P2P: improve RX/TX flow control

1) "optimistic write": Push each message to kernel socket buffer immediately.

2) If there is write data at select time, that implies send() blocked
   during optimistic write.  Drain write queue, before receiving
   any more messages.

This avoids needlessly queueing received data, if the remote peer
is not themselves receiving data.

Result: write buffer (and thus memory usage) is kept small, DoS
potential is slightly lower, and TCP flow control signalling is
properly utilized.

The kernel will queue data into the socket buffer, then signal the
remote peer to stop sending data, until we resume reading again.

1 files changed, 10 insertions, 6 deletions

diff --git a/src/net.cpp b/src/net.cpp
index 96719367ce..eafb335642 100644
--- a/src/net.cpp
+++ b/src/net.cpp
@@ -855,14 +855,18 @@ void ThreadSocketHandler2(void* parg)
             {
                 if (pnode->hSocket == INVALID_SOCKET)
                     continue;
-                FD_SET(pnode->hSocket, &fdsetRecv);
-                FD_SET(pnode->hSocket, &fdsetError);
-                hSocketMax = max(hSocketMax, pnode->hSocket);
-                have_fds = true;
                 {
                     TRY_LOCK(pnode->cs_vSend, lockSend);
-                    if (lockSend && !pnode->vSend.empty())
-                        FD_SET(pnode->hSocket, &fdsetSend);
+                    if (lockSend) {
+                        // do not read, if draining write queue
+                        if (!pnode->vSend.empty())
+                            FD_SET(pnode->hSocket, &fdsetSend);
+                        else
+                            FD_SET(pnode->hSocket, &fdsetRecv);
+                        FD_SET(pnode->hSocket, &fdsetError);
+                        hSocketMax = max(hSocketMax, pnode->hSocket);
+                        have_fds = true;
+                    }
                 }
             }
         }