[CIR] Upstream handling of integral-to-pointer casts #161653

kimsh02 · 2025-10-02T11:40:21Z

Fix #153658

Handle integral to pointer casts and port the relevant cast.cpp test cases from incubator.

llvmbot · 2025-10-02T11:40:56Z

@llvm/pr-subscribers-clangir

@llvm/pr-subscribers-clang

Author: Shawn K (kimsh02)

Changes

Fix #153658

Handle integral to pointer casts and port the relevant cast.cpp test cases from incubator.

Full diff: https://github.com/llvm/llvm-project/pull/161653.diff

3 Files Affected:

(modified) clang/include/clang/CIR/Dialect/IR/CIRDataLayout.h (+18)
(modified) clang/lib/CIR/CodeGen/CIRGenExprScalar.cpp (+19)
(modified) clang/test/CIR/CodeGen/cast.cpp (+22)

diff --git a/clang/include/clang/CIR/Dialect/IR/CIRDataLayout.h b/clang/include/clang/CIR/Dialect/IR/CIRDataLayout.h
index 417a226e44cbf..c7450d8770714 100644
--- a/clang/include/clang/CIR/Dialect/IR/CIRDataLayout.h
+++ b/clang/include/clang/CIR/Dialect/IR/CIRDataLayout.h
@@ -14,6 +14,11 @@
 
 #include "mlir/Dialect/DLTI/DLTI.h"
 #include "mlir/IR/BuiltinOps.h"
+#include "clang/CIR/Dialect/IR/CIRAttrs.h"
+#include "clang/CIR/Dialect/IR/CIRTypes.h"
+#include "llvm/IR/DataLayout.h"
+#include "llvm/Support/Alignment.h"
+#include "llvm/Support/TypeSize.h"
 
 namespace cir {
 
@@ -81,6 +86,19 @@ class CIRDataLayout {
   }
 
   llvm::TypeSize getTypeSizeInBits(mlir::Type ty) const;
+
+    llvm::TypeSize getPointerTypeSizeInBits(mlir::Type Ty) const {
+    assert(mlir::isa<cir::PointerType>(Ty) &&
+           "This should only be called with a pointer type");
+    return layout.getTypeSizeInBits(Ty);
+  }
+
+  mlir::Type getIntPtrType(mlir::Type Ty) const {
+    assert(mlir::isa<cir::PointerType>(Ty) && "Expected pointer type");
+    auto IntTy =
+        cir::IntType::get(Ty.getContext(), getPointerTypeSizeInBits(Ty), false);
+    return IntTy;
+  }
 };
 
 } // namespace cir
diff --git a/clang/lib/CIR/CodeGen/CIRGenExprScalar.cpp b/clang/lib/CIR/CodeGen/CIRGenExprScalar.cpp
index 500007f6f241b..c64c5f7d7a9c8 100644
--- a/clang/lib/CIR/CodeGen/CIRGenExprScalar.cpp
+++ b/clang/lib/CIR/CodeGen/CIRGenExprScalar.cpp
@@ -1889,6 +1889,25 @@ mlir::Value ScalarExprEmitter::VisitCastExpr(CastExpr *ce) {
     }
     return v;
   }
+  case CK_IntegralToPointer: {
+    auto DestCIRTy = cgf.convertType(destTy);
+    mlir::Value Src = Visit(const_cast<Expr *>(subExpr));
+
+    // Properly resize by casting to an int of the same size as the pointer.
+    // Clang's IntegralToPointer includes 'bool' as the source, but in CIR
+    // 'bool' is not an integral type.  So check the source type to get the
+    // correct CIR conversion.
+    auto MiddleTy = cgf.cgm.getDataLayout().getIntPtrType(DestCIRTy);
+    auto MiddleVal = builder.createCast(subExpr->getType()->isBooleanType()
+                                            ? cir::CastKind::bool_to_int
+                                            : cir::CastKind::integral,
+                                        Src, MiddleTy);
+
+    if (cgf.cgm.getCodeGenOpts().StrictVTablePointers)
+      llvm_unreachable("NYI");
+
+    return builder.createIntToPtr(MiddleVal, DestCIRTy);
+  }  
 
   case CK_ArrayToPointerDecay:
     return cgf.emitArrayToPointerDecay(subExpr).getPointer();
diff --git a/clang/test/CIR/CodeGen/cast.cpp b/clang/test/CIR/CodeGen/cast.cpp
index 7afa955cf3bcf..a1fd92612fe37 100644
--- a/clang/test/CIR/CodeGen/cast.cpp
+++ b/clang/test/CIR/CodeGen/cast.cpp
@@ -131,3 +131,25 @@ void bitcast() {
 
