ACK timeouts: done in firmware or software?

[Jouni Malinen]

On Tue, Dec 03, 2002 at 03:29:48AM -0500, Brian Capouch wrote:

> I'm not aware enough to know whether I should assume that my 
> drivers--HostAP and Agere's AP-1000 bins--currently only wait for a SIFS 
> period before giving up on an ACK.   If that's what it is could I change 
> it to length DIFS instead and thereby gain the benefit they posit?

ACK handling is quite lowlevel task in firmware/hardware. I'm not aware
of any way to configure this from the driver code for Prism2-based
cards. It would be possible to set the maximum retry count to zero so
that the firmware does not do any retransmissions and then retransmit
the frames from the driver code based on some criteria. However, it
might not be possible to get any signal about the possible ACK received
after the original timeout.

-- 
Jouni Malinen                                            PGP id EFC895FA