Statistical stack profiler for Windows x64

Provides the collection part of a client stack collection pipeline for
Win x64. This is intended to be used initially to provide real-world
stack trace data to UMA to facilitate implementation of server-side
processing. This CL has just the profiler implementation. Instantiation
and use of the profiler in Chrome, and handling by UMA, will be
addressed in follow-on CLs.

BUG=464929

Review URL: https://codereview.chromium.org/1016563004

Cr-Commit-Position: refs/heads/master@{#321658}

8 files changed, 1140 insertions, 0 deletions

diff --git a/base/BUILD.gn b/base/BUILD.gn
index b95ff72..8b9b5bf 100644
--- a/base/BUILD.gn
+++ b/base/BUILD.gn
@@ -370,6 +370,10 @@ component("base") {
     "profiler/scoped_profile.h",
     "profiler/scoped_tracker.cc",
     "profiler/scoped_tracker.h",
+    "profiler/stack_sampling_profiler.cc",
+    "profiler/stack_sampling_profiler.h",
+    "profiler/stack_sampling_profiler_posix.cc",
+    "profiler/stack_sampling_profiler_win.cc",
     "profiler/tracked_time.cc",
     "profiler/tracked_time.h",
     "rand_util.cc",
@@ -1191,6 +1195,7 @@ test("base_unittests") {
     "process/process_unittest.cc",
     "process/process_util_unittest.cc",
     "process/process_util_unittest_ios.cc",
+    "profiler/stack_sampling_profiler_unittest.cc",
     "profiler/tracked_time_unittest.cc",
     "rand_util_unittest.cc",
     "scoped_clear_errno_unittest.cc",
diff --git a/base/base.gyp b/base/base.gyp
index 6f8102c..ac37a89 100644
--- a/base/base.gyp
+++ b/base/base.gyp
@@ -595,6 +595,7 @@
         'process/process_metrics_unittest_ios.cc',
         'process/process_unittest.cc',
         'process/process_util_unittest.cc',
+        'profiler/stack_sampling_profiler_unittest.cc',
         'profiler/tracked_time_unittest.cc',
         'rand_util_unittest.cc',
         'scoped_clear_errno_unittest.cc',
diff --git a/base/base.gypi b/base/base.gypi
index 40c7d35..825bca1 100644
--- a/base/base.gypi
+++ b/base/base.gypi
@@ -488,6 +488,10 @@
           'profiler/scoped_profile.h',
           'profiler/scoped_tracker.cc',
           'profiler/scoped_tracker.h',
+          'profiler/stack_sampling_profiler.cc',
+          'profiler/stack_sampling_profiler.h',
+          'profiler/stack_sampling_profiler_posix.cc',
+          'profiler/stack_sampling_profiler_win.cc',
           'profiler/tracked_time.cc',
           'profiler/tracked_time.h',
           'rand_util.cc',
diff --git a/base/profiler/stack_sampling_profiler.cc b/base/profiler/stack_sampling_profiler.cc
new file mode 100644
index 0000000..57b7b35
--- /dev/null
+++ b/base/profiler/stack_sampling_profiler.cc
@@ -0,0 +1,261 @@
+// Copyright 2015 The Chromium Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style license that can be
+// found in the LICENSE file.
+
+#include "base/profiler/stack_sampling_profiler.h"
+
+#include <algorithm>
+
+#include "base/bind.h"
+#include "base/callback.h"
+#include "base/memory/singleton.h"
+#include "base/synchronization/lock.h"
+#include "base/synchronization/waitable_event.h"
+#include "base/timer/elapsed_timer.h"
+
+template <typename T> struct DefaultSingletonTraits;
+
+namespace base {
+
+namespace {
+
+// Thread-safe singleton class that stores collected profiles waiting to be
+// processed.
+class PendingProfiles {
+ public:
+  PendingProfiles();
+  ~PendingProfiles();
+
+  static PendingProfiles* GetInstance();
+
+  // Appends |profiles|. This function is thread safe.
+  void PutProfiles(const std::vector<StackSamplingProfiler::Profile>& profiles);
+  // Gets the pending profiles into *|profiles|. This function is thread safe.
+  void GetProfiles(std::vector<StackSamplingProfiler::Profile>* profiles);
+
+ private:
+  Lock profiles_lock_;
+  std::vector<StackSamplingProfiler::Profile> profiles_;
+
+  DISALLOW_COPY_AND_ASSIGN(PendingProfiles);
+};
+
+PendingProfiles::PendingProfiles() {}
+
+PendingProfiles::~PendingProfiles() {}
+
+// static
+PendingProfiles* PendingProfiles::GetInstance() {
+  return Singleton<PendingProfiles>::get();
+}
+
+void PendingProfiles::PutProfiles(
+    const std::vector<StackSamplingProfiler::Profile>& profiles) {
+  AutoLock scoped_lock(profiles_lock_);
+  profiles_.insert(profiles_.end(), profiles.begin(), profiles.end());
+}
+
+void PendingProfiles::GetProfiles(
+    std::vector<StackSamplingProfiler::Profile>* profiles) {
+  profiles->clear();
+
+  AutoLock scoped_lock(profiles_lock_);
+  profiles_.swap(*profiles);
+}
+}  // namespace
+
+StackSamplingProfiler::Module::Module() : base_address(nullptr) {}
+
+StackSamplingProfiler::Module::~Module() {}
+
+StackSamplingProfiler::Frame::Frame()
+    : instruction_pointer(nullptr),
+      module_index(-1) {}
+
+StackSamplingProfiler::Frame::~Frame() {}
+
+StackSamplingProfiler::Profile::Profile() : preserve_sample_ordering(false) {}
+
+StackSamplingProfiler::Profile::~Profile() {}
+
+class StackSamplingProfiler::SamplingThread : public PlatformThread::Delegate {
+ public:
+  // Samples stacks using |native_sampler|. When complete, invokes
+  // |profiles_callback| with the collected profiles. |profiles_callback| must
+  // be thread-safe and may consume the contents of the vector.
