/*

 * Copyright (c) 1999, 2017, Oracle and/or its affiliates. All rights reserved.

 * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.

 *

 * This code is free software; you can redistribute it and/or modify it

 * under the terms of the GNU General Public License version 2 only, as

 * published by the Free Software Foundation.

 *

 * This code is distributed in the hope that it will be useful, but WITHOUT

 * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or

 * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License

 * version 2 for more details (a copy is included in the LICENSE file that

 * accompanied this code).

 *

 * You should have received a copy of the GNU General Public License version

 * 2 along with this work; if not, write to the Free Software Foundation,

 * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA.

 *

 * Please contact Oracle, 500 Oracle Parkway, Redwood Shores, CA 94065 USA

 * or visit www.oracle.com if you need additional information or have any

 * questions.

 *

 */

// no precompiled headers

#include "jvm.h"

#include "asm/macroAssembler.hpp"

#include "classfile/classLoader.hpp"

#include "classfile/systemDictionary.hpp"

#include "classfile/vmSymbols.hpp"

#include "code/icBuffer.hpp"

#include "code/vtableStubs.hpp"

#include "interpreter/interpreter.hpp"

#include "memory/allocation.inline.hpp"

#include "memory/resourceArea.hpp"

#include "nativeInst_x86.hpp"

#include "os_share_windows.hpp"

#include "prims/jniFastGetField.hpp"

#include "prims/jvm_misc.hpp"

#include "runtime/arguments.hpp"

#include "runtime/extendedPC.hpp"

#include "runtime/frame.inline.hpp"

#include "runtime/interfaceSupport.inline.hpp"

#include "runtime/java.hpp"

#include "runtime/javaCalls.hpp"

#include "runtime/mutexLocker.hpp"

#include "runtime/osThread.hpp"

#include "runtime/sharedRuntime.hpp"

#include "runtime/stubRoutines.hpp"

#include "runtime/thread.inline.hpp"

#include "runtime/timer.hpp"

#include "symbolengine.hpp"

#include "unwind_windows_x86.hpp"

#include "utilities/events.hpp"

#include "utilities/vmError.hpp"

#include "windbghelp.hpp"

#undef REG_SP

#undef REG_FP

#undef REG_PC

#ifdef AMD64

#define REG_SP Rsp

#define REG_FP Rbp

#define REG_PC Rip

#else

#define REG_SP Esp

#define REG_FP Ebp

#define REG_PC Eip

#endif // AMD64

extern LONG WINAPI topLevelExceptionFilter(_EXCEPTION_POINTERS* );

// Install a win32 structured exception handler around thread.

void os::os_exception_wrapper(java_call_t f, JavaValue* value, const methodHandle& method, JavaCallArguments* args, Thread* thread) {

  __try {

#ifndef AMD64

    // We store the current thread in this wrapperthread location

    // and determine how far away this address is from the structured

    // execption pointer that FS:[0] points to.  This get_thread

    // code can then get the thread pointer via FS.

    //

    // Warning:  This routine must NEVER be inlined since we'd end up with

    //           multiple offsets.

    //

    volatile Thread* wrapperthread = thread;

    if (os::win32::get_thread_ptr_offset() == 0) {

      int thread_ptr_offset;

      __asm {

        lea eax, dword ptr wrapperthread;

        sub eax, dword ptr FS:[0H];

        mov thread_ptr_offset, eax

      };

      os::win32::set_thread_ptr_offset(thread_ptr_offset);

    }

#ifdef ASSERT

    // Verify that the offset hasn't changed since we initally captured

    // it. This might happen if we accidentally ended up with an

    // inlined version of this routine.

    else {

      int test_thread_ptr_offset;

      __asm {

        lea eax, dword ptr wrapperthread;

        sub eax, dword ptr FS:[0H];

        mov test_thread_ptr_offset, eax

      };

      assert(test_thread_ptr_offset == os::win32::get_thread_ptr_offset(),

             "thread pointer offset from SEH changed");

    }

#endif // ASSERT

#endif // !AMD64

    f(value, method, args, thread);

  } __except(topLevelExceptionFilter((_EXCEPTION_POINTERS*)_exception_info())) {

      // Nothing to do.