 // LLVM: %[[D_VEC:.*]] = load <2 x double>, ptr {{.*}}, align 16
 // LLVM: %[[I_VEC:.*]] = bitcast <2 x double> %[[D_VEC]] to <4 x i32>
+
+void f(long int start) {
+  void *p = (void*)start;
+}
+// CIR: %[[L:.*]] = cir.load {{.*}} : !cir.ptr<!s64i>, !s64i
+// CIR: %[[MID:.*]] = cir.cast integral %[[L]] : !s64i -> !u64i
+// CIR:          cir.cast int_to_ptr %[[MID]] : !u64i -> !cir.ptr<!void>
+
+struct A { int x; };
+
+void int_cast(long ptr) {
+  ((A *)ptr)->x = 0;
+}
+// CIR: cir.cast int_to_ptr {{.*}} : !u64i -> !cir.ptr<!rec_A>
+// LLVM: inttoptr {{.*}} to ptr
+
+void null_cast(long) {
+  *(int *)0 = 0;
+  ((A *)0)->x = 0;
+}
+// CIR: #cir.ptr<null> : !cir.ptr<!s32i>
+// CIR: #cir.ptr<null> : !cir.ptr<!rec_A>

github-actions · 2025-10-02T11:42:48Z

⚠️ C/C++ code formatter, clang-format found issues in your code. ⚠️

You can test this locally with the following command:

git-clang-format --diff origin/main HEAD --extensions h,cpp -- clang/include/clang/CIR/Dialect/IR/CIRDataLayout.h clang/lib/CIR/CodeGen/CIRGenExprScalar.cpp clang/test/CIR/CodeGen/cast.cpp

⚠️
The reproduction instructions above might return results for more than one PR
in a stack if you are using a stacked PR workflow. You can limit the results by
changing origin/main to the base branch/commit you want to compare against.
⚠️

View the diff from clang-format here.

diff --git a/clang/include/clang/CIR/Dialect/IR/CIRDataLayout.h b/clang/include/clang/CIR/Dialect/IR/CIRDataLayout.h
index 4eeb27d15..0ada74c69 100644
--- a/clang/include/clang/CIR/Dialect/IR/CIRDataLayout.h
+++ b/clang/include/clang/CIR/Dialect/IR/CIRDataLayout.h
@@ -95,7 +95,8 @@ public:
 
   mlir::Type getIntPtrType(mlir::Type Ty) const {
     assert(mlir::isa<cir::PointerType>(Ty) && "Expected pointer type");
-    return cir::IntType::get(ty.getContext(), getPointerTypeSizeInBits(ty), false);
+    return cir::IntType::get(ty.getContext(), getPointerTypeSizeInBits(ty),
+                             false);
   }
 };
 
diff --git a/clang/lib/CIR/CodeGen/CIRGenExprScalar.cpp b/clang/lib/CIR/CodeGen/CIRGenExprScalar.cpp
index 7ee8c3b08..05ed0c4ee 100644
--- a/clang/lib/CIR/CodeGen/CIRGenExprScalar.cpp
+++ b/clang/lib/CIR/CodeGen/CIRGenExprScalar.cpp
@@ -1890,7 +1890,8 @@ mlir::Value ScalarExprEmitter::VisitCastExpr(CastExpr *ce) {
     return v;
   }
   case CK_IntegralToPointer: {
-    mlir:Type destCIRTy = cgf.convertType(destTy);
+  mlir:
+    Type destCIRTy = cgf.convertType(destTy);
     mlir::Value src = Visit(const_cast<Expr *>(subExpr));
 
     // Properly resize by casting to an int of the same size as the pointer.
@@ -1898,13 +1899,14 @@ mlir::Value ScalarExprEmitter::VisitCastExpr(CastExpr *ce) {
     // 'bool' is not an integral type.  So check the source type to get the
     // correct CIR conversion.
     mlirType middleTy = cgf.cgm.getDataLayout().getIntPtrType(destCIRTy);
-    cir::CastOp middleVal = builder.createCast(subExpr->getType()->isBooleanType()
-                                            ? cir::CastKind::bool_to_int
+    cir::CastOp middleVal = builder.createCast(
+        subExpr->getType()->isBooleanType() ? cir::CastKind::bool_to_int
                                             : cir::CastKind::integral,
-                                        src, middleTy);
+        src, middleTy);
 
     if (cgf.cgm.getCodeGenOpts().StrictVTablePointers) {
-      cgf.cgm.errorNYI(subExpr->getSourceRange(), "IntegralToPointer: strict vtable pointers");
+      cgf.cgm.errorNYI(subExpr->getSourceRange(),
+                       "IntegralToPointer: strict vtable pointers");
       return {};
     }

andykaylor

Thanks for the PR!

