Arch Linux

Please read this before reporting a bug:
https://wiki.archlinux.org/title/Bug_reporting_guidelines

Do NOT report bugs when a package is just outdated, or it is in the AUR. Use the 'flag out of date' link on the package page, or the Mailing List.

REPEAT: Do NOT report bugs for outdated packages!
Tasklist

FS#79017 - [linux] sddm-greater crash

Attached to Project: Arch Linux
Opened by Alexandre ZANNI (noraj) - Saturday, 08 July 2023, 12:49 GMT
Last edited by Toolybird (Toolybird) - Tuesday, 11 July 2023, 07:31 GMT
Task Type Bug Report
Category Packages: Core
Status Waiting on Response   Reopened
Assigned To No-one
Architecture x86_64
Severity Low
Priority Normal
Reported Version
Due in Version Undecided
Due Date Undecided
Percent Complete 0%
Votes 1
Private No

Details

Description:


Since linux 6.4.0 (6.4.1 and 6.4.2 too):

1. boot to vanilla kernel
2. enter luks password
3. systemd starts
4. when reaching the graphical target, sddm-greater crashes, I no longer have any video signal, my displays are going into sleep mode

While using linux-lts 6.1.38

1. boot to lts kernel
2. enter luks password
3. systemd starts
4. sddm greater displays correctly

In journalctl I can see "systemd-coredump[1035]: [🡕] Process 986 (sddm-greeter) of user 975 dumped core."
Full stack trace: https://pastebin.com/raw/W4Qfyv3w

GPU

```
➜ lspci -k | grep -A 3 -E "(VGA|3D)"
0a:00.0 VGA compatible controller: Advanced Micro Devices, Inc. [AMD/ATI] Ellesmere [Radeon RX 470/480/570/570X/580/580X/590] (rev c7)
Subsystem: Micro-Star International Co., Ltd. [MSI] Radeon RX 480 Gaming X 8GB
Kernel driver in use: amdgpu
Kernel modules: amdgpu
```
This task depends upon

Comment by Toolybird (Toolybird) - Saturday, 08 July 2023, 22:34 GMT
> no longer have any video signal

You should be checking the journal for signs of GPU driver crashes.

> Full stack trace

Basically useless without debug symbols [1]. It's usually as simple as:

$ coredumpctl gdb (then answer y when it asks "Enable debuginfod for this session?")
(gdb) set logging enabled
(gdb) bt (or bt full)

Then post gdb.txt

[1] https://wiki.archlinux.org/title/Debugging/Getting_traces
Comment by Alexandre ZANNI (noraj) - Sunday, 09 July 2023, 19:25 GMT
The thing is that despite the crash, I see no error (error level) from the previous boot on the logs with `journalctl -b -1`.

However, I can see some warning level logs `journalctl -b -1 -p warning` (boot-1_warnings.log) for amdgpu, sddm-greeter, sddm-helper and sddm. (the 2h timing shift in the log timestamp is due to the hardware clock and the ntp sync in between during the init).

Note: the previous pastebin was actually from the LTS boot, not from the crashing vanilla boot.

Full last boot (journalctl -b -1) logs (boot-1_full.log).
   boot-1_warnings.log (4 KiB)
   boot-1_full.log (160.5 KiB)
   gdb.txt (4.1 KiB)
   coredump-gdb.log (13.6 KiB)
Comment by Alexandre ZANNI (noraj) - Sunday, 09 July 2023, 20:12 GMT
Sorry, I hadn't the debug symbols installed, here is the proper gdb trace.

   gdb.txt (14.2 KiB)
Comment by Toolybird (Toolybird) - Monday, 10 July 2023, 02:13 GMT
You are running some third party "Sweet" theme. This is not from Arch and therefore not supported...so this is definitely not an Arch packaging bug.

sddm[640]: Loading theme configuration from "/usr/share/sddm/themes/Sweet/theme.conf"
Comment by Alexandre ZANNI (noraj) - Tuesday, 11 July 2023, 07:31 GMT
This has nothing to do with the theme used. The proof is I set back the default breeze theme and the same crash occurs.
Comment by Toolybird (Toolybird) - Tuesday, 11 July 2023, 07:35 GMT
Well, this is a fairly popular pkg and nobody else appears to be having the same problem. You should probably report this upstream. Please ensure to give them a backtrace with debug symbols.
Comment by Alexandre ZANNI (noraj) - Tuesday, 11 July 2023, 09:17 GMT
Maybe this is not something between the kernel and SDDM but between the kernel and my graphic drivers for my specific hardware and the SDDM crash is just a consequence.

