https://github.com/rnk closed https://github.com/llvm/llvm-project/pull/135430
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
rnk wrote:
I just happened to look into the implementation of AutoConvert.h, and I see the
entire interface is defined away when not targeting zOS. This means that all
call sites need to be conditional on `ifdef __MVS__`, which means we have ugly
ifdef droppings all over the codebase, which ha
@@ -13,10 +15,41 @@ struct Bar {
struct Baz {
virtual ~Baz() {}
- static void operator delete[](void* ptr) = delete; // expected-note
{{explicitly marked deleted here}}
+ static void operator delete[](void* ptr) = delete; // expected-note
{{explicitly marked deleted here
https://github.com/rnk edited https://github.com/llvm/llvm-project/pull/135041
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/rnk updated https://github.com/llvm/llvm-project/pull/134196
>From 528be44ebae8ea5cd7f23c51ad11c314a73f152b Mon Sep 17 00:00:00 2001
From: Reid Kleckner
Date: Fri, 28 Mar 2025 22:11:05 -0700
Subject: [PATCH 1/8] [clang] Merge gtest binaries into AllClangUnitTests
This reduces
https://github.com/rnk ready_for_review
https://github.com/llvm/llvm-project/pull/134196
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/rnk updated https://github.com/llvm/llvm-project/pull/134196
>From 528be44ebae8ea5cd7f23c51ad11c314a73f152b Mon Sep 17 00:00:00 2001
From: Reid Kleckner
Date: Fri, 28 Mar 2025 22:11:05 -0700
Subject: [PATCH 1/7] [clang] Merge gtest binaries into AllClangUnitTests
This reduces
https://github.com/rnk updated https://github.com/llvm/llvm-project/pull/134196
>From 528be44ebae8ea5cd7f23c51ad11c314a73f152b Mon Sep 17 00:00:00 2001
From: Reid Kleckner
Date: Fri, 28 Mar 2025 22:11:05 -0700
Subject: [PATCH 1/6] [clang] Merge gtest binaries into AllClangUnitTests
This reduces
https://github.com/rnk updated https://github.com/llvm/llvm-project/pull/134196
>From 528be44ebae8ea5cd7f23c51ad11c314a73f152b Mon Sep 17 00:00:00 2001
From: Reid Kleckner
Date: Fri, 28 Mar 2025 22:11:05 -0700
Subject: [PATCH 1/5] [clang] Merge gtest binaries into AllClangUnitTests
This reduces
https://github.com/rnk updated https://github.com/llvm/llvm-project/pull/134196
>From 528be44ebae8ea5cd7f23c51ad11c314a73f152b Mon Sep 17 00:00:00 2001
From: Reid Kleckner
Date: Fri, 28 Mar 2025 22:11:05 -0700
Subject: [PATCH 1/4] [clang] Merge gtest binaries into AllClangUnitTests
This reduces
https://github.com/rnk updated https://github.com/llvm/llvm-project/pull/134196
>From 528be44ebae8ea5cd7f23c51ad11c314a73f152b Mon Sep 17 00:00:00 2001
From: Reid Kleckner
Date: Fri, 28 Mar 2025 22:11:05 -0700
Subject: [PATCH 1/3] [clang] Merge gtest binaries into AllClangUnitTests
This reduces
@@ -11045,15 +11045,18 @@ bool Sema::CheckDestructor(CXXDestructorDecl
*Destructor) {
DiagnoseUseOfDecl(OperatorDelete, Loc);
MarkFunctionReferenced(Loc, OperatorDelete);
Destructor->setOperatorDelete(OperatorDelete, ThisArg);
- // Lookup delete[] too in
https://github.com/rnk approved this pull request.
