Export of internal Abseil changes

--
f28b989d5161230c6561e923b458c797a96bcf90 by Greg Falcon <gfalcon@google.com>:

Import of CCTZ from GitHub.

PiperOrigin-RevId: 263586488

--
8259484025b7de45358719fc6182a48cac8044c6 by Andy Soffer <asoffer@google.com>:

Internal changes and combine namespaces into a single namespace.

PiperOrigin-RevId: 263560576

--
8d19f41661984a600d1f8bbfeb8a30fcb4dee7d6 by Mark Barolak <mbar@google.com>:

Inside of absl::string_view::copy, use absl::string_view::traits_type::copy instead of std:copy to do the actual work.  This both follows the C++ standard more closely and avoids avoid MSVC unchecked iterator warnings.

PiperOrigin-RevId: 263430502

--
c06bf74236e12c7c1c97bfcbbc9d29bd65d6b36c by Andy Soffer <asoffer@google.com>:

Remove force-inlining attributes. Benchmarking results indicate that they are creating meaningful performance differences.

PiperOrigin-RevId: 263364896

--
ec4fa6eac958a9521456201b138784f55d3b17bc by Abseil Team <absl-team@google.com>:

Make BM_Fill benchmarks more representative.

PiperOrigin-RevId: 263349482

--
4ae280b4eb31d9cb58e847eb670473340f7778c1 by Derek Mauro <dmauro@google.com>:

Fix new -Wdeprecated-copy warning in gcc9

PiperOrigin-RevId: 263348118

--
d238a92f452a5c35686f9c71596fdd1fe62090a2 by Matt Calabrese <calabrese@google.com>:

The std::is_trivially_xxx fail on versions of GCC up until 7.4 due to faulty underlying intrinsics, but our emulation succeeds. Update our traits to not compare against the standard library implementation in these versions.

PiperOrigin-RevId: 263209457
GitOrigin-RevId: f28b989d5161230c6561e923b458c797a96bcf90
Change-Id: I4c41db5928ba71e243aeace4420e06d1a2df0b5b
This commit is contained in:
Abseil Team 2019-08-15 10:06:39 -07:00 committed by vslashg
parent d9aa92d7fb
commit aae8143cf9
9 changed files with 98 additions and 116 deletions

View file

@ -125,7 +125,7 @@
#error ABSL_HAVE_STD_IS_TRIVIALLY_ASSIGNABLE cannot directly set
#elif (defined(__clang__) && defined(_LIBCPP_VERSION)) || \
(!defined(__clang__) && defined(__GNUC__) && \
(__GNUC__ > 5 || (__GNUC__ == 5 && __GNUC_MINOR__ >= 1)) && \
(__GNUC__ > 7 || (__GNUC__ == 7 && __GNUC_MINOR__ >= 4)) && \
(defined(_LIBCPP_VERSION) || defined(__GLIBCXX__))) || \
(defined(_MSC_VER) && !defined(__NVCC__))
#define ABSL_HAVE_STD_IS_TRIVIALLY_CONSTRUCTIBLE 1

View file

@ -323,6 +323,7 @@ class unique_checker : public base_checker<TreeType, CheckerType> {
unique_checker(const unique_checker &x) : super_type(x) {}
template <class InputIterator>
unique_checker(InputIterator b, InputIterator e) : super_type(b, e) {}
unique_checker& operator=(const unique_checker&) = default;
// Insertion routines.
std::pair<iterator, bool> insert(const value_type &x) {
@ -370,6 +371,7 @@ class multi_checker : public base_checker<TreeType, CheckerType> {
multi_checker(const multi_checker &x) : super_type(x) {}
template <class InputIterator>
multi_checker(InputIterator b, InputIterator e) : super_type(b, e) {}
multi_checker& operator=(const multi_checker&) = default;
// Insertion routines.
iterator insert(const value_type &x) {