Upstream issue on SDDM tracker: https://github.com/sddm/sddm/issues/1761
   gdb.txt (14.8 KiB)
Comment by Alexandre ZANNI (noraj) - Tuesday, 11 July 2023, 10:06 GMT
The more I think, the more I start to believe the fact that I have no signal on linux 6.4 is unrelated to SDDM but related to the AMD graphics and maybe display port changes in linux 6.4 https://kernelnewbies.org/Linux_6.4. With the AMD graphics changes on 6.4 many people reports bad gaming performance but for my case it may as well end up with no signal at all. Anyway looks like an upstream issue. I have no clue hopw to troubleshoot that since I have no amdgpu error in journalctl.
Comment by loqs (loqs) - Tuesday, 11 July 2023, 10:16 GMT
Is the issue also present using the 6.5-rc1 from https://bbs.archlinux.org/viewtopic.php?pid=2108998#p2108998 you could also try 6.4.3 with amdgpu patches from https://bbs.archlinux.org/viewtopic.php?pid=2107227#p2107227
See also https://wiki.archlinux.org/title/Bisecting_bugs_with_Git
Comment by Taylor (shootguard44) - Friday, 14 July 2023, 21:36 GMT
I have am having a weird issue with SDDM crashing on 6.4+ as well, with amdgpu drivers. GDM has no issues. Not sure if it is specific to certain hardware configs (7900xt, 12700k). I also haven't tried changing sddm back to a x11 session yet. Just been fine using gdm for the time being. I can boot on older kernels before 6.4 as well.
Im new to the community and have been using arch for about 7 months now so it could be something I just messed up on my end. Your experience does seem extremely similar to mine, so I thought Id share.
Comment by Alexandre ZANNI (noraj) - Monday, 17 July 2023, 18:06 GMT
@Taylor: Unrelated to X11 or wayland, happens when both are predefined as default and anyway the crash occurs before opening a session.
Comment by loqs (loqs) - Wednesday, 19 July 2023, 20:26 GMT
Please try with linux 6.4.4. If the issue is still present please try 6.5-rc2 or rc3 if available. If that still has the issue please try [1] if the issue is still present please bisect [2] between 6.3 and 6.4 and report the issue upstream [3].

[1] https://gitlab.freedesktop.org/agd5f/linux
[2] https://wiki.archlinux.org/title/Kernel#Debugging_regressions
[3] https://gitlab.freedesktop.org/drm/amd/-/issues
Comment by Alexandre ZANNI (noraj) - Friday, 21 July 2023, 22:44 GMT
linux-6.4.4 and linux-zen-6.4.4 not working.
linux-mainline (6.5-rc2) not working either (kernel drm warning logs attached).
linux-drm-tip-git (6.5.r1200134.003f860e5577-1) and linux-drm-next-git (6.4.r1188951.6725f3322807-1) are not working either.
linux-amd-drm-next (6.4.2023.05.31-1) and linux-amd-staging-drm-next-git (6.2.r1160341.2514cae86fce-1) are not working. Note: having inux-amd-staging-drm-next-git not working while it is 6.2.x indicates that the crash must comes from an AMDGPU update that was introduces on 6.4.0.
   drm.log (16 KiB)
Comment by Alexandre ZANNI (noraj) - Saturday, 22 July 2023, 22:31 GMT
Since none of the kernel worked, I have to git bisect to report upstream.
If I understand correctly to bisect I'll have to re-build the kernel from git for each commit. Since building `linux-amd-staging-drm-next-git` took me an entire day, bisecting will take me an eternity. I know all unnecessary drivers (network, sound, printer, etc.) could be disabled using modprobed-db to speed up the build time but I don't have the skills to do that.
Comment by loqs (loqs) - Sunday, 23 July 2023, 15:57 GMT
$ git bisect start
status: waiting for both good and bad commits
$ git bisect bad v6.4
status: waiting for good commit(s), bad commit known
$ git bisect good v6.3
Bisecting: 8012 revisions left to test after this (roughly 13 steps)
[d42b1c47570eb2ed818dc3fe94b2678124af109d] Merge tag 'devicetree-for-6.4-1' of git://git.kernel.org/pub/scm/linux/kernel/git/robh/linux

https://drive.google.com/file/d/1LgdkT9cgo4msW7A9ws09rNLlvy3jj-8_/view?usp=sharing linux-6.3.r7998.gd42b1c47570e-1-x86_64.pkg.tar.zst
https://drive.google.com/file/d/1yjHu9Xsn5Be3u9f9Z-oXLZkQDElSNim2/view?usp=sharing linux-headers-6.3.r7998.gd42b1c47570e-1-x86_64.pkg.tar.zst
Comment by Alexandre ZANNI (noraj) - Sunday, 23 July 2023, 22:35 GMT
It doesn't work with linux-6.3.r7998.gd42b1c47570e.
   journalctl-kernel.log (13.5 KiB)
Comment by loqs (loqs) - Monday, 24 July 2023, 08:23 GMT
$ git bisect bad
Bisecting: 4005 revisions left to test after this (roughly 12 steps)
[088e0c188513b58a0056a488cf5b7df094a8a48a] Merge tag 'platform-drivers-x86-v6.4-1' of git://git.kernel.org/pub/scm/linux/kernel/git/pdx86/platform-drivers-x86