I have some minor suggestions, but please go ahead and land this without
additional review since it is a forward-fix.
https://github.com/llvm/llvm-project/pull/135041
___
cfe-commits mailing list
cfe-
@@ -11045,15 +11045,18 @@ bool Sema::CheckDestructor(CXXDestructorDecl
*Destructor) {
DiagnoseUseOfDecl(OperatorDelete, Loc);
MarkFunctionReferenced(Loc, OperatorDelete);
Destructor->setOperatorDelete(OperatorDelete, ThisArg);
- // Lookup delete[] too in
https://github.com/rnk commented:
The premerge tests failed on a modules crash recovery test case on other
platforms, and those seem like true positives that need to be fixed.
Overall, I think the main risk with this approach is that it will break
in-process, FP-based stack unwinders that vali
https://github.com/rnk updated https://github.com/llvm/llvm-project/pull/134196
>From 528be44ebae8ea5cd7f23c51ad11c314a73f152b Mon Sep 17 00:00:00 2001
From: Reid Kleckner
Date: Fri, 28 Mar 2025 22:11:05 -0700
Subject: [PATCH 1/2] [clang] Merge gtest binaries into AllClangUnitTests
This reduces
https://github.com/rnk closed https://github.com/llvm/llvm-project/pull/134195
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/rnk converted_to_draft
https://github.com/llvm/llvm-project/pull/134196
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
Author: Reid Kleckner
Date: 2025-04-04T22:10:19-07:00
New Revision: a1935fd3809772c06f9a09fa151181642ae92b20
URL:
https://github.com/llvm/llvm-project/commit/a1935fd3809772c06f9a09fa151181642ae92b20
DIFF:
https://github.com/llvm/llvm-project/commit/a1935fd3809772c06f9a09fa151181642ae92b20.diff
rnk wrote:
We have internal reports of incorrect mangling substitutions after updating to
537b6541e8067d7ef7aa38791989fca6303b7fd from
799e9053641a6478d3144866a97737b37b87c260, and this change seems like the likely
culprit because it's touching the Itanium mangling code. We'll see if that can
@@ -0,0 +1,115 @@
+//===--- RunOnNewStack.cpp - Crash Recovery
---===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM
Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Ap
@@ -0,0 +1,115 @@
+//===--- RunOnNewStack.cpp - Crash Recovery
---===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM
Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Ap
Markus =?utf-8?q?Gschoßmann?= ,
Markus =?utf-8?q?Gschoßmann?=
Message-ID:
In-Reply-To:
@@ -0,0 +1,16 @@
+// RUN: %clang_cc1 -triple x86_64-none-linux-gnu -emit-llvm
-debug-info-kind=limited %s -o - | FileCheck %s
rnk wrote:
Thanks, I looked at the files, and
@@ -3897,6 +3897,13 @@ void CodeGenFunction::EmitFunctionEpilog(const
CGFunctionInfo &FI,
return;
}
+ // If there is no valid insert point, we won't emit a return.
+ // The insert point could be null if we have already emitted a return
+ // (e.g. if musttail)
+ if (
https://github.com/rnk created https://github.com/llvm/llvm-project/pull/134196
This reduces the size of the clang/unittests build directory by 64% and my
overall build dir size by 5%. Static linking is the real driving factor here,
but even if the default build configuration used shared librar
https://github.com/rnk created https://github.com/llvm/llvm-project/pull/134195
This reapplies 5ffd9bdb50b57 (#133545) with fixes.
The BUILD_SHARED_LIBS=ON build was fixed by adding missing LLVM
dependencies to the InterpTests binary in
unittests/AST/ByteCode/CMakeLists.txt .
