#209: panic on 2.6.14.3 with r1347 & r1350
--------------------------------------+-------------------------------------
Reporter: [EMAIL PROTECTED] | Owner: mrenzmann
Type: defect | Status: assigned
Priority: major | Milestone: version 0.9.0 - move to
new codebase
Component: madwifi: other | Version: trunk
Resolution: | Keywords: patch
Patch_attached: 1 |
--------------------------------------+-------------------------------------
Changes (by mrenzmann):
* status: new => assigned
* owner: => mrenzmann
Comment:
It seems to me that the patch isn't correct. {{{IEEE80211_CHAN_ANYC}}} is
used upon initialisation until a better value can be determined. It seems
to be wrong for the callers to assume that {{{ieee80211_chan2mode}}} is
going to handle that somehow - rather the caller should IMO take care to
not call that function with an invalid value.
For now I added a KASSERT to catch such cases. Note that this will still
cause an oops (but this time it's a wanted behaviour rather than a bug).
I'll negotiate with Sam about that issue - FreeBSDs
[http://www.freebsd.org/cgi/cvsweb.cgi/src/sys/net80211/ieee80211.c?rev=1.27
&content-type=text/x-cvsweb-markup ieee80211_chan2mode] doesn't catch this
situation as well.
@seniorr: Thanks for tracking this issue down. It has already been tried
to convince you of the need of the dual licensing scheme (and also that
the current developers are not the people who could decide about a license
change). If you don't agree to the current situation then please don't
provide patches - it wastes your time and ours, and probably makes it
harder for us to fix the bug than if you hadn't submitted your patch.
Please note that this does NOT mean that your help isn't welcome at all.
It's just related to submitting patches that you are not willing to sign
off.
--
Ticket URL: <http://madwifi.org/ticket/209>
MadWifi <http://madwifi.org/>
Multiband Atheros Driver for Wireless Fidelity
