From 2e60048641e86b7a414aec51d920bc4e1e3fbeb6 Mon Sep 17 00:00:00 2001
From: Yingwei Zheng <dtcxzyw2333@gmail.com>
Date: Thu, 21 Nov 2024 20:47:24 +0800
Subject: [PATCH] [InstCombine] Fold zext(X) + C2 pred C -> X + C3 pred C4
 (#110511)

Motivating case from
https://github.com/torvalds/linux/blob/9852d85ec9d492ebef56dc5f229416c925758edc/drivers/gpu/drm/drm_edid.c#L5238-L5240:
```
define i1 @src(i8 noundef %v13) {
entry:
  %conv1 = zext i8 %v13 to i32
  %add = add nsw i32 %conv1, -4
  %cmp = icmp ult i32 %add, 3
  %cmp4 = icmp slt i8 %v13, 4
  %cond = select i1 %cmp4, i1 true, i1 %cmp
  ret i1 %cond
}

define i1 @tgt(i8 noundef %v13) {
entry:
  %cmp4 = icmp slt i8 %v13, 7
  ret i1 %cmp4
}
```
---
 .../InstCombine/InstCombineCompares.cpp       |  24 ++++
 llvm/test/Transforms/InstCombine/icmp-add.ll  | 106 ++++++++++++++++++
 2 files changed, 130 insertions(+)

diff --git a/llvm/lib/Transforms/InstCombine/InstCombineCompares.cpp b/llvm/lib/Transforms/InstCombine/InstCombineCompares.cpp
index d602a907e72bcd..4be576c8ce3b0f 100644
--- a/llvm/lib/Transforms/InstCombine/InstCombineCompares.cpp
+++ b/llvm/lib/Transforms/InstCombine/InstCombineCompares.cpp
@@ -3183,6 +3183,30 @@ Instruction *InstCombinerImpl::foldICmpAddConstant(ICmpInst &Cmp,
                         Builder.CreateAdd(X, ConstantInt::get(Ty, *C2 - C - 1)),
                         ConstantInt::get(Ty, ~C));
 
+  // zext(V) + C2 pred C -> V + C3 pred' C4
+  Value *V;
+  if (match(X, m_ZExt(m_Value(V)))) {
+    Type *NewCmpTy = V->getType();
+    unsigned CmpBW = Ty->getScalarSizeInBits();
+    unsigned NewCmpBW = NewCmpTy->getScalarSizeInBits();
+    if (shouldChangeType(Ty, NewCmpTy)) {
+      if (CR.getActiveBits() <= NewCmpBW) {
+        ConstantRange SrcCR = CR.truncate(NewCmpBW);
+        CmpInst::Predicate EquivPred;
+        APInt EquivInt;
+        APInt EquivOffset;
+
+        SrcCR.getEquivalentICmp(EquivPred, EquivInt, EquivOffset);
+        return new ICmpInst(
+            EquivPred,
+            EquivOffset.isZero()
+                ? V
+                : Builder.CreateAdd(V, ConstantInt::get(NewCmpTy, EquivOffset)),
+            ConstantInt::get(NewCmpTy, EquivInt));
+      }
+    }
+  }
+
   return nullptr;
 }
 