View file

@ -25,42 +25,45 @@
namespace {
void BM_InlinedVectorFill(benchmark::State& state) {
const int len = state.range(0);
absl::InlinedVector<int, 8> v;
int val = 10;
v.reserve(len);
for (auto _ : state) {
v.resize(0); // Use resize(0) as InlinedVector releases storage on clear().
for (int i = 0; i < len; ++i) {
v.push_back(i);
}
benchmark::DoNotOptimize(v);
v.push_back(val);
}
}
BENCHMARK(BM_InlinedVectorFill)->Range(0, 1024);
BENCHMARK(BM_InlinedVectorFill)->Range(1, 256);
void BM_InlinedVectorFillRange(benchmark::State& state) {
const int len = state.range(0);
std::unique_ptr<int[]> ia(new int[len]);
for (int i = 0; i < len; i++) {
ia[i] = i;
}
auto* from = ia.get();
auto* to = from + len;
const std::vector<int> src(len, len);
absl::InlinedVector<int, 8> v;
v.reserve(len);
for (auto _ : state) {
benchmark::DoNotOptimize(from);
benchmark::DoNotOptimize(to);
absl::InlinedVector<int, 8> v(from, to);
benchmark::DoNotOptimize(src);
v.assign(src.begin(), src.end());
benchmark::DoNotOptimize(v);
}
}
BENCHMARK(BM_InlinedVectorFillRange)->Range(0, 1024);
BENCHMARK(BM_InlinedVectorFillRange)->Range(1, 256);
void BM_StdVectorFill(benchmark::State& state) {
const int len = state.range(0);
std::vector<int> v;
int val = 10;
v.reserve(len);
for (auto _ : state) {
v.clear();
for (int i = 0; i < len; ++i) {
v.push_back(i);
}
benchmark::DoNotOptimize(v);
benchmark::DoNotOptimize(val);
v.push_back(val);
}
}
BENCHMARK(BM_StdVectorFill)->Range(0, 1024);
BENCHMARK(BM_StdVectorFill)->Range(1, 256);
// The purpose of the next two benchmarks is to verify that
// absl::InlinedVector is efficient when moving is more efficent than

View file

@ -546,6 +546,28 @@ TEST(TypeTraitsTest, TestTrivialDefaultCtor) {
#endif
}
// GCC prior to 7.4 had a bug in its trivially-constructible traits
// (https://gcc.gnu.org/bugzilla/show_bug.cgi?id=80654).
// This test makes sure that we do not depend on the trait in these cases when
// implementing absl triviality traits.
template <class T>
struct BadConstructors {
BadConstructors() { static_assert(T::value, ""); }
BadConstructors(BadConstructors&&) { static_assert(T::value, ""); }
BadConstructors(const BadConstructors&) { static_assert(T::value, ""); }
};
TEST(TypeTraitsTest, TestTrivialityBadConstructors) {
using BadType = BadConstructors<int>;
EXPECT_FALSE(absl::is_trivially_default_constructible<BadType>::value);
EXPECT_FALSE(absl::is_trivially_move_constructible<BadType>::value);
EXPECT_FALSE(absl::is_trivially_copy_constructible<BadType>::value);
}
TEST(TypeTraitsTest, TestTrivialMoveCtor) {
// Verify that arithmetic types and pointers have trivial move
// constructors.

View file

@ -306,7 +306,10 @@ cc_library(
# anyway and thus there wouldn't be any gain from using it as a module.
features = ["-header_modules"],
linkopts = ABSL_DEFAULT_LINKOPTS,
deps = [":platform"],
deps = [
":platform",
"//absl/base:core_headers",
],
)
cc_binary(

