Issue needs an owner and prioritized if RBB.

Hi, issue still needs an owner and fix prioritized. Has this been validated on the latest version? There were related issues that were fixed. Please update, Thanks!

Able to reproduce the issue on 69.0.3497.21/10895.9.0 - Reks.  

seems it is duplicated of issue/850431

Issue appears to be limited to Strago boards, these can be skipped during Beta. Boards in  issue 850431  included, as it is also unresolved at the moment. 

Should merge  issue 850431  if it is a duplicate and update bug to reflect it is a Strago issue.

As developer mode is not fully supported (we strictly speaking only support normal mode), if we wanted we could push out a beta with this anyway.

Particularly since developer mode mostly still appears to work, and one can get to a console using ctrl+alt+t then running `shell` after logging in.

 Issue 850431  has been merged into this issue.

So, it looks like developer console works underneath but the screen is black.
Modeset seems to have worked, backlight is on so there must be something with display (PSR?) that goes wrong.

Developer mode will not block beta, moving to RBS. 

Well, with debugging enabled it looks like everything in modeset succeeds.
Transfering drm master from chrome to frecon works as well.
If Chrome is stopped and frecon started from scratch it works.
Also, if Chrome is running after switch to frecon and killing frecon modetest is unable to set mode either.
I suspect that something goes awry in the driver when we switch from
CONNECTOR:47 - ENCODER:46 - CRTC:26 that Chrome uses to
CONNECTOR:47 - ENCODER:46 - CRTC:31 that frecon uses.

Able to reproduce the issue on 69.0.3497.35/10895.21.0 dev Channel Celes device.

I think it happens on all Braswell. 
Modeset succeeds but the screen remains black.
I have a patch for frecon that tries to use the same CRTC for connector as Chrome used here https://chromium-review.googlesource.com/c/chromiumos/platform/frecon/+/1168643
But weirdly enough, while this works for first switch to console it still has the same problem in subsequent switches to developer console.

Can you please add "intel@chromium.org" to the CC list.

Results of regression analysis 

                                     Result	Release date
ChromeOS-test-R69-10719.0.0-reks
(68.0.3438.0)	                      Pass	2018-05-26

ChromeOS-test-R69-10740.0.0-reks
(69.0.3445.0)	                      Pass	2018-06-01

ChromeOS-test-R69-10760.0.0-reks
 (69.0.3452.0)	                     Failed	2018-06-07

There is no patch commit on cros kernel 3.18 branch from 2018-06-01 to 2018-06-07, suspecting regression from user space.

Additional data on regression. Data pointing to user space regression (no commit on cros kernel 3.18 branch between pass and fail build).

ChromeOS-test-R69-10758.0.0-reks(69.0.3451.0)      Pass	        2018-06-07
ChromeOS-test-R69-10760.0.0-reks(69.0.3452.0)      Failed	2018-06-07

I do not think it is a regression. 
I also suspect the user space changed behavior when this has happened. Probably Chrome started doing "more interesting" things like using overlays or setting video mode differently or using some different tiled mode.

When you enable drm debugging in kernel mode set ioctls succeed without any error, yet the screen remains black. This I think means something silently fails in kernel or some display register is not programmed correctly so the screen remains blank.

1. We look into kernel log, eDP link training works well after switch to VT2 but nothing display.
2. Developer console can works by manual configure on image 69.0.3491.0/10880.0.0. Here are the steps.
   a. Boot DUT up with wire/wirless network connection.
   b. Create ssh connection to DUT and enter below commands.
      # stop ui
      # frecon

Looks like some unexpected behaviors to cause this symptom after cros suspend then switch VT2 on. Do you have any comment for further debug?

Is this reproducible in the latest M69?

Issue is reproducible on 69.0.3497.76/10895.42.0 beta channel Reks and Celes.

@shawn.c.lee, #23 what type of debugging information would you need?

1. The issue of blank screen is happening because the color space 
   coefficients (ctm) are being set to 0 during switch from chrome ui to 
   developer console. 

