diff --git a/docs/src/code/writing-tests.adoc b/docs/src/code/writing-tests.adoc
index c50061f1a4a..02e73c5465f 100644
--- a/docs/src/code/writing-tests.adoc
+++ b/docs/src/code/writing-tests.adoc
@@ -68,9 +68,6 @@ checkresult::
    The exit code of this program controls the result of the test.
    If both `expected` and `checkresult` exist, only `checkresult` is consulted to validate the test output.
 
- xfail::
-   If this file exist, a test failure is expected and does not cause runtests to return an exit code signaling an error.
-
  skip::
    If this file exist, the test is skipped and not executed at all.
 
diff --git a/scripts/runtests.in b/scripts/runtests.in
index 0b2fc95901a..76ccab8ce3d 100755
--- a/scripts/runtests.in
+++ b/scripts/runtests.in
@@ -45,7 +45,6 @@ export RUNTESTS
 
 NUM=0
 FAIL=0; FAIL_NAMES=""
-XFAIL=0
 SHMERR=0
 SKIP=0
 VERBOSE=0
@@ -286,7 +285,7 @@ run_tests () {
 	    fi
 	fi
 	if [ "$exitcode" -ne 0 ]; then
-	    echo "*** $testdir: XFAIL: $reason"
+	    echo "*** $testdir: FAIL: $reason"
             if test $PRINT = 1; then
                 echo "************** result:"
                 tail -500 "$testdir/result" | sed 's/^/        /'
@@ -294,27 +293,15 @@ run_tests () {
                 tail -500 "$testdir/stderr" | sed 's/^/        /'
                 echo "**************"
             fi
-	    if [ -f "$testdir/xfail" ]; then
-		XFAIL=$((XFAIL + 1))
-		if [ $NOCLEAN -eq 0 ]; then
-		    rm -f "$testdir/stderr" "$testdir/result" \
-			"$testdir"/*.var "$testdir"/*.var.bak
-		fi
-	    else
-		FAIL=$((FAIL + 1))
-		FAIL_NAMES="$FAIL_NAMES"$'\n'"$testdir"
-	    fi
+	    FAIL=$((FAIL + 1))
+	    FAIL_NAMES="$FAIL_NAMES"$'\n'"$testdir"
             if test $STOP = 1; then
 	        break
 	    fi
 	else
-	    if [ -f "$testdir/xfail" ]; then
-		echo "*** $testdir: XPASS: Passed, but was expected to fail"
-	    else
-		if [ $NOCLEAN -eq 0 ]; then
-		    rm -f "$testdir/stderr" "$testdir/result" \
-			"$testdir"/*.var "$testdir"/*.var.bak
-		fi
+	    if [ $NOCLEAN -eq 0 ]; then
+		rm -f "$testdir/stderr" "$testdir/result" \
+		    "$testdir"/*.var "$testdir"/*.var.bak
 	    fi
 	fi
 
@@ -327,8 +314,8 @@ run_tests () {
         fi
     done < "$TMPDIR/alltests"
 
-    SUCC=$((NUM-FAIL-XFAIL))
-    echo "Runtest: $NUM tests run, $SUCC successful, $FAIL failed + $XFAIL expected, $SKIP skipped, $SHMERR shmem errors"
+    SUCC=$((NUM-FAIL))
+    echo "Runtest: $NUM tests run, $SUCC successful, $FAIL failed, $SKIP skipped, $SHMERR shmem errors"
     if [ "$FAIL" -ne 0 ]; then
 	echo "Failed: $FAIL_NAMES"
 	exit 1;
diff --git a/tests/interp/m98m99/12-M99-endless-main-program/xfail b/tests/interp/m98m99/12-M99-endless-main-program/xfail
deleted file mode 100644
index e69de29bb2d..00000000000
diff --git a/tests/overrun/README b/tests/overrun/README
index 3c44e21e20b..414a47e7a19 100644
--- a/tests/overrun/README
+++ b/tests/overrun/README
@@ -1,2 +1,2 @@
-This test makes sure that the "overrun" detection works.  It is an XFAIL
-because the testing infrastructure will believe it is a failure.
+This test makes sure that the "overrun" detection works.  The inner runtests
+run is expected to fail, so test.sh negates its exit code.
diff --git a/tests/overrun/checkresult b/tests/overrun/checkresult
index b64faf30722..414899285da 100755
--- a/tests/overrun/checkresult
+++ b/tests/overrun/checkresult
@@ -2,4 +2,4 @@
 set -e
 grep -q "overrun detected in sampler" "$1"
 grep -q "10 overruns detected, giving up" "$1"
-grep -q "1 tests run, 0 successful, 1 failed . 0 expected" "$1"
+grep -q "1 tests run, 0 successful, 1 failed," "$1"