[clang] [ObjC] Fix offsets following `[[no_unique_address]]` for `@encode()` (PR #71321)
https://github.com/BertalanD closed https://github.com/llvm/llvm-project/pull/71321 ___ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[clang] [ObjC] Fix offsets following `[[no_unique_address]]` for `@encode()` (PR #71321)
https://github.com/BertalanD created
https://github.com/llvm/llvm-project/pull/71321
Commit 46ca880fca made `@encode` skip fields that are made zero-sized by
`[[no_unique_address]]`. When iterating the fields, the index which is passed
to `getFieldOffset` failed to be incremented for those due to the use of an
early `continue`, so subsequent fields reported an incorrect offset. This
caused an assertion to be triggered in `getObjCEncodingForStructureImpl`.
Fixes #71250
>From 2ad862a174685a679d398b628769e88c5eebbee6 Mon Sep 17 00:00:00 2001
From: Daniel Bertalan
Date: Sun, 5 Nov 2023 16:19:50 +0100
Subject: [PATCH] [ObjC] Fix offsets following `[[no_unique_address]]` for
`@encode()`
Commit 46ca880fca made `@encode` skip fields that are made zero-sized by
`[[no_unique_address]]`. When iterating the fields, the index which is
passed to `getFieldOffset` failed to be incremented for those due to the
use of an early `continue`, so subsequent fields reported an incorrect
offset. This caused an assertion to be triggered in
`getObjCEncodingForStructureImpl`.
Fixes #71250
---
clang/lib/AST/ASTContext.cpp | 4 +---
clang/test/CodeGenObjCXX/encode.mm | 14 ++
2 files changed, 15 insertions(+), 3 deletions(-)
diff --git a/clang/lib/AST/ASTContext.cpp b/clang/lib/AST/ASTContext.cpp
index da90136752210b6..60146e9901bc44d 100644
--- a/clang/lib/AST/ASTContext.cpp
+++ b/clang/lib/AST/ASTContext.cpp
@@ -8535,14 +8535,12 @@ void
ASTContext::getObjCEncodingForStructureImpl(RecordDecl *RDecl,
}
}
- unsigned i = 0;
for (FieldDecl *Field : RDecl->fields()) {
if (!Field->isZeroLengthBitField(*this) && Field->isZeroSize(*this))
continue;
-uint64_t offs = layout.getFieldOffset(i);
+uint64_t offs = layout.getFieldOffset(Field->getFieldIndex());
FieldOrBaseOffsets.insert(FieldOrBaseOffsets.upper_bound(offs),
std::make_pair(offs, Field));
-++i;
}
if (CXXRec && includeVBases) {
diff --git a/clang/test/CodeGenObjCXX/encode.mm
b/clang/test/CodeGenObjCXX/encode.mm
index f382e7f23d77335..cad70e379c386bf 100644
--- a/clang/test/CodeGenObjCXX/encode.mm
+++ b/clang/test/CodeGenObjCXX/encode.mm
@@ -339,3 +339,17 @@ @implementation N
const char *inner0 = @encode(Outer0::Inner0 *);
const char *inner1 = @encode(Outer0::Inner1 *);
}
+
+#if __cplusplus >= 202002L
+namespace GH71250 {
+ struct Empty {};
+ struct S {
+[[no_unique_address]] Empty a;
+long b;
+long c;
+ };
+
+ // CHECKCXX20: @_ZN7GH712501sE = constant [7 x i8] c"{S=qq}\00", align 1
+ extern const char s[] = @encode(S);
+}
+#endif
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[clang] [ObjC] Fix offsets following `[[no_unique_address]]` for `@encode()` (PR #71321)
https://github.com/BertalanD unassigned https://github.com/llvm/llvm-project/pull/71321 ___ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[clang] [ObjC] Fix offsets following `[[no_unique_address]]` for `@encode()` (PR #71321)
https://github.com/BertalanD unassigned https://github.com/llvm/llvm-project/pull/71321 ___ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[clang] [clang] Implement constexpr evaluation for `__builtin_{add,sub}c` (PR #66005)
https://github.com/BertalanD created
https://github.com/llvm/llvm-project/pull/66005:
GCC has gained support for these multiprecision arithmetic builtins in
`r14-1896-g2b4e0415ad6`, and although they aren't explicitly specified as such
in the documentation, they are usable in a constexpr context.
This commit adds constexpr evaluation support to Clang to match GCC's behavior.
The implementation mirrors how the builtins are lowered to a pair of
`u{add,sub}.with.overflow` operations and the carryout is set to 1 if either of
those result in an overflow.
