Bug 138882

Summary:	NetworkManager unusable with atmel_cs
Product:	[Fedora] Fedora	Reporter:	Bastien Nocera <bnocera>
Component:	NetworkManager	Assignee:	Christopher Aillon <caillon>
Status:	CLOSED CANTFIX	QA Contact:
Severity:	medium	Docs Contact:
Priority:	medium
Version:	3
Target Milestone:	---
Target Release:	---
Hardware:	All
OS:	Linux
Whiteboard:
Fixed In Version:		Doc Type:	Bug Fix
Doc Text:		Story Points:	---
Clone Of:		Environment:
Last Closed:	2006-10-30 22:17:16 UTC	Type:	---
Regression:	---	Mount Type:	---
Documentation:	---	CRM:
Verified Versions:		Category:	---
oVirt Team:	---	RHEL 7.3 requirements from Atomic Host:
Cloudforms Team:	---	Target Upstream Version:
Embargoed:

Description Bastien Nocera 2004-11-11 19:45:18 UTC

Every time NetworkManager does a scan, the network drops.
See upstream bug:
http://bugme.osdl.org/show_bug.cgi?id=3338

Should there be a blacklist for such badly behaved drivers?

Comment 1 Dan Williams 2004-11-11 19:47:10 UTC

Hmm, I've got an atmel CS card (Belkin F5D6020) that seems to work
alright...  I'll have to test it more.

Comment 2 Dan Williams 2004-11-11 19:49:36 UTC

However, be aware that no network traffic usually flows out when the
card is scanning, that's expected and a limitation of most cards. 
Therefore we try not to scan so often...

Comment 3 Bastien Nocera 2005-01-17 17:33:20 UTC

I understand that. It's only more visible with the atmel_cs, you can see the
netstatus applet go to its "!" state, and traffic stopping for the best part of
2 seconds. On my laptop with an ipw2200 card, netstatus doesn't even have time
to switch to a different icon.

Comment 4 Dan Williams 2005-01-17 17:41:36 UTC

I did back the scan interval off to 20s (from 10s) but that of course doesn't
help when you're _in_ a scan.  Cards usually stay on a particular channel for a
certain period of time so they can intercept the beacons from the APs, and in
the case of the atmel driver it may be longer.  AFAIK the default for PrismII
cards (like Netgear MA111) using the wlan-ng driver is min=200ms max=250ms for
each channel for scanning, I don't know what other cards' intervals are.  For
the # of channels that the card has to support (up to 14 since AFAIK atmel_cs
doesn't support a or g modes) multiplied by 200ms you of course get a couple
second timeout.  I'm not quite sure why other cards cope with this better...

Atheros cards seem to freeze up traffic during the scan too, but Cisco and
prism54 seem to handle scanning + traffic better.

Comment 5 Dan Williams 2005-01-29 21:59:25 UTC

Bastien,

Does it seem to be any better with newer NetworkManager?  I've
introduced a backoff algorithm that does scanning later and later (max
60s interval though) if the list of networks doesn't really change on
each scan.  Drops back to 20s when the list changes.

Comment 6 John Thacker 2006-10-30 22:17:16 UTC

Closing per lack of response to previous request for information.
This bug was originally filed against a much earlier version of Fedora
Core, and significant changes have taken place since the last version
for which this bug is confirmed.

Note that FC3 and FC4 are supported by Fedora Legacy for security
fixes only.  Please install a still supported version and retest.  If
it still occurs on FC5 or FC6, please reopen and assign to the correct
version.  Otherwise, if this a security issue, please change the
product to Fedora Legacy.  Thanks, and we are sorry that we did not
get to this bug earlier.