The BLX instruction is encoded differently than the BL, because why not? In

[oota-llvm.git] / test / Analysis / ScalarEvolution / trip-count3.ll

diff --git a/test/Analysis/ScalarEvolution/trip-count3.ll b/test/Analysis/ScalarEvolution/trip-count3.ll
index e4ef955b6d7ad90dc79a30218378b82da781b7ca..10b798b5067dfc4fa96770a394770363b69092e6 100644 (file)
--- a/test/Analysis/ScalarEvolution/trip-count3.ll
+++ b/test/Analysis/ScalarEvolution/trip-count3.ll
@@ -1,5 +1,5 @@
-; RUN: opt %s -scalar-evolution -analyze -disable-output \
-; RUN:  | grep {Loop bb3\\.i: Unpredictable backedge-taken count\\.}
+; RUN: opt < %s -scalar-evolution -analyze \
+; RUN:  | grep {Loop %bb3\\.i: Unpredictable backedge-taken count\\.}
 
 ; ScalarEvolution can't compute a trip count because it doesn't know if
 ; dividing by the stride will have a remainder. This could theoretically