[flang-commits] [compiler-rt] [llvm] [libcxxabi] [clang-tools-extra] [clang] [libunwind] [libc] [flang] [lldb] [libcxx] [lld] [libc] Add more functions in CPP/bit.h (PR #73814)

Guillaume Chatelet via flang-commits flang-commits at lists.llvm.org
Thu Nov 30 04:50:42 PST 2023


https://github.com/gchatelet updated https://github.com/llvm/llvm-project/pull/73814

>From bdd0a0855c0dd98c93977db6982e480ab270b3cd Mon Sep 17 00:00:00 2001
From: Guillaume Chatelet <gchatelet at google.com>
Date: Wed, 29 Nov 2023 16:29:12 +0000
Subject: [PATCH 1/8] [libc] Add more functions in CPP/bit.h

---
 libc/src/__support/CPP/CMakeLists.txt         |   1 +
 libc/src/__support/CPP/bit.h                  | 242 ++++++++++++++++--
 libc/src/__support/CPP/limits.h               |  18 +-
 libc/test/src/__support/CPP/CMakeLists.txt    |  11 +
 libc/test/src/__support/CPP/bit_test.cpp      | 216 ++++++++++++++++
 .../llvm-project-overlay/libc/BUILD.bazel     |   1 +
 .../libc/test/src/__support/CPP/BUILD.bazel   |   9 +
 7 files changed, 478 insertions(+), 20 deletions(-)
 create mode 100644 libc/test/src/__support/CPP/bit_test.cpp

diff --git a/libc/src/__support/CPP/CMakeLists.txt b/libc/src/__support/CPP/CMakeLists.txt
index 10bcebf9b04f61d..5c47b2a815356b9 100644
--- a/libc/src/__support/CPP/CMakeLists.txt
+++ b/libc/src/__support/CPP/CMakeLists.txt
@@ -15,6 +15,7 @@ add_header_library(
   HDRS
     bit.h
   DEPENDS
+    .limits
     .type_traits
     libc.src.__support.macros.attributes
     libc.src.__support.macros.config
diff --git a/libc/src/__support/CPP/bit.h b/libc/src/__support/CPP/bit.h
index f6f3131c1ccfd81..f744d804398e213 100644
--- a/libc/src/__support/CPP/bit.h
+++ b/libc/src/__support/CPP/bit.h
@@ -1,57 +1,261 @@
-//===-- Freestanding version of bit_cast  -----------------------*- C++ -*-===//
+//===-- Implementation of the C++20 bit header  -----------------*- 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
 //
 //===----------------------------------------------------------------------===//
+// This is inspired from LLVM ADT/bit.h header.
 
 #ifndef LLVM_LIBC_SRC___SUPPORT_CPP_BIT_H
 #define LLVM_LIBC_SRC___SUPPORT_CPP_BIT_H
 
+#include "src/__support/CPP/limits.h" // numeric_limits
 #include "src/__support/CPP/type_traits.h"
 #include "src/__support/macros/attributes.h"
 #include "src/__support/macros/config.h" // LIBC_HAS_BUILTIN
 #include "src/__support/macros/sanitizer.h"
 
-namespace LIBC_NAMESPACE::cpp {
+#include <stdint.h>
 
-#if LIBC_HAS_BUILTIN(__builtin_bit_cast)
-#define LLVM_LIBC_HAS_BUILTIN_BIT_CAST
-#endif
+namespace LIBC_NAMESPACE::cpp {
 
 #if LIBC_HAS_BUILTIN(__builtin_memcpy_inline)
 #define LLVM_LIBC_HAS_BUILTIN_MEMCPY_INLINE
 #endif
 
