Take yet more care not to be dividing by zero when calculating the bit

rate of the channel/UE. Times four... 

svn path=/trunk/; revision=43578

1 files changed, 8 insertions, 0 deletions

diff --git a/ui/cli/tap-macltestat.c b/ui/cli/tap-macltestat.c
index af230da51f..cf2df41e09 100644
--- a/ui/cli/tap-macltestat.c
+++ b/ui/cli/tap-macltestat.c
@@ -395,9 +395,16 @@ mac_lte_stat_packet(void *phs, packet_info *pinfo, epan_dissect_t *edt _U_,
 /* Calculate and return a bandwidth figure, in Mbs */
 static float calculate_bw(nstime_t *start_time, nstime_t *stop_time, guint32 bytes)
 {
+    /* Can only calculate bandwidth if have time delta */
     if (memcmp(start_time, stop_time, sizeof(nstime_t)) != 0) {
         float elapsed_ms = (((float)stop_time->secs - (float)start_time->secs) * 1000) +
                            (((float)stop_time->nsecs - (float)start_time->nsecs) / 1000000);
+
+        /* Only really meaningful if have a few frames spread over time...
+           For now at least avoid dividing by something very close to 0.0 */
+        if (elapsed_ms < 2.0) {
+           return 0.0;
+        }
         return ((bytes * 8) / elapsed_ms) / 1000;
     }
     else {
@@ -406,6 +413,7 @@ static float calculate_bw(nstime_t *start_time, nstime_t *stop_time, guint32 byt
 }
 
 
+
 /* Output the accumulated stats */
 static void
 mac_lte_stat_draw(void *phs)