2. I do not see the blank screen issue while doing modeset using modetest 
   tool on the target. Ie.. 
   a. stop ui
   b. looping through the following modetest scenarios.
      modetest -s 47@26:1366x768 
      modetest -s 47@31:1366x768

dbehr@ is OOO. Per chat, gurchetansingh@ will take a look. Thanks!

Is this reproducible on M70 or ToT?

Issue is reproducible on  70.0.3538.15/11021.12.0 dev channel Reks and 71.0.3544.0/11054.0.0 dev channel Reks 

Comment #27 is correct, it looks like it's the CTM.  Ignoring the property when programming the display fixes the issue:

diff --git a/drivers/gpu/drm/i915/intel_color.c b/drivers/gpu/drm/i915/intel_color.c
index ac7028659310b..e5d530aa77935 100644
--- a/drivers/gpu/drm/i915/intel_color.c
+++ b/drivers/gpu/drm/i915/intel_color.c
@@ -100,7 +100,7 @@ static void i9xx_load_csc_matrix(struct drm_crtc *crtc)
        int i, pipe = intel_crtc->pipe;
        uint16_t coeffs[9] = { 0, };
 
-       if (crtc_state->ctm) {
+       if (false) {
                struct drm_color_ctm *ctm =
                        (struct drm_color_ctm *)crtc_state->ctm->data;
                uint64_t input[9] = { 0, };
@@ -210,7 +210,7 @@ static void cherryview_load_csc_matrix(struct drm_crtc *crtc)
        int pipe = to_intel_crtc(crtc)->pipe;
        uint32_t mode;
 
-       if (state->ctm) {
+       if (false) {
                struct drm_color_ctm *ctm =
                        (struct drm_color_ctm *) state->ctm->data;
                uint16_t coeffs[9] = { 0, };
@@ -243,7 +243,7 @@ static void cherryview_load_csc_matrix(struct drm_crtc *crtc)
                I915_WRITE(CGM_PIPE_CSC_COEFF8(pipe), coeffs[8]);
        }
 
-       mode = (state->ctm ? CGM_PIPE_MODE_CSC : 0);
+       mode = (false ? CGM_PIPE_MODE_CSC : 0);
        if (!crtc_state_is_legacy(state)) {
                mode |= (state->degamma_lut ? CGM_PIPE_MODE_DEGAMMA : 0) |
                        (state->gamma_lut ? CGM_PIPE_MODE_GAMMA : 0);
@@ -433,7 +433,7 @@ static void cherryview_load_luts(struct drm_crtc *crtc)
        if (crtc_state_is_legacy(state)) {
                /* Turn off degamma/gamma on CGM block. */
                I915_WRITE(CGM_PIPE_MODE(pipe),
-                          (state->ctm ? CGM_PIPE_MODE_CSC : 0));
+                          (false ? CGM_PIPE_MODE_CSC : 0));
                i9xx_load_luts_internal(crtc, state->gamma_lut);
                return;
        }
@@ -469,7 +469,7 @@ static void cherryview_load_luts(struct drm_crtc *crtc)
        }
 
        I915_WRITE(CGM_PIPE_MODE(pipe),
-                  (state->ctm ? CGM_PIPE_MODE_CSC : 0) |
+                  (false ? CGM_PIPE_MODE_CSC : 0) |
                   (state->degamma_lut ? CGM_PIPE_MODE_DEGAMMA : 0) |
                   (state->gamma_lut ? CGM_PIPE_MODE_GAMMA : 0));


Some Chrome change between 69.0.3451.0 and 69.0.3452.0 must've altered how the property is set.

In theory, we can modify Frecon to set the property correctly.  crrev.com/c/1220601 is the WiP CL.

Kernel change is made to resolve the issue. Back porting the patch to 3.18 is in progress. 

https://chromium-review.googlesource.com/c/chromiumos/third_party/kernel/+/1220974

Thanks! Let me know when you are ready for merge approval for M69.

Still seeing this issue on Reks with today's beta build 69.0.3497.95/10895.56.0

This fix is the CQ for ToT.  I've tested it, requesting merge to M69.

