Xenia-Canary/src/xenia/kernel/objects/xthread.cc

/**
 ******************************************************************************
 * Xenia : Xbox 360 Emulator Research Project                                 *
 ******************************************************************************
 * Copyright 2013 Ben Vanik. All rights reserved.                             *
 * Released under the BSD license - see LICENSE in the root for more details. *
 ******************************************************************************
 */

#include <xenia/kernel/objects/xthread.h>

#include <poly/math.h>

#include <xenia/cpu/cpu.h>
#include <xenia/kernel/native_list.h>
#include <xenia/kernel/xboxkrnl_threading.h>
#include <xenia/kernel/objects/xevent.h>
#include <xenia/kernel/objects/xuser_module.h>


using namespace alloy;
using namespace xe;
using namespace xe::cpu;
using namespace xe::kernel;


namespace {
  static uint32_t next_xthread_id = 0;
  static uint32_t current_thread_tls = xeKeTlsAlloc();
  static xe_mutex_t* critical_region_ = xe_mutex_alloc(10000);
  static XThread* shared_kernel_thread_ = 0;
}


XThread::XThread(KernelState* kernel_state,
                 uint32_t stack_size,
                 uint32_t xapi_thread_startup,
                 uint32_t start_address, uint32_t start_context,
                 uint32_t creation_flags) :
    XObject(kernel_state, kTypeThread),
    thread_id_(++next_xthread_id),
    thread_handle_(0),
    thread_state_address_(0),
    thread_state_(0),
    event_(NULL),
    name_(0),
    irql_(0) {
  creation_params_.stack_size           = stack_size;
  creation_params_.xapi_thread_startup  = xapi_thread_startup;
  creation_params_.start_address        = start_address;
  creation_params_.start_context        = start_context;

  // top 8 bits = processor ID (or 0 for default)
  // bit 0 = 1 to create suspended
  creation_params_.creation_flags       = creation_flags;

  // Adjust stack size - min of 16k.
  if (creation_params_.stack_size < 16 * 1024) {
    creation_params_.stack_size = 16 * 1024;
  }

  apc_lock_ = xe_mutex_alloc();
  apc_list_ = new NativeList(kernel_state->memory());

  event_ = new XEvent(kernel_state);
  event_->Initialize(true, false);

  // The kernel does not take a reference. We must unregister in the dtor.
  kernel_state_->RegisterThread(this);
}

XThread::~XThread() {
  // Unregister first to prevent lookups while deleting.
  kernel_state_->UnregisterThread(this);

  delete apc_list_;
  xe_mutex_free(apc_lock_);

  event_->Release();

  PlatformDestroy();

  if (thread_state_) {
    delete thread_state_;
  }
  if (scratch_address_) {
    kernel_state()->memory()->HeapFree(scratch_address_, 0);
  }
  if (tls_address_) {
    kernel_state()->memory()->HeapFree(tls_address_, 0);
  }
  if (thread_state_address_) {
    kernel_state()->memory()->HeapFree(thread_state_address_, 0);
  }
  if (name_) {
    xe_free(name_);
  }

  if (thread_handle_) {
    // TODO(benvanik): platform kill
    XELOGE("Thread disposed without exiting");
  }
}

XThread* XThread::GetCurrentThread() {
  XThread* thread = (XThread*)xeKeTlsGetValue(current_thread_tls);
  if (!thread) {
    // Assume this is some shared interrupt thread/etc.
    XThread::EnterCriticalRegion();
    thread = shared_kernel_thread_;
    if (!thread) {
      thread = new XThread(
          KernelState::shared(), 32 * 1024, 0, 0, 0, 0);
      shared_kernel_thread_ = thread;
      xeKeTlsSetValue(current_thread_tls, (uint64_t)thread);
    }
    XThread::LeaveCriticalRegion();
  }
  return thread;
}

uint32_t XThread::GetCurrentThreadHandle() {
  XThread* thread = XThread::GetCurrentThread();
  return thread->handle();
}

uint32_t XThread::GetCurrentThreadId(const uint8_t* thread_state_block) {
  return poly::load_and_swap<uint32_t>(thread_state_block + 0x14C);
}

