Abseil Common Libraries (C++) (grcp 依赖) https://abseil.io/
You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.

207 lines
7.7 KiB

Export of internal Abseil changes -- f012012ef78234a6a4585321b67d7b7c92ebc266 by Laramie Leavitt <lar@google.com>: Slight restructuring of absl/random/internal randen implementation. Convert round-keys.inc into randen_round_keys.cc file. Consistently use a 128-bit pointer type for internal method parameters. This allows simpler pointer arithmetic in C++ & permits removal of some constants and casts. Remove some redundancy in comments & constexpr variables. Specifically, all references to Randen algorithm parameters use RandenTraits; duplication in RandenSlow removed. PiperOrigin-RevId: 312190313 -- dc8b42e054046741e9ed65335bfdface997c6063 by Abseil Team <absl-team@google.com>: Internal change. PiperOrigin-RevId: 312167304 -- f13d248fafaf206492c1362c3574031aea3abaf7 by Matthew Brown <matthewbr@google.com>: Cleanup StrFormat extensions a little. PiperOrigin-RevId: 312166336 -- 9d9117589667afe2332bb7ad42bc967ca7c54502 by Derek Mauro <dmauro@google.com>: Internal change PiperOrigin-RevId: 312105213 -- 9a12b9b3aa0e59b8ee6cf9408ed0029045543a9b by Abseil Team <absl-team@google.com>: Complete IGNORE_TYPE macro renaming. PiperOrigin-RevId: 311999699 -- 64756f20d61021d999bd0d4c15e9ad3857382f57 by Gennadiy Rozental <rogeeff@google.com>: Switch to fixed bytes specific default value. This fixes the Abseil Flags for big endian platforms. PiperOrigin-RevId: 311844448 -- bdbe6b5b29791dbc3816ada1828458b3010ff1e9 by Laramie Leavitt <lar@google.com>: Change many distribution tests to use pcg_engine as a deterministic source of entropy. It's reasonable to test that the BitGen itself has good entropy, however when testing the cross product of all random distributions x all the architecture variations x all submitted changes results in a large number of tests. In order to account for these failures while still using good entropy requires that our allowed sigma need to account for all of these independent tests. Our current sigma values are too restrictive, and we see a lot of failures, so we have to either relax the sigma values or convert some of the statistical tests to use deterministic values. This changelist does the latter. PiperOrigin-RevId: 311840096 GitOrigin-RevId: f012012ef78234a6a4585321b67d7b7c92ebc266 Change-Id: Ic84886f38ff30d7d72c126e9b63c9a61eb729a1a
5 years ago
#ifndef ABSL_DEBUGGING_INTERNAL_STACKTRACE_AARCH64_INL_H_
#define ABSL_DEBUGGING_INTERNAL_STACKTRACE_AARCH64_INL_H_
// Generate stack tracer for aarch64
#if defined(__linux__)
#include <sys/mman.h>
#include <ucontext.h>
#include <unistd.h>
#endif
#include <atomic>
#include <cassert>
#include <cstdint>
#include <iostream>
#include "absl/base/attributes.h"
#include "absl/debugging/internal/address_is_readable.h"
#include "absl/debugging/internal/vdso_support.h" // a no-op on non-elf or non-glibc systems
#include "absl/debugging/stacktrace.h"
static const size_t kUnknownFrameSize = 0;
Export of internal Abseil changes -- f012012ef78234a6a4585321b67d7b7c92ebc266 by Laramie Leavitt <lar@google.com>: Slight restructuring of absl/random/internal randen implementation. Convert round-keys.inc into randen_round_keys.cc file. Consistently use a 128-bit pointer type for internal method parameters. This allows simpler pointer arithmetic in C++ & permits removal of some constants and casts. Remove some redundancy in comments & constexpr variables. Specifically, all references to Randen algorithm parameters use RandenTraits; duplication in RandenSlow removed. PiperOrigin-RevId: 312190313 -- dc8b42e054046741e9ed65335bfdface997c6063 by Abseil Team <absl-team@google.com>: Internal change. PiperOrigin-RevId: 312167304 -- f13d248fafaf206492c1362c3574031aea3abaf7 by Matthew Brown <matthewbr@google.com>: Cleanup StrFormat extensions a little. PiperOrigin-RevId: 312166336 -- 9d9117589667afe2332bb7ad42bc967ca7c54502 by Derek Mauro <dmauro@google.com>: Internal change PiperOrigin-RevId: 312105213 -- 9a12b9b3aa0e59b8ee6cf9408ed0029045543a9b by Abseil Team <absl-team@google.com>: Complete IGNORE_TYPE macro renaming. PiperOrigin-RevId: 311999699 -- 64756f20d61021d999bd0d4c15e9ad3857382f57 by Gennadiy Rozental <rogeeff@google.com>: Switch to fixed bytes specific default value. This fixes the Abseil Flags for big endian platforms. PiperOrigin-RevId: 311844448 -- bdbe6b5b29791dbc3816ada1828458b3010ff1e9 by Laramie Leavitt <lar@google.com>: Change many distribution tests to use pcg_engine as a deterministic source of entropy. It's reasonable to test that the BitGen itself has good entropy, however when testing the cross product of all random distributions x all the architecture variations x all submitted changes results in a large number of tests. In order to account for these failures while still using good entropy requires that our allowed sigma need to account for all of these independent tests. Our current sigma values are too restrictive, and we see a lot of failures, so we have to either relax the sigma values or convert some of the statistical tests to use deterministic values. This changelist does the latter. PiperOrigin-RevId: 311840096 GitOrigin-RevId: f012012ef78234a6a4585321b67d7b7c92ebc266 Change-Id: Ic84886f38ff30d7d72c126e9b63c9a61eb729a1a
5 years ago
#if defined(__linux__)
// Returns the address of the VDSO __kernel_rt_sigreturn function, if present.