>From 00050a01970e47bbf3a9c8cdf29c41b699887900 Mon Sep 17 00:00:00 2001
From: Daniel Bertalan
Date: Sat, 22 Jul 2023 08:20:45 +0200
Subject: [PATCH] [clang] Implement constexpr evaluation for
`__builtin_{add,sub}c`
GCC has gained support for these multiprecision arithmetic builtins in
`r14-1896-g2b4e0415ad6`, and although they aren't explicitly specified
as such in the documentation, they are usable in a constexpr context.
This commit adds constexpr evaluation support to Clang to match GCC's
behavior. The implementation mirrors how the builtins are lowered to a
pair of `u{add,sub}.with.overflow` operations and the carryout is set to
1 if either of those result in an overflow.
---
clang/include/clang/Basic/Builtins.def| 20 ++--
clang/lib/AST/ExprConstant.cpp| 50 +
.../test/SemaCXX/builtins-multiprecision.cpp | 105 ++
3 files changed, 165 insertions(+), 10 deletions(-)
create mode 100644 clang/test/SemaCXX/builtins-multiprecision.cpp
diff --git a/clang/include/clang/Basic/Builtins.def
b/clang/include/clang/Basic/Builtins.def
index 586dcf05170eb58..8a54850cca672ba 100644
--- a/clang/include/clang/Basic/Builtins.def
+++ b/clang/include/clang/Basic/Builtins.def
@@ -1619,16 +1619,16 @@ BUILTIN(__builtin_assume, "vb", "nE")
BUILTIN(__builtin_assume_separate_storage, "vvCD*vCD*", "nE")
// Multiprecision Arithmetic Builtins.
-BUILTIN(__builtin_addcb, "UcUcCUcCUcCUc*", "n")
-BUILTIN(__builtin_addcs, "UsUsCUsCUsCUs*", "n")
-BUILTIN(__builtin_addc, "UiUiCUiCUiCUi*", "n")
-BUILTIN(__builtin_addcl, "ULiULiCULiCULiCULi*", "n")
-BUILTIN(__builtin_addcll, "ULLiULLiCULLiCULLiCULLi*", "n")
-BUILTIN(__builtin_subcb, "UcUcCUcCUcCUc*", "n")
-BUILTIN(__builtin_subcs, "UsUsCUsCUsCUs*", "n")
-BUILTIN(__builtin_subc, "UiUiCUiCUiCUi*", "n")
-BUILTIN(__builtin_subcl, "ULiULiCULiCULiCULi*", "n")
-BUILTIN(__builtin_subcll, "ULLiULLiCULLiCULLiCULLi*", "n")
+BUILTIN(__builtin_addcb, "UcUcCUcCUcCUc*", "nE")
+BUILTIN(__builtin_addcs, "UsUsCUsCUsCUs*", "nE")
+BUILTIN(__builtin_addc, "UiUiCUiCUiCUi*", "nE")
+BUILTIN(__builtin_addcl, "ULiULiCULiCULiCULi*", "nE")
+BUILTIN(__builtin_addcll, "ULLiULLiCULLiCULLiCULLi*", "nE")
+BUILTIN(__builtin_subcb, "UcUcCUcCUcCUc*", "nE")
+BUILTIN(__builtin_subcs, "UsUsCUsCUsCUs*", "nE")
+BUILTIN(__builtin_subc, "UiUiCUiCUiCUi*", "nE")
+BUILTIN(__builtin_subcl, "ULiULiCULiCULiCULi*", "nE")
+BUILTIN(__builtin_subcll, "ULLiULLiCULLiCULLiCULLi*", "nE")
// Checked Arithmetic Builtins for Security.
