From: Sanjay Patel <spatel@rotateright.com>
Date: Wed, 24 May 2017 14:21:31 +0000 (+0000)
Subject: [InstCombine] add tests to document bitcast + bitwise-logic behavior; NFC
X-Git-Url: https://granicus.if.org/sourcecode?a=commitdiff_plain;h=852ab094ff77c05ade3a20e414f092840b53bcb7;p=llvm

[InstCombine] add tests to document bitcast + bitwise-logic behavior; NFC

The solution for PR26702 ( https://bugs.llvm.org/show_bug.cgi?id=26702 )
added a canonicalization rule, but the minimal regression tests don't
demonstrate how that rule interacts with other folds.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@303750 91177308-0d34-0410-b5e6-96231b3b80d8
---

diff --git a/test/Transforms/InstCombine/bitcast.ll b/test/Transforms/InstCombine/bitcast.ll
index 2e7f30fee14..4cf3f27ab01 100644
--- a/test/Transforms/InstCombine/bitcast.ll
+++ b/test/Transforms/InstCombine/bitcast.ll
@@ -70,6 +70,51 @@ define <2 x i32> @or_bitcast_int_to_vec(i64 %a) {
   ret <2 x i32> %t2
 }
 
+; PR26702 - https://bugs.llvm.org//show_bug.cgi?id=26702
+; Bitcast is canonicalized below logic, so we can see the not-not pattern.
+
+define <2 x i64> @is_negative(<4 x i32> %x) {
+; CHECK-LABEL: @is_negative(
+; CHECK-NEXT:    [[LOBIT:%.*]] = ashr <4 x i32> %x, <i32 31, i32 31, i32 31, i32 31>
+; CHECK-NEXT:    [[NOTNOT:%.*]] = bitcast <4 x i32> [[LOBIT]] to <2 x i64>
+; CHECK-NEXT:    ret <2 x i64> [[NOTNOT]]
+;
+  %lobit = ashr <4 x i32> %x, <i32 31, i32 31, i32 31, i32 31>
+  %not = xor <4 x i32> %lobit, <i32 -1, i32 -1, i32 -1, i32 -1>
+  %bc = bitcast <4 x i32> %not to <2 x i64>
+  %notnot = xor <2 x i64> %bc, <i64 -1, i64 -1>
+  ret <2 x i64> %notnot
+}
+
+; This variation has an extra bitcast at the end. This means that the 2nd xor
+; can be done in <4 x i32> to eliminate a bitcast regardless of canonicalizaion.
+
+define <4 x i32> @is_negative_bonus_bitcast(<4 x i32> %x) {
+; CHECK-LABEL: @is_negative_bonus_bitcast(
+; CHECK-NEXT:    [[LOBIT:%.*]] = ashr <4 x i32> %x, <i32 31, i32 31, i32 31, i32 31>
+; CHECK-NEXT:    ret <4 x i32> [[LOBIT]]
+;
+  %lobit = ashr <4 x i32> %x, <i32 31, i32 31, i32 31, i32 31>
+  %not = xor <4 x i32> %lobit, <i32 -1, i32 -1, i32 -1, i32 -1>
+  %bc = bitcast <4 x i32> %not to <2 x i64>
+  %notnot = xor <2 x i64> %bc, <i64 -1, i64 -1>
+  %bc2 = bitcast <2 x i64> %notnot to <4 x i32>
+  ret <4 x i32> %bc2
+}
+
+; Negative test: bitcasts are canonicalized below bitwise logic. No changes here.
+
+define <2 x i8> @canonicalize_bitcast_logic_with_constant(<4 x i4> %x) {
+; CHECK-LABEL: @canonicalize_bitcast_logic_with_constant(
+; CHECK-NEXT:    [[A:%.*]] = and <4 x i4> %x, <i4 0, i4 -8, i4 0, i4 -8>
+; CHECK-NEXT:    [[B:%.*]] = bitcast <4 x i4> [[A]] to <2 x i8>
+; CHECK-NEXT:    ret <2 x i8> [[B]]
+;
+  %a = and <4 x i4> %x, <i4 0, i4 8, i4 0, i4 8>
+  %b = bitcast <4 x i4> %a to <2 x i8>
+  ret <2 x i8> %b
+}
+
 ; PR27925 - https://llvm.org/bugs/show_bug.cgi?id=27925
 
 define <4 x i32> @bitcasts_and_bitcast(<4 x i32> %a, <8 x i16> %b) {