uint32_t XThread::thread_state() {
  return thread_state_address_;
}

uint32_t XThread::thread_id() {
  return thread_id_;
}

uint32_t XThread::last_error() {
  uint8_t *p = memory()->Translate(thread_state_address_);
  return poly::load_and_swap<uint32_t>(p + 0x160);
}

void XThread::set_last_error(uint32_t error_code) {
  uint8_t *p = memory()->Translate(thread_state_address_);
  poly::store_and_swap<uint32_t>(p + 0x160, error_code);
}

void XThread::set_name(const char* name) {
  if (name == name_) {
    return;
  }
  if (name_) {
    xe_free(name_);
  }
  name_ = xestrdupa(name);

#if XE_PLATFORM_WIN32
  // Do the nasty set for us.
  #pragma pack(push, 8)
  typedef struct tagTHREADNAME_INFO {
    DWORD dwType; // must be 0x1000
    LPCSTR szName; // pointer to name (in user addr space)
    DWORD dwThreadID; // thread ID (-1=caller thread)
    DWORD dwFlags; // reserved for future use, must be zero
  } THREADNAME_INFO;
  #pragma pack(pop)
  THREADNAME_INFO info;
  info.dwType = 0x1000;
  info.szName = name_;
  info.dwThreadID = ::GetThreadId(thread_handle_);
  info.dwFlags = 0;
  __try {
    RaiseException(0x406D1388, 0, sizeof(info) / sizeof(ULONG_PTR), (ULONG_PTR*)&info);
  } __except(EXCEPTION_CONTINUE_EXECUTION) {
  }
#endif  // WIN32
}

X_STATUS XThread::Create() {
  // Allocate thread state block from heap.
  // This is set as r13 for user code and some special inlined Win32 calls
  // (like GetLastError/etc) will poke it directly.
  // We try to use it as our primary store of data just to keep things all
  // consistent.
  // 0x000: pointer to tls data
  // 0x100: pointer to self?
  // 0x14C: thread id
  // 0x150: if >0 then error states don't get set
  // 0x160: last error
  // So, at offset 0x100 we have a 4b pointer to offset 200, then have the
  // structure.
  thread_state_address_ = (uint32_t)memory()->HeapAlloc(
      0, 2048, MEMORY_FLAG_ZERO);
  if (!thread_state_address_) {
    XELOGW("Unable to allocate thread state block");
    return X_STATUS_NO_MEMORY;
  }

  // Set native info.
  SetNativePointer(thread_state_address_);

  XUserModule* module = kernel_state()->GetExecutableModule();

  // Allocate thread scratch.
  // This is used by interrupts/APCs/etc so we can round-trip pointers through.
  scratch_size_ = 4 * 16;
  scratch_address_ = (uint32_t)memory()->HeapAlloc(
      0, scratch_size_, MEMORY_FLAG_ZERO);

  // Allocate TLS block.
  const xe_xex2_header_t* header = module->xex_header();
  uint32_t tls_size = header->tls_info.slot_count * header->tls_info.data_size;
  tls_address_ = (uint32_t)memory()->HeapAlloc(
      0, tls_size, MEMORY_FLAG_ZERO);
  if (!tls_address_) {
    XELOGW("Unable to allocate thread local storage block");
    module->Release();
    return X_STATUS_NO_MEMORY;
  }

  // Copy in default TLS info.
  // TODO(benvanik): is this correct?
  memory()->Copy(
      tls_address_, header->tls_info.raw_data_address, tls_size);

  // Setup the thread state block (last error/etc).
  uint8_t *p = memory()->Translate(thread_state_address_);
  poly::store_and_swap<uint32_t>(p + 0x000, tls_address_);
  poly::store_and_swap<uint32_t>(p + 0x100, thread_state_address_);
  poly::store_and_swap<uint32_t>(p + 0x14C, thread_id_);
  poly::store_and_swap<uint32_t>(p + 0x150, 0); // ?
  poly::store_and_swap<uint32_t>(p + 0x160, 0); // last error

  // Allocate processor thread state.
  // This is thread safe.
  thread_state_ = new XenonThreadState(
      kernel_state()->processor()->runtime(),
      thread_id_, creation_params_.stack_size, thread_state_address_);

