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Do NOT report bugs when a package is just outdated, or it is in the AUR. Use the 'flag out of date' link on the package page, or the Mailing List.

REPEAT: Do NOT report bugs for outdated packages!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#54998 - [systemd] 234.11 - fix available for getpid() cache

Attached to Project: Arch Linux
Opened by Leryan (Leryan) - Tuesday, 01 August 2017, 07:35 GMT
Last edited by Christian Hesse (eworm) - Wednesday, 02 August 2017, 05:53 GMT
Task Type Bug Report
Category Upstream Bugs
Status Closed
Assigned To Dave Reisner (falconindy)
Christian Hesse (eworm)
Architecture All
Severity Very Low
Priority Normal
Reported Version
Due in Version Undecided
Due Date Undecided
Percent Complete 100%
Votes 0
Private No

Details

Description:

systemd has a fix for getpid() cache: https://github.com/systemd/systemd/pull/6413

Without that, we have a process /usr/lib/systemd/systemd --user that eats 6% CPU making getpid() calls. This leads to freezes and hangs on any subsequently started programs (shell, terminal, firefox…).

Additional info:

* systemd 234.11-1


Steps to reproduce:

lightdm + xorg + i3
This task depends upon

Closed by  Christian Hesse (eworm)
Wednesday, 02 August 2017, 05:53 GMT
Reason for closing:  Fixed
Additional comments about closing:  systemd 234.11-4
Comment by Christian Hesse (eworm) - Tuesday, 01 August 2017, 13:54 GMT
We would need upstream commits:

5c30a6d2b805ae9b5dd0ad003b9ee86b8965bc47
df0ff127758809a45105893772de76082d12a26d

Dave, your opinion?
Comment by Dave Reisner (falconindy) - Tuesday, 01 August 2017, 14:19 GMT
Sure, go for it... Worst case, we get some really strange bug reports to push upstream.

Edit: though, FWIW, I cannot replicate the high CPU usage. That sounds like a separate bug if the user manager is doing something other than sitting on epoll_wait() for most of its life.
Comment by Leryan (Leryan) - Tuesday, 01 August 2017, 16:55 GMT
@Dave: sometimes i have this bug, sometimes no.

In fact, i don't reproduce the bug at all *if* and only if i wait like 20 seconds before login with lightdm. Otherwise, since i'm on SSD, as i stay in front of the machine to login immediately, the bug triggers almost every time.

Very strange but reproducible quite easily.
Comment by Dave Reisner (falconindy) - Tuesday, 01 August 2017, 17:34 GMT
Right, and that seems to indicate an actual bug rather than an inefficiency.

Loading...