FS#51950 - [systemd] Crashes when partion name contains odd characters

Attached to Project: Arch Linux
Opened by Julian Knauer (jpk) - Thursday, 24 November 2016, 19:45 GMT
Last edited by Dave Reisner (falconindy) - Monday, 28 November 2016, 00:33 GMT
Task Type Bug Report
Category Upstream Bugs
Status Closed
Assigned To Dave Reisner (falconindy)
Architecture x86_64
Severity Medium
Priority Normal
Reported Version
Due in Version Undecided
Due Date Undecided
Percent Complete 100%
Votes 0
Private No

Details

Description:
When upgrading to the current version 232 (revision 1 to 3) of systemd. The PID1 will crash during update, when one or more partition names in /dev/disk/by-partlabel contains invalid characters. The systemctl command cannot be run anymore. The system will not be able to boot again until downgraded to version 231 (any revision)


Additional info:
+ Works fine until systemd-231-4
- Broken since systemd-232-1 until systemd-232-3

Steps to reproduce:
Name a NTFS partition "†††††††††††††††††††††††††††††††††††"

$ blkid |grep sda6
/dev/sda6: LABEL="Work" UUID="E63E05143E04E003" TYPE="ntfs" PARTLABEL="M-bM-^@M- M-bM-^@M- M-bM-^@M- M-bM-^@M- M-bM-^@M- M-bM-^@M- M-bM-^@M- M-bM-^@M- M-bM-^@M- M-bM-^@M- M-bM-^@M- M-bM-^@M- M-bM-^@M- M-bM-^@M- M-bM-^@M- M-bM-^@M- M-bM-^@M- M-bM-^@M- M-bM-^@M- M-bM-^@M- M-bM-^@M- M-bM-^@M- M-bM-^@M- M-bM-^@M- M-bM-^@M- M-bM-^@M- M-bM-^@M- M-bM-^@M- M-bM-^@M- M-bM-^@M- M-bM-^@M- M-bM-^@M- M-bM-^@M- M-bM-^@M- M-bM-^@M- " PARTUUID="6b316bcc-450c-11e5-985b-806e6f6e6963"

Run:
$ /usr/lib/systemd/systemd --test --system
Aborded (core dumped)
This task depends upon

Closed by  Dave Reisner (falconindy)
Monday, 28 November 2016, 00:33 GMT
Reason for closing:  Fixed
Additional comments about closing:  systemd-232-4
Comment by Doug Newgard (Scimmia) - Friday, 25 November 2016, 00:07 GMT
I can't see how this could be a packaging issue. You'll want to take it upstream.
Comment by Dave Reisner (falconindy) - Friday, 25 November 2016, 19:51 GMT
You'll need to provide the assertion failure at minimum, but ideally a backtrace from the abort. I can't reproduce this with the info given... my guess is there's invalid utf8 somewhere in that PARTLABEL.
Comment by Dave Reisner (falconindy) - Saturday, 26 November 2016, 17:43 GMT
Okay... managed to reproduce this. Should be easy enough to bisect.
Comment by Dave Reisner (falconindy) - Saturday, 26 November 2016, 18:47 GMT

Loading...