View file

@ -22,39 +22,9 @@
#include <cstdint>
#include <cstring>
#include "absl/base/attributes.h"
#include "absl/random/internal/platform.h"
// ABSL_HAVE_ATTRIBUTE
#if !defined(ABSL_HAVE_ATTRIBUTE)
#ifdef __has_attribute
#define ABSL_HAVE_ATTRIBUTE(x) __has_attribute(x)
#else
#define ABSL_HAVE_ATTRIBUTE(x) 0
#endif
#endif
#if ABSL_HAVE_ATTRIBUTE(always_inline) || \
(defined(__GNUC__) && !defined(__clang__))
#define ABSL_RANDOM_INTERNAL_ATTRIBUTE_ALWAYS_INLINE \
__attribute__((always_inline))
#elif defined(_MSC_VER)
// We can achieve something similar to attribute((always_inline)) with MSVC by
// using the __forceinline keyword, however this is not perfect. MSVC is
// much less aggressive about inlining, and even with the __forceinline keyword.
#define ABSL_RANDOM_INTERNAL_ATTRIBUTE_ALWAYS_INLINE __forceinline
#else
#define ABSL_RANDOM_INTERNAL_ATTRIBUTE_ALWAYS_INLINE
#endif
// ABSL_ATTRIBUTE_FLATTEN enables much more aggressive inlining within
// the indicated function.
#undef ABSL_ATTRIBUTE_FLATTEN
#if ABSL_HAVE_ATTRIBUTE(flatten) || (defined(__GNUC__) && !defined(__clang__))
#define ABSL_ATTRIBUTE_FLATTEN __attribute__((flatten))
#else
#define ABSL_ATTRIBUTE_FLATTEN
#endif
// ABSL_RANDEN_HWAES_IMPL indicates whether this file will contain
// a hardware accelerated implementation of randen, or whether it
// will contain stubs that exit the process.
@ -146,18 +116,6 @@ void RandenHwAes::Generate(const void*, void*) {
#include "absl/random/internal/randen_traits.h"
// ABSL_FUNCTION_ALIGN32 defines a 32-byte alignment attribute
// for the functions in this file.
//
// NOTE: Determine whether we actually have any wins from ALIGN32
// using microbenchmarks. If not, remove.
#undef ABSL_FUNCTION_ALIGN32
#if ABSL_HAVE_ATTRIBUTE(aligned) || (defined(__GNUC__) && !defined(__clang__))
#define ABSL_FUNCTION_ALIGN32 __attribute__((aligned(32)))
#else
#define ABSL_FUNCTION_ALIGN32
#endif
// TARGET_CRYPTO defines a crypto attribute for each architecture.
//
// NOTE: Evaluate whether we should eliminate ABSL_TARGET_CRYPTO.
@ -191,8 +149,7 @@ using Vector128 = __vector unsigned long long; // NOLINT(runtime/int)
namespace {
inline ABSL_TARGET_CRYPTO ABSL_RANDOM_INTERNAL_ATTRIBUTE_ALWAYS_INLINE Vector128
ReverseBytes(const Vector128& v) {
inline ABSL_TARGET_CRYPTO Vector128 ReverseBytes(const Vector128& v) {
// Reverses the bytes of the vector.
const __vector unsigned char perm = {15, 14, 13, 12, 11, 10, 9, 8,
7, 6, 5, 4, 3, 2, 1, 0};
@ -202,26 +159,26 @@ ReverseBytes(const Vector128& v) {
// WARNING: these load/store in native byte order. It is OK to load and then
// store an unchanged vector, but interpreting the bits as a number or input
// to AES will have undefined results.
inline ABSL_TARGET_CRYPTO ABSL_RANDOM_INTERNAL_ATTRIBUTE_ALWAYS_INLINE Vector128
inline ABSL_TARGET_CRYPTO Vector128
Vector128Load(const void* ABSL_RANDOM_INTERNAL_RESTRICT from) {
return vec_vsx_ld(0, reinterpret_cast<const Vector128*>(from));
}
inline ABSL_TARGET_CRYPTO ABSL_RANDOM_INTERNAL_ATTRIBUTE_ALWAYS_INLINE void
Vector128Store(const Vector128& v, void* ABSL_RANDOM_INTERNAL_RESTRICT to) {
inline ABSL_TARGET_CRYPTO void Vector128Store(
const Vector128& v, void* ABSL_RANDOM_INTERNAL_RESTRICT to) {
vec_vsx_st(v, 0, reinterpret_cast<Vector128*>(to));
}
// One round of AES. "round_key" is a public constant for breaking the
// symmetry of AES (ensures previously equal columns differ afterwards).
inline ABSL_TARGET_CRYPTO ABSL_RANDOM_INTERNAL_ATTRIBUTE_ALWAYS_INLINE Vector128
AesRound(const Vector128& state, const Vector128& round_key) {
inline ABSL_TARGET_CRYPTO Vector128 AesRound(const Vector128& state,
const Vector128& round_key) {
