On Thu, 26 May 2022 21:49:17 +0200, johannes@xxxxxxxxxxxxxxxx wrote: > I'm really surprised to see this callback in a modern driver - wouldn't > you want to support some form of multi-channel operation? Even just > using the chanctx callbacks might make some of the DFS things you have > there easier? That is a good point. We picked the “old” channel API, which took less time to implement – the benefits of channel context were not clear enough during that time. > This feels ... odd. You really shouldn't have to look into the beacon to > figure out these things? > > And SGI etc. are per-STA rate control parameters anyway? Hmm. Information from this dynamic parsing is required for our driver and FW to function properly. > You have all this hardware/firmware and you implement this? Interesting > design choice. One that I'm sure you'll revisit for WiFi 7 ;-) Actually, the driver is doing that in cl_tx_handle_beacon_tim function as a result of TBTT interrupt. However, the next version of the RFC will be without that code – we are moving the routine to the FW side due to timing issues in the multiclient setup. Best regards, Viktor Barna
