Re: [CFT] if_ath updates - ar5416 (macbook pro, etc)

On 08/10/2010 03:58, Adrian Chadd wrote:
> On 10 August 2010 15:51, Adrian Chadd <adrian.chadd_at_gmail.com> wrote:
>> Hi everyone,
>>
>> I'm committing some updates to the if_ath and ath_hal code.
>>
>> I've just committed updates to the AR5416 register setup values. I'd
>> appreciate some testing by AR5416 users - eg macbook pro users - to
>> ensure this hasn't broken functionality.
> 
> I've also started merging in general if_ath changes; so please test
> the ath code in -head and let me know if I've introduced any
> regressions.
> 

I have applied the following (r211136) to my local stable/8 branch and
has proven to be an improvement with no drawbacks.

diff -r 7d9664d4ac99 sys/dev/ath/if_ath.c
--- a/sys/dev/ath/if_ath.c      Tue Aug 10 16:41:55 2010 -0400
+++ b/sys/dev/ath/if_ath.c      Tue Aug 10 19:39:25 2010 -0400
_at__at_ -5418,6 +5418,7 _at__at_
        longCal = (ticks - sc->sc_lastlongcal >= ath_longcalinterval*hz);
        if (longCal) {
                sc->sc_stats.ast_per_cal++;
+               sc->sc_lastlongcal = ticks;
                if (ath_hal_getrfgain(ah) == HAL_RFGAIN_NEED_CHANGE) {
                        /*
                         * Rfgain is out of bounds, reset the chip
_at__at_ -5466,7 +5467,6 _at__at_
                        nextcal *= 10;
        } else {
                nextcal = ath_longcalinterval*hz;
-               sc->sc_lastlongcal = ticks;
                if (sc->sc_lastcalreset == 0)
                        sc->sc_lastcalreset = sc->sc_lastlongcal;
                else if (ticks - sc->sc_lastcalreset >=
ath_resetcalinterval*hz)

As for the rest I can not vouch for unless you give me a specific patch
to test but I am willing to merge from head to my local tree at any
given point in time if you can give some revisions you would like
tested. You "I believe" made it possible for me to use this card within
a week that I had reported it ~2007-2008 so I and my card are at your
disposal.


Thank you.

-- 

 jhell,v