[clang] improve consistency with GCC vector comparison #148954

ImpleLee · 2025-07-15T20:37:16Z

fixes issue #132604

GCC's preference of vector comparison types should be int > signed char > short > long > long long, so I just duplicate that.
reference: gcc-help mailing list: https://gcc.gnu.org/pipermail/gcc-help/2025-July/144357.html

fixes issue llvm#132604

github-actions · 2025-07-15T20:37:37Z

Thank you for submitting a Pull Request (PR) to the LLVM Project!

This PR will be automatically labeled and the relevant teams will be notified.

If you wish to, you can add reviewers by using the "Reviewers" section on this page.

If this is not working for you, it is probably because you do not have write permissions for the repository. In which case you can instead tag reviewers by name in a comment by using @ followed by their GitHub username.

If you have received no comments on your PR for a week, you can request a review by "ping"ing the PR by adding a comment “Ping”. The common courtesy "ping" rate is once a week. Please remember that you are asking for valuable time from other developers.

If you have further questions, they may be answered by the LLVM GitHub User Guide.

You can also ask questions in a comment on this PR, on the LLVM Discord or on the forums.

llvmbot · 2025-07-15T20:38:09Z

@llvm/pr-subscribers-clang

Author: Imple Lee (ImpleLee)

Changes

fixes issue #132604

GCC's order of types of vector comparison should be int > signed char > short > long > long long, so I just duplicate that.
reference: gcc-help mailing list: https://gcc.gnu.org/pipermail/gcc-help/2025-July/144357.html

Full diff: https://github.com/llvm/llvm-project/pull/148954.diff

3 Files Affected:

(modified) clang/lib/Sema/SemaExpr.cpp (+11-11)
(modified) clang/test/Sema/vector-gcc-compat.c (+4-2)
(modified) clang/test/Sema/vector-gcc-compat.cpp (+4-2)

diff --git a/clang/lib/Sema/SemaExpr.cpp b/clang/lib/Sema/SemaExpr.cpp
index a3f534ee6712e..8a81cda9912f7 100644
--- a/clang/lib/Sema/SemaExpr.cpp
+++ b/clang/lib/Sema/SemaExpr.cpp
@@ -12901,24 +12901,24 @@ QualType Sema::GetSignedVectorType(QualType V) {
     return Context.getExtVectorType(Context.LongLongTy, VTy->getNumElements());
   }
 
-  if (TypeSize == Context.getTypeSize(Context.Int128Ty))
-    return Context.getVectorType(Context.Int128Ty, VTy->getNumElements(),
-                                 VectorKind::Generic);
-  if (TypeSize == Context.getTypeSize(Context.LongLongTy))
-    return Context.getVectorType(Context.LongLongTy, VTy->getNumElements(),
-                                 VectorKind::Generic);
-  if (TypeSize == Context.getTypeSize(Context.LongTy))
-    return Context.getVectorType(Context.LongTy, VTy->getNumElements(),
-                                 VectorKind::Generic);
   if (TypeSize == Context.getTypeSize(Context.IntTy))
     return Context.getVectorType(Context.IntTy, VTy->getNumElements(),
                                  VectorKind::Generic);
+  if (TypeSize == Context.getTypeSize(Context.SignedCharTy))
+    return Context.getVectorType(Context.SignedCharTy, VTy->getNumElements(),
+                                 VectorKind::Generic);
   if (TypeSize == Context.getTypeSize(Context.ShortTy))
     return Context.getVectorType(Context.ShortTy, VTy->getNumElements(),
                                  VectorKind::Generic);
-  assert(TypeSize == Context.getTypeSize(Context.CharTy) &&
+  if (TypeSize == Context.getTypeSize(Context.LongTy))
+    return Context.getVectorType(Context.LongTy, VTy->getNumElements(),
+                                 VectorKind::Generic);
+  if (TypeSize == Context.getTypeSize(Context.LongLongTy))
+    return Context.getVectorType(Context.LongLongTy, VTy->getNumElements(),
+                                 VectorKind::Generic);
+  assert(TypeSize == Context.getTypeSize(Context.Int128Ty) &&
          "Unhandled vector element size in vector compare");
-  return Context.getVectorType(Context.CharTy, VTy->getNumElements(),
+  return Context.getVectorType(Context.Int128Ty, VTy->getNumElements(),
                                VectorKind::Generic);
 }
 
diff --git a/clang/test/Sema/vector-gcc-compat.c b/clang/test/Sema/vector-gcc-compat.c
index 7764b3bf686ad..7d9d25019666c 100644
--- a/clang/test/Sema/vector-gcc-compat.c
+++ b/clang/test/Sema/vector-gcc-compat.c
@@ -3,6 +3,8 @@
 // Test the compatibility of clang's vector extensions with gcc's vector
 // extensions for C. Notably &&, ||, ?: and ! are not available.
 typedef long long v2i64 __attribute__((vector_size(16)));
+// this type is specifically used as the result type of comparison of v2i64
+typedef long v2i_long __attribute__((vector_size(16)));
 typedef int v2i32 __attribute__((vector_size(8)));
 typedef short v2i16 __attribute__((vector_size(4)));
 typedef char v2i8 __attribute__((vector_size(2)));
@@ -81,7 +83,7 @@ void arithmeticTest(void) {
 