  X_STATUS return_code = PlatformCreate();
  if (XFAILED(return_code)) {
    XELOGW("Unable to create platform thread (%.8X)", return_code);
    module->Release();
    return return_code;
  }

  char thread_name[32];
  xesnprintfa(thread_name, XECOUNT(thread_name), "XThread%04X", handle());
  set_name(thread_name);

  uint32_t proc_mask = creation_params_.creation_flags >> 24;
  if (proc_mask) {
    SetAffinity(proc_mask);
  }

  module->Release();
  return X_STATUS_SUCCESS;
}

X_STATUS XThread::Exit(int exit_code) {
  // TODO(benvanik): set exit code in thread state block

  // TODO(benvanik); dispatch events? waiters? etc?
  event_->Set(0, false);
  RundownAPCs();

  // NOTE: unless PlatformExit fails, expect it to never return!
  X_STATUS return_code = PlatformExit(exit_code);
  if (XFAILED(return_code)) {
    return return_code;
  }
  return X_STATUS_SUCCESS;
}

#if XE_PLATFORM_WIN32

static uint32_t __stdcall XThreadStartCallbackWin32(void* param) {
  XThread* thread = reinterpret_cast<XThread*>(param);
  xe::Profiler::ThreadEnter(thread->name());
  xeKeTlsSetValue(current_thread_tls, (uint64_t)thread);
  thread->Execute();
  xeKeTlsSetValue(current_thread_tls, NULL);
  thread->Release();
  xe::Profiler::ThreadExit();
  return 0;
}

X_STATUS XThread::PlatformCreate() {
  bool suspended = creation_params_.creation_flags & 0x1;
  thread_handle_ = CreateThread(
      NULL,
      creation_params_.stack_size,
      (LPTHREAD_START_ROUTINE)XThreadStartCallbackWin32,
      this,
      suspended ? CREATE_SUSPENDED : 0,
      NULL);
  if (!thread_handle_) {
    uint32_t last_error = GetLastError();
    // TODO(benvanik): translate?
    XELOGE("CreateThread failed with %d", last_error);
    return last_error;
  }

  return X_STATUS_SUCCESS;
}

void XThread::PlatformDestroy() {
  CloseHandle(reinterpret_cast<HANDLE>(thread_handle_));
  thread_handle_ = NULL;
}

X_STATUS XThread::PlatformExit(int exit_code) {
  // NOTE: does not return.
  ExitThread(exit_code);
  return X_STATUS_SUCCESS;
}

#else

static void* XThreadStartCallbackPthreads(void* param) {
  XThread* thread = reinterpret_cast<XThread*>(param);
  xe::Profiler::ThreadEnter(thread->name());
  xeKeTlsSetValue(current_thread_tls, (uint64_t)thread);
  thread->Execute();
  xeKeTlsSetValue(current_thread_tls, NULL);
  thread->Release();
  xe::Profiler::ThreadExit();
  return 0;
}

X_STATUS XThread::PlatformCreate() {
  pthread_attr_t attr;

  pthread_attr_init(&attr);
  pthread_attr_setstacksize(&attr, creation_params_.stack_size);

  int result_code;
  if (creation_params_.creation_flags & 0x1) {
#if XE_PLATFORM_OSX
    result_code = pthread_create_suspended_np(
        reinterpret_cast<pthread_t*>(&thread_handle_),
        &attr,
        &XThreadStartCallbackPthreads,
        this);
#else
    // TODO(benvanik): pthread_create_suspended_np on linux
    assert_always();
#endif  // OSX
  } else {
    result_code = pthread_create(
        reinterpret_cast<pthread_t*>(&thread_handle_),
        &attr,
        &XThreadStartCallbackPthreads,
        this);
  }

  pthread_attr_destroy(&attr);

  switch (result_code) {
    case 0:
      // Succeeded!
      return X_STATUS_SUCCESS;
    default:
    case EAGAIN:
      return X_STATUS_NO_MEMORY;
    case EINVAL:
    case EPERM:
      return X_STATUS_INVALID_PARAMETER;
  }
}

void XThread::PlatformDestroy() {
  // No-op?
}

X_STATUS XThread::PlatformExit(int exit_code) {
  // NOTE: does not return.
  pthread_exit((void*)exit_code);
  return X_STATUS_SUCCESS;
}