return Vector128(__builtin_crypto_vcipher(state, round_key));
}
// Enables native loads in the round loop by pre-swapping.
inline ABSL_TARGET_CRYPTO ABSL_RANDOM_INTERNAL_ATTRIBUTE_ALWAYS_INLINE void
SwapEndian(uint64_t* ABSL_RANDOM_INTERNAL_RESTRICT state) {
inline ABSL_TARGET_CRYPTO void SwapEndian(
uint64_t* ABSL_RANDOM_INTERNAL_RESTRICT state) {
using absl::random_internal::RandenTraits;
constexpr size_t kLanes = 2;
constexpr size_t kFeistelBlocks = RandenTraits::kFeistelBlocks;
@ -273,20 +230,20 @@ using Vector128 = uint8x16_t;
namespace {
inline ABSL_TARGET_CRYPTO ABSL_RANDOM_INTERNAL_ATTRIBUTE_ALWAYS_INLINE Vector128
inline ABSL_TARGET_CRYPTO Vector128
Vector128Load(const void* ABSL_RANDOM_INTERNAL_RESTRICT from) {
return vld1q_u8(reinterpret_cast<const uint8_t*>(from));
}
inline ABSL_TARGET_CRYPTO ABSL_RANDOM_INTERNAL_ATTRIBUTE_ALWAYS_INLINE void
Vector128Store(const Vector128& v, void* ABSL_RANDOM_INTERNAL_RESTRICT to) {
inline ABSL_TARGET_CRYPTO void Vector128Store(
const Vector128& v, void* ABSL_RANDOM_INTERNAL_RESTRICT to) {
vst1q_u8(reinterpret_cast<uint8_t*>(to), v);
}
// One round of AES. "round_key" is a public constant for breaking the
// symmetry of AES (ensures previously equal columns differ afterwards).
inline ABSL_TARGET_CRYPTO ABSL_RANDOM_INTERNAL_ATTRIBUTE_ALWAYS_INLINE Vector128
AesRound(const Vector128& state, const Vector128& round_key) {
inline ABSL_TARGET_CRYPTO Vector128 AesRound(const Vector128& state,
const Vector128& round_key) {
// It is important to always use the full round function - omitting the
// final MixColumns reduces security [https://eprint.iacr.org/2010/041.pdf]
// and does not help because we never decrypt.
@ -297,8 +254,8 @@ AesRound(const Vector128& state, const Vector128& round_key) {
return vaesmcq_u8(vaeseq_u8(state, uint8x16_t{})) ^ round_key;
}
inline ABSL_TARGET_CRYPTO ABSL_RANDOM_INTERNAL_ATTRIBUTE_ALWAYS_INLINE void
SwapEndian(uint64_t* ABSL_RANDOM_INTERNAL_RESTRICT) {}
inline ABSL_TARGET_CRYPTO void SwapEndian(
uint64_t* ABSL_RANDOM_INTERNAL_RESTRICT) {}
} // namespace
@ -313,16 +270,11 @@ namespace {
class Vector128 {
public:
// Convert from/to intrinsics.
inline ABSL_RANDOM_INTERNAL_ATTRIBUTE_ALWAYS_INLINE explicit Vector128(
const __m128i& Vector128)
: data_(Vector128) {}
inline explicit Vector128(const __m128i& Vector128) : data_(Vector128) {}
inline ABSL_RANDOM_INTERNAL_ATTRIBUTE_ALWAYS_INLINE __m128i data() const {
return data_;
}
inline __m128i data() const { return data_; }
inline ABSL_RANDOM_INTERNAL_ATTRIBUTE_ALWAYS_INLINE Vector128& operator^=(
const Vector128& other) {
inline Vector128& operator^=(const Vector128& other) {
data_ = _mm_xor_si128(data_, other.data());
return *this;
}
@ -331,29 +283,29 @@ class Vector128 {
__m128i data_;
};
inline ABSL_TARGET_CRYPTO ABSL_RANDOM_INTERNAL_ATTRIBUTE_ALWAYS_INLINE Vector128
inline ABSL_TARGET_CRYPTO Vector128
Vector128Load(const void* ABSL_RANDOM_INTERNAL_RESTRICT from) {
return Vector128(_mm_load_si128(reinterpret_cast<const __m128i*>(from)));
}
inline ABSL_TARGET_CRYPTO ABSL_RANDOM_INTERNAL_ATTRIBUTE_ALWAYS_INLINE void
Vector128Store(const Vector128& v, void* ABSL_RANDOM_INTERNAL_RESTRICT to) {
inline ABSL_TARGET_CRYPTO void Vector128Store(
const Vector128& v, void* ABSL_RANDOM_INTERNAL_RESTRICT to) {
_mm_store_si128(reinterpret_cast<__m128i * ABSL_RANDOM_INTERNAL_RESTRICT>(to),
v.data());
}
// One round of AES. "round_key" is a public constant for breaking the
// symmetry of AES (ensures previously equal columns differ afterwards).
inline ABSL_TARGET_CRYPTO ABSL_RANDOM_INTERNAL_ATTRIBUTE_ALWAYS_INLINE Vector128
AesRound(const Vector128& state, const Vector128& round_key) {
inline ABSL_TARGET_CRYPTO Vector128 AesRound(const Vector128& state,