-// This function guarantees the bitcast to be optimized away by the compiler for
-// GCC >= 8 and Clang >= 6.
-template <class To, class From>
+// This implementation of bit_cast requires trivially-constructible To, to avoid
+// UB in the implementation.
+template <
+    typename To, typename From,
+    typename = cpp::enable_if_t<sizeof(To) == sizeof(From)>,
+    typename = cpp::enable_if_t<cpp::is_trivially_constructible<To>::value>,
+    typename = cpp::enable_if_t<cpp::is_trivially_copyable<To>::value>,
+    typename = cpp::enable_if_t<cpp::is_trivially_copyable<From>::value>>
 LIBC_INLINE constexpr To bit_cast(const From &from) {
-  static_assert(sizeof(To) == sizeof(From), "To and From must be of same size");
-  static_assert(cpp::is_trivially_copyable<To>::value &&
-                    cpp::is_trivially_copyable<From>::value,
-                "Cannot bit-cast instances of non-trivially copyable classes.");
   MSAN_UNPOISON(&from, sizeof(From));
-#if defined(LLVM_LIBC_HAS_BUILTIN_BIT_CAST)
+#if LIBC_HAS_BUILTIN(__builtin_bit_cast)
   return __builtin_bit_cast(To, from);
 #else
-  static_assert(cpp::is_trivially_constructible<To>::value,
-                "This implementation additionally requires destination type to "
-                "be trivially constructible");
   To to;
   char *dst = reinterpret_cast<char *>(&to);
   const char *src = reinterpret_cast<const char *>(&from);
-#if defined(LLVM_LIBC_HAS_BUILTIN_MEMCPY_INLINE)
+#if LIBC_HAS_BUILTIN(__builtin_memcpy_inline)
   __builtin_memcpy_inline(dst, src, sizeof(To));
 #else
   for (unsigned i = 0; i < sizeof(To); ++i)
     dst[i] = src[i];
-#endif // defined(LLVM_LIBC_HAS_BUILTIN_MEMCPY_INLINE)
+#endif // LIBC_HAS_BUILTIN(__builtin_memcpy_inline)
   return to;
-#endif // defined(LLVM_LIBC_HAS_BUILTIN_BIT_CAST)
+#endif // LIBC_HAS_BUILTIN(__builtin_bit_cast)
+}
+
+template <typename T, typename = cpp::enable_if_t<cpp::is_unsigned_v<T>>>
+[[nodiscard]] LIBC_INLINE constexpr bool has_single_bit(T value) {
+  return (value != 0) && ((value & (value - 1)) == 0);
+}
+
+// A temporary macro to add template function specialization when compiler
+// builtin is available.
+#define ADD_SPECIALIZATION(NAME, TYPE, BUILTIN)                                \
+  template <> [[nodiscard]] LIBC_INLINE constexpr int NAME<TYPE>(TYPE value) { \
+    static_assert(cpp::is_unsigned_v<TYPE>);                                   \
+    return value == 0 ? cpp::numeric_limits<TYPE>::digits : BUILTIN(value);    \
+  }
+
+/// Count number of 0's from the least significant bit to the most
+///   stopping at the first 1.
+///
+/// Only unsigned integral types are allowed.
+///
+/// Returns cpp::numeric_limits<T>::digits on an input of 0.
+template <typename T>
+[[nodiscard]] LIBC_INLINE constexpr int countr_zero(T value) {
+  static_assert(cpp::is_unsigned_v<T>);
+  if (!value)
+    return cpp::numeric_limits<T>::digits;
+  if (value & 0x1)
+    return 0;
+  // Bisection method.
+  unsigned zero_bits = 0;
+  T shift = cpp::numeric_limits<T>::digits >> 1;
+  T mask = cpp::numeric_limits<T>::max() >> shift;
+  while (shift) {
+    if ((value & mask) == 0) {
+      value >>= shift;
+      zero_bits |= shift;
+    }
+    shift >>= 1;
+    mask >>= shift;
+  }
+  return zero_bits;
+}
+#if LIBC_HAS_BUILTIN(__builtin_ctzs)
+ADD_SPECIALIZATION(countr_zero, unsigned short, __builtin_ctzs)
+#endif
+#if LIBC_HAS_BUILTIN(__builtin_ctz)
+ADD_SPECIALIZATION(countr_zero, unsigned int, __builtin_ctz)
+#endif
+#if LIBC_HAS_BUILTIN(__builtin_ctzl)
+ADD_SPECIALIZATION(countr_zero, unsigned long, __builtin_ctzl)
+#endif
+#if LIBC_HAS_BUILTIN(__builtin_ctzll)
+ADD_SPECIALIZATION(countr_zero, unsigned long long, __builtin_ctzll)
+#endif
+
+/// Count number of 0's from the most significant bit to the least
+///   stopping at the first 1.
+///
+/// Only unsigned integral types are allowed.
+///
+/// Returns cpp::numeric_limits<T>::digits on an input of 0.
+template <typename T>
+[[nodiscard]] LIBC_INLINE constexpr int countl_zero(T value) {
+  static_assert(cpp::is_unsigned_v<T>);
+  if (!value)
+    return cpp::numeric_limits<T>::digits;
+  // Bisection method.
+  unsigned zero_bits = 0;
+  for (T shift = cpp::numeric_limits<T>::digits >> 1; shift; shift >>= 1) {
+    T tmp = value >> shift;
+    if (tmp)
+      value = tmp;
+    else
+      zero_bits |= shift;
+  }
+  return zero_bits;
+}
+#if LIBC_HAS_BUILTIN(__builtin_clzs)
+ADD_SPECIALIZATION(countl_zero, unsigned short, __builtin_clzs)
+#endif
+#if LIBC_HAS_BUILTIN(__builtin_clz)
+ADD_SPECIALIZATION(countl_zero, unsigned int, __builtin_clz)
+#endif
+#if LIBC_HAS_BUILTIN(__builtin_clzl)
+ADD_SPECIALIZATION(countl_zero, unsigned long, __builtin_clzl)
+#endif
+#if LIBC_HAS_BUILTIN(__builtin_clzll)
+ADD_SPECIALIZATION(countl_zero, unsigned long long, __builtin_clzll)
+#endif
+
+#undef ADD_SPECIALIZATION
+
+/// Count the number of ones from the most significant bit to the first
+/// zero bit.
+///
+/// Ex. countl_one(0xFF0FFF00) == 8.
+/// Only unsigned integral types are allowed.
+///
+/// Returns cpp::numeric_limits<T>::digits on an input of all ones.
+template <typename T>
+[[nodiscard]] LIBC_INLINE constexpr int countl_one(T value) {
+  static_assert(cpp::is_unsigned_v<T>);
+  return cpp::countl_zero<T>(~value);
+}
+
+/// Count the number of ones from the least significant bit to the first
+/// zero bit.
+///
+/// Ex. countr_one(0x00FF00FF) == 8.
+/// Only unsigned integral types are allowed.
+///
+/// Returns cpp::numeric_limits<T>::digits on an input of all ones.
+template <typename T>
+[[nodiscard]] LIBC_INLINE constexpr int countr_one(T value) {
+  static_assert(cpp::is_unsigned_v<T>);
+  return cpp::countr_zero<T>(~value);
+}
+
+/// Returns the number of bits needed to represent value if value is nonzero.
+/// Returns 0 otherwise.
+///
+/// Ex. bit_width(5) == 3.
+template <typename T>
+[[nodiscard]] LIBC_INLINE constexpr int bit_width(T value) {
+  static_assert(cpp::is_unsigned_v<T>);
+  return cpp::numeric_limits<T>::digits - cpp::countl_zero(value);
+}
+
+/// Returns the largest integral power of two no greater than value if value is
+/// nonzero.  Returns 0 otherwise.
+///
+/// Ex. bit_floor(5) == 4.
+template <typename T> [[nodiscard]] LIBC_INLINE constexpr T bit_floor(T value) {
+  static_assert(cpp::is_unsigned_v<T>);
+  if (!value)
+    return 0;
+  return T(1) << (cpp::bit_width(value) - 1);
+}
+
+/// Returns the smallest integral power of two no smaller than value if value is
+/// nonzero.  Returns 1 otherwise.
+///
+/// Ex. bit_ceil(5) == 8.
+///
+/// The return value is undefined if the input is larger than the largest power
+/// of two representable in T.
+template <typename T> [[nodiscard]] LIBC_INLINE constexpr T bit_ceil(T value) {
+  static_assert(cpp::is_unsigned_v<T>);
+  if (value < 2)
+    return 1;
+  return T(1) << cpp::bit_width<T>(value - 1u);
+}
+
+/// Count the number of set bits in a value.
+/// Ex. popcount(0xF000F000) = 8
+/// Returns 0 if the word is zero.
+template <typename T, typename = cpp::enable_if_t<cpp::is_unsigned_v<T>>>
+[[nodiscard]] LIBC_INLINE constexpr int popcount(T value) {
+  if constexpr (sizeof(T) == 8) {
+#if LIBC_HAS_BUILTIN(__builtin_popcountll)
+    return (int)__builtin_popcountll(value);
+#endif
+    uint64_t v = value;
+    v = v - ((v >> 1) & 0x5555555555555555ULL);
+    v = (v & 0x3333333333333333ULL) + ((v >> 2) & 0x3333333333333333ULL);
+    v = (v + (v >> 4)) & 0x0F0F0F0F0F0F0F0FULL;
+    return int((uint64_t)(v * 0x0101010101010101ULL) >> 56);
+  } else if constexpr (sizeof(T) <= 4) {
+#if LIBC_HAS_BUILTIN(__builtin_popcount)
+    return (int)__builtin_popcount(value);
+#endif
+    uint32_t v = value;
+    v = v - ((v >> 1) & 0x55555555);
+    v = (v & 0x33333333) + ((v >> 2) & 0x33333333);
+    return int(((v + (v >> 4) & 0xF0F0F0F) * 0x1010101) >> 24);
+  } else {
+    static_assert(cpp::always_false<T>);
+  }
+}
+
+// Forward-declare rotr so that rotl can use it.
+template <typename T, typename = cpp::enable_if_t<cpp::is_unsigned_v<T>>>
+[[nodiscard]] LIBC_INLINE constexpr T rotr(T value, int rotate);
+
+template <typename T, typename = cpp::enable_if_t<cpp::is_unsigned_v<T>>>
+[[nodiscard]] LIBC_INLINE constexpr T rotl(T value, int rotate) {
+  constexpr unsigned N = cpp::numeric_limits<T>::digits;
+  rotate = rotate % N;
+  if (!rotate)
+    return value;
+  if (rotate < 0)
+    return cpp::rotr(value, -rotate);
+  return (value << rotate) | (value >> (N - rotate));
+}
+
+template <typename T, typename>
+[[nodiscard]] LIBC_INLINE constexpr T rotr(T value, int rotate) {
+  constexpr unsigned N = cpp::numeric_limits<T>::digits;
+  rotate = rotate % N;
+  if (!rotate)
+    return value;
+  if (rotate < 0)
+    return cpp::rotl(value, -rotate);
+  return (value >> rotate) | (value << (N - rotate));
 }
 