#endif  // WIN32

void XThread::Execute() {
  // If a XapiThreadStartup value is present, we use that as a trampoline.
  // Otherwise, we are a raw thread.
  if (creation_params_.xapi_thread_startup) {
    uint64_t args[] = {
      creation_params_.start_address,
      creation_params_.start_context
    };
    kernel_state()->processor()->Execute(
        thread_state_,
        creation_params_.xapi_thread_startup, args, XECOUNT(args));
  } else {
    // Run user code.
    uint64_t args[] = {
      creation_params_.start_context
    };
    int exit_code = (int)kernel_state()->processor()->Execute(
        thread_state_,
        creation_params_.start_address, args, XECOUNT(args));
    // If we got here it means the execute completed without an exit being called.
    // Treat the return code as an implicit exit code.
    Exit(exit_code);
  }
}

void XThread::EnterCriticalRegion() {
  // Global critical region. This isn't right, but is easy.
  xe_mutex_lock(critical_region_);
}

void XThread::LeaveCriticalRegion() {
  xe_mutex_unlock(critical_region_);
}

uint32_t XThread::RaiseIrql(uint32_t new_irql) {
  return irql_.exchange(new_irql);
}

void XThread::LowerIrql(uint32_t new_irql) {
  irql_ = new_irql;
}

void XThread::LockApc() {
  xe_mutex_lock(apc_lock_);
}

void XThread::UnlockApc() {
  bool needs_apc = apc_list_->HasPending();
  xe_mutex_unlock(apc_lock_);
  if (needs_apc) {
    QueueUserAPC(reinterpret_cast<PAPCFUNC>(DeliverAPCs),
                 thread_handle_,
                 reinterpret_cast<ULONG_PTR>(this));
  }
}

void XThread::DeliverAPCs(void* data) {
  // http://www.drdobbs.com/inside-nts-asynchronous-procedure-call/184416590?pgno=1
  // http://www.drdobbs.com/inside-nts-asynchronous-procedure-call/184416590?pgno=7
  XThread* thread = reinterpret_cast<XThread*>(data);
  auto membase = thread->memory()->membase();
  auto processor = thread->kernel_state()->processor();
  auto apc_list = thread->apc_list();
  thread->LockApc();
  while (apc_list->HasPending()) {
    // Get APC entry (offset for LIST_ENTRY offset) and cache what we need.
    // Calling the routine may delete the memory/overwrite it.
    uint32_t apc_address = apc_list->Shift() - 8;
    uint8_t* apc_ptr = membase + apc_address;
    uint32_t kernel_routine = poly::load_and_swap<uint32_t>(apc_ptr + 16);
    uint32_t normal_routine = poly::load_and_swap<uint32_t>(apc_ptr + 24);
    uint32_t normal_context = poly::load_and_swap<uint32_t>(apc_ptr + 28);
    uint32_t system_arg1 = poly::load_and_swap<uint32_t>(apc_ptr + 32);
    uint32_t system_arg2 = poly::load_and_swap<uint32_t>(apc_ptr + 36);

    // Mark as uninserted so that it can be reinserted again by the routine.
    uint32_t old_flags = poly::load_and_swap<uint32_t>(apc_ptr + 40);
    poly::store_and_swap<uint32_t>(apc_ptr + 40, old_flags & ~0xFF00);

    // Call kernel routine.
    // The routine can modify all of its arguments before passing it on.
    // Since we need to give guest accessible pointers over, we copy things
    // into and out of scratch.
    uint8_t* scratch_ptr = membase + thread->scratch_address_;
    poly::store_and_swap<uint32_t>(scratch_ptr + 0, normal_routine);
    poly::store_and_swap<uint32_t>(scratch_ptr + 4, normal_context);
    poly::store_and_swap<uint32_t>(scratch_ptr + 8, system_arg1);
    poly::store_and_swap<uint32_t>(scratch_ptr + 12, system_arg2);
    // kernel_routine(apc_address, &normal_routine, &normal_context, &system_arg1, &system_arg2)
    uint64_t kernel_args[] = {
      apc_address,
      thread->scratch_address_ + 0,
      thread->scratch_address_ + 4,
      thread->scratch_address_ + 8,
      thread->scratch_address_ + 12,
    };
    processor->ExecuteInterrupt(
        0, kernel_routine, kernel_args, XECOUNT(kernel_args));
    normal_routine = poly::load_and_swap<uint32_t>(scratch_ptr + 0);
    normal_context = poly::load_and_swap<uint32_t>(scratch_ptr + 4);
    system_arg1 = poly::load_and_swap<uint32_t>(scratch_ptr + 8);
    system_arg2 = poly::load_and_swap<uint32_t>(scratch_ptr + 12);