static const unsigned char* GetKernelRtSigreturnAddress() {
constexpr uintptr_t kImpossibleAddress = 1;
ABSL_CONST_INIT static std::atomic<uintptr_t> memoized{kImpossibleAddress};
uintptr_t address = memoized.load(std::memory_order_relaxed);
if (address != kImpossibleAddress) {
return reinterpret_cast<const unsigned char*>(address);
}
address = reinterpret_cast<uintptr_t>(nullptr);
#ifdef ABSL_HAVE_VDSO_SUPPORT
absl::debugging_internal::VDSOSupport vdso;
if (vdso.IsPresent()) {
absl::debugging_internal::VDSOSupport::SymbolInfo symbol_info;
auto lookup = [&](int type) {
return vdso.LookupSymbol("__kernel_rt_sigreturn", "LINUX_2.6.39", type,
&symbol_info);
};
if ((!lookup(STT_FUNC) && !lookup(STT_NOTYPE)) ||
Export of internal Abseil changes -- f012012ef78234a6a4585321b67d7b7c92ebc266 by Laramie Leavitt <lar@google.com>: Slight restructuring of absl/random/internal randen implementation. Convert round-keys.inc into randen_round_keys.cc file. Consistently use a 128-bit pointer type for internal method parameters. This allows simpler pointer arithmetic in C++ & permits removal of some constants and casts. Remove some redundancy in comments & constexpr variables. Specifically, all references to Randen algorithm parameters use RandenTraits; duplication in RandenSlow removed. PiperOrigin-RevId: 312190313 -- dc8b42e054046741e9ed65335bfdface997c6063 by Abseil Team <absl-team@google.com>: Internal change. PiperOrigin-RevId: 312167304 -- f13d248fafaf206492c1362c3574031aea3abaf7 by Matthew Brown <matthewbr@google.com>: Cleanup StrFormat extensions a little. PiperOrigin-RevId: 312166336 -- 9d9117589667afe2332bb7ad42bc967ca7c54502 by Derek Mauro <dmauro@google.com>: Internal change PiperOrigin-RevId: 312105213 -- 9a12b9b3aa0e59b8ee6cf9408ed0029045543a9b by Abseil Team <absl-team@google.com>: Complete IGNORE_TYPE macro renaming. PiperOrigin-RevId: 311999699 -- 64756f20d61021d999bd0d4c15e9ad3857382f57 by Gennadiy Rozental <rogeeff@google.com>: Switch to fixed bytes specific default value. This fixes the Abseil Flags for big endian platforms. PiperOrigin-RevId: 311844448 -- bdbe6b5b29791dbc3816ada1828458b3010ff1e9 by Laramie Leavitt <lar@google.com>: Change many distribution tests to use pcg_engine as a deterministic source of entropy. It's reasonable to test that the BitGen itself has good entropy, however when testing the cross product of all random distributions x all the architecture variations x all submitted changes results in a large number of tests. In order to account for these failures while still using good entropy requires that our allowed sigma need to account for all of these independent tests. Our current sigma values are too restrictive, and we see a lot of failures, so we have to either relax the sigma values or convert some of the statistical tests to use deterministic values. This changelist does the latter. PiperOrigin-RevId: 311840096 GitOrigin-RevId: f012012ef78234a6a4585321b67d7b7c92ebc266 Change-Id: Ic84886f38ff30d7d72c126e9b63c9a61eb729a1a
5 years ago
symbol_info.address == nullptr) {
// Unexpected: VDSO is present, yet the expected symbol is missing
// or null.
