[clang] [Clang] Fix assertion failure when storing to ext_vector_type bool elements (PR #189305)
Kartik Ohlan via cfe-commits
cfe-commits at lists.llvm.org
Sun Apr 5 22:13:03 PDT 2026
https://github.com/Ko496-glitch updated https://github.com/llvm/llvm-project/pull/189305
>From 2d47db0ff28f869363ed3808a82d6996dd07c2c5 Mon Sep 17 00:00:00 2001
From: kartikohlan <kartik7ohlan at gmail.com>
Date: Sun, 29 Mar 2026 23:49:47 -0400
Subject: [PATCH 1/4] [Clang] Fix assertion failure (#189260)
---
clang/lib/CodeGen/CGExpr.cpp | 4 ++++
1 file changed, 4 insertions(+)
diff --git a/clang/lib/CodeGen/CGExpr.cpp b/clang/lib/CodeGen/CGExpr.cpp
index 23802cdeb4811..737d4fd145a4d 100644
--- a/clang/lib/CodeGen/CGExpr.cpp
+++ b/clang/lib/CodeGen/CGExpr.cpp
@@ -2781,6 +2781,10 @@ void CodeGenFunction::EmitStoreThroughLValue(RValue Src, LValue Dst,
Builder.getInt1Ty(), IRStoreTy->getPrimitiveSizeInBits());
Vec = Builder.CreateBitCast(Vec, IRVecTy);
// iN --> <N x i1>.
+
+ if(SrcVal->getType() != Builder.getInt1Ty())
+ SrcVal = Builder.CreateTrunc(SrcVal,Builder.getInt1Ty());
+ }
}
// Allow inserting `<1 x T>` into an `<N x T>`. It can happen with scalar
>From 45a41f0bc15e448ab22921d14c7f39acae8d83b4 Mon Sep 17 00:00:00 2001
From: kartikohlan <kartik7ohlan at gmail.com>
Date: Mon, 30 Mar 2026 18:46:21 -0400
Subject: [PATCH 2/4] Added the exisiting practise and fixed the format
---
clang/lib/CodeGen/CGExpr.cpp | 5 ++---
clang/test/CodeGen/ext-vector-bool-read.cpp | 21 +++++++++++++++++++++
2 files changed, 23 insertions(+), 3 deletions(-)
create mode 100644 clang/test/CodeGen/ext-vector-bool-read.cpp
diff --git a/clang/lib/CodeGen/CGExpr.cpp b/clang/lib/CodeGen/CGExpr.cpp
index 737d4fd145a4d..20a38b6367df5 100644
--- a/clang/lib/CodeGen/CGExpr.cpp
+++ b/clang/lib/CodeGen/CGExpr.cpp
@@ -2782,9 +2782,8 @@ void CodeGenFunction::EmitStoreThroughLValue(RValue Src, LValue Dst,
Vec = Builder.CreateBitCast(Vec, IRVecTy);
// iN --> <N x i1>.
- if(SrcVal->getType() != Builder.getInt1Ty())
- SrcVal = Builder.CreateTrunc(SrcVal,Builder.getInt1Ty());
- }
+ SrcVal = Builder.CreateIntCast(SrcVal, Builder.getInt1Ty(),
+ /*isSigned=*/false);
}
// Allow inserting `<1 x T>` into an `<N x T>`. It can happen with scalar
diff --git a/clang/test/CodeGen/ext-vector-bool-read.cpp b/clang/test/CodeGen/ext-vector-bool-read.cpp
new file mode 100644
index 0000000000000..358dc5919b245
--- /dev/null
+++ b/clang/test/CodeGen/ext-vector-bool-read.cpp
@@ -0,0 +1,21 @@
+// RUN: %clang_cc1 -triple x86_64-unknown-linux-gnu -emit-llvm -o - %s | FileCheck %s
+// RUN: %clang_cc1 -triple x86_64-unknown-linux-gnu -emit-llvm -O2 -o - %s | FileCheck %s
+
+// Regression test for GH#189260: Clang crashed with an assertion failure
+// in InsertElementInst when storing to an element of an ext_vector_type
+// with bool element type.