What you've done here is basically correct. I have a lot of comments, but they're all just a matter of adapting to the expectations for upstreaming CIR code. I hope you won't be discouraged.

andykaylor · 2025-10-02T15:39:18Z

clang/include/clang/CIR/Dialect/IR/CIRDataLayout.h

+#include "clang/CIR/Dialect/IR/CIRAttrs.h"
+#include "clang/CIR/Dialect/IR/CIRTypes.h"
+#include "llvm/IR/DataLayout.h"
+#include "llvm/Support/Alignment.h"
+#include "llvm/Support/TypeSize.h"


Are all of these headers needed for your change?

https://llvm.org/docs/CodingStandards.html#include-as-little-as-possible

clang/lib/CIR/CodeGen/CIRGenExprScalar.cpp

andykaylor · 2025-10-02T15:53:30Z

clang/test/CIR/CodeGen/cast.cpp

+}
+// CIR: %[[L:.*]] = cir.load {{.*}} : !cir.ptr<!s64i>, !s64i
+// CIR: %[[MID:.*]] = cir.cast integral %[[L]] : !s64i -> !u64i
+// CIR:          cir.cast int_to_ptr %[[MID]] : !u64i -> !cir.ptr<!void>


Can you add LLVM checks here?

andykaylor · 2025-10-02T15:56:53Z

clang/test/CIR/CodeGen/cast.cpp

+// CIR: #cir.ptr<null> : !cir.ptr<!s32i>
+// CIR: #cir.ptr<null> : !cir.ptr<!rec_A>


These checks need to be expanded. The CIR generated should look like this:

%1 = cir.const #cir.int<0> : !s32i %2 = cir.const #cir.ptr<null> : !cir.ptr<!s32i> cir.store align(4) %1, %2 : !s32i, !cir.ptr<!s32i> %3 = cir.const #cir.int<0> : !s32i %4 = cir.const #cir.ptr<null> : !cir.ptr<!rec_A> %5 = cir.get_member %4[0] {name = "x"} : !cir.ptr<!rec_A> -> !cir.ptr<!s32i> cir.store align(4) %3, %5 : !s32i, !cir.ptr<!s32i>

The point of the test is to verify that we're using a null pointer in the expected instructions, so I'd check at least that (and always use regex substitutions for SSA values):

// CIR: %[[NULLPTR:.*]] = cir.const #cir.ptr<null> : !cir.ptr<!s32i> // CIR: cir.store{{.*}} %{{.*}}, %[[NULLPTR]] : !s32i, !cir.ptr<!s32i> // CIR: %[[NULLPTR_A:.*]] = cir.const #cir.ptr<null> : !cir.ptr<!rec_A> // CIR: %[[A_X:.*]] = cir.get_member %[[NULLPTR_A]][0] {name = "x"} : !cir.ptr<!rec_A> -> !cir.ptr<!s32i>

Note that I've used %{{.*}} for the value being stored in the second line, because we don't need to test that here. Also note, I used {{.*}} following cir.store so that this test will pass with or without an alignment specifier, because that's not relevant to this test and we wouldn't want to have to update this test if alignment behavior is changed in the future.

Nothing specific to casts is being lowered to LLVM IR, so I don't think LLVM checks are needed here.

Co-authored-by: Andy Kaylor <akaylor@nvidia.com>

kimsh02 · 2025-10-02T17:20:05Z

Thanks for all the helpful feedback! I've applied the suggested commits, and I'll revisit the headers and IR checks.

andykaylor · 2025-10-02T19:14:01Z

Thanks for all the helpful feedback! I've applied the suggested commits, and I'll revisit the headers and IR checks.

My suggested changes didn't include all uses of the variables that were renamed, so you'll need to do that manually, as well as fixing formatting issues.

AmrDeveloper

Thank you @kimsh02 for the PR!