assert(false && "VDSO is present, but doesn't have expected symbol");
} else {
if (reinterpret_cast<uintptr_t>(symbol_info.address) !=
kImpossibleAddress) {
address = reinterpret_cast<uintptr_t>(symbol_info.address);
} else {
assert(false && "VDSO returned invalid address");
}
}
}
#endif
memoized.store(address, std::memory_order_relaxed);
return reinterpret_cast<const unsigned char*>(address);
}
#endif // __linux__
// Compute the size of a stack frame in [low..high). We assume that
// low < high. Return size of kUnknownFrameSize.
template<typename T>
static inline size_t ComputeStackFrameSize(const T* low,
const T* high) {
Export of internal Abseil changes -- f012012ef78234a6a4585321b67d7b7c92ebc266 by Laramie Leavitt <lar@google.com>: Slight restructuring of absl/random/internal randen implementation. Convert round-keys.inc into randen_round_keys.cc file. Consistently use a 128-bit pointer type for internal method parameters. This allows simpler pointer arithmetic in C++ & permits removal of some constants and casts. Remove some redundancy in comments & constexpr variables. Specifically, all references to Randen algorithm parameters use RandenTraits; duplication in RandenSlow removed. PiperOrigin-RevId: 312190313 -- dc8b42e054046741e9ed65335bfdface997c6063 by Abseil Team <absl-team@google.com>: Internal change. PiperOrigin-RevId: 312167304 -- f13d248fafaf206492c1362c3574031aea3abaf7 by Matthew Brown <matthewbr@google.com>: Cleanup StrFormat extensions a little. PiperOrigin-RevId: 312166336 -- 9d9117589667afe2332bb7ad42bc967ca7c54502 by Derek Mauro <dmauro@google.com>: Internal change PiperOrigin-RevId: 312105213 -- 9a12b9b3aa0e59b8ee6cf9408ed0029045543a9b by Abseil Team <absl-team@google.com>: Complete IGNORE_TYPE macro renaming. PiperOrigin-RevId: 311999699 -- 64756f20d61021d999bd0d4c15e9ad3857382f57 by Gennadiy Rozental <rogeeff@google.com>: Switch to fixed bytes specific default value. This fixes the Abseil Flags for big endian platforms. PiperOrigin-RevId: 311844448 -- bdbe6b5b29791dbc3816ada1828458b3010ff1e9 by Laramie Leavitt <lar@google.com>: Change many distribution tests to use pcg_engine as a deterministic source of entropy. It's reasonable to test that the BitGen itself has good entropy, however when testing the cross product of all random distributions x all the architecture variations x all submitted changes results in a large number of tests. In order to account for these failures while still using good entropy requires that our allowed sigma need to account for all of these independent tests. Our current sigma values are too restrictive, and we see a lot of failures, so we have to either relax the sigma values or convert some of the statistical tests to use deterministic values. This changelist does the latter. PiperOrigin-RevId: 311840096 GitOrigin-RevId: f012012ef78234a6a4585321b67d7b7c92ebc266 Change-Id: Ic84886f38ff30d7d72c126e9b63c9a61eb729a1a
5 years ago
const char* low_char_ptr = reinterpret_cast<const char *>(low);
const char* high_char_ptr = reinterpret_cast<const char *>(high);
return low < high ? static_cast<size_t>(high_char_ptr - low_char_ptr)
: kUnknownFrameSize;
Export of internal Abseil changes -- f012012ef78234a6a4585321b67d7b7c92ebc266 by Laramie Leavitt <lar@google.com>: Slight restructuring of absl/random/internal randen implementation. Convert round-keys.inc into randen_round_keys.cc file. Consistently use a 128-bit pointer type for internal method parameters. This allows simpler pointer arithmetic in C++ & permits removal of some constants and casts. Remove some redundancy in comments & constexpr variables. Specifically, all references to Randen algorithm parameters use RandenTraits; duplication in RandenSlow removed. PiperOrigin-RevId: 312190313 -- dc8b42e054046741e9ed65335bfdface997c6063 by Abseil Team <absl-team@google.com>: Internal change. PiperOrigin-RevId: 312167304 -- f13d248fafaf206492c1362c3574031aea3abaf7 by Matthew Brown <matthewbr@google.com>: Cleanup StrFormat extensions a little. PiperOrigin-RevId: 312166336 -- 9d9117589667afe2332bb7ad42bc967ca7c54502 by Derek Mauro <dmauro@google.com>: Internal change PiperOrigin-RevId: 312105213 -- 9a12b9b3aa0e59b8ee6cf9408ed0029045543a9b by Abseil Team <absl-team@google.com>: Complete IGNORE_TYPE macro renaming. PiperOrigin-RevId: 311999699 -- 64756f20d61021d999bd0d4c15e9ad3857382f57 by Gennadiy Rozental <rogeeff@google.com>: Switch to fixed bytes specific default value. This fixes the Abseil Flags for big endian platforms. PiperOrigin-RevId: 311844448 -- bdbe6b5b29791dbc3816ada1828458b3010ff1e9 by Laramie Leavitt <lar@google.com>: Change many distribution tests to use pcg_engine as a deterministic source of entropy. It's reasonable to test that the BitGen itself has good entropy, however when testing the cross product of all random distributions x all the architecture variations x all submitted changes results in a large number of tests. In order to account for these failures while still using good entropy requires that our allowed sigma need to account for all of these independent tests. Our current sigma values are too restrictive, and we see a lot of failures, so we have to either relax the sigma values or convert some of the statistical tests to use deterministic values. This changelist does the latter. PiperOrigin-RevId: 311840096 GitOrigin-RevId: f012012ef78234a6a4585321b67d7b7c92ebc266 Change-Id: Ic84886f38ff30d7d72c126e9b63c9a61eb729a1a