+
+typedef __attribute__((ext_vector_type(32))) bool v32bool;
+v32bool v32b = {};
+
+// CHECK-LABEL: @_Z5test1v
+void test1() {
+ // CHECK: insertelement <32 x i1>
+ v32b[0] = true;
+}
+
+// CHECK-LABEL: @_Z5test2v
+void test2() {
+ // CHECK: insertelement <32 x i1>
+ v32b[31] = true;
+}
>From be6ade7e8e79586bcf3ad73fe9b981a679e972b0 Mon Sep 17 00:00:00 2001
From: kartikohlan <kartik7ohlan at gmail.com>
Date: Tue, 31 Mar 2026 08:32:31 -0400
Subject: [PATCH 3/4] Added release notes for #189260
---
clang/docs/ReleaseNotes.rst | 13 +++++++------
1 file changed, 7 insertions(+), 6 deletions(-)
diff --git a/clang/docs/ReleaseNotes.rst b/clang/docs/ReleaseNotes.rst
index fc17eead02589..1ba369ff9f925 100644
--- a/clang/docs/ReleaseNotes.rst
+++ b/clang/docs/ReleaseNotes.rst
@@ -113,7 +113,7 @@ Clang Python Bindings Potentially Breaking Changes
``locations`` argument are passed. Previousy, ``locations`` took precedence.
- ``_CXUnsavedFile`` will be renamed to ``UnsavedFile`` for consistency.
``UnsavedFile`` is already available to use and existing uses should
- be adapted to refer to it instead. ``_CXUnsavedFile`` will be removed in a
+ be adapted to refer to it instead. ``_CXUnsavedFile`` will be removed in a
future release.
What's New in Clang |release|?
@@ -324,7 +324,7 @@ Improvements to Clang's diagnostics
when accessing a member function on a past-the-end array element.
(#GH179128)
-- Added a missing space to the FixIt for the ``implicit-int`` group of diagnostics and
+- Added a missing space to the FixIt for the ``implicit-int`` group of diagnostics and
made sure that only one such diagnostic and FixIt is emitted per declaration group. (#GH179354)
- Fixed the Fix-It insertion point for ``expected ';' after alias declaration``
@@ -337,7 +337,7 @@ Improvements to Clang's diagnostics
- The ``-Wloop-analysis`` warning has been extended to catch more cases of
variable modification inside lambda expressions (#GH132038).
-- Clang now emits ``-Wsizeof-pointer-memaccess`` when snprintf/vsnprintf use the sizeof
+- Clang now emits ``-Wsizeof-pointer-memaccess`` when snprintf/vsnprintf use the sizeof
the destination buffer(dynamically allocated) in the len parameter(#GH162366)
- Added ``-Wmodule-map-path-outside-directory`` (off by default) to warn on
@@ -384,6 +384,7 @@ Bug Fixes in This Version
- Correctly diagnosing and no longer crashing when ``export module foo``
(without a semicolon) are the final tokens in a module file. (#GH187771)
- Fixed a crash in duplicate attribute checking caused by comparing constant arguments with different integer signedness. (#GH188259)
+- Fixed a crash when assigning to an element of an ``ext_vector_type`` with ``bool`` element type. (#GH189260)
Bug Fixes to Compiler Builtins
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
@@ -406,7 +407,7 @@ Bug Fixes to C++ Support
