delve/pkg/proc/native/proc_windows.go
Alessandro Arzilli 0843376018
proc/*: remove proc.Thread.Blocked, refactor memory access (#2206)
On linux we can not read memory if the thread we use to do it is
occupied doing certain system calls. The exact conditions when this
happens have never been clear.

This problem was worked around by using the Blocked method which
recognized the most common circumstances where this would happen.

However this is a hack: Blocked returning true doesn't mean that the
problem will manifest and Blocked returning false doesn't necessarily
mean the problem will not manifest. A side effect of this is issue
#2151 where sometimes we can't read the memory of a thread and find its
associated goroutine.

This commit fixes this problem by always reading memory using a thread
we know to be good for this, specifically the one returned by
ContinueOnce. In particular the changes are as follows:

1. Remove (ProcessInternal).CurrentThread and
(ProcessInternal).SetCurrentThread, the "current thread" becomes a
field of Target, CurrentThread becomes a (*Target) method and
(*Target).SwitchThread basically just sets a field Target.

2. The backends keep track of their own internal idea of what the
current thread is, to use it to read memory, this is the thread they
return from ContinueOnce as trapthread

3. The current thread in the backend and the current thread in Target
only ever get synchronized in two places: when the backend creates a
Target object the currentThread field of Target is initialized with the
backend's current thread and when (*Target).Restart gets called (when a
recording is rewound the currentThread used by Target might not exist
anymore).

4. We remove the MemoryReadWriter interface embedded in Thread and
instead add a Memory method to Process that returns a MemoryReadWriter.
The  backends will return something here that will read memory using
the current thread saved by the backend.

5. The Thread.Blocked method is removed

One possible problem with this change is processes that have threads
with different memory maps. As far as I can determine this could happen
on old versions of linux but this option was removed in linux 2.5.

