path: root/tools/profiler/core/platform-macos.cpp

/* -*- Mode: C++; tab-width: 8; indent-tabs-mode: nil; c-basic-offset: 2 -*- */
/* vim: set ts=8 sts=2 et sw=2 tw=80: */
/* This Source Code Form is subject to the terms of the Mozilla Public
 * License, v. 2.0. If a copy of the MPL was not distributed with this
 * file, You can obtain one at http://mozilla.org/MPL/2.0/. */

#include <unistd.h>
#include <sys/mman.h>
#include <mach/mach_init.h>
#include <mach-o/getsect.h>

#include <AvailabilityMacros.h>

#include <pthread.h>
#include <semaphore.h>
#include <signal.h>
#include <libkern/OSAtomic.h>
#include <libproc.h>
#include <mach/mach.h>
#include <mach/semaphore.h>
#include <mach/task.h>
#include <mach/thread_act.h>
#include <mach/vm_statistics.h>
#include <sys/time.h>
#include <sys/resource.h>
#include <sys/syscall.h>
#include <sys/types.h>
#include <sys/sysctl.h>
#include <stdarg.h>
#include <stdlib.h>
#include <string.h>
#include <errno.h>
#include <math.h>

// this port is based off of v8 svn revision 9837

mozilla::profiler::PlatformData::PlatformData(ProfilerThreadId aThreadId)
    : mProfiledThread(mach_thread_self()) {}

mozilla::profiler::PlatformData::~PlatformData() {
  // Deallocate Mach port for thread.
  mach_port_deallocate(mach_task_self(), mProfiledThread);
}

////////////////////////////////////////////////////////////////////////
// BEGIN Sampler target specifics

Sampler::Sampler(PSLockRef aLock) {}

void Sampler::Disable(PSLockRef aLock) {}

static void StreamMetaPlatformSampleUnits(PSLockRef aLock,
                                          SpliceableJSONWriter& aWriter) {
  // Microseconds.
  aWriter.StringProperty("threadCPUDelta", "\u00B5s");
}

/* static */
uint64_t RunningTimes::ConvertRawToJson(uint64_t aRawValue) {
  return aRawValue;
}

namespace mozilla::profiler {
bool GetCpuTimeSinceThreadStartInNs(
    uint64_t* aResult, const mozilla::profiler::PlatformData& aPlatformData) {
  thread_extended_info_data_t threadInfoData;
  mach_msg_type_number_t count = THREAD_EXTENDED_INFO_COUNT;
  if (thread_info(aPlatformData.ProfiledThread(), THREAD_EXTENDED_INFO,
                  (thread_info_t)&threadInfoData, &count) != KERN_SUCCESS) {
    return false;
  }

  *aResult = threadInfoData.pth_user_time + threadInfoData.pth_system_time;
  return true;
}
}  // namespace mozilla::profiler

static RunningTimes GetProcessRunningTimesDiff(
    PSLockRef aLock, RunningTimes& aPreviousRunningTimesToBeUpdated) {
  AUTO_PROFILER_STATS(GetProcessRunningTimes);

  RunningTimes newRunningTimes;
  {
    AUTO_PROFILER_STATS(GetProcessRunningTimes_task_info);

    static const auto pid = getpid();
    struct proc_taskinfo pti;
    if ((unsigned long)proc_pidinfo(pid, PROC_PIDTASKINFO, 0, &pti,
                                    PROC_PIDTASKINFO_SIZE) >=
        PROC_PIDTASKINFO_SIZE) {
      newRunningTimes.SetThreadCPUDelta(pti.pti_total_user +
                                        pti.pti_total_system);
    }
    newRunningTimes.SetPostMeasurementTimeStamp(TimeStamp::Now());
  };

  const RunningTimes diff = newRunningTimes - aPreviousRunningTimesToBeUpdated;
  aPreviousRunningTimesToBeUpdated = newRunningTimes;
  return diff;
}

static RunningTimes GetThreadRunningTimesDiff(
    PSLockRef aLock,
    ThreadRegistration::UnlockedRWForLockedProfiler& aThreadData) {
  AUTO_PROFILER_STATS(GetRunningTimes);

