On Fri, Aug 18, 2017 at 09:23:19AM -0400, Daniel Wilkins wrote: > On Fri, Aug 18, 2017 at 12:49:54PM +0200, Stefan Sperling wrote: > > We might have to blacklist AR5424 indeed. > > However, this hardware is becoming obsolete anyway so not very many > > people run into this issue anymore. > > > > Long ago there was an attempt to get it working but that never succeeded. > > See this long thread: http://marc.info/?t=126437919200002&r=1&w=2 > > > > > >Actual Fix: Unknown. > > > > I would welcome a diff that makes it work. > > Do you think you could give that go? > > > > I've never given any kernel or driver hacking a shot, but I guess I could > take a look. FreeBSD wiki seems to imply that it has full support for the > 5424, would it be a good idea to take a look there or have the driver > bases diverged enough that seeing how they handle the quirks not help me > out? >
FreeBSD is a good reference. You could also try Linux ath5k as a reference. Neither of these drivers shares ancestry with OpenBSD's driver, except for some high-level parts which were ported from FreeBSD/NetBSD very long ago. What we do know is that 5424 has never worked with OpenBSD's driver. And there are other known problems, for instance 11g mode does not work with OpenBSD's driver, only 11a and 11b modes do. This prevents device models which are 11g only (such as the AR5BXB63) from working. This driver is inherently complicated because it has to do a lot of things which are usually implement in firmware nowadays. So don't get discouraged if it takes a while to make progress ;-) You don't have to be exceptionally smart to work on this stuff, but an unhealthy degree of stubbornness helps.