+  SamplingThread(
+      scoped_ptr<NativeStackSampler> native_sampler,
+      const SamplingParams& params,
+      Callback<void(const std::vector<Profile>&)> completed_callback);
+  ~SamplingThread() override;
+
+  // Implementation of PlatformThread::Delegate:
+  void ThreadMain() override;
+
+  void Stop();
+
+ private:
+  // Collects a profile from a single burst. Returns true if the profile was
+  // collected, or false if collection was stopped before it completed.
+  bool CollectProfile(Profile* profile, TimeDelta* elapsed_time);
+  // Collects profiles from all bursts, or until the sampling is stopped. If
+  // stopped before complete, |profiles| will contains only full bursts.
+  void CollectProfiles(std::vector<Profile>* profiles);
+
+  scoped_ptr<NativeStackSampler> native_sampler_;
+
+  const SamplingParams params_;
+
+  WaitableEvent stop_event_;
+
+  Callback<void(const std::vector<Profile>&)> completed_callback_;
+
+  DISALLOW_COPY_AND_ASSIGN(SamplingThread);
+};
+
+StackSamplingProfiler::SamplingThread::SamplingThread(
+    scoped_ptr<NativeStackSampler> native_sampler,
+    const SamplingParams& params,
+    Callback<void(const std::vector<Profile>&)> completed_callback)
+    : native_sampler_(native_sampler.Pass()),
+      params_(params),
+      stop_event_(false, false),
+      completed_callback_(completed_callback) {
+}
+
+StackSamplingProfiler::SamplingThread::~SamplingThread() {}
+
+void StackSamplingProfiler::SamplingThread::ThreadMain() {
+  PlatformThread::SetName("Chrome_SamplingProfilerThread");
+
+  std::vector<Profile> profiles;
+  CollectProfiles(&profiles);
+  completed_callback_.Run(profiles);
+}
+
+bool StackSamplingProfiler::SamplingThread::CollectProfile(
+    Profile* profile,
+    TimeDelta* elapsed_time) {
+  ElapsedTimer profile_timer;
+  Profile current_profile;
+  native_sampler_->ProfileRecordingStarting(&current_profile);
+  current_profile.sampling_period = params_.sampling_interval;
+  bool stopped_early = false;
+  for (int i = 0; i < params_.samples_per_burst; ++i) {
+    ElapsedTimer sample_timer;
+    current_profile.samples.push_back(Sample());
+    native_sampler_->RecordStackSample(&current_profile.samples.back());
+    TimeDelta elapsed_sample_time = sample_timer.Elapsed();
+    if (i != params_.samples_per_burst - 1) {
+      if (stop_event_.TimedWait(
+              std::max(params_.sampling_interval - elapsed_sample_time,
+                       TimeDelta()))) {
+        stopped_early = true;
+        break;
+      }
+    }
+  }
+
+  *elapsed_time = profile_timer.Elapsed();
+  current_profile.profile_duration = *elapsed_time;
+  native_sampler_->ProfileRecordingStopped();
+
+  if (!stopped_early)
+    *profile = current_profile;
+
+  return !stopped_early;
+}
+
+void StackSamplingProfiler::SamplingThread::CollectProfiles(
+    std::vector<Profile>* profiles) {
+  if (stop_event_.TimedWait(params_.initial_delay))
+    return;
+
+  for (int i = 0; i < params_.bursts; ++i) {
+    Profile profile;
+    TimeDelta elapsed_profile_time;
+    if (CollectProfile(&profile, &elapsed_profile_time))
+      profiles->push_back(profile);
+    else
+      return;
+
+    if (stop_event_.TimedWait(
+            std::max(params_.burst_interval - elapsed_profile_time,
+                     TimeDelta())))
+      return;
+  }
+}
+
+void StackSamplingProfiler::SamplingThread::Stop() {
+  stop_event_.Signal();
+}
+
+void StackSamplingProfiler::SamplingThreadDeleter::operator()(
+    SamplingThread* thread) const {
+  delete thread;
+}
+
+StackSamplingProfiler::NativeStackSampler::NativeStackSampler() {}
+
+StackSamplingProfiler::NativeStackSampler::~NativeStackSampler() {}
+
+StackSamplingProfiler::SamplingParams::SamplingParams()
+    : initial_delay(TimeDelta::FromMilliseconds(0)),
+      bursts(1),
+      burst_interval(TimeDelta::FromMilliseconds(10000)),
+      samples_per_burst(300),
+      sampling_interval(TimeDelta::FromMilliseconds(100)),
+      preserve_sample_ordering(false) {
+}
+
+StackSamplingProfiler::StackSamplingProfiler(PlatformThreadId thread_id,
+                                             const SamplingParams& params)
+    : thread_id_(thread_id), params_(params) {}
+
+StackSamplingProfiler::~StackSamplingProfiler() {}
+
+void StackSamplingProfiler::Start() {
+  native_sampler_ = NativeStackSampler::Create(thread_id_);
+  if (!native_sampler_)
+    return;
+
+  sampling_thread_.reset(
+      new SamplingThread(
+          native_sampler_.Pass(), params_,
+          (custom_completed_callback_.is_null() ?