>From fe6604dc283
https://github.com/rnk updated https://github.com/llvm/llvm-project/pull/133545
>From e662d8d9483fdf82030ddec6969bc89ae2404060 Mon Sep 17 00:00:00 2001
From: Reid Kleckner
Date: Fri, 28 Mar 2025 16:49:09 -0700
Subject: [PATCH 1/2] [cmake] Refactor clang unittest cmake
Pass all the dependencies
https://github.com/rnk updated https://github.com/llvm/llvm-project/pull/133545
>From e662d8d9483fdf82030ddec6969bc89ae2404060 Mon Sep 17 00:00:00 2001
From: Reid Kleckner
Date: Fri, 28 Mar 2025 16:49:09 -0700
Subject: [PATCH 1/2] [cmake] Refactor clang unittest cmake
Pass all the dependencies
https://github.com/rnk created https://github.com/llvm/llvm-project/pull/133545
Pass all the dependencies into add_clang_unittest. This is consistent with how
it is done for LLDB. I borrowed the same named argument list structure from
add_lldb_unittest. This is a necessary step towards consolid
Markus =?utf-8?q?Gschoßmann?= ,
Markus =?utf-8?q?Gschoßmann?=
Message-ID:
In-Reply-To:
@@ -0,0 +1,16 @@
+// RUN: %clang_cc1 -triple x86_64-none-linux-gnu -emit-llvm
-debug-info-kind=limited %s -o - | FileCheck %s
rnk wrote:
Please add a test for the MSVC ABI,
@@ -0,0 +1,115 @@
+//===--- RunOnNewStack.cpp - Crash Recovery
---===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM
Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Ap
@@ -52,6 +56,7 @@ else()
check_include_file(mach/mach.h HAVE_MACH_MACH_H)
check_include_file(malloc/malloc.h HAVE_MALLOC_MALLOC_H)
check_include_file(pthread.h HAVE_PTHREAD_H)
+ check_include_file(sys/resource.h HAVE_SYS_RESOURCE_H)
rnk wrote:
Do we nee
rnk wrote:
I'll try to take a look at this when I get a chance, but that looks like it's
not happening today.
I read the comments a bit and found some codesearch links worth sharing for
reference. Here's the original UnicodeString array new operation:
https://source.chromium.org/chromium/chrom
@@ -0,0 +1,115 @@
+//===--- RunOnNewStack.cpp - Crash Recovery
---===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM
Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Ap
@@ -0,0 +1,115 @@
+//===--- RunOnNewStack.cpp - Crash Recovery
---===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM
Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Ap
https://github.com/rnk commented:
If split stacks are negatively impacting profiling, debugging, or other
compiler development tasks, I have to ask, have we considered optimizing clang
stack usage?
There are multiple issues:
* Clang is a recursive descent parser. This isn't going to change, it
https://github.com/rnk edited https://github.com/llvm/llvm-project/pull/133173
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
rnk wrote:
I think the `.SCOV$M` section globals should be in the `??_EClassA@@UEAAPEXI@Z`
comdat group. I think sanitizer coverage (sancov) is one of the lesser-used
sanitizers here, and it may lack some sophistication when it comes to comdat
groups. I wasn't aware of this creative use of `bl
@@ -6519,6 +6519,13 @@ def warn_signed_bitfield_enum_conversion : Warning<
InGroup, DefaultIgnore;
def note_change_bitfield_sign : Note<
"consider making the bit-field type %select{unsigned|signed}0">;
+def warn_ms_bitfield_mismatched_storage_packing : Warning<
+ "bit-fiel
@@ -4769,31 +4769,32 @@ ExprResult Sema::CheckOSLogFormatStringArg(Expr *Arg) {
/// Check that the user is calling the appropriate va_start builtin for the
/// target and calling convention.
static bool checkVAStartABI(Sema &S, unsigned BuiltinID, Expr *Fn) {
- const llvm::Tri
https://github.com/rnk approved this pull request.
