FIXED
Alias: None
Product: DRI
Classification: Unclassified
Component: DRM/Intel (show other bugs)
Version: DRI git
Hardware: Other All
: highest critical
Assignee: shashank.sharma@intel.com
QA Contact: Intel GFX Bugs mailing list
URL:
Whiteboard: ReadyForDev
Keywords:
: 103026 (view as bug list)
Depends on:
Blocks:
 
Reported: 2017-10-10 10:14 UTC by Marta Löfstedt
Modified: 2018-01-05 17:21 UTC (History)
2 users (show)

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


Attachments

Description Marta Löfstedt 2017-10-10 10:14:01 UTC
CI_DRM_3201 fi-cfl-s dmesg-warn on igt@drv_module_reload@basic-reloa

[  519.812385] Setting dangerous option reset - tainting kernel
[  524.508633] to colour frame buffer device 240x67
[  527.542339] Setting dangerous option reset - tainting kernel
[  528.816057] [drm:intel_dp_get_link_train_fallback_values [i915]] *ERROR* Link Training Unsuccessful
[  529.109133] [drm:intel_dp_get_link_train_fallback_values [i915]] *ERROR* Link Training Unsuccessful

https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_3201/fi-cfl-s/igt@drv_module_reload@basic-reload.html
Comment 1 Marta Löfstedt 2017-10-10 14:11:23 UTC
*** Bug 103026 has been marked as a duplicate of this bug. ***
Comment 2 Marta Löfstedt 2017-10-10 14:13:09 UTC
Note duplicate of 103026 is only due to link training issues. The lockdep was solved from 4.14.0-rc3.
Comment 3 Manasi 2017-10-12 21:35:39 UTC
Is this a LSPCON port on CFL?
Not sure if this is LSPCON related issue or another DP timeout issue on CFL.

Manasi
Comment 4 shashank.sharma@intel.com 2017-10-13 08:10:59 UTC
This can be a LSPCON issue, so it needs testing with my patchset https://patchwork.freedesktop.org/series/31639/

Only when this patch series is merged, we would be able to comment on if this is a new issue, or one among the existing ones. 

- Shashank
Comment 5 Jani Nikula 2017-10-13 09:21:13 UTC
Probably fixed by these commits in drm-intel-next-queued:

commit f687e25a7a245952349f1f9f9cc238ac5a3be258
Author: Shashank Sharma <shashank.sharma@intel.com>
Date:   Thu Oct 12 22:10:08 2017 +0530

    drm: Add retries for lspcon mode detection

commit d18aef0f75436abb95894a230b504432df26c167
Author: Shashank Sharma <shashank.sharma@intel.com>
Date:   Tue Oct 10 15:37:43 2017 +0530

    drm/i915: Don't give up waiting on INVALID_MODE

commit a2fc4bd61e7ec3bb1f7c8b3d47272be813f88aea
Author: Shashank Sharma <shashank.sharma@intel.com>
Date:   Tue Oct 10 15:37:44 2017 +0530

    drm/i915: Add retries for LSPCON detection
Comment 6 Jani Saarinen 2017-10-16 10:04:36 UTC
This system needs still fw update.
Comment 7 Marta Löfstedt 2017-10-16 10:45:54 UTC
This is is still reproducible:
https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_3240/fi-cfl-s/igt@gem_exec_suspend@basic-s3.html
Comment 8 Marta Löfstedt 2017-10-31 08:12:25 UTC
new subtests:CI_DRM_3294 fi-cfl-s

It appear to have started on:
igt@debugfs_test@read_all_entries

https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_3294/fi-cfl-s/igt@debugfs_test@read_all_entries.html

[   51.206919] [drm:intel_dp_start_link_train [i915]] *ERROR* failed to start channel equalization

igt@drv_getparams_basic@basic-eu-total

https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_3294/fi-cfl-s/igt@drv_getparams_basic@basic-eu-total.html

[   51.622067] [drm:intel_dp_start_link_train [i915]] *ERROR* failed to enable link training
[   51.704281] [drm:intel_dp_start_link_train [i915]] *ERROR* failed to enable link training

igt@drv_getparams_basic@basic-subslice-total

https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_3294/fi-cfl-s/igt@drv_getparams_basic@basic-subslice-total.html

[   52.147955] [drm:intel_dp_start_link_train [i915]] *ERROR* failed to enable link training

igt@drv_hangman@error-state-basic.html

https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_3294/fi-cfl-s/igt@drv_hangman@error-state-basic.html

[   52.269911] i915 0000:00:02.0: Resetting rcs0 after gpu hang
[   52.523708] [drm:intel_dp_start_link_train [i915]] *ERROR* failed to enable link training
[   52.523798] [drm:intel_dp_get_link_train_fallback_values [i915]] *ERROR* Link Training Unsuccessful
Comment 10 shashank.sharma@intel.com 2017-11-03 08:09:22 UTC
Marta,
Can you please confirm that the LSPCON FW version is 1.70 not 1.63 and older ? 

