Closed LazyDodo closed 10 years ago
Is this a bug? I am not sure what the action should be from the library standpoint if a request came back with less than the actual length being transferred for a bulk transfer. Do you resubmit with the rest?
This is a bug. The library needs to handle shorter returns. This is nominal behavior when the FX3 is out of data to send.
So is the solution to just resubmit the transfer of the shorter length? What about minimum transfer sizes? Do those need to be upheld with the leftover data?
What we should definitely NOT do is bother the end user with a big ass popup with this exception with a abort/retry/ignore options. (also, assertions on in release builds?) I say pass the proper length to the callback and remove the assert.
https://github.com/Nuand/bladeRF/blob/master/host/libraries/libbladeRF/src/backend/libusb.c#L1725