5 years ago
}
// Given a pointer to a stack frame, locate and return the calling
// stackframe, or return null if no stackframe can be found. Perform sanity
// checks (the strictness of which is controlled by the boolean parameter
// "STRICT_UNWINDING") to reduce the chance that a bad pointer is returned.
template<bool STRICT_UNWINDING, bool WITH_CONTEXT>
ABSL_ATTRIBUTE_NO_SANITIZE_ADDRESS // May read random elements from stack.
ABSL_ATTRIBUTE_NO_SANITIZE_MEMORY // May read random elements from stack.
static void **NextStackFrame(void **old_frame_pointer, const void *uc) {
void **new_frame_pointer = reinterpret_cast<void**>(*old_frame_pointer);
bool check_frame_size = true;
#if defined(__linux__)
if (WITH_CONTEXT && uc != nullptr) {
// Check to see if next frame's return address is __kernel_rt_sigreturn.
if (old_frame_pointer[1] == GetKernelRtSigreturnAddress()) {
const ucontext_t *ucv = static_cast<const ucontext_t *>(uc);
// old_frame_pointer[0] is not suitable for unwinding, look at
// ucontext to discover frame pointer before signal.
void **const pre_signal_frame_pointer =
reinterpret_cast<void **>(ucv->uc_mcontext.regs[29]);
// Check that alleged frame pointer is actually readable. This is to
// prevent "double fault" in case we hit the first fault due to e.g.
// stack corruption.
if (!absl::debugging_internal::AddressIsReadable(
pre_signal_frame_pointer))
return nullptr;
// Alleged frame pointer is readable, use it for further unwinding.
new_frame_pointer = pre_signal_frame_pointer;
// Skip frame size check if we return from a signal. We may be using a
// an alternate stack for signals.
check_frame_size = false;
}
}
#endif
// The frame pointer should be 8-byte aligned.
if ((reinterpret_cast<uintptr_t>(new_frame_pointer) & 7) != 0)
Export of internal Abseil changes -- f012012ef78234a6a4585321b67d7b7c92ebc266 by Laramie Leavitt <lar@google.com>: Slight restructuring of absl/random/internal randen implementation. Convert round-keys.inc into randen_round_keys.cc file. Consistently use a 128-bit pointer type for internal method parameters. This allows simpler pointer arithmetic in C++ & permits removal of some constants and casts. Remove some redundancy in comments & constexpr variables. Specifically, all references to Randen algorithm parameters use RandenTraits; duplication in RandenSlow removed. PiperOrigin-RevId: 312190313 -- dc8b42e054046741e9ed65335bfdface997c6063 by Abseil Team <absl-team@google.com>: Internal change. PiperOrigin-RevId: 312167304 -- f13d248fafaf206492c1362c3574031aea3abaf7 by Matthew Brown <matthewbr@google.com>: Cleanup StrFormat extensions a little. PiperOrigin-RevId: 312166336 -- 9d9117589667afe2332bb7ad42bc967ca7c54502 by Derek Mauro <dmauro@google.com>: Internal change PiperOrigin-RevId: 312105213 -- 9a12b9b3aa0e59b8ee6cf9408ed0029045543a9b by Abseil Team <absl-team@google.com>: Complete IGNORE_TYPE macro renaming. PiperOrigin-RevId: 311999699 -- 64756f20d61021d999bd0d4c15e9ad3857382f57 by Gennadiy Rozental <rogeeff@google.com>: Switch to fixed bytes specific default value. This fixes the Abseil Flags for big endian platforms. PiperOrigin-RevId: 311844448 -- bdbe6b5b29791dbc3816ada1828458b3010ff1e9 by Laramie Leavitt <lar@google.com>: Change many distribution tests to use pcg_engine as a deterministic source of entropy. It's reasonable to test that the BitGen itself has good entropy, however when testing the cross product of all random distributions x all the architecture variations x all submitted changes results in a large number of tests. In order to account for these failures while still using good entropy requires that our allowed sigma need to account for all of these independent tests. Our current sigma values are too restrictive, and we see a lot of failures, so we have to either relax the sigma values or convert some of the statistical tests to use deterministic values. This changelist does the latter. PiperOrigin-RevId: 311840096 GitOrigin-RevId: f012012ef78234a6a4585321b67d7b7c92ebc266 Change-Id: Ic84886f38ff30d7d72c126e9b63c9a61eb729a1a
5 years ago
return nullptr;
// Check frame size. In strict mode, we assume frames to be under
// 100,000 bytes. In non-strict mode, we relax the limit to 1MB.
