Closed zandrey closed 4 years ago
Some additional Weston logging information from the image built from zeus
branch:
Date: 2019-12-19 UTC
[08:14:43.775] weston 5.0.0
https://wayland.freedesktop.org
Bug reports to: https://gitlab.freedesktop.org/wayland/weston/issues/
Build: 5.0.0-33-gfb563901-dirty MGS-4627 [#ccc] weston will turn off when panel-psotion=none and use gplay (2019-03-14 11:30:30 +0800)
[08:14:43.775] Command line: /usr/bin/weston --log=/var/log/weston.log
[08:14:43.775] OS: Linux, 4.14.98-imx+g1175b5961153, #1 SMP PREEMPT Thu Dec 19 08:36:24 UTC 2019, aarch64
[08:14:43.776] Using config file '/etc/xdg/weston/weston.ini'
[08:14:43.779] Output repaint window is 16 ms maximum.
[08:14:43.780] Loading module '/usr/lib/libweston-5/drm-backend.so'
[08:14:43.787] initializing drm backend
[08:14:43.787] logind: not running in a systemd session
[08:14:43.787] logind: cannot setup systemd-logind helper (-61), using legacy fallback
[08:14:43.791] using /dev/dri/card0
[08:14:43.791] DRM: supports universal planes
[08:14:43.791] DRM: supports atomic modesetting
[08:14:43.791] DRM: does not support picture aspect ratio
[08:14:43.802] Loading module '/usr/lib/libweston-5/g2d-renderer.so'
[08:14:43.922] event0 - 30370000.snvs:snvs-powerkey: is tagged by udev as: Keyboard
[08:14:43.922] event0 - 30370000.snvs:snvs-powerkey: device is a keyboard
[08:14:43.925] event3 - Chicony HP Business Slim Keyboard: is tagged by udev as: Keyboard
[08:14:43.925] event3 - Chicony HP Business Slim Keyboard: device is a keyboard
[08:14:43.929] event4 - Chicony HP Business Slim Keyboard: is tagged by udev as: Keyboard
[08:14:43.929] event4 - Chicony HP Business Slim Keyboard: device is a keyboard
[08:14:43.931] event1 - gpio_ir_recv: is tagged by udev as: Keyboard
[08:14:43.931] event1 - gpio_ir_recv: device is a keyboard
[08:14:43.933] event2 - MCE IR Keyboard/Mouse (gpio-rc-recv): is tagged by udev as: Keyboard Mouse
[08:14:43.933] event2 - MCE IR Keyboard/Mouse (gpio-rc-recv): device is a pointer
[08:14:43.933] event2 - MCE IR Keyboard/Mouse (gpio-rc-recv): device is a keyboard
[08:14:43.987] DRM: head 'HDMI-A-1' found, connector 33 is connected, EDID make 'SAM', model 'SMS24A650', serial 'HTPC800767'
[08:14:43.987] Registered plugin API 'weston_drm_output_api_v1' of size 24
[08:14:43.996] No backlight control for output 'HDMI-A-1'
[08:14:43.996] Output HDMI-A-1 (crtc 31) video modes:
1920x1080@60.0, preferred, current, 148.5 MHz
1920x1080@59.9, 148.4 MHz
1920x1080@50.0, 148.5 MHz
1600x1200@60.0, 162.0 MHz
1680x1050@59.9, 119.0 MHz
1280x1024@75.0, 135.0 MHz
1280x1024@60.0, 108.0 MHz
1440x900@75.0, 136.8 MHz
1440x900@59.9, 88.8 MHz
1280x960@60.0, 108.0 MHz
1280x800@59.9, 71.0 MHz
1152x864@75.0, 108.0 MHz
1280x720@60.0, 74.2 MHz
1280x720@59.9, 74.2 MHz
1280x720@50.0, 74.2 MHz
1024x768@75.0, 78.8 MHz
1024x768@70.1, 75.0 MHz
1024x768@60.0, 65.0 MHz
832x624@74.6, 57.3 MHz
800x600@75.0, 49.5 MHz
800x600@72.2, 50.0 MHz
800x600@60.3, 40.0 MHz
800x600@56.2, 36.0 MHz
720x576@50.0, 27.0 MHz
720x480@60.0, 27.0 MHz
720x480@59.9, 27.0 MHz
640x480@75.0, 31.5 MHz
640x480@72.8, 31.5 MHz
640x480@66.7, 30.2 MHz
640x480@60.0, 25.2 MHz
640x480@59.9, 25.2 MHz
720x400@70.1, 28.3 MHz
[08:14:43.997] associating input device event0 with output HDMI-A-1 (none by udev)
[08:14:43.997] associating input device event3 with output HDMI-A-1 (none by udev)
[08:14:43.997] associating input device event4 with output HDMI-A-1 (none by udev)
[08:14:43.997] associating input device event1 with output HDMI-A-1 (none by udev)
[08:14:43.997] associating input device event2 with output HDMI-A-1 (none by udev)
[08:14:43.997] Output 'HDMI-A-1' enabled with head(s) HDMI-A-1
[08:14:43.997] Compositor capabilities:
arbitrary surface rotation: no
screen capture uses y-flip: no
presentation clock: CLOCK_MONOTONIC, id 1
presentation clock resolution: 0.000000001 s
[08:14:44.003] Loading module '/usr/lib/weston/desktop-shell.so'
[08:14:44.006] launching '/usr/libexec/weston-keyboard'
[08:14:44.008] launching '/usr/libexec/weston-desktop-shell'
[08:14:44.362] deactivating session
[08:14:44.362] event0 - 30370000.snvs:snvs-powerkey: device removed
[08:14:44.410] event3 - Chicony HP Business Slim Keyboard: device removed
[08:14:44.454] event4 - Chicony HP Business Slim Keyboard: device removed
[08:14:44.474] event1 - gpio_ir_recv: device removed
[08:14:44.530] event2 - MCE IR Keyboard/Mouse (gpio-rc-recv): device removed
Here is could be seen that Weston is starting perfectly fine.
