ath(4) now defaults to 802.11n on GENERIC/i386 and GENERIC/amd64

[Justin Hibbits]

Hi,

On Fri, 15 Jun 2012 01:17:31 -0700
Adrian Chadd <adrian at freebsd.org> wrote:

> Hi,
> 
> I've flipped on ATH_ENABLE_11N and the interrupt mitigation in i386
> and amd64 GENERIC.
> 
> I'll flip it on on PPC when someone (chmee?) verifies that 802.11n
> works on PPC.

I did my duty as PPC guinea pig, and can happily say that unlike other
cards *cough*bw{i,n},wi*cough* I can actually use ath on my PowerBook.
AR5416, about as stable as I've seen so far.  I think you can flip it
on PPC now.
> 
> It's still delicate. I still don't know if RTS/CTS HT frame protection
> is working quite right on all chips. But all the basics are there
> (software TX aggregation, RX AMPDU reordering in net80211, BAR
> transmission, software queue pause and unpause, frame retransmission.)
> 
> Don't be surprised if your 802.11n mobile devices perform poorly as
> the power queue handling is very broken and will result in all kinds
> of weird traffic stalls if things are too aggressive. I'll look into
> fixing that soon.
> 
> If you have problems, please read the wiki article:
> http://wiki.freebsd.org/dev/ath(4) - specifically the bits where I
> tell you to compile with the debugging and diagnostic stuff in your
> -HEAD kernel, including the HAL/driver diagnostic APIs and tools.
> 
> Enjoy,
> 
> 
> 
> Adrian

If I can reproduce that crash I mentioned on IRC, I'll post what I can
about it.

- Justin