This bug requires manual review: Request affecting a post-stable build
Please contact the milestone owner if you have questions.
Owners: amineer@(Android), kariahda@(iOS), cindyb@(ChromeOS), govind@(Desktop)

For more details visit https://www.chromium.org/issue-tracking/autotriage - Your friendly Sheriffbot

update: Issue is reproducible on Google Chrome 69.0.3497.95 Platform 10895.56.0 Reks beta build

The following revision refers to this bug:
  https://chromium.googlesource.com/chromiumos/third_party/kernel/+/0d4ac83af53d59701876508eae9d11d54c21a2ef

commit 0d4ac83af53d59701876508eae9d11d54c21a2ef
Author: P Raviraj Sitaram <raviraj.p.sitaram@intel.com>
Date: Fri Sep 14 00:43:26 2018

BACKPORT: FROMGIT: drm/i915/chv: Update csc coefficient matrix during modeset

During modeset, previously configured csc coefficient matrix,if any, will
not persist. This can result in blank screen as csc mode will be programmed
while loading LUT but csc coefficient matrix remains unprogrammed.

Changes since V1:
- Removed platform check

Signed-off-by: P Raviraj Sitaram <raviraj.p.sitaram@intel.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/1536589634-29680-1-git-send-email-raviraj.p.sitaram@intel.com
(cherry picked from commit c59d2da8ec189489a30b351d7fdb5923bcb9ca5f
git://anongit.freedesktop.org/drm-tip drm-tip)

Conflicts:
   drivers/gpu/drm/i915/intel_display.c
(Manually resolved conflicts because of code divergence in intel_display.c
from upstream and function signature change in valleyview_crtc_enable())

Signed-off-by: P Raviraj Sitaram <raviraj.p.sitaram@intel.com>

BUG= chromium:863853 
TEST=on cherryview devices, switch to dev console from chrome ui.
     dev console should come up properly

Change-Id: I15b9158d00e3cc898171b9fa74d5112491740c62
Reviewed-on: https://chromium-review.googlesource.com/1225891
Reviewed-by: Robert Tarasov <tutankhamen@chromium.org>
Commit-Queue: Gurchetan Singh <gurchetansingh@chromium.org>
Tested-by: Gurchetan Singh <gurchetansingh@chromium.org>
Trybot-Ready: Gurchetan Singh <gurchetansingh@chromium.org>

[modify] https://crrev.com/0d4ac83af53d59701876508eae9d11d54c21a2ef/drivers/gpu/drm/i915/intel_display.c

If this is just VT2 which is just in dev mode, it does not impact the normal user experience, and there is a trivial workaround (ctrl+alt+t; shell), I don't think we need to hold back stable for this, since we have a fix merged we can pick it up in the second stable.

 Issue 880588  has been merged into this issue.

The following revision refers to this bug:
  https://chromium.googlesource.com/chromiumos/third_party/kernel/+/422cda132b4bd0677f6fadea218730048b05b294

commit 422cda132b4bd0677f6fadea218730048b05b294
Author: P Raviraj Sitaram <raviraj.p.sitaram@intel.com>
Date: Sat Sep 15 00:05:57 2018

BACKPORT: FROMGIT: drm/i915/chv: Update csc coefficient matrix during modeset

During modeset, previously configured csc coefficient matrix,if any, will
not persist. This can result in blank screen as csc mode will be programmed
while loading LUT but csc coefficient matrix remains unprogrammed.

Changes since V1:
- Removed platform check

Signed-off-by: P Raviraj Sitaram <raviraj.p.sitaram@intel.com>
Signed-off-by: Ville Syrjl <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/1536589634-29680-1-git-send-email-raviraj.p.sitaram@intel.com
(cherry picked from commit c59d2da8ec189489a30b351d7fdb5923bcb9ca5f
git://anongit.freedesktop.org/drm-tip drm-tip)

Conflicts:
   drivers/gpu/drm/i915/intel_display.c
(Manually resolved conflicts because of code divergence in intel_display.c
from upstream and function signature change in valleyview_crtc_enable())

Signed-off-by: P Raviraj Sitaram <raviraj.p.sitaram@intel.com>

BUG= chromium:863853 
TEST=on cherryview devices, switch to dev console from chrome ui.
     dev console should come up properly

