[InstSimplify] Support ptrtoaddr in pointer subtraction fold #162672
Merged
+39
−1
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Add a new m_PtrToIntOrAddr() matcher which matches both ptrtoint and ptrtoaddr. Pointer arithmetic only works on the address bits, so supporting ptrtoaddr is always fine here.
llvmbot
added
llvm:instcombine
|
@llvm/pr-subscribers-llvm-ir @llvm/pr-subscribers-llvm-transforms Author: Nikita Popov (nikic) ChangesAdd a new m_PtrToIntOrAddr() matcher which matches both ptrtoint and ptrtoaddr. Pointer arithmetic only works on the address bits, so supporting ptrtoaddr is always fine here. Full diff: https://github.com/llvm/llvm-project/pull/162672.diff 3 Files Affected:
diff --git a/llvm/include/llvm/IR/PatternMatch.h b/llvm/include/llvm/IR/PatternMatch.h
index 2e31fe5bb431e..79342777aba4b 100644
--- a/llvm/include/llvm/IR/PatternMatch.h
+++ b/llvm/include/llvm/IR/PatternMatch.h
@@ -2184,6 +2184,18 @@ inline PtrToIntSameSize_match<OpTy> m_PtrToIntSameSize(const DataLayout &DL,
return PtrToIntSameSize_match<OpTy>(DL, Op);
}
+/// Matches PtrToAddr.
+template <typename OpTy>
+inline CastOperator_match<OpTy, Instruction::PtrToAddr>
+m_PtrToAddr(const OpTy &Op) {
+ return CastOperator_match<OpTy, Instruction::PtrToAddr>(Op);
+}
+
+/// Matches PtrToInt or PtrToAddr.
+template <typename OpTy> inline auto m_PtrToIntOrAddr(const OpTy &Op) {
+ return m_CombineOr(m_PtrToInt(Op), m_PtrToAddr(Op));
+}
+
/// Matches IntToPtr.
template <typename OpTy>
inline CastOperator_match<OpTy, Instruction::IntToPtr>
diff --git a/llvm/lib/Analysis/InstructionSimplify.cpp b/llvm/lib/Analysis/InstructionSimplify.cpp
index d1977f0b0daf9..71050a4117d5a 100644
--- a/llvm/lib/Analysis/InstructionSimplify.cpp
+++ b/llvm/lib/Analysis/InstructionSimplify.cpp
@@ -853,10 +853,12 @@ static Value *simplifySubInst(Value *Op0, Value *Op1, bool IsNSW, bool IsNUW,
return W;
// Variations on GEP(base, I, ...) - GEP(base, i, ...) -> GEP(null, I-i, ...).
- if (match(Op0, m_PtrToInt(m_Value(X))) && match(Op1, m_PtrToInt(m_Value(Y))))
+ if (match(Op0, m_PtrToIntOrAddr(m_Value(X))) &&
+ match(Op1, m_PtrToIntOrAddr(m_Value(Y)))) {
if (Constant *Result = computePointerDifference(Q.DL, X, Y))
return ConstantFoldIntegerCast(Result, Op0->getType(), /*IsSigned*/ true,
Q.DL);
+ }
// i1 sub -> xor.
if (MaxRecurse && Op0->getType()->isIntOrIntVectorTy(1))
diff --git a/llvm/test/Transforms/InstCombine/ptrtoaddr.ll b/llvm/test/Transforms/InstCombine/ptrtoaddr.ll
index 5211fbd8b897a..1de8ea7a662c2 100644
--- a/llvm/test/Transforms/InstCombine/ptrtoaddr.ll
+++ b/llvm/test/Transforms/InstCombine/ptrtoaddr.ll
@@ -130,3 +130,27 @@ define i32 @ptrtoaddr_sub_consts_offset_addrsize() {
;
ret i32 sub (i32 ptrtoaddr (ptr addrspace(1) getelementptr (i8, ptr addrspace(1) @g.as1, i32 42) to i32), i32 ptrtoaddr (ptr addrspace(1) @g.as1 to i32))
}
+
+define i64 @ptrtoaddr_sub_known_offset(ptr %p) {
+; CHECK-LABEL: define i64 @ptrtoaddr_sub_known_offset(
+; CHECK-SAME: ptr [[P:%.*]]) {
+; CHECK-NEXT: ret i64 42
+;
+ %p2 = getelementptr inbounds i8, ptr %p, i64 42
+ %p.addr = ptrtoaddr ptr %p to i64
+ %p2.addr = ptrtoaddr ptr %p2 to i64
+ %sub = sub i64 %p2.addr, %p.addr
+ ret i64 %sub
+}
+
+define i32 @ptrtoaddr_sub_known_offset_addrsize(ptr addrspace(1) %p) {
+; CHECK-LABEL: define i32 @ptrtoaddr_sub_known_offset_addrsize(
+; CHECK-SAME: ptr addrspace(1) [[P:%.*]]) {
+; CHECK-NEXT: ret i32 42
+;
+ %p2 = getelementptr inbounds i8, ptr addrspace(1) %p, i32 42
+ %p.addr = ptrtoaddr ptr addrspace(1) %p to i32
+ %p2.addr = ptrtoaddr ptr addrspace(1) %p2 to i32
+ %sub = sub i32 %p2.addr, %p.addr
+ ret i32 %sub
+}
|
|
@llvm/pr-subscribers-llvm-analysis Author: Nikita Popov (nikic) ChangesAdd a new m_PtrToIntOrAddr() matcher which matches both ptrtoint and ptrtoaddr. Pointer arithmetic only works on the address bits, so supporting ptrtoaddr is always fine here. Full diff: https://github.com/llvm/llvm-project/pull/162672.diff 3 Files Affected:
diff --git a/llvm/include/llvm/IR/PatternMatch.h b/llvm/include/llvm/IR/PatternMatch.h
index 2e31fe5bb431e..79342777aba4b 100644
--- a/llvm/include/llvm/IR/PatternMatch.h
+++ b/llvm/include/llvm/IR/PatternMatch.h
@@ -2184,6 +2184,18 @@ inline PtrToIntSameSize_match<OpTy> m_PtrToIntSameSize(const DataLayout &DL,
return PtrToIntSameSize_match<OpTy>(DL, Op);
}
+/// Matches PtrToAddr.