Thanks, I spent a while thinking if this was correct, and then convincing
myself that we ignore the relevant calling conventions on the appropriate
platforms (we ignore `__fastcall` on x64 and ignore `__vectorcall` on Win ARM).
https://github.
rnk wrote:
If you take the two PRs together, we're going to implement the indirection
logic at both the frontend and backend level. If we do it in the backend, I
think it's technically necessary to do it in the frontend, but it is consistent
with how we handle all other types. I'd like to see
https://github.com/rnk approved this pull request.
https://github.com/llvm/llvm-project/pull/115052
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
@@ -190,3 +210,38 @@ method
some_class::good_coroutine_calls_custom_constructor(float) {
// CHECK: invoke void
@_ZNSt16coroutine_traitsIJ6methodR10some_classfEE12promise_typeC1ES2_f(ptr
{{[^,]*}} %__promise, ptr noundef nonnull align 1 dereferenceable(1) %{{.+}},
float
c
@@ -855,6 +855,16 @@ void CodeGenFunction::EmitCoroutineBody(const
CoroutineBodyStmt &S) {
// Create parameter copies. We do it before creating a promise, since an
// evolution of coroutine TS may allow promise constructor to observe
// parameter copies.
+for (
https://github.com/rnk edited https://github.com/llvm/llvm-project/pull/127653
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/rnk approved this pull request.
I think correctness comes first, so this prioritization makes sense.
https://github.com/llvm/llvm-project/pull/127653
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-
https://github.com/rnk approved this pull request.
Thanks! I approved it.
https://github.com/llvm/llvm-project/pull/126240
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/rnk commented:
I resolved two conversations, but there are two actionable comments.
https://github.com/llvm/llvm-project/pull/126240
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listi
https://github.com/rnk commented:
Hm, these comments didn't post. They may be stale. I will post them and revisit
them.
https://github.com/llvm/llvm-project/pull/126240
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi
@@ -7919,3 +7919,38 @@ void CodeGenModule::moveLazyEmissionStates(CodeGenModule
*NewBuilder) {
NewBuilder->ABI->MangleCtx = std::move(ABI->MangleCtx);
}
+
+bool CodeGenModule::classNeedsVectorDestructor(const CXXRecordDecl *RD) {
+ CXXDestructorDecl *Dtor = RD->getDestruct
@@ -7919,3 +7919,38 @@ void CodeGenModule::moveLazyEmissionStates(CodeGenModule
*NewBuilder) {
NewBuilder->ABI->MangleCtx = std::move(ABI->MangleCtx);
}
+
+bool CodeGenModule::classNeedsVectorDestructor(const CXXRecordDecl *RD) {
+ CXXDestructorDecl *Dtor = RD->getDestruct
@@ -575,6 +576,12 @@ class CGCXXABI {
QualType ElementType, llvm::Value *&NumElements,
llvm::Value *&AllocPtr, CharUnits &CookieSize);
+ /// Reads the array cookie associated with the given pointer,
+ /// that sho
@@ -2657,7 +2657,10 @@ class VFTableBuilder {
MethodVFTableLocation Loc(MI.VBTableIndex, WhichVFPtr.getVBaseWithVPtr(),
WhichVFPtr.NonVirtualOffset, MI.VFTableIndex);
if (const CXXDestructorDecl *DD = dyn_cast(MD)) {
-MethodVF
https://github.com/rnk edited https://github.com/llvm/llvm-project/pull/126240
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/rnk closed https://github.com/llvm/llvm-project/pull/128592
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/rnk created https://github.com/llvm/llvm-project/pull/128592
cvise reimplements creduce in Python and bundles clang-delta and other tools.
In my experience, it is generally a more robust reduction tool that is better
maintained. I renamed the script to make it tool-neutral, w
rnk wrote:
Re: sroa/mem2reg, that's a valid concern with Hans's intrinsic approach.
> So it's not really a question of "improving" the existing algorithm; we need
> markers in the IR, like coro_outside_frame, and the algorithm should be based
> on that.
To check my understanding, by explicit
https://github.com/rnk commented:
In the issue (#127499) you pointed out that this issue applies to the MSVC ABI
where all parameters are destroyed in the callee. Is it reasonable to construct
the analogous test case in that environment? I tried to do this locally, but
can't for unrelated buil
rnk wrote:
Thanks for the updates, I think this is almost done.