+// TODO: Do we need this function at all? How is it different from
+// 'static_cast'?
 template <class To, class From>
 LIBC_INLINE constexpr To bit_or_static_cast(const From &from) {
   if constexpr (sizeof(To) == sizeof(From)) {
diff --git a/libc/src/__support/CPP/limits.h b/libc/src/__support/CPP/limits.h
index b1a1203df79674a..b56f2557a737cab 100644
--- a/libc/src/__support/CPP/limits.h
+++ b/libc/src/__support/CPP/limits.h
@@ -11,7 +11,7 @@
 
 #include "src/__support/macros/attributes.h" // LIBC_INLINE
 
-#include <limits.h>
+#include <limits.h> // CHAR_BIT
 
 namespace LIBC_NAMESPACE {
 namespace cpp {
@@ -36,41 +36,53 @@ template <> class numeric_limits<int> {
 public:
   LIBC_INLINE static constexpr int max() { return INT_MAX; }
   LIBC_INLINE static constexpr int min() { return INT_MIN; }
+  LIBC_INLINE_VAR static constexpr int digits = CHAR_BIT * sizeof(int) - 1;
 };
 template <> class numeric_limits<unsigned int> {
 public:
   LIBC_INLINE static constexpr unsigned int max() { return UINT_MAX; }
   LIBC_INLINE static constexpr unsigned int min() { return 0; }
+  LIBC_INLINE_VAR static constexpr int digits = CHAR_BIT * sizeof(unsigned int);
 };
 template <> class numeric_limits<long> {
 public:
   LIBC_INLINE static constexpr long max() { return LONG_MAX; }
   LIBC_INLINE static constexpr long min() { return LONG_MIN; }
+  LIBC_INLINE_VAR static constexpr int digits = CHAR_BIT * sizeof(long) - 1;
 };
 template <> class numeric_limits<unsigned long> {
 public:
   LIBC_INLINE static constexpr unsigned long max() { return ULONG_MAX; }
   LIBC_INLINE static constexpr unsigned long min() { return 0; }
+  LIBC_INLINE_VAR static constexpr int digits =
+      CHAR_BIT * sizeof(unsigned long);
 };
 template <> class numeric_limits<long long> {
 public:
   LIBC_INLINE static constexpr long long max() { return LLONG_MAX; }
   LIBC_INLINE static constexpr long long min() { return LLONG_MIN; }
+  LIBC_INLINE_VAR static constexpr int digits =
+      CHAR_BIT * sizeof(long long) - 1;
 };
 template <> class numeric_limits<unsigned long long> {
 public:
   LIBC_INLINE static constexpr unsigned long long max() { return ULLONG_MAX; }
   LIBC_INLINE static constexpr unsigned long long min() { return 0; }
+  LIBC_INLINE_VAR static constexpr int digits =
+      CHAR_BIT * sizeof(unsigned long long);
 };
 template <> class numeric_limits<short> {
 public:
   LIBC_INLINE static constexpr short max() { return SHRT_MAX; }
   LIBC_INLINE static constexpr short min() { return SHRT_MIN; }
+  LIBC_INLINE_VAR static constexpr int digits = CHAR_BIT * sizeof(short) - 1;
 };
 template <> class numeric_limits<unsigned short> {
 public:
   LIBC_INLINE static constexpr unsigned short max() { return USHRT_MAX; }
   LIBC_INLINE static constexpr unsigned short min() { return 0; }
+  LIBC_INLINE_VAR static constexpr int digits =
+      CHAR_BIT * sizeof(unsigned short);
 };
 template <> class numeric_limits<char> {
 public:
@@ -81,11 +93,13 @@ template <> class numeric_limits<signed char> {
 public:
   LIBC_INLINE static constexpr signed char max() { return SCHAR_MAX; }
   LIBC_INLINE static constexpr signed char min() { return SCHAR_MIN; }
+  LIBC_INLINE_VAR static constexpr int digits = CHAR_BIT - 1;
 };
 template <> class numeric_limits<unsigned char> {
 public:
   LIBC_INLINE static constexpr unsigned char max() { return UCHAR_MAX; }
   LIBC_INLINE static constexpr unsigned char min() { return 0; }
+  LIBC_INLINE_VAR static constexpr int digits = CHAR_BIT;
 };
 #ifdef __SIZEOF_INT128__
 // On platform where UInt128 resolves to __uint128_t, this specialization
@@ -94,6 +108,8 @@ template <> class numeric_limits<__uint128_t> {
 public:
   LIBC_INLINE static constexpr __uint128_t max() { return ~__uint128_t(0); }
   LIBC_INLINE static constexpr __uint128_t min() { return 0; }
+  LIBC_INLINE_VAR static constexpr int digits =
+      CHAR_BIT * sizeof(__uint128_t) - 1;
 };
 #endif
 
diff --git a/libc/test/src/__support/CPP/CMakeLists.txt b/libc/test/src/__support/CPP/CMakeLists.txt
index 5772a83a65cad13..6927579289bc2f3 100644
--- a/libc/test/src/__support/CPP/CMakeLists.txt
+++ b/libc/test/src/__support/CPP/CMakeLists.txt
@@ -1,5 +1,16 @@
 add_custom_target(libc-cpp-utils-tests)
 