if (check_frame_size) {
const size_t max_size = STRICT_UNWINDING ? 100000 : 1000000;
const size_t frame_size =
Export of internal Abseil changes -- f012012ef78234a6a4585321b67d7b7c92ebc266 by Laramie Leavitt <lar@google.com>: Slight restructuring of absl/random/internal randen implementation. Convert round-keys.inc into randen_round_keys.cc file. Consistently use a 128-bit pointer type for internal method parameters. This allows simpler pointer arithmetic in C++ & permits removal of some constants and casts. Remove some redundancy in comments & constexpr variables. Specifically, all references to Randen algorithm parameters use RandenTraits; duplication in RandenSlow removed. PiperOrigin-RevId: 312190313 -- dc8b42e054046741e9ed65335bfdface997c6063 by Abseil Team <absl-team@google.com>: Internal change. PiperOrigin-RevId: 312167304 -- f13d248fafaf206492c1362c3574031aea3abaf7 by Matthew Brown <matthewbr@google.com>: Cleanup StrFormat extensions a little. PiperOrigin-RevId: 312166336 -- 9d9117589667afe2332bb7ad42bc967ca7c54502 by Derek Mauro <dmauro@google.com>: Internal change PiperOrigin-RevId: 312105213 -- 9a12b9b3aa0e59b8ee6cf9408ed0029045543a9b by Abseil Team <absl-team@google.com>: Complete IGNORE_TYPE macro renaming. PiperOrigin-RevId: 311999699 -- 64756f20d61021d999bd0d4c15e9ad3857382f57 by Gennadiy Rozental <rogeeff@google.com>: Switch to fixed bytes specific default value. This fixes the Abseil Flags for big endian platforms. PiperOrigin-RevId: 311844448 -- bdbe6b5b29791dbc3816ada1828458b3010ff1e9 by Laramie Leavitt <lar@google.com>: Change many distribution tests to use pcg_engine as a deterministic source of entropy. It's reasonable to test that the BitGen itself has good entropy, however when testing the cross product of all random distributions x all the architecture variations x all submitted changes results in a large number of tests. In order to account for these failures while still using good entropy requires that our allowed sigma need to account for all of these independent tests. Our current sigma values are too restrictive, and we see a lot of failures, so we have to either relax the sigma values or convert some of the statistical tests to use deterministic values. This changelist does the latter. PiperOrigin-RevId: 311840096 GitOrigin-RevId: f012012ef78234a6a4585321b67d7b7c92ebc266 Change-Id: Ic84886f38ff30d7d72c126e9b63c9a61eb729a1a
5 years ago
ComputeStackFrameSize(old_frame_pointer, new_frame_pointer);
if (frame_size == kUnknownFrameSize || frame_size > max_size)
return nullptr;
}
return new_frame_pointer;
}
template <bool IS_STACK_FRAMES, bool IS_WITH_CONTEXT>
ABSL_ATTRIBUTE_NO_SANITIZE_ADDRESS // May read random elements from stack.