  }

}

#ifdef AMD64

// This is the language specific handler for exceptions

// originating from dynamically generated code.

// We call the standard structured exception handler

// We only expect Continued Execution since we cannot unwind

// from generated code.

LONG HandleExceptionFromCodeCache(

  IN PEXCEPTION_RECORD ExceptionRecord,

  IN ULONG64 EstablisherFrame,

  IN OUT PCONTEXT ContextRecord,

  IN OUT PDISPATCHER_CONTEXT DispatcherContext) {

  EXCEPTION_POINTERS ep;

  LONG result;

  ep.ExceptionRecord = ExceptionRecord;

  ep.ContextRecord = ContextRecord;

  result = topLevelExceptionFilter(&ep);

  // We better only get a CONTINUE_EXECUTION from our handler

  // since we don't have unwind information registered.

  guarantee( result == EXCEPTION_CONTINUE_EXECUTION,

             "Unexpected result from topLevelExceptionFilter");

  return(ExceptionContinueExecution);

}

// Structure containing the Windows Data Structures required

// to register our Code Cache exception handler.

// We put these in the CodeCache since the API requires

// all addresses in these structures are relative to the Code

// area registered with RtlAddFunctionTable.

typedef struct {

  char ExceptionHandlerInstr[16];  // jmp HandleExceptionFromCodeCache

  RUNTIME_FUNCTION rt;

  UNWIND_INFO_EH_ONLY unw;

} DynamicCodeData, *pDynamicCodeData;

#endif // AMD64

//

// Register our CodeCache area with the OS so it will dispatch exceptions

// to our topLevelExceptionFilter when we take an exception in our

// dynamically generated code.

//

// Arguments:  low and high are the address of the full reserved

// codeCache area

//

bool os::register_code_area(char *low, char *high) {

#ifdef AMD64

  ResourceMark rm;

  pDynamicCodeData pDCD;

  PRUNTIME_FUNCTION prt;

  PUNWIND_INFO_EH_ONLY punwind;

  BufferBlob* blob = BufferBlob::create("CodeCache Exception Handler", sizeof(DynamicCodeData));

  CodeBuffer cb(blob);

  MacroAssembler* masm = new MacroAssembler(&cb);

  pDCD = (pDynamicCodeData) masm->pc();

  masm->jump(ExternalAddress((address)&HandleExceptionFromCodeCache));

  masm->flush();

  // Create an Unwind Structure specifying no unwind info

  // other than an Exception Handler

  punwind = &pDCD->unw;

  punwind->Version = 1;

  punwind->Flags = UNW_FLAG_EHANDLER;

  punwind->SizeOfProlog = 0;

  punwind->CountOfCodes = 0;

  punwind->FrameRegister = 0;

  punwind->FrameOffset = 0;

  punwind->ExceptionHandler = (char *)(&(pDCD->ExceptionHandlerInstr[0])) -

                              (char*)low;

  punwind->ExceptionData[0] = 0;

  // This structure describes the covered dynamic code area.

  // Addresses are relative to the beginning on the code cache area

  prt = &pDCD->rt;

  prt->BeginAddress = 0;

  prt->EndAddress = (ULONG)(high - low);

  prt->UnwindData = ((char *)punwind - low);

  guarantee(RtlAddFunctionTable(prt, 1, (ULONGLONG)low),

            "Failed to register Dynamic Code Exception Handler with RtlAddFunctionTable");

#endif // AMD64

  return true;