+add_libc_test(
+  bit_test
+  SUITE
+    libc-cpp-utils-tests
+  SRCS
+    bit_test.cpp
+  DEPENDS
+    libc.src.__support.CPP.bit
+    libc.src.__support.uint
+)
+
 add_libc_test(
   bitset_test
   SUITE
diff --git a/libc/test/src/__support/CPP/bit_test.cpp b/libc/test/src/__support/CPP/bit_test.cpp
new file mode 100644
index 000000000000000..caec621c5a6f64a
--- /dev/null
+++ b/libc/test/src/__support/CPP/bit_test.cpp
@@ -0,0 +1,216 @@
+//===-- Unittests for Bit -------------------------------------------------===//
+//
+// 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/__support/CPP/bit.h"
+#include "src/__support/UInt.h"
+#include "test/UnitTest/Test.h"
+
+#include <stdint.h>
+
+namespace LIBC_NAMESPACE::cpp {
+
+using UnsignedTypes =
+    testing::TypeList<unsigned char, unsigned short, unsigned int,
+                      unsigned long, unsigned long long>;
+
+TYPED_TEST(LlvmLibcBitTest, HasSingleBit, UnsignedTypes) {
+  EXPECT_FALSE(has_single_bit<T>(T(0)));
+  EXPECT_FALSE(has_single_bit<T>(~T(0)));
+  for (T value = 1; value; value <<= 1)
+    EXPECT_TRUE(has_single_bit<T>(value));
+}
+
+TYPED_TEST(LlvmLibcBitTest, CountLZero, UnsignedTypes) {
+  EXPECT_EQ(countl_zero<T>(T(0)), cpp::numeric_limits<T>::digits);
+  int expected = 0;
+  for (T value = ~T(0); value; value >>= 1, ++expected)
+    EXPECT_EQ(countl_zero<T>(value), expected);
+}
+
+TYPED_TEST(LlvmLibcBitTest, CountRZero, UnsignedTypes) {
+  EXPECT_EQ(countr_zero<T>(T(0)), cpp::numeric_limits<T>::digits);
+  int expected = 0;
+  for (T value = ~T(0); value; value <<= 1, ++expected)
+    EXPECT_EQ(countr_zero<T>(value), expected);
+}
+
+TYPED_TEST(LlvmLibcBitTest, CountLOne, UnsignedTypes) {
+  EXPECT_EQ(countl_one<T>(T(0)), 0);
+  int expected = cpp::numeric_limits<T>::digits;
+  for (T value = ~T(0); value; value <<= 1, --expected)
+    EXPECT_EQ(countl_one<T>(value), expected);
+}
+
+TYPED_TEST(LlvmLibcBitTest, CountROne, UnsignedTypes) {
+  EXPECT_EQ(countr_one<T>(T(0)), 0);
+  int expected = cpp::numeric_limits<T>::digits;
+  for (T value = ~T(0); value; value >>= 1, --expected)
+    EXPECT_EQ(countr_one<T>(value), expected);
+}
+
+TYPED_TEST(LlvmLibcBitTest, BitWidth, UnsignedTypes) {
+  EXPECT_EQ(bit_width<T>(T(0)), 0);
+  int one_index = 0;
+  for (T value = 1; value; value <<= 1, ++one_index)
+    EXPECT_EQ(bit_width<T>(value), one_index + 1);
+}
+
+TEST(LlvmLibcBitTest, BitCeil) {
+  EXPECT_EQ(uint8_t(1), bit_ceil(uint8_t(0)));
+  EXPECT_EQ(uint16_t(1), bit_ceil(uint16_t(0)));
+  EXPECT_EQ(uint32_t(1), bit_ceil(uint32_t(0)));
+  EXPECT_EQ(uint64_t(1), bit_ceil(uint64_t(0)));
+
+  EXPECT_EQ(uint8_t(1), bit_ceil(uint8_t(1)));
+  EXPECT_EQ(uint16_t(1), bit_ceil(uint16_t(1)));
+  EXPECT_EQ(uint32_t(1), bit_ceil(uint32_t(1)));
+  EXPECT_EQ(uint64_t(1), bit_ceil(uint64_t(1)));
+
+  EXPECT_EQ(uint8_t(2), bit_ceil(uint8_t(2)));
+  EXPECT_EQ(uint16_t(2), bit_ceil(uint16_t(2)));
+  EXPECT_EQ(uint32_t(2), bit_ceil(uint32_t(2)));
+  EXPECT_EQ(uint64_t(2), bit_ceil(uint64_t(2)));
+
+  EXPECT_EQ(uint8_t(4), bit_ceil(uint8_t(3)));
+  EXPECT_EQ(uint16_t(4), bit_ceil(uint16_t(3)));
+  EXPECT_EQ(uint32_t(4), bit_ceil(uint32_t(3)));
+  EXPECT_EQ(uint64_t(4), bit_ceil(uint64_t(3)));
+
+  EXPECT_EQ(uint8_t(4), bit_ceil(uint8_t(4)));
+  EXPECT_EQ(uint16_t(4), bit_ceil(uint16_t(4)));
+  EXPECT_EQ(uint32_t(4), bit_ceil(uint32_t(4)));
+  EXPECT_EQ(uint64_t(4), bit_ceil(uint64_t(4)));
+
+  // The result is the largest representable value for each type.
+  EXPECT_EQ(uint8_t(0x80), bit_ceil(uint8_t(0x7f)));
+  EXPECT_EQ(uint16_t(0x8000), bit_ceil(uint16_t(0x7fff)));
+  EXPECT_EQ(uint32_t(0x80000000), bit_ceil(uint32_t(0x7fffffff)));
+  EXPECT_EQ(uint64_t(0x8000000000000000),
+            bit_ceil(uint64_t(0x7fffffffffffffff)));
+}
+
+TEST(LlvmLibcBitTest, BitFloor) {
+  EXPECT_EQ(uint8_t(0), bit_floor(uint8_t(0)));
+  EXPECT_EQ(uint16_t(0), bit_floor(uint16_t(0)));
+  EXPECT_EQ(uint32_t(0), bit_floor(uint32_t(0)));
+  EXPECT_EQ(uint64_t(0), bit_floor(uint64_t(0)));
+
+  EXPECT_EQ(uint8_t(1), bit_floor(uint8_t(1)));
+  EXPECT_EQ(uint16_t(1), bit_floor(uint16_t(1)));
+  EXPECT_EQ(uint32_t(1), bit_floor(uint32_t(1)));
+  EXPECT_EQ(uint64_t(1), bit_floor(uint64_t(1)));
+
+  EXPECT_EQ(uint8_t(2), bit_floor(uint8_t(2)));
+  EXPECT_EQ(uint16_t(2), bit_floor(uint16_t(2)));
+  EXPECT_EQ(uint32_t(2), bit_floor(uint32_t(2)));
+  EXPECT_EQ(uint64_t(2), bit_floor(uint64_t(2)));
+
+  EXPECT_EQ(uint8_t(2), bit_floor(uint8_t(3)));
+  EXPECT_EQ(uint16_t(2), bit_floor(uint16_t(3)));
+  EXPECT_EQ(uint32_t(2), bit_floor(uint32_t(3)));
+  EXPECT_EQ(uint64_t(2), bit_floor(uint64_t(3)));
+
+  EXPECT_EQ(uint8_t(4), bit_floor(uint8_t(4)));
+  EXPECT_EQ(uint16_t(4), bit_floor(uint16_t(4)));
+  EXPECT_EQ(uint32_t(4), bit_floor(uint32_t(4)));
+  EXPECT_EQ(uint64_t(4), bit_floor(uint64_t(4)));
+
+  EXPECT_EQ(uint8_t(0x40), bit_floor(uint8_t(0x7f)));
