merge main into amd-staging

reverts: MC fails to build, needs downstream work
9abb574 [MC] Make MCCFIInstruction smaller

Change-Id: Ie1090b05dc092956fff4412d92b684625423886e

clang-tools-extra/clangd/support/ThreadsafeFS.cpp

@@ -41,7 +41,7 @@ class VolatileFileSystem : public llvm::vfs::ProxyFileSystem {
     llvm::StringRef FileName = llvm::sys::path::filename(Path);
     if (FileName.starts_with("preamble-") && FileName.ends_with(".pch"))
       return File;
-    return std::unique_ptr<VolatileFile>(new VolatileFile(std::move(*File)));
+    return std::make_unique<VolatileFile>(std::move(*File));
   }
 
 private:

clang/docs/ReleaseNotes.rst

@@ -979,6 +979,8 @@ Bug Fixes to C++ Support
   of the address of operator. (#GH97483).
 - Fixed an assertion failure about a constant expression which is a known integer but is not
   evaluated to an integer. (#GH96670).
+- Fixed a bug where references to lambda capture inside a ``noexcept`` specifier were not correctly
+  instantiated. (#GH95735).
 
 Bug Fixes to AST Handling
 ^^^^^^^^^^^^^^^^^^^^^^^^^

clang/lib/Sema/SemaTemplateInstantiateDecl.cpp

@@ -4704,6 +4704,12 @@ void Sema::InstantiateExceptionSpec(SourceLocation PointOfInstantiation,
     return;
   }
 
+  // The noexcept specification could reference any lambda captures. Ensure
+  // those are added to the LocalInstantiationScope.
+  LambdaScopeForCallOperatorInstantiationRAII PushLambdaCaptures(
+      *this, Decl, TemplateArgs, Scope,
+      /*ShouldAddDeclsFromParentScope=*/false);
+
   SubstExceptionSpec(Decl, Template->getType()->castAs<FunctionProtoType>(),
                      TemplateArgs);
 }

clang/test/SemaTemplate/generic-lambda.cpp

@@ -60,3 +60,26 @@ template<class T1> C1<X<X<T1>>> auto t3() {
 template C1<X<X<int>>> auto t3<int>();
 static_assert(is_same<decltype(t3<int>()), X<X<X<int>>>>);
 #endif
+
+namespace GH95735 {
+
+int g(int fn) {
+  return [f = fn](auto tpl) noexcept(noexcept(f)) { return f; }(0);
+}
+
+int foo(auto... fn) {
+  // FIXME: This one hits the assertion "if the exception specification is dependent,
+  // then the noexcept expression should be value-dependent" in the constructor of
+  // FunctionProtoType.
+  // One possible solution is to update Sema::canThrow() to consider expressions
+  // (e.g. DeclRefExpr/FunctionParmPackExpr) involving unexpanded parameters as Dependent.
+  // This would effectively add an extra value-dependent flag to the noexcept expression.
+  // However, I'm afraid that would also cause ABI breakage.
+  // [...f = fn](auto tpl) noexcept(noexcept(f)) { return 0; }(0);
+  [...f = fn](auto tpl) noexcept(noexcept(g(fn...))) { return 0; }(0);
+  return [...f = fn](auto tpl) noexcept(noexcept(g(f...))) { return 0; }(0);
+}
+
+int v = foo(42);
+
+} // namespace GH95735

libc/benchmarks/gpu/LibcGpuBenchmark.cpp

@@ -17,7 +17,8 @@ void Benchmark::add_benchmark(Benchmark *benchmark) {
   benchmarks.push_back(benchmark);
 }
 
-BenchmarkResult reduce_results(cpp::array<BenchmarkResult, 1024> &results) {
+BenchmarkResult
+reduce_results(const cpp::array<BenchmarkResult, 1024> &results) {
   BenchmarkResult result;
   uint64_t cycles_sum = 0;
   double standard_deviation_sum = 0;
@@ -51,16 +52,16 @@ void Benchmark::run_benchmarks() {
   uint64_t id = gpu::get_thread_id();
   gpu::sync_threads();
 
-  for (Benchmark *benchmark : benchmarks)
-    results[id] = benchmark->run();
+  for (Benchmark *b : benchmarks)
+    results[id] = b->run();
   gpu::sync_threads();
   if (id == 0) {
-    for (Benchmark *benchmark : benchmarks) {
+    for (Benchmark const *b : benchmarks) {
       BenchmarkResult all_results = reduce_results(results);
       constexpr auto GREEN = "\033[32m";
       constexpr auto RESET = "\033[0m";
-      log << GREEN << "[ RUN      ] " << RESET << benchmark->get_name() << '\n';
-      log << GREEN << "[       OK ] " << RESET << benchmark->get_name() << ": "
+      log << GREEN << "[ RUN      ] " << RESET << b->get_name() << '\n';
+      log << GREEN << "[       OK ] " << RESET << b->get_name() << ": "
           << all_results.cycles << " cycles, " << all_results.min << " min, "
           << all_results.max << " max, " << all_results.total_iterations
           << " iterations, " << all_results.total_time << " ns, "
@@ -82,7 +83,6 @@ BenchmarkResult benchmark(const BenchmarkOptions &options,
   uint32_t samples = 0;
   uint64_t total_time = 0;
   uint64_t best_guess = 0;
-  uint64_t total_cycles = 0;
   uint64_t cycles_squared = 0;
   uint64_t min = UINT64_MAX;
   uint64_t max = 0;
@@ -92,23 +92,22 @@ BenchmarkResult benchmark(const BenchmarkOptions &options,
   for (int i = 0; i < overhead_iterations; i++)
     overhead = cpp::min(overhead, LIBC_NAMESPACE::overhead());
 
