diff options
author | misterg <misterg@google.com> | 2017-09-19T20·54-0400 |
---|---|---|
committer | misterg <misterg@google.com> | 2017-09-19T20·54-0400 |
commit | c2e754829628d1e9b7a16b3389cfdace76950fdf (patch) | |
tree | 5a7f056f44e27c30e10025113b644f0b3b5801fc /absl/base/internal/unaligned_access.h |
Initial Commit
Diffstat (limited to 'absl/base/internal/unaligned_access.h')
-rw-r--r-- | absl/base/internal/unaligned_access.h | 256 |
1 files changed, 256 insertions, 0 deletions
diff --git a/absl/base/internal/unaligned_access.h b/absl/base/internal/unaligned_access.h new file mode 100644 index 000000000000..ea30829b0e18 --- /dev/null +++ b/absl/base/internal/unaligned_access.h @@ -0,0 +1,256 @@ +// +// Copyright 2017 The Abseil Authors. +// +// Licensed under the Apache License, Version 2.0 (the "License"); +// you may not use this file except in compliance with the License. +// You may obtain a copy of the License at +// +// http://www.apache.org/licenses/LICENSE-2.0 +// +// Unless required by applicable law or agreed to in writing, software +// distributed under the License is distributed on an "AS IS" BASIS, +// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. +// See the License for the specific language governing permissions and +// limitations under the License. +// + +#ifndef ABSL_BASE_INTERNAL_UNALIGNED_ACCESS_H_ +#define ABSL_BASE_INTERNAL_UNALIGNED_ACCESS_H_ + +#include <string.h> +#include <cstdint> + +#include "absl/base/attributes.h" + +// unaligned APIs + +// Portable handling of unaligned loads, stores, and copies. +// On some platforms, like ARM, the copy functions can be more efficient +// then a load and a store. +// +// It is possible to implement all of these these using constant-length memcpy +// calls, which is portable and will usually be inlined into simple loads and +// stores if the architecture supports it. However, such inlining usually +// happens in a pass that's quite late in compilation, which means the resulting +// loads and stores cannot participate in many other optimizations, leading to +// overall worse code. + +// The unaligned API is C++ only. The declarations use C++ features +// (namespaces, inline) which are absent or incompatible in C. +#if defined(__cplusplus) + +#if defined(ADDRESS_SANITIZER) || defined(THREAD_SANITIZER) ||\ + defined(MEMORY_SANITIZER) +// Consider we have an unaligned load/store of 4 bytes from address 0x...05. +// AddressSanitizer will treat it as a 3-byte access to the range 05:07 and +// will miss a bug if 08 is the first unaddressable byte. +// ThreadSanitizer will also treat this as a 3-byte access to 05:07 and will +// miss a race between this access and some other accesses to 08. +// MemorySanitizer will correctly propagate the shadow on unaligned stores +// and correctly report bugs on unaligned loads, but it may not properly +// update and report the origin of the uninitialized memory. +// For all three tools, replacing an unaligned access with a tool-specific +// callback solves the problem. + +// Make sure uint16_t/uint32_t/uint64_t are defined. +#include <stdint.h> + +extern "C" { +uint16_t __sanitizer_unaligned_load16(const void *p); +uint32_t __sanitizer_unaligned_load32(const void *p); +uint64_t __sanitizer_unaligned_load64(const void *p); +void __sanitizer_unaligned_store16(void *p, uint16_t v); +void __sanitizer_unaligned_store32(void *p, uint32_t v); +void __sanitizer_unaligned_store64(void *p, uint64_t v); +} // extern "C" + +namespace absl { + +inline uint16_t UnalignedLoad16(const void *p) { + return __sanitizer_unaligned_load16(p); +} + +inline uint32_t UnalignedLoad32(const void *p) { + return __sanitizer_unaligned_load32(p); +} + +inline uint64_t UnalignedLoad64(const void *p) { + return __sanitizer_unaligned_load64(p); +} + +inline void UnalignedStore16(void *p, uint16_t v) { + __sanitizer_unaligned_store16(p, v); +} + +inline void UnalignedStore32(void *p, uint32_t v) { + __sanitizer_unaligned_store32(p, v); +} + +inline void UnalignedStore64(void *p, uint64_t v) { + __sanitizer_unaligned_store64(p, v); +} + +} // namespace absl + +#define ABSL_INTERNAL_UNALIGNED_LOAD16(_p) (absl::UnalignedLoad16(_p)) +#define ABSL_INTERNAL_UNALIGNED_LOAD32(_p) (absl::UnalignedLoad32(_p)) +#define ABSL_INTERNAL_UNALIGNED_LOAD64(_p) (absl::UnalignedLoad64(_p)) + +#define ABSL_INTERNAL_UNALIGNED_STORE16(_p, _val) \ + (absl::UnalignedStore16(_p, _val)) +#define ABSL_INTERNAL_UNALIGNED_STORE32(_p, _val) \ + (absl::UnalignedStore32(_p, _val)) +#define ABSL_INTERNAL_UNALIGNED_STORE64(_p, _val) \ + (absl::UnalignedStore64(_p, _val)) + +#elif defined(__x86_64__) || defined(_M_X64) || defined(__i386) || \ + defined(_M_IX86) || defined(__ppc__) || defined(__PPC__) || \ + defined(__ppc64__) || defined(__PPC64__) + +// x86 and x86-64 can perform unaligned loads/stores directly; +// modern PowerPC hardware can also do unaligned integer loads and stores; +// but note: the FPU still sends unaligned loads and stores to a trap handler! + +#define ABSL_INTERNAL_UNALIGNED_LOAD16(_p) \ + (*reinterpret_cast<const uint16_t *>(_p)) +#define ABSL_INTERNAL_UNALIGNED_LOAD32(_p) \ + (*reinterpret_cast<const uint32_t *>(_p)) +#define ABSL_INTERNAL_UNALIGNED_LOAD64(_p) \ + (*reinterpret_cast<const uint64_t *>(_p)) + +#define ABSL_INTERNAL_UNALIGNED_STORE16(_p, _val) \ + (*reinterpret_cast<uint16_t *>(_p) = (_val)) +#define ABSL_INTERNAL_UNALIGNED_STORE32(_p, _val) \ + (*reinterpret_cast<uint32_t *>(_p) = (_val)) +#define ABSL_INTERNAL_UNALIGNED_STORE64(_p, _val) \ + (*reinterpret_cast<uint64_t *>(_p) = (_val)) + +#elif defined(__arm__) && \ + !defined(__ARM_ARCH_5__) && \ + !defined(__ARM_ARCH_5T__) && \ + !defined(__ARM_ARCH_5TE__) && \ + !defined(__ARM_ARCH_5TEJ__) && \ + !defined(__ARM_ARCH_6__) && \ + !defined(__ARM_ARCH_6J__) && \ + !defined(__ARM_ARCH_6K__) && \ + !defined(__ARM_ARCH_6Z__) && \ + !defined(__ARM_ARCH_6ZK__) && \ + !defined(__ARM_ARCH_6T2__) + + +// ARMv7 and newer support native unaligned accesses, but only of 16-bit +// and 32-bit values (not 64-bit); older versions either raise a fatal signal, +// do an unaligned read and rotate the words around a bit, or do the reads very +// slowly (trip through kernel mode). There's no simple #define that says just +// “ARMv7 or higher”, so we have to filter away all ARMv5 and ARMv6 +// sub-architectures. Newer gcc (>= 4.6) set an __ARM_FEATURE_ALIGNED #define, +// so in time, maybe we can move on to that. +// +// This is a mess, but there's not much we can do about it. +// +// To further complicate matters, only LDR instructions (single reads) are +// allowed to be unaligned, not LDRD (two reads) or LDM (many reads). Unless we +// explicitly tell the compiler that these accesses can be unaligned, it can and +// will combine accesses. On armcc, the way to signal this is done by accessing +// through the type (uint32_t __packed *), but GCC has no such attribute +// (it ignores __attribute__((packed)) on individual variables). However, +// we can tell it that a _struct_ is unaligned, which has the same effect, +// so we do that. + +namespace absl { +namespace internal { + +struct Unaligned16Struct { + uint16_t value; + uint8_t dummy; // To make the size non-power-of-two. +} ABSL_ATTRIBUTE_PACKED; + +struct Unaligned32Struct { + uint32_t value; + uint8_t dummy; // To make the size non-power-of-two. +} ABSL_ATTRIBUTE_PACKED; + +} // namespace internal +} // namespace absl + +#define ABSL_INTERNAL_UNALIGNED_LOAD16(_p) \ + ((reinterpret_cast<const ::absl::internal::Unaligned16Struct *>(_p))->value) +#define ABSL_INTERNAL_UNALIGNED_LOAD32(_p) \ + ((reinterpret_cast<const ::absl::internal::Unaligned32Struct *>(_p))->value) + +#define ABSL_INTERNAL_UNALIGNED_STORE16(_p, _val) \ + ((reinterpret_cast< ::absl::internal::Unaligned16Struct *>(_p))->value = \ + (_val)) +#define ABSL_INTERNAL_UNALIGNED_STORE32(_p, _val) \ + ((reinterpret_cast< ::absl::internal::Unaligned32Struct *>(_p))->value = \ + (_val)) + +namespace absl { + +inline uint64_t UnalignedLoad64(const void *p) { + uint64_t t; + memcpy(&t, p, sizeof t); + return t; +} + +inline void UnalignedStore64(void *p, uint64_t v) { memcpy(p, &v, sizeof v); } + +} // namespace absl + +#define ABSL_INTERNAL_UNALIGNED_LOAD64(_p) (absl::UnalignedLoad64(_p)) +#define ABSL_INTERNAL_UNALIGNED_STORE64(_p, _val) \ + (absl::UnalignedStore64(_p, _val)) + +#else + +// ABSL_INTERNAL_NEED_ALIGNED_LOADS is defined when the underlying platform +// doesn't support unaligned access. +#define ABSL_INTERNAL_NEED_ALIGNED_LOADS + +// These functions are provided for architectures that don't support +// unaligned loads and stores. + +namespace absl { + +inline uint16_t UnalignedLoad16(const void *p) { + uint16_t t; + memcpy(&t, p, sizeof t); + return t; +} + +inline uint32_t UnalignedLoad32(const void *p) { + uint32_t t; + memcpy(&t, p, sizeof t); + return t; +} + +inline uint64_t UnalignedLoad64(const void *p) { + uint64_t t; + memcpy(&t, p, sizeof t); + return t; +} + +inline void UnalignedStore16(void *p, uint16_t v) { memcpy(p, &v, sizeof v); } + +inline void UnalignedStore32(void *p, uint32_t v) { memcpy(p, &v, sizeof v); } + +inline void UnalignedStore64(void *p, uint64_t v) { memcpy(p, &v, sizeof v); } + +} // namespace absl + +#define ABSL_INTERNAL_UNALIGNED_LOAD16(_p) (absl::UnalignedLoad16(_p)) +#define ABSL_INTERNAL_UNALIGNED_LOAD32(_p) (absl::UnalignedLoad32(_p)) +#define ABSL_INTERNAL_UNALIGNED_LOAD64(_p) (absl::UnalignedLoad64(_p)) + +#define ABSL_INTERNAL_UNALIGNED_STORE16(_p, _val) \ + (absl::UnalignedStore16(_p, _val)) +#define ABSL_INTERNAL_UNALIGNED_STORE32(_p, _val) \ + (absl::UnalignedStore32(_p, _val)) +#define ABSL_INTERNAL_UNALIGNED_STORE64(_p, _val) \ + (absl::UnalignedStore64(_p, _val)) + +#endif + +#endif // defined(__cplusplus), end of unaligned API + +#endif // ABSL_BASE_INTERNAL_UNALIGNED_ACCESS_H_ |