btl/usnic: increase default retrans_timeout

Significantly increase the default retrans timeout.  If the
retrans timeout is too soon, we can end up in a retransmission storm
where the logic will continually re-transmit the same frames during a
single run through the usNIC progress function (because the timer for
a single frame expires before we have run through re-transmitting all
the frames pending re-transmission).

Signed-off-by: Jeff Squyres <[email protected]>

Author: jsquyres

opal/mca/btl/usnic/btl_usnic_mca.c

@@ -246,7 +246,7 @@ int opal_btl_usnic_component_register(void)
     mca_btl_usnic_component.udp_port_base = (int) udp_port_base;
 
     CHECK(reg_int("retrans_timeout", "Number of microseconds before retransmitting a frame",
-                  5000, &mca_btl_usnic_component.retrans_timeout,
+                  100000, &mca_btl_usnic_component.retrans_timeout,
                   REGINT_GE_ONE, OPAL_INFO_LVL_5));
 
     CHECK(reg_int("ack_iteration_delay", "Minimum number of times through usNIC \"progress\" function before checking to see if standalone ACKs need to be sent",