Build is NRD20B

The first set of logs from Marc show:

<3>[  802.063323] tegra-sor 54580000.sor: failed to compute configuration: -34

Which occurs just after link training. I've asked Marc to generate a cleaner set of logs just in case this isn't the root cause, since this is a number of minutes into the boot.

I've attached the kernel logs from the bug report for completeness.

Deleted: imagepro-II_kernel_logs.txt 235 KB

imagepro-II_kernel_logs.txt 235 KB View Download

Re #4:

The error "failed to compute configuration -34" means the selected link rate and numbers may not meet the output requirement. But it should not be the root cause and maybe some other error cause.

Sean,

May we have a chance to enable the drm debug print? It will tell me more things about the fail.

The patch for enabling debug print is:

diff --git a/arch/arm64/boot/dts/tegra/tegra210-smaug.dtsi b/arch/arm64/boot/dts/tegra/tegra210-smaug.dtsi
index ed29fb4aa0fa..ecb32e03ddfd 100644
--- a/arch/arm64/boot/dts/tegra/tegra210-smaug.dtsi
+++ b/arch/arm64/boot/dts/tegra/tegra210-smaug.dtsi
@@ -16,7 +16,7 @@
        compatible = "google,smaug", "nvidia,tegra210";
 
         chosen {
-               bootargs = "cros_legacy console= init=/init rw ignore_loglevel=1 drm.atomic=1 firmware_class.path=/vendor/firmware kern_guid=%U no_console_suspend";
+               bootargs = "cros_legacy console= init=/init rw ignore_loglevel=1 drm.atomic=1 firmware_class.path=/vendor/firmware kern_guid=%U no_console_suspend drm.debug=0x4";
        };
 
        memory {

Thanks

> it should not be the root cause

I suppose it depends on what mode we're trying to set, and what speed we trained the link at. I got another set of logs from Marc with the same failure, so at least it's reproducible.

I'm going to head to CA next week to get hands on with the projector. Hopefully something will pop out.

Alright. So the issue is that we're getting a 0 value for color depth from the ImagePro EDID. This causes the DP helper code to incorrectly choose 1 lane @ 1.62GHz for training. This is obviously insufficient and things fail.

I've posted https://chromium-review.googlesource.com/#/c/341705 to use default bpc@8 if not supplied by the EDID. With this patch, we can do 1080p reliably.

The following revision refers to this bug:
  https://chromium.googlesource.com/chromiumos/third_party/kernel/+/0f8aa36a86e0e390887c96244ca33660e48e4234

commit 0f8aa36a86e0e390887c96244ca33660e48e4234
Author: Sean Paul <seanpaul@chromium.org>
Date: Mon May 02 21:36:37 2016

drm/tegra: Use default bpc if not provided in EDID

If the EDID provides an invalid EDID (either too high or 0),
use the default bpc of 8 for link training purposes.

BUG= chromium:605723 
TEST=Tested on ImagePRO-II 1920x1080@60, works

Change-Id: Iffbec7dc4f3f0dfeac2e9130219d45c44b835254
Signed-off-by: Sean Paul <seanpaul@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/341800
Reviewed-by: Andrew Bresticker <abrestic@chromium.org>
Commit-Queue: Andrew Bresticker <abrestic@chromium.org>
Tested-by: Andrew Bresticker <abrestic@chromium.org>

[modify] https://crrev.com/0f8aa36a86e0e390887c96244ca33660e48e4234/drivers/gpu/drm/tegra/sor.c

Fixed in build 2830460

This is great. Nice work.


From: seanpaul@chromium.org via Monorail [mailto:monorail@chromium.org]
Sent: Tuesday, May 03, 2016 10:40 AM
To: Larry Robinson
Subject:  Issue 605723  in chromium: ryu: DP doesn't work on Folsom ImagePRO-II

The following revision refers to this bug:
  https://chromium.googlesource.com/chromiumos/third_party/kernel/+/6520c2d13974ca019dc627169d99bce19c0d50e0

commit 6520c2d13974ca019dc627169d99bce19c0d50e0
Author: Sean Paul <seanpaul@chromium.org>
Date: Mon May 02 21:36:37 2016

drm/tegra: Use default bpc if not provided in EDID

If the EDID provides an invalid EDID (either too high or 0),
use the default bpc of 8 for link training purposes.

BUG= chromium:605723 
TEST=Tested on ImagePRO-II 1920x1080@60, works

Change-Id: Iffbec7dc4f3f0dfeac2e9130219d45c44b835254
Signed-off-by: Sean Paul <seanpaul@chromium.org>
Reviewed-on: https://chromium-review.googlesource.com/341705
Commit-Ready: Sean Paul <seanpaul@google.com>
Tested-by: Sean Paul <seanpaul@google.com>
Reviewed-by: Stéphane Marchesin <marcheu@chromium.org>
Reviewed-by: Zach Reizner <zachr@chromium.org>
Reviewed-by: Haixia Shi <hshi@chromium.org>

[modify] https://crrev.com/6520c2d13974ca019dc627169d99bce19c0d50e0/drivers/gpu/drm/tegra/sor.c