Cat /sys/kernel/debug/dri/0/i915_display_info
Look for connector DP-1/2 (where we have LSPCON attached)
There would be branch device info like this:
>                DP branch device present: yes
>                Type: HDMI
>                ID: MC2800
>                HW: 2.2
>                SW: 1.63
Comment 11 Marta Löfstedt 2017-11-03 08:26:47 UTC
(In reply to shashank.sharma@intel.com from comment #10)
> Marta,
> Can you please confirm that the LSPCON FW version is 1.70 not 1.63 and older
> ? 
> 
> Cat /sys/kernel/debug/dri/0/i915_display_info
> Look for connector DP-1/2 (where we have LSPCON attached)
> There would be branch device info like this:
> >                DP branch device present: yes
> >                Type: HDMI
> >                ID: MC2800
> >                HW: 2.2
> >                SW: 1.63

From IRC:
<marta_> tsa, shashank_sharma wanted to verify LSPCON FW version on CFL-s, I tried to ssh to it but got "Connection closed...", tsa, could you cat /sys/kernel/debug/dri/0/i915_display_info and update in https://bugs.freedesktop.org/show_bug.cgi?id=103186
<tsa> marta_: 1.63 still. I haven't been able to update the FW
Comment 12 shashank.sharma@intel.com 2017-11-03 08:50:13 UTC
Thanks for the confirmation Marta. 
We need to test this issue with LSPCON FW V1.70, as this issue is fixed on several KBL devices using this FW version.

- Shashank
Comment 13 Marta Löfstedt 2017-11-06 06:51:06 UTC
https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_3313/fi-cfl-s/igt@gem_basic@create-close.html

[   25.491005] [drm:intel_dp_get_link_train_fallback_values [i915]] *ERROR* Link Training Unsuccessful
[   25.820995] [drm:intel_dp_get_link_train_fallback_values [i915]] *ERROR* Link Training Unsuccessful
Comment 14 Marta Löfstedt 2017-11-06 06:51:23 UTC
https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_3313/fi-cfl-s/igt@gem_basic@create-fd-close.html

[   26.134438] [drm:intel_dp_get_link_train_fallback_values [i915]] *ERROR* Link Training Unsuccessful
Comment 16 Marta Löfstedt 2017-11-10 07:07:19 UTC
https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_3329/fi-cfl-s/igt@prime_busy@basic-after-default.html

[  501.968918] [drm:intel_dp_get_link_train_fallback_values [i915]] *ERROR* Link Training Unsuccessful

https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_3329/fi-cfl-s/igt@prime_vgem@basic-fence-flip.html

[  505.388885] [drm:intel_dp_get_link_train_fallback_values [i915]] *ERROR* Link Training Unsuccessful
Comment 19 Marta Löfstedt 2017-11-16 14:28:50 UTC
https://intel-gfx-ci.01.org/tree/drm-tip/IGTPW_502/
Comment 20 Jani Saarinen 2017-11-23 08:49:54 UTC
CFL-S no replaced with CFL-S2, is this issue seen there. 
On CFL-2" now:
DP branch device present: yes
		Type: HDMI
		ID: MC2800
		HW: 2.2
		SW: 1.72
		Max TMDS clock: 600000 kHz
		Max bpc: 16
Comment 21 Marta Löfstedt 2017-11-23 09:21:20 UTC
(In reply to Jani Saarinen from comment #20)
> CFL-S no replaced with CFL-S2, is this issue seen there. 
> On CFL-2" now:
> DP branch device present: yes
> 		Type: HDMI
> 		ID: MC2800
> 		HW: 2.2
> 		SW: 1.72
> 		Max TMDS clock: 600000 kHz
> 		Max bpc: 16

If it was an issue on the cfl-s2 I would have filed that issue on this bug. SO, NO this and no other of the cfl-s bugs have so far not affected cfl-s2.
Comment 22 Jani Saarinen 2017-11-23 09:28:19 UTC
I meant to say:
"CFL-S now replaced with CFL-S2, is this issue seen there."
Comment 23 Jani Saarinen 2017-12-13 07:10:36 UTC
So was this seen only on Old FW version (1.63) and not surfaced later?
Comment 24 Marta Löfstedt 2017-12-13 07:15:08 UTC
(In reply to Jani Saarinen from comment #23)
> So was this seen only on Old FW version (1.63) and not surfaced later?

Again this bug was for the old cfl-s, this bug was archived from cibuglog when cfl-s was replaced with cfl-s2. 

The similar bug for cfl-s2 is bug 104056
Comment 25 Jani Saarinen 2017-12-13 08:44:10 UTC
OK, resolving this then.


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.