+           Bind(&PendingProfiles::PutProfiles,
+                Unretained(PendingProfiles::GetInstance())) :
+           custom_completed_callback_)));
+  if (!PlatformThread::CreateNonJoinable(0, sampling_thread_.get()))
+    LOG(ERROR) << "failed to create thread";
+}
+
+void StackSamplingProfiler::Stop() {
+  if (sampling_thread_)
+    sampling_thread_->Stop();
+}
+
+// static
+void StackSamplingProfiler::GetPendingProfiles(std::vector<Profile>* profiles) {
+  PendingProfiles::GetInstance()->GetProfiles(profiles);
+}
+
+void StackSamplingProfiler::SetCustomCompletedCallback(
+    Callback<void(const std::vector<Profile>&)> callback) {
+  custom_completed_callback_ = callback;
+}
+
+bool operator==(const StackSamplingProfiler::Frame &a,
+                const StackSamplingProfiler::Frame &b) {
+  return a.instruction_pointer == b.instruction_pointer &&
+      a.module_index == b.module_index;
+}
+
+bool operator<(const StackSamplingProfiler::Frame &a,
+               const StackSamplingProfiler::Frame &b) {
+  return (a.module_index < b.module_index) ||
+      (a.module_index == b.module_index &&
+       a.instruction_pointer < b.instruction_pointer);
+}
+
+}  // namespace base
diff --git a/base/profiler/stack_sampling_profiler.h b/base/profiler/stack_sampling_profiler.h
new file mode 100644
index 0000000..60faa51
--- /dev/null
+++ b/base/profiler/stack_sampling_profiler.h
@@ -0,0 +1,206 @@
+// Copyright 2015 The Chromium 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 BASE_PROFILER_STACK_SAMPLING_PROFILER_H_
+#define BASE_PROFILER_STACK_SAMPLING_PROFILER_H_
+
+#include <string>
+#include <vector>
+
+#include "base/base_export.h"
+#include "base/callback.h"
+#include "base/files/file_path.h"
+#include "base/memory/scoped_ptr.h"
+#include "base/strings/string16.h"
+#include "base/threading/platform_thread.h"
+#include "base/time/time.h"
+
+namespace base {
+
+// StackSamplingProfiler periodically stops a thread to sample its stack, for
+// the purpose of collecting information about which code paths are
+// executing. This information is used in aggregate by UMA to identify hot
+// and/or janky code paths.
+//
+// Sample StackStackSamplingProfiler usage:
+//
+//   // Create and customize params as desired.
+//   base::StackStackSamplingProfiler::SamplingParams params;
+//   // Any thread's ID may be passed as the target.
+//   base::StackSamplingProfiler profiler(base::PlatformThread::CurrentId()),
+//       params);
+//
+//   // To process the profiles within Chrome rather than via UMA, set a custom
+//   // completed callback:
+//   base::Callback<void(const std::vector<Profile>&)>
+//       thread_safe_callback = ...;
+//   profiler.SetCustomCompletedCallback(thread_safe_callback);
+//
+//   profiler.Start();
+//   // ... work being done on the target thread here ...
+//   profiler.Stop();  // optional, stops collection before complete per params
+//
+// When all profiles are complete or the profiler is stopped, if the custom
+// completed callback was set it will be called from the profiler thread with
+// the completed profiles. If no callback was set, the profiles are stored
+// internally and retrieved for UMA through
+// GetPendingProfiles(). GetPendingProfiles() should never be called by other
+// code; to retrieve profiles for in-process processing, set a completed
+// callback.
+class BASE_EXPORT StackSamplingProfiler {
+ public:
+  // Module represents the module (DLL or exe) corresponding to a stack frame.
+  struct Module {
+    Module();
+    ~Module();
+
+    // Points to the base address of the module.
+    const void* base_address;
+    // An opaque binary string that uniquely identifies a particular program
+    // version with high probability. This is parsed from headers of the loaded
+    // module.
+    // For binaries generated by GNU tools:
+    //   Contents of the .note.gnu.build-id field.
+    // On Windows:
+    //   GUID + AGE in the debug image headers of a module.
+    std::string id;
+    // The filename of the module.
+    FilePath filename;
+  };
+
+  // Frame represents an individual sampled stack frame with module information.
+  struct Frame {
+    Frame();
+    ~Frame();
+
+    // The sampled instruction pointer within the function.
+    const void* instruction_pointer;
+    // Index of the module in the array of modules. We don't represent module
+    // state directly here to save space.
+    int module_index;
+  };
+
+  // Sample represents a set of stack frames.
+  using Sample = std::vector<Frame>;
+
+  // Profile represents a set of samples.
+  struct BASE_EXPORT Profile {
+    Profile();
+    ~Profile();
+
+    std::vector<Module> modules;
+    std::vector<Sample> samples;
+    // Duration of this profile.
+    TimeDelta profile_duration;
+    // Time between samples.
+    TimeDelta sampling_period;
+    // True if sample ordering is important and should be preserved if and when
+    // this profile is compressed and processed.
+    bool preserve_sample_ordering;
+  };
+
+  // NativeStackSampler abstracts the native implementation required to record a
+  // stack sample for a given thread.
+  class NativeStackSampler {
+   public:
+    virtual ~NativeStackSampler();
+
+    // Create a stack sampler that records samples for |thread_handle|. Returns
+    // null if this platform does not support stack sampling.
+    static scoped_ptr<NativeStackSampler> Create(PlatformThreadId thread_id);
+
+    // Notify the sampler that we're starting to record a new profile. This
+    // function is called on the SamplingThread.
+    virtual void ProfileRecordingStarting(Profile* profile) = 0;
+
+    // Record a stack sample. This function is called on the SamplingThread.
+    virtual void RecordStackSample(Sample* sample) = 0;
+
+    // Notify the sampler that we've stopped recording the current profile. This
+    // function is called on the SamplingThread.
+    virtual void ProfileRecordingStopped() = 0;
+
+   protected:
+    NativeStackSampler();
+
+   private:
+    DISALLOW_COPY_AND_ASSIGN(NativeStackSampler);
+  };
+
+  // Represents parameters that configure the sampling.
+  struct BASE_EXPORT SamplingParams {
+    SamplingParams();
+
+    // Time to delay before first samples are taken. Defaults to 0.
+    TimeDelta initial_delay;
+    // Number of sampling bursts to perform. Defaults to 1.
+    int bursts;
+    // Interval between sampling bursts. This is the desired duration from the
+    // start of one burst to the start of the next burst. Defaults to 10s.
+    TimeDelta burst_interval;
+    // Number of samples to record per burst. Defaults to 300.
