FS#33467 - [systemd] [dhcpcd] Predictable interface names are inconsistent

Attached to Project: Arch Linux
Opened by Candy (CandyAngel) - Sunday, 20 January 2013, 00:20 GMT
Last edited by Dave Reisner (falconindy) - Sunday, 27 January 2013, 22:39 GMT
Task Type Bug Report
Category Packages: Core
Status Closed
Assigned To Ronald van Haren (pressh)
Dave Reisner (falconindy)
Tom Gundersen (tomegun)
Architecture i686
Severity Low
Priority Normal
Reported Version
Due in Version Undecided
Due Date Undecided
Percent Complete 100%
Votes 0
Private No

Details

Description:

Recent changes to systemd/udev described here (http://www.freedesktop.org/wiki/Software/systemd/PredictableNetworkInterfaceNames) are intended to cause networking interfaces to get consistent naming.

However, depending on when they are added, their names are different. They seem to get the predicatable names only if they are present during boot. Added them later results in the unpredictable names (e.g. eth0, wlan0). This is a problem for mobile users who may plug/unplug adapters while the machine is booted.

Additional info:
* Motion Computing LE1600
* Fresh Arch installation using 2012-12 media
* systemd 194-4

Steps causing discovery:

1) Observe predictable name (wlp6s2)
2) rmmod/modprobe module for interface (e.g. ipw2200)
3) Observe non-predictable name (eth0)

Alternative steps during testing of bug:

1) Plug in adapter
2) Observe non-predictable name (wlan0)
3) Boot machine with USB wireless adapter
4) Observe predictable name (wlp0s29f7u2)
5) Unplug adapter
6) Replug adapter
7) Observe non-predictable name (wlan0)
This task depends upon

Closed by  Dave Reisner (falconindy)
Sunday, 27 January 2013, 22:39 GMT
Reason for closing:  Upstream
Additional comments about closing:  Not a packaging bug. Upstream needs to deal with this.
Comment by Candy (CandyAngel) - Sunday, 20 January 2013, 02:13 GMT
Closure request link does not address the problem that I can discern. I have raised this upstream here: https://bugs.freedesktop.org/show_bug.cgi?id=59610
Comment by Dave Reisner (falconindy) - Wednesday, 23 January 2013, 02:01 GMT
And upstream is the way to handle this since there's clearly no packaging bug to be fixed here (based on Kay's comments).

Loading...