Just the usual day started with some trouble. I logged into my Fedora 40 system (running on a Hyper-V Windows 11 host) Note: my GPU is Nvidia 3080ti All I got was Black screen! After few hours of problem solving, it seems the culprit is Wayland 24.1.0-1.fc40 I got the problem after a recent update (only 68 packages total, all except 1 was update) As soon as I downgraded this single package to 23.2.6-1.fc40.x86_64 using the sudo dnf downgrade command.. problem fixed! Reproducible: Always Steps to Reproduce: 1. Install 24.1.0-1.fc40 xorg-x11-server-Xwayland 2. Boot, login into ANY user account 3. Black Screen! Actual Results: After doing a dnf downgrade to the previous version, everything works fine Expected Results: expectation was that this is a stable release, so usually I shouldn't have run into any problem, given my hardware is not old either, nothing fancy.
Can you provide some more info on your setup? More precisely which Wayland compositor you use (GNOME, KDE, Sway, etc.), which driver (nouveau, NVIDIA proprietary driver, if so, what exact version) and also attach the journalctl logs of a user session which fails to show?
I'm using: KDE Plasma 6.0.5 QT: 6.7.1 Graphic Processor is listed as : 'llvmpipe' What should be my command for getting journalctl logs for Wayland ?
(In reply to ssabs from comment #2) > I'm using: > > KDE Plasma 6.0.5 > QT: 6.7.1 Just a quick note, Xwayland is a compatibility tool, it's an Xserver for legacy X11 clients which cannot run on Wayland natively, it is not the display server like Xorg was. With Wayland, the display server is the Wayland compositor. So Xwayland itself cannot cause the whole screen to go black at startup. However, if Xwayland fails to start for whatever reason, it is possible that it breaks KDE even with a Wayland session. > Graphic Processor is listed as : 'llvmpipe' Humm that does not seem right. Did you try to install the NVIDIA proprietary driver at some point in the past? What gives "cat /proc/cmdline" ? > What should be my command for getting journalctl logs for Wayland ? Wayland is just a set of protocol definitions, it's not a tool in itself. Let's capture everything, i.e. something like "journalctl -b0 | gzip -9 -c > journactl.gz" as the problem may lie anywhere.
1. Thank you for the quick note, it is valuable insight 2. cat /proc/cmdline - gives me this: BOOT_IMAGE=(hd0,gpt2)/vmlinuz-6.8.11-300.fc40.x86_64 root=UUID=e94c20b3-6ded-49d5-b9fa-b97c83d044d6 ro rootflags=subvol=root rhgb quiet 3. journalctl -b0 | gzip -9 -c > testfile.gz won't work for my case to provide the correct information you might be looking for, since I have booted multiple times since then (its running inside a Hyper-V VM, on a Windows Host). I tried journalctl -S 2024-05-28 -U 2024-05-31 | gzip -9 -c > testfile.gz But I'm getting journal file is truncated, ignoring file error. And no gzip file gets created in that case.
I'm not sure, I don't think I explicitly tried to install any video drivers. I just installed Fedora Workstation with defaults.
I managed to generate a log as root, here is the link to the gzip file: https://filetransfer.io/data-package/lYpElkAx#link
From the very first line in that log file: > May 28 23:02:27 fedora kwin_wayland[2075]: kwin_wayland_drm: Failed to create framebuffer: Invalid argument And that error show hundreds of thousands times: > $ zgrep "Failed to create framebuffer: Invalid argument" gziplog.gz | wc -l > 809325 That's what is causing the screen to appear black. I don't think this has anything to do with Xwayland, nor with NVIDIA actually (nowhere "nvidia" appears in that log): > $ zgrep -i nvidia gziplog.gz | wc -l > 0 It's using "hyperv_drm" : > May 31 14:55:12 fedora kernel: hyperv_drm 5620e0c7-8062-4dce-aeb7-520c7ef76171: [drm] Synthvid Version major 3, minor 5 > […] > May 31 14:55:12 fedora kernel: [drm] Initialized hyperv_drm 1.0.0 2020 for 5620e0c7-8062-4dce-aeb7-520c7ef76171 on minor 1 > May 31 14:55:12 fedora kernel: fbcon: Deferring console take-over > May 31 14:55:12 fedora kernel: hyperv_drm 5620e0c7-8062-4dce-aeb7-520c7ef76171: [drm] fb0: hyperv_drmdrmfb frame buffer device Unfortunately, I do not konw anything about Hyper-V or even Windows so I won't be able to help much there. I do see Xwayland crashing though: > May 31 16:39:50 fedora kwin_wayland_wrapper[2169]: (EE) > May 31 16:39:50 fedora kwin_wayland_wrapper[2169]: (EE) Backtrace: > May 31 16:39:50 fedora kwin_wayland_wrapper[2169]: (EE) 0: /usr/bin/Xwayland (0x55bf28ec0000+0x1660c2) [0x55bf290260c2] > May 31 16:39:50 fedora kwin_wayland_wrapper[2169]: (EE) 1: /usr/bin/Xwayland (0x55bf28ec0000+0x1661c5) [0x55bf290261c5] > May 31 16:39:50 fedora kwin_wayland_wrapper[2169]: (EE) 2: /lib64/libc.so.6 (0x7f4506013000+0x40710) [0x7f4506053710] > May 31 16:39:50 