On Thu, 18 Mar 2021, Stefan Sperling wrote:

> Our iwn(4) driver tries to be a bit smart and clears already acknowledged
> frames from Tx aggregation queues, even if those frames are still in the
> firmware's current block ack window.
> 
> Instead, the driver can simply clear frames before the firmware's BA window.
> This matches what the Linux driver does and makes our driver code simpler.
> 
> The firmware, not our driver/stack, controls the BA window and shifts it
> forwards. our driver and stack can simply follow along. The firmware should
> be smart enough to avoid needless retransmissions of frames which have
> already been acknowledged.
> 
> Also, our Tx rate control code relies on sequence numbers falling into
> the BA window, so skip rate control for frames before this window.
> 
> Please test this patch on iwn devices and check for regressions. Thanks!
> iwn's Tx aggregation code will eventually be ported to iwm(4), so testing
> this patch will benefit iwm(4) indirectly.

Tested on a t430s.  All seems good to me.  Also did a suspend/resume
changing APs without issues (I do this quite often at home).  APs used
were a Linksys E1200 which gives me around 35Mbps both directions (more
or less like before) and a Ubiquiti AC Pro which maxes at 52Mbps both
directions.

Cheers,

-- 
Paco Esteban.
0x5818130B8A6DBC03

Reply via email to