const Vector128& round_key) {
// It is important to always use the full round function - omitting the
// final MixColumns reduces security [https://eprint.iacr.org/2010/041.pdf]
// and does not help because we never decrypt.
return Vector128(_mm_aesenc_si128(state.data(), round_key.data()));
}
inline ABSL_TARGET_CRYPTO ABSL_RANDOM_INTERNAL_ATTRIBUTE_ALWAYS_INLINE void
SwapEndian(uint64_t* ABSL_RANDOM_INTERNAL_RESTRICT) {}
inline ABSL_TARGET_CRYPTO void SwapEndian(
uint64_t* ABSL_RANDOM_INTERNAL_RESTRICT) {}
} // namespace
@ -450,8 +402,8 @@ constexpr size_t kLanes = 2;
// Block shuffles applies a shuffle to the entire state between AES rounds.
// Improved odd-even shuffle from "New criterion for diffusion property".
inline ABSL_RANDOM_INTERNAL_ATTRIBUTE_ALWAYS_INLINE ABSL_TARGET_CRYPTO void
BlockShuffle(uint64_t* ABSL_RANDOM_INTERNAL_RESTRICT state) {
inline ABSL_TARGET_CRYPTO void BlockShuffle(
uint64_t* ABSL_RANDOM_INTERNAL_RESTRICT state) {
static_assert(kFeistelBlocks == 16, "Expecting 16 FeistelBlocks.");
constexpr size_t shuffle[kFeistelBlocks] = {7, 2, 13, 4, 11, 8, 3, 6,
@ -499,9 +451,8 @@ BlockShuffle(uint64_t* ABSL_RANDOM_INTERNAL_RESTRICT state) {
// per 16 bytes (vs. 10 for AES-CTR). Computing eight round functions in
// parallel hides the 7-cycle AESNI latency on HSW. Note that the Feistel
// XORs are 'free' (included in the second AES instruction).
inline ABSL_RANDOM_INTERNAL_ATTRIBUTE_ALWAYS_INLINE ABSL_TARGET_CRYPTO const
u64x2*
FeistelRound(uint64_t* ABSL_RANDOM_INTERNAL_RESTRICT state,
inline ABSL_TARGET_CRYPTO const u64x2* FeistelRound(
uint64_t* ABSL_RANDOM_INTERNAL_RESTRICT state,
const u64x2* ABSL_RANDOM_INTERNAL_RESTRICT keys) {
static_assert(kFeistelBlocks == 16, "Expecting 16 FeistelBlocks.");
@ -561,8 +512,8 @@ inline ABSL_RANDOM_INTERNAL_ATTRIBUTE_ALWAYS_INLINE ABSL_TARGET_CRYPTO const
// Indistinguishable from ideal by chosen-ciphertext adversaries using less than
// 2^64 queries if the round function is a PRF. This is similar to the b=8 case
// of Simpira v2, but more efficient than its generic construction for b=16.
inline ABSL_RANDOM_INTERNAL_ATTRIBUTE_ALWAYS_INLINE ABSL_TARGET_CRYPTO void
Permute(const void* ABSL_RANDOM_INTERNAL_RESTRICT keys,
inline ABSL_TARGET_CRYPTO void Permute(
const void* ABSL_RANDOM_INTERNAL_RESTRICT keys,
uint64_t* ABSL_RANDOM_INTERNAL_RESTRICT state) {
const u64x2* ABSL_RANDOM_INTERNAL_RESTRICT keys128 =
static_cast<const u64x2*>(keys);
@ -584,16 +535,15 @@ namespace random_internal {
bool HasRandenHwAesImplementation() { return true; }
const void* ABSL_TARGET_CRYPTO ABSL_FUNCTION_ALIGN32 ABSL_ATTRIBUTE_FLATTEN
RandenHwAes::GetKeys() {
const void* ABSL_TARGET_CRYPTO RandenHwAes::GetKeys() {
// Round keys for one AES per Feistel round and branch.
// The canonical implementation uses first digits of Pi.
return round_keys;
}
// NOLINTNEXTLINE
void ABSL_TARGET_CRYPTO ABSL_FUNCTION_ALIGN32 ABSL_ATTRIBUTE_FLATTEN
RandenHwAes::Absorb(const void* seed_void, void* state_void) {
void ABSL_TARGET_CRYPTO RandenHwAes::Absorb(const void* seed_void,
void* state_void) {
uint64_t* ABSL_RANDOM_INTERNAL_RESTRICT state =
reinterpret_cast<uint64_t*>(state_void);
const uint64_t* ABSL_RANDOM_INTERNAL_RESTRICT seed =
@ -669,8 +619,8 @@ RandenHwAes::Absorb(const void* seed_void, void* state_void) {
}
// NOLINTNEXTLINE
void ABSL_TARGET_CRYPTO ABSL_FUNCTION_ALIGN32 ABSL_ATTRIBUTE_FLATTEN
RandenHwAes::Generate(const void* keys, void* state_void) {
void ABSL_TARGET_CRYPTO RandenHwAes::Generate(const void* keys,
void* state_void) {
static_assert(kCapacityBytes == sizeof(Vector128), "Capacity mismatch");
uint64_t* ABSL_RANDOM_INTERNAL_RESTRICT state =