Some more information:
I've ran G2D test application provided by imx-gpu-g2d
and it is also appears to be broken. The result of execution of /opt/g2d_samples/g2d_test
is:
g2d_open: fail with status -7
g2d_open fail.
It seems to me that commit f4d05aa18c0192fea2eef3fbfb3390bdb5033c64 brought a faulty version of G2D, which in turn breaks the Weston by failing to initialize.
Since the imx-gpu-g2d
is a binary package - there is not much room for further investigation left to be performed, and I would have to relay it further to NXP for investigation...
-- andrey
@zandrey thanks for the details. I'm working on upgrading the recipe for imx-gpu-viv, meanwhile, can you give it a try? See my fork.
@mihaix thanks for the update!
I've tried your fork and it looks like Weston and G2D are fully operable!
I've seen that the difference is only 2 commits [fae1d46f and 9ce3a55e] from the master
of your fork; the rest looks aligned with master
branch of meta-freescale
.
There is a QA hiccup in your fork that should be addressed:
ERROR: imx-gpu-viv-1_6.4.0.p1.0-aarch64-r0 do_package_qa: QA Issue: libgbm-imx rdepends on imx-gpu-viv-dev [dev-deps]
ERROR: imx-gpu-viv-1_6.4.0.p1.0-aarch64-r0 do_package_qa: QA Issue: libegl-imx rdepends on imx-gpu-viv-dev [dev-deps]
ERROR: imx-gpu-viv-1_6.4.0.p1.0-aarch64-r0 do_package_qa: QA run found fatal errors. Please consider fixing them.
For my personal build to run through, I've patched it with following statements:
+INSANE_SKIP_libegl-imx += "dev-so dev-deps"
+INSANE_SKIP_libgbm-imx += "dev-deps"
This is however not a preferred way, as it is quite unclear why ligegl
and libgbm
require a -dev package from imx-gpu-viv
... It might be that since do_package_qa is executed in parallel and this recipe provides both libegl and libgbm - the QA kicks a bit earlier than expected, which causes this issue...
Nevertheless, I would be waiting now till you'd make a PR and test the layer itself after it would be merged.
Thanks a lot!
-- andrey
Thank you for the update, that's great! You're right, that's not a pretty way of fixing those QA issues, I've seen those later and was still working on them. I've updated the commit, it should be fine now. There were newly added libraries that were not covered in any of the packages and ended up in imx-gpu-viv-dev, which led to the dev-deps QA issue.
Thanks, -- Mihai
Thank you for the update, that's great!
Thanks for solving this issue pretty quick! :)
You're right, that's not a pretty way of fixing those QA issues, I've seen those later and was still working on them. I've updated the commit, it should be fine now. There were newly added libraries that were not covered in any of the packages and ended up in imx-gpu-viv-dev, which led to the dev-deps QA issue.
True, this totally slipped my mind: this is actually a default behavior. If nothing consumes files - they are landing in -dev
package and leading to QA Error afterwards. I'd note this for myself for the future.
Anyways, I guess once the PR would come - this issue can be closed.
-- andrey
Hello all,
I've tested the latest master
and with PR #249 this issue is resolved, Weston is fully operable.
Thanks for taking care of this.
-- andrey
Duplicating the original thread from meta-freescale yocto mailing list:
After the recent upgrade of Weston (commit: a4fdb4074644bda09b46490163886065b4c00aab) I started to receive the error from G2D on i.MX8M Mini at the start with following output:
If the G2D support is disabled in the /etc/xdg/weston/weston.ini by commenting the line:
#use-g2d=1
then Weston falls back to use the GL backend, which also fails with the following output:
zeus
branch is not affected by this issue, there the Weston is operable.Following are SRCREVs of components in question that could've potentially impact the Weston. From
master
branch build (which is failing):From
zeus
branch build (which is operable):To me it looks like graphics drivers are broken in the 4.19.y version (which were also recently upgraded to match kernel major/minor).
Looks like we're hitting a regression in Weston with all recent upgrades from NXP of all components (Kernel, drivers, Weston, ...).
@otavio @thochstein @mihaix: can you please have a look at this? I believe a lot of people would be affected by this once Yocto 3.1 would roll-out and everyone would start to upgrade.
Thanks a lot!
-- andrey