On Mon, 2008-10-20 at 16:04 +0800, Zhu Yi wrote: > On Mon, 2008-10-20 at 01:09 -0600, Johannes Berg wrote: > > Yes, I know that much, but can you tell us how exactly? > > I don't know the details either. Will follow up on this. Thanks. > > It would be good to actually report as accurate as possible (up to > > MAX_RATES) the retries the firmware will have done. We're going to > > export that information to userland in radiotap too. > > Sounds like a good idea. The actual number of retried rates should > depend on hardware/firmware. But '5' doesn't look like the possible MAX. Yeah, we don't have more space than for five entries. The question is whether the firmware will really go lower for each try, that would seem harmful for throughput in most cases. And if it doesn't ever use more than five different rates we can obviously add the count... If it does use more than five rates then we'll just have to live with imperfect reporting I guess. johannes
Attachment:
signature.asc
Description: This is a digitally signed message part