+  EXPECT_EQ(uint16_t(0x4000), bit_floor(uint16_t(0x7fff)));
+  EXPECT_EQ(uint32_t(0x40000000), bit_floor(uint32_t(0x7fffffff)));
+  EXPECT_EQ(uint64_t(0x4000000000000000),
+            bit_floor(uint64_t(0x7fffffffffffffffull)));
+
+  EXPECT_EQ(uint8_t(0x80), bit_floor(uint8_t(0x80)));
+  EXPECT_EQ(uint16_t(0x8000), bit_floor(uint16_t(0x8000)));
+  EXPECT_EQ(uint32_t(0x80000000), bit_floor(uint32_t(0x80000000)));
+  EXPECT_EQ(uint64_t(0x8000000000000000),
+            bit_floor(uint64_t(0x8000000000000000)));
+
+  EXPECT_EQ(uint8_t(0x80), bit_floor(uint8_t(0xff)));
+  EXPECT_EQ(uint16_t(0x8000), bit_floor(uint16_t(0xffff)));
+  EXPECT_EQ(uint32_t(0x80000000), bit_floor(uint32_t(0xffffffff)));
+  EXPECT_EQ(uint64_t(0x8000000000000000),
+            bit_floor(uint64_t(0xffffffffffffffff)));
+}
+
+TEST(LlvmLibcBitTest, PopCount) {
+  EXPECT_EQ(0, popcount(0U));
+  EXPECT_EQ(0, popcount(0ULL));
+
+  EXPECT_EQ(32, popcount(~0U));
+  EXPECT_EQ(64, popcount(~0ULL));
+
+  for (int i = 0; i != 32; ++i)
+    EXPECT_EQ(1, popcount(1U << i));
+}
+
+TYPED_TEST(LlvmLibcBitTest, RotateIsInvariantForZeroAndOne, UnsignedTypes) {
+  constexpr T all_zeros = T(0);
+  constexpr T all_ones = ~T(0);
+  for (int i = 0; i < cpp::numeric_limits<T>::digits; ++i) {
+    EXPECT_EQ(rotl<T>(all_zeros, i), all_zeros);
+    EXPECT_EQ(rotl<T>(all_ones, i), all_ones);
+    EXPECT_EQ(rotr<T>(all_zeros, i), all_zeros);
+    EXPECT_EQ(rotr<T>(all_ones, i), all_ones);
+  }
+}
+
+TEST(LlvmLibcBitTest, Rotl) {
+  EXPECT_EQ(uint8_t(0x53U), rotl<uint8_t>(0x53, 0));
+  EXPECT_EQ(uint8_t(0x4dU), rotl<uint8_t>(0x53, 2));
+  EXPECT_EQ(uint8_t(0xa6U), rotl<uint8_t>(0x53, 9));
+  EXPECT_EQ(uint8_t(0x9aU), rotl<uint8_t>(0x53, -5));
+
+  EXPECT_EQ(uint16_t(0xabcdU), rotl<uint16_t>(0xabcd, 0));
+  EXPECT_EQ(uint16_t(0xf36aU), rotl<uint16_t>(0xabcd, 6));
+  EXPECT_EQ(uint16_t(0xaf36U), rotl<uint16_t>(0xabcd, 18));
+  EXPECT_EQ(uint16_t(0xf36aU), rotl<uint16_t>(0xabcd, -10));
+
+  EXPECT_EQ(uint32_t(0xdeadbeefU), rotl<uint32_t>(0xdeadbeef, 0));
+  EXPECT_EQ(uint32_t(0x7ddfbd5bU), rotl<uint32_t>(0xdeadbeef, 17));
+  EXPECT_EQ(uint32_t(0x5b7ddfbdU), rotl<uint32_t>(0xdeadbeef, 41));
+  EXPECT_EQ(uint32_t(0xb6fbbf7aU), rotl<uint32_t>(0xdeadbeef, -22));
+
+  EXPECT_EQ(uint64_t(0x12345678deadbeefULL),
+            rotl<uint64_t>(0x12345678deadbeefULL, 0));
+  EXPECT_EQ(uint64_t(0xf56df77891a2b3c6ULL),
+            rotl<uint64_t>(0x12345678deadbeefULL, 35));
+  EXPECT_EQ(uint64_t(0x8d159e37ab6fbbc4ULL),
+            rotl<uint64_t>(0x12345678deadbeefULL, 70));
+  EXPECT_EQ(uint64_t(0xb7dde2468acf1bd5ULL),
+            rotl<uint64_t>(0x12345678deadbeefULL, -19));
+}
+
+TEST(LlvmLibcBitTest, Rotr) {
+  EXPECT_EQ(uint8_t(0x53U), rotr<uint8_t>(0x53, 0));
+  EXPECT_EQ(uint8_t(0xd4U), rotr<uint8_t>(0x53, 2));
+  EXPECT_EQ(uint8_t(0xa9U), rotr<uint8_t>(0x53, 9));
+  EXPECT_EQ(uint8_t(0x6aU), rotr<uint8_t>(0x53, -5));
+
+  EXPECT_EQ(uint16_t(0xabcdU), rotr<uint16_t>(0xabcd, 0));
+  EXPECT_EQ(uint16_t(0x36afU), rotr<uint16_t>(0xabcd, 6));
+  EXPECT_EQ(uint16_t(0x6af3U), rotr<uint16_t>(0xabcd, 18));
+  EXPECT_EQ(uint16_t(0x36afU), rotr<uint16_t>(0xabcd, -10));
+
+  EXPECT_EQ(uint32_t(0xdeadbeefU), rotr<uint32_t>(0xdeadbeef, 0));
+  EXPECT_EQ(uint32_t(0xdf77ef56U), rotr<uint32_t>(0xdeadbeef, 17));
+  EXPECT_EQ(uint32_t(0x77ef56dfU), rotr<uint32_t>(0xdeadbeef, 41));
+  EXPECT_EQ(uint32_t(0xbbf7ab6fU), rotr<uint32_t>(0xdeadbeef, -22));
+
+  EXPECT_EQ(uint64_t(0x12345678deadbeefULL),
+            rotr<uint64_t>(0x12345678deadbeefULL, 0));
+  EXPECT_EQ(uint64_t(0x1bd5b7dde2468acfULL),
+            rotr<uint64_t>(0x12345678deadbeefULL, 35));
+  EXPECT_EQ(uint64_t(0xbc48d159e37ab6fbULL),
+            rotr<uint64_t>(0x12345678deadbeefULL, 70));
+  EXPECT_EQ(uint64_t(0xb3c6f56df77891a2ULL),
+            rotr<uint64_t>(0x12345678deadbeefULL, -19));
+}
+
+} // namespace LIBC_NAMESPACE::cpp
diff --git a/utils/bazel/llvm-project-overlay/libc/BUILD.bazel b/utils/bazel/llvm-project-overlay/libc/BUILD.bazel
index ffb2652ed1419d0..4146fb4cb36f3d3 100644
--- a/utils/bazel/llvm-project-overlay/libc/BUILD.bazel
+++ b/utils/bazel/llvm-project-overlay/libc/BUILD.bazel
@@ -161,6 +161,7 @@ libc_support_library(
     name = "__support_cpp_bit",
     hdrs = ["src/__support/CPP/bit.h"],
     deps = [
+        ":__support_cpp_limits",
         ":__support_cpp_type_traits",
         ":__support_macros_attributes",
         ":__support_macros_config",
diff --git a/utils/bazel/llvm-project-overlay/libc/test/src/__support/CPP/BUILD.bazel b/utils/bazel/llvm-project-overlay/libc/test/src/__support/CPP/BUILD.bazel
index 59dcfd6ca93ca2a..07d9ac1d500820b 100644
--- a/utils/bazel/llvm-project-overlay/libc/test/src/__support/CPP/BUILD.bazel
+++ b/utils/bazel/llvm-project-overlay/libc/test/src/__support/CPP/BUILD.bazel
@@ -22,6 +22,15 @@ libc_test(
     deps = ["//libc:__support_cpp_bitset"],
 )
 
