svn commit: r286410 - in head: etc etc/rc.d sys/dev/ath sys/dev/ath/ath_rate/sample sys/dev/bwi sys/dev/bwn sys/dev/if_ndis sys/dev/ipw sys/dev/iwi sys/dev/iwn sys/dev/malo sys/dev/mwl sys/dev/ral ...

[Adrian Chadd]

Gleb,

At least for ath(4), TDMA doesn't work, ATH_DEBUG doesn't work,
IFF_DUMPPKTS() has an extra brace, and that's what I found in the last
15 minutes of this, compiling one of the many MIPS kernels that we
have in -HEAD for hardware that people are using. I left this up to
you to run to completion, and a single 'make universe' would've found
these issues. Instead, head is actively broken leading into a weekend.
There's at least one report of a hostapd problem, which indicates that
hostapd didn't get enough testing. I don't know if adhoc or monitor
mode got tested. I understand if mesh/tdma weren't tested - I'm slowly
chipping away at those.

I've just reverted the work. Let's get all the debug stuff more
thoroughly tested and get more exposure before we try to land this
again.

If there are any follow up issues with the reversion then please poke
me and I'll get it tested.

Thanks,



-adrian


On 7 August 2015 at 16:49, Adrian Chadd <adrian.chadd at gmail.com> wrote:
> Hi,
>
> Since I'm on holidays, and I really don't have time to tidy up after
> this commit:
>
> * everyone can commit to the wireless stuff to get this fixed up and working;
> * if hostap mode stays broken for more than 4 days, I'm going to back
> all the changes out.
>
> (And yes, gleb did run this stuff by me a few times and I'm okay with
> the general theory of this, but we'll see how it pans out. I'm totally
> fine with completely rolling it back if it's a failed experiment.)
>
>
>
> -adrian