Index | Thread | Search

From:
Stefan Sperling <stsp@stsp.name>
Subject:
Re: urtwn(4) firmware commands, what should get vs. what is written
To:
Mikolaj Kucharski <mikolaj@kucharski.name>
Cc:
tech@openbsd.org
Date:
Mon, 19 Oct 2026 21:06:22 +0200

Download raw body.

Thread
On Mon, Oct 19, 2026 at 06:42:38PM +0000, Mikolaj Kucharski wrote:
> With tcpdump on access point I see without my diff `1Mbit/s`, but with
> my diff I see `1Mbit/s  6Mbit/s  9Mbit/s  12Mbit/s  18Mbit/s`.

Rate-scaling being broken on R92C is a long-standing issue.
So what you did to the driver definitely has a good effect.
I'm too tired to properly review this diff right now, though.

Regarding device timeouts, those should be caused by missing
Tx completion notifications. Maybe our driver is still doing
something wrong in that regard.

Have you looked at FreeBSD's version? They have significantly improved
their forks of OpenBSD's urtwn and rtwn, including the addition of 11n
support. If I was trying to fix our driver, I would start by looking
for fixes which can be picked up from there.