I added some small suggestions

clang/include/clang/CIR/Dialect/IR/CIRDataLayout.h

clang/lib/CIR/CodeGen/CIRGenExprScalar.cpp

Co-authored-by: Amr Hesham <amr96@programmer.net>

kimsh02 · 2025-10-03T11:25:05Z

Thanks, @AmrDeveloper ! I've applied the suggested commits.

kimsh02 · 2025-10-03T11:27:07Z

@andykaylor Yes, thanks for heads up. I will be properly revisiting everything.

AmrDeveloper · 2025-10-04T11:23:37Z

@kimsh02 For code formatting, you can check it locally using git clang-format HEAD~1 (Change 1 for the number of commits you want to check)

https://llvm.org/docs/Contributing.html#how-to-submit-a-patch

Note: For the "Build and Test Linux" the current issue is not related to your PR

kimsh02 · 2025-10-04T16:23:48Z

@AmrDeveloper Thanks!

[CIR] Upstream handling of integral-to-pointer casts

eec6869

kimsh02 requested review from lanza, bcardosolopes, xlauko and andykaylor as code owners October 2, 2025 11:40

llvmbot added clang Clang issues not falling into any other category ClangIR Anything related to the ClangIR project labels Oct 2, 2025

Clang-format

54e154c

andykaylor reviewed Oct 2, 2025

View reviewed changes

kimsh02 and others added 5 commits October 2, 2025 10:10

Update clang/lib/CIR/CodeGen/CIRGenExprScalar.cpp

2807388

Co-authored-by: Andy Kaylor <akaylor@nvidia.com>

Update clang/lib/CIR/CodeGen/CIRGenExprScalar.cpp

d328d27

Co-authored-by: Andy Kaylor <akaylor@nvidia.com>

Update clang/lib/CIR/CodeGen/CIRGenExprScalar.cpp

ec1ca90

Co-authored-by: Andy Kaylor <akaylor@nvidia.com>

Update clang/lib/CIR/CodeGen/CIRGenExprScalar.cpp

2b3d7cb

Co-authored-by: Andy Kaylor <akaylor@nvidia.com>

Update clang/lib/CIR/CodeGen/CIRGenExprScalar.cpp

9a9e05d

Co-authored-by: Andy Kaylor <akaylor@nvidia.com>

AmrDeveloper reviewed Oct 3, 2025

View reviewed changes

kimsh02 and others added 5 commits October 3, 2025 04:16

Update clang/include/clang/CIR/Dialect/IR/CIRDataLayout.h

119873c

Co-authored-by: Amr Hesham <amr96@programmer.net>

Update clang/include/clang/CIR/Dialect/IR/CIRDataLayout.h

c321248

Co-authored-by: Amr Hesham <amr96@programmer.net>

Update clang/lib/CIR/CodeGen/CIRGenExprScalar.cpp

5866289

Co-authored-by: Amr Hesham <amr96@programmer.net>

Update clang/lib/CIR/CodeGen/CIRGenExprScalar.cpp

728ec84

Co-authored-by: Amr Hesham <amr96@programmer.net>

Update clang/lib/CIR/CodeGen/CIRGenExprScalar.cpp

142e70f

Co-authored-by: Amr Hesham <amr96@programmer.net>

		// CIR: #cir.ptr<null> : !cir.ptr<!s32i>
		// CIR: #cir.ptr<null> : !cir.ptr<!rec_A>

[CIR] Upstream handling of integral-to-pointer casts #161653

Are you sure you want to change the base?

[CIR] Upstream handling of integral-to-pointer casts #161653

Conversation

kimsh02 commented Oct 2, 2025

Uh oh!

llvmbot commented Oct 2, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

github-actions bot commented Oct 2, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

andykaylor left a comment

Choose a reason for hiding this comment

Uh oh!

andykaylor Oct 2, 2025

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

andykaylor Oct 2, 2025

Choose a reason for hiding this comment

Uh oh!

andykaylor Oct 2, 2025

Choose a reason for hiding this comment

Uh oh!

kimsh02 commented Oct 2, 2025

Uh oh!

andykaylor commented Oct 2, 2025

Uh oh!

AmrDeveloper left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

kimsh02 commented Oct 3, 2025

Uh oh!

kimsh02 commented Oct 3, 2025

Uh oh!

AmrDeveloper commented Oct 4, 2025

Uh oh!

kimsh02 commented Oct 4, 2025

Uh oh!

Uh oh!

llvmbot commented Oct 2, 2025 •

edited

Loading

github-actions bot commented Oct 2, 2025 •

edited

Loading