 void comparisonTest(void) {
   v2i64 v2i64_a = (v2i64){0, 1};
-  v2i64 v2i64_r;
+  v2i_long v2i64_r;
 
   v2i64_r = v2i64_a == 1;
   v2i64_r = v2i64_a != 1;
@@ -166,7 +168,7 @@ void floatTestConstantComparison(void) {
 
 void doubleTestConstantComparison(void) {
   v2f64 v2f64_a = {0.4, 0.4};
-  v2i64 v2i64_r;
+  v2i_long v2i64_r;
   v2i64_r = v2f64_a > 0.4;
   v2i64_r = v2f64_a >= 0.4;
   v2i64_r = v2f64_a < 0.4;
diff --git a/clang/test/Sema/vector-gcc-compat.cpp b/clang/test/Sema/vector-gcc-compat.cpp
index 42c24d91ea8f3..e66cee128cd90 100644
--- a/clang/test/Sema/vector-gcc-compat.cpp
+++ b/clang/test/Sema/vector-gcc-compat.cpp
@@ -5,6 +5,8 @@
 // || operators work on vector types.
 
 typedef long long v2i64 __attribute__((vector_size(16))); // expected-warning {{'long long' is incompatible with C++98}}
+// this type is specifically used as the result type of comparison of v2i64
+typedef long v2i_long __attribute__((vector_size(16)));
 typedef int v2i32 __attribute__((vector_size(8)));
 typedef short v2i16 __attribute__((vector_size(4)));
 typedef char v2i8 __attribute__((vector_size(2)));
@@ -60,7 +62,7 @@ void arithmeticTest(void) {
 
 void comparisonTest(void) {
   v2i64 v2i64_a = (v2i64){0, 1}; // expected-warning {{compound literals are a C99-specific feature}}
-  v2i64 v2i64_r;
+  v2i_long v2i64_r;
 
   v2i64_r = v2i64_a == 1;
   v2i64_r = v2i64_a != 1;
@@ -141,7 +143,7 @@ void floatTestConstantComparison(void) {
 
 void doubleTestConstantComparison(void) {
   v2f64 v2f64_a = {0.4, 0.4};
-  v2i64 v2i64_r;
+  v2i_long v2i64_r;
   v2i64_r = v2f64_a > 0.4;
   v2i64_r = v2f64_a >= 0.4;
   v2i64_r = v2f64_a < 0.4;

clang/lib/Sema/SemaExpr.cpp

ImpleLee · 2025-07-17T19:12:28Z

Compared to int > signed char > short > long > long long, inconsistency happens when sizeof(int) == sizeof(short), which is true only in LP32 among currently widely accepted data models.

Looks like GCC uses the same order for __attribute__((mode)), so we get that attribute wrong in exactly the same way for LP32.

Maybe we should add a getGCCCompatibleIntTypeForBitwidth that prefers int over everything else, and use it in both places?

Originally posted by @zygoloid in #132604

I now prefer this proposal over the current PR. I will convert this PR into a draft PR when implementing it, and ask for review again after that.

They seem to be the same in GCC

ImpleLee · 2025-07-19T16:21:06Z

Now I make a utility function getGCCCompatibleIntTypeForBitwidth for this type preference. It should also be used for mode attribute.

There is currently no test about this, but I don't know how to or even whether should I write relevant tests.

There are many other places in clang that uses ASTContext::getIntTypeForBitwidth or TargetInfo::getIntTypeByWidth, but I don't know whether they should switch to this function.

erichkeane · 2025-07-21T13:35:25Z

clang/include/clang/AST/ASTContext.h

+  /// int > signed char > short > long > long long > int128_t
+  /// Returns empty type if there is no appropriate target types.
+  QualType getGCCCompatibleIntTypeForBitwidth(unsigned DestWidth,
+                                              unsigned Signed) const;


Suggested change

unsigned Signed) const;

bool Signed) const;

The parameters are intentionally the same as the parameters of getIntTypeForBitwidth.

llvm-project/clang/include/clang/AST/ASTContext.h

Lines 877 to 882 in abce4e9

/// getIntTypeForBitwidth -

/// sets integer QualTy according to specified details:

/// bitwidth, signed/unsigned.

/// Returns empty type if there is no appropriate target types.

QualType getIntTypeForBitwidth(unsigned DestWidth,

unsigned Signed) const;

Is this a tradition in LLVM? Or I should just use bool here?

erichkeane · 2025-07-21T13:35:42Z

clang/lib/AST/ASTContext.cpp

+/// int > signed char > short > long > long long > int128_t
+/// Returns empty type if there is no appropriate target types.
+QualType ASTContext::getGCCCompatibleIntTypeForBitwidth(unsigned DestWidth,
+                                                        unsigned Signed) const {


Suggested change

unsigned Signed) const {

bool Signed) const {

erichkeane · 2025-07-21T13:36:10Z

clang/lib/AST/ASTContext.cpp

+    return Signed ? LongLongTy : UnsignedLongLongTy;
+  if (DestWidth == 128)
+    return Signed ? Int128Ty : UnsignedInt128Ty;
+  return {};


We should probably assert if we don't find one.

