Dennis Windisch wrote: >>>> This is conforming again: MSG_DONTWAIT passed as flag to rt_dev_recv* >>>> will have the non-blocking effect specifically for this call. >>> I've done this but error remains the same (-11 = EAGAIN). >>> That seems strange to me because I know there is data available at the >>> time >>> the recv is executed (I even tried some kind of polling, namely making a >>> loop over the recv function polling it 1000 times every cycle). >>> Where is my error in reasoning? >>> >> RTAI 3.3 release or vulcano/magma CVS? The former contains a bug which >> could explain this behaviour. > > I've updated to the latest magma CVS, but the error remains the same. Do you > have any other ideas/suggestions? >
Sorry for the late reply. I have actually three ideas: o Double-check with rtcap on the receiver side if there are actually the expected packets arriving (no typos in the port number e.g.). o Does changing the timeout to something greater or equal 0 change anything? o Can you reproduce the same issue over rt-loopback? If yes, it would be very helpful to derive a simple test case and post it to the list. Jan
signature.asc
Description: OpenPGP digital signature
