[libc] add strdup implementation

[michaelrj]

Summary

Add an implementation for strdup.

[michaelrj]

update and add a new test

[michaelrj]

small fix

libc/include/stdlib.h.def

17–22 ↗

(On Diff #379188)

This feels wrong, but I'm not sure how else to get these into the header

[sivachandra]

libc/include/stdlib.h.def

17–22 ↗

(On Diff #379188)

This is a good find. Assuming that allocators are but one of the "external" functions that LLVM libc will provide, I think we should have a scalable solution for this. So, may be add a new cmake rule called add_external_entrypoint which does nothing but create a dummy target that can be listed in places like entrypoints.txt?

libc/src/string/strdup.cpp

25

This is a C library implementation. So, the style should be snake_case for variable names. Or, you can completely avoid this conundrum by renaming the src to old and newStr to str.

[lntue]

libc/include/stdlib.h.def

17–22 ↗

(On Diff #379188)

+1. I wasn't able to find a better way, and was waiting for Siva for a better suggestion.

[michaelrj]

add a new path for external functions to be included in the appropriate header file.

[sivachandra]

Can we separate the CMake and allocator changes into a different patch? Rest of the patch is OK but will accept after we separate.

[michaelrj]

split the cmake changes into their own patch

[michaelrj]

missed the changes in stdc.td

[michaelrj]

fix cmake rules to match the new way external entrypoints are handled

[lntue]

libc/test/src/string/CMakeLists.txt

100

nit: alignment.

[michaelrj]

fix indent

libc/config/linux/x86_64/entrypoints.txt

set(TARGET_LIBC_ENTRYPOINTS
     libc.src.string.mempcpy
     libc.src.string.memrchr
     libc.src.string.memset
     libc.src.string.strcat
     libc.src.string.strchr
     libc.src.string.strcmp
     libc.src.string.strcpy
     libc.src.string.strcspn
+    libc.src.string.strdup
     libc.src.string.strlen
     libc.src.string.strncat
     libc.src.string.strncmp
     libc.src.string.strncpy
     libc.src.string.strnlen
     libc.src.string.strpbrk
     libc.src.string.strrchr
     libc.src.string.strspn
if(LLVM_LIBC_INCLUDE_SCUDO)
   list(APPEND TARGET_LIBC_ENTRYPOINTS
 
     # stdlib.h external entrypoints
     libc.src.stdlib.malloc
     libc.src.stdlib.calloc
     libc.src.stdlib.realloc
     libc.src.stdlib.free
 
-
+    # string.h entrypoints that depend on malloc
+    libc.src.string.strdup
   )
 endif()
 
 set(TARGET_LLVMLIBC_ENTRYPOINTS
   ${TARGET_LIBC_ENTRYPOINTS}
   ${TARGET_LIBM_ENTRYPOINTS}
 )

libc/spec/stdc.td

HeaderSpec String = HeaderSpec<
               [ArgSpec<ConstCharPtr>, ArgSpec<IntType>]
           >,
           FunctionSpec<
               "strcspn",
               RetValSpec<SizeTType>,
               [ArgSpec<ConstCharPtr>, ArgSpec<ConstCharPtr>]
           >,
           FunctionSpec<
+              "strdup",
+              RetValSpec<CharPtr>,
+              [ArgSpec<ConstCharPtr>]
+          >,
+          FunctionSpec<
               "strpbrk",
               RetValSpec<CharPtr>,
               [ArgSpec<ConstCharPtr>, ArgSpec<ConstCharPtr>]
           >,
           FunctionSpec<
               "strrchr",
               RetValSpec<CharPtr>,
               [ArgSpec<ConstCharPtr>, ArgSpec<IntType>]

libc/src/string/CMakeLists.txt

SRCS
     strcspn.cpp
   HDRS
     strcspn.h
   DEPENDS
     .string_utils
 )
 
 add_entrypoint_object(
+  strdup
+  SRCS
+    strdup.cpp
+  HDRS
+    strdup.h
+  DEPENDS
+    .memcpy
+    .string_utils
+    libc.include.stdlib
+)
+
+add_entrypoint_object(
   strlen
   SRCS
     strlen.cpp
   HDRS
     strlen.h
   DEPENDS
     libc.include.string
 )

libc/src/string/strdup.h

+//===-- Implementation header for strdup ------------------------*- 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_STRING_STRDUP_H
+#define LLVM_LIBC_SRC_STRING_STRDUP_H
+
+#include <string.h>
+
+namespace __llvm_libc {
+
+char *strdup(const char *src);
+
+} // namespace __llvm_libc
+
+#endif // LLVM_LIBC_SRC_STRING_STRDUP_H

libc/src/string/strdup.cpp

+//===-- Implementation of strdup ------------------------------------------===//
+//
+// 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 "src/string/strdup.h"
+#include "src/string/memcpy.h"
+#include "src/string/string_utils.h"
+
+#include "src/__support/common.h"
+
+#include <stdlib.h>
+
+namespace __llvm_libc {
+
+LLVM_LIBC_FUNCTION(char *, strdup, (const char *src)) {
+  if (src == nullptr) {
+    return nullptr;
+  }
+  size_t len = internal::string_length(src) + 1;
+  char *dest = reinterpret_cast<char *>(::malloc(len)); // NOLINT
+  if (dest == nullptr) {
+    return nullptr;
+  }
+  char *result = reinterpret_cast<char *>(__llvm_libc::memcpy(dest, src, len));
+  return result;
+}
+
+} // namespace __llvm_libc

libc/test/src/string/CMakeLists.txt

SUITE
     libc_string_unittests
   SRCS
     strcspn_test.cpp
   DEPENDS
     libc.src.string.strcspn
 )
 
 add_libc_unittest(
+  strdup_test
+  SUITE
+    libc_string_unittests
+  SRCS
+    strdup_test.cpp
+  DEPENDS
+    libc.include.stdlib
+    libc.src.string.strdup
+)
+
+add_libc_unittest(
   strlen_test
   SUITE
     libc_string_unittests
   SRCS
     strlen_test.cpp
   DEPENDS
     libc.src.string.strlen
 )

libc/test/src/string/strdup_test.cpp

+//===-- Unittests for strdup ----------------------------------------------===//
+//
+// 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 "src/string/strdup.h"
+#include "utils/UnitTest/Test.h"
+#include <stdlib.h>
+
+TEST(LlvmLibcStrDupTest, EmptyString) {
+  const char *empty = "";
+
+  char *result = __llvm_libc::strdup(empty);
+  ASSERT_NE(result, static_cast<char *>(nullptr));
+  ASSERT_NE(empty, const_cast<const char *>(result));
+  ASSERT_STREQ(empty, result);
+  ::free(result);
+}
+
+TEST(LlvmLibcStrDupTest, AnyString) {
+  const char *abc = "abc";
+
+  char *result = __llvm_libc::strdup(abc);
+
+  ASSERT_NE(result, static_cast<char *>(nullptr));
+  ASSERT_NE(abc, const_cast<const char *>(result));
+  ASSERT_STREQ(abc, result);
+  ::free(result);
+}
+
+TEST(LlvmLibcStrDupTest, NullPtr) {
+
+  char *result = __llvm_libc::strdup(nullptr);
+
+  ASSERT_EQ(result, static_cast<char *>(nullptr));
+}