Open tkna91 opened 5 months ago
[wlr] [backend/drm/legacy.c:175] connector HDMI-A-4: drmModePageFlip failed: Device or resource busy
So our attempt to enable the output fails. If you try to enable it again with swaymsg, does it come alive?
Please test with https://github.com/swaywm/sway/pull/8095 to see if that resolves the issue.
@kennylevinsen I checked below and both did not seem to work.
Procedure:
$mod+enter
footswaymsg -t get_outputs
swaymsg -t get_outputs
swaymsg output HDMI-A-3 power on
-> Display not recognizedswaymsg output HDMI-A-3 power off
swaymsg output HDMI-A-3 power on
-> Display not recognizedSo our attempt to enable the output fails. If you try to enable it again with swaymsg, does it come alive?
Sway version 1.10-dev-646019cad (Apr 23 2024, branch 'master')
:Please test with #8095 to see if that resolves the issue.
Sway version 1.10-dev-3e2047e7 (Apr 24 2024, branch 'search_configs')
:It just keeps failing with:
00:00:36.366 [DEBUG] [wlr] [types/output/cursor.c:169] Failed to get cursor display formats
00:00:36.366 [DEBUG] [wlr] [types/output/cursor.c:207] Failed to pick cursor format
00:00:36.366 [DEBUG] [wlr] [types/output/cursor.c:286] Failed to render cursor buffer
00:00:36.366 [DEBUG] [wlr] [types/output/cursor.c:386] Falling back to software cursor on output 'HDMI-A-4'
00:00:36.367 [ERROR] [wlr] [backend/drm/legacy.c:175] connector DVI-D-1: drmModePageFlip failed: Device or resource busy
I followed the following page and it solved the problem.
Apart from this problem, I have been troubled with the following problem since a certain version of sway (sway-git-1.10.r7328.dcb142b-1-x86_64.pkg.tar.zst
).
However, I noticed that when I connected to the other of the two GPUs it would show up. It seems the problem was not with the display, but with the video card. I am using the following two GPUs
# lspci | grep -i "disp\|vga"
00:02.0 Display controller: Intel Corporation Alder Lake-S GT1 [UHD Graphics 730] (rev 0c)
01:00.0 VGA compatible controller: Advanced Micro Devices, Inc. [AMD/ATI] Caicos [Radeon HD 6450/7450/8450 / R5 230 OEM]
#
# udevadm info --attribute-walk --name=/dev/dri/card0 | grep DRIVERS==\"[^\"]
DRIVERS=="radeon"
DRIVERS=="pcieport"
# udevadm info --attribute-walk --name=/dev/dri/card1 | grep DRIVERS==\"[^\"]
DRIVERS=="i915"
#
The following was added to the script that starts udev-rule and Sway, and now it appears correctly in the latest version.
# cat /etc/udev/rules.d/set-dri-links.rules
KERNEL=="card*", SUBSYSTEM=="drm", ATTRS{vendor}=="0x1002", ATTRS{device}=="0x6779", SYMLINK+="dri/by-name/radeon-0x6779"
KERNEL=="card*", SUBSYSTEM=="drm", ATTRS{vendor}=="0x8086", ATTRS{device}=="0x4692", SYMLINK+="dri/by-name/i915-0x4692"
#
if [ "`hostname`" == "gb1" ] ; then
export WLR_DRM_DEVICES=/dev/dri/by-name/i915-0x4692:/dev/dri/by-name/radeon-0x6779
fi
Power off/on is also no problem. There is no problem to display the video in full screen. Perhaps this correspondence could close some of the other existing issues.
You can close it if you wish.
$ pacman -Q sway-git wlroots-git mesa linux
sway-git 1.10.r7348.2e9139d-1
wlroots-git 0.18.0.r7063.6e6c4408d-1
mesa 1:24.1.0-1
linux 6.9.2.arch1-1
$
Any input on the last comment, I have only one GPU (iGPU Intel) but was able to reproduce this issue with an ASUS monitor and a Dell monitor @kennylevinsen under Sway
I have not found a clear commit range where this occurs, but it did not occur at dc9f217. It occurs from dcb142b through 1267e47 .
Reproduction procedure
sway-run-debug
and login$mod+enter
footswaymsg -t get_outputs
swaymsg -t get_outputs
Environment
sway version 1.10-dev-087226d99 (Apr 13 2024, branch 'master')