+libc_test(
+    name = "bit_test",
+    srcs = ["bit_test.cpp"],
+    deps = [
+        "//libc:__support_cpp_bit",
+        "//libc:__support_uint",
+    ],
+)
+
 libc_test(
     name = "cstddef_test",
     srcs = ["cstddef_test.cpp"],

>From c91367d4d99e0857efbd8c51cc743e74bee750aa Mon Sep 17 00:00:00 2001
From: Guillaume Chatelet <gchatelet at google.com>
Date: Thu, 30 Nov 2023 09:31:44 +0000
Subject: [PATCH 2/8] SFINAE all the functions

---
 libc/src/__support/CPP/bit.h | 25 ++++++++++---------------
 1 file changed, 10 insertions(+), 15 deletions(-)

diff --git a/libc/src/__support/CPP/bit.h b/libc/src/__support/CPP/bit.h
index f744d804398e213..0059ad1ef7e3559 100644
--- a/libc/src/__support/CPP/bit.h
+++ b/libc/src/__support/CPP/bit.h
@@ -5,7 +5,7 @@
 // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
 //
 //===----------------------------------------------------------------------===//
-// This is inspired from LLVM ADT/bit.h header.
+// This is inspired by LLVM ADT/bit.h header.
 
 #ifndef LLVM_LIBC_SRC___SUPPORT_CPP_BIT_H
 #define LLVM_LIBC_SRC___SUPPORT_CPP_BIT_H
@@ -69,9 +69,8 @@ template <typename T, typename = cpp::enable_if_t<cpp::is_unsigned_v<T>>>
 /// Only unsigned integral types are allowed.
 ///
 /// Returns cpp::numeric_limits<T>::digits on an input of 0.
-template <typename T>
+template <typename T, typename = cpp::enable_if_t<cpp::is_unsigned_v<T>>>
 [[nodiscard]] LIBC_INLINE constexpr int countr_zero(T value) {
-  static_assert(cpp::is_unsigned_v<T>);
   if (!value)
     return cpp::numeric_limits<T>::digits;
   if (value & 0x1)
@@ -109,9 +108,8 @@ ADD_SPECIALIZATION(countr_zero, unsigned long long, __builtin_ctzll)
 /// Only unsigned integral types are allowed.
 ///
 /// Returns cpp::numeric_limits<T>::digits on an input of 0.
-template <typename T>
+template <typename T, typename = cpp::enable_if_t<cpp::is_unsigned_v<T>>>
 [[nodiscard]] LIBC_INLINE constexpr int countl_zero(T value) {
-  static_assert(cpp::is_unsigned_v<T>);
   if (!value)
     return cpp::numeric_limits<T>::digits;
   // Bisection method.
@@ -147,9 +145,8 @@ ADD_SPECIALIZATION(countl_zero, unsigned long long, __builtin_clzll)
 /// Only unsigned integral types are allowed.
 ///
 /// Returns cpp::numeric_limits<T>::digits on an input of all ones.
-template <typename T>
+template <typename T, typename = cpp::enable_if_t<cpp::is_unsigned_v<T>>>
 [[nodiscard]] LIBC_INLINE constexpr int countl_one(T value) {
-  static_assert(cpp::is_unsigned_v<T>);
   return cpp::countl_zero<T>(~value);
 }
 
@@ -160,9 +157,8 @@ template <typename T>
 /// Only unsigned integral types are allowed.
 ///
 /// Returns cpp::numeric_limits<T>::digits on an input of all ones.
-template <typename T>
+template <typename T, typename = cpp::enable_if_t<cpp::is_unsigned_v<T>>>
 [[nodiscard]] LIBC_INLINE constexpr int countr_one(T value) {
-  static_assert(cpp::is_unsigned_v<T>);
   return cpp::countr_zero<T>(~value);
 }
 
@@ -170,9 +166,8 @@ template <typename T>
 /// Returns 0 otherwise.
 ///
 /// Ex. bit_width(5) == 3.
-template <typename T>
+template <typename T, typename = cpp::enable_if_t<cpp::is_unsigned_v<T>>>
 [[nodiscard]] LIBC_INLINE constexpr int bit_width(T value) {
-  static_assert(cpp::is_unsigned_v<T>);
   return cpp::numeric_limits<T>::digits - cpp::countl_zero(value);
 }
 
@@ -180,8 +175,8 @@ template <typename T>
 /// nonzero.  Returns 0 otherwise.
 ///
 /// Ex. bit_floor(5) == 4.
-template <typename T> [[nodiscard]] LIBC_INLINE constexpr T bit_floor(T value) {
-  static_assert(cpp::is_unsigned_v<T>);
+template <typename T, typename = cpp::enable_if_t<cpp::is_unsigned_v<T>>>
+[[nodiscard]] LIBC_INLINE constexpr T bit_floor(T value) {
   if (!value)
     return 0;
   return T(1) << (cpp::bit_width(value) - 1);
@@ -194,8 +189,8 @@ template <typename T> [[nodiscard]] LIBC_INLINE constexpr T bit_floor(T value) {
 ///
 /// The return value is undefined if the input is larger than the largest power
 /// of two representable in T.
-template <typename T> [[nodiscard]] LIBC_INLINE constexpr T bit_ceil(T value) {
-  static_assert(cpp::is_unsigned_v<T>);
+template <typename T, typename = cpp::enable_if_t<cpp::is_unsigned_v<T>>>
+[[nodiscard]] LIBC_INLINE constexpr T bit_ceil(T value) {
   if (value < 2)
     return 1;
   return T(1) << cpp::bit_width<T>(value - 1u);

>From 95ade514c90aaf5caf9e1b442adc2da94c15a9a9 Mon Sep 17 00:00:00 2001
From: Guillaume Chatelet <gchatelet at google.com>
Date: Thu, 30 Nov 2023 09:56:02 +0000
Subject: [PATCH 3/8] Add more documentation

---
 libc/src/__support/CPP/bit.h | 7 +++++++
 1 file changed, 7 insertions(+)

diff --git a/libc/src/__support/CPP/bit.h b/libc/src/__support/CPP/bit.h
index 0059ad1ef7e3559..01cec4ce6b01dd0 100644
--- a/libc/src/__support/CPP/bit.h
+++ b/libc/src/__support/CPP/bit.h
@@ -199,6 +199,10 @@ template <typename T, typename = cpp::enable_if_t<cpp::is_unsigned_v<T>>>
 /// Count the number of set bits in a value.
 /// Ex. popcount(0xF000F000) = 8
 /// Returns 0 if the word is zero.
+///
+/// The non-builtin version uses an algorithm that counts bit set by dividing
+/// the integer into multiple lanes and reducing the result.
+/// https://graphics.stanford.edu/~seander/bithacks.html#CountBitsSetParallel
 template <typename T, typename = cpp::enable_if_t<cpp::is_unsigned_v<T>>>
 [[nodiscard]] LIBC_INLINE constexpr int popcount(T value) {
   if constexpr (sizeof(T) == 8) {
@@ -223,6 +227,9 @@ template <typename T, typename = cpp::enable_if_t<cpp::is_unsigned_v<T>>>
   }
 }
 
+// Rotate algorithms make use of "Safe, Efficient, and Portable Rotate in C/C++"
+// from https://blog.regehr.org/archives/1063.
+
 // Forward-declare rotr so that rotl can use it.
 template <typename T, typename = cpp::enable_if_t<cpp::is_unsigned_v<T>>>
 [[nodiscard]] LIBC_INLINE constexpr T rotr(T value, int rotate);

>From 0404e52cd15b5893fddce0d7551a0f03ae02bd3a Mon Sep 17 00:00:00 2001
From: Guillaume Chatelet <gchatelet at google.com>
Date: Thu, 30 Nov 2023 10:07:49 +0000
Subject: [PATCH 4/8] Refactor numeric_limits

---
 libc/src/__support/CPP/limits.h | 136 ++++++++++++++------------------
 1 file changed, 58 insertions(+), 78 deletions(-)

diff --git a/libc/src/__support/CPP/limits.h b/libc/src/__support/CPP/limits.h
index b56f2557a737cab..3126b38ec48e612 100644
--- a/libc/src/__support/CPP/limits.h
+++ b/libc/src/__support/CPP/limits.h
@@ -9,6 +9,7 @@
 #ifndef LLVM_LIBC_SRC___SUPPORT_CPP_LIMITS_H
 #define LLVM_LIBC_SRC___SUPPORT_CPP_LIMITS_H
 
