Bug 917038 – x11-misc/sddm-0.20.0-r1: sddm-greeter[1335]: memfd_create() called without MFD_EXEC or MFD_NOEXEC_SEAL set

• Status: UNCONFIRMED
• Product: Gentoo Linux

[Description lekto 2023-11-08 14:57:14 UTC]

Created attachment 874330 [details]
emerge --info

I don't know if this is important, but I noticed warning in dmesg:
> [   11.499990] sddm-greeter[1335]: memfd_create() called without MFD_EXEC or MFD_NOEXEC_SEAL set

This messages are appearing since 8th of October:
>lekto@lapek ~ $ journalctl | grep memfd_create
> Oct 08 08:38:24 localhost kernel: systemd[1]: memfd_create() called without MFD_EXEC or MFD_NOEXEC_SEAL set
> Oct 08 09:08:08 localhost kernel: systemd[1]: memfd_create() called without MFD_EXEC or MFD_NOEXEC_SEAL set
> Oct 09 15:47:50 localhost kernel: systemd[1]: memfd_create() called without MFD_EXEC or MFD_NOEXEC_SEAL set
> [...]
> Nov 06 15:44:50 lapek kernel: sddm-greeter[1331]: memfd_create() called without MFD_EXEC or MFD_NOEXEC_SEAL set
> Nov 07 15:54:10 lapek kernel: sddm-greeter[1315]: memfd_create() called without MFD_EXEC or MFD_NOEXEC_SEAL set
> Nov 08 15:44:52 lapek kernel: sddm-greeter[1335]: memfd_create() called without MFD_EXEC or MFD_NOEXEC_SEAL set

I don't think this is related to update of sddm, because it was updated 3rd of September. This are packaged that were updated day before first message:
>     Sat Oct  7 18:22:49 2023 >>> sys-kernel/linux-headers-6.5-r1
>     Sat Oct  7 18:23:11 2023 >>> dev-python/rpds-py-0.10.3
>     Sat Oct  7 18:23:24 2023 >>> dev-python/msgpack-1.0.7
>     Sat Oct  7 18:23:48 2023 >>> net-misc/yt-dlp-2023.10.07

[Comment 1 lekto 2023-11-08 16:35:07 UTC]

I'm running lxqt-1.3.0. On other machine, that is running OpenRC, I don't see this message, nor in dmesg, nor in /var/log/messages.

[Comment 2 Andreas Sturmlechner 2024-03-07 14:33:25 UTC]

What about 0.21.0?

[Comment 3 lekto 2024-03-07 15:25:58 UTC]

I'm still seeing it with 0.21.0.
> [   14.254734] sddm-greeter[1236]: memfd_create() called without MFD_EXEC or MFD_NOEXEC_SEAL set

[Comment 4 Andreas Sturmlechner 2024-03-07 15:33:56 UTC]

Could you add more information on the upstream issue, possibly including a backtrace with debug symbols?

[Comment 5 lekto 2024-03-07 17:24:25 UTC]

I can't provide a backtrace because there isn't coredump or at least I don't see any. That line appears in dmesg, but everything is working.

[Comment 6 Erik 2024-05-23 11:30:24 UTC]

(In reply to Andreas Sturmlechner from comment #2)
> What about 0.21.0?

Confirmed with 6.6.31-gentoo but the error is not in dmesg from 6.1.91-gentoo.

See also: https://bbs.archlinux.org/viewtopic.php?id=288830

[Comment 7 Erik 2024-05-23 19:06:23 UTC]

(In reply to Andreas Sturmlechner from comment #4)
> Could you add more information on the upstream issue, possibly including a
> backtrace with debug symbols?

How to make a backtrace of that dmesg message? Is it possible to set a breakpoint?

I can't find memfd_create in in the sddm source on github. Is it called via a library? If so, which?

Should sddm depend on kernel < 6.5?

[Comment 8 Erik 2024-05-23 20:57:47 UTC]

The change might be in Linux 6.3.

See also: https://lwn.net/Articles/918106/