On Wed, 8 Jun 2011 07:22:04 +0800 Adrian Chadd <adr...@freebsd.org> wrote:
> On 8 June 2011 04:54, Tony Houghton <h...@realh.co.uk> wrote: > > > My bisect log showed the last "good" commit was > > d5cdfacb35ed886271d1ccfffbded98d3447da17 and the first "bad" > > 53bc7aa08b48e5cd745f986731cc7dc24eef2a9f. > > > > I tested those again and it seemed to confirm the initial finding. > > With the "good" one I suspended/resumed at least twice, did rmmod > > ath9k at least twice, and disabled wireless with NetworkManager, > > all without incident. The "bad" one crashed the first time I tried > > rmmod ath9k. > > Ok. Let's try the patches in turn, ok? > > Go back to the first good commit, then apply the patch to reg.h to > pick up the change to AR9285_WA_DEFAULT. > Then test with that. > > The only change in your tree should be to reg.h: > > http://git390.marist.edu/cgi-bin/gitweb.cgi?p=linux-2.6.git;a=blobdiff;f=drivers/net/wireless/ath/ath9k/reg.h;h=7e36ad7421b77f6464c8bc810c9b1276209fe329;hp=198e41dd38a6142e232a22807d665f78dbecae72;hb=53bc7aa08b48e5cd745f986731cc7dc24eef2a9f;hpb=d5cdfacb35ed886271d1ccfffbded98d3447da17 Right, it did crash with this patch. This time I was able to rmmod ath9k once but it crashed on rmload ath9k. I also had a freeze while booting up. That was on AC power. With this version, running on battery power seems to greatly reduce the chances of a crash, if not prevent it altogether. I had noticed this on previous stock kernels, but recent ones crash on battery power too. _______________________________________________ ath9k-devel mailing list ath9k-devel@lists.ath9k.org https://lists.ath9k.org/mailman/listinfo/ath9k-devel