+    int samples_per_burst;
+    // Interval between samples during a sampling burst. This is the desired
+    // duration from the start of one burst to the start of the next
+    // burst. Defaults to 100ms.
+    TimeDelta sampling_interval;
+    // True if sample ordering is important and should be preserved if and when
+    // this profile is compressed and processed. Defaults to false.
+    bool preserve_sample_ordering;
+  };
+
+  StackSamplingProfiler(PlatformThreadId thread_id,
+                        const SamplingParams& params);
+  ~StackSamplingProfiler();
+
+  // Initializes the profiler and starts sampling.
+  void Start();
+  // Stops the profiler and any ongoing sampling. Calling this function is
+  // optional; if not invoked profiling will terminate when all the profiling
+  // bursts specified in the SamplingParams are completed.
+  void Stop();
+
+  // Gets the pending profiles into *|profiles| and clears the internal
+  // storage. This function is thread safe.
+  //
+  // ***This is intended for use only by UMA.*** Callers who want to process the
+  // collected profiles should use SetCustomCompletedCallback.
+  static void GetPendingProfiles(std::vector<Profile>* profiles);
+
+  // By default, collected profiles are stored internally and can be retrieved
+  // by GetPendingProfiles. If a callback is provided via this function,
+  // however, it will be called with the collected profiles instead. Note that
+  // this call to the callback occurs *on the profiler thread*.
+  void SetCustomCompletedCallback(
+      Callback<void(const std::vector<Profile>&)> callback);
+
+ private:
+  class SamplingThread;
+  struct SamplingThreadDeleter {
+    void operator() (SamplingThread* thread) const;
+  };
+
+  // The thread whose stack will be sampled.
+  PlatformThreadId thread_id_;
+
+  const SamplingParams params_;
+
+  scoped_ptr<SamplingThread, SamplingThreadDeleter> sampling_thread_;
+  scoped_ptr<NativeStackSampler> native_sampler_;
+
+  Callback<void(const std::vector<Profile>&)> custom_completed_callback_;
+
+  DISALLOW_COPY_AND_ASSIGN(StackSamplingProfiler);
+};
+
+// Defined to allow equality check of Samples.
+BASE_EXPORT bool operator==(const StackSamplingProfiler::Frame& a,
+                            const StackSamplingProfiler::Frame& b);
+// Defined to allow ordering of Samples.
+BASE_EXPORT bool operator<(const StackSamplingProfiler::Frame& a,
+                           const StackSamplingProfiler::Frame& b);
+
+}  // namespace base
+
+#endif  // BASE_PROFILER_STACK_SAMPLING_PROFILER_H_
diff --git a/base/profiler/stack_sampling_profiler_posix.cc b/base/profiler/stack_sampling_profiler_posix.cc
new file mode 100644
index 0000000..6a44d7e
--- /dev/null
+++ b/base/profiler/stack_sampling_profiler_posix.cc
@@ -0,0 +1,14 @@
+// Copyright 2015 The Chromium Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style license that can be
+// found in the LICENSE file.
+
+#include "base/profiler/stack_sampling_profiler.h"
+
+namespace base {
+
+scoped_ptr<StackSamplingProfiler::NativeStackSampler>
+StackSamplingProfiler::NativeStackSampler::Create(PlatformThreadId thread_id) {
+  return scoped_ptr<NativeStackSampler>();
+}
+
+}  // namespace base
diff --git a/base/profiler/stack_sampling_profiler_unittest.cc b/base/profiler/stack_sampling_profiler_unittest.cc
new file mode 100644
index 0000000..dfccab4
--- /dev/null
+++ b/base/profiler/stack_sampling_profiler_unittest.cc
@@ -0,0 +1,324 @@
+// Copyright 2015 The Chromium Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style license that can be
+// found in the LICENSE file.
+
+#include <sstream>
+
+#include "base/bind.h"
+#include "base/compiler_specific.h"
+#include "base/path_service.h"
+#include "base/profiler/stack_sampling_profiler.h"
+#include "base/synchronization/waitable_event.h"
+#include "base/threading/platform_thread.h"
+#include "base/time/time.h"
+#include "testing/gtest/include/gtest/gtest.h"
+
+namespace base {
+
+using Frame = StackSamplingProfiler::Frame;
+using Module = StackSamplingProfiler::Module;
+using Sample = StackSamplingProfiler::Sample;
+using Profile = StackSamplingProfiler::Profile;
+
+namespace {
+// A thread to target for profiling, whose stack is guaranteed to contain
+// SignalAndWaitUntilSignaled() when coordinated with the main thread.
+class TargetThread : public PlatformThread::Delegate {
+ public:
+  TargetThread();
+
+  // Implementation of PlatformThread::Delegate:
+  void ThreadMain() override;
+
+  // Wait for the thread to have started and be executing in
+  // SignalAndWaitUntilSignaled().
+  void WaitForThreadStart();
+  // Allow the thread to return from SignalAndWaitUntilSignaled() and finish
+  // execution.
+  void SignalThreadToFinish();
+
+  // This function is guaranteed to be executing between calls to
+  // WaitForThreadStart() and SignalThreadToFinish().
+  static void SignalAndWaitUntilSignaled(WaitableEvent* thread_started_event,
+                                         WaitableEvent* finish_event);
+
+  PlatformThreadId id() const { return id_; }
+
+ private:
+  WaitableEvent thread_started_event_;
+  WaitableEvent finish_event_;
+  PlatformThreadId id_;
+
+  DISALLOW_COPY_AND_ASSIGN(TargetThread);
+};
+
+TargetThread::TargetThread()
+    : thread_started_event_(false, false), finish_event_(false, false),
+      id_(0) {}
+
+void TargetThread::ThreadMain() {
+  id_ = PlatformThread::CurrentId();
+  SignalAndWaitUntilSignaled(&thread_started_event_, &finish_event_);
+}
+
+void TargetThread::WaitForThreadStart() {
+  thread_started_event_.Wait();
+}
+
+void TargetThread::SignalThreadToFinish() {
+  finish_event_.Signal();
+}
+
+// static
+#if defined(_WIN64)
+// Disable optimizations for this function so that it gets its own stack frame.