Change-Id: I15b9158d00e3cc898171b9fa74d5112491740c62
Reviewed-on: https://chromium-review.googlesource.com/1220974
Commit-Ready: ChromeOS CL Exonerator Bot <chromiumos-cl-exonerator@appspot.gserviceaccount.com>
Tested-by: Raviraj P Sitaram <raviraj.p.sitaram@intel.com>
Reviewed-by: Anand Naik <anand.naik@intel.corp-partner.google.com>
Reviewed-by: Stéphane Marchesin <marcheu@chromium.org>
Reviewed-by: Gurchetan Singh <gurchetansingh@chromium.org>

[modify] https://crrev.com/422cda132b4bd0677f6fadea218730048b05b294/drivers/gpu/drm/i915/intel_display.c

The following revision refers to this bug:
  https://chromium.googlesource.com/chromiumos/third_party/kernel/+/8cb27e42b4a859b8f346cd58d31f71411bc72de6

commit 8cb27e42b4a859b8f346cd58d31f71411bc72de6
Author: P Raviraj Sitaram <raviraj.p.sitaram@intel.com>
Date: Sat Sep 15 00:14:50 2018

BACKPORT: FROMGIT: drm/i915/chv: Update csc coefficient matrix during modeset

During modeset, previously configured csc coefficient matrix,if any, will
not persist. This can result in blank screen as csc mode will be programmed
while loading LUT but csc coefficient matrix remains unprogrammed.

Changes since V1:
- Removed platform check

Signed-off-by: P Raviraj Sitaram <raviraj.p.sitaram@intel.com>
Signed-off-by: Ville Syrjl <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/1536589634-29680-1-git-send-email-raviraj.p.sitaram@intel.com
(cherry picked from commit c59d2da8ec189489a30b351d7fdb5923bcb9ca5f
git://anongit.freedesktop.org/drm-tip drm-tip)

Conflicts:
   drivers/gpu/drm/i915/intel_display.c
(Manually resolved conflicts because of code divergence in intel_display.c
from upstream and function signature change in valleyview_crtc_enable())

Signed-off-by: P Raviraj Sitaram <raviraj.p.sitaram@intel.com>

BUG= chromium:863853 
TEST=on cherryview devices, switch to dev console from chrome ui.
     dev console should come up properly

Change-Id: I15b9158d00e3cc898171b9fa74d5112491740c62
Reviewed-on: https://chromium-review.googlesource.com/1220974
Commit-Ready: ChromeOS CL Exonerator Bot <chromiumos-cl-exonerator@appspot.gserviceaccount.com>
Tested-by: Raviraj P Sitaram <raviraj.p.sitaram@intel.com>
Reviewed-by: Anand Naik <anand.naik@intel.corp-partner.google.com>
Reviewed-by: Stéphane Marchesin <marcheu@chromium.org>
Reviewed-by: Gurchetan Singh <gurchetansingh@chromium.org>
(cherry picked from commit 422cda132b4bd0677f6fadea218730048b05b294)
Reviewed-on: https://chromium-review.googlesource.com/1226403
Commit-Queue: Gurchetan Singh <gurchetansingh@chromium.org>
Tested-by: Gurchetan Singh <gurchetansingh@chromium.org>
Trybot-Ready: Gurchetan Singh <gurchetansingh@chromium.org>

[modify] https://crrev.com/8cb27e42b4a859b8f346cd58d31f71411bc72de6/drivers/gpu/drm/i915/intel_display.c

This issue has been approved for a merge. Please merge the fix to any appropriate branches as soon as possible!

If all merges have been completed, please remove any remaining Merge-Approved labels from this issue.

Thanks for your time! To disable nags, add the Disable-Nags label.

For more details visit https://www.chromium.org/issue-tracking/autotriage - Your friendly Sheriffbot

The following revision refers to this bug:
  https://chromium.googlesource.com/chromiumos/platform/frecon/+/57b5c72ce23adb3c1ad45f6d471c8fce84641ced

commit 57b5c72ce23adb3c1ad45f6d471c8fce84641ced
Author: Dominik Behr <dbehr@chromium.org>
Date: Wed Oct 03 20:59:07 2018

frecon: refactor modeset and pretty much all drm module

When frecon sets video mode on cyan it ends up picking different encoder/crtc
for connector from Chrome. Setting video mode succeeds but the display is blank.

