On 1/28/07, Ulrich Kunitz <[EMAIL PROTECTED]> wrote:
> On 07-01-27 22:26 Jon Smirl wrote:
> > In order to learn more about how the zd1211 works I though I could try
> > implementing a few missing functions in the dscape zd1211 driver. The
> > easiest ones looked to be the statistics functions.
> >
> > get_stats needs this filled:
> >
> > struct ieee80211_low_level_stats {
> >         unsigned int dot11ACKFailureCount;
> >         unsigned int dot11RTSFailureCount;
> >         unsigned int dot11FCSErrorCount;
> >         unsigned int dot11RTSSuccessCount;
> > }
> >
> > get_tx_stats needs this filled, one per queue:
> >
> > struct ieee80211_tx_queue_stats_data {
> >         unsigned int len; /* num packets in queue */
> >         unsigned int limit; /* queue len (soft) limit */
> >         unsigned int count; /* total num frames sent */
> > };
> >
> > I've been poking around in the vendor driver amd zd1211rw and don't
> > see an obvious way to collect this data. For example I didn't see
> > where RTS success/failure is tracked. I also haven't located the exact
> > spot to separate FCS (CRC errors) from other classes of errors.
> We do/emulate ACK confirmation for packet tx in d80211, so it
> should be possible to estimate the error count. However it is not
> obvious to me, what the semantics is here. Are ACK failures for
> retransmissions also counted here?

I'll look around for more specific documentation.

> There is no simple way to count the RTS failures and successes. You could
> however get all the RTS packets and work from that. However I
> would regard this as waste of device-to-host bandwidth, which is
> shared of all USB devices on the same host adapter.

I'll look in the dscape code and see what they do with these errors.

> FCS errors are in struct rx_status (ZD_RX_CRC32_ERROR) if ZD_RX_ERROR is set.
> Queue stats might be possible, if we start to understand how
> queuing does work for the ZD1211B. I guess that the high byte of packet
> length in zd_ctrlset is used for this. But we don't know how
> exactly. Notify there is no support for queuing in ZD1211 devices.

If I understand things, these stats are what is needed to do rate
adaptation in the dscape stack.

> --
> Uli Kunitz

Jon Smirl

Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys - and earn cash
Zd1211-devs mailing list - http://zd1211.ath.cx/
Unsubscribe: https://lists.sourceforge.net/lists/listinfo/zd1211-devs

Reply via email to