}

// Atomics and Stub Functions

typedef int32_t   xchg_func_t            (int32_t,  volatile int32_t*);

typedef int64_t   xchg_long_func_t       (int64_t,  volatile int64_t*);

typedef int32_t   cmpxchg_func_t         (int32_t,  volatile int32_t*, int32_t);

typedef int8_t    cmpxchg_byte_func_t    (int8_t,   volatile int8_t*,  int8_t);

typedef int64_t   cmpxchg_long_func_t    (int64_t,  volatile int64_t*, int64_t);

typedef int32_t   add_func_t             (int32_t,  volatile int32_t*);

typedef int64_t   add_long_func_t        (int64_t,  volatile int64_t*);

#ifdef AMD64

int32_t os::atomic_xchg_bootstrap(int32_t exchange_value, volatile int32_t* dest) {

  // try to use the stub:

  xchg_func_t* func = CAST_TO_FN_PTR(xchg_func_t*, StubRoutines::atomic_xchg_entry());

  if (func != NULL) {

    os::atomic_xchg_func = func;

    return (*func)(exchange_value, dest);

  }

  assert(Threads::number_of_threads() == 0, "for bootstrap only");

  int32_t old_value = *dest;

  *dest = exchange_value;

  return old_value;

}

int64_t os::atomic_xchg_long_bootstrap(int64_t exchange_value, volatile int64_t* dest) {

  // try to use the stub:

  xchg_long_func_t* func = CAST_TO_FN_PTR(xchg_long_func_t*, StubRoutines::atomic_xchg_long_entry());

  if (func != NULL) {

    os::atomic_xchg_long_func = func;

    return (*func)(exchange_value, dest);

  }

  assert(Threads::number_of_threads() == 0, "for bootstrap only");

  int64_t old_value = *dest;

  *dest = exchange_value;

  return old_value;

}

int32_t os::atomic_cmpxchg_bootstrap(int32_t exchange_value, volatile int32_t* dest, int32_t compare_value) {

  // try to use the stub:

  cmpxchg_func_t* func = CAST_TO_FN_PTR(cmpxchg_func_t*, StubRoutines::atomic_cmpxchg_entry());

  if (func != NULL) {

    os::atomic_cmpxchg_func = func;

    return (*func)(exchange_value, dest, compare_value);

  }

  assert(Threads::number_of_threads() == 0, "for bootstrap only");

  int32_t old_value = *dest;

  if (old_value == compare_value)

    *dest = exchange_value;

  return old_value;

}

int8_t os::atomic_cmpxchg_byte_bootstrap(int8_t exchange_value, volatile int8_t* dest, int8_t compare_value) {

  // try to use the stub:

  cmpxchg_byte_func_t* func = CAST_TO_FN_PTR(cmpxchg_byte_func_t*, StubRoutines::atomic_cmpxchg_byte_entry());

  if (func != NULL) {

    os::atomic_cmpxchg_byte_func = func;

    return (*func)(exchange_value, dest, compare_value);

  }

  assert(Threads::number_of_threads() == 0, "for bootstrap only");

  int8_t old_value = *dest;

  if (old_value == compare_value)

    *dest = exchange_value;

  return old_value;

}

#endif // AMD64

int64_t os::atomic_cmpxchg_long_bootstrap(int64_t exchange_value, volatile int64_t* dest, int64_t compare_value) {

  // try to use the stub:

  cmpxchg_long_func_t* func = CAST_TO_FN_PTR(cmpxchg_long_func_t*, StubRoutines::atomic_cmpxchg_long_entry());

  if (func != NULL) {

    os::atomic_cmpxchg_long_func = func;

    return (*func)(exchange_value, dest, compare_value);

  }

  assert(Threads::number_of_threads() == 0, "for bootstrap only");

  int64_t old_value = *dest;

  if (old_value == compare_value)

    *dest = exchange_value;

  return old_value;

