Login issue immediately after poweron and boot. I am using a vanilla system, with a very old graphics card. My system, with Wayland works 100% for all versions of Fedora up to and including 38. Fedora 40 works well, and all competitive distros (linux mint, Debian, Suse, and gasp,"ubunu" work well with Wayland BUT NOT FEDORA 39 except if If I log into the Fedora 39 nightly, and install same (I do a very clean installation), Fedora 39 Beta comes up with a distorted smear. I have to manually select the background image. Today (Oct 3), with updates, I am unable to post boot log into Fedora 39 Wayland directly. I have to first boot into xorg, then set the background image, and logout, and log in again, to get Fedora39 Wayland to function. I am using a vanilla nighly workstation release. It is my view that some graphics card initialization code is missing from the Wayland startup. I am trying to discover what it is. Why do I not have login issues with Fedora 38 and Rawhide (Fedora 40) but only with Fedora 39. Reproducible: Always Steps to Reproduce: 1.Use Fedora Media writer to create a USB bootable using a current Fedora Nightly 2.Boot the nightly after completing the sha256 check 3.Reboot and after logging in. Mixed issues as described above. Actual Results: Using the Oct 1 Nightly, On reboot, the image is smeared. I need to visit the setup page to refresh the background image. Success. Unsuccess follows: However, following today (Oct 3, ) beta update, the system will not allow a login to Fedora 39 Wayland directly. I must first log into Fedora using xorg, then exit and log in again with Wayland. Expected Results: Expect Fedora 39 beta to act as does all Fedora versions .38 with Wayland and with Rawhide (Fedora 40). Please review my opening comments. I use a fresh btrfs installation for / and /home (30gigs and 10gigs). Fresh is Fresh.
Today, Oct 4th, following sudo dnf update. From a power-off state (30-60 secs no power), to boot, login screen appears, and I am able to log in using Wayland. the image presented was one I selected at installation time. On logout, I am not able to log in again. I lose mouse and keyboard access, and I have the black screen of death. On reboot, after any (un)sessful login, I am unable to log into Fedora 39. Summary. Only from cold poweron am I able to log into Fedora 39 By the way, original installation was done using the "Everything.iso" of October 3.
October 5, Today, as in other days, one bug I am experiencing is gnome not correcly picking up the default image. There is an addressing error there, and I am getting bizzare results. I get half of the same image repeated. or a smearing of the top half of the screen with the bottom half normal of the default image. I am doing daily installations using the koji lists. My experience with workspace.iso has not been good, My experience with "everything.iso" seems to provide better results? I do not know how to define "better".
Well, another day of testing and non-release type corruption. I decided to compare Gnome startup to KDE on the same partitions (Dual boot). Gnome Wayland is able to run from cold power-on to login, but not able to be logged in if one is logged out and tries to relog. KDE Wayland seems to be much better, but firefox seems to be corrupted for both Gnome and KDE. (KDE installed via dnf group install "KDE .... " XORG version works 100% for both KDE and Gnome. There is another bug that is worrysome, here is a paste of an email On Thu, 2023-10-05 at 16:36 -0400, Neal Gompa wrote: > On Thu, Oct 5, 2023 at 4:03 AM Matthias Saou <matthias> wrote: > > > > Hi, > > > > I continued digging some more, and it turns out there is a one liner > > fix for the issue that has been merged upstream: > > > > https://github.com/dracutdevs/dracut/pull/2196 > > > > Would there be any hope to get this backported to be included in Fedora > > 39? The scope of the change is ridiculously tiny, limited to this > > single modules.d/90dmsquash-live/iso-scan.sh file which is exclusively > > used for this broken feature. > > > > It would be really great to get the final Fedora 39 iso working again. > > > > An update has been proposed: > https://bodhi.fedoraproject.org/updates/FEDORA-2023-f71507e281 The bug still needs to be accepted as an FE for this to make final, though. https://pagure.io/fedora-qa/blocker-review/issue/1382 -- Adam Williamson (he/him/his) Fedora QA Fedora Chat: @adamwill:fedora.im | Mastodon: @adamw https://www.happyassassin.net
Gnome 45 in problem KDE 39 OK.
October 7. Results of Everything.iso for Oct 7. Some progress When I boot Fedora 39, The background image is fully distorted and smeared. I must visit settings to select a clean background image. On logout and relog (Fedora 39 does not anymore crash) but the selected background image is smeared and distorted. On this also shows a secondary effect when the list of installed applications show all extra applications as having a default image. Fix the problem of faulty background image and that will probably fix everything else.
Created attachment 1992948 [details] partial log display Partial log display.
Oct 07 15:38:36 Lockwood kernel: pcpu-alloc: s225280 r8192 d28672 u262144 alloc=1*2097152 Oct 07 15:38:36 Lockwood kernel: pcpu-alloc: [0] 00 01 02 03 04 05 06 07 [0] 08 09 10 11 12 13 14 15 Oct 07 15:38:36 Lockwood kernel: pcpu-alloc: [0] 16 17 18 19 20 21 22 23 [0] 24 25 26 27 28 29 30 31 Oct 07 15:38:36 Lockwood kernel: Kernel command line: BOOT_IMAGE=(hd10,gpt12)/vmlinuz-6.5.5-300.fc39.x86_64 root=UUID=72740d33-c04c-488c-8437-6460a45b7c78 ro resume=UUID=a391dafa-4d2f-4071-9656-932f9215efc6 rhgb quiet Oct 07 15:38:36 Lockwood kernel: Unknown kernel command line parameters "rhgb BOOT_IMAGE=(hd10,gpt12)/vmlinuz-6.5.5-300.fc39.x86_64", will be passed to user space. Oct 07 15:38:36 Lockwood kernel: Dentry cache hash table entries: 2097152 (order: 12, 16777216 bytes, linear) Oct 07 15:38:36 Lockwood kernel: Inode-cache hash table entries: 1048576 (order: 11, 8388608 bytes, linear) Oct 07 15:38:36 Lockwood kernel: Fallback order for Node 0: 0 Oct 07 15:38:36 Lockwood kernel: Built 1 zonelists, mobility grouping on. Total pages: 4105599 Oct 07 15:38:36 Lockwood kernel: Policy zone: Normal Unknown kernel command on line 5
Did a fresh install with Oct 9 Everything.iso No improvement. Black screen of death
This is my setup. ----------- # System Details Report --- ## Report details - **Date generated:** 2023-10-09 15:06:41 ## Hardware Information: - **Hardware Model:** ASUSTeK COMPUTER INC. PRIME X470-PRO - **Memory:** 16.0 GiB - **Processor:** AMD Ryzen™ 7 2700X × 16 - **Graphics:** NVA8 - **Disk Capacity:** 5.6 TB ## Software Information: - **Firmware Version:** 6042 - **OS Name:** Fedora Linux 39 (Workstation Edition Prerelease) - **OS Build:** (null) - **OS Type:** 64-bit - **GNOME Version:** 45.0 - **Windowing System:** X11 - **Kernel Version:** Linux 6.5.6-300.fc39.x86_64 -------------- The graphics card is about 10 years old, and has been used for Fedora 20 through Fedora 38 (10 years of Fedora updates). The system works with Rawhide, same kernel, same graphics install, same gnome version.
