You are not logged in.
Upgraded extra-testing/{mesa,vulkan-mesa-layers ,vulkan-radeon} to 1:24.3.1-1
Kitty (0.37.0-1) started immediately coredumping, even after a reboot.
Downgraded back to extra-testing/{mesa,vulkan-mesa-layers ,vulkan-radeon} back to 1:24.3.0-1, problem stopped.
I'll try to find the time today to run a debug on the coredump.
Last edited by merlock (2024-12-06 23:43:59)
Eenie meenie, chili beanie, the spirits are about to speak -- Bullwinkle J. Moose
It's a big club...and you ain't in it -- George Carlin
Registered Linux user #149839
perl -e 'print$i=pack(c5,(41*2),sqrt(7056),(unpack(c,H)-2),oct(115),10); '
Offline
This is not the case on my setup, but I also don't usually use kitty ... Could you post the output of "pacman -Qm" please? It might be that some foreign package is interfering.
Offline
a core dump is the result of an error, it is not an error itself. It's done strictly for debugging purposes. To have any idea of anything, we need the actual error to start with.
Offline
Similar here. alacritty 0.14.0-1 is producing core dumps without any other output, even after I rebuilt that package. mullvad-vpn-beta 2024.7-stable and 2024.8-stable also lost the GUI - the window is all black.
Downgrading mesa 1:24.3.0-1 resolved all issues, even with vulkan-{radeon,mesa-layers} at 1:24.3.1-1.
Extended renderer info (GLX_MESA_query_renderer):
Vendor: AMD (0x1002)
Device: AMD Radeon Pro WX 5100 Graphics (radeonsi, polaris10, LLVM 18.1.8, DRM 3.59, 6.11.9-zen1-1-zen) (0x67c7)
[ranguvar@khufu alacritty]$ alacritty
Segmentation fault (core dumped)
[ranguvar@khufu alacritty]$ coredumpctl info alacritty.
Last edited by Ranguvar (2024-12-05 19:13:20)
Offline
Last edited by Hanabishi (2024-12-05 19:38:11)
Offline
SOLUTION: Reinstalling KDE completely after update solved this problem.
Same problem here but I'm not on Nvidia, I'm on Radeon.
Plasma does not start correctly in Wayland but it dows in X11.
@QUL What exactly do you mean by "reinstalling completely" ? Can you list the packages?
Offline
I'm on AMD, yet still have the same problem. X11 works fine.
EDIT: did the reinstall. As in complete install of plasma meta package. Didn't work.
I'm assuming that downgrading either drivers or mesa might fix this issue, but I'll probably test it tomorrow.
Last edited by Primoz (2024-12-06 12:01:23)
Arch x86_64 ATI AMD APU KDE frameworks 5
---------------------------------
Whatever I do, I always end up with something horribly mis-configured.
Offline
<Same problem here but I'm not on Nvidia, I'm on Radeon> +1
X11 works with no problems
Last edited by dbaxps2024 (2024-12-06 11:56:53)
Offline
I'm on radeon mesa, reinstalling plasma entirely like this can help:
pacman -S plasma
But after logging in (which it took a little while) when I called "glxinfo -B" in a terminal, the command was hanging showing a possible problem with the latest mesa driver (24.3.1) as a workaround I rolled back my mesa packages to 24.2.7.
These are the packages I rolled back:
lib32-mesa-1:24.2.7-1-x86_64.pkg.tar.zst
lib32-vulkan-mesa-layers-1:24.2.7-1-x86_64.pkg.tar.zst
mesa-1:24.2.7-1-x86_64.pkg.tar.zst
opencl-rusticl-mesa-1:24.2.7-1-x86_64.pkg.tar.zst
vulkan-mesa-layers-1:24.2.7-1-x86_64.pkg.tar.zst
This seems to solve the problem for now.
