Chrome unable to modest internal display on cyan as of ChromeOS 9096.0.0 |
|||||||
Issue description
Chrome Version: 57.0.2954.0
OS: 9096.0.0 (Official Build) dev-channel cyan test
What steps will reproduce the problem?
(1) cros flash {ip} --clobber-stateful xbuddy://remote/cyan/R57-9096.0.0/test
(2) wait for reboot
What is the expected result? OOBE flow appears on internal display
What happens instead? Boot splash screen remains.
/var/log/ui/ui.LATEST contains:
libva info: VA-API version 0.39.2
libva info: va_getDriverName() returns 0
libva info: Trying to open /usr/lib64/va/drivers/i965_drv_video.so
libva info: Found init function __vaDriverInit_0_39
libva info: va_openDriver() returns 0
[1565:1575:1219/142527.880474:ERROR:crtc_controller.cc(43)] Failed to modeset: crtc=27 connector=45 framebuffer_id=57 mode=1366x768@60: Invalid argument
[1565:1575:1219/142527.880569:ERROR:screen_manager.cc(381)] Failed to modeset controller
[1415:1415:1219/142527.881674:ERROR:display_error_observer_chromeos.cc(22)] Failed to configure the following display(s):
[1415:1415:1219/142527.881718:ERROR:display_error_observer_chromeos.cc(24)] - Display with ID = 1881264395124736, and EDID = 00FFFFFFFFFFFF0006AF5C400000000000190104901A0E78029985955556922822505400000001010101010101010101010101010101121B565A50001930302046000090100000180000000F0000000000000000000000000020000000FE0041554F0A202020202020202020000000FE004231313658414E30342E30200A00CC.
[1415:1415:1219/142532.617863:ERROR:device_event_log_impl.cc(140)] [14:25:32.617] USB: usb_service_linux.cc:226 Got duplicate add event for path: /dev/bus/usb/001/003
[1415:1415:1219/142532.662448:ERROR:device_event_log_impl.cc(140)] [14:25:32.662] USB: usb_service_linux.cc:226 Got duplicate add event for path: /dev/bus/usb/001/002
[1415:1415:1219/142532.708685:ERROR:device_event_log_impl.cc(140)] [14:25:32.708] USB: usb_service_linux.cc:226 Got duplicate add event for path: /dev/bus/usb/001/004
[1415:1415:1219/142533.430115:ERROR:geolocation_handler.cc(107)] Failed to get Geolocation data: 0
I discovered (through bisecting) that this started happening with 9096.0.0 (Chrome 57.0.2954.0), but was not happening with 9095.0.0 (Chrome 57.0.2953.0). I will soon take a look at changes in Chrome-side changelog https://chromium.googlesource.com/chromium/src/+log/57.0.2953.0..57.0.2954.0?pretty=fuller&n=10000 but wanted to file this first.
,
Dec 19 2016
I suspect it might be related to crrev.com/2420083003 Do you mind trying to deploy chrome without that patch?
,
Dec 20 2016
I just tried ToT chrome with crrev.com/2420083003 reverted locally, and the device no longer has this issue.
,
Dec 20 2016
I'll take a look tomorrow, I have a cyan.
,
Dec 20 2016
,
Dec 20 2016
Setting priority to 1. I'll revert crrev.com/2420083003 if we can't fix it by tomorrow.
,
Dec 21 2016
I can reproduce it here, but I'm out of time for today. Feel free to revert, I'll continue looking tomorrow.
,
Dec 21 2016
My edgar on 57.0.2955.0 has simalar behavior: crbug.com/675656 Please dupe mine if it appears to be the same issue this.
,
Dec 21 2016
Reverting while we figure out what's the problem: https://codereview.chromium.org/2597683002/
,
Jan 9 2017
Issue 675656 has been merged into this issue.
,
Jan 9 2017
https://codereview.chromium.org/2597683002/ seems to have triggered a bug in intel drm code on 3.18. crtc_state can end up active and not enabled, making drm_atomic_crtc_check fail. https://chromium.googlesource.com/chromiumos/third_party/kernel/+/9b5edbf7a462b7c846e9d4675b9b72b7a4e84de9%5E%21/ and https://chromium.googlesource.com/chromiumos/third_party/kernel/+/c30f55a7b1336cdfeac74c7931ebff40a32e72b7%5E%21/ seems to address the problem and are already in newer kernels. With those two patches modesetting succeeds, but the screen flashes once in the transition between the splash screen and the login screen since we are setting a different crtc.
,
Jan 10 2017
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/dcb4b63e4b07bf55e5ad9c20e9346e42c7eeba84 commit dcb4b63e4b07bf55e5ad9c20e9346e42c7eeba84 Author: dcastagna <dcastagna@chromium.org> Date: Tue Jan 10 04:39:53 2017 ozone: Pick CRTCs with majority of planes. When we pick a CRTC to use for a Display, we should prioritize CRTCs that support the majority of overlays. This CL also sorts the connectors before assigning CRTCs. If a connector A supports only a subset of CRTCs supported by connector B, A should be processed first. BUG= 675755 Committed: https://crrev.com/93a8c0a676c220b4031af879a2691cb96f053480 Cr-Original-Commit-Position: refs/heads/master@{#439248} Review-Url: https://codereview.chromium.org/2420083003 Cr-Commit-Position: refs/heads/master@{#442472} [modify] https://crrev.com/dcb4b63e4b07bf55e5ad9c20e9346e42c7eeba84/ui/ozone/platform/drm/common/drm_util.cc
,
Jan 17 2017
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/7ee2088a61386825bb9544c8e1fe9a11742868bf commit 7ee2088a61386825bb9544c8e1fe9a11742868bf Author: dcastagna <dcastagna@chromium.org> Date: Tue Jan 17 16:55:54 2017 ozone: Pick CRTCs with majority of planes. When we pick a CRTC to use for a Display, we should prioritize CRTCs that support the majority of overlays. This CL also sorts the connectors before assigning CRTCs. If a connector A supports only a subset of CRTCs supported by connector B, A should be processed first. BUG= 675755 , 680517 NOTRY=true NOPRESUBMIT=true Committed: https://crrev.com/93a8c0a676c220b4031af879a2691cb96f053480 Cr-Original-Commit-Position: refs/heads/master@{#439248} Review-Url: https://codereview.chromium.org/2420083003 Cr-Commit-Position: refs/heads/master@{#442472} (cherry picked from commit dcb4b63e4b07bf55e5ad9c20e9346e42c7eeba84) Review-Url: https://codereview.chromium.org/2625813005 Cr-Commit-Position: refs/branch-heads/2924@{#776} Cr-Branched-From: 3a87aecc31cd1ffe751dd72c04e5a96a1fc8108a-refs/heads/master@{#433059} [modify] https://crrev.com/7ee2088a61386825bb9544c8e1fe9a11742868bf/ui/ozone/platform/drm/common/drm_util.cc
,
Mar 9 2018
,
Mar 9 2018
I think this is long fixed... Let's mark it as such so we can get verification. |
|||||||
►
Sign in to add a comment |
|||||||
Comment 1 by jhorwich@chromium.org
, Dec 19 2016