ABSL_ATTRIBUTE_NO_SANITIZE_MEMORY // May read random elements from stack.
static int UnwindImpl(void** result, int* sizes, int max_depth, int skip_count,
const void *ucp, int *min_dropped_frames) {
#ifdef __GNUC__
void **frame_pointer = reinterpret_cast<void**>(__builtin_frame_address(0));
#else
# error reading stack point not yet supported on this platform.
#endif
skip_count++; // Skip the frame for this function.
int n = 0;
// The frame pointer points to low address of a frame. The first 64-bit
// word of a frame points to the next frame up the call chain, which normally
// is just after the high address of the current frame. The second word of
// a frame contains return adress of to the caller. To find a pc value
// associated with the current frame, we need to go down a level in the call
// chain. So we remember return the address of the last frame seen. This
// does not work for the first stack frame, which belongs to UnwindImp() but
// we skip the frame for UnwindImp() anyway.
void* prev_return_address = nullptr;
while (frame_pointer && n < max_depth) {
// The absl::GetStackFrames routine is called when we are in some
// informational context (the failure signal handler for example).
// Use the non-strict unwinding rules to produce a stack trace
// that is as complete as possible (even if it contains a few bogus
// entries in some rare cases).
void **next_frame_pointer =
NextStackFrame<!IS_STACK_FRAMES, IS_WITH_CONTEXT>(frame_pointer, ucp);
if (skip_count > 0) {
skip_count--;
} else {
result[n] = prev_return_address;
if (IS_STACK_FRAMES) {
sizes[n] = static_cast<int>(
ComputeStackFrameSize(frame_pointer, next_frame_pointer));
Export of internal Abseil changes -- f012012ef78234a6a4585321b67d7b7c92ebc266 by Laramie Leavitt <lar@google.com>: Slight restructuring of absl/random/internal randen implementation. Convert round-keys.inc into randen_round_keys.cc file. Consistently use a 128-bit pointer type for internal method parameters. This allows simpler pointer arithmetic in C++ & permits removal of some constants and casts. Remove some redundancy in comments & constexpr variables. Specifically, all references to Randen algorithm parameters use RandenTraits; duplication in RandenSlow removed. PiperOrigin-RevId: 312190313 -- dc8b42e054046741e9ed65335bfdface997c6063 by Abseil Team <absl-team@google.com>: Internal change. PiperOrigin-RevId: 312167304 -- f13d248fafaf206492c1362c3574031aea3abaf7 by Matthew Brown <matthewbr@google.com>: Cleanup StrFormat extensions a little. PiperOrigin-RevId: 312166336 -- 9d9117589667afe2332bb7ad42bc967ca7c54502 by Derek Mauro <dmauro@google.com>: Internal change PiperOrigin-RevId: 312105213 -- 9a12b9b3aa0e59b8ee6cf9408ed0029045543a9b by Abseil Team <absl-team@google.com>: Complete IGNORE_TYPE macro renaming. PiperOrigin-RevId: 311999699 -- 64756f20d61021d999bd0d4c15e9ad3857382f57 by Gennadiy Rozental <rogeeff@google.com>: Switch to fixed bytes specific default value. This fixes the Abseil Flags for big endian platforms. PiperOrigin-RevId: 311844448 -- bdbe6b5b29791dbc3816ada1828458b3010ff1e9 by Laramie Leavitt <lar@google.com>: Change many distribution tests to use pcg_engine as a deterministic source of entropy. It's reasonable to test that the BitGen itself has good entropy, however when testing the cross product of all random distributions x all the architecture variations x all submitted changes results in a large number of tests. In order to account for these failures while still using good entropy requires that our allowed sigma need to account for all of these independent tests. Our current sigma values are too restrictive, and we see a lot of failures, so we have to either relax the sigma values or convert some of the statistical tests to use deterministic values. This changelist does the latter. PiperOrigin-RevId: 311840096 GitOrigin-RevId: f012012ef78234a6a4585321b67d7b7c92ebc266 Change-Id: Ic84886f38ff30d7d72c126e9b63c9a61eb729a1a
5 years ago
}
n++;
}
prev_return_address = frame_pointer[1];
frame_pointer = next_frame_pointer;
}
if (min_dropped_frames != nullptr) {
// Implementation detail: we clamp the max of frames we are willing to
// count, so as not to spend too much time in the loop below.
