From fa8c75182fbfdeddb2485fc0d53baeda3f40b7a3 Mon Sep 17 00:00:00 2001 From: Abseil Team Date: Wed, 13 Nov 2019 08:54:32 -0800 Subject: Export of internal Abseil changes -- 049ac45508e335c6f010f2d28d71016b9fa65b4e by Derek Mauro : Fix librt detection PiperOrigin-RevId: 280207723 -- 6382c3a9fb2643af9dc031f92ca846c4a78e249c by Andy Getzendanner : Fix Conan builds Import of https://github.com/abseil/abseil-cpp/pull/400 PiperOrigin-RevId: 280025424 -- aebcd52b1686ac82663a8d0193b60d0122a43372 by Samuel Benzaquen : Enable the assertion in the iterator's operator== and operator!= PiperOrigin-RevId: 279998951 -- 5b61d909e2159ac6fd45e0e456818db1e725ecd1 by Derek Mauro : Add best effort support for compiling much of Abseil with MinGW. This involves disabling ABSL_ATTRIBUTE_WEAK and adding link flags. A change to CCTZ is still necessary. Tests were not run yet, but most of them now build. PiperOrigin-RevId: 279966541 -- 4336f8b10cff906e2defdd7d1d449cde4907da5d by Abseil Team : Add comments and relax memory orders in base_internal::CallOnceImpl. Add a comment to document the memory order guarantee if base_internal::SpinLockWait() is called and returns kOnceDone. Add a comment for the load/store sequence in base_internal::CallOnceImpl based on Mike Burrows' explanation. The atomic load of 'control' in the #ifndef NDEBUG block does not need std::memory_order_acquire. It can use std::memory_order_relaxed. The atomic compare_exchange_strong of 'control' does not need std::memory_order_acquire in the success case. It can use std::memory_order_relaxed. PiperOrigin-RevId: 279814155 -- 407de3a5e9af957cded54a136ca0468bde620d4d by Abseil Team : Added a script to generate abseil.podspec from all BUILD.bazel files automatically. PiperOrigin-RevId: 279811441 -- 26139497d4a363d6c7bc989c554da593e8819a07 by Derek Mauro : Add missing copyright and Apache License to //absl/functional/BUILD.bazel PiperOrigin-RevId: 279795227 -- 98ed625b02af6e5834edf52a920d8ca2dab4cd90 by Matt Kulukundis : Switch the implementation of hashtablez to *only* work on platforms that have a PER_THREAD_TLS. The old case is very slow (global mutex) and nobody collects data from that configuration anyway. PiperOrigin-RevId: 279775149 -- 07225900ef672c005c38f467ad3f92f38d0922b3 by Derek Mauro : Remove the minumum glibc version check PiperOrigin-RevId: 279750412 -- ec09956a951b4f52228ecc81968b8db7ae19ed15 by Derek Mauro : CMake only: link with -lrt to support older glibc versions PiperOrigin-RevId: 279741661 -- 97b113fb2e8246f6152c36330ba13793b37154b6 by Xiaoyi Zhang : Internal change. PiperOrigin-RevId: 279390188 -- ca8f72f2721546cc9b01bd01b2ea144962e6e0c5 by Andy Getzendanner : Expose PutTwoDigits for internal use within Abseil. PiperOrigin-RevId: 279374239 -- 14c6384cc03bbdfdefd2e4b635f104af5dd7e026 by Derek Mauro : Remove log_severity sources from the base target. They are already compiled as part of a separate library. PiperOrigin-RevId: 279372619 -- 3c5d926c718f8bf394e3bee87b6ba8d94601e0d3 by Abseil Team : s/indepdent/independent/g in SimpleAtof's documentation. PiperOrigin-RevId: 279350836 -- de2c44be8a8edf9efa1fe2007cba3564f3e5b0b8 by Abseil Team : Internal change PiperOrigin-RevId: 279346990 -- 2ba078341423fcf6d0ba5ca1831f86570a26e615 