Make tcp worker multithreaded, too

This shamelessly steals the model established in the UDP workers, i.e.
if the kernel is able to bind multiple listening sockets to the same
endpoint, we will share them across threads. If it can't, all threads
will accept on the same socket.

The fast path handles most of the connections, only the rare case of
no data yet on the socket is offloaded to the slow path (which used to
be the mainloop) for completion. This scales nicely, as the higher the
load is, the higher chances are that an accept happens only after the
network roundtrip to the peer completes, leading to more connections
that already have data waiting for us.

We also consolidate the workstruct init and deinit code, preparation
for later unifying all thread state structs for all protocols (mca,
udc and tcp).

The config file gains a per tcp connection worker count and some docs
explaining what this knob buys and costs, nudging the user to either
keep it 0 or dial it up to at least 2.

1 files changed, 6 insertions, 15 deletions

diff --git a/ot_udp.c b/ot_udp.c
index 1ec5148..00ec566 100644
--- a/ot_udp.c
+++ b/ot_udp.c
@@ -220,26 +220,17 @@ int handle_udp6(int64_t serversocket, struct ot_workstruct *ws) {
 static void *udp_worker(void *args) {
   struct udp_worker   *me = args;
   struct ot_workstruct ws;
-  memset(&ws, 0, sizeof(ws));
 
-  stats_counters_register(&ws);
-  ws.inbuf  = malloc(G_INBUF_SIZE);
-  ws.outbuf = malloc(G_OUTBUF_SIZE);
-#ifdef _DEBUG_HTTPERROR
-  ws.debugbuf = malloc(G_DEBUGBUF_SIZE);
-#endif
+  if (workstruct_init(&ws) == -1) {
+    fprintf(stderr, "Warning: a udp worker could not get its buffers and will not run.\n");
+    atomic_store(&me->left, 1);
+    return NULL;
+  }
 
   while (g_opentracker_running)
     handle_udp6(me->sock, &ws);
 
-  stats_counters_retire(&ws);
-
-  free(ws.inbuf);
-  free(ws.outbuf);
-#ifdef _DEBUG_HTTPERROR
-  free(ws.debugbuf);
-#endif
-
+  workstruct_deinit(&ws);
   atomic_store(&me->left, 1);
   return NULL;
 }