The behavior of returning an empty or null type is also intentionally kept the same as getIntTypeForBitwidth. Maybe keeping this behavior can make this function a better in-place replacement of getIntTypeForBitwidth?

clang/lib/Sema/SemaExpr.cpp

mizvekov

LGTM sans the nits already pointed out. Thanks!

clang/lib/Sema/SemaExpr.cpp

mizvekov · 2025-07-21T17:53:25Z

clang/include/clang/AST/ASTContext.h

+  /// bitwidth, signed/unsigned.
+  /// this function is compatible with GCC's preference:
+  /// int > signed char > short > long > long long > int128_t
+  /// Returns empty type if there is no appropriate target types.


Suggested change

/// Returns empty type if there is no appropriate target types.

/// Returns null type if there is no appropriate target types.

The term empty type is also used in documentation of getIntTypeForBitwidth and getRealTypeForBitwidth series. I guess keeping them unified is probably better?

llvm-project/clang/include/clang/AST/ASTContext.h

Lines 877 to 888 in abce4e9

/// getIntTypeForBitwidth -

/// sets integer QualTy according to specified details:

/// bitwidth, signed/unsigned.

/// Returns empty type if there is no appropriate target types.

QualType getIntTypeForBitwidth(unsigned DestWidth,

unsigned Signed) const;

/// getRealTypeForBitwidth -

/// sets floating point QualTy according to specified bitwidth.

/// Returns empty type if there is no appropriate target types.

QualType getRealTypeForBitwidth(unsigned DestWidth,

FloatModeKind ExplicitType) const;

Yeah okay, that's pre-existing inconsistency (because in QualType itself we call this state 'Null'), so leaving it as 'empty' here is fine.

[clang] improve consistency with GCC vector comparison

3b19742

fixes issue llvm#132604

llvmbot added clang Clang issues not falling into any other category clang:frontend Language frontend issues, e.g. anything involving "Sema" labels Jul 15, 2025

hstk30-hw requested review from erichkeane and mizvekov July 16, 2025 08:14

erichkeane reviewed Jul 16, 2025

View reviewed changes

clang/lib/Sema/SemaExpr.cpp Outdated Show resolved Hide resolved

ImpleLee marked this pull request as draft July 17, 2025 19:12

ImpleLee added 2 commits July 20, 2025 00:11

use the logic of determining vector type for mode attribute

9f51aca

They seem to be the same in GCC

clang-format

dcde7fd

ImpleLee marked this pull request as ready for review July 19, 2025 16:14

erichkeane reviewed Jul 21, 2025

View reviewed changes

mizvekov approved these changes Jul 21, 2025

View reviewed changes

add proper comments for literal argument

53e63bc

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

[clang] improve consistency with GCC vector comparison #148954

[clang] improve consistency with GCC vector comparison #148954

ImpleLee commented Jul 15, 2025 •

edited

Loading

Uh oh!

github-actions bot commented Jul 15, 2025

Uh oh!

llvmbot commented Jul 15, 2025

Uh oh!

Uh oh!

ImpleLee commented Jul 17, 2025

Uh oh!

ImpleLee commented Jul 19, 2025 •

edited

Loading

Uh oh!

erichkeane Jul 21, 2025

Uh oh!

ImpleLee Jul 21, 2025

Uh oh!

erichkeane Jul 21, 2025

Uh oh!

erichkeane Jul 21, 2025

Uh oh!

ImpleLee Jul 21, 2025

Uh oh!

Uh oh!

mizvekov left a comment

Uh oh!

Uh oh!

mizvekov Jul 21, 2025

Uh oh!

ImpleLee Jul 21, 2025

Uh oh!

mizvekov Jul 21, 2025

Uh oh!

Uh oh!

	/// getIntTypeForBitwidth -
	/// sets integer QualTy according to specified details:
	/// bitwidth, signed/unsigned.
	/// Returns empty type if there is no appropriate target types.
	QualType getIntTypeForBitwidth(unsigned DestWidth,
	unsigned Signed) const;

	/// Returns empty type if there is no appropriate target types.
	/// Returns null type if there is no appropriate target types.

[clang] improve consistency with GCC vector comparison #148954

Are you sure you want to change the base?

[clang] improve consistency with GCC vector comparison #148954

Conversation

ImpleLee commented Jul 15, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

github-actions bot commented Jul 15, 2025

Uh oh!

llvmbot commented Jul 15, 2025

Uh oh!

Uh oh!

ImpleLee commented Jul 17, 2025

Uh oh!

ImpleLee commented Jul 19, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

mizvekov left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

ImpleLee commented Jul 15, 2025 •

edited

Loading

ImpleLee commented Jul 19, 2025 •

edited

Loading