-  for (uint64_t time_budget = options.max_duration; time_budget >= 0;) {
+  for (int64_t time_budget = options.max_duration; time_budget >= 0;) {
     uint64_t sample_cycles = 0;
     const clock_t start = static_cast<double>(clock());
     for (uint32_t i = 0; i < iterations; i++) {
       auto wrapper_intermediate = wrapper_func();
-      uint64_t result = wrapper_intermediate - overhead;
-      max = cpp::max(max, result);
-      min = cpp::min(min, result);
-      sample_cycles += result;
+      uint64_t current_result = wrapper_intermediate - overhead;
+      max = cpp::max(max, current_result);
+      min = cpp::min(min, current_result);
+      sample_cycles += current_result;
     }
     const clock_t end = clock();
     const clock_t duration_ns =
         ((end - start) * 1000 * 1000 * 1000) / CLOCKS_PER_SEC;
     total_time += duration_ns;
     time_budget -= duration_ns;
     samples++;
-    total_cycles += sample_cycles;
     cycles_squared += sample_cycles * sample_cycles;
 
     total_iterations += iterations;

libc/benchmarks/gpu/LibcGpuBenchmark.h

@@ -19,8 +19,8 @@ struct BenchmarkOptions {
   uint32_t max_iterations = 10000000;
   uint32_t min_samples = 4;
   uint32_t max_samples = 1000;
-  uint64_t min_duration = 0;                  // in nanoseconds (ns)
-  uint64_t max_duration = 1000 * 1000 * 1000; // 1e9 nanoseconds = 1 second
+  int64_t min_duration = 0;                  // in nanoseconds (ns)
+  int64_t max_duration = 1000 * 1000 * 1000; // 1e9 nanoseconds = 1 second
   double epsilon = 0.01;
   double scaling_factor = 1.4;
 };

libc/config/linux/aarch64/entrypoints.txt

@@ -17,6 +17,12 @@ set(TARGET_LIBC_ENTRYPOINTS
     libc.src.ctype.tolower
     libc.src.ctype.toupper
 
+    # dlfcn.h entrypoints
+    libc.src.dlfcn.dlclose
+    libc.src.dlfcn.dlerror
+    libc.src.dlfcn.dlopen
+    libc.src.dlfcn.dlsym
+
     # errno.h entrypoints
     libc.src.errno.errno
 

libc/config/linux/x86_64/entrypoints.txt

@@ -17,6 +17,12 @@ set(TARGET_LIBC_ENTRYPOINTS
     libc.src.ctype.tolower
     libc.src.ctype.toupper
 
+    # dlfcn.h entrypoints
+    libc.src.dlfcn.dlclose
+    libc.src.dlfcn.dlerror
+    libc.src.dlfcn.dlopen
+    libc.src.dlfcn.dlsym
+
     # errno.h entrypoints
     libc.src.errno.errno
 

libc/docs/dev/undefined_behavior.rst

@@ -89,3 +89,7 @@ The C23 standard states that if the value of the ``rnd`` argument of the
 the value of a math rounding direction macro, the direction of rounding is
 unspecified. LLVM's libc chooses to use the ``FP_INT_TONEAREST`` rounding
 direction in this case.
+
+Non-const Constant Return Values
+--------------------------------
+Some libc functions, like ``dlerror()``, return ``char *`` instead of ``const char *`` and then tell the caller they promise not to to modify this value. Any modification of this value is undefined behavior.

libc/spec/posix.td

@@ -222,6 +222,40 @@ def POSIX : StandardSpec<"POSIX"> {
       []  // Functions
   >;
 
+  HeaderSpec DlFcn = HeaderSpec<
+    "dlfcn.h",
+    [
+      Macro<"RTLD_LAZY">,
+      Macro<"RTLD_NOW">,
+      Macro<"RTLD_GLOBAL">,
+      Macro<"RTLD_LOCAL">,
+    ],
+    [],  // Types
+    [], // Enumerations
+    [
+      FunctionSpec<
+          "dlclose",
+          RetValSpec<IntType>,
+          [ArgSpec<VoidPtr>]
+      >,
+      FunctionSpec<
+          "dlerror",
+          RetValSpec<CharPtr>,
+          []
+      >,
+      FunctionSpec<
+          "dlopen",
+          RetValSpec<VoidPtr>,
+          [ArgSpec<ConstCharPtr>, ArgSpec<IntType>]
+      >,
+      FunctionSpec<
+          "dlsym",
+          RetValSpec<VoidPtr>,
+          [ArgSpec<VoidRestrictedPtr>, ArgSpec<ConstCharRestrictedPtr>]
+      >,
+    ]
+  >;
+
   HeaderSpec FCntl = HeaderSpec<
     "fcntl.h",
     [], // Macros
@@ -1690,6 +1724,7 @@ def POSIX : StandardSpec<"POSIX"> {
     ArpaInet,
     CType,
     Dirent,
+    DlFcn,
     Errno,
     FCntl,
     PThread,

