comm_tcp: enable TCP keepalive on host side as well

artiq/coredevice/comm_tcp.py

@@ -1,5 +1,6 @@
 import logging
 import socket
+import sys
 
 from artiq.coredevice.comm_generic import CommGeneric
 
@@ -7,6 +8,22 @@ from artiq.coredevice.comm_generic import CommGeneric
 logger = logging.getLogger(__name__)
 
 
+def set_keepalive(sock, after_idle, interval, max_fails):
+    if sys.platform.startswith("linux"):
+        sock.setsockopt(socket.SOL_SOCKET, socket.SO_KEEPALIVE, 1)
+        sock.setsockopt(socket.IPPROTO_TCP, socket.TCP_KEEPIDLE, after_idle)
+        sock.setsockopt(socket.IPPROTO_TCP, socket.TCP_KEEPINTVL, interval)
+        sock.setsockopt(socket.IPPROTO_TCP, socket.TCP_KEEPCNT, max_fails)
+    elif sys.platform.startswith("win") or sys.platform.startswith("cygwin"):
+        # setting max_fails is not supported, typically ends up being 5 or 10
+        # depending on Windows version
+        sock.ioctl(socket.SIO_KEEPALIVE_VALS,
+                   (1, after_idle*1000, interval*1000))
+    else:
+        logger.warning("TCP keepalive not supported on platform '%s', ignored",
+                       sys.platform)
+
+
 class Comm(CommGeneric):
     def __init__(self, dmgr, host, port=1381):
         self.host = host
@@ -16,6 +33,7 @@ class Comm(CommGeneric):
         if hasattr(self, "socket"):
             return
         self.socket = socket.create_connection((self.host, self.port))
+        set_keepalive(self.socket, 3, 2, 3)
         logger.debug("connected to host %s on port %d", self.host, self.port)
         self.write(b"ARTIQ coredev\n")