BUILTIN(__builtin_add_overflow, "b.", "ntE")
diff --git a/clang/lib/AST/ExprConstant.cpp b/clang/lib/AST/ExprConstant.cpp
index dfa48e9c030b6a3..c8f837f7fc1f688 100644
--- a/clang/lib/AST/ExprConstant.cpp
+++ b/clang/lib/AST/ExprConstant.cpp
@@ -12630,6 +12630,56 @@ bool IntExprEvaluator::VisitBuiltinCallExpr(const
CallExpr *E,
return false;
return Success(DidOverflow, E);
}
+ case Builtin::BI__builtin_addcb:
+ case Builtin::BI__builtin_addcs:
+ case Builtin::BI__builtin_addc:
+ case Builtin::BI__builtin_addcl:
+ case Builtin::BI__builtin_addcll:
+ case Builtin::BI__builtin_subcb:
+ case Builtin::BI__builtin_subcs:
+ case Builtin::BI__builtin_subc:
+ case Builtin::BI__builtin_subcl:
+ case Builtin::BI__builtin_subcll: {
+APSInt X, Y, CarryIn;
+LValue CarryOut;
+
+QualType ResultType = E->getArg(3)->getType()->getPointeeType();
+if (!EvaluateInteger(E->getArg(0), X, Info) ||
+!EvaluateInteger(E->getArg(1), Y, Info) ||
+!EvaluateInteger(E->getArg(2), CarryIn, Info) ||
+!EvaluatePointer(E->getArg(3), CarryOut, Info))
+ return false;
+
+APInt Result;
+bool DidOverflow1 = false;
+bool DidOverflow2 = false;
+
+switch (BuiltinOp) {
+default:
+ llvm_unreachable("Invalid value for BuiltinOp");
+case Builtin::BI__builtin_addcb:
+case Builtin::BI__builtin_addcs:
+case Builtin::BI__builtin_addc:
+case Builtin::BI__builtin_addcl:
+case Builtin::BI__builtin_addcll:
+ Result = X.uadd_ov(Y, DidOverflow1).uadd_ov(CarryIn, DidOverflow2);
+ break;
+case Builtin::BI__builtin_subcb:
+case Builtin::BI__builtin_subcs:
+case Builtin::BI__builtin_subc:
+case Builtin::BI__builtin_subcl:
+case Builtin::BI__builtin_subcll:
+ Result = X.usub_ov(Y, DidOverflow1).usub_ov(CarryIn, DidOverflow2);
+ break;
+}
+
+APSInt DidOverflow(
+APInt(Result.getBitWidth(), D
[clang] [clang] Implement constexpr evaluation for `__builtin_{add,sub}c` (PR #66005)
https://github.com/BertalanD review_requested https://github.com/llvm/llvm-project/pull/66005 ___ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[clang] [clang] Implement constexpr evaluation for `__builtin_{add,sub}c` (PR #66005)
https://github.com/BertalanD updated
https://github.com/llvm/llvm-project/pull/66005:
>From 1a1f9ce09efb51a26440f378334df97ec014cad5 Mon Sep 17 00:00:00 2001
From: Daniel Bertalan
Date: Sat, 22 Jul 2023 08:20:45 +0200
Subject: [PATCH] [clang] Implement constexpr evaluation for
`__builtin_{add,sub}c`
GCC has gained support for these multiprecision arithmetic builtins in
`r14-1896-g2b4e0415ad6`, and although they aren't explicitly specified
as such in the documentation, they are usable in a constexpr context.
This commit adds constexpr evaluation support to Clang to match GCC's
behavior. The implementation mirrors how the builtins are lowered to a
pair of `u{add,sub}.with.overflow` operations and the carryout is set to
1 if either of those result in an overflow.
---
clang/docs/LanguageExtensions.rst | 12 ++
clang/include/clang/Basic/Builtins.def| 20 ++--
clang/lib/AST/ExprConstant.cpp| 50 +
.../test/SemaCXX/builtins-multiprecision.cpp | 105 ++
4 files changed, 177 insertions(+), 10 deletions(-)
create mode 100644 clang/test/SemaCXX/builtins-multiprecision.cpp
diff --git a/clang/docs/LanguageExtensions.rst
b/clang/docs/LanguageExtensions.rst
index 11cbdca7a268fc3..8541558fa4c1ec7 100644
--- a/clang/docs/LanguageExtensions.rst
+++ b/clang/docs/LanguageExtensions.rst
@@ -3449,6 +3449,8 @@ The complete list of builtins are:
unsigned long __builtin_subcl (unsigned long x, unsigned long y,
unsigned long carryin, unsigned long *carryout);
unsigned long long __builtin_subcll(unsigned long long x, unsigned long long
y, unsigned long long carryin, unsigned long long *carryout);
+These builtins can be used in constant expressions.