libc/src/CMakeLists.txt

@@ -1,6 +1,7 @@
 add_subdirectory(__support)
 
 add_subdirectory(ctype)
+add_subdirectory(dlfcn)
 add_subdirectory(errno)
 add_subdirectory(fenv)
 add_subdirectory(inttypes)

libc/src/dlfcn/CMakeLists.txt

@@ -0,0 +1,40 @@
+add_entrypoint_object(
+  dlclose
+  SRCS
+    dlclose.cpp
+  HDRS
+    dlclose.h
+)
+
+add_entrypoint_object(
+  dlerror
+  SRCS
+    dlerror.cpp
+  HDRS
+    dlerror.h
+  DEPENDS
+    libc.include.dlfcn
+    libc.src.errno.errno
+)
+
+add_entrypoint_object(
+  dlopen
+  SRCS
+    dlopen.cpp
+  HDRS
+    dlopen.h
+  DEPENDS
+    libc.include.dlfcn
+    libc.src.errno.errno
+)
+
+add_entrypoint_object(
+  dlsym
+  SRCS
+    dlsym.cpp
+  HDRS
+    dlsym.h
+  DEPENDS
+    libc.include.dlfcn
+    libc.src.errno.errno
+)

libc/src/dlfcn/dlclose.cpp

@@ -0,0 +1,18 @@
+//===-- Implementation of dlclose -----------------------------------------===//
+//
+// 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: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+
+#include "dlclose.h"
+
+#include "src/__support/common.h"
+
+namespace LIBC_NAMESPACE {
+
+// TODO(@izaakschroeder): https://github.com/llvm/llvm-project/issues/97917
+LLVM_LIBC_FUNCTION(int, dlclose, (void *)) { return -1; }
+
+} // namespace LIBC_NAMESPACE

libc/src/dlfcn/dlclose.h

@@ -0,0 +1,18 @@
+//===-- Implementation header of dlclose ------------------------*- C++ -*-===//
+//
+// 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: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+
+#ifndef LLVM_LIBC_SRC_DLFCN_DLCLOSE_H
+#define LLVM_LIBC_SRC_DLFCN_DLCLOSE_H
+
+namespace LIBC_NAMESPACE {
+
+int dlclose(void *);
+
+} // namespace LIBC_NAMESPACE
+
+#endif // LLVM_LIBC_SRC_DLFCN_DLCLOSE_H

libc/src/dlfcn/dlerror.cpp

@@ -0,0 +1,20 @@
+//===-- Implementation of delerror ----------------------------------------===//
+//
+// 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: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+
+#include "dlerror.h"
+
+#include "src/__support/common.h"
+
+namespace LIBC_NAMESPACE {
+
+// TODO(@izaakschroeder): https://github.com/llvm/llvm-project/issues/97918
+LLVM_LIBC_FUNCTION(char *, dlerror, ()) {
+  return const_cast<char *>("unsupported");
+}
+
+} // namespace LIBC_NAMESPACE

libc/src/dlfcn/dlerror.h

@@ -0,0 +1,18 @@
+//===-- Implementation header of dlerror ------------------------*- C++ -*-===//
+//
+// 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: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+
+#ifndef LLVM_LIBC_SRC_DLFCN_DLERROR_H
+#define LLVM_LIBC_SRC_DLFCN_DLERROR_H
+
+namespace LIBC_NAMESPACE {
+
+char *dlerror();
+
+} // namespace LIBC_NAMESPACE
+
+#endif // LLVM_LIBC_SRC_DLFCN_DLERROR_H

libc/src/dlfcn/dlopen.cpp

@@ -0,0 +1,18 @@
+//===-- Implementation of dlopen -----------------------------------------===//
+//
+// 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: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+
+#include "dlopen.h"
+
+#include "src/__support/common.h"
+
+namespace LIBC_NAMESPACE {
+
+// TODO(@izaakschroeder): https://github.com/llvm/llvm-project/issues/97919
+LLVM_LIBC_FUNCTION(void *, dlopen, (const char *, int)) { return nullptr; }
+
+} // namespace LIBC_NAMESPACE

libc/src/dlfcn/dlopen.h

@@ -0,0 +1,18 @@
+//===-- Implementation header of dlopen -------------------------*- C++ -*-===//
+//
+// 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: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+
+#ifndef LLVM_LIBC_SRC_DLFCN_DLOPEN_H
+#define LLVM_LIBC_SRC_DLFCN_DLOPEN_H
+
+namespace LIBC_NAMESPACE {
+
+void *dlopen(const char *, int);
+
+} // namespace LIBC_NAMESPACE
+
+#endif // LLVM_LIBC_SRC_DLFCN_DLOPEN_H