View file

@ -35,9 +35,6 @@ struct IntervalClosedClosedT {};
struct IntervalClosedOpenT {};
struct IntervalOpenClosedT {};
struct IntervalOpenOpenT {};
} // namespace random_internal
namespace random_internal {
// The functions
// uniform_lower_bound(tag, a, b)

View file

@ -342,7 +342,7 @@ class string_view {
size_type rlen = (std::min)(length_ - pos, n);
if (rlen > 0) {
const char* start = ptr_ + pos;
std::copy(start, start + rlen, buf);
traits_type::copy(buf, start, rlen);
}
return rlen;
}

View file

@ -33,7 +33,12 @@ using TimeZoneImplByName =
TimeZoneImplByName* time_zone_map = nullptr;
// Mutual exclusion for time_zone_map.
std::mutex time_zone_mutex;
std::mutex& TimeZoneMutex() {
// This mutex is intentionally "leaked" to avoid the static deinitialization
// order fiasco (std::mutex's destructor is not trivial on many platforms).
static std::mutex* time_zone_mutex = new std::mutex;
return *time_zone_mutex;
}
} // namespace
@ -54,7 +59,7 @@ bool time_zone::Impl::LoadTimeZone(const std::string& name, time_zone* tz) {
// Then check, under a shared lock, whether the time zone has already
// been loaded. This is the common path. TODO: Move to shared_mutex.
{
std::lock_guard<std::mutex> lock(time_zone_mutex);
std::lock_guard<std::mutex> lock(TimeZoneMutex());
if (time_zone_map != nullptr) {
TimeZoneImplByName::const_iterator itr = time_zone_map->find(name);
if (itr != time_zone_map->end()) {
@ -65,7 +70,7 @@ bool time_zone::Impl::LoadTimeZone(const std::string& name, time_zone* tz) {
}
// Now check again, under an exclusive lock.
std::lock_guard<std::mutex> lock(time_zone_mutex);
std::lock_guard<std::mutex> lock(TimeZoneMutex());
if (time_zone_map == nullptr) time_zone_map = new TimeZoneImplByName;
const Impl*& impl = (*time_zone_map)[name];
if (impl == nullptr) {
@ -84,7 +89,7 @@ bool time_zone::Impl::LoadTimeZone(const std::string& name, time_zone* tz) {
}
void time_zone::Impl::ClearTimeZoneMapTestOnly() {
std::lock_guard<std::mutex> lock(time_zone_mutex);
std::lock_guard<std::mutex> lock(TimeZoneMutex());
if (time_zone_map != nullptr) {
// Existing time_zone::Impl* entries are in the wild, so we simply
// leak them. Future requests will result in reloading the data.