Wayland is a set of protocol definitions and a small IPC library, as a rule of thumb, such bugs with the desktops environments are never Wayland. Based on the description and comments above, what you describe here is more of a bug in the compositor (GNOME Shell / mutter). That does nto seem to be a bug with the display drivers since you mentioned Xorg and KDE on Wayland both work fine. I am therefore moving this bug to GNOME Shell. Could you please also attach the unaltered journalctl logs from the boot sequence and desktop session (not just attachment 1992948 [details] which is almost empty and contains not meaningful data)?
Created attachment 1993357 [details] ascii output of journalctl -xk I ran sudo journalctl -xk If you need a more comprehensive journal please advise.
That's the kernel journal; what would be useful is also the system journal, i.e. without -k. Also see the manual about -x ("Note: when attaching journalctl output to bug reports, please do not use -x.").
Created attachment 1993530 [details] journalctl >oct11
The logs contain this: Oct 10 22:37:19 Lockwood gnome-shell[6920]: Connection to xwayland lost Oct 10 22:37:19 Lockwood gnome-shell[6920]: Xwayland terminated, exiting since it was mandatory Does `coredumpctl list` contain either gnome-shell or Xwayland?
Just before that, there was those messages in the log: Oct 10 22:53:04 Lockwood kernel: nouveau 0000:09:00.0: fifo: DMA_PUSHER - ch 7 [gnome-shell[7435]] get 000041f150 put 000041f15c ib_get 0000000f ib_put 00000010 state 40000004 (err: INVALID_MTHD) push 00400040 Oct 10 22:53:04 Lockwood kernel: nouveau 0000:09:00.0: AMD-Vi: Event logged [IO_PAGE_FAULT domain=0x0010 address=0xfbd000e0 flags=0x0000] Oct 10 22:53:04 Lockwood kernel: nouveau 0000:09:00.0: fifo: CACHE_ERROR - ch 7 [gnome-shell[7435]] subc 0 mthd 0000 data 04380000 Oct 10 22:53:04 Lockwood kernel: nouveau 0000:09:00.0: AMD-Vi: Event logged [IO_PAGE_FAULT domain=0x0010 address=0xfbd00140 flags=0x0000] Oct 10 22:53:04 Lockwood kernel: nouveau 0000:09:00.0: fifo: DMA_PUSHER - ch 7 [gnome-shell[7435]] get 000041f170 put 000041f238 ib_get 00000013 ib_put 00000014 state 80000024 (err: INVALID_CMD) push 00400040 Oct 10 22:53:04 Lockwood kernel: nouveau 0000:09:00.0: AMD-Vi: Event logged [IO_PAGE_FAULT domain=0x0010 address=0xfbd00160 flags=0x0000] Oct 10 22:53:04 Lockwood kernel: nouveau 0000:09:00.0: fifo: DMA_PUSHER - ch 7 [gnome-shell[7435]] get 000041f238 put 000041f24c ib_get 00000015 ib_put 00000016 state 80000024 (err: INVALID_CMD) push 00400040 Oct 10 22:53:04 Lockwood kernel: nouveau 0000:09:00.0: AMD-Vi: Event logged [IO_PAGE_FAULT domain=0x0010 address=0xfbd00220 flags=0x0000] Oct 10 22:53:04 Lockwood kernel: nouveau 0000:09:00.0: fifo: DMA_PUSHER - ch 7 [gnome-shell[7435]] get 000041f2a8 put 000041f2ac ib_get 00000017 ib_put 00000018 state 80006f04 (err: INVALID_CMD) push 00400040 Oct 10 22:53:04 Lockwood kernel: nouveau 0000:09:00.0: AMD-Vi: Event logged [IO_PAGE_FAULT domain=0x0010 address=0xfbd00240 flags=0x0000] Oct 10 22:53:04 Lockwood kernel: nouveau 0000:09:00.0: AMD-Vi: Event logged [IO_PAGE_FAULT domain=0x0010 address=0xfbd002a0 flags=0x0000] Oct 10 22:53:04 Lockwood kernel: nouveau 0000:09:00.0: fifo: CACHE_ERROR - ch 7 [gnome-shell[7435]] subc 0 mthd 0400 data 00000000 Oct 10 22:53:04 Lockwood kernel: nouveau 0000:09:00.0: AMD-Vi: Event logged [IO_PAGE_FAULT domain=0x0010 address=0xfbd00700 flags=0x0000] Oct 10 22:53:04 Lockwood kernel: nouveau 0000:09:00.0: AMD-Vi: Event logged [IO_PAGE_FAULT domain=0x0010 address=0xfbd00b00 flags=0x0000] Oct 10 22:53:04 Lockwood kernel: nouveau 0000:09:00.0: fifo: CACHE_ERROR - ch 7 [gnome-shell[7435]] subc 0 mthd 