Use OpenSSL's SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER flag.

This disables an entirely unnecessary "sanity check" that causes failures
in nonblocking mode, because OpenSSL complains if we move or compact the
write buffer.  The only actual requirement is that we not modify pending
data once we've attempted to send it, which we don't.  Per testing and
research by Martin Pihlak, though this fix is a lot simpler than his patch.

I put the same change into the backend, although it's less clear whether
it's necessary there.  We do use nonblock mode in some situations in
streaming replication, so seems best to keep the same behavior in the
backend as in libpq.

Back-patch to all supported releases.

src/backend/libpq/be-secure.c

@@ -727,6 +727,12 @@ initialize_SSL(void)
 					(errmsg("could not create SSL context: %s",
 							SSLerrmessage())));
 
+		/*
+		 * Disable OpenSSL's moving-write-buffer sanity check, because it
+		 * causes unnecessary failures in nonblocking send cases.
+		 */
+		SSL_CTX_set_mode(SSL_context, SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER);
+
 		/*
 		 * Load and verify certificate and private key
 		 */

src/interfaces/libpq/fe-secure.c

@@ -900,6 +900,12 @@ init_ssl_system(PGconn *conn)
 #endif
 			return -1;
 		}
+
+		/*
+		 * Disable OpenSSL's moving-write-buffer sanity check, because it
+		 * causes unnecessary failures in nonblocking send cases.
+		 */
+		SSL_CTX_set_mode(SSL_context, SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER);
 	}
 #ifdef ENABLE_THREAD_SAFETY
 	pthread_mutex_unlock(&init_mutex);