}

#ifdef AMD64

int32_t os::atomic_add_bootstrap(int32_t add_value, volatile int32_t* dest) {

  // try to use the stub:

  add_func_t* func = CAST_TO_FN_PTR(add_func_t*, StubRoutines::atomic_add_entry());

  if (func != NULL) {

    os::atomic_add_func = func;

    return (*func)(add_value, dest);

  }

  assert(Threads::number_of_threads() == 0, "for bootstrap only");

  return (*dest) += add_value;

}

int64_t os::atomic_add_long_bootstrap(int64_t add_value, volatile int64_t* dest) {

  // try to use the stub:

  add_long_func_t* func = CAST_TO_FN_PTR(add_long_func_t*, StubRoutines::atomic_add_long_entry());

  if (func != NULL) {

    os::atomic_add_long_func = func;

    return (*func)(add_value, dest);

  }

  assert(Threads::number_of_threads() == 0, "for bootstrap only");

  return (*dest) += add_value;

}

xchg_func_t*         os::atomic_xchg_func         = os::atomic_xchg_bootstrap;

xchg_long_func_t*    os::atomic_xchg_long_func    = os::atomic_xchg_long_bootstrap;

cmpxchg_func_t*      os::atomic_cmpxchg_func      = os::atomic_cmpxchg_bootstrap;

cmpxchg_byte_func_t* os::atomic_cmpxchg_byte_func = os::atomic_cmpxchg_byte_bootstrap;

add_func_t*          os::atomic_add_func          = os::atomic_add_bootstrap;

add_long_func_t*     os::atomic_add_long_func     = os::atomic_add_long_bootstrap;

#endif // AMD64

cmpxchg_long_func_t* os::atomic_cmpxchg_long_func = os::atomic_cmpxchg_long_bootstrap;

#ifdef AMD64

/*

 * Windows/x64 does not use stack frames the way expected by Java:

 * [1] in most cases, there is no frame pointer. All locals are addressed via RSP

 * [2] in rare cases, when alloca() is used, a frame pointer is used, but this may

 *     not be RBP.

 * See http://msdn.microsoft.com/en-us/library/ew5tede7.aspx

 *

 * So it's not possible to print the native stack using the

 *     while (...) {...  fr = os::get_sender_for_C_frame(&fr); }

 * loop in vmError.cpp. We need to roll our own loop.

 */

bool os::platform_print_native_stack(outputStream* st, const void* context,

                                     char *buf, int buf_size)

{

  CONTEXT ctx;

  if (context != NULL) {

    memcpy(&ctx, context, sizeof(ctx));

  } else {

    RtlCaptureContext(&ctx);

  }

  st->print_cr("Native frames: (J=compiled Java code, j=interpreted, Vv=VM code, C=native code)");

  STACKFRAME stk;

  memset(&stk, 0, sizeof(stk));

  stk.AddrStack.Offset    = ctx.Rsp;

  stk.AddrStack.Mode      = AddrModeFlat;

  stk.AddrFrame.Offset    = ctx.Rbp;

  stk.AddrFrame.Mode      = AddrModeFlat;

  stk.AddrPC.Offset       = ctx.Rip;

  stk.AddrPC.Mode         = AddrModeFlat;

  int count = 0;

  address lastpc = 0;

  while (count++ < StackPrintLimit) {

    intptr_t* sp = (intptr_t*)stk.AddrStack.Offset;

    intptr_t* fp = (intptr_t*)stk.AddrFrame.Offset; // NOT necessarily the same as ctx.Rbp!

    address pc = (address)stk.AddrPC.Offset;

    if (pc != NULL) {

      if (count == 2 && lastpc == pc) {

        // Skip it -- StackWalk64() may return the same PC

        // (but different SP) on the first try.