0404 data 00000001 Oct 10 22:53:05 Lockwood kernel: nouveau 0000:09:00.0: AMD-Vi: Event logged [IO_PAGE_FAULT domain=0x0010 address=0xfbd00f80 flags=0x0000] Oct 10 22:53:05 Lockwood kernel: nouveau 0000:09:00.0: fifo: CACHE_ERROR - ch 7 [gnome-shell[7435]] subc 0 mthd 0408 data 00446974 [...] Oct 10 22:53:05 Lockwood kernel: nouveau 0000:09:00.0: fifo: CACHE_ERROR - ch 7 [gnome-shell[7435]] subc 0 mthd 040c data 00002800 Oct 10 22:53:05 Lockwood kernel: nouveau 0000:09:00.0: fifo: CACHE_ERROR - ch 7 [gnome-shell[7435]] subc 0 mthd 0410 data 0000a340 Oct 10 22:53:05 Lockwood kernel: nouveau 0000:09:00.0: fifo: CACHE_ERROR - ch 7 [gnome-shell[7435]] subc 0 mthd 0414 data 00002000 Oct 10 22:53:05 Lockwood kernel: nouveau 0000:09:00.0: fifo: CACHE_ERROR - ch 7 [gnome-shell[7435]] subc 0 mthd 0418 data 00422a10 Oct 10 22:53:05 Lockwood kernel: nouveau 0000:09:00.0: fifo: CACHE_ERROR - ch 7 [gnome-shell[7435]] subc 0 mthd 041c data 00005800 Oct 10 22:53:05 Lockwood kernel: nouveau 0000:09:00.0: fifo: CACHE_ERROR - ch 7 [gnome-shell[7435]] subc 0 mthd 0420 data 000076a0 Oct 10 22:53:05 Lockwood kernel: nouveau 0000:09:00.0: fifo: CACHE_ERROR - ch 7 [gnome-shell[7435]] subc 0 mthd 0424 data 00002000 Oct 10 22:53:05 Lockwood kernel: nouveau 0000:09:00.0: fifo: CACHE_ERROR - ch 7 [gnome-shell[7435]] subc 0 mthd 0428 data 00446974 Oct 10 22:53:05 Lockwood kernel: nouveau 0000:09:00.0: fifo: CACHE_ERROR - ch 7 [gnome-shell[7435]] subc 0 mthd 042c data 00002800 Oct 10 22:53:05 Lockwood kernel: nouveau 0000:09:00.0: fifo: CACHE_ERROR - ch 7 [gnome-shell[7435]] subc 0 mthd 0430 data 0000a340 Oct 10 22:53:05 Lockwood kernel: nouveau 0000:09:00.0: fifo: CACHE_ERROR - ch 7 [gnome-shell[7435]] subc 0 mthd 0434 data 00002000 Oct 10 22:53:05 Lockwood kernel: nouveau 0000:09:00.0: fifo: CACHE_ERROR - ch 7 [gnome-shell[7435]] subc 0 mthd 0438 data 00422a10 Oct 10 22:53:05 Lockwood kernel: nouveau 0000:09:00.0: fifo: CACHE_ERROR - ch 7 [gnome-shell[7435]] subc 0 mthd 043c data 00005800 Oct 10 22:53:05 Lockwood kernel: nouveau 0000:09:00.0: fifo: CACHE_ERROR - ch 7 [gnome-shell[7435]] subc 0 mthd 0440 data 000076a0 Oct 10 22:53:05 Lockwood kernel: nouveau 0000:09:00.0: fifo: CACHE_ERROR - ch 7 [gnome-shell[7435]] subc 0 mthd 0444 data 00002000 Oct 10 22:53:05 Lockwood kernel: nouveau 0000:09:00.0: fifo: CACHE_ERROR - ch 7 [gnome-shell[7435]] subc 0 mthd 0448 data 00086d74 Oct 10 22:53:05 Lockwood kernel: nouveau 0000:09:00.0: fifo: CACHE_ERROR - ch 7 [gnome-shell[7435]] subc 0 mthd 044c data 00000000 Oct 10 22:53:05 Lockwood kernel: nouveau 0000:09:00.0: fifo: CACHE_ERROR - ch 7 [gnome-shell[7435]] subc 0 mthd 0450 data 00000004 Oct 10 22:53:05 Lockwood kernel: nouveau 0000:09:00.0: fifo: CACHE_ERROR - ch 7 [gnome-shell[7435]] subc 0 mthd 0454 data 000475e0 Oct 10 22:53:05 Lockwood kernel: nouveau 0000:09:00.0: fifo: CACHE_ERROR - ch 7 [gnome-shell[7435]] subc 0 mthd 0458 data 00000000 Oct 10 22:53:05 Lockwood kernel: nouveau 0000:09:00.0: fifo: CACHE_ERROR - ch 7 [gnome-shell[7435]] subc 0 mthd 045c data 00107b00 Oct 10 22:53:05 Lockwood kernel: nouveau 0000:09:00.0: fifo: CACHE_ERROR - ch 7 [gnome-shell[7435]] subc 0 mthd 0460 data 00000000 Oct 10 22:53:05 Lockwood kernel: nouveau 0000:09:00.0: fifo: CACHE_ERROR - ch 7 [gnome-shell[7435]] subc 0 mthd 0464 data 00675000 Oct 10 22:53:05 Lockwood kernel: nouveau 0000:09:00.0: fifo: CACHE_ERROR - ch 7 [gnome-shell[7435]] subc 0 mthd 0468 data 00086d74 Oct 10 22:53:05 Lockwood kernel: nouveau 0000:09:00.0: fifo: CACHE_ERROR - ch 7 [gnome-shell[7435]] subc 0 mthd 046c data 00000000 Oct 10 22:53:05 Lockwood kernel: nouveau 0000:09:00.0: fifo: CACHE_ERROR - ch 7 [gnome-shell[7435]] subc 