On Monday 2 January 2006 15:43, Jiri Benc wrote: > On Mon, 2 Jan 2006 15:22:37 +0100, Ivo van Doorn wrote: > > I don't know if there is a list of known issues with the stack, > > but apparently the 80211.ko module doesn't load on a 64bit machine. > > My todo list currently starts with these items: > > - Don't copy ieee80211_tx_control and ieee80211_rx_status structures to > skb->cb.
This should indeed fix the 64bit problem. > - Get rid of tons of net_devices. Use only as few devices as needed > (i.e. in most cases just one). > - Add missing wireless extensions for WPA support. > - Allow ieee80211_hw->tx() to return NETDEV_TX_BUSY (even between > fragments). I was wondering about the ieee80211_hw->tx() method, perhaps I have missed something in the ieee80211 code, but MGMT frames are not indicated as a seperate field or queue right? RT2x00 has for example besides the TX ring the PRIO ring which is intended for MGMT frames to be send with a higher priority then regular TX. Should each packet be checked for being a management frame and put to the seperate queue, or should an attempt be made to make this PRIO ring a higher priority TX ring which will allow regular TX frames to be send with that ring as well. Another thing I notice was that fragments are handed 1 frame at a time, but making the device kick the TX-ring after each frame. Shouldn't for fragmented frames the ring be kicked after the last fragment has been given to the device? > - Easier handling of configuration requests (ieee80211_hw->config()). In the configuration the antenna_sel value is not really clear to me. The options are Diversity, Ant0 en Ant1 which is clear, but would this value be applied to TX or RX antenna? It might be Ralink specific but in those devices the RX and TX antenna can both be configured. > - Make the stack SMP safe. > - PLCP Length calculation. IvD
pgpta8aS3A6no.pgp
Description: PGP signature