This change changes how we determine which encoder and crtc to use, preferring
one that are already configured. Along the way much of the code for mode set
has been refactored and simplified. Connectors are disabled/configured in a
single loop that has full understanding of how to configure all of them.

BUG= chromium:863853 
TEST=run frecon on cyan and kevin

Change-Id: If4df2889814c840777a3f0e62379a1e2f6f06054
Signed-off-by: Dominik Behr <dbehr@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/1168643
Commit-Ready: ChromeOS CL Exonerator Bot <chromiumos-cl-exonerator@appspot.gserviceaccount.com>

[modify] https://crrev.com/57b5c72ce23adb3c1ad45f6d471c8fce84641ced/term.c
[modify] https://crrev.com/57b5c72ce23adb3c1ad45f6d471c8fce84641ced/drm.h
[modify] https://crrev.com/57b5c72ce23adb3c1ad45f6d471c8fce84641ced/fb.c
[modify] https://crrev.com/57b5c72ce23adb3c1ad45f6d471c8fce84641ced/drm.c

The following revision refers to this bug:
  https://chromium.googlesource.com/chromiumos/platform/frecon/+/7bb94df09196e7c863e83dec9e4105dd13fad14b

commit 7bb94df09196e7c863e83dec9e4105dd13fad14b
Author: Dominik Behr <dbehr@chromium.org>
Date: Wed Oct 10 04:14:27 2018

frecon: add atomic modeset

This should make switch to console potentially faster on kernels that
support atomic ioctl.

BUG= chromium:863853 ,b:110258421
TEST=switch to console on Grunt

Change-Id: I792e962d897df894994035c3267e0d0f5136698e
Signed-off-by: Dominik Behr <dbehr@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/1176396
Commit-Ready: ChromeOS CL Exonerator Bot <chromiumos-cl-exonerator@appspot.gserviceaccount.com>
Reviewed-by: Stéphane Marchesin <marcheu@chromium.org>

[modify] https://crrev.com/7bb94df09196e7c863e83dec9e4105dd13fad14b/drm.h
[modify] https://crrev.com/7bb94df09196e7c863e83dec9e4105dd13fad14b/drm.c

The following revision refers to this bug:
  https://chromium.googlesource.com/chromiumos/platform/frecon/+/4601b90d0472ac826f1c11ec5a2c1fafb19192c5

commit 4601b90d0472ac826f1c11ec5a2c1fafb19192c5
Author: Ross Zwisler <zwisler@chromium.org>
Date: Fri Oct 12 11:47:17 2018

Revert "frecon: add atomic modeset"

This reverts commit 7bb94df09196e7c863e83dec9e4105dd13fad14b.

Reason for revert:
  Broke virtual terminals on octopus.  Reverting this one change
  fixed the issue.

BUG=b:117590941,b:117559913

Original change's description:
> frecon: add atomic modeset
>
> This should make switch to console potentially faster on kernels that
> support atomic ioctl.
>
> BUG= chromium:863853 ,b:110258421
> TEST=switch to console on Grunt
>
> Change-Id: I792e962d897df894994035c3267e0d0f5136698e
> Signed-off-by: Dominik Behr <dbehr@chromium.org>
> Reviewed-on: https://chromium-review.googlesource.com/1176396
> Commit-Ready: ChromeOS CL Exonerator Bot <chromiumos-cl-exonerator@appspot.gserviceaccount.com>
> Reviewed-by: Stphane Marchesin <marcheu@chromium.org>

Bug:  chromium:863853 , b:110258421
Change-Id: I5305b183b90340bbca97a37e6ab6465240febe15
Reviewed-on: https://chromium-review.googlesource.com/1277045
Commit-Ready: Daniel Kurtz <djkurtz@chromium.org>
Tested-by: Ross Zwisler <zwisler@chromium.org>
Reviewed-by: Dominik Behr <dbehr@chromium.org>
Reviewed-by: Benson Leung <bleung@chromium.org>
Reviewed-by: Drew Davenport <ddavenport@chromium.org>
Reviewed-by: Ross Zwisler <zwisler@chromium.org>