const int kMaxUnwind = 200;
Export of internal Abseil changes -- 389189dbb322df0d0468ab13edf7dc185dc63833 by Abseil Team <absl-team@google.com>: absl str_format.h can compile with -Wconversion and -Wsign-compare PiperOrigin-RevId: 420799960 -- 762e5adc429fc143756c42fe92fe8073c87c075f by Abseil Team <absl-team@google.com>: GetStackTraceWithContext: Fix min_dropped_frames when no frames are recorded Previously, if there were still frames to skip, they would be included in min_dropped_frames. PiperOrigin-RevId: 420766341 -- 7d4374b8eaa410f4f98ec03d6a8997dccadfb271 by Abseil Team <absl-team@google.com>: absl::flags compiles with -Wconversion and -Wsign-compare PiperOrigin-RevId: 420476807 -- 5f00f7805419d725fa1ff57b388e4c0750d1d6b0 by Derek Mauro <dmauro@google.com>: Internal change PiperOrigin-RevId: 420282152 -- bd5471fc34956acf3888bf90287b2aee4415c96d by Derek Mauro <dmauro@google.com>: Internal change PiperOrigin-RevId: 420282033 -- 61c78020804e4290e9b2fe151aeaf99b198716ee by Derek Mauro <dmauro@google.com>: Internal change PiperOrigin-RevId: 420281867 -- dbe3aad24b65ea11664401a307ea3d2f28e8a7b9 by Derek Mauro <dmauro@google.com>: Remove the incorrect test for the availability of [[nodiscard]] It should have been ABSL_HAVE_CPP_ATTRIBUTE(nodiscard) instead of ABSL_HAVE_ATTRIBUTE(nodiscard). As a result, some code is not compliant with [[nodiscard]], so we cannot simply correct the availability test. Recommend that C++17-only code use the standard [[nodiscard] directly. PiperOrigin-RevId: 420150702 GitOrigin-RevId: 389189dbb322df0d0468ab13edf7dc185dc63833 Change-Id: Idf6ebae3c4edd945c9032c7db3d0ab32d16e8078
3 years ago
int num_dropped_frames = 0;
for (int j = 0; frame_pointer != nullptr && j < kMaxUnwind; j++) {
if (skip_count > 0) {
skip_count--;
} else {
num_dropped_frames++;
}
Export of internal Abseil changes -- f012012ef78234a6a4585321b67d7b7c92ebc266 by Laramie Leavitt <lar@google.com>: Slight restructuring of absl/random/internal randen implementation. Convert round-keys.inc into randen_round_keys.cc file. Consistently use a 128-bit pointer type for internal method parameters. This allows simpler pointer arithmetic in C++ & permits removal of some constants and casts. Remove some redundancy in comments & constexpr variables. Specifically, all references to Randen algorithm parameters use RandenTraits; duplication in RandenSlow removed. PiperOrigin-RevId: 312190313 -- dc8b42e054046741e9ed65335bfdface997c6063 by Abseil Team <absl-team@google.com>: Internal change. PiperOrigin-RevId: 312167304 -- f13d248fafaf206492c1362c3574031aea3abaf7 by Matthew Brown <matthewbr@google.com>: Cleanup StrFormat extensions a little. PiperOrigin-RevId: 312166336 -- 9d9117589667afe2332bb7ad42bc967ca7c54502 by Derek Mauro <dmauro@google.com>: Internal change PiperOrigin-RevId: 312105213 -- 9a12b9b3aa0e59b8ee6cf9408ed0029045543a9b by Abseil Team <absl-team@google.com>: Complete IGNORE_TYPE macro renaming. PiperOrigin-RevId: 311999699 -- 64756f20d61021d999bd0d4c15e9ad3857382f57 by Gennadiy Rozental <rogeeff@google.com>: Switch to fixed bytes specific default value. This fixes the Abseil Flags for big endian platforms. PiperOrigin-RevId: 311844448 -- bdbe6b5b29791dbc3816ada1828458b3010ff1e9 by Laramie Leavitt <lar@google.com>: Change many distribution tests to use pcg_engine as a deterministic source of entropy. It's reasonable to test that the BitGen itself has good entropy, however when testing the cross product of all random distributions x all the architecture variations x all submitted changes results in a large number of tests. In order to account for these failures while still using good entropy requires that our allowed sigma need to account for all of these independent tests. Our current sigma values are too restrictive, and we see a lot of failures, so we have to either relax the sigma values or convert some of the statistical tests to use deterministic values. This changelist does the latter. PiperOrigin-RevId: 311840096 GitOrigin-RevId: f012012ef78234a6a4585321b67d7b7c92ebc266 Change-Id: Ic84886f38ff30d7d72c126e9b63c9a61eb729a1a
5 years ago
frame_pointer =
NextStackFrame<!IS_STACK_FRAMES, IS_WITH_CONTEXT>(frame_pointer, ucp);
}
