On Tue, 2008-10-21 at 14:23 +0800, Zhu Yi wrote: > On Mon, 2008-10-20 at 22:07 -0600, Zhu Yi wrote: > > The max retry rates number for iwl4965 is 16. See iwl-commands.h for > > more details: > > > > a) Use the initial rate for only one entry. > > b) For each following entry, reduce the rate to next lower available > > rate, until reaching the lowest available rate. > > c) When reducing rate, also switch antenna selection. > > d) Once lowest available rate is reached, repeat this rate until > > rate table is filled (16 entries), switching antenna each entry. Thanks for the info. > Now I even wonder if abstract the common MRR info is possible. Yes, the > algorithm itself is hardware independent. But the one we use best > utilizes the hardware capability and it communicates with the hardware > frequently. I don't know about other hardware. But I highly doubt the > interface is so easy to be defined within IEEE80211_SKB_CB. Yeah, we definitely don't have enough space for 16 rates. If you want to properly report all that to userspace you'd probably have to add some radiotap information in the driver, but that's not currently possible... For other HW, all I know is Atheros (4 different rates with up to 16 tries each iirc), Broadcom (2 different rates, but we can rewrite ucode) and p54 (8 different rates, but no firmware influence) I'll clean up the three things you pointed out in this patch, I don't think we can achieve perfect reporting of what your hardware does at this time. johannes
Attachment:
signature.asc
Description: This is a digitally signed message part