+template <typename OpTy>
+inline CastOperator_match<OpTy, Instruction::PtrToAddr>
+m_PtrToAddr(const OpTy &Op) {
+ return CastOperator_match<OpTy, Instruction::PtrToAddr>(Op);
+}
+
+/// Matches PtrToInt or PtrToAddr.
+template <typename OpTy> inline auto m_PtrToIntOrAddr(const OpTy &Op) {
+ return m_CombineOr(m_PtrToInt(Op), m_PtrToAddr(Op));
+}
+
/// Matches IntToPtr.
template <typename OpTy>
inline CastOperator_match<OpTy, Instruction::IntToPtr>
diff --git a/llvm/lib/Analysis/InstructionSimplify.cpp b/llvm/lib/Analysis/InstructionSimplify.cpp
index d1977f0b0daf9..71050a4117d5a 100644
--- a/llvm/lib/Analysis/InstructionSimplify.cpp
+++ b/llvm/lib/Analysis/InstructionSimplify.cpp
@@ -853,10 +853,12 @@ static Value *simplifySubInst(Value *Op0, Value *Op1, bool IsNSW, bool IsNUW,
return W;
// Variations on GEP(base, I, ...) - GEP(base, i, ...) -> GEP(null, I-i, ...).
- if (match(Op0, m_PtrToInt(m_Value(X))) && match(Op1, m_PtrToInt(m_Value(Y))))
+ if (match(Op0, m_PtrToIntOrAddr(m_Value(X))) &&
+ match(Op1, m_PtrToIntOrAddr(m_Value(Y)))) {
if (Constant *Result = computePointerDifference(Q.DL, X, Y))
return ConstantFoldIntegerCast(Result, Op0->getType(), /*IsSigned*/ true,
Q.DL);
+ }
// i1 sub -> xor.
if (MaxRecurse && Op0->getType()->isIntOrIntVectorTy(1))
diff --git a/llvm/test/Transforms/InstCombine/ptrtoaddr.ll b/llvm/test/Transforms/InstCombine/ptrtoaddr.ll
index 5211fbd8b897a..1de8ea7a662c2 100644
--- a/llvm/test/Transforms/InstCombine/ptrtoaddr.ll
+++ b/llvm/test/Transforms/InstCombine/ptrtoaddr.ll
@@ -130,3 +130,27 @@ define i32 @ptrtoaddr_sub_consts_offset_addrsize() {
;
ret i32 sub (i32 ptrtoaddr (ptr addrspace(1) getelementptr (i8, ptr addrspace(1) @g.as1, i32 42) to i32), i32 ptrtoaddr (ptr addrspace(1) @g.as1 to i32))
}
+
+define i64 @ptrtoaddr_sub_known_offset(ptr %p) {
+; CHECK-LABEL: define i64 @ptrtoaddr_sub_known_offset(
+; CHECK-SAME: ptr [[P:%.*]]) {
+; CHECK-NEXT: ret i64 42
+;
+ %p2 = getelementptr inbounds i8, ptr %p, i64 42
+ %p.addr = ptrtoaddr ptr %p to i64
+ %p2.addr = ptrtoaddr ptr %p2 to i64
+ %sub = sub i64 %p2.addr, %p.addr
+ ret i64 %sub
+}
+
+define i32 @ptrtoaddr_sub_known_offset_addrsize(ptr addrspace(1) %p) {
+; CHECK-LABEL: define i32 @ptrtoaddr_sub_known_offset_addrsize(
+; CHECK-SAME: ptr addrspace(1) [[P:%.*]]) {
+; CHECK-NEXT: ret i32 42
+;
+ %p2 = getelementptr inbounds i8, ptr addrspace(1) %p, i32 42
+ %p.addr = ptrtoaddr ptr addrspace(1) %p to i32
+ %p2.addr = ptrtoaddr ptr addrspace(1) %p2 to i32
+ %sub = sub i32 %p2.addr, %p.addr
+ ret i32 %sub
+}
|
arichardson
approved these changes
Oct 9, 2025
dtcxzyw
approved these changes
Oct 9, 2025
XChy
approved these changes
Oct 10, 2025
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
5 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Add a new m_PtrToIntOrAddr() matcher which matches both ptrtoint and ptrtoaddr. Pointer arithmetic only works on the address bits, so supporting ptrtoaddr is always fine here.