Fixes #2151
2020-11-09 11:28:40 -08:00

490 lines
12 KiB
Go

package native
import (
"fmt"
"os"
"path/filepath"
"syscall"
"unsafe"
sys "golang.org/x/sys/windows"
"github.com/go-delve/delve/pkg/proc"
)
// osProcessDetails holds Windows specific information.
type osProcessDetails struct {
hProcess syscall.Handle
breakThread int
entryPoint uint64
running bool
}
// Launch creates and begins debugging a new process.
func Launch(cmd []string, wd string, flags proc.LaunchFlags, _ []string, _ string, redirects [3]string) (*proc.Target, error) {
argv0Go, err := filepath.Abs(cmd[0])
if err != nil {
return nil, err
}
env := proc.DisableAsyncPreemptEnv()
stdin, stdout, stderr, closefn, err := openRedirects(redirects, true)
if err != nil {
return nil, err
}
var p *os.Process
dbp := newProcess(0)
dbp.execPtraceFunc(func() {
attr := &os.ProcAttr{
Dir: wd,
Files: []*os.File{stdin, stdout, stderr},
Sys: &syscall.SysProcAttr{
CreationFlags: _DEBUG_ONLY_THIS_PROCESS,
},
Env: env,
}
p, err = os.StartProcess(argv0Go, cmd, attr)
})
closefn()
if err != nil {
return nil, err
}
defer p.Release()
dbp.pid = p.Pid
dbp.childProcess = true
tgt, err := dbp.initialize(argv0Go, []string{})
if err != nil {
dbp.Detach(true)
return nil, err
}
return tgt, nil
}
func initialize(dbp *nativeProcess) error {
// It should not actually be possible for the
// call to waitForDebugEvent to fail, since Windows
// will always fire a CREATE_PROCESS_DEBUG_EVENT event
// immediately after launching under DEBUG_ONLY_THIS_PROCESS.
// Attaching with DebugActiveProcess has similar effect.
var err error
var tid, exitCode int
dbp.execPtraceFunc(func() {
tid, exitCode, err = dbp.waitForDebugEvent(waitBlocking)
})
if err != nil {
return err
}
if tid == 0 {
dbp.postExit()
return proc.ErrProcessExited{Pid: dbp.pid, Status: exitCode}
}
// Suspend all threads so that the call to _ContinueDebugEvent will
// not resume the target.
for _, thread := range dbp.threads {
_, err := _SuspendThread(thread.os.hThread)
if err != nil {
return err
}
}
dbp.execPtraceFunc(func() {
err = _ContinueDebugEvent(uint32(dbp.pid), uint32(dbp.os.breakThread), _DBG_CONTINUE)
})
return err
}
// findExePath searches for process pid, and returns its executable path.
func findExePath(pid int) (string, error) {
// Original code suggested different approach (see below).
// Maybe it could be useful in the future.
//
// Find executable path from PID/handle on Windows:
// https://msdn.microsoft.com/en-us/library/aa366789(VS.85).aspx
p, err := syscall.OpenProcess(syscall.PROCESS_QUERY_INFORMATION, false, uint32(pid))
if err != nil {
return "", err
}
defer syscall.CloseHandle(p)
n := uint32(128)
for {
buf := make([]uint16, int(n))
err = _QueryFullProcessImageName(p, 0, &buf[0], &n)
switch err {
case syscall.ERROR_INSUFFICIENT_BUFFER:
// try bigger buffer
n *= 2
// but stop if it gets too big
if n > 10000 {
return "", err
}
case nil:
return syscall.UTF16ToString(buf[:n]), nil
default:
return "", err
}
}
}
// Attach to an existing process with the given PID.
func Attach(pid int, _ []string) (*proc.Target, error) {
dbp := newProcess(pid)
var err error
dbp.execPtraceFunc(func() {
// TODO: Probably should have SeDebugPrivilege before starting here.
err = _DebugActiveProcess(uint32(pid))
})
if err != nil {
return nil, err
}
exepath, err := findExePath(pid)
if err != nil {
return nil, err
}
tgt, err := dbp.initialize(exepath, []string{})
if err != nil {
dbp.Detach(true)
return nil, err
}
return tgt, nil
}
// kill kills the process.
func (dbp *nativeProcess) kill() error {
if dbp.exited {
return nil
}
p, err := os.FindProcess(dbp.pid)
if err != nil {
return err
}
defer p.Release()
// TODO: Should not have to ignore failures here,
// but some tests appear to Kill twice causing
// this to fail on second attempt.