@zmodem , this will probably have some impact on Chrome code size and will
probably churn many crash stack traces. I think it's safe to rely on all the
existing automated testing, but this is a high-risk change to keep in mind if
https://github.com/rnk approved this pull request.
Thanks!
https://github.com/llvm/llvm-project/pull/126952
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/rnk approved this pull request.
Thanks, I'm happy with this patch.
https://github.com/llvm/llvm-project/pull/124834
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
@@ -2950,12 +2950,11 @@ static void handleSectionAttr(Sema &S, Decl *D, const
ParsedAttr &AL) {
}
}
-static bool isValidCodeModelAttr(Sema &S, StringRef Str) {
- if (S.Context.getTargetInfo().getTriple().isLoongArch()) {
+static bool isValidCodeModelAttr(llvm::Triple Tripl
@@ -2964,20 +2963,36 @@ static bool isValidCodeModelAttr(Sema &S, StringRef
Str) {
static void handleCodeModelAttr(Sema &S, Decl *D, const ParsedAttr &AL) {
StringRef Str;
SourceLocation LiteralLoc;
+ auto IsTripleSupported = [](const llvm::Triple Triple) {
+return Tr
@@ -2964,20 +2963,36 @@ static bool isValidCodeModelAttr(Sema &S, StringRef
Str) {
static void handleCodeModelAttr(Sema &S, Decl *D, const ParsedAttr &AL) {
StringRef Str;
SourceLocation LiteralLoc;
+ auto IsTripleSupported = [](const llvm::Triple Triple) {
+return Tr
rnk wrote:
This is an observable behavior change, so it should have a test. There's a test
for this header at clang/test/Headers/ms-intrins.cpp that you can extend. You
probably need to add a new RUN line that adds `-mrtm` or some other
micro-architectural feature so that rtmintrin.h declares
@@ -2949,15 +2950,34 @@ static void handleSectionAttr(Sema &S, Decl *D, const
ParsedAttr &AL) {
}
}
+static bool isValidCodeModelAttr(Sema &S, StringRef Str) {
+ if (S.Context.getTargetInfo().getTriple().isLoongArch()) {
+return Str == "normal" || Str == "medium" || St
@@ -5193,12 +5193,12 @@ bool Sema::CheckCallingConvAttr(const ParsedAttr
&Attrs, CallingConv &CC,
CC = CC_X86RegCall;
break;
case ParsedAttr::AT_MSABI:
-CC = Context.getTargetInfo().getTriple().isOSWindows() ? CC_C :
-
@@ -74,7 +74,7 @@ static CCMangling getCallingConvMangling(const ASTContext
&Context,
if (FD->isMain() && FD->getNumParams() == 2)
return CCM_WasmMainArgcArgv;
- if (!Triple.isOSWindows() || !Triple.isX86())
+ if (!Triple.isOSWindowsOrUEFI() || !Triple.isX86()
rnk wrote:
> The build is still broken (e.g.
> https://lab.llvm.org/buildbot/#/builders/63/builds/3861). I'll back it out.
Thank you for doing that. I missed all notifications because they got lost in
the storm of failures due to the mismatch between my local build and what was
on github.
ht
rnk wrote:
> But one specific question: would you prefer me to land as a series of commits
> or a single squashed commit for the entire PR? I'm happy either way. My mild
> preference is to prefer the series of commits, but open to suggestions here.
I would land it as the six-patch stack in thi
https://github.com/rnk edited https://github.com/llvm/llvm-project/pull/120534
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
@@ -482,17 +488,42 @@ void clang::EmitClangBuiltins(const RecordKeeper
&Records, raw_ostream &OS) {
for (const auto *BuiltinRecord :
Records.getAllDerivedDefinitions("AtomicBuiltin"))
collectBuiltins(BuiltinRecord, Builtins);
-
unsigned NumAtomicBuiltins = Buil
https://github.com/rnk approved this pull request.