[modify] https://crrev.com/4601b90d0472ac826f1c11ec5a2c1fafb19192c5/drm.h
[modify] https://crrev.com/4601b90d0472ac826f1c11ec5a2c1fafb19192c5/drm.c

The following revision refers to this bug:
  https://chromium.googlesource.com/chromiumos/platform/frecon/+/bd14c48930f102edb0b2c7d47a02082aa9459a63

commit bd14c48930f102edb0b2c7d47a02082aa9459a63
Author: Ross Zwisler <zwisler@chromium.org>
Date: Fri Oct 12 21:51:28 2018

Revert "frecon: add atomic modeset"

This reverts commit 7bb94df09196e7c863e83dec9e4105dd13fad14b.

Reason for revert:
  Broke virtual terminals on octopus.  Reverting this one change
  fixed the issue.

BUG=b:117590941,b:117559913

Original change's description:
> frecon: add atomic modeset
>
> This should make switch to console potentially faster on kernels that
> support atomic ioctl.
>
> BUG= chromium:863853 ,b:110258421
> TEST=switch to console on Grunt
>
> Change-Id: I792e962d897df894994035c3267e0d0f5136698e
> Signed-off-by: Dominik Behr <dbehr@chromium.org>
> Reviewed-on: https://chromium-review.googlesource.com/1176396
> Commit-Ready: ChromeOS CL Exonerator Bot <chromiumos-cl-exonerator@appspot.gserviceaccount.com>
> Reviewed-by: Stphane Marchesin <marcheu@chromium.org>

Bug:  chromium:863853 , b:110258421
Reviewed-on: https://chromium-review.googlesource.com/1277045
Commit-Ready: Daniel Kurtz <djkurtz@chromium.org>
Tested-by: Ross Zwisler <zwisler@chromium.org>
Reviewed-by: Dominik Behr <dbehr@chromium.org>
Reviewed-by: Benson Leung <bleung@chromium.org>
Reviewed-by: Drew Davenport <ddavenport@chromium.org>
Reviewed-by: Ross Zwisler <zwisler@chromium.org>
(cherry picked from commit 4601b90d0472ac826f1c11ec5a2c1fafb19192c5)

Change-Id: I5305b183b90340bbca97a37e6ab6465240febe15
Reviewed-on: https://chromium-review.googlesource.com/c/1279112
Reviewed-by: Kirtika Ruchandani <kirtika@chromium.org>
Commit-Queue: Benson Leung <bleung@chromium.org>
Tested-by: Benson Leung <bleung@chromium.org>
Trybot-Ready: Benson Leung <bleung@chromium.org>

[modify] https://crrev.com/bd14c48930f102edb0b2c7d47a02082aa9459a63/drm.h
[modify] https://crrev.com/bd14c48930f102edb0b2c7d47a02082aa9459a63/drm.c

The following revision refers to this bug:
  https://chromium.googlesource.com/chromiumos/platform/frecon/+/bf08a557cafc68e63c03fe9d87d6b19de202c14b

commit bf08a557cafc68e63c03fe9d87d6b19de202c14b
Author: Dominik Behr <dbehr@chromium.org>
Date: Sat Oct 27 00:19:00 2018

frecon: add atomic modeset V2

This should make switch to console potentially faster on kernels that
support atomic ioctl.

v2: fixed error check condition, added reset of gamma tables and color matrix.

BUG= chromium:863853 ,b:110258421
TEST=switch to console on Grunt

Change-Id: I9c1bf863959e3840157d5c3675454e238e226041
Signed-off-by: Dominik Behr <dbehr@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/1300254
Commit-Ready: ChromeOS CL Exonerator Bot <chromiumos-cl-exonerator@appspot.gserviceaccount.com>
Reviewed-by: Stéphane Marchesin <marcheu@chromium.org>

[modify] https://crrev.com/bf08a557cafc68e63c03fe9d87d6b19de202c14b/drm.h
[modify] https://crrev.com/bf08a557cafc68e63c03fe9d87d6b19de202c14b/drm.c