  const mozilla::profiler::PlatformData& platformData =
      aThreadData.PlatformDataCRef();

  const RunningTimes newRunningTimes = GetRunningTimesWithTightTimestamp(
      [&platformData](RunningTimes& aRunningTimes) {
        AUTO_PROFILER_STATS(GetRunningTimes_thread_info);
        thread_basic_info_data_t threadBasicInfo;
        mach_msg_type_number_t basicCount = THREAD_BASIC_INFO_COUNT;
        if (thread_info(platformData.ProfiledThread(), THREAD_BASIC_INFO,
                        reinterpret_cast<thread_info_t>(&threadBasicInfo),
                        &basicCount) == KERN_SUCCESS &&
            basicCount == THREAD_BASIC_INFO_COUNT) {
          uint64_t userTimeUs =
              uint64_t(threadBasicInfo.user_time.seconds) *
                  uint64_t(USEC_PER_SEC) +
              uint64_t(threadBasicInfo.user_time.microseconds);
          uint64_t systemTimeUs =
              uint64_t(threadBasicInfo.system_time.seconds) *
                  uint64_t(USEC_PER_SEC) +
              uint64_t(threadBasicInfo.system_time.microseconds);
          aRunningTimes.ResetThreadCPUDelta(userTimeUs + systemTimeUs);
        } else {
          aRunningTimes.ClearThreadCPUDelta();
        }
      });

  ProfiledThreadData* profiledThreadData =
      aThreadData.GetProfiledThreadData(aLock);
  MOZ_ASSERT(profiledThreadData);
  RunningTimes& previousRunningTimes =
      profiledThreadData->PreviousThreadRunningTimesRef();
  const RunningTimes diff = newRunningTimes - previousRunningTimes;
  previousRunningTimes = newRunningTimes;
  return diff;
}

static void DiscardSuspendedThreadRunningTimes(
    PSLockRef aLock,
    ThreadRegistration::UnlockedRWForLockedProfiler& aThreadData) {
  // Nothing to do!
  // On macOS, suspending a thread doesn't make that thread work.
}

template <typename Func>
void Sampler::SuspendAndSampleAndResumeThread(
    PSLockRef aLock,
    const ThreadRegistration::UnlockedReaderAndAtomicRWOnThread& aThreadData,
    const TimeStamp& aNow, const Func& aProcessRegs) {
  thread_act_t samplee_thread = aThreadData.PlatformDataCRef().ProfiledThread();

  //----------------------------------------------------------------//
  // Suspend the samplee thread and get its context.

  // We're using thread_suspend on OS X because pthread_kill (which is what we
  // at one time used on Linux) has less consistent performance and causes
  // strange crashes, see bug 1166778 and bug 1166808.  thread_suspend
  // is also just a lot simpler to use.

  if (KERN_SUCCESS != thread_suspend(samplee_thread)) {
    return;
  }

  //----------------------------------------------------------------//
  // Sample the target thread.

  // WARNING WARNING WARNING WARNING WARNING WARNING WARNING WARNING
  //
  // The profiler's "critical section" begins here.  We must be very careful
  // what we do here, or risk deadlock.  See the corresponding comment in
  // platform-linux-android.cpp for details.

#if defined(__x86_64__)
  thread_state_flavor_t flavor = x86_THREAD_STATE64;
  x86_thread_state64_t state;
  mach_msg_type_number_t count = x86_THREAD_STATE64_COUNT;
#  if __DARWIN_UNIX03
#    define REGISTER_FIELD(name) __r##name
#  else
#    define REGISTER_FIELD(name) r##name
#  endif  // __DARWIN_UNIX03
#elif defined(__aarch64__)
  thread_state_flavor_t flavor = ARM_THREAD_STATE64;
  arm_thread_state64_t state;
  mach_msg_type_number_t count = ARM_THREAD_STATE64_COUNT;
#  if __DARWIN_UNIX03
#    define REGISTER_FIELD(name) __##name
#  else
#    define REGISTER_FIELD(name) name
#  endif  // __DARWIN_UNIX03
#else
#  error "unknown architecture"
#endif

