2009/2/10 Bob Copeland :
> So, I'm seeing more reports of mysterious lockups -- some of these
> are stuck beacons perhaps, but I think there could be some bugs
> with the way we deal with interrupts as well. I've yet to come up
> with an obvious scenario that leads to a hang but if anyone has
> on
The
>> patchset needs some cleanup and testing so I'll hold off on patch
>> bombing for now, but you can find it at:
>>
>> http://bobcopeland.com/kernel/wl/20090211/
>
> Most patches (before the "ath" directory creation) look pretty
> uncontroversial
ch
> bombing for now, but you can find it at:
>
> http://bobcopeland.com/kernel/wl/20090211/
Most patches (before the "ath" directory creation) look pretty
uncontroversial and could be submitted so that they get some testing
early.
--
Regards,
Pavel Roskin
Hi all,
I have some preliminary code that takes the ath9k regulatory stuff and
puts it into a shared module that ath5k and ath9k can both use. The
patchset needs some cleanup and testing so I'll hold off on patch
bombing for now, but you can find it at:
http://bobcopeland.com/kernel/wl/200
Bob Copeland wrote:
> 3) For mib interrupts and beacon timer, we program the card within the
> isr. This strikes me as maybe not the best idea, though maybe it's no
> problem. It would be potentially cleaner and safer to use a single
> bottom half in the isr.
We moved beacon processing in ath9k
> For AR5413 cards there was a bug when programming PLL, it's being
> fixed right now (the patch that updates reset sequence), so they had
> bad performance on 5GHz channels. This can be an explanation why your
> AR5413 card couldn't work.
OK, now I using wireless-testing git and I'll try again...