0 mthd 0470 data 00000004 Oct 10 22:53:05 Lockwood kernel: nouveau 0000:09:00.0: fifo: CACHE_ERROR - ch 7 [gnome-shell[7435]] subc 0 mthd 0474 data 000475e0 Oct 10 22:53:05 Lockwood kernel: nouveau 0000:09:00.0: fifo: CACHE_ERROR - ch 7 [gnome-shell[7435]] subc 0 mthd 0478 data 00000000 Oct 10 22:53:05 Lockwood kernel: nouveau 0000:09:00.0: fifo: CACHE_ERROR - ch 7 [gnome-shell[7435]] subc 0 mthd 047c data 00107b00 Oct 10 22:53:05 Lockwood kernel: nouveau 0000:09:00.0: fifo: CACHE_ERROR - ch 7 [gnome-shell[7435]] subc 0 mthd 0480 data 00000000 Oct 10 22:53:05 Lockwood kernel: nouveau 0000:09:00.0: fifo: CACHE_ERROR - ch 7 [gnome-shell[7435]] subc 0 mthd 0484 data 00675000 Oct 10 22:53:05 Lockwood kernel: nouveau 0000:09:00.0: fifo: CACHE_ERROR - ch 7 [gnome-shell[7435]] subc 0 mthd 0488 data 00000001 Oct 10 22:53:05 Lockwood kernel: nouveau 0000:09:00.0: fifo: CACHE_ERROR - ch 7 [gnome-shell[7435]] subc 0 mthd 048c data 00005002 Oct 10 22:53:05 Lockwood kernel: nouveau 0000:09:00.0: fifo: CACHE_ERROR - ch 7 [gnome-shell[7435]] subc 0 mthd 0490 data 00107b00 Oct 10 22:53:05 Lockwood kernel: nouveau 0000:09:00.0: fifo: CACHE_ERROR - ch 7 [gnome-shell[7435]] subc 0 mthd 0494 data 00000000 Oct 10 22:53:05 Lockwood kernel: nouveau 0000:09:00.0: fifo: CACHE_ERROR - ch 7 [gnome-shell[7435]] subc 0 mthd 0498 data 00215000 Oct 10 22:53:05 Lockwood kernel: nouveau 0000:09:00.0: fifo: CACHE_ERROR - ch 7 [gnome-shell[7435]] subc 0 mthd 049c data 000002de Oct 10 22:53:05 Lockwood kernel: nouveau 0000:09:00.0: fifo: CACHE_ERROR - ch 7 [gnome-shell[7435]] subc 0 mthd 04a0 data 1000f010 Oct 10 22:53:05 Lockwood kernel: nouveau 0000:09:00.0: fifo: CACHE_ERROR - ch 7 [gnome-shell[7435]] subc 0 mthd 04a4 data 00000000 Oct 10 22:53:05 Lockwood kernel: nouveau 0000:09:00.0: fifo: CACHE_ERROR - ch 7 [gnome-shell[7435]] subc 0 mthd 04a8 data 00000001 Oct 10 22:53:05 Lockwood kernel: nouveau 0000:09:00.0: fifo: CACHE_ERROR - ch 7 [gnome-shell[7435]] subc 0 mthd 04ac data 00005002 Oct 10 22:53:05 Lockwood kernel: nouveau 0000:09:00.0: fifo: CACHE_ERROR - ch 7 [gnome-shell[7435]] subc 0 mthd 04b0 data 00107b00 Oct 10 22:53:05 Lockwood kernel: nouveau 0000:09:00.0: fifo: CACHE_ERROR - ch 7 [gnome-shell[7435]] subc 0 mthd 04b4 data 00000000 Oct 10 22:53:05 Lockwood kernel: nouveau 0000:09:00.0: fifo: CACHE_ERROR - ch 7 [gnome-shell[7435]] subc 0 mthd 04b8 data 00215000 Oct 10 22:53:05 Lockwood kernel: nouveau 0000:09:00.0: fifo: CACHE_ERROR - ch 7 [gnome-shell[7435]] subc 0 mthd 04bc data 000002de Oct 10 22:53:05 Lockwood kernel: nouveau 0000:09:00.0: fifo: CACHE_ERROR - ch 7 [gnome-shell[7435]] subc 0 mthd 04c0 data 1000f010 Oct 10 22:53:05 Lockwood kernel: nouveau 0000:09:00.0: fifo: CACHE_ERROR - ch 7 [gnome-shell[7435]] subc 0 mthd 04c4 data 00000000 Oct 10 22:53:05 Lockwood kernel: nouveau 0000:09:00.0: fifo: CACHE_ERROR - ch 7 [gnome-shell[7435]] subc 0 mthd 04c8 data 00040060 Oct 10 22:53:05 Lockwood kernel: nouveau 0000:09:00.0: fifo: CACHE_ERROR - ch 7 [gnome-shell[7435]] subc 0 mthd 04cc data beef0201 Oct 10 22:53:05 Lockwood kernel: nouveau 0000:09:00.0: fifo: CACHE_ERROR - ch 7 [gnome-shell[7435]] subc 0 mthd 04d0 data 00140010 Oct 10 22:53:05 Lockwood kernel: nouveau 0000:09:00.0: fifo: CACHE_ERROR - ch 7 [gnome-shell[7435]] subc 0 mthd 04d4 data 00000000 Oct 10 22:53:05 Lockwood kernel: nouveau 0000:09:00.0: fifo: CACHE_ERROR - ch 7 [gnome-shell[7435]] subc 0 mthd 04d8 data 00013030 Oct 10 22:53:05 Lockwood kernel: nouveau 0000:09:00.0: fifo: CACHE_ERROR - ch 7 [gnome-shell[7435]] subc 0 mthd 04dc data 000031cc Oct 10 22:53:05 Lockwood kernel: nouveau 0000:09:00.0: fifo: CACHE_ERROR - ch 7 [gnome-shell[7435]] subc 