by Samuel Benzaquen : Add hash support for std::wstring, std::u16string and std::u32string. PiperOrigin-RevId: 279320672 -- 3272d3ffcfa55283a04f90e5868701912da95ef7 by Andy Soffer : Removing a bunch of __restricts that amount to no performance differences. One of these is the cause of https://github.com/abseil/abseil-cpp/issues/396. In particular, in one of the Vector128Store functions, restricts on two pointers that were indeed aliased seems to be the root cause of the issues. Closes #396 PiperOrigin-RevId: 279318999 -- 342f338ab31cc24344d5de8f28cf455bbb629a17 by Jorg Brown : Support uint128 in SimpleAtoi PiperOrigin-RevId: 279234038 -- 81cb0a04cf2dc4515d303679fc60968712191571 by Derek Mauro : Change the check for futex availability to support older Linux systems PiperOrigin-RevId: 279147079 -- cb4ca4aa4c8d2d710a5d483c56c4ce4f979e14b1 by Abseil Team : Add IWYU pragma: export for int128 .inc files. PiperOrigin-RevId: 279107098 -- b8df86ef610c366729f07326c726f3e34817b4dd by Abseil Team : An optimization for Waiter::Post() in the SEM waiter mode. Like the FUTEX waiter mode, Waiter::Post() only needs to call Poke() if it incremented the atomic variable from 0. PiperOrigin-RevId: 279086133 GitOrigin-RevId: 049ac45508e335c6f010f2d28d71016b9fa65b4e Change-Id: I4c1a4073fff62cb6a1fcb1c104aa7d62dad588c2 --- absl/random/internal/randen_hwaes.cc | 48 ++++++++++++------------------------ 1 file changed, 16 insertions(+), 32 deletions(-) (limited to 'absl/random/internal/randen_hwaes.cc') diff --git a/absl/random/internal/randen_hwaes.cc b/absl/random/internal/randen_hwaes.cc index 7d5b2b7..6cc36fd 100644 --- a/absl/random/internal/randen_hwaes.cc +++ b/absl/random/internal/randen_hwaes.cc @@ -159,13 +159,11 @@ inline ABSL_TARGET_CRYPTO Vector128 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 Vector128 -Vector128Load(const void* ABSL_RANDOM_INTERNAL_RESTRICT from) { +inline ABSL_TARGET_CRYPTO Vector128 Vector128Load(const void* from) { return vec_vsx_ld(0, reinterpret_cast(from)); } -inline ABSL_TARGET_CRYPTO void Vector128Store( - const Vector128& v, void* ABSL_RANDOM_INTERNAL_RESTRICT to) { +inline ABSL_TARGET_CRYPTO void Vector128Store(const Vector128& v, void* to) { vec_vsx_st(v, 0, reinterpret_cast(to)); } @@ -177,8 +175,7 @@ inline ABSL_TARGET_CRYPTO Vector128 AesRound(const Vector128& state, } // Enables native loads in the round loop by pre-swapping. -inline ABSL_TARGET_CRYPTO void SwapEndian( - uint64_t* ABSL_RANDOM_INTERNAL_RESTRICT state) { +inline ABSL_TARGET_CRYPTO void SwapEndian(uint64_t* state) { using absl::random_internal::RandenTraits; constexpr size_t kLanes = 2; constexpr size_t kFeistelBlocks = RandenTraits::kFeistelBlocks; @@ -230,13 +227,11 @@ using Vector128 = uint8x16_t; namespace { -inline ABSL_TARGET_CRYPTO Vector128 -Vector128Load(const void* ABSL_RANDOM_INTERNAL_RESTRICT from) { +inline ABSL_TARGET_CRYPTO Vector128 Vector128Load(const void* from) { return vld1q_u8(reinterpret_cast(from)); } -inline ABSL_TARGET_CRYPTO void Vector128Store( - const Vector128& v, void* ABSL_RANDOM_INTERNAL_RESTRICT to) { +inline ABSL_TARGET_CRYPTO void Vector128Store(const Vector128& v, void* to) { vst1q_u8(reinterpret_cast(to), v); } @@ -254,8 +249,7 @@ inline ABSL_TARGET_CRYPTO