Make integrity tests more robust against timeouts

If the first test in a series for a given MAC happens to modify the low
bytes of a packet length, then ssh will time out and this will be
interpreted as a test failure.  Handle this failure mode.

Bug: https://bugzilla.mindrot.org/show_bug.cgi?id=2658
Patch-Name: regress-integrity-robust.patch

Last-Update: 2017-01-01

1 files changed, 4 insertions, 2 deletions

diff --git a/regress/integrity.sh b/regress/integrity.sh
index 1df2924f5..ed3783372 100644
--- a/regress/integrity.sh
+++ b/regress/integrity.sh
@@ -60,14 +60,16 @@ for m in $macs; do
                 Corrupted?MAC* | *message?authentication?code?incorrect*)
                                 emac=`expr $emac + 1`; skip=0;;
                 padding*)       epad=`expr $epad + 1`; skip=0;;
+                *Timeout,?server*)
+                                etmo=`expr $etmo + 1`; skip=0;;
                 *)              fail "unexpected error mac $m at $off: $out";;
                 esac
         done
-        verbose "test $tid: $ecnt errors: mac $emac padding $epad length $elen"
+        verbose "test $tid: $ecnt errors: mac $emac padding $epad length $elen timeout $etmo"
         if [ $emac -eq 0 ]; then
                 fail "$m: no mac errors"
         fi
-        expect=`expr $ecnt - $epad - $elen`
+        expect=`expr $ecnt - $epad - $elen - $etmo`
         if [ $emac -ne $expect ]; then
                 fail "$m: expected $expect mac errors, got $emac"
         fi