Thanks for optimizing the builtins! I feel like builtins have grown
significantly since adding RISC-V and ARM MVE intrinsics, and few people until
now have stopped to re-evaluate how we represent these things.
While I was watching these clearl
https://github.com/rnk closed https://github.com/llvm/llvm-project/pull/115099
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
@@ -146,8 +146,13 @@ let Attributes = [Const, NoThrow,
RequiredVectorWidth<256>], Features = "avx" in
// current formulation is based on what was easiest to recognize from the
// pre-TableGen version.
-let Features = "mmx", Attributes = [NoThrow, Const] in {
- def _mm_prefet
https://github.com/rnk updated https://github.com/llvm/llvm-project/pull/115099
>From e5f485ad8000c296229794346fdd627b90e504d2 Mon Sep 17 00:00:00 2001
From: Reid Kleckner
Date: Tue, 5 Nov 2024 16:05:53 -0800
Subject: [PATCH 1/7] [Win/X86] Make _m_prefetch[w] builtins to avoid winnt.h
conflicts
rnk wrote:
I assume it is safe for all device side compilations to ignore this attribute.
Presumably device code uses fully general patterns to access global data that
could live in any shared object. Most device ISAs are virtual anyway, it's not
like they're going to use PC relative addressin
https://github.com/rnk updated https://github.com/llvm/llvm-project/pull/115099
>From e5f485ad8000c296229794346fdd627b90e504d2 Mon Sep 17 00:00:00 2001
From: Reid Kleckner
Date: Tue, 5 Nov 2024 16:05:53 -0800
Subject: [PATCH 1/6] [Win/X86] Make _m_prefetch[w] builtins to avoid winnt.h
conflicts
@@ -2225,9 +2221,10 @@ _mm_storer_ps(float *__p, __m128 __a)
///be generated. \n
///_MM_HINT_T2: Move data using the T2 hint. The PREFETCHT2 instruction
will
///be generated.
-#define _mm_prefetch(a, sel) (__builtin_prefetch((const void *)(a), \
-
https://github.com/rnk updated https://github.com/llvm/llvm-project/pull/115099
>From e5f485ad8000c296229794346fdd627b90e504d2 Mon Sep 17 00:00:00 2001
From: Reid Kleckner
Date: Tue, 5 Nov 2024 16:05:53 -0800
Subject: [PATCH 1/5] [Win/X86] Make _m_prefetch[w] builtins to avoid winnt.h
conflicts
@@ -146,8 +146,13 @@ let Attributes = [Const, NoThrow,
RequiredVectorWidth<256>], Features = "avx" in
// current formulation is based on what was easiest to recognize from the
// pre-TableGen version.
-let Features = "mmx", Attributes = [NoThrow, Const] in {
- def _mm_prefet
https://github.com/rnk updated https://github.com/llvm/llvm-project/pull/115099
>From e5f485ad8000c296229794346fdd627b90e504d2 Mon Sep 17 00:00:00 2001
From: Reid Kleckner
Date: Tue, 5 Nov 2024 16:05:53 -0800
Subject: [PATCH 1/4] [Win/X86] Make _m_prefetch[w] builtins to avoid winnt.h
conflicts
https://github.com/rnk commented:
I had to rebase this over the table-gen-ification of the builtins, but I think
this is still relevant.
https://github.com/llvm/llvm-project/pull/115099
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://l
@@ -31,10 +31,13 @@
// All MMX instructions will be generated via builtins. Any MMX vector
// types (<1 x i64>, <2 x i32>, etc.) that aren't used by these builtins will
be
// expanded by the back-end.