_ = syscall.TerminateProcess(dbp.os.hProcess, 1)
dbp.execPtraceFunc(func() {
dbp.waitForDebugEvent(waitBlocking | waitDontHandleExceptions)
})
p.Wait()
dbp.postExit()
return nil
}
func (dbp *nativeProcess) requestManualStop() error {
if !dbp.os.running {
return nil
}
dbp.os.running = false
return _DebugBreakProcess(dbp.os.hProcess)
}
func (dbp *nativeProcess) updateThreadList() error {
// We ignore this request since threads are being
// tracked as they are created/killed in waitForDebugEvent.
return nil
}
func (dbp *nativeProcess) addThread(hThread syscall.Handle, threadID int, attach, suspendNewThreads bool) (*nativeThread, error) {
if thread, ok := dbp.threads[threadID]; ok {
return thread, nil
}
thread := &nativeThread{
ID: threadID,
dbp: dbp,
os: new(osSpecificDetails),
}
thread.os.hThread = hThread
dbp.threads[threadID] = thread
if dbp.memthread == nil {
dbp.memthread = dbp.threads[threadID]
}
if suspendNewThreads {
_, err := _SuspendThread(thread.os.hThread)
if err != nil {
return nil, err
}
}
return thread, nil
}
func findExecutable(path string, pid int) string {
return path
}
type waitForDebugEventFlags int
const (
waitBlocking waitForDebugEventFlags = 1 << iota
waitSuspendNewThreads
waitDontHandleExceptions
)
const _MS_VC_EXCEPTION = 0x406D1388 // part of VisualC protocol to set thread names
func (dbp *nativeProcess) waitForDebugEvent(flags waitForDebugEventFlags) (threadID, exitCode int, err error) {
var debugEvent _DEBUG_EVENT
shouldExit := false
for {
continueStatus := uint32(_DBG_CONTINUE)
var milliseconds uint32 = 0
if flags&waitBlocking != 0 {
milliseconds = syscall.INFINITE
}
// Wait for a debug event...
err := _WaitForDebugEvent(&debugEvent, milliseconds)
if err != nil {
return 0, 0, err
}
// ... handle each event kind ...
unionPtr := unsafe.Pointer(&debugEvent.U[0])
switch debugEvent.DebugEventCode {
case _CREATE_PROCESS_DEBUG_EVENT:
debugInfo := (*_CREATE_PROCESS_DEBUG_INFO)(unionPtr)
hFile := debugInfo.File
if hFile != 0 && hFile != syscall.InvalidHandle {
err = syscall.CloseHandle(hFile)
if err != nil {
return 0, 0, err
}
}
dbp.os.entryPoint = uint64(debugInfo.BaseOfImage)
dbp.os.hProcess = debugInfo.Process
_, err = dbp.addThread(debugInfo.Thread, int(debugEvent.ThreadId), false, flags&waitSuspendNewThreads != 0)
if err != nil {
return 0, 0, err
}
break
case _CREATE_THREAD_DEBUG_EVENT:
debugInfo := (*_CREATE_THREAD_DEBUG_INFO)(unionPtr)
_, err = dbp.addThread(debugInfo.Thread, int(debugEvent.ThreadId), false, flags&waitSuspendNewThreads != 0)
if err != nil {
return 0, 0, err
}
break
case _EXIT_THREAD_DEBUG_EVENT:
delete(dbp.threads, int(debugEvent.ThreadId))
break
case _OUTPUT_DEBUG_STRING_EVENT:
//TODO: Handle debug output strings
break
case _LOAD_DLL_DEBUG_EVENT:
debugInfo := (*_LOAD_DLL_DEBUG_INFO)(unionPtr)
hFile := debugInfo.File
if hFile != 0 && hFile != syscall.InvalidHandle {
err = syscall.CloseHandle(hFile)
if err != nil {
return 0, 0, err
}
}
break
case _UNLOAD_DLL_DEBUG_EVENT:
break
case _RIP_EVENT:
break
case _EXCEPTION_DEBUG_EVENT:
if flags&waitDontHandleExceptions != 0 {
continueStatus = _DBG_EXCEPTION_NOT_HANDLED
break
}
exception := (*_EXCEPTION_DEBUG_INFO)(unionPtr)
tid := int(debugEvent.ThreadId)
switch code := exception.ExceptionRecord.ExceptionCode; code {
case _EXCEPTION_BREAKPOINT:
// check if the exception address really is a breakpoint instruction, if
// it isn't we already removed that breakpoint and we can't deal with
// this exception anymore.
atbp := true
if thread, found := dbp.threads[tid]; found {
data := make([]byte, dbp.bi.Arch.BreakpointSize())
if _, err := thread.ReadMemory(data, uint64(exception.ExceptionRecord.ExceptionAddress)); err == nil {
instr := dbp.bi.Arch.BreakpointInstruction()
for i := range instr {
if data[i] != instr[i] {
atbp = false
break
}
}
}
if !atbp {