+#pragma optimize("", off)
+#endif
+void TargetThread::SignalAndWaitUntilSignaled(
+    WaitableEvent* thread_started_event,
+    WaitableEvent* finish_event) {
+  thread_started_event->Signal();
+  finish_event->Wait();
+}
+#if defined(_WIN64)
+#pragma optimize("", on)
+#endif
+
+// Called on the profiler thread when complete. Collects profiles produced by
+// the profiler, and signals an event to allow the main thread to know that that
+// the profiler is done.
+void SaveProfilesAndSignalEvent(std::vector<Profile>* profiles,
+                                WaitableEvent* event,
+                                const std::vector<Profile>& pending_profiles) {
+  *profiles = pending_profiles;
+  event->Signal();
+}
+
+// Captures profiles as specified by |params| on the TargetThread, and returns
+// them in |profiles|. Waits up to |profiler_wait_time| for the profiler to
+// complete.
+void CaptureProfiles(const StackSamplingProfiler::SamplingParams& params,
+                     std::vector<Profile>* profiles,
+                     TimeDelta profiler_wait_time) {
+  TargetThread target_thread;
+  PlatformThreadHandle target_thread_handle;
+  EXPECT_TRUE(PlatformThread::Create(0, &target_thread, &target_thread_handle));
+
+  target_thread.WaitForThreadStart();
+
+  WaitableEvent sampling_thread_completed(true, false);
+  profiles->clear();
+  StackSamplingProfiler profiler(target_thread.id(), params);
+  profiler.SetCustomCompletedCallback(
+      Bind(&SaveProfilesAndSignalEvent, Unretained(profiles),
+           Unretained(&sampling_thread_completed)));
+  profiler.Start();
+  sampling_thread_completed.TimedWait(profiler_wait_time);
+  profiler.Stop();
+  sampling_thread_completed.Wait();
+
+  target_thread.SignalThreadToFinish();
+
+  PlatformThread::Join(target_thread_handle);
+}
+
+// If this executable was linked with /INCREMENTAL (the default for non-official
+// debug and release builds on Windows), function addresses do not correspond to
+// function code itself, but instead to instructions in the Incremental Link
+// Table that jump to the functions. Check for a jump instruction and if present
+// do a little decompilation to find the function's actual starting address.
+const void* MaybeFixupFunctionAddressForILT(const void* function_address) {
+#if defined(_WIN64)
+  const unsigned char* opcode =
+      reinterpret_cast<const unsigned char*>(function_address);
+  if (*opcode == 0xe9) {
+    // This is a relative jump instruction. Assume we're in the ILT and compute
+    // the function start address from the instruction offset.
+    const unsigned char* offset = opcode + 1;
+    const unsigned char* next_instruction = opcode + 5;
+    return next_instruction +
+        static_cast<int64>(*reinterpret_cast<const int32*>(offset));
+  }
+#endif
+  return function_address;
+}
+
+// Searches through the frames in |sample|, returning an iterator to the first
+// frame that has an instruction pointer between |function_address| and
+// |function_address| + |size|. Returns sample.end() if no such frames are
+// found.
+Sample::const_iterator FindFirstFrameWithinFunction(
+    const Sample& sample,
+    const void* function_address,
+    int function_size) {
+  function_address = MaybeFixupFunctionAddressForILT(function_address);
+  for (auto it = sample.begin(); it != sample.end(); ++it) {
+    if ((reinterpret_cast<const unsigned char*>(it->instruction_pointer) >=
+         reinterpret_cast<const unsigned char*>(function_address)) &&
+        (reinterpret_cast<const unsigned char*>(it->instruction_pointer) <
+         (reinterpret_cast<const unsigned char*>(function_address) +
+          function_size)))
+      return it;
+  }
+  return sample.end();
+}
+
+// Formats a sample into a string that can be output for test diagnostics.
+std::string FormatSampleForDiagnosticOutput(
+    const Sample& sample,
+    const std::vector<Module>& modules) {
+  std::ostringstream stream;
+  for (const Frame& frame: sample) {
+    stream << frame.instruction_pointer << " "
+           << modules[frame.module_index].filename.value() << std::endl;
+  }
+  return stream.str();
+}
+
+// Returns a duration that is longer than the test timeout. We would use
+// TimeDelta::Max() but https://crbug.com/465948.
+TimeDelta AVeryLongTimeDelta() { return TimeDelta::FromDays(1); }
+}  // namespace
+
+
+// The tests below are enabled for Win x64 only, pending implementation of the
+// tested functionality on other platforms/architectures.
+
+// Checks that the basic expected information is present in a sampled profile.
+#if defined(_WIN64)
+#define MAYBE_Basic Basic
+#else
+#define MAYBE_Basic DISABLED_Basic
+#endif
+TEST(StackSamplingProfilerTest, MAYBE_Basic) {
+  StackSamplingProfiler::SamplingParams params;
+  params.initial_delay = params.burst_interval = params.sampling_interval =
+      TimeDelta::FromMilliseconds(0);
+  params.bursts = 1;
+  params.samples_per_burst = 1;
+
+  std::vector<Profile> profiles;
+  CaptureProfiles(params, &profiles, AVeryLongTimeDelta());
+
+  // Check that the profile and samples sizes are correct, and the module
+  // indices are in range.
+
+  ASSERT_EQ(1u, profiles.size());
+  const Profile& profile = profiles[0];
+  ASSERT_EQ(1u, profile.samples.size());
+  EXPECT_EQ(params.sampling_interval, profile.sampling_period);
+  const Sample& sample = profile.samples[0];
+  for (const auto& frame : sample) {
+    ASSERT_GE(frame.module_index, 0);
+    ASSERT_LT(frame.module_index, static_cast<int>(profile.modules.size()));
+  }
+
+  // Check that the stack contains a frame for
+  // TargetThread::SignalAndWaitUntilSignaled() and that the frame has this
+  // executable's module.
