Hi all, I've been talking privately with Tony and I think we've narrowed down what's going on.
In short - the pcie powersave stuff is somehow messed up. On 25 February 2011 04:02, Jonathan Nieder <jrnie...@gmail.com> wrote: >> 53bc7aa08b48e5cd745f986731cc7dc24eef2a9f is the first bad commit >> commit 53bc7aa08b48e5cd745f986731cc7dc24eef2a9f >> Author: Vivek Natarajan <vnatara...@atheros.com> >> Date: Mon Apr 5 14:48:04 2010 +0530 >> >> ath9k: Add support for newer AR9285 chipsets. The long: the AR9285_WA_DEFAULT change in the above patch causes Tony's NIC to become unstable when powered down. The long(er): When he tested the latest wireless-testing tree with the AR9285_WA_DEFAULT value reverted, he had the same problem. Since the pcie powersave function has changed considerably between that above commit and today, there's also very likely something subtly evil going on. Given that we've now narrowed down the specific change that's causing this, and that the latest ath9k ath9k_hw_configpcipowersave() function is doing quite a bit different (and I think disabling those bits too), I think we can conclude that there's something strange going on. What I haven't yet tried is getting Tony to test wireless-testing but with an older ath9k_hw_configpcipowersave(). Since this is affecting a variety of people, would it be possible to get it looked at by someone with a better understanding of the PCIe power saving/APSM stuff? I hereby punt this to Atheros/Qualcomm to resolve. :) Adrian _______________________________________________ ath9k-devel mailing list ath9k-devel@lists.ath9k.org https://lists.ath9k.org/mailman/listinfo/ath9k-devel
