Hello, I have a client/server networking application that exhibits TCP socket handling errors. This only happens on FreeBSD, while NetBSD, Linux, Solaris, etc. all seem to work correctly. I was hoping to get some advice on what could be the root cause.
I have two processes - client and server, sending and receiving data to/from each other on 127.0.0.1 Client connects to server and calls send(2)/recv(2) in a loop. This is a bidirectional data exchange. When all send data is transferred, client calls shutdown(sockfd, SHUT_WR) and continues receiving data on the same socket until recv(2) returns 0 bytes, which signals end of receive data. At this stage client calls close(sockfd) and terminates. Server has the same data transfer loop as the client. I frequently get ECONNRESET when calling close(2), sometimes from the server and sometimes from the client process. This should not be happening, but I'm not sure what could be causing it. The client logic is as follows: 1. Set sockfd nonblocking. 2. Call send(2)/recv(2) in a loop until N bytes have been transferred in each direction. 3. Set sockfd blocking. 4. Call send_buf() to send control handshake to server. 5. Call shutdown(sockfd, SHUT_WR) to signal end of send data from client. 6. Call recv_buf() to receive control handshake from server. 7. Call recv_buf() and verify it returned 0 bytes to indicate end of data from server. 8. Call close(sockfd) and verify success. Step 8 sometimes fails and returns ECONNRESET. Functions send_buf() and recv_buf() are wrappers around send(2) and recv(2) which restart those system calls until the specified number of buffer bytes have been fully transferred or 0 is returned in the case of recv_buf() indicating end of data. They are designed to work with blocking file descriptors and avoid short reads/writes. I don't understand why close(2) sometimes returns ECONNRESET when the previous recv(2) call at step 7 returned 0 bytes, indicating the remote TCP end sent us a FIN. I don't set SO_LINGER socket option and when I checked the default on FreeBSD it reports l_onoff=0, l_linger=0 so there should be no immediate RST on socket close(2). Does anyone have any suggestions or ideas? Thanks.