+
+  // Since we don't have a good way to know the function size, use 100 bytes as
+  // a reasonable window to locate the instruction pointer.
+  Sample::const_iterator loc = FindFirstFrameWithinFunction(
+      sample,
+      reinterpret_cast<const void*>(&TargetThread::SignalAndWaitUntilSignaled),
+      100);
+  ASSERT_TRUE(loc != sample.end())
+      << "Function at "
+      << MaybeFixupFunctionAddressForILT(
+          reinterpret_cast<const void*>(
+              &TargetThread::SignalAndWaitUntilSignaled))
+      << " was not found in stack:" << std::endl
+      << FormatSampleForDiagnosticOutput(sample, profile.modules);
+
+  FilePath executable_path;
+  bool got_executable_path = PathService::Get(FILE_EXE, &executable_path);
+  EXPECT_TRUE(got_executable_path);
+  EXPECT_EQ(executable_path, profile.modules[loc->module_index].filename);
+}
+
+// Checks that the expected number of profiles and samples are present in the
+// profiles produced.
+#if defined(_WIN64)
+#define MAYBE_MultipleProfilesAndSamples MultipleProfilesAndSamples
+#else
+#define MAYBE_MultipleProfilesAndSamples DISABLED_MultipleProfilesAndSamples
+#endif
+TEST(StackSamplingProfilerTest, MAYBE_MultipleProfilesAndSamples) {
+  StackSamplingProfiler::SamplingParams params;
+  params.initial_delay = params.burst_interval = params.sampling_interval =
+      TimeDelta::FromMilliseconds(0);
+  params.bursts = 2;
+  params.samples_per_burst = 3;
+
+  std::vector<Profile> profiles;
+  CaptureProfiles(params, &profiles, AVeryLongTimeDelta());
+
+  ASSERT_EQ(2u, profiles.size());
+  EXPECT_EQ(3u, profiles[0].samples.size());
+  EXPECT_EQ(3u, profiles[1].samples.size());
+}
+
+// Checks that no profiles are captured if the profiling is stopped during the
+// initial delay.
+#if defined(_WIN64)
+#define MAYBE_StopDuringInitialDelay StopDuringInitialDelay
+#else
+#define MAYBE_StopDuringInitialDelay DISABLED_StopDuringInitialDelay
+#endif
+TEST(StackSamplingProfilerTest, MAYBE_StopDuringInitialDelay) {
+  StackSamplingProfiler::SamplingParams params;
+  params.burst_interval = params.sampling_interval =
+      TimeDelta::FromMilliseconds(0);
+  params.initial_delay = TimeDelta::FromSeconds(60);
+  params.bursts = params.samples_per_burst = 1;
+
+  std::vector<Profile> profiles;
+  CaptureProfiles(params, &profiles, TimeDelta::FromMilliseconds(0));
+
+  EXPECT_TRUE(profiles.empty());
+}
+
+// Checks that the single completed profile is captured if the profiling is
+// stopped between bursts.
+#if defined(_WIN64)
+#define MAYBE_StopDuringInterBurstInterval StopDuringInterBurstInterval
+#else
+#define MAYBE_StopDuringInterBurstInterval DISABLED_StopDuringInterBurstInterval
+#endif
+TEST(StackSamplingProfilerTest, MAYBE_StopDuringInterBurstInterval) {
+  StackSamplingProfiler::SamplingParams params;
+  params.initial_delay = params.sampling_interval =
+      TimeDelta::FromMilliseconds(0);
+  params.burst_interval = TimeDelta::FromSeconds(60);
+  params.bursts = 2;
+  params.samples_per_burst = 1;
+
+  std::vector<Profile> profiles;
+  CaptureProfiles(params, &profiles, TimeDelta::FromMilliseconds(50));
+
+  ASSERT_EQ(1u, profiles.size());
+  EXPECT_EQ(1u, profiles[0].samples.size());
+}
+
+// Checks that only completed profiles are captured.
+#if defined(_WIN64)
+#define MAYBE_StopDuringInterSampleInterval StopDuringInterSampleInterval
+#else
+#define MAYBE_StopDuringInterSampleInterval \
+  DISABLED_StopDuringInterSampleInterval
+#endif
+TEST(StackSamplingProfilerTest, MAYBE_StopDuringInterSampleInterval) {
+  StackSamplingProfiler::SamplingParams params;
+  params.initial_delay = params.burst_interval = TimeDelta::FromMilliseconds(0);
+  params.sampling_interval = TimeDelta::FromSeconds(60);
+  params.bursts = 1;
+  params.samples_per_burst = 2;
+
+  std::vector<Profile> profiles;
+  CaptureProfiles(params, &profiles, TimeDelta::FromMilliseconds(50));
+
+  EXPECT_TRUE(profiles.empty());
+}
+
+}  // namespace tracked_objects
diff --git a/base/profiler/stack_sampling_profiler_win.cc b/base/profiler/stack_sampling_profiler_win.cc
new file mode 100644
index 0000000..ba46cf0
--- /dev/null
+++ b/base/profiler/stack_sampling_profiler_win.cc
@@ -0,0 +1,325 @@
+// Copyright 2015 The Chromium Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style license that can be
+// found in the LICENSE file.
