On Thu, Mar 17, 2011 at 7:45 PM, Luis R. Rodriguez <lrodriguez@xxxxxxxxxxx> wrote: > Defer device registration to cfg80211 until it is ready. > This fixes a race against udev whereby device events > (both udev and RTNL) might be emitted with an unregistered > device specification (wlan%d). > > This fixes a race on ath6kl on systems where the device > boots up really quickly to the the point udev and the > respective userspace daemon tries to start mucking with > the device prior to the device being initialized properly. > > This fixes this bug: > > http://code.google.com/p/chromium-os/issues/detail?id=13159 > > Without this loading of ath6kl will fail. > > Cc: stable@xxxxxxxxxx > Cc: Naveen Singh <naveen.singh@xxxxxxxxxxx> > Signed-off-by: Luis R. Rodriguez <lrodriguez@xxxxxxxxxxx> Please disregard this for now, I am testing ath6k right now on x86_64 where it has a failure upon load and it oopses, likely due to this patch and what I believe is wiphy_unregister() being called twice. I'll fix the oops first and then resubmit a v2. If you do not hit a failure upon load though then the patch would help your issues with udev, but still we must handle the failure cases better. My failure are because I'm using an x86_64 box to test ath6kl and its buggy there. Luis _______________________________________________ devel mailing list devel@xxxxxxxxxxxxxxxxxxxxxx http://driverdev.linuxdriverproject.org/mailman/listinfo/devel