+#include "src/__support/CPP/type_traits/is_signed.h"
 #include "src/__support/macros/attributes.h" // LIBC_INLINE
 
 #include <limits.h> // CHAR_BIT
@@ -24,93 +25,72 @@ constexpr long long LLONG_MIN = 1LL << (LLONG_BIT_WIDTH - 1);
 constexpr unsigned long long ULLONG_MAX = ~0ULL;
 #endif
 
-template <class T> class numeric_limits {
-public:
-  LIBC_INLINE static constexpr T max();
-  LIBC_INLINE static constexpr T min();
+namespace internal {
+
+template <typename T, T min_value, T max_value> struct numeric_limits_impl {
+  LIBC_INLINE static constexpr T max() { return max_value; }
+  LIBC_INLINE static constexpr T min() { return min_value; }
+  LIBC_INLINE_VAR static constexpr int digits =
+      CHAR_BIT * sizeof(T) - cpp::is_signed_v<T>;
 };
 
+} // namespace internal
+
+template <class T> struct numeric_limits {};
+
 // TODO: Add numeric_limits specializations as needed for new types.
+template <>
+struct numeric_limits<short>
+    : public internal::numeric_limits_impl<short, SHRT_MIN, SHRT_MAX> {};
 
-template <> class numeric_limits<int> {
-public:
-  LIBC_INLINE static constexpr int max() { return INT_MAX; }
-  LIBC_INLINE static constexpr int min() { return INT_MIN; }
-  LIBC_INLINE_VAR static constexpr int digits = CHAR_BIT * sizeof(int) - 1;
-};
-template <> class numeric_limits<unsigned int> {
-public:
-  LIBC_INLINE static constexpr unsigned int max() { return UINT_MAX; }
-  LIBC_INLINE static constexpr unsigned int min() { return 0; }
-  LIBC_INLINE_VAR static constexpr int digits = CHAR_BIT * sizeof(unsigned int);
-};
-template <> class numeric_limits<long> {
-public:
-  LIBC_INLINE static constexpr long max() { return LONG_MAX; }
-  LIBC_INLINE static constexpr long min() { return LONG_MIN; }
-  LIBC_INLINE_VAR static constexpr int digits = CHAR_BIT * sizeof(long) - 1;
-};
-template <> class numeric_limits<unsigned long> {
-public:
-  LIBC_INLINE static constexpr unsigned long max() { return ULONG_MAX; }
-  LIBC_INLINE static constexpr unsigned long min() { return 0; }
-  LIBC_INLINE_VAR static constexpr int digits =
-      CHAR_BIT * sizeof(unsigned long);
-};
-template <> class numeric_limits<long long> {
-public:
-  LIBC_INLINE static constexpr long long max() { return LLONG_MAX; }
-  LIBC_INLINE static constexpr long long min() { return LLONG_MIN; }
-  LIBC_INLINE_VAR static constexpr int digits =
-      CHAR_BIT * sizeof(long long) - 1;
-};
-template <> class numeric_limits<unsigned long long> {
-public:
-  LIBC_INLINE static constexpr unsigned long long max() { return ULLONG_MAX; }
-  LIBC_INLINE static constexpr unsigned long long min() { return 0; }
-  LIBC_INLINE_VAR static constexpr int digits =
-      CHAR_BIT * sizeof(unsigned long long);
-};
-template <> class numeric_limits<short> {
-public:
-  LIBC_INLINE static constexpr short max() { return SHRT_MAX; }
-  LIBC_INLINE static constexpr short min() { return SHRT_MIN; }
-  LIBC_INLINE_VAR static constexpr int digits = CHAR_BIT * sizeof(short) - 1;
-};
-template <> class numeric_limits<unsigned short> {
-public:
-  LIBC_INLINE static constexpr unsigned short max() { return USHRT_MAX; }
-  LIBC_INLINE static constexpr unsigned short min() { return 0; }
-  LIBC_INLINE_VAR static constexpr int digits =
-      CHAR_BIT * sizeof(unsigned short);
-};
-template <> class numeric_limits<char> {
-public:
-  LIBC_INLINE static constexpr char max() { return CHAR_MAX; }
-  LIBC_INLINE static constexpr char min() { return CHAR_MIN; }
-};
-template <> class numeric_limits<signed char> {
-public:
-  LIBC_INLINE static constexpr signed char max() { return SCHAR_MAX; }
-  LIBC_INLINE static constexpr signed char min() { return SCHAR_MIN; }
-  LIBC_INLINE_VAR static constexpr int digits = CHAR_BIT - 1;
+template <>
+struct numeric_limits<unsigned short>
+    : public internal::numeric_limits_impl<unsigned short, 0, USHRT_MAX> {};
+
+template <>
+struct numeric_limits<int>
+    : public internal::numeric_limits_impl<int, INT_MIN, INT_MAX> {};
+
+template <>
+struct numeric_limits<unsigned int>
+    : public internal::numeric_limits_impl<unsigned int, 0, UINT_MAX> {};
+
+template <>
+struct numeric_limits<long>
+    : public internal::numeric_limits_impl<long, LONG_MIN, LONG_MAX> {};
+
+template <>
+struct numeric_limits<unsigned long>
+    : public internal::numeric_limits_impl<unsigned long, 0, ULONG_MAX> {};
+
+template <>
+struct numeric_limits<long long>
+    : public internal::numeric_limits_impl<long long, LLONG_MIN, LLONG_MAX> {};
+
+template <>
+struct numeric_limits<unsigned long long>
+    : public internal::numeric_limits_impl<unsigned long long, 0, ULLONG_MAX> {
 };
-template <> class numeric_limits<unsigned char> {
-public:
-  LIBC_INLINE static constexpr unsigned char max() { return UCHAR_MAX; }
-  LIBC_INLINE static constexpr unsigned char min() { return 0; }
-  LIBC_INLINE_VAR static constexpr int digits = CHAR_BIT;
+
+template <>
+struct numeric_limits<char>
+    : public internal::numeric_limits_impl<char, CHAR_MIN, CHAR_MAX> {};
+
+template <>
+struct numeric_limits<signed char>
+    : public internal::numeric_limits_impl<signed char, SCHAR_MIN, SCHAR_MAX> {
 };
+
+template <>
+struct numeric_limits<unsigned char>
+    : public internal::numeric_limits_impl<unsigned char, 0, UCHAR_MAX> {};
+
 #ifdef __SIZEOF_INT128__
 // On platform where UInt128 resolves to __uint128_t, this specialization
 // provides the limits of UInt128.
-template <> class numeric_limits<__uint128_t> {
-public:
-  LIBC_INLINE static constexpr __uint128_t max() { return ~__uint128_t(0); }
-  LIBC_INLINE static constexpr __uint128_t min() { return 0; }
-  LIBC_INLINE_VAR static constexpr int digits =
-      CHAR_BIT * sizeof(__uint128_t) - 1;
-};
+template <>
+struct numeric_limits<__uint128_t>
+    : public internal::numeric_limits_impl<__uint128_t, 0, ~__uint128_t(0)> {};
 #endif
 