0 mthd 04e0 data 00000002 Oct 10 22:53:05 Lockwood kernel: nouveau 0000:09:00.0: fifo: CACHE_ERROR - ch 7 [gnome-shell[7435]] subc 0 mthd 04e4 data 00000000 Oct 10 22:53:05 Lockwood kernel: nouveau 0000:09:00.0: fifo: CACHE_ERROR - ch 7 [gnome-shell[7435]] subc 0 mthd 04e8 data 00446974 Oct 10 22:53:05 Lockwood kernel: nouveau 0000:09:00.0: fifo: CACHE_ERROR - ch 7 [gnome-shell[7435]] subc 0 mthd 04ec data 00002800 Oct 10 22:53:05 Lockwood kernel: nouveau 0000:09:00.0: fifo: CACHE_ERROR - ch 7 [gnome-shell[7435]] subc 0 mthd 04f0 data 0000a340 Oct 10 22:53:05 Lockwood kernel: nouveau 0000:09:00.0: fifo: CACHE_ERROR - ch 7 [gnome-shell[7435]] subc 0 mthd 04f4 data 00002000 Oct 10 22:53:05 Lockwood kernel: nouveau 0000:09:00.0: fifo: CACHE_ERROR - ch 7 [gnome-shell[7435]] subc 0 mthd 04f8 data 0044699c Oct 10 22:53:05 Lockwood kernel: nouveau 0000:09:00.0: fifo: CACHE_ERROR - ch 7 [gnome-shell[7435]] subc 0 mthd 04fc data 00001400 Oct 10 22:53:05 Lockwood kernel: nouveau 0000:09:00.0: fifo: CACHE_ERROR - ch 7 [gnome-shell[7435]] subc 0 mthd 0500 data 0000a360 Oct 10 22:53:05 Lockwood kernel: nouveau 0000:09:00.0: fifo: CACHE_ERROR - ch 7 [gnome-shell[7435]] subc 0 mthd 0504 data 00002000 Oct 10 22:53:05 Lockwood kernel: nouveau 0000:09:00.0: fifo: CACHE_ERROR - ch 7 [gnome-shell[7435]] subc 0 mthd 0508 data 00446974 Oct 10 22:53:05 Lockwood kernel: nouveau 0000:09:00.0: fifo: CACHE_ERROR - ch 7 [gnome-shell[7435]] subc 0 mthd 050c data 00002800 Oct 10 22:53:05 Lockwood kernel: nouveau 0000:09:00.0: fifo: CACHE_ERROR - ch 7 [gnome-shell[7435]] subc 0 mthd 0510 data 0000a340 Oct 10 22:53:05 Lockwood kernel: nouveau 0000:09:00.0: fifo: CACHE_ERROR - ch 7 [gnome-shell[7435]] subc 0 mthd 0514 data 00002000 Oct 10 22:53:05 Lockwood kernel: nouveau 0000:09:00.0: fifo: CACHE_ERROR - ch 7 [gnome-shell[7435]] subc 0 mthd 0518 data 0044699c Oct 10 22:53:05 Lockwood kernel: nouveau 0000:09:00.0: fifo: CACHE_ERROR - ch 7 [gnome-shell[7435]] subc 0 mthd 051c data 00001400 Oct 10 22:53:05 Lockwood kernel: nouveau 0000:09:00.0: fifo: CACHE_ERROR - ch 7 [gnome-shell[7435]] subc 0 mthd 0520 data 0000a360 Oct 10 22:53:05 Lockwood kernel: nouveau 0000:09:00.0: fifo: CACHE_ERROR - ch 7 [gnome-shell[7435]] subc 0 mthd 0524 data 00002000 Oct 10 22:53:05 Lockwood kernel: nouveau 0000:09:00.0: fifo: CACHE_ERROR - ch 7 [gnome-shell[7435]] subc 0 mthd 0528 data 00000001 Oct 10 22:53:05 Lockwood kernel: nouveau 0000:09:00.0: fifo: CACHE_ERROR - ch 7 [gnome-shell[7435]] subc 0 mthd 052c data 00005002 Oct 10 22:53:05 Lockwood kernel: nouveau 0000:09:00.0: fifo: DMA_PUSHER - ch 7 [gnome-shell[7435]] get 000041fbf0 put 0000420ac0 ib_get 00000019 ib_put 0000001a state 8000a320 (err: INVALID_CMD) push 00502011 Oct 10 22:53:05 Lockwood kernel: nouveau 0000:09:00.0: fifo: CACHE_ERROR - ch 7 [gnome-shell[7435]] subc 0 mthd 0530 data 00107b00 [...] Oct 10 22:53:05 Lockwood kernel: nouveau 0000:09:00.0: fifo: CACHE_ERROR - ch 7 [gnome-shell[7435]] subc 0 mthd 0cf8 data 00000001 Oct 10 22:53:05 Lockwood kernel: nouveau 0000:09:00.0: fifo: CACHE_ERROR - ch 7 [gnome-shell[7435]] subc 0 mthd 0cfc data 0004a314 Oct 10 22:53:05 Lockwood kernel: nouveau 0000:09:00.0: gr: DATA_ERROR 00000005 [INVALID_ENUM] Oct 10 22:53:05 Lockwood kernel: nouveau 0000:09:00.0: gr: 00100000 [] ch 7 [003f778000 gnome-shell[7435]] subc 5 class 5039 mthd 0220 data 00000101 Oct 10 22:53:05 Lockwood kernel: nouveau 0000:09:00.0: gr: DATA_ERROR 00000004 [INVALID_VALUE] Oct 10 22:53:05 Lockwood kernel: nouveau 0000:09:00.0: gr: 00100000 [] ch 7 [003f778000 gnome-shell[7435]] subc 5 class 5039 mthd 0224 data 046bc000 Something doesn't look quite right with nouveau here.