    // Call the normal routine. Note that it may have been killed by the kernel
    // routine.
    if (normal_routine) {
      thread->UnlockApc();
      // normal_routine(normal_context, system_arg1, system_arg2)
      uint64_t normal_args[] = { normal_context, system_arg1, system_arg2 };
      processor->ExecuteInterrupt(
          0, normal_routine, normal_args, XECOUNT(normal_args));
      thread->LockApc();
    }
  }
  thread->UnlockApc();
}

void XThread::RundownAPCs() {
  auto membase = memory()->membase();
  LockApc();
  while (apc_list_->HasPending()) {
    // Get APC entry (offset for LIST_ENTRY offset) and cache what we need.
    // Calling the routine may delete the memory/overwrite it.
    uint32_t apc_address = apc_list_->Shift() - 8;
    uint8_t* apc_ptr = membase + apc_address;
    uint32_t rundown_routine = poly::load_and_swap<uint32_t>(apc_ptr + 20);

    // Mark as uninserted so that it can be reinserted again by the routine.
    uint32_t old_flags = poly::load_and_swap<uint32_t>(apc_ptr + 40);
    poly::store_and_swap<uint32_t>(apc_ptr + 40, old_flags & ~0xFF00);

    // Call the rundown routine.
    if (rundown_routine) {
      // rundown_routine(apc)
      uint64_t args[] = { apc_address };
      kernel_state()->processor()->ExecuteInterrupt(
          0, rundown_routine, args, XECOUNT(args));
    }
  }
  UnlockApc();
}

int32_t XThread::QueryPriority() {
  return GetThreadPriority(thread_handle_);
}

void XThread::SetPriority(int32_t increment) {
  SetThreadPriority(thread_handle_, increment);
}

void XThread::SetAffinity(uint32_t affinity) {
  // TODO(benvanik): implement.
  XELOGW("KeSetAffinityThread not implemented");
}

X_STATUS XThread::Resume(uint32_t* out_suspend_count) {
  DWORD result = ResumeThread(thread_handle_);
  if (result >= 0) {
    if (out_suspend_count) {
      *out_suspend_count = result;
    }
    return X_STATUS_SUCCESS;
  } else {
    return X_STATUS_UNSUCCESSFUL;
  }
}

X_STATUS XThread::Suspend(uint32_t* out_suspend_count) {
  DWORD result = SuspendThread(thread_handle_);
  if (result >= 0) {
    if (out_suspend_count) {
      *out_suspend_count = result;
    }
    return X_STATUS_SUCCESS;
  } else {
    return X_STATUS_UNSUCCESSFUL;
  }
}

X_STATUS XThread::Delay(
  uint32_t processor_mode, uint32_t alertable, uint64_t interval) {
  int64_t timeout_ticks = interval;
  DWORD timeout_ms;
  if (timeout_ticks > 0) {
    // Absolute time, based on January 1, 1601.
    // TODO(benvanik): convert time to relative time.
    assert_always();
    timeout_ms = 0;
  } else if (timeout_ticks < 0) {
    // Relative time.
    timeout_ms = (DWORD)(-timeout_ticks / 10000); // Ticks -> MS
  } else {
    timeout_ms = 0;
  }
  DWORD result = SleepEx(timeout_ms, alertable ? TRUE : FALSE);
  switch (result) {
  case 0:
    return X_STATUS_SUCCESS;
  case WAIT_IO_COMPLETION:
    return X_STATUS_USER_APC;
  default:
    return X_STATUS_ALERTED;
  }
}

void* XThread::GetWaitHandle() {
  return event_->GetWaitHandle();
}