Bug 682337 - assumes all network devices are named ethX or trX
Summary: assumes all network devices are named ethX or trX
Keywords:
Status: CLOSED RAWHIDE
Alias: None
Product: Fedora
Classification: Fedora
Component: amtu
Version: 15
Hardware: Unspecified
OS: Unspecified
unspecified
unspecified
Target Milestone: ---
Assignee: Steve Grubb
QA Contact: Fedora Extras Quality Assurance
URL:
Whiteboard:
Depends On:
Blocks: 682334 682339
TreeView+ depends on / blocked
 
Reported: 2011-03-04 21:24 UTC by Bill Nottingham
Modified: 2014-03-17 03:26 UTC (History)
4 users (show)

Fixed In Version:
Doc Type: Bug Fix
Doc Text:
Clone Of:
: 682339 (view as bug list)
Environment:
Last Closed: 2012-05-14 15:49:41 UTC
Type: ---
Embargoed:


Attachments (Terms of Use)
Patch... just read all devices with proper addresses (949 bytes, patch)
2011-03-04 21:24 UTC, Bill Nottingham
no flags Details | Diff

Description Bill Nottingham 2011-03-04 21:24:52 UTC
Created attachment 482379 [details]
Patch... just read all devices with proper addresses

Description of problem:

Network devices can have arbitrary names, and due to http://fedoraproject.org/wiki/Features/ConsistentNetworkDeviceNaming, will have different names in Fedora 15.

amtu assumes all devices are ethX or trX.

Version-Release number of selected component (if applicable):

amtu-1.0.8

How reproducible:

Via source code inspection

Comment 1 Steve Grubb 2011-06-24 19:03:21 UTC
Not sure this patch is adequate. The system could have tunnels and bridges and other things that amtu should not touch. How do you identify just ethernet and tokenring interfaces in the new naming scheme?

Comment 2 Bill Nottingham 2011-06-24 19:42:19 UTC
What is amtu verifying about the adapter that makes a bridge or bonded adapter invalid?

Comment 3 Steve Grubb 2011-06-24 20:07:39 UTC
The code is just to check that the basic security assumptions are met, meaning you have lossless IO. There is a programming note that the test does not work for async network devices. So, I think the code looks for ethernet and tokenring knowing they are not async.

Comment 4 Bill Nottingham 2011-06-24 20:16:20 UTC
Probably keying off of the header type (/sys/class/net/<device>/type, see include/linux/if_arp.h for the list) is best then?

Comment 5 Steve Grubb 2012-05-14 15:49:41 UTC
Applied patch.


Note You need to log in before you can comment on or make changes to this bug.