Tue 2023-10-10 21:48:48 EDT 11771 1000 1000 SIGABRT present /usr/libexec/xdg-document-portal 164.2K Wed 2023-10-11 16:23:17 EDT 9942 1000 1000 SIGSEGV present /usr/bin/xfce4-terminal 1.6M We both need a clean Fedora 39 to debug. Would you like me to do a fresh installation of Fedora 39, gnome, and log into it with Wayland? I can also just reboot, and try the coredumpctl list again. By the way, up until October 10, 2023 Fedora40 Rawhide was perfect. By the way, Yesterday, October 11, Following a reboot, after dnf update, Fedora40 rawhide shows the same issues as does Fedora 39 beta. I have wiped Fedora 40 and will install a clean Workstation.iso install. Bonjour Olivier. Si la correspodance vous convient, je pourrais continuer en française.
Some further information. My bug is against Fedora 39 beta but.... Yesterday's updates to Fedora 38 (current supported) has caused intermittant failures. vis: Fom power on to successful login, A OK, everything looks good BUT.... Mozilla comes up on blue. Fedora media writer comes up in black. Next, doing logon -> logoff -> logon (about 3 to 4 iterations results in black screen, loss of keyboard and loss of mouse. By the way, no issue with the above if selecting gnome on xorg. (applies to Fedora 38 and Fedora 39)
From another Fedora user named John. NVIDIA 9899GT Using F39, Wayland, with Ext4 System drive and Stratis data volume on my demo system. I only ran for a few minutes with the NVidia 9800GT and got a hard failure to login to Gnome Wayland ( lockup). My regular card on this system is a AMD radeon 6770HD card using the 'Radeon" driver. I just ran a contious set of 20 login/out activities without an issue. I have experienced no issues when running the Radeon device and driver. Here is a possible help I have 4 entries re fedora 38 (Same problem as 39) The first two are the fedora 38 updated with some faulty gnome library The next two are the Fedora 38 not updated to include those libraries. I can produce sha1sums of directories that can be used to determine what gnome interface has changed
Hi Florian, Using fedoraforum.org, I asked users to list their graphics card and the card size. Fedora 39 with xorg works 100% for my needs. There were some changes made that screwed up the video and login/logout sequences from Fedora 38 Wayland, Fedora 39 Wayland and yes Rawhide 40. My grub.cfg shows 4 version of Fedora 38 therein. The latter two boot in Wayland mode with zero issues. The first two will provide the "black screen of death", with full loss of keyboard and mouse. However, terminal mode is available to take some sha1sums, if that is of any help to identifying the culprit functions. NVA8 graphics card is 256meg, the 9899GT card is similar in size. Thank you for your investigation grandpa Leslie A Fedora user and Fedora bigot, My default distro beginning Fedora-core (pre-Fedora-01 some 20+years non-stop )
From another user: The Nvidia 9800 GT that I tested in an earlier post has 1GB memory. It failed to login at all on Gnome Wayland and suceeded on Xorg. Ditto for the NVA8 Fedora 38, 39 work well via xorg, I would like to test Rawhide(Fedora 40), but 40 will be without xorg. I am assuming that the delivery of Fedora 39 is a priority, taking away resources. Hopefully, this issue can be tackled very shortly thereafter. Other Fedora users suggest that the problem may be with the "nouveau driver". FYI, KDE works great if installed from the live ISO, but not so from Everything.iso. I will create a new bug report.
One negative gnome feedback. Same kernel level, same Wayland level, Same graphics cards, all works with Ubuntu 23.10 gnome KDE works well in Fedora 39. I will be testing the two versions with Fedora 40 (Rawhide).
Will the following help? With WAYLAND The issue appears to be some address miscalculation when, with Wayland, the background image is to be presented as first image after system boot at login. On first login, the background image is smeared. Attempting to change the image (from gnome list), causes a second image, also to show smeared, to overlay the existing image. On logout, One cannot log in a second or third time using Wayland. But I have no issues with using xorg. Item two. If I ignore the background image, the rest of gnome (clicking on an icon to launch a program, works as designed.
Leslie, this seems most likely to be a kernel issue. All those "different versions" on the boot screen are different kernel versions. They will be something like "6.5.6-200.fc38". Can you please tell us which kernel versions give which results?
Some very good news and some not. Today I installed dnf updates for Fedora 38 and Fedora 39 beta. The problem I was reporting with this bug report is solved for me. GREAT!! I would like to thank the individual(s) who worked to resolve the issue. I will continue testing, as I want to see if the fix also works for Rawhide (Fedora 40) I will close this bug report by 1 November 2023 The not good news Fedora 38 Rawhide on the screen is in half or somewhat less display intensity. It is somewhat (low priority) important to me as I do make use of some gnome extensions that have not as yet been ported to Fedora 39.
Hi Adam, the issues I reported and for which I am still having problems are obtained from within https://kojipkgs.fedoraproject.org/compose/branched/latest-Fedora-39/compose/ The problems started about October 5th, and have persisted to Oct 24. There were some minor improvements today with dnf updates. The everything.iso and the workstation.io. Failure with Wayland, but not with xorg. I also test and happy to say, without issue, KDE Wayland and KDE xorg. By the way, the same issues for Fedora 38 and Fedora 39. I am prepared to take dumps, provide logs etc, for anyone who needs technical information. As a reminder. The same hardware works with Ubuntu 23:10 and leap (1.55 and 1.56). I would suggest a test "Everything.iso" dated 23 Oct would suffice to use for debugging Wayland. Use Gnome.
vmlinuz-6.5.6-300.fc39.x86_64 causes me serious issues. Cant relog after logout, smeared background image. I can present a cellphone image of the screen, it that will help.
