v8/include/v8-unwinder.h

133 lines
4.6 KiB
C
Raw Normal View History

Reland "[include] Split out v8.h" This is a reland of d1b27019d3bf86360ea838c317f8505fac6d3a7e Fixes include: Adding missing file to bazel build Forward-declaring classing before friend-classing them to fix win/gcc Add missing v8-isolate.h include for vtune builds Original change's description: > [include] Split out v8.h > > This moves every single class/function out of include/v8.h into a > separate header in include/, which v8.h then includes so that > externally nothing appears to have changed. > > Every include of v8.h from inside v8 has been changed to a more > fine-grained include. > > Previously inline functions defined at the bottom of v8.h would call > private non-inline functions in the V8 class. Since that class is now > in v8-initialization.h and is rarely included (as that would create > dependency cycles), this is not possible and so those methods have been > moved out of the V8 class into the namespace v8::api_internal. > > None of the previous files in include/ now #include v8.h, which means > if embedders were relying on this transitive dependency then it will > give compile failures. > > v8-inspector.h does depend on v8-scripts.h for the time being to ensure > that Chrome continue to compile but that change will be reverted once > those transitive #includes in chrome are changed to include it directly. > > Full design: > https://docs.google.com/document/d/1rTD--I8hCAr-Rho1WTumZzFKaDpEp0IJ8ejZtk4nJdA/edit?usp=sharing > > Bug: v8:11965 > Change-Id: I53b84b29581632710edc80eb11f819c2097a2877 > Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3097448 > Reviewed-by: Yang Guo <yangguo@chromium.org> > Reviewed-by: Camillo Bruni <cbruni@chromium.org> > Reviewed-by: Jakob Kummerow <jkummerow@chromium.org> > Reviewed-by: Leszek Swirski <leszeks@chromium.org> > Reviewed-by: Michael Lippautz <mlippautz@chromium.org> > Commit-Queue: Dan Elphick <delphick@chromium.org> > Cr-Commit-Position: refs/heads/main@{#76424} Cq-Include-Trybots: luci.v8.try:v8_linux_vtunejit Bug: v8:11965 Change-Id: I99f5d3a73bf8fe25b650adfaf9567dc4e44a09e6 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3113629 Reviewed-by: Leszek Swirski <leszeks@chromium.org> Reviewed-by: Camillo Bruni <cbruni@chromium.org> Reviewed-by: Michael Lippautz <mlippautz@chromium.org> Reviewed-by: Jakob Kummerow <jkummerow@chromium.org> Reviewed-by: Simon Zünd <szuend@chromium.org> Commit-Queue: Dan Elphick <delphick@chromium.org> Cr-Commit-Position: refs/heads/main@{#76460}
2021-08-23 13:01:06 +00:00
// Copyright 2021 the V8 project authors. All rights reserved.
// Use of this source code is governed by a BSD-style license that can be
// found in the LICENSE file.
#ifndef INCLUDE_V8_UNWINDER_H_
#define INCLUDE_V8_UNWINDER_H_
#include <memory>
Reland "[profiler] Surface VM & Embedder State" This is a reland of 2d087f237eadd78f5545548675642f013fdfe675 The changes are : * Fix redundant reinterpret_cast in test file for MSVC failure https://crbug.com/v8/12476 * Fix flaky test https://crbug.com/v8/12475 If a sample is captured during a GC, no embedder context is obtained defaulting to EMPTY. This is the expected behavior, made it in clear in implementation and in test. * Synchronized the embedder context filter behavior with existing native context filter. Original change's description: > Add APIs to surface VMState and new EmbedderState to CpuProfile samples. > > EmbedderState: > * An EmbedderState is defined as a value uint8_t and a v8::context used > for filtering. > * EmbedderStates are stack allocated by the embedder, construction and > destruction set/unset the state to the isolate thread local top. > * A v8::context is used to filter states that are added to a CpuProfile, > if the CpuProfile do not have a ContextFilter set or if contexts do not > match, state defaults to Empty. > > * v8:StateTag is already propagated all the way to a Sample, simply add > an API to surface it. > > VMState: > Change-Id: I7eed08907360b99b0ad20ddcff59c95c7076c85e > Bug: chromium:1263871 > Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3188072 > Auto-Submit: Corentin Pescheloche <cpescheloche@fb.com> > Reviewed-by: Camillo Bruni <cbruni@chromium.org> > Reviewed-by: Dominik Inführ <dinfuehr@chromium.org> > Reviewed-by: Igor Sheludko <ishell@chromium.org> > Commit-Queue: Camillo Bruni <cbruni@chromium.org> > Cr-Commit-Position: refs/heads/main@{#78250} Bug: chromium:1263871 Change-Id: Ief891b05da99c695e9fb70f94ed7ebdecc6c3b7b Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3320037 Auto-Submit: Corentin Pescheloche <cpescheloche@fb.com> Reviewed-by: Camillo Bruni <cbruni@chromium.org> Reviewed-by: Igor Sheludko <ishell@chromium.org> Reviewed-by: Dominik Inführ <dinfuehr@chromium.org> Commit-Queue: Dominik Inführ <dinfuehr@chromium.org> Cr-Commit-Position: refs/heads/main@{#78281}
2021-12-07 07:28:08 +00:00
#include "v8-embedder-state-scope.h" // NOLINT(build/include_directory)
#include "v8config.h" // NOLINT(build/include_directory)
Reland "[include] Split out v8.h" This is a reland of d1b27019d3bf86360ea838c317f8505fac6d3a7e Fixes include: Adding missing file to bazel build Forward-declaring classing before friend-classing them to fix win/gcc Add missing v8-isolate.h include for vtune builds Original change's description: > [include] Split out v8.h > > This moves every single class/function out of include/v8.h into a > separate header in include/, which v8.h then includes so that > externally nothing appears to have changed. > > Every include of v8.h from inside v8 has been changed to a more > fine-grained include. > > Previously inline functions defined at the bottom of v8.h would call > private non-inline functions in the V8 class. Since that class is now > in v8-initialization.h and is rarely included (as that would create > dependency cycles), this is not possible and so those methods have been > moved out of the V8 class into the namespace v8::api_internal. > > None of the previous files in include/ now #include v8.h, which means > if embedders were relying on this transitive dependency then it will > give compile failures. > > v8-inspector.h does depend on v8-scripts.h for the time being to ensure > that Chrome continue to compile but that change will be reverted once > those transitive #includes in chrome are changed to include it directly. > > Full design: > https://docs.google.com/document/d/1rTD--I8hCAr-Rho1WTumZzFKaDpEp0IJ8ejZtk4nJdA/edit?usp=sharing > > Bug: v8:11965 > Change-Id: I53b84b29581632710edc80eb11f819c2097a2877 > Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3097448 > Reviewed-by: Yang Guo <yangguo@chromium.org> > Reviewed-by: Camillo Bruni <cbruni@chromium.org> > Reviewed-by: Jakob Kummerow <jkummerow@chromium.org> > Reviewed-by: Leszek Swirski <leszeks@chromium.org> > Reviewed-by: Michael Lippautz <mlippautz@chromium.org> > Commit-Queue: Dan Elphick <delphick@chromium.org> > Cr-Commit-Position: refs/heads/main@{#76424} Cq-Include-Trybots: luci.v8.try:v8_linux_vtunejit Bug: v8:11965 Change-Id: I99f5d3a73bf8fe25b650adfaf9567dc4e44a09e6 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3113629 Reviewed-by: Leszek Swirski <leszeks@chromium.org> Reviewed-by: Camillo Bruni <cbruni@chromium.org> Reviewed-by: Michael Lippautz <mlippautz@chromium.org> Reviewed-by: Jakob Kummerow <jkummerow@chromium.org> Reviewed-by: Simon Zünd <szuend@chromium.org> Commit-Queue: Dan Elphick <delphick@chromium.org> Cr-Commit-Position: refs/heads/main@{#76460}
2021-08-23 13:01:06 +00:00
namespace v8 {
// Holds the callee saved registers needed for the stack unwinder. It is the
// empty struct if no registers are required. Implemented in
// include/v8-unwinder-state.h.
struct CalleeSavedRegisters;
// A RegisterState represents the current state of registers used
// by the sampling profiler API.
struct V8_EXPORT RegisterState {
RegisterState();
~RegisterState();
RegisterState(const RegisterState& other);
RegisterState& operator=(const RegisterState& other);
void* pc; // Instruction pointer.
void* sp; // Stack pointer.
void* fp; // Frame pointer.
void* lr; // Link register (or nullptr on platforms without a link register).