https://drive.google.com/file/d/1AwDtQfg6wNgNM5v44-q_-0T3eK1tnPU4/view?usp=sharing linux-6.3.r3992.g088e0c188513-1-x86_64.pkg.tar.zst
https://drive.google.com/file/d/1HjykYgs5C1A0AGGAbPd4fbFDXLR21euD/view?usp=sharing linux-headers-6.3.r3992.g088e0c188513-1-x86_64.pkg.tar.zst
Comment by Alexandre ZANNI (noraj) - Thursday, 27 July 2023, 21:47 GMT
6.4.6 not working.
linux-6.3.r3992.g088e0c188513 not working.
6.1.39-1-lts working.
Comment by loqs (loqs) - Friday, 28 July 2023, 00:54 GMT
$ git bisect bad
Bisecting: 2178 revisions left to test after this (roughly 11 steps)
[736b378b29d89c8c3567fa4b2e948be5568aebb8] Merge tag 'slab-for-6.4' of git://git.kernel.org/pub/scm/linux/kernel/git/vbabka/slab

https://drive.google.com/file/d/117YtgDAR0dVFyOnr3OG27lKR1TABhN59/view?usp=sharing linux-6.3.r1813.g736b378b29d8-1-x86_64.pkg.tar.zst
https://drive.google.com/file/d/1tzLQqSeJLInvCj5mn1hE6aDgwdXCuKT6/view?usp=sharing linux-headers-6.3.r1813.g736b378b29d8-1-x86_64.pkg.tar.zst
Comment by Alexandre ZANNI (noraj) - Saturday, 29 July 2023, 09:54 GMT
6.4.7 not working.
linux-6.3.r1813.g736b378b29d8 IS working!
Comment by loqs (loqs) - Saturday, 29 July 2023, 13:11 GMT
$ git bisect good
Bisecting: 1079 revisions left to test after this (roughly 10 steps)
[52b113e968be66b57f792b2e2a9b8b77f382bd5f] Merge tag 'drm-misc-next-2023-04-06' of git://anongit.freedesktop.org/drm/drm-misc into drm-next

https://drive.google.com/file/d/1edca4oXfpK8eaILq9FqtemtANMO2WaVM/view?usp=sharing linux-6.3rc4.r1099.g52b113e968be-1-x86_64.pkg.tar.zst
https://drive.google.com/file/d/1xf7AQFpz8SV3lRP091bPnADKdQ3TO5N8/view?usp=sharing linux-headers-6.3rc4.r1099.g52b113e968be-1-x86_64.pkg.tar.zst
Comment by Alexandre ZANNI (noraj) - Sunday, 30 July 2023, 21:38 GMT
linux-6.3rc4.r1099.g52b113e968be is not working.
Comment by loqs (loqs) - Monday, 31 July 2023, 11:24 GMT
$ git bisect bad
Bisecting: 539 revisions left to test after this (roughly 9 steps)
[c6265f5c2f502e442c4f339f121bedbc990c12e7] Merge tag 'drm-misc-next-2023-03-16' of git://anongit.freedesktop.org/drm/drm-misc into drm-next

https://drive.google.com/file/d/1b-rVBiw_xDE1QtfKU5J0rXB9YEA1WP5v/view?usp=sharing linux-6.3rc2.r559.gc6265f5c2f50-1-x86_64.pkg.tar.zst
https://drive.google.com/file/d/1AmbGjZq0q_h2x3F3RDJKVH6C91zRKni1/view?usp=sharing linux-headers-6.3rc2.r559.gc6265f5c2f50-1-x86_64.pkg.tar.zst
Comment by Alexandre ZANNI (noraj) - Monday, 07 August 2023, 16:54 GMT
linux-6.3rc2.r559.gc6265f5c2f50 is not working.
Comment by loqs (loqs) - Monday, 07 August 2023, 17:47 GMT
$ git bisect bad
Bisecting: 261 revisions left to test after this (roughly 8 steps)
[8bf6e20253b2d2b614f2c0b491f840e956fa6b05] Merge tag 'drm-intel-next-2023-03-07' of git://anongit.freedesktop.org/drm/drm-intel into drm-next

https://drive.google.com/file/d/1yyheaxz27qtQqLIonwg1ylBjctradA51/view?usp=sharing linux-6.3rc2.r297.g8bf6e20253b2-1-x86_64.pkg.tar.zst
https://drive.google.com/file/d/1N_FOiy9kJ1UYQFMypETkdMymFPBmM5qm/view?usp=sharing linux-headers-6.3rc2.r297.g8bf6e20253b2-1-x86_64.pkg.tar.zst

Loading...