At some point between 2.5.46 and 2.5.47, the LTP tests for recvfrom01
and recvmsg01 began to fail. The test common to both that is failing is
passing fromlen == -1. Calling recvfrom or recvmsg like this passes on
2.4 and 2.5.46, but returns -EINVAL on 2.5.47 and later. If this was
intentional, I could understand it since that is certainly an invalid
argument for fromlen, but I would like to have that confirmed, and know
what the plans are for backporting that to 2.4. I'll probably be
changing the test before the next ltp release (should be on Dec. 10),
but it will just flipflop from the test failing on 2.5 all the time to
failing on 2.4 all the time.
Thanks,
Paul Larson