+
+#include "base/profiler/stack_sampling_profiler.h"
+
+#include <dbghelp.h>
+#include <map>
+#include <utility>
+#include <windows.h>
+
+#include "base/logging.h"
+#include "base/time/time.h"
+#include "base/win/pe_image.h"
+#include "base/win/scoped_handle.h"
+
+namespace base {
+
+namespace {
+
+class NativeStackSamplerWin : public StackSamplingProfiler::NativeStackSampler {
+ public:
+  explicit NativeStackSamplerWin(win::ScopedHandle thread_handle);
+  ~NativeStackSamplerWin() override;
+
+  // StackSamplingProfiler::NativeStackSampler:
+  void ProfileRecordingStarting(
+      StackSamplingProfiler::Profile* profile) override;
+  void RecordStackSample(StackSamplingProfiler::Sample* sample) override;
+  void ProfileRecordingStopped() override;
+
+ private:
+  static bool GetModuleInfo(HMODULE module,
+                            StackSamplingProfiler::Module* module_info);
+
+  void CopyToSample(const void* const instruction_pointers[],
+                    const HMODULE modules[],
+                    int stack_depth,
+                    StackSamplingProfiler::Sample* sample,
+                    std::vector<StackSamplingProfiler::Module>* module_infos);
+
+  win::ScopedHandle thread_handle_;
+  // Weak. Points to the profile being recorded between
+  // ProfileRecordingStarting() and ProfileRecordingStopped().
+  StackSamplingProfiler::Profile* current_profile_;
+  // Maps a module to the module's index within current_profile_->modules.
+  std::map<HMODULE, int> profile_module_index_;
+
+  DISALLOW_COPY_AND_ASSIGN(NativeStackSamplerWin);
+};
+
+// Walk the stack represented by |context| from the current frame downwards,
+// recording the instruction pointers for each frame in |instruction_pointers|.
+int RecordStack(CONTEXT* context,
+                int max_stack_size,
+                const void* instruction_pointers[],
+                bool* last_frame_is_unknown_function) {
+#ifdef _WIN64
+  *last_frame_is_unknown_function = false;
+
+  IMAGEHLP_SYMBOL64 sym;
+  sym.SizeOfStruct = sizeof(sym);
+  sym.MaxNameLength = 0;
+
+  for (int i = 0; i < max_stack_size; ++i) {
+    // Try to look up unwind metadata for the current function.
+    ULONG64 image_base;
+    PRUNTIME_FUNCTION runtime_function =
+        RtlLookupFunctionEntry(context->Rip, &image_base, nullptr);
+
+    instruction_pointers[i] = reinterpret_cast<void*>(context->Rip);
+
+    if (runtime_function) {
+      KNONVOLATILE_CONTEXT_POINTERS nvcontext = {0};
+      void* handler_data;
+      ULONG64 establisher_frame;
+      RtlVirtualUnwind(0, image_base, context->Rip, runtime_function, context,
+          &handler_data, &establisher_frame, &nvcontext);
+    } else {
+      // If we don't have a RUNTIME_FUNCTION, then we've encountered
+      // a leaf function.  Adjust the stack appropriately.
+      context->Rip = *reinterpret_cast<PDWORD64>(context->Rsp);
+      context->Rsp += 8;
+      *last_frame_is_unknown_function = true;
+    }
+
+    if (!context->Rip)
+      return i;
+  }
+  return max_stack_size;
+#else
+  return 0;
+#endif
+}
+
+// Fills in |modules| corresponding to the pointers to code in |addresses|. The
+// modules are returned with reference counts incremented should be freed with
+// FreeModules.
+void FindModulesForAddresses(const void* const addresses[], HMODULE modules[],
+                             int stack_depth,
+                             bool last_frame_is_unknown_function) {
+  const int module_frames = last_frame_is_unknown_function ? stack_depth - 1 :
+      stack_depth;
+  for (int i = 0; i < module_frames; ++i) {
+    HMODULE module = NULL;
+    if (GetModuleHandleEx(GET_MODULE_HANDLE_EX_FLAG_FROM_ADDRESS,
+                          reinterpret_cast<LPCTSTR>(addresses[i]),
+                          &module)) {
+      // HMODULE is the base address of the module.
+      DCHECK_LT(reinterpret_cast<const void*>(module), addresses[i]);
+      modules[i] = module;
+    }
+  }
+}
+
+// Free the modules returned by FindModulesForAddresses.
+void FreeModules(int stack_depth, HMODULE modules[]) {
+  for (int i = 0; i < stack_depth; ++i) {
+    if (modules[i])
+      ::FreeLibrary(modules[i]);
+  }
+}
+
+// Disables priority boost on a thread for the lifetime of the object.
+class ScopedDisablePriorityBoost {
+ public:
+  ScopedDisablePriorityBoost(HANDLE thread_handle);
+  ~ScopedDisablePriorityBoost();
+
+ private:
+  HANDLE thread_handle_;
+  BOOL got_previous_boost_state_;
+  BOOL boost_state_was_disabled_;
+
+  DISALLOW_COPY_AND_ASSIGN(ScopedDisablePriorityBoost);
+};
+
+ScopedDisablePriorityBoost::ScopedDisablePriorityBoost(HANDLE thread_handle)
+    : thread_handle_(thread_handle),
+      got_previous_boost_state_(false),
+      boost_state_was_disabled_(false) {
+  got_previous_boost_state_ =
+      ::GetThreadPriorityBoost(thread_handle_, &boost_state_was_disabled_);
+  if (got_previous_boost_state_ && !boost_state_was_disabled_) {
+    // Confusingly, TRUE disables priority boost ...
+    ::SetThreadPriorityBoost(thread_handle_, TRUE);
+  }
+}
+
+ScopedDisablePriorityBoost::~ScopedDisablePriorityBoost() {
+  if (got_previous_boost_state_ && !boost_state_was_disabled_) {
+    // ... and FALSE enables priority boost.
+    ::SetThreadPriorityBoost(thread_handle_, FALSE);
+  }
+}
+
+// Suspends the thread with |thread_handle|, records the stack into
+// |instruction_pointers|, then resumes the thread. Returns the size of the
+// stack.