fedora kwin_wayland_wrapper[2169]: (EE) 3: /usr/bin/Xwayland (0x55bf28ec0000+0x41075) [0x55bf28f01075] > May 31 16:39:50 fedora kwin_wayland_wrapper[2169]: (EE) 4: /usr/bin/Xwayland (0x55bf28ec0000+0x3466f) [0x55bf28ef466f] > May 31 16:39:50 fedora kwin_wayland_wrapper[2169]: (EE) 5: /usr/bin/Xwayland (0x55bf28ec0000+0x1fa76) [0x55bf28edfa76] > May 31 16:39:50 fedora kwin_wayland_wrapper[2169]: (EE) 6: /lib64/libc.so.6 (0x7f4506013000+0x2a088) [0x7f450603d088] > May 31 16:39:50 fedora kwin_wayland_wrapper[2169]: (EE) 7: /lib64/libc.so.6 (__libc_start_main+0x8b) [0x7f450603d14b] > May 31 16:39:50 fedora kwin_wayland_wrapper[2169]: (EE) 8: /usr/bin/Xwayland (0x55bf28ec0000+0x224c5) [0x55bf28ee24c5] > May 31 16:39:50 fedora kwin_wayland_wrapper[2169]: (EE) > May 31 16:39:50 fedora kwin_wayland_wrapper[2169]: (EE) Segmentation fault at address 0x8 > May 31 16:39:50 fedora kwin_wayland_wrapper[2169]: (EE) > May 31 16:39:50 fedora kwin_wayland_wrapper[2169]: Fatal server error: > May 31 16:39:50 fedora kwin_wayland_wrapper[2169]: (EE) Caught signal 11 (Segmentation fault). Server aborting > May 31 16:39:50 fedora kwin_wayland_wrapper[2169]: (EE) And that's coming from GLAMOR; Stack trace of thread 2169: #0 0x00007f45060ab144 __pthread_kill_implementation (libc.so.6 + 0x98144) #1 0x00007f450605365e raise (libc.so.6 + 0x4065e) #2 0x00007f450603b902 abort (libc.so.6 + 0x28902) #3 0x000055bf2902f86a OsAbort (Xwayland + 0x16f86a) #4 0x000055bf2902faee FatalError (Xwayland + 0x16faee) #5 0x000055bf29026222 OsSigHandler (Xwayland + 0x166222) #6 0x00007f4506053710 __restore_rt (libc.so.6 + 0x40710) #7 0x000055bf28f01075 xwl_glamor_init (Xwayland + 0x41075) #8 0x000055bf28ef466f xwl_screen_init (Xwayland + 0x3466f) #9 0x000055bf28edfa76 main (Xwayland + 0x1fa76) #10 0x00007f450603d088 __libc_start_call_main (libc.so.6 + 0x2a088) #11 0x00007f450603d14b __libc_start_main@@GLIBC_2.34 (libc.so.6 + 0x2a14b) #12 0x000055bf28ee24c5 _start (Xwayland + 0x224c5) Can you please install the debuginfo packages for Xwayland 24.1.0 and generate a backtrace? > $ coredumpctl debug Xwayland > (gdb) bt full And post the output in this bug?
Humm, so if my maths is correct here, the crash occurs in hw/xwayland/xwayland-glamor-gbm.c at line 1517: 1516 main_dev = xwl_screen->default_feedback.main_dev; 1517 if (!(main_dev->available_nodes & (1 << DRM_NODE_RENDER))) { 1518 ErrorF("Main linux-dmabuf device has no render node\n"); 1519 return FALSE; 1520 } So most likely "main_dev" in NULL (that should not happen), hence the crash. Can you try that scratch build and report if that helps with Xwayland crashing (as reported in the logs)? https://koji.fedoraproject.org/koji/taskinfo?taskID=118641931 Note: This is a scratch build, meaning that it will be removed after a short while, so please make sure to download it while it lasts.
*** Bug 2291294 has been marked as a duplicate of this bug. ***
I just wanted to say what appears to be similar behavior occurs with Fedora 39 upgrading to 23.2.7-1.fc39 from 23.2.1-1.fc39. This is on physical hardware using i915 GPU kernel module Now that I've reverted, many features are now requesting I enter credentials (attach/detach monitor, request system suspend)
I see this fix has reached xwayland-24.1-backports https://gitlab.freedesktop.org/ofourdan/xserver/-/commit/fc785d83685677dddcfecc86db87028cacca155e Can you share anything on when an updated release would be cut upstream, and/or in Fedora? I seem to be hitting the issue in both Fedora 40 and OpenSuse Kalpa
*** Bug 2292853 has been marked as a duplicate of this bug. ***
(In reply to Patrick Lang from comment #12) > Can you share anything on when an updated release would be cut upstream, > and/or in Fedora? I seem to be hitting the issue in both Fedora 40 and > OpenSuse Kalpa I agree it;s taking too long, I'll file a new update downstream with these fixes backported. As for upstream, this is pending some additional fixes before I can make a new release.
FEDORA-2024-bd81b79a0b (xorg-x11-server-Xwayland-24.1.0-2.fc40) has been submitted as an update to Fedora 40. https://bodhi.fedoraproject.org/updates/FEDORA-2024-bd81b79a0b
FEDORA-2024-bd81b79a0b has been pushed to the Fedora 40 testing repository. Soon you'll be able to install the update with the following command: `sudo dnf upgrade --enablerepo=updates-testing --refresh --advisory=FEDORA-2024-bd81b79a0b` You can provide feedback for this update here: https://bodhi.fedoraproject.org/updates/FEDORA-2024-bd81b79a0b See also https://fedoraproject.org/wiki/QA:Updates_Testing for more information on how to test updates.
FEDORA-2024-bd81b79a0b (xorg-x11-server-Xwayland-24.1.0-2.fc40) has been pushed to the Fedora 40 stable repository. If problem still persists, please make note of it in this bug report.