One more bit of information. The graphics card nvidia NVA8 has 256megs ram. The NV 9800 suppoedly has 1 gig. (comment 20 above)
In reponse to a note from Adam W, here is what i have and what I do for testing Fedora xx My system 2 x 1tb nvme + 4 SSDs ram 16gig 8 core amd cpu 2700x What I do is allocate, using the nvme 1 400megs for /boot/efi efi (fat32) 2 1 gig for /boot ext4 3 30 gigs for / btrfs 4 10 gigs for /home btrfs Preparation ----------------- I then download the latest Everything.iso, I do the appropriate check, and I use the Fedora Media Writer to load a USB. Installation ------------- I boot the USB created from the Everything.iso and run any precheck, and then anaconda. Using anaconda, I recreate the 4 partitions. With the Everything.iso I choose and create the workstation options, including development tools, printing support cargo, meld, vim, and software development support After Installation ================== Wayland Use =========== Two modes. Cold power on boot and warm reboot. On boot from a power-off state, During login, I check for Wayland before hitting enter Result: I will get a badly smeared background image, but much of what I want to use, from firefox or other apps may work. (Terminal mode works). After logout, the normal login screen appears, and if I try to log in a second time, the system is entirely locked up black screen, no keyboard keys recognized, and no mouse activity. If I hit system reset, (no computer power off) and I am doing the same boot steps as above, After accepting the login / password, I have the black screen of death. Xorg USE. However, If in lieu of Wayland, I choose xorg to replace Wayland, all items I would do with Wayland, xorg handles appropriately. Every login is functional and clean. All apps display correctly. KDE test ======== Much of Wayland mode appears almost ok, but black text on white background is fuzzy and almost not readable. ======== Xorg use is what we should have. Two different users, two different systems, I with NVA8, experience the same results. Our graphics cards are 6+ years old from windows 8 era.
Re comment 23. Prior to Oct 5, 2023, I had no issues with Wayland for F38,39 and Rawhide 40. Today Oct 26, I did a fresh install of the Everything.iso from Oct 24. Same problem. Here is the initramfs timestamp for that installation. initramfs-6.5.6-300.fc39.x86_64.img The kernel: vmlinuz-6.5.6-300.fc39.x86_64 Disappointedly, I had kept 3 versions of F38, with the latter two being with the working kernel. That kernel that worked may have been created around Sept 28, 2023 (using the "r 38 img" rescue date. Time and dnf updates killed their functionality. I am able to use xorg. Fedora 40 is 6 months away, So, do not let my issue and that of two others, not by a showstopper. (One user removed some files that he called Plymouthxxx. I do not know if they are relevant to this problem. His system symptoms were the similar to what I was experiencing with F38,F39.
Hint about this bug. A conjecture The graphics card is of smaller size 256megs When the login screen is presented, it appears that the video card buffer is overrun by gnome.The screen is smeared. If I visit gnome settings to change the background, that happens properly, no overrun. More modern cards are 1gig in size. Second comment All the gnome applications work as described. Reminder xorg with this card works 100% Other Info. Bios is up-to-date
F39 KDE with Wayland=100% F39 Gnome with Wayland=smeared poweron login provides background image smeared, and believe graphics output of gnome background image overruns graphics card. reboot login with Wayland = lockup. black screen, no mouse, no keyboard only works with xorg Since xorg is avail for F39, I can continue with Fedora. For 40, no xorg, I will switch to KDE for F40 Please try to repair same before Fed40
Here is an update with betas. The first Workspace beta for F39 failed, no itramfs and checksum failure The first Beta for KDE loads, with Wayland, lettering is faded and abscure for text within panels. Please note: I have been a bigoted Fedora user since version 0.1 (20 years as my only Daily Distro), I am not imagining the issues. If there are no fixes, I will be forced to switch. All other distros I tested that provide full Wayland support are functioning without issues, and snappily. I fix the gnome-Wayland bridge and the new one with KDE beta released Oct 24 have issues. Do you need cellphone photos?
Fedora-KDE-Live-x86_64-39-20231019.n.0.iso works well Fedora-KDE-Live-x86_64-39_Beta-1.1.iso does not.
What we need is clearer information, more logs, and less guesswork. Rather than trying different desktops with different images constantly, we need to know clearly what is the latest-dated thing you have that works reliably - without installing any updates, just booting several times and confirming everything works as expected - and what is the earliest-dated thing you have that does *not* work reliably under the same conditions. And then we need logs from both cases. You've thrown out a lot of version numbers and dates and it is confusing. If you have problems in both KDE and GNOME, there are only a limited amount of things that could cause that - mainly the kernel and mesa. The stable F39 kernel has not changed since October 9 - it has been kernel-6.5.6-300.fc39 ever since then. The kernel in updates-testing went to 6.5.7 on October 12, 6.5.8 on October 21, and 6.5.9 on October 26. mesa has been 23.2.1-2 in stable since October 9, with no newer version in updates-testing. Beta 1.1 is much older than 20231019.n.0. If 20231019.n.0 works and Beta 1.1 does not, that means something got fixed between Beta 1.1 and 20231019.n.0.
Hi Adam and others. I have some problems with your request (comment 34). For every beta and existing Fedora 38, the system boots without error, I watched ) boot messages scroll by (and viewed the log's), and there are no red flags. Once any Fedora linux presents the login prompt, I do the following: Terminal Modes ============== The direct Linux terminal mode works. Vis. Before logging in, I ctl-alt-fx and enter terminal mode. Log-in and usage and logout is normal. Wonderful, fantastic. I exit and return to the graphical login. If I log in as anything but Gnome-Wayland, there is zero, repeat, zero errors. All is well, I can use KDE, and other spins as x11. If I boot the new KDE and log in as KDE-Wayland, there are display issues (I have cellphone images). If I log out and return as x11, all is well. The issue is not the boot of current kernel. The issue is the interface of User software to Wayland, Of that, is there a log that is produced that follows my login as a Wayland user to the lockup problem? I have tested with Spins and Workstation, with Wayland/x11 and the issue is the userspace wayland interface. Where do I feel the problem arises? When the gnome45 Wayland moves to display the background image on the terminal. If it shows(no lockup), that screen is "garbled" with what appears is buffer overrun/wraparound. The background image can be made right by using gnome-settings. And if I logout or let the system time out the session, and attempt to relog -- black screen of death. Reboot necessary. One other note. a relog after reboot is directly to black screen of death, unless I power off, wait 10 seconds and reboot, Leslie Footnote. -rw-r--r--. 1 leslie leslie 2480287744 Oct 6 16:48 Fedora-KDE-Live-x86_64-39-20231006.n.0.iso Worked most cleanly with Wayland, no noticable garbling of text. -rw-rw-rw-. 1 leslie leslie 2480699392 Nov 1 20:23 Fedora-KDE-Live-x86_64-39-20231101.n.0.iso Cannot use some functions in Wayland mode. Text not readable. Footnote2. I run on real hardware. Footnote3 The other distros, Debian,Mint,Leap155/leap156,Ubuntu, all work well with Wayland. I believe it is Wayland user interface that gets invoked at login time. Believe me, I want to remain with Fedora. It has been my absolute desktop system for 19 years.