Offline
I confirm, I installed the previous version of mesa and it works as a temporary fix
$ sudo pacman -U /var/cache/pacman/pkg/mesa-1\:24.2.7-1-x86_64.pkg.tar.zst
Greetings from Venezuela
Offline
On fresh Arch Plasma install or just as another option to downgrade mesa :-
$ git clone https://aur.archlinux.org/yay.git
$ cd yay
$ makepkg -si
$ which yay
$ yay -S downgrade
$ sudo downgrade mesa ( => 24.2.7.1 )
Offline
Also having this issue with KDE on AMD graphics, I managed to fix it by downgrading the mesa package to 24.3.0, so the issue is with 24.3.1 specifically.
Also got this error on dmesg:
[ 636.240892] QSGRenderThread[5356]: segfault at 0 ip 0000785e28ff8478 sp 0000785e13f476a0 error 4 in libgallium-24.3.1-arch1.1.so[3f478,785e28fd1000+1779000] likely on CPU 2 (core 2, socket 0)
[ 636.240910] Code: 0f 84 fa 00 00 00 41 83 fe 0c 0f 87 f0 00 00 00 48 8d 35 e7 d1 8d 01 44 89 f1 48 63 0c 8e 48 01 f1 3e ff e1 66 0f 1f 44 00 00 <48> 8b 07 8b 40 40 41 89 04 24 66 66 2e 0f 1f 84 00 00 00 00 00 0f
Offline
Which actual GPUs do the radeon people here have? There's an investigation in a bug that only appears on Polaris and earlier: https://gitlab.freedesktop.org/mesa/mesa/-/issues/12253
And by extension; https://bbs.archlinux.org/viewtopic.php?id=301550 -- since all of the radeon reports are highly unlikely to be related to the original nvidia issue (that sounded like an issue in shader generation, which can happen on first boot after a new Qt/KDE release) I'll merge you there. ... And move this topic to Upgrade issues
Last edited by V1del (2024-12-06 16:40:20)
Offline
Which actual GPUs do the radeon people here have? There's an investigation in a bug that only appears on Polaris and earlier: https://gitlab.freedesktop.org/mesa/mesa/-/issues/12253
And by extension; https://bbs.archlinux.org/viewtopic.php?id=301550 -- since all of the radeon reports are highly unlikely to be related to the original nvidia issue (that sounded like an issue in shader generation, which can happen on first boot after a new Qt/KDE release) I'll merge you there. ... And move this topic to Upgrade issues
I'm on Polaris, so it's probably related to the problem.
Offline
Which actual GPUs do the radeon people here have? There's an investigation in a bug that only appears on Polaris and earlier: https://gitlab.freedesktop.org/mesa/mesa/-/issues/12253
And by extension; https://bbs.archlinux.org/viewtopic.php?id=301550 -- since all of the radeon reports are highly unlikely to be related to the original nvidia issue (that sounded like an issue in shader generation, which can happen on first boot after a new Qt/KDE release) I'll merge you there. ... And move this topic to Upgrade issues
Yup, I'm on Polaris as well. RX 560.
Offline
Thanks for the help. KDE crashed entirely after restarting with an RX580. Reverting mesa to 1:24.2.7-1 fixed it.
Offline
mesa 1:24.3.1-2 has the offending commit reverted and is in [extra-testing]. Confirmed that it fixes the issue for me.
Big thanks to Hanabishi for reporting upstream.
Last edited by Ranguvar (2024-12-06 18:08:06)
Offline
Which actual GPUs do the radeon people here have?
Radeon PRO WX 5100 same problem. (Ellesmere)
Last edited by junakreiter (2024-12-06 20:42:57)
Offline
Confirm that extra-testing/mesa 1:24.3.1-2 fixed my problem.
RX-560 for me as well (on Sway); this did seem to be a wayland issue-was able to use xterm to do the downgrade.
Another big shout-out to @Hanabishi for the trace and bisect reported upstream!!
Marking as solved.
Last edited by merlock (2024-12-07 02:28:21)
Eenie meenie, chili beanie, the spirits are about to speak -- Bullwinkle J. Moose
It's a big club...and you ain't in it -- George Carlin
Registered Linux user #149839
perl -e 'print$i=pack(c5,(41*2),sqrt(7056),(unpack(c,H)-2),oct(115),10); '
Offline