thread.SetPC(uint64(exception.ExceptionRecord.ExceptionAddress))
}
}
if atbp {
dbp.os.breakThread = tid
return tid, 0, nil
} else {
continueStatus = _DBG_CONTINUE
}
case _EXCEPTION_SINGLE_STEP:
dbp.os.breakThread = tid
return tid, 0, nil
case _MS_VC_EXCEPTION:
// This exception is sent to set the thread name in VisualC, we should
// mask it or it might crash the program.
continueStatus = _DBG_CONTINUE
default:
continueStatus = _DBG_EXCEPTION_NOT_HANDLED
}
case _EXIT_PROCESS_DEBUG_EVENT:
debugInfo := (*_EXIT_PROCESS_DEBUG_INFO)(unionPtr)
exitCode = int(debugInfo.ExitCode)
shouldExit = true
default:
return 0, 0, fmt.Errorf("unknown debug event code: %d", debugEvent.DebugEventCode)
}
// .. and then continue unless we received an event that indicated we should break into debugger.
err = _ContinueDebugEvent(debugEvent.ProcessId, debugEvent.ThreadId, continueStatus)
if err != nil {
return 0, 0, err
}
if shouldExit {
return 0, exitCode, nil
}
}
}
func (dbp *nativeProcess) trapWait(pid int) (*nativeThread, error) {
var err error
var tid, exitCode int
dbp.execPtraceFunc(func() {
tid, exitCode, err = dbp.waitForDebugEvent(waitBlocking)
})
if err != nil {
return nil, err
}
if tid == 0 {
dbp.postExit()
return nil, proc.ErrProcessExited{Pid: dbp.pid, Status: exitCode}
}
th := dbp.threads[tid]
return th, nil
}
func (dbp *nativeProcess) wait(pid, options int) (int, *sys.WaitStatus, error) {
return 0, nil, fmt.Errorf("not implemented: wait")
}
func (dbp *nativeProcess) exitGuard(err error) error {
return err
}
func (dbp *nativeProcess) resume() error {
for _, thread := range dbp.threads {
if thread.CurrentBreakpoint.Breakpoint != nil {
if err := thread.StepInstruction(); err != nil {
return err
}
thread.CurrentBreakpoint.Clear()
}
}
for _, thread := range dbp.threads {
_, err := _ResumeThread(thread.os.hThread)
if err != nil {
return err
}
}
dbp.os.running = true
return nil
}
// stop stops all running threads threads and sets breakpoints
func (dbp *nativeProcess) stop(trapthread *nativeThread) (*nativeThread, error) {
if dbp.exited {
return nil, &proc.ErrProcessExited{Pid: dbp.Pid()}
}
dbp.os.running = false
// While the debug event that stopped the target was being propagated
// other target threads could generate other debug events.
// After this function we need to know about all the threads
// stopped on a breakpoint. To do that we first suspend all target
// threads and then repeatedly call _ContinueDebugEvent followed by
// waitForDebugEvent in non-blocking mode.
// We need to explicitly call SuspendThread because otherwise the
// call to _ContinueDebugEvent will resume execution of some of the
// target threads.
err := trapthread.SetCurrentBreakpoint(true)
if err != nil {
return nil, err
}
for _, thread := range dbp.threads {
_, err := _SuspendThread(thread.os.hThread)
if err != nil {
return nil, err
}
}
for {
var err error
var tid int
dbp.execPtraceFunc(func() {
err = _ContinueDebugEvent(uint32(dbp.pid), uint32(dbp.os.breakThread), _DBG_CONTINUE)
if err == nil {
tid, _, _ = dbp.waitForDebugEvent(waitSuspendNewThreads)
}
})
if err != nil {
return nil, err
}
if tid == 0 {
break
}
err = dbp.threads[tid].SetCurrentBreakpoint(true)
if err != nil {
return nil, err
}
}
return trapthread, nil
}
func (dbp *nativeProcess) detach(kill bool) error {
if !kill {
//TODO(aarzilli): when debug.Target exist Detach should be moved to
// debug.Target and the call to RestoreAsyncPreempt should be moved there.
for _, thread := range dbp.threads {
_, err := _ResumeThread(thread.os.hThread)
if err != nil {
return err
}
}
}
return _DebugActiveProcessStop(uint32(dbp.pid))
}
func (dbp *nativeProcess) EntryPoint() (uint64, error) {
return dbp.os.entryPoint, nil
}
func killProcess(pid int) error {
p, err := os.FindProcess(pid)
if err != nil {
return err
}
defer p.Release()
return p.Kill()
}