On Sat, 2009-03-14 at 09:30 +0100, Johannes Berg wrote: > On Fri, 2009-03-13 at 13:18 -0400, Dan Williams wrote: > > > It doesn't look like wpa_supplicant cares about the returned channel > > list from SIOCGIWRANGE. > > > > NM, however, uses the returned channel list to figure out a free > > supported adhoc frequency to use when creating an adhoc network (like > > for connection sharing). If it can't find an intersection between an > > internal "safe" frequency list and what the card supports (which it > > wouldn't be able to with this patch), it will simply use 2462 or 5180 > > MHz. > > > > Thus, I don't think you'll *break* anything specifically, but this patch > > would cause the default frequency for an IBSS created by NM to move > > from channel 1 to channel 11. > > In that case, let's just drop the patch, I have no problem with that. > > We just couldn't figure out why some channels were missing in iwlist > when they were in iw, but that's due to the 32 channel limit. Since most > people probably don't use this particular functionality, I guess it > doesn't really matter. Eventually NM should just use nl80211, or at > least try to, because there it gets more information like "cannot do > IBSS on channel XY", something that wext doesn't give you, but mac80211 > enforces. Yup; nl80211 is the way to go here. Or better yet, make wpa_supplicant card capabilities over D-Bus so NM doesn't have to touch the driver directly. Dan -- To unsubscribe from this list: send the line "unsubscribe linux-wireless" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html