On Thu, Mar 20, 2014 at 12:15 AM, Oliver Neukum <oneukum@xxxxxxx> wrote:
...
> I have an idea. Could you test this patch?
...
> - if (dev->wait) {
..
> + if (waitqueue_active(&dev->wait)) {
Yes - building new image now (and transfer to USB and boot from USB).
Should know in an hour or so (doing other things in parallel).
I was sure the problem is in usbnet_bh() since that's the only code
change I'm actually exercising (so far). The way I was reading the
code, we might see extra wake_up calls...but there is clearly more
going on.
Can you please explain why we need to check if the waitqueue is active?
This patch should also add a comment to hint why waitqueue_active()
must be called.
Why? Several experienced people looked at the code and didn't see the
problem including the original author of the patch.
thanks,
grant
--
To unsubscribe from this list: send the line "unsubscribe linux-usb" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
