It's not necessary to do rounding for alloca operations when the requested

alignment is equal to the stack alignment.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@40004 91177308-0d34-0410-b5e6-96231b3b80d8

1 files changed, 26 insertions, 0 deletions

diff --git a/test/Transforms/LoopStrengthReduce/dont-hoist-simple-loop-constants.ll b/test/Transforms/LoopStrengthReduce/dont-hoist-simple-loop-constants.ll
new file mode 100644
index 0000000..3c1003b
--- /dev/null
+++ b/test/Transforms/LoopStrengthReduce/dont-hoist-simple-loop-constants.ll
@@ -0,0 +1,26 @@
+; RUN: llvm-upgrade < %s | llvm-as | opt -loop-reduce | llvm-dis | \
+; RUN:   not grep {cast uint 1 to uint}
+; END.
+; The setlt wants to use a value that is incremented one more than the dominant
+; IV.  Don't insert the 1 outside the loop, preventing folding it into the add.
+
+implementation   ; Functions:
+
+void %test([700 x int]* %nbeaux_.0__558, int* %i_.16574) {
+then.0:
+        br label %no_exit.2
+
+no_exit.2:              ; preds = %no_exit.2, %then.0
+        %indvar630 = phi uint [ 0, %then.0 ], [ %indvar.next631, %no_exit.2 ]           ; <uint> [#uses=3]
+        %indvar630 = cast uint %indvar630 to int                ; <int> [#uses=1]
+        %tmp.38 = getelementptr [700 x int]* %nbeaux_.0__558, int 0, uint %indvar630            ; <int*> [#uses=1]
+        store int 0, int* %tmp.38
+        %inc.2 = add int %indvar630, 2          ; <int> [#uses=2]
+        %tmp.34 = setlt int %inc.2, 701         ; <bool> [#uses=1]
+        %indvar.next631 = add uint %indvar630, 1                ; <uint> [#uses=1]
+        br bool %tmp.34, label %no_exit.2, label %loopexit.2.loopexit
+
+loopexit.2.loopexit:            ; preds = %no_exit.2
+        store int %inc.2, int* %i_.16574
+        ret void
+}