Hi Ulrich,
On Wednesday 16 Aug 2017 15:23:52 Ulrich Hecht wrote:
> On Wed, Jul 19, 2017 at 5:00 PM, Wolfram Sang <wsa@xxxxxxxxxxxxx> wrote:
> >> +{
> >> + wait_event_interruptible_timeout(dev->rx_wq,
> >> + dev->rx_state <= RX_FRAME_ERROR,
> >> + HZ/2);
> >
> > I'd suggest to drop the interruptible. It can be done but it is usually
> > not trivial to abort the operation gracefully when a signal comes in.
> >
> > Also, timeout is superfluous since you don't get the return value?
>
> Be that as it may, I still want a timeout; wouldn't wait_event() block
> forever?
I think that Wolfram's point is that you should report the timeout error up to
the upper layers.
--
Regards,
Laurent Pinchart