+//
// FIXME: _mm_prefetch must be a built-in because it takes a compile-ti
https://github.com/rnk edited https://github.com/llvm/llvm-project/pull/115099
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/rnk updated https://github.com/llvm/llvm-project/pull/115099
>From e5f485ad8000c296229794346fdd627b90e504d2 Mon Sep 17 00:00:00 2001
From: Reid Kleckner
Date: Tue, 5 Nov 2024 16:05:53 -0800
Subject: [PATCH 1/3] [Win/X86] Make _m_prefetch[w] builtins to avoid winnt.h
conflicts
@@ -637,6 +637,10 @@ static llvm::Triple computeTargetTriple(const Driver &D,
}
}
+ // Currently the only architecture supported by *-uefi triples are x86_64.
+ if(Target.isUEFI() && Target.getArch() != llvm::Triple::x86_64)
rnk wrote:
```suggestion
rnk wrote:
+1 for a test
https://github.com/llvm/llvm-project/pull/124824
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/rnk approved this pull request.
https://github.com/llvm/llvm-project/pull/123548
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
@@ -51,28 +57,71 @@ class StringToOffsetTable {
return II->second;
}
- // Emit the string using string literal concatenation, for better readability
- // and searchability.
- void EmitStringLiteralDef(raw_ostream &OS, const Twine &Decl,
-co
@@ -51,28 +57,71 @@ class StringToOffsetTable {
return II->second;
}
- // Emit the string using string literal concatenation, for better readability
- // and searchability.
- void EmitStringLiteralDef(raw_ostream &OS, const Twine &Decl,
-co
rnk wrote:
+1 to separating implementation from enablement. Enabling the warning by
default will predictably generate pushback and folks may ask to revert, and
it's a good best practice to be able to easily do that while leaving the
implementation in place.
https://github.com/llvm/llvm-projec
@@ -2316,6 +2316,20 @@ void
ASTDeclReader::VisitCXXConstructorDecl(CXXConstructorDecl *D) {
}
VisitCXXMethodDecl(D);
+
+ // Microsoft CXX ABI specific:
+ // Restore the RecordToCopyCtor sidecar LUT entry so that `throw` expressions
+ // find the correct copy constructo
@@ -2316,6 +2316,20 @@ void
ASTDeclReader::VisitCXXConstructorDecl(CXXConstructorDecl *D) {
}
VisitCXXMethodDecl(D);
+
+ // Microsoft CXX ABI specific:
+ // Restore the RecordToCopyCtor sidecar LUT entry so that `throw` expressions
+ // find the correct copy constructo
@@ -249,6 +249,11 @@ FIELD(HasDeclaredCopyAssignmentWithConstParam, 1, MERGE_OR)
/// base classes or fields have a no-return destructor
FIELD(IsAnyDestructorNoReturn, 1, NO_MERGE)
+/// Microsoft CXX ABI specific:
+/// Whether the copy constructor is used by a `throw` expressio
https://github.com/rnk edited https://github.com/llvm/llvm-project/pull/114075
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
@@ -0,0 +1,19 @@
+// REQUIRES: system-windows, target={{.*-windows-msvc}}
+// RUN: %clang_cc1 -x c++ -std=c++17 -fcxx-exceptions -fexceptions
-triple=%ms_abi_triple -emit-pch -building-pch-with-obj -fmodules-codegen -o
%t.pch %S/cxx-exception-copy-ctor-crash.h
+// RUN: %clang_cc
@@ -2316,6 +2316,20 @@ void
ASTDeclReader::VisitCXXConstructorDecl(CXXConstructorDecl *D) {
}
VisitCXXMethodDecl(D);
+
+ // Microsoft CXX ABI specific:
+ // Restore the RecordToCopyCtor sidecar LUT entry so that `throw` expressions
+ // find the correct copy constructo
https://github.com/rnk commented:
I think @majnemer added the lookup table originally. I think we did things this
way because, while it creates obvious serialization problems, it's the most
straightforwardly correct approach if you think finding copy constructors is an
ambiguous operation. It'
1 - 100 of 1471 matches
Mail list logo