  if (thread_get_state(samplee_thread, flavor,
                       reinterpret_cast<natural_t*>(&state),
                       &count) == KERN_SUCCESS) {
    Registers regs;
#if defined(__x86_64__)
    regs.mPC = reinterpret_cast<Address>(state.REGISTER_FIELD(ip));
    regs.mSP = reinterpret_cast<Address>(state.REGISTER_FIELD(sp));
    regs.mFP = reinterpret_cast<Address>(state.REGISTER_FIELD(bp));
    regs.mLR = 0;
#elif defined(__aarch64__)
    regs.mPC = reinterpret_cast<Address>(state.REGISTER_FIELD(pc));
    regs.mSP = reinterpret_cast<Address>(state.REGISTER_FIELD(sp));
    regs.mFP = reinterpret_cast<Address>(state.REGISTER_FIELD(fp));
    regs.mLR = reinterpret_cast<Address>(state.REGISTER_FIELD(lr));
#else
#  error "unknown architecture"
#endif

    aProcessRegs(regs, aNow);
  }

#undef REGISTER_FIELD

  //----------------------------------------------------------------//
  // Resume the target thread.

  thread_resume(samplee_thread);

  // The profiler's critical section ends here.
  //
  // WARNING WARNING WARNING WARNING WARNING WARNING WARNING WARNING
}

// END Sampler target specifics
////////////////////////////////////////////////////////////////////////

////////////////////////////////////////////////////////////////////////
// BEGIN SamplerThread target specifics

static void* ThreadEntry(void* aArg) {
  auto thread = static_cast<SamplerThread*>(aArg);
  thread->Run();
  return nullptr;
}

SamplerThread::SamplerThread(PSLockRef aLock, uint32_t aActivityGeneration,
                             double aIntervalMilliseconds, uint32_t aFeatures)
    : mSampler(aLock),
      mActivityGeneration(aActivityGeneration),
      mIntervalMicroseconds(
          std::max(1, int(floor(aIntervalMilliseconds * 1000 + 0.5)))),
      mThread{nullptr} {
  pthread_attr_t* attr_ptr = nullptr;
  if (pthread_create(&mThread, attr_ptr, ThreadEntry, this) != 0) {
    MOZ_CRASH("pthread_create failed");
  }
}

SamplerThread::~SamplerThread() {
  pthread_join(mThread, nullptr);
  // Just in the unlikely case some callbacks were added between the end of the
  // thread and now.
  InvokePostSamplingCallbacks(std::move(mPostSamplingCallbackList),
                              SamplingState::JustStopped);
}

void SamplerThread::SleepMicro(uint32_t aMicroseconds) {
  usleep(aMicroseconds);
  // FIXME: the OSX 10.12 page for usleep says "The usleep() function is
  // obsolescent.  Use nanosleep(2) instead."  This implementation could be
  // merged with the linux-android version.  Also, this doesn't handle the
  // case where the usleep call is interrupted by a signal.
}

void SamplerThread::Stop(PSLockRef aLock) { mSampler.Disable(aLock); }

// END SamplerThread target specifics
////////////////////////////////////////////////////////////////////////

static void PlatformInit(PSLockRef aLock) {}

// clang-format off
#if defined(HAVE_NATIVE_UNWIND)
// Derive the stack pointer from the frame pointer. The 0x10 offset is
// 8 bytes for the previous frame pointer and 8 bytes for the return
// address both stored on the stack after at the beginning of the current
// frame.
#  define REGISTERS_SYNC_POPULATE(regs)                                       \
    regs.mSP = reinterpret_cast<Address>(__builtin_frame_address(0)) + 0x10;  \
    _Pragma("GCC diagnostic push")                                            \
    _Pragma("GCC diagnostic ignored \"-Wframe-address\"")                     \
    regs.mFP = reinterpret_cast<Address>(__builtin_frame_address(1));         \
    _Pragma("GCC diagnostic pop")                                             \
    regs.mPC = reinterpret_cast<Address>(                                     \
        __builtin_extract_return_addr(__builtin_return_address(0)));          \
    regs.mLR = 0;
#endif
// clang-format on