On Wednesday 17 December 2008 17:01:09 Larry Finger wrote: > Christian Lamparter wrote: > > Ohh, it always gets better. ;) > > Anyway, I'll see what I can do tomorrow. > > > > In the meantime, I remembered that I put a heat sink on my chip a while ago, > > but I'm not sure if this had any effect here... But back in the summer, it allowed me > > fix the iperf stalls I had, but that was after days with 60Gb traffic... > > That would be a good idea; however, it is winter in my hemisphere (I don't care > what the calender says.), and the room is cool. In addition, my device is a Dell > Wireless 1450, which comes with a pretty hefty heat sink on its own. Your had already one? Perhaps, I got a very old one then. (Rev A01) > I'll throw printk's into every place I can think of and try to provide more info. maybe this printk could shed some light into the dark... (In station mode: "status" should be either 0(ACKed) or 1(not ACKed/timeout) and (2 [ device is sleeping - but we don't set the PSM yet, so this is not expected]) Regards, Chr
diff -Nurp p54-free-on-tx/p54common.c p54-eee/p54common.c --- p54-free-on-tx/p54common.c 2008-12-14 14:49:03.000000000 +0100 +++ p54-eee/p54common.c 2008-12-17 17:28:44.153744168 +0100 @@ -800,6 +800,12 @@ static void p54_rx_frame_sent(struct iee info->status.ack_signal = p54_rssi_to_dbm(dev, (int)payload->ack_rssi); skb_pull(entry, sizeof(*hdr) + pad + sizeof(*entry_data)); + if (payload->status) { + struct ieee80211_hdr *hd = (void*) skb->data; + printk(KERN_DEBUG "frame_control:%x status %d\n", + hd->frame_control, + payload->status); + } ieee80211_tx_status_irqsafe(dev, entry); goto out; }