S390: Disable gcc generation of mul-add in cctests

Generated code performs distinct floating multiply and add/subtract
operations. Tests fail when GCC uses mul-add to calculate the
expected result since these instructions provide higher accuracy due
to the lack of an intermediate round.

R=machenbach@chromium.org, mbrandy@us.ibm.com, joransiu@ca.ibm.com, michael_dawson@ca.ibm.com
BUG=

Review URL: https://codereview.chromium.org/1812823008

Cr-Commit-Position: refs/heads/master@{#34953}

test/cctest/cctest.gyp

@@ -314,7 +314,8 @@
           },
         }],
         ['v8_target_arch=="ppc" or v8_target_arch=="ppc64" \
-          or v8_target_arch=="arm" or v8_target_arch=="arm64"', {
+          or v8_target_arch=="arm" or v8_target_arch=="arm64" \
+          or v8_target_arch=="s390" or v8_target_arch=="s390x"', {
           # disable fmadd/fmsub so that expected results match generated code in
           # RunFloat64MulAndFloat64Add1 and friends.
           'cflags': ['-ffp-contract=off'],