// Callee saved registers (or null if no callee saved registers were stored)
std::unique_ptr<CalleeSavedRegisters> callee_saved;
};
// A StateTag represents a possible state of the VM.
Reland "[profiler] Surface VM & Embedder State" This is a reland of 2d087f237eadd78f5545548675642f013fdfe675 The changes are : * Fix redundant reinterpret_cast in test file for MSVC failure https://crbug.com/v8/12476 * Fix flaky test https://crbug.com/v8/12475 If a sample is captured during a GC, no embedder context is obtained defaulting to EMPTY. This is the expected behavior, made it in clear in implementation and in test. * Synchronized the embedder context filter behavior with existing native context filter. Original change's description: > Add APIs to surface VMState and new EmbedderState to CpuProfile samples. > > EmbedderState: > * An EmbedderState is defined as a value uint8_t and a v8::context used > for filtering. > * EmbedderStates are stack allocated by the embedder, construction and > destruction set/unset the state to the isolate thread local top. > * A v8::context is used to filter states that are added to a CpuProfile, > if the CpuProfile do not have a ContextFilter set or if contexts do not > match, state defaults to Empty. > > * v8:StateTag is already propagated all the way to a Sample, simply add > an API to surface it. > > VMState: > Change-Id: I7eed08907360b99b0ad20ddcff59c95c7076c85e > Bug: chromium:1263871 > Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3188072 > Auto-Submit: Corentin Pescheloche <cpescheloche@fb.com> > Reviewed-by: Camillo Bruni <cbruni@chromium.org> > Reviewed-by: Dominik Inführ <dinfuehr@chromium.org> > Reviewed-by: Igor Sheludko <ishell@chromium.org> > Commit-Queue: Camillo Bruni <cbruni@chromium.org> > Cr-Commit-Position: refs/heads/main@{#78250} Bug: chromium:1263871 Change-Id: Ief891b05da99c695e9fb70f94ed7ebdecc6c3b7b Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3320037 Auto-Submit: Corentin Pescheloche <cpescheloche@fb.com> Reviewed-by: Camillo Bruni <cbruni@chromium.org> Reviewed-by: Igor Sheludko <ishell@chromium.org> Reviewed-by: Dominik Inführ <dinfuehr@chromium.org> Commit-Queue: Dominik Inführ <dinfuehr@chromium.org> Cr-Commit-Position: refs/heads/main@{#78281}
2021-12-07 07:28:08 +00:00
enum StateTag : int {
Reland "[include] Split out v8.h" This is a reland of d1b27019d3bf86360ea838c317f8505fac6d3a7e Fixes include: Adding missing file to bazel build Forward-declaring classing before friend-classing them to fix win/gcc Add missing v8-isolate.h include for vtune builds Original change's description: > [include] Split out v8.h > > This moves every single class/function out of include/v8.h into a > separate header in include/, which v8.h then includes so that > externally nothing appears to have changed. > > Every include of v8.h from inside v8 has been changed to a more > fine-grained include. > > Previously inline functions defined at the bottom of v8.h would call > private non-inline functions in the V8 class. Since that class is now > in v8-initialization.h and is rarely included (as that would create > dependency cycles), this is not possible and so those methods have been > moved out of the V8 class into the namespace v8::api_internal. > > None of the previous files in include/ now #include v8.h, which means > if embedders were relying on this transitive dependency then it will > give compile failures. > > v8-inspector.h does depend on v8-scripts.h for the time being to ensure > that Chrome continue to compile but that change will be reverted once > those transitive #includes in chrome are changed to include it directly. > > Full design: > https://docs.google.com/document/d/1rTD--I8hCAr-Rho1WTumZzFKaDpEp0IJ8ejZtk4nJdA/edit?usp=sharing > > Bug: v8:11965 > Change-Id: I53b84b29581632710edc80eb11f819c2097a2877 > Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3097448 > Reviewed-by: Yang Guo <yangguo@chromium.org> > Reviewed-by: Camillo Bruni <cbruni@chromium.org> > Reviewed-by: Jakob Kummerow <jkummerow@chromium.org> > Reviewed-by: Leszek Swirski <leszeks@chromium.org> > Reviewed-by: Michael Lippautz <mlippautz@chromium.org> > Commit-Queue: Dan Elphick <delphick@chromium.org> > Cr-Commit-Position: refs/heads/main@{#76424} Cq-Include-Trybots: luci.v8.try:v8_linux_vtunejit Bug: v8:11965 Change-Id: I99f5d3a73bf8fe25b650adfaf9567dc4e44a09e6 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3113629 Reviewed-by: Leszek Swirski <leszeks@chromium.org> Reviewed-by: Camillo Bruni <cbruni@chromium.org> Reviewed-by: Michael Lippautz <mlippautz@chromium.org> Reviewed-by: Jakob Kummerow <jkummerow@chromium.org> Reviewed-by: Simon Zünd <szuend@chromium.org> Commit-Queue: Dan Elphick <delphick@chromium.org> Cr-Commit-Position: refs/heads/main@{#76460}
2021-08-23 13:01:06 +00:00
JS,
GC,
PARSER,
BYTECODE_COMPILER,
COMPILER,
OTHER,
EXTERNAL,
ATOMICS_WAIT,
IDLE
};
// The output structure filled up by GetStackSample API function.
