Re: [ath5k-devel] Race condition in CRDA calls?

["Luis R. Rodriguez" <lrodriguez@atheros.com>]

On Mon, Oct 11, 2010 at 11:13:03AM -0700, Lukáš Turek wrote:
> On Friday 08 October 2010 09:57:17 you wrote:
> > root@RMR1:~# 1284335259.046129: phy #1: regulatory domain change:
> > intersection  used due to a request made by a driver on phy1
> > 1284335263.403515: phy #1: regulatory domain change: intersection used due
> > to  a request made by a driver on phy1
> > 
> > interesting that phy #1 appears twice...
> 
> When I hit this bug my idea was that the problem lies in the CRDA API: the 
> wiphy identifier is not sent to CRDA, instead a global variable last_request 
> is used. However, if a new request is started before the previous one is 
> finished, the old last_request value is replaced, both replies use the same 
> last_request and strange think happen...
> 
> I hoped someone who knows the kernel-userspace interaction better will confirm 
> or disprove it, so I didn't investigate it further, and then I just forgot 
> about it...

Sorry about the delay on a response on this, I've just been caught up
with other higher priority issues, but as soon as I have some cycles I will
give it a shot with two cards.

  Luis