 } // namespace cpp

>From d2c83406d1e5dedd7b4ca87a8d6cc7abde967ee4 Mon Sep 17 00:00:00 2001
From: Guillaume Chatelet <gchatelet at google.com>
Date: Thu, 30 Nov 2023 10:09:33 +0000
Subject: [PATCH 5/8] Fix comment

---
 libc/test/src/__support/CPP/bit_test.cpp | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/libc/test/src/__support/CPP/bit_test.cpp b/libc/test/src/__support/CPP/bit_test.cpp
index caec621c5a6f64a..c64ee3d232f09e8 100644
--- a/libc/test/src/__support/CPP/bit_test.cpp
+++ b/libc/test/src/__support/CPP/bit_test.cpp
@@ -86,7 +86,7 @@ TEST(LlvmLibcBitTest, BitCeil) {
   EXPECT_EQ(uint32_t(4), bit_ceil(uint32_t(4)));
   EXPECT_EQ(uint64_t(4), bit_ceil(uint64_t(4)));
 
-  // The result is the largest representable value for each type.
+  // The result is the representable power of 2 for each type.
   EXPECT_EQ(uint8_t(0x80), bit_ceil(uint8_t(0x7f)));
   EXPECT_EQ(uint16_t(0x8000), bit_ceil(uint16_t(0x7fff)));
   EXPECT_EQ(uint32_t(0x80000000), bit_ceil(uint32_t(0x7fffffff)));

>From 8de6a9058244c6797b1f00068f31c4dfc2643e1f Mon Sep 17 00:00:00 2001
From: Guillaume Chatelet <gchatelet at google.com>
Date: Thu, 30 Nov 2023 10:22:04 +0000
Subject: [PATCH 6/8] Remove popcount as we don't have a use case for it now.

---
 libc/src/__support/CPP/bit.h             | 31 ------------------------
 libc/test/src/__support/CPP/bit_test.cpp | 11 ---------
 2 files changed, 42 deletions(-)

diff --git a/libc/src/__support/CPP/bit.h b/libc/src/__support/CPP/bit.h
index 01cec4ce6b01dd0..abea20b67537125 100644
--- a/libc/src/__support/CPP/bit.h
+++ b/libc/src/__support/CPP/bit.h
@@ -196,37 +196,6 @@ template <typename T, typename = cpp::enable_if_t<cpp::is_unsigned_v<T>>>
   return T(1) << cpp::bit_width<T>(value - 1u);
 }
 
-/// Count the number of set bits in a value.
-/// Ex. popcount(0xF000F000) = 8
-/// Returns 0 if the word is zero.
-///
-/// The non-builtin version uses an algorithm that counts bit set by dividing
-/// the integer into multiple lanes and reducing the result.
-/// https://graphics.stanford.edu/~seander/bithacks.html#CountBitsSetParallel
-template <typename T, typename = cpp::enable_if_t<cpp::is_unsigned_v<T>>>
-[[nodiscard]] LIBC_INLINE constexpr int popcount(T value) {
-  if constexpr (sizeof(T) == 8) {
-#if LIBC_HAS_BUILTIN(__builtin_popcountll)
-    return (int)__builtin_popcountll(value);
-#endif
-    uint64_t v = value;
-    v = v - ((v >> 1) & 0x5555555555555555ULL);
-    v = (v & 0x3333333333333333ULL) + ((v >> 2) & 0x3333333333333333ULL);
-    v = (v + (v >> 4)) & 0x0F0F0F0F0F0F0F0FULL;
-    return int((uint64_t)(v * 0x0101010101010101ULL) >> 56);
-  } else if constexpr (sizeof(T) <= 4) {
-#if LIBC_HAS_BUILTIN(__builtin_popcount)
-    return (int)__builtin_popcount(value);
-#endif
-    uint32_t v = value;
-    v = v - ((v >> 1) & 0x55555555);
-    v = (v & 0x33333333) + ((v >> 2) & 0x33333333);
-    return int(((v + (v >> 4) & 0xF0F0F0F) * 0x1010101) >> 24);
-  } else {
-    static_assert(cpp::always_false<T>);
-  }
-}
-
 // Rotate algorithms make use of "Safe, Efficient, and Portable Rotate in C/C++"
 // from https://blog.regehr.org/archives/1063.
 
diff --git a/libc/test/src/__support/CPP/bit_test.cpp b/libc/test/src/__support/CPP/bit_test.cpp
index c64ee3d232f09e8..b7cd99d169fc52e 100644
--- a/libc/test/src/__support/CPP/bit_test.cpp
+++ b/libc/test/src/__support/CPP/bit_test.cpp
@@ -139,17 +139,6 @@ TEST(LlvmLibcBitTest, BitFloor) {
             bit_floor(uint64_t(0xffffffffffffffff)));
 }
 
-TEST(LlvmLibcBitTest, PopCount) {
-  EXPECT_EQ(0, popcount(0U));
-  EXPECT_EQ(0, popcount(0ULL));
-
-  EXPECT_EQ(32, popcount(~0U));
-  EXPECT_EQ(64, popcount(~0ULL));
-
-  for (int i = 0; i != 32; ++i)
-    EXPECT_EQ(1, popcount(1U << i));
-}
-
 TYPED_TEST(LlvmLibcBitTest, RotateIsInvariantForZeroAndOne, UnsignedTypes) {
   constexpr T all_zeros = T(0);
   constexpr T all_ones = ~T(0);

>From 4cfc2df0370ef3c33df983cdf6bde60eee5d0499 Mon Sep 17 00:00:00 2001
From: Guillaume Chatelet <gchatelet at google.com>
Date: Thu, 30 Nov 2023 10:23:43 +0000
Subject: [PATCH 7/8] Add comment

---
 libc/src/__support/CPP/bit.h | 1 +
 1 file changed, 1 insertion(+)

diff --git a/libc/src/__support/CPP/bit.h b/libc/src/__support/CPP/bit.h
index abea20b67537125..2091c3662d5076b 100644
--- a/libc/src/__support/CPP/bit.h
+++ b/libc/src/__support/CPP/bit.h
@@ -6,6 +6,7 @@
 //
 //===----------------------------------------------------------------------===//
 // This is inspired by LLVM ADT/bit.h header.
+// Some functions are missing, we can add them as needed (popcount, byteswap).
 
 #ifndef LLVM_LIBC_SRC___SUPPORT_CPP_BIT_H
 #define LLVM_LIBC_SRC___SUPPORT_CPP_BIT_H

>From d5969755d9892f4eab01bc51165f9e7eca99f4c6 Mon Sep 17 00:00:00 2001
From: Guillaume Chatelet <gchatelet at google.com>
Date: Thu, 30 Nov 2023 12:50:18 +0000
Subject: [PATCH 8/8] Commit newline at EOF

---
 libc/src/__support/CPP/limits.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/libc/src/__support/CPP/limits.h b/libc/src/__support/CPP/limits.h
index f471a6016a0fa81..e92ad00f80e0379 100644
--- a/libc/src/__support/CPP/limits.h
+++ b/libc/src/__support/CPP/limits.h
@@ -96,4 +96,4 @@ struct numeric_limits<__uint128_t>
 } // namespace cpp
 } // namespace LIBC_NAMESPACE
 
-#endif // LLVM_LIBC_SRC___SUPPORT_CPP_LIMITS_H
\ No newline at end of file
+#endif // LLVM_LIBC_SRC___SUPPORT_CPP_LIMITS_H



More information about the flang-commits mailing list