Some followup I was getting some nouveau error messages when I tried to use Wayland. I did not find a "nouveau.log", My messages appeared (3 repeats within one half second on the console Then a lockup. It might be "nouveau -3 ..." FWI, no other distro that I use to test my programs has any Wayland lockup
I am an end-user of Fedora (since 200x (more than 20 years devoted). The nouveau interface is flawed. I say that because.. a) I reinstalled my Fedora 38 ISO from my backups. My ISO is dated 1 August. It is the ISO that has an older version of Wayland interface b) I did not as yet run dnf update to bring the workstation ISO up to date. If I do so, I will be back to my complaint (see comment 1) . c) Wayland works fine, with this ISO. If I run sudo dnf update, I will be back to Fedora with only xorg access. d) Suggestions about what to do?
Comment about comment 37 To try to localize the problem, I indicated that I used an early Fedora 38 ISO (official release), and without kernel updates it works just fine. I blocked kernel updates and updated everything else, and the result works just fine with Wayland and xorg. I subsequently updated the kernel and I are back to comment 1, the original complaint. I have two grub boot entries, One with the older kernel (this one works) and the current kernel (broken Wayland). I would like to use https://kojipkgs.fedoraproject.org/compose/branched/ daily access to determine the date things went sour for Fedora 38,39,and beyond. If there is a log kept of the daily updates, that information will be most useful. Reminder. problem began "end september" for both 38 and 39 and 40 releases.
SOME GOOD NEWS. The Fedora Everything.iso for Fedora 40 was successfully installed. Wayland issues noted above no longer apply to Fedora 40 Rawhide ( Fedora-Everything-netinst-x86_64-Rawhide-20230929.n.0.iso has zero Wayland desktop/mouse lockup)
Sad news Wayland does not work with the released version of Fedora 39 (Everything.iso). Wayland with all other distros works flawlessly. Take the vanilla Fedora 38, and block kernel updates. It works with Wayland. Ditto for the Fedora 39 beta up to end-July (29 July). Works with Wayland. Some change with "nouveau" driver is what I suspect. PS. I am not alone with this issue. I am doing reporting on behalf of two others that I met on the Fedoraforum.org site. I am willing and ready to test any potential fixes you have. I am even willing to buy a new graphics card if that solves my Fedora39/Fedora40 Wayland issue.
I normally use the Everything.iso. I tried to use the Workstation.iso and watched the checksum succeed, but the installation of the live ISO fail gnome service -- Gnome Display manager displayed on the log, but coinciding with it -- no mouse, no keyboard, no screen activity.
This bug applies to Fedora 39 workstation and Fedora Everything.iso With modern processor but 256 graphics card, Workstation will not run because the installation is using Wayland, and therefore with my system, Fedora39 workstation cannot be installed With Fedora 39 Everything.iso, I can boot the new system if I first power-off the desktop for 10 seconds. I will be able to log in (to a faulty wayland display), once. On normal logout, I am not able to log in a second time (Wayland only). If I succeed the second time with Wayland, I will not be able to succeed a third time. xorg works 100% I can do everything with xorg. All gnome and other Linux functions work properly I have tested with KDE xfce, SWAY and gnome(xorg). I do know that Fedora 38, 39 and Rawhide 40 worked on my system until End September. Thereafter, an update to Gnome at that time has caused problems for systems with minimal graphics cards. I have stated in earlier comments that I will install any debug modules you need to resolve this issue. I am still offering to be available for this debugging.
The Fedora Wayland system cannot operate with a graphics card of size 256meg, which is the size of the graphics card installed in many systems owned by Linux users. However, other distributions do work with this sized graphics card, including Fedora 38 initial release, and Fedora 39, initial release. Then sometime at end september 2023, updates were installed, which caused memory overruns or other issues that resulted in a black screen, and unresponsive mouse/keyboard. But the competition's Linux versions worked A1. Today, I replaced the existing graphics card with a high-end graphics card (5 ports, multiple 8 gig ddr5 ram) and the problem of system lockup has disappeared. However, on YouTube, one of the presenter's of Systems, had problems similar to what I experienced. he has 2000+ viewers. I am closing my bug report. Adam, I tried my best to identify the issue. My efforts were not good enough. I had help from fedoraforum.org. I can mail the card I removed to a developer whom I hope would be able to debug the problem at a more detailed level than can I. On the plus side, With the replacement card, With Wayland issue resolved, I will be able to participate in the review of Fedora 40.
"The Fedora Wayland system cannot operate with a graphics card of size 256meg, which is the size of the graphics card installed in many systems owned by Linux users." The amount of memory likely has absolutely nothing to do with it at all. There is clearly a problem with this specific card, but you never really provided enough information to figure out what the problem actually is, and buried us in a flood of other information that just makes it impossible to do anything useful with this bug report due to the sheer volume of confusing messages. Sorry about that. Distributions all use fairly similar kernels, but update at different speeds. Fedora's kernel is updated quite aggressively. It is quite likely you will also see the same problem on other distributions as they reach a kernel version that has the bug, whatever it is.
The needinfo request[s] on this closed bug have been removed as they have been unresolved for 120 days