+int SuspendThreadAndRecordStack(HANDLE thread_handle, int max_stack_size,
+                                const void* instruction_pointers[],
+                                bool* last_frame_is_unknown_function) {
+#if defined(_WIN64)
+  if (RtlVirtualUnwind == nullptr || RtlLookupFunctionEntry == nullptr)
+    return 0;
+#endif
+
+  if (::SuspendThread(thread_handle) == -1) {
+    LOG(ERROR) << "SuspendThread failed: " << GetLastError();
+    return 0;
+  }
+
+  CONTEXT thread_context = {0};
+  thread_context.ContextFlags = CONTEXT_FULL;
+  if (!::GetThreadContext(thread_handle, &thread_context)) {
+    LOG(ERROR) << "GetThreadContext failed: " << GetLastError();
+  }
+
+  int stack_depth = RecordStack(&thread_context, max_stack_size,
+                                instruction_pointers,
+                                last_frame_is_unknown_function);
+
+  {
+    ScopedDisablePriorityBoost disable_priority_boost(thread_handle);
+    if (::ResumeThread(thread_handle) == -1)
+      LOG(ERROR) << "ResumeThread failed: " << GetLastError();
+  }
+
+  return stack_depth;
+}
+
+}  // namespace
+
+scoped_ptr<StackSamplingProfiler::NativeStackSampler>
+StackSamplingProfiler::NativeStackSampler::Create(PlatformThreadId thread_id) {
+#if _WIN64
+  // Get the thread's handle.
+  HANDLE thread_handle = ::OpenThread(
+      THREAD_GET_CONTEXT | THREAD_SUSPEND_RESUME | THREAD_QUERY_INFORMATION,
+      FALSE,
+      thread_id);
+  DCHECK(thread_handle) << "OpenThread failed";
+
+  return scoped_ptr<NativeStackSampler>(new NativeStackSamplerWin(
+      win::ScopedHandle(thread_handle)));
+#else
+  return scoped_ptr<NativeStackSampler>();
+#endif
+}
+
+NativeStackSamplerWin::NativeStackSamplerWin(win::ScopedHandle thread_handle)
+    : thread_handle_(thread_handle.Take()) {
+#ifdef _WIN64
+  if (RtlVirtualUnwind == nullptr && RtlLookupFunctionEntry == nullptr) {
+    const HMODULE nt_dll_handle = ::GetModuleHandle(L"ntdll.dll");
+    // This should always be non-null, but handle just in case.
+    if (nt_dll_handle) {
+      reinterpret_cast<void*&>(RtlVirtualUnwind) =
+          ::GetProcAddress(nt_dll_handle, "RtlVirtualUnwind");
+      reinterpret_cast<void*&>(RtlLookupFunctionEntry) =
+          ::GetProcAddress(nt_dll_handle, "RtlLookupFunctionEntry");
+    }
+  }
+#endif
+}
+
+NativeStackSamplerWin::~NativeStackSamplerWin() {
+}
+
+void NativeStackSamplerWin::ProfileRecordingStarting(
+    StackSamplingProfiler::Profile* profile) {
+  current_profile_ = profile;
+  profile_module_index_.clear();
+}
+
+void NativeStackSamplerWin::RecordStackSample(
+    StackSamplingProfiler::Sample* sample) {
+  DCHECK(current_profile_);
+
+  const int max_stack_size = 64;
+  const void* instruction_pointers[max_stack_size] = {0};
+  HMODULE modules[max_stack_size] = {0};
+
+  bool last_frame_is_unknown_function = false;
+  int stack_depth = SuspendThreadAndRecordStack(
+      thread_handle_.Get(), max_stack_size, instruction_pointers,
+      &last_frame_is_unknown_function);
+  FindModulesForAddresses(instruction_pointers, modules, stack_depth,
+                          last_frame_is_unknown_function);
+  CopyToSample(instruction_pointers, modules, stack_depth, sample,
+               &current_profile_->modules);
+  FreeModules(stack_depth, modules);
+}
+
+void NativeStackSamplerWin::ProfileRecordingStopped() {
+  current_profile_ = nullptr;
+}
+
+// static
+bool NativeStackSamplerWin::GetModuleInfo(
+    HMODULE module,
+    StackSamplingProfiler::Module* module_info) {
+  wchar_t module_name[MAX_PATH];
+  DWORD result_length =
+      GetModuleFileName(module, module_name, arraysize(module_name));
+  if (result_length == 0)
+    return false;
+
+  module_info->filename = base::FilePath(module_name);
+
+  module_info->base_address = reinterpret_cast<const void*>(module);
+
+  GUID guid;
+  DWORD age;
+  win::PEImage(module).GetDebugId(&guid, &age);
+  module_info->id.insert(module_info->id.end(),
+                         reinterpret_cast<char*>(&guid),
+                         reinterpret_cast<char*>(&guid + 1));
+  module_info->id.insert(module_info->id.end(),
+                         reinterpret_cast<char*>(&age),
+                         reinterpret_cast<char*>(&age + 1));
+
+  return true;
+}
+
+void NativeStackSamplerWin::CopyToSample(
+    const void* const instruction_pointers[],
+    const HMODULE modules[],
+    int stack_depth,
+    StackSamplingProfiler::Sample* sample,
+    std::vector<StackSamplingProfiler::Module>* module_infos) {
+  sample->clear();
+  sample->reserve(stack_depth);
+
+  for (int i = 0; i < stack_depth; ++i) {
+    sample->push_back(StackSamplingProfiler::Frame());
+    StackSamplingProfiler::Frame& frame = sample->back();
+
+    frame.instruction_pointer = instruction_pointers[i];
+
+    // Record an invalid module index if we don't have a valid module.
+    if (!modules[i]) {
+      frame.module_index = -1;
+      continue;
+    }
+
+    auto loc = profile_module_index_.find(modules[i]);
+    if (loc == profile_module_index_.end()) {
+      StackSamplingProfiler::Module module_info;
+      // Record an invalid module index if we have a module but can't find
+      // information on it.
+      if (!GetModuleInfo(modules[i], &module_info)) {
+        frame.module_index = -1;
+        continue;
+      }
+      module_infos->push_back(module_info);
+      loc = profile_module_index_.insert(std::make_pair(
+          modules[i], static_cast<int>(module_infos->size() - 1))).first;
+    }
+
+    frame.module_index = loc->second;
+  }
+}
+
+}  // namespace base