+
Checked Arithmetic Builtins
---
@@ -5173,6 +5175,11 @@ Intrinsics Support within Constant Expressions
The following builtin intrinsics can be used in constant expressions:
+* ``__builtin_addc``
+* ``__builtin_addcb``
+* ``__builtin_addcl``
+* ``__builtin_addcll``
+* ``__builtin_addcs``
* ``__builtin_bitreverse8``
* ``__builtin_bitreverse16``
* ``__builtin_bitreverse32``
@@ -5219,6 +5226,11 @@ The following builtin intrinsics can be used in constant
expressions:
* ``__builtin_rotateright16``
* ``__builtin_rotateright32``
* ``__builtin_rotateright64``
+* ``__builtin_subc``
+* ``__builtin_subcb``
+* ``__builtin_subcl``
+* ``__builtin_subcll``
+* ``__builtin_subcs``
The following x86-specific intrinsics can be used in constant expressions:
diff --git a/clang/include/clang/Basic/Builtins.def
b/clang/include/clang/Basic/Builtins.def
index 586dcf05170eb58..8a54850cca672ba 100644
--- a/clang/include/clang/Basic/Builtins.def
+++ b/clang/include/clang/Basic/Builtins.def
@@ -1619,16 +1619,16 @@ BUILTIN(__builtin_assume, "vb", "nE")
BUILTIN(__builtin_assume_separate_storage, "vvCD*vCD*", "nE")
// Multiprecision Arithmetic Builtins.
-BUILTIN(__builtin_addcb, "UcUcCUcCUcCUc*", "n")
-BUILTIN(__builtin_addcs, "UsUsCUsCUsCUs*", "n")
-BUILTIN(__builtin_addc, "UiUiCUiCUiCUi*", "n")
-BUILTIN(__builtin_addcl, "ULiULiCULiCULiCULi*", "n")
-BUILTIN(__builtin_addcll, "ULLiULLiCULLiCULLiCULLi*", "n")
-BUILTIN(__builtin_subcb, "UcUcCUcCUcCUc*", "n")
-BUILTIN(__builtin_subcs, "UsUsCUsCUsCUs*", "n")
-BUILTIN(__builtin_subc, "UiUiCUiCUiCUi*", "n")
-BUILTIN(__builtin_subcl, "ULiULiCULiCULiCULi*", "n")
-BUILTIN(__builtin_subcll, "ULLiULLiCULLiCULLiCULLi*", "n")
+BUILTIN(__builtin_addcb, "UcUcCUcCUcCUc*", "nE")
+BUILTIN(__builtin_addcs, "UsUsCUsCUsCUs*", "nE")
+BUILTIN(__builtin_addc, "UiUiCUiCUiCUi*", "nE")
+BUILTIN(__builtin_addcl, "ULiULiCULiCULiCULi*", "nE")
+BUILTIN(__builtin_addcll, "ULLiULLiCULLiCULLiCULLi*", "nE")
+BUILTIN(__builtin_subcb, "UcUcCUcCUcCUc*", "nE")
+BUILTIN(__builtin_subcs, "UsUsCUsCUsCUs*", "nE")
+BUILTIN(__builtin_subc, "UiUiCUiCUiCUi*", "nE")
+BUILTIN(__builtin_subcl, "ULiULiCULiCULiCULi*", "nE")
+BUILTIN(__builtin_subcll, "ULLiULLiCULLiCULLiCULLi*", "nE")
// Checked Arithmetic Builtins for Security.
BUILTIN(__builtin_add_overflow, "b.", "ntE")
diff --git a/clang/lib/AST/ExprConstant.cpp b/clang/lib/AST/ExprConstant.cpp
index dfa48e9c030b6a3..c8f837f7fc1f688 100644
--- a/clang/lib/AST/ExprConstant.cpp
+++ b/clang/lib/AST/ExprConstant.cpp
@@ -12630,6 +12630,56 @@ bool IntExprEvaluator::VisitBuiltinCallExpr(const
CallExpr *E,
return false;
return Success(DidOverflow, E);
}
+ case Builtin::BI__builtin_addcb:
+ case Builtin::BI__builtin_addcs:
+ case Builtin::BI__builtin_addc:
+ case Builtin::BI__builtin_addcl:
+ case Builtin::BI__builtin_addcll:
+ case Builtin::BI__builtin_subcb:
+ case Builtin::BI__builtin_subcs:
+ case Builtin::BI__builtin_subc:
+ case Builtin::BI__builtin_subcl:
+ case Builtin::BI__builtin_subcll: {
+APSInt X, Y, CarryIn;
+LValue CarryOut;
+
+QualType ResultType = E->getArg(3)->getType()->getPointeeType();
+if (!EvaluateInteger(E->getArg(0), X, Info) ||
+!EvaluateInteger(E->getArg(1), Y
[clang] [clang] Implement constexpr evaluation for `__builtin_{add,sub}c` (PR #66005)
https://github.com/BertalanD closed https://github.com/llvm/llvm-project/pull/66005 ___ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
[clang] [clang] Implement constexpr evaluation for `__builtin_{add,sub}c` (PR #66005)
BertalanD wrote: Obsoleted by #81656 https://github.com/llvm/llvm-project/pull/66005 ___ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