- Fixed a crash when diagnosing an invalid static member function with an explicit object parameter (#GH177741)
- Fixed a crash when instantiating an invalid out-of-line static data member definition in a local class. (#GH176152)
- Fixed a crash when pack expansions are used as arguments for non-pack parameters of built-in templates. (#GH180307)
-- Fixed a bug where captured variables in non-mutable lambdas were incorrectly treated as mutable
+- Fixed a bug where captured variables in non-mutable lambdas were incorrectly treated as mutable
when used inside decltype in the return type. (#GH180460)
- Fixed a crash when evaluating uninitialized GCC vector/ext_vector_type vectors in ``constexpr``. (#GH180044)
- Fixed a crash when `explicit(bool)` is used with an incomplete enumeration. (#GH183887)
@@ -414,8 +415,8 @@ Bug Fixes to C++ Support
- Fixed a crash when an immediate-invoked ``consteval`` lambda is used as an invalid initializer. (#GH185270)
- Fixed an assertion failure when using a global destructor with a target with a non-default program address space. (#GH186484)
-- Inherited constructors in ``dllexport`` classes are now exported for ABI-compatible cases, matching
- MSVC behavior. Constructors with variadic arguments or callee-cleanup parameters are not yet supported
+- Inherited constructors in ``dllexport`` classes are now exported for ABI-compatible cases, matching
+ MSVC behavior. Constructors with variadic arguments or callee-cleanup parameters are not yet supported
and produce a warning. (#GH162640)
- Fix initialization of GRO when GRO-return type mismatches, as part of CWG2563. (#GH98744)
>From 7c0ffc888eb6320b9799277e1fddc216ca8b5f06 Mon Sep 17 00:00:00 2001
From: kartikohlan <kartik7ohlan at gmail.com>
Date: Fri, 3 Apr 2026 12:59:39 -0400
Subject: [PATCH 4/4] Added the update_cc_test_checks.py checks and added
VecTy->isVectorTy()
---
clang/lib/CodeGen/CGExpr.cpp | 7 +--
clang/test/CodeGen/ext-vector-bool-read.cpp | 47 ++++++++++++++++++---
2 files changed, 43 insertions(+), 11 deletions(-)
diff --git a/clang/lib/CodeGen/CGExpr.cpp b/clang/lib/CodeGen/CGExpr.cpp
index 20a38b6367df5..2ef399a7e7309 100644
--- a/clang/lib/CodeGen/CGExpr.cpp
+++ b/clang/lib/CodeGen/CGExpr.cpp
@@ -2771,8 +2771,8 @@ void CodeGenFunction::EmitStoreThroughLValue(RValue Src, LValue Dst,
llvm::Type *VecTy = Vec->getType();
llvm::Value *SrcVal = Src.getScalarVal();
- if (SrcVal->getType()->getPrimitiveSizeInBits() <
- VecTy->getScalarSizeInBits())
+ if (VecTy->isVectorTy() && SrcVal->getType()->getPrimitiveSizeInBits() <
+ VecTy->getScalarSizeInBits())
SrcVal = Builder.CreateZExt(SrcVal, VecTy->getScalarType());
auto *IRStoreTy = dyn_cast<llvm::IntegerType>(Vec->getType());
@@ -2781,9 +2781,6 @@ void CodeGenFunction::EmitStoreThroughLValue(RValue Src, LValue Dst,
Builder.getInt1Ty(), IRStoreTy->getPrimitiveSizeInBits());
Vec = Builder.CreateBitCast(Vec, IRVecTy);
// iN --> <N x i1>.