struct SampleInfo {
Reland "[profiler] Surface VM & Embedder State" This is a reland of 2d087f237eadd78f5545548675642f013fdfe675 The changes are : * Fix redundant reinterpret_cast in test file for MSVC failure https://crbug.com/v8/12476 * Fix flaky test https://crbug.com/v8/12475 If a sample is captured during a GC, no embedder context is obtained defaulting to EMPTY. This is the expected behavior, made it in clear in implementation and in test. * Synchronized the embedder context filter behavior with existing native context filter. Original change's description: > Add APIs to surface VMState and new EmbedderState to CpuProfile samples. > > EmbedderState: > * An EmbedderState is defined as a value uint8_t and a v8::context used > for filtering. > * EmbedderStates are stack allocated by the embedder, construction and > destruction set/unset the state to the isolate thread local top. > * A v8::context is used to filter states that are added to a CpuProfile, > if the CpuProfile do not have a ContextFilter set or if contexts do not > match, state defaults to Empty. > > * v8:StateTag is already propagated all the way to a Sample, simply add > an API to surface it. > > VMState: > Change-Id: I7eed08907360b99b0ad20ddcff59c95c7076c85e > Bug: chromium:1263871 > Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3188072 > Auto-Submit: Corentin Pescheloche <cpescheloche@fb.com> > Reviewed-by: Camillo Bruni <cbruni@chromium.org> > Reviewed-by: Dominik Inführ <dinfuehr@chromium.org> > Reviewed-by: Igor Sheludko <ishell@chromium.org> > Commit-Queue: Camillo Bruni <cbruni@chromium.org> > Cr-Commit-Position: refs/heads/main@{#78250} Bug: chromium:1263871 Change-Id: Ief891b05da99c695e9fb70f94ed7ebdecc6c3b7b Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3320037 Auto-Submit: Corentin Pescheloche <cpescheloche@fb.com> Reviewed-by: Camillo Bruni <cbruni@chromium.org> Reviewed-by: Igor Sheludko <ishell@chromium.org> Reviewed-by: Dominik Inführ <dinfuehr@chromium.org> Commit-Queue: Dominik Inführ <dinfuehr@chromium.org> Cr-Commit-Position: refs/heads/main@{#78281}
2021-12-07 07:28:08 +00:00
size_t frames_count; // Number of frames collected.
void* external_callback_entry; // External callback address if VM is
// executing an external callback.
void* context; // Incumbent native context address.
void* embedder_context; // Native context address for embedder state
StateTag vm_state; // Current VM state.