      } else {

        // Don't try to create a frame(sp, fp, pc) -- on WinX64, stk.AddrFrame

        // may not contain what Java expects, and may cause the frame() constructor

        // to crash. Let's just print out the symbolic address.

        frame::print_C_frame(st, buf, buf_size, pc);

        // print source file and line, if available

        char buf[128];

        int line_no;

        if (SymbolEngine::get_source_info(pc, buf, sizeof(buf), &line_no)) {

          st->print("  (%s:%d)", buf, line_no);

        }

        st->cr();

      }

      lastpc = pc;

    }

    PVOID p = WindowsDbgHelp::symFunctionTableAccess64(GetCurrentProcess(), stk.AddrPC.Offset);

    if (!p) {

      // StackWalk64() can't handle this PC. Calling StackWalk64 again may cause crash.

      break;

    }

    BOOL result = WindowsDbgHelp::stackWalk64(

        IMAGE_FILE_MACHINE_AMD64,  // __in      DWORD MachineType,

        GetCurrentProcess(),       // __in      HANDLE hProcess,

        GetCurrentThread(),        // __in      HANDLE hThread,

        &stk,                      // __inout   LP STACKFRAME64 StackFrame,

        &ctx);                     // __inout   PVOID ContextRecord,

    if (!result) {

      break;

    }

  }

  if (count > StackPrintLimit) {

    st->print_cr("...<more frames>...");

  }

  st->cr();

  return true;

}

#endif // AMD64

ExtendedPC os::fetch_frame_from_context(const void* ucVoid,

                    intptr_t** ret_sp, intptr_t** ret_fp) {

  ExtendedPC  epc;

  CONTEXT* uc = (CONTEXT*)ucVoid;

  if (uc != NULL) {

    epc = ExtendedPC((address)uc->REG_PC);

    if (ret_sp) *ret_sp = (intptr_t*)uc->REG_SP;

    if (ret_fp) *ret_fp = (intptr_t*)uc->REG_FP;

  } else {

    // construct empty ExtendedPC for return value checking

    epc = ExtendedPC(NULL);

    if (ret_sp) *ret_sp = (intptr_t *)NULL;

    if (ret_fp) *ret_fp = (intptr_t *)NULL;

  }

  return epc;

}

frame os::fetch_frame_from_context(const void* ucVoid) {

  intptr_t* sp;

  intptr_t* fp;

  ExtendedPC epc = fetch_frame_from_context(ucVoid, &sp, &fp);

  return frame(sp, fp, epc.pc());

}

// VC++ does not save frame pointer on stack in optimized build. It

// can be turned off by /Oy-. If we really want to walk C frames,

// we can use the StackWalk() API.

frame os::get_sender_for_C_frame(frame* fr) {

  return frame(fr->sender_sp(), fr->link(), fr->sender_pc());

}

#ifndef AMD64

// Ignore "C4172: returning address of local variable or temporary" on 32bit

PRAGMA_DIAG_PUSH

PRAGMA_DISABLE_MSVC_WARNING(4172)

// Returns an estimate of the current stack pointer. Result must be guaranteed

// to point into the calling threads stack, and be no lower than the current

// stack pointer.

address os::current_stack_pointer() {

  int dummy;

  address sp = (address)&dummy;

  return sp;

}

PRAGMA_DIAG_POP

#else

// Returns the current stack pointer. Accurate value needed for

// os::verify_stack_alignment().

address os::current_stack_pointer() {

  typedef address get_sp_func();

  get_sp_func* func = CAST_TO_FN_PTR(get_sp_func*,

                                     StubRoutines::x86::get_previous_sp_entry());

  return (*func)();

}

#endif

#ifndef AMD64

intptr_t* _get_previous_fp() {

  intptr_t **frameptr;

  __asm {

    mov frameptr, ebp

  };

  // ebp (frameptr) is for this frame (_get_previous_fp). We want the ebp for the

  // caller of os::current_frame*(), so go up two frames. However, for

  // optimized builds, _get_previous_fp() will be inlined, so only go

  // up 1 frame in that case.

#ifdef _NMT_NOINLINE_

  return **(intptr_t***)frameptr;

#else

  return *frameptr;

#endif

}

#endif // !AMD64