On Tue, Aug 25, 2009 at 4:38 PM, Luis R. Rodriguez<lrodriguez@xxxxxxxxxxx> wrote: > On Tue, Aug 25, 2009 at 12:22:55PM -0700, Bob Copeland wrote: >> On Tue, Aug 25, 2009 at 2:58 PM, Luis R. >> Rodriguez<lrodriguez@xxxxxxxxxxx> wrote: >> > OK how about the band info, think that's useful? >> >> Yeah, band is useful. We don't have to pretty print it though >> (unless you really feel like it), %d would work just as well. >> >> Also WARN_ON_ONCE might be a good idea. > > Alright, first let me see if I can reproduce after cleaning > curband crap a little. > >> >> which could indicate some >> >> race condition with flushing the rx queue on channel changes. >> > >> > I'll see if I can reproduce somehow. >> > >> >> I haven't yet seen a hw rate we didn't know about. >> > >> > So you've seen this lately as well? >> >> I saw it some time ago, but then changed the order of how the >> curchan/curband variables were set when we change channels and >> haven't seen since. But kerneloops says a lot of other people >> are still hitting it as well :( > > OK that's good to know too. > >> Hmm, ath5k_rx_stop probably wants an equivalent to txq_drainq in >> there somewhere. > > True. > > First thing I saw upon a quick review was we were relying on our own > curband for RX instead of the cfg80211 hw->conf band. Now granted > that *may* be updated properly, and it seems that may be the case, > but since do not tend to trust drivers here's a removal of all that > crap first. I'll test this first to see if I run into the same warning. > If this doesn't cure it then we are indeed processing old frames on RX > and need to flush 'em out prior to a channel change first. > > Good thing is after a few hours I can reproduce at least. Whah, the warn is still there. [ 7281.492622] ------------[ cut here ]------------ [ 7281.492655] WARNING: at drivers/net/wireless/ath/ath5k/base.c:1133 ath5k_tasklet_rx+0x57b/0x720 [ath5k]() [ 7281.492663] Hardware name: 7660A14 [ 7281.492668] invalid driver rate index mapped for hw_rix: 1b on freq: 5540 Luis -- 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