EmbedderStateTag embedder_state; // Current Embedder state
Reland "[include] Split out v8.h" This is a reland of d1b27019d3bf86360ea838c317f8505fac6d3a7e Fixes include: Adding missing file to bazel build Forward-declaring classing before friend-classing them to fix win/gcc Add missing v8-isolate.h include for vtune builds Original change's description: > [include] Split out v8.h > > This moves every single class/function out of include/v8.h into a > separate header in include/, which v8.h then includes so that > externally nothing appears to have changed. > > Every include of v8.h from inside v8 has been changed to a more > fine-grained include. > > Previously inline functions defined at the bottom of v8.h would call > private non-inline functions in the V8 class. Since that class is now > in v8-initialization.h and is rarely included (as that would create > dependency cycles), this is not possible and so those methods have been > moved out of the V8 class into the namespace v8::api_internal. > > None of the previous files in include/ now #include v8.h, which means > if embedders were relying on this transitive dependency then it will > give compile failures. > > v8-inspector.h does depend on v8-scripts.h for the time being to ensure > that Chrome continue to compile but that change will be reverted once > those transitive #includes in chrome are changed to include it directly. > > Full design: > https://docs.google.com/document/d/1rTD--I8hCAr-Rho1WTumZzFKaDpEp0IJ8ejZtk4nJdA/edit?usp=sharing > > Bug: v8:11965 > Change-Id: I53b84b29581632710edc80eb11f819c2097a2877 > Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3097448 > Reviewed-by: Yang Guo <yangguo@chromium.org> > Reviewed-by: Camillo Bruni <cbruni@chromium.org> > Reviewed-by: Jakob Kummerow <jkummerow@chromium.org> > Reviewed-by: Leszek Swirski <leszeks@chromium.org> > Reviewed-by: Michael Lippautz <mlippautz@chromium.org> > Commit-Queue: Dan Elphick <delphick@chromium.org> > Cr-Commit-Position: refs/heads/main@{#76424} Cq-Include-Trybots: luci.v8.try:v8_linux_vtunejit Bug: v8:11965 Change-Id: I99f5d3a73bf8fe25b650adfaf9567dc4e44a09e6 Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3113629 Reviewed-by: Leszek Swirski <leszeks@chromium.org> Reviewed-by: Camillo Bruni <cbruni@chromium.org> Reviewed-by: Michael Lippautz <mlippautz@chromium.org> Reviewed-by: Jakob Kummerow <jkummerow@chromium.org> Reviewed-by: Simon Zünd <szuend@chromium.org> Commit-Queue: Dan Elphick <delphick@chromium.org> Cr-Commit-Position: refs/heads/main@{#76460}
2021-08-23 13:01:06 +00:00
};
struct MemoryRange {
const void* start = nullptr;
size_t length_in_bytes = 0;
};
struct JSEntryStub {
MemoryRange code;
};
struct JSEntryStubs {
JSEntryStub js_entry_stub;
JSEntryStub js_construct_entry_stub;
JSEntryStub js_run_microtasks_entry_stub;
};
/**
* Various helpers for skipping over V8 frames in a given stack.
*
* The unwinder API is only supported on the x64, ARM64 and ARM32 architectures.
*/
class V8_EXPORT Unwinder {
public:
/**
* Attempt to unwind the stack to the most recent C++ frame. This function is
* signal-safe and does not access any V8 state and thus doesn't require an
* Isolate.
*
* The unwinder needs to know the location of the JS Entry Stub (a piece of
* code that is run when C++ code calls into generated JS code). This is used
* for edge cases where the current frame is being constructed or torn down
* when the stack sample occurs.
*
* The unwinder also needs the virtual memory range of all possible V8 code
* objects. There are two ranges required - the heap code range and the range
* for code embedded in the binary.
*
* Available on x64, ARM64 and ARM32.
*
* \param code_pages A list of all of the ranges in which V8 has allocated
* executable code. The caller should obtain this list by calling
* Isolate::CopyCodePages() during the same interrupt/thread suspension that
* captures the stack.
* \param register_state The current registers. This is an in-out param that
* will be overwritten with the register values after unwinding, on success.
* \param stack_base The resulting stack pointer and frame pointer values are
* bounds-checked against the stack_base and the original stack pointer value
* to ensure that they are valid locations in the given stack. If these values
* or any intermediate frame pointer values used during unwinding are ever out
* of these bounds, unwinding will fail.
*
* \return True on success.
*/
static bool TryUnwindV8Frames(const JSEntryStubs& entry_stubs,
size_t code_pages_length,
const MemoryRange* code_pages,
RegisterState* register_state,
const void* stack_base);
/**
* Whether the PC is within the V8 code range represented by code_pages.
*
* If this returns false, then calling UnwindV8Frames() with the same PC
* and unwind_state will always fail. If it returns true, then unwinding may
* (but not necessarily) be successful.
*
* Available on x64, ARM64 and ARM32
*/
static bool PCIsInV8(size_t code_pages_length, const MemoryRange* code_pages,
void* pc);
};
} // namespace v8
#endif // INCLUDE_V8_UNWINDER_H_