Export of internal Abseil changes -- 389189dbb322df0d0468ab13edf7dc185dc63833 by Abseil Team <absl-team@google.com>: absl str_format.h can compile with -Wconversion and -Wsign-compare PiperOrigin-RevId: 420799960 -- 762e5adc429fc143756c42fe92fe8073c87c075f by Abseil Team <absl-team@google.com>: GetStackTraceWithContext: Fix min_dropped_frames when no frames are recorded Previously, if there were still frames to skip, they would be included in min_dropped_frames. PiperOrigin-RevId: 420766341 -- 7d4374b8eaa410f4f98ec03d6a8997dccadfb271 by Abseil Team <absl-team@google.com>: absl::flags compiles with -Wconversion and -Wsign-compare PiperOrigin-RevId: 420476807 -- 5f00f7805419d725fa1ff57b388e4c0750d1d6b0 by Derek Mauro <dmauro@google.com>: Internal change PiperOrigin-RevId: 420282152 -- bd5471fc34956acf3888bf90287b2aee4415c96d by Derek Mauro <dmauro@google.com>: Internal change PiperOrigin-RevId: 420282033 -- 61c78020804e4290e9b2fe151aeaf99b198716ee by Derek Mauro <dmauro@google.com>: Internal change PiperOrigin-RevId: 420281867 -- dbe3aad24b65ea11664401a307ea3d2f28e8a7b9 by Derek Mauro <dmauro@google.com>: Remove the incorrect test for the availability of [[nodiscard]] It should have been ABSL_HAVE_CPP_ATTRIBUTE(nodiscard) instead of ABSL_HAVE_ATTRIBUTE(nodiscard). As a result, some code is not compliant with [[nodiscard]], so we cannot simply correct the availability test. Recommend that C++17-only code use the standard [[nodiscard] directly. PiperOrigin-RevId: 420150702 GitOrigin-RevId: 389189dbb322df0d0468ab13edf7dc185dc63833 Change-Id: Idf6ebae3c4edd945c9032c7db3d0ab32d16e8078
3 years ago
*min_dropped_frames = num_dropped_frames;
Export of internal Abseil changes -- f012012ef78234a6a4585321b67d7b7c92ebc266 by Laramie Leavitt <lar@google.com>: Slight restructuring of absl/random/internal randen implementation. Convert round-keys.inc into randen_round_keys.cc file. Consistently use a 128-bit pointer type for internal method parameters. This allows simpler pointer arithmetic in C++ & permits removal of some constants and casts. Remove some redundancy in comments & constexpr variables. Specifically, all references to Randen algorithm parameters use RandenTraits; duplication in RandenSlow removed. PiperOrigin-RevId: 312190313 -- dc8b42e054046741e9ed65335bfdface997c6063 by Abseil Team <absl-team@google.com>: Internal change. PiperOrigin-RevId: 312167304 -- f13d248fafaf206492c1362c3574031aea3abaf7 by Matthew Brown <matthewbr@google.com>: Cleanup StrFormat extensions a little. PiperOrigin-RevId: 312166336 -- 9d9117589667afe2332bb7ad42bc967ca7c54502 by Derek Mauro <dmauro@google.com>: Internal change PiperOrigin-RevId: 312105213 -- 9a12b9b3aa0e59b8ee6cf9408ed0029045543a9b by Abseil Team <absl-team@google.com>: Complete IGNORE_TYPE macro renaming. PiperOrigin-RevId: 311999699 -- 64756f20d61021d999bd0d4c15e9ad3857382f57 by Gennadiy Rozental <rogeeff@google.com>: Switch to fixed bytes specific default value. This fixes the Abseil Flags for big endian platforms. PiperOrigin-RevId: 311844448 -- bdbe6b5b29791dbc3816ada1828458b3010ff1e9 by Laramie Leavitt <lar@google.com>: Change many distribution tests to use pcg_engine as a deterministic source of entropy. It's reasonable to test that the BitGen itself has good entropy, however when testing the cross product of all random distributions x all the architecture variations x all submitted changes results in a large number of tests. In order to account for these failures while still using good entropy requires that our allowed sigma need to account for all of these independent tests. Our current sigma values are too restrictive, and we see a lot of failures, so we have to either relax the sigma values or convert some of the statistical tests to use deterministic values. This changelist does the latter. PiperOrigin-RevId: 311840096 GitOrigin-RevId: f012012ef78234a6a4585321b67d7b7c92ebc266 Change-Id: Ic84886f38ff30d7d72c126e9b63c9a61eb729a1a
5 years ago
}
return n;
}
namespace absl {
ABSL_NAMESPACE_BEGIN
namespace debugging_internal {
bool StackTraceWorksForTest() {
return true;
}
} // namespace debugging_internal
ABSL_NAMESPACE_END
} // namespace absl
#endif // ABSL_DEBUGGING_INTERNAL_STACKTRACE_AARCH64_INL_H_