-
- SrcVal = Builder.CreateIntCast(SrcVal, Builder.getInt1Ty(),
- /*isSigned=*/false);
}
// Allow inserting `<1 x T>` into an `<N x T>`. It can happen with scalar
diff --git a/clang/test/CodeGen/ext-vector-bool-read.cpp b/clang/test/CodeGen/ext-vector-bool-read.cpp
index 358dc5919b245..527cf160c1614 100644
--- a/clang/test/CodeGen/ext-vector-bool-read.cpp
+++ b/clang/test/CodeGen/ext-vector-bool-read.cpp
@@ -1,5 +1,6 @@
-// RUN: %clang_cc1 -triple x86_64-unknown-linux-gnu -emit-llvm -o - %s | FileCheck %s
-// RUN: %clang_cc1 -triple x86_64-unknown-linux-gnu -emit-llvm -O2 -o - %s | FileCheck %s
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --version 6
+// RUN: %clang_cc1 -triple x86_64-unknown-linux-gnu -emit-llvm -o - %s | FileCheck %s --check-prefix=CHECK
+// RUN: %clang_cc1 -triple x86_64-unknown-linux-gnu -emit-llvm -O2 -o - %s | FileCheck %s --check-prefix=OPT
// Regression test for GH#189260: Clang crashed with an assertion failure
// in InsertElementInst when storing to an element of an ext_vector_type
@@ -8,14 +9,48 @@
typedef __attribute__((ext_vector_type(32))) bool v32bool;
v32bool v32b = {};
-// CHECK-LABEL: @_Z5test1v
+// CHECK-LABEL: define dso_local void @_Z5test1v(
+// CHECK-SAME: ) #[[ATTR0:[0-9]+]] {
+// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: [[TMP0:%.*]] = load i32, ptr @v32b, align 4
+// CHECK-NEXT: [[TMP1:%.*]] = bitcast i32 [[TMP0]] to <32 x i1>
+// CHECK-NEXT: [[VECINS:%.*]] = insertelement <32 x i1> [[TMP1]], i1 true, i32 0
+// CHECK-NEXT: [[TMP2:%.*]] = bitcast <32 x i1> [[VECINS]] to i32
+// CHECK-NEXT: store i32 [[TMP2]], ptr @v32b, align 4
+// CHECK-NEXT: ret void
+//
+// OPT-LABEL: define dso_local void @_Z5test1v(
+// OPT-SAME: ) local_unnamed_addr #[[ATTR0:[0-9]+]] {
+// OPT-NEXT: [[ENTRY:.*:]]
+// OPT-NEXT: [[TMP0:%.*]] = load <32 x i1>, ptr @v32b, align 4
+// OPT-NEXT: [[VECINS:%.*]] = insertelement <32 x i1> [[TMP0]], i1 true, i64 0
+// OPT-NEXT: store <32 x i1> [[VECINS]], ptr @v32b, align 4
+// OPT-NEXT: ret void
+//
void test1() {
- // CHECK: insertelement <32 x i1>
+
v32b[0] = true;
}
-// CHECK-LABEL: @_Z5test2v
+// CHECK-LABEL: define dso_local void @_Z5test2v(
+// CHECK-SAME: ) #[[ATTR0]] {
+// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: [[TMP0:%.*]] = load i32, ptr @v32b, align 4
+// CHECK-NEXT: [[TMP1:%.*]] = bitcast i32 [[TMP0]] to <32 x i1>
+// CHECK-NEXT: [[VECINS:%.*]] = insertelement <32 x i1> [[TMP1]], i1 true, i32 31
+// CHECK-NEXT: [[TMP2:%.*]] = bitcast <32 x i1> [[VECINS]] to i32
+// CHECK-NEXT: store i32 [[TMP2]], ptr @v32b, align 4
+// CHECK-NEXT: ret void
+//
+// OPT-LABEL: define dso_local void @_Z5test2v(
+// OPT-SAME: ) local_unnamed_addr #[[ATTR0]] {
+// OPT-NEXT: [[ENTRY:.*:]]
+// OPT-NEXT: [[TMP0:%.*]] = load <32 x i1>, ptr @v32b, align 4
+// OPT-NEXT: [[VECINS:%.*]] = insertelement <32 x i1> [[TMP0]], i1 true, i64 31
+// OPT-NEXT: store <32 x i1> [[VECINS]], ptr @v32b, align 4
+// OPT-NEXT: ret void
+//
void test2() {
- // CHECK: insertelement <32 x i1>
+
v32b[31] = true;
}
More information about the cfe-commits
mailing list