[ath9k-devel] ath9k (AR9485) failing with PCI errors after hours of uptime with current ath9k.ko from linux-wireless

From: Adrian Chadd <adrian@freebsd.org>
To: ath9k-devel@lists.ath9k.org
Subject: [ath9k-devel] ath9k (AR9485) failing with PCI errors after hours of uptime with current ath9k.ko from linux-wireless
Date: Mon, 13 May 2013 06:29:12 -0700	[thread overview]
Message-ID: <CAJ-VmokY31FRdQmtKPFQzeDaAFL9Mj+tE9qZNnFUrUQ2UQqeuQ@mail.gmail.com> (raw)
In-Reply-To: <1368444534.15473.9.camel@debian>

On 13 May 2013 04:28, Joel Diaz <joeldiaz@earthlink.net> wrote:
> On Wed, 2013-05-08 at 22:10 -0700, Adrian Chadd wrote:
>> No, I mean - there's an iw command to disable entering power save on
>> your linux station. Find that and disable it.
>>
> It has now been running for about 4 and half days without seeing the
> wireless connection drop after the 'iw dev wlan0 set power_save off'. I
> hope it's not premature to say thanks, as this is the longest I've seen
> it go without a failure.
>
> I'm not at all familiar with the history behind why power saving needs
> to be disabled (as this machine is a desktop, it's not terribly
> important to me), but if there's any hope of having it fixed through
> patches, I'd be more than willing to help out so other people don't have
> to go through what I went through.

Hi!

please file a kernel.org bugzilla bug.

There's two overarching possibilities:

* there's a driver bug (eg setting up DMA when the chip is asleep);
trying to put the NIC to sleep during a DMA, etc;
* that some PCIe quirk needs to be added to the ath9k driver to work
around some oddities with PCie sleep state(s).

Thanks,

Adrian