diff --git a/llvm/test/Transforms/InstCombine/icmp-add.ll b/llvm/test/Transforms/InstCombine/icmp-add.ll
index dd8e9c1a45ea10..579247aaccf282 100644
--- a/llvm/test/Transforms/InstCombine/icmp-add.ll
+++ b/llvm/test/Transforms/InstCombine/icmp-add.ll
@@ -3183,3 +3183,109 @@ define i1 @icmp_of_ucmp_plus_const_with_const(i32 %x, i32 %y) {
   %cmp2 = icmp ult i8 %add, 2
   ret i1 %cmp2
 }
+
+define i1 @zext_range_check_ult(i8 %x) {
+; CHECK-LABEL: @zext_range_check_ult(
+; CHECK-NEXT:  entry:
+; CHECK-NEXT:    [[TMP0:%.*]] = add i8 [[X:%.*]], -4
+; CHECK-NEXT:    [[CMP:%.*]] = icmp ult i8 [[TMP0]], 3
+; CHECK-NEXT:    ret i1 [[CMP]]
+;
+entry:
+  %conv = zext i8 %x to i32
+  %add = add i32 %conv, -4
+  %cmp = icmp ult i32 %add, 3
+  ret i1 %cmp
+}
+
+; TODO: should be canonicalized to (x - 4) u> 2
+define i1 @zext_range_check_ugt(i8 %x) {
+; CHECK-LABEL: @zext_range_check_ugt(
+; CHECK-NEXT:  entry:
+; CHECK-NEXT:    [[CONV:%.*]] = zext i8 [[X:%.*]] to i32
+; CHECK-NEXT:    [[TMP0:%.*]] = add nsw i32 [[CONV]], -7
+; CHECK-NEXT:    [[CMP:%.*]] = icmp ult i32 [[TMP0]], -3
+; CHECK-NEXT:    ret i1 [[CMP]]
+;
+entry:
+  %conv = zext i8 %x to i32
+  %add = add i32 %conv, -4
+  %cmp = icmp ugt i32 %add, 2
+  ret i1 %cmp
+}
+
+; TODO: should be canonicalized to (x - 4) u> 2
+define i1 @zext_range_check_ult_alter(i8 %x) {
+; CHECK-LABEL: @zext_range_check_ult_alter(
+; CHECK-NEXT:  entry:
+; CHECK-NEXT:    [[CONV:%.*]] = zext i8 [[X:%.*]] to i32
+; CHECK-NEXT:    [[ADD:%.*]] = add nsw i32 [[CONV]], -7
+; CHECK-NEXT:    [[CMP:%.*]] = icmp ult i32 [[ADD]], -3
+; CHECK-NEXT:    ret i1 [[CMP]]
+;
+entry:
+  %conv = zext i8 %x to i32
+  %add = add i32 %conv, -7
+  %cmp = icmp ult i32 %add, -3
+  ret i1 %cmp
+}
+
+define i1 @zext_range_check_mergable(i8 %x) {
+; CHECK-LABEL: @zext_range_check_mergable(
+; CHECK-NEXT:    [[COND:%.*]] = icmp slt i8 [[X:%.*]], 7
+; CHECK-NEXT:    ret i1 [[COND]]
+;
+  %conv = zext i8 %x to i32
+  %add = add nsw i32 %conv, -4
+  %cmp1 = icmp ult i32 %add, 3
+  %cmp2 = icmp slt i8 %x, 4
+  %cond = select i1 %cmp2, i1 true, i1 %cmp1
+  ret i1 %cond
+}
+
+; Negative tests
+
+define i1 @sext_range_check_ult(i8 %x) {
+; CHECK-LABEL: @sext_range_check_ult(
+; CHECK-NEXT:  entry:
+; CHECK-NEXT:    [[CONV:%.*]] = sext i8 [[X:%.*]] to i32
+; CHECK-NEXT:    [[ADD:%.*]] = add nsw i32 [[CONV]], -4
+; CHECK-NEXT:    [[CMP:%.*]] = icmp ult i32 [[ADD]], 3
+; CHECK-NEXT:    ret i1 [[CMP]]
+;
+entry:
+  %conv = sext i8 %x to i32
+  %add = add i32 %conv, -4
+  %cmp = icmp ult i32 %add, 3
+  ret i1 %cmp
+}
+
+define i1 @zext_range_check_ult_illegal_type(i7 %x) {
+; CHECK-LABEL: @zext_range_check_ult_illegal_type(
+; CHECK-NEXT:  entry:
+; CHECK-NEXT:    [[CONV:%.*]] = zext i7 [[X:%.*]] to i32
+; CHECK-NEXT:    [[ADD:%.*]] = add nsw i32 [[CONV]], -4
+; CHECK-NEXT:    [[CMP:%.*]] = icmp ult i32 [[ADD]], 3
+; CHECK-NEXT:    ret i1 [[CMP]]
+;
+entry:
+  %conv = zext i7 %x to i32
+  %add = add i32 %conv, -4
+  %cmp = icmp ult i32 %add, 3
+  ret i1 %cmp
+}
+
+define i1 @zext_range_check_ult_range_check_failure(i8 %x) {
+; CHECK-LABEL: @zext_range_check_ult_range_check_failure(
+; CHECK-NEXT:  entry:
+; CHECK-NEXT:    [[CONV:%.*]] = zext i8 [[X:%.*]] to i32
+; CHECK-NEXT:    [[ADD:%.*]] = add nsw i32 [[CONV]], -4
+; CHECK-NEXT:    [[CMP:%.*]] = icmp ult i32 [[ADD]], 253
+; CHECK-NEXT:    ret i1 [[CMP]]
+;
+entry:
+  %conv = zext i8 %x to i32
+  %add = add i32 %conv, -4
+  %cmp = icmp ult i32 %add, 253
+  ret i1 %cmp
+}