Vector128 AesRound(const Vector128& state, return vaesmcq_u8(vaeseq_u8(state, uint8x16_t{})) ^ round_key; } -inline ABSL_TARGET_CRYPTO void SwapEndian( - uint64_t* ABSL_RANDOM_INTERNAL_RESTRICT) {} +inline ABSL_TARGET_CRYPTO void SwapEndian(uint64_t*) {} } // namespace @@ -283,15 +277,12 @@ class Vector128 { __m128i data_; }; -inline ABSL_TARGET_CRYPTO Vector128 -Vector128Load(const void* ABSL_RANDOM_INTERNAL_RESTRICT from) { +inline ABSL_TARGET_CRYPTO Vector128 Vector128Load(const void* from) { return Vector128(_mm_load_si128(reinterpret_cast(from))); } -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()); +inline ABSL_TARGET_CRYPTO void Vector128Store(const Vector128& v, void* to) { + _mm_store_si128(reinterpret_cast<__m128i*>(to), v.data()); } // One round of AES. "round_key" is a public constant for breaking the @@ -304,8 +295,7 @@ inline ABSL_TARGET_CRYPTO Vector128 AesRound(const Vector128& state, return Vector128(_mm_aesenc_si128(state.data(), round_key.data())); } -inline ABSL_TARGET_CRYPTO void SwapEndian( - uint64_t* ABSL_RANDOM_INTERNAL_RESTRICT) {} +inline ABSL_TARGET_CRYPTO void SwapEndian(uint64_t*) {} } // namespace @@ -402,8 +392,7 @@ 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_TARGET_CRYPTO void BlockShuffle( - uint64_t* ABSL_RANDOM_INTERNAL_RESTRICT state) { +inline ABSL_TARGET_CRYPTO void BlockShuffle(uint64_t* state) { static_assert(kFeistelBlocks == 16, "Expecting 16 FeistelBlocks."); constexpr size_t shuffle[kFeistelBlocks] = {7, 2, 13, 4, 11, 8, 3, 6, @@ -452,8 +441,7 @@ inline ABSL_TARGET_CRYPTO void BlockShuffle( // parallel hides the 7-cycle AESNI latency on HSW. Note that the Feistel // XORs are 'free' (included in the second AES instruction). inline ABSL_TARGET_CRYPTO const u64x2* FeistelRound( - uint64_t* ABSL_RANDOM_INTERNAL_RESTRICT state, - const u64x2* ABSL_RANDOM_INTERNAL_RESTRICT keys) { + uint64_t* state, const u64x2* ABSL_RANDOM_INTERNAL_RESTRICT keys) { static_assert(kFeistelBlocks == 16, "Expecting 16 FeistelBlocks."); // MSVC does a horrible job at unrolling loops. @@ -513,8 +501,7 @@ inline ABSL_TARGET_CRYPTO const u64x2* FeistelRound( // 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_TARGET_CRYPTO void Permute( - const void* ABSL_RANDOM_INTERNAL_RESTRICT keys, - uint64_t* ABSL_RANDOM_INTERNAL_RESTRICT state) { + const void* ABSL_RANDOM_INTERNAL_RESTRICT keys, uint64_t* state) { const u64x2* ABSL_RANDOM_INTERNAL_RESTRICT keys128 = static_cast(keys); @@ -544,10 +531,8 @@ const void* ABSL_TARGET_CRYPTO RandenHwAes::GetKeys() { // NOLINTNEXTLINE void ABSL_TARGET_CRYPTO RandenHwAes::Absorb(const void* seed_void, void* state_void) { - uint64_t* ABSL_RANDOM_INTERNAL_RESTRICT state = - reinterpret_cast(state_void); - const uint64_t* ABSL_RANDOM_INTERNAL_RESTRICT seed = - reinterpret_cast(seed_void); + auto* state = static_cast(state_void); + const auto* seed = static_cast(seed_void); constexpr size_t kCapacityBlocks = kCapacityBytes / sizeof(Vector128); constexpr size_t kStateBlocks = kStateBytes / sizeof(Vector128); @@ -623,8 +608,7 @@ 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 = - reinterpret_cast(state_void); + auto* state = static_cast(state_void); const Vector128 prev_inner = Vector128Load(state); -- cgit v1.2.3