hyprwm / Hyprland

Hyprland is an independent, highly customizable, dynamic tiling Wayland compositor that doesn't sacrifice on its looks.
https://hyprland.org
BSD 3-Clause "New" or "Revised" License
21.84k stars 908 forks source link

Screen not rendering on Intel graphics #8541

Closed themonglord closed 2 days ago

themonglord commented 4 days ago

Already reported ? *

Regression?

Yes

System Info and Version

Hyprland 0.45.0 built from branch at commit dirty (). Date: Tag: , commits: built against aquamarine 0.5.0

flags set: debug

System Information: System name: Linux Node name: FreakyArchBTW Release: 6.11.9-arch1-1 Version: #1 SMP PREEMPT_DYNAMIC Sun, 17 Nov 2024 16:06:17 +0000

GPU information: 00:02.0 VGA compatible controller [0300]: Intel Corporation Skylake GT2 [HD Graphics 520] [8086:1916] (rev 07) (prog-if 00 [VGA controller])

os-release: NAME="Arch Linux" PRETTY_NAME="Arch Linux" ID=arch BUILD_ID=rolling ANSI_COLOR="38;2;23;147;209" HOME_URL="https://archlinux.org/" DOCUMENTATION_URL="https://wiki.archlinux.org/" SUPPORT_URL="https://bbs.archlinux.org/" BUG_REPORT_URL="https://gitlab.archlinux.org/groups/archlinux/-/issues" PRIVACY_POLICY_URL="https://terms.archlinux.org/docs/privacy-policy/" LOGO=archlinux-logo

plugins: unknown: not runtime

Description

This has started happening after I updated hyprland to 0.45.1. Whenever I start hyprland (either via tty or sddm), the screen turns off and stays off. When I switch ttys and go back to the hyprland tty, it shows one frame and it freezes. I am running an Intel I5-6300u with a 520 integrated gpu.

How to reproduce

Start Hyprland (either via sddm or tty) on the cpu specified on v0.45.1 and above

Crash reports, logs, images, videos

No response

vaxerski commented 3 days ago

can you bisect this? https://wiki.hyprland.org/Crashes-and-Bugs/#bisecting-an-issue

themonglord commented 3 days ago

ok, between 0.45.0 and 0.45.1 (where this started happening), it output: [c4a77b8da7783d95345935cd2468d83cd54d182b] core: guard pmonitor in focuswindow

vaxerski commented 2 days ago

I don't think you did it correctly if that's the output. Check the git head (latest git) and if it's bad, select that as bad and put 0.45 as good

themonglord commented 2 days ago

between 0.45.0 and head [83be2480c45bdead9110fdf137a1cd2f5a203378] workspace: fix missing name via focusworkspaceoncurrentmonitor (#8484)

MightyPlaza commented 2 days ago

are you sure this couldn't be a aquamarine regression? https://gitlab.archlinux.org/archlinux/packaging/packages/hyprland/-/commit/a6f593168059910617ec760ff3e49f9fbafdc787 was released a day before 0.45 on arch

themonglord commented 2 days ago

it is aquamarine. i installed that version of hyprland and it still isn't rendering. pacman doesn't wanna downgrade aquamarine because it apparently breaks this "libaquamarine" dependency that hyprland uses.

edit: bisect for aquamarine between 0.4.5 and head [3b00e96f90cb0040de6d88ad99bf5f4d443f0c59] drm: scan only cards and not outputs, safeguard against null renderer (#106)

m4rch3n1ng commented 2 days ago

what version of mesa are you using @themonglord? i noticed the same, also on intel, when upgrading mesa to version 24.3.0 (and reported it to upstream mesa), and the fix for me was downgrading mesa to 24.2.7 (by, in my case, disabling the arch testing repos and running pacman -Syuu)

themonglord commented 2 days ago

what version of mesa are you using @themonglord? i noticed the same, also on intel, when upgrading mesa to version 24.3.0 (and reported it to upstream mesa), and the fix for me was downgrading mesa to 24.2.7 (by, in my case, disabling the arch testing repos and running pacman -Syuu)

well i guess it was my mesa version. i was using amber instead of the main build. thank you

m4rch3n1ng commented 1 day ago

i think it's still worth keeping this issue open, because hyprland is the only wayland compositor (that i tested) that is affected by this mesa issue, niri, sway, cosmic all work fine for me

vaxerski commented 1 day ago

unless mesa can point to a culprit being hyprland doing something wrong, I disagree.