delve/pkg/proc/native/proc.go
Alessandro Arzilli 1a9e38aa0c
proc,terminal: Implement reverse step, next and stepout (#1785)
* proc: move defer breakpoint code into a function

Moves the code that sets a breakpoint on the first deferred function,
used by both next and StepOut, to its function.

* proc: implement reverse step/next/stepout

When the direction of execution is reversed (on a recording) Step, Next and
StepOut will behave similarly to their forward version. However there are
some subtle interactions between their behavior, prologue skipping, deferred
calls and normal calls. Specifically:

- when stepping backwards we need to set a breakpoint on the first
  instruction after each CALL instruction, once this breakpoint is reached we
  need to execute a single StepInstruction operation to reverse step into the
  CALL.
- to insure that the prologue is skipped reverse next needs to check if it
  is on the first instruction after the prologue, and if it is behave like
  reverse stepout.
- there is no reason to set breakpoints on deferred calls when reverse
  nexting or reverse stepping out, they will never be hit.
- reverse step out should generally place its breakpoint on the CALL
  instruction that created the current stack frame (which will be the CALL
  instruction immediately preceding the instruction at the return address).
- reverse step out needs to treat panic calls and deferreturn calls
  specially.

* service,terminal: implement reverse step, next, stepout
2020-03-11 15:40:41 -07:00

352 lines
9.7 KiB
Go

package native
import (
"go/ast"
"runtime"
"sync"
"github.com/go-delve/delve/pkg/proc"
)
// Process represents all of the information the debugger
// is holding onto regarding the process we are debugging.
type Process struct {
bi *proc.BinaryInfo
pid int // Process Pid
// Breakpoint table, holds information on breakpoints.
// Maps instruction address to Breakpoint struct.
breakpoints proc.BreakpointMap
// List of threads mapped as such: pid -> *Thread
threads map[int]*Thread
// Active thread
currentThread *Thread
os *OSProcessDetails
firstStart bool
stopMu sync.Mutex
resumeChan chan<- struct{}
ptraceChan chan func()
ptraceDoneChan chan interface{}
childProcess bool // this process was launched, not attached to
manualStopRequested bool
exited, detached bool
}
var _ proc.ProcessInternal = &Process{}
// New returns an initialized Process struct. Before returning,
// it will also launch a goroutine in order to handle ptrace(2)
// functions. For more information, see the documentation on
// `handlePtraceFuncs`.
func New(pid int) *Process {
dbp := &Process{
pid: pid,
threads: make(map[int]*Thread),
breakpoints: proc.NewBreakpointMap(),
firstStart: true,
os: new(OSProcessDetails),
ptraceChan: make(chan func()),
ptraceDoneChan: make(chan interface{}),
bi: proc.NewBinaryInfo(runtime.GOOS, runtime.GOARCH),
}
go dbp.handlePtraceFuncs()
return dbp
}
// BinInfo will return the binary info struct associated with this process.
func (dbp *Process) BinInfo() *proc.BinaryInfo {
return dbp.bi
}
// Recorded always returns false for the native proc backend.
func (dbp *Process) Recorded() (bool, string) { return false, "" }
// Restart will always return an error in the native proc backend, only for
// recorded traces.
func (dbp *Process) Restart(string) error { return proc.ErrNotRecorded }
// ChangeDirection will always return an error in the native proc backend, only for
// recorded traces.
func (dbp *Process) ChangeDirection(dir proc.Direction) error {
if dir != proc.Forward {
return proc.ErrNotRecorded
}
return nil
}
// GetDirection will always return Forward.
func (p *Process) GetDirection() proc.Direction { return proc.Forward }
// When will always return an empty string and nil, not supported on native proc backend.
func (dbp *Process) When() (string, error) { return "", nil }
// Checkpoint will always return an error on the native proc backend,
// only supported for recorded traces.
func (dbp *Process) Checkpoint(string) (int, error) { return -1, proc.ErrNotRecorded }
// Checkpoints will always return an error on the native proc backend,
// only supported for recorded traces.
func (dbp *Process) Checkpoints() ([]proc.Checkpoint, error) { return nil, proc.ErrNotRecorded }
// ClearCheckpoint will always return an error on the native proc backend,
// only supported in recorded traces.
func (dbp *Process) ClearCheckpoint(int) error { return proc.ErrNotRecorded }
// Detach from the process being debugged, optionally killing it.
func (dbp *Process) Detach(kill bool) (err error) {
if dbp.exited {
return nil
}
if kill && dbp.childProcess {
err := dbp.kill()
if err != nil {
return err
}
dbp.bi.Close()
return nil
}
// Clean up any breakpoints we've set.
for _, bp := range dbp.breakpoints.M {
if bp != nil {
_, err := dbp.ClearBreakpoint(bp.Addr)
if err != nil {
return err
}
}
}
dbp.execPtraceFunc(func() {
err = dbp.detach(kill)
if err != nil {
return
}
if kill {
err = killProcess(dbp.pid)
}
})
dbp.detached = true
dbp.postExit()
return
}
// Valid returns whether the process is still attached to and
// has not exited.
func (dbp *Process) Valid() (bool, error) {
if dbp.detached {
return false, &proc.ProcessDetachedError{}
}
if dbp.exited {
return false, &proc.ErrProcessExited{Pid: dbp.Pid()}
}
return true, nil
}
// ResumeNotify specifies a channel that will be closed the next time
// ContinueOnce finishes resuming the target.
func (dbp *Process) ResumeNotify(ch chan<- struct{}) {
dbp.resumeChan = ch
}
// Pid returns the process ID.
func (dbp *Process) Pid() int {
return dbp.pid
}
// ThreadList returns a list of threads in the process.
func (dbp *Process) ThreadList() []proc.Thread {
r := make([]proc.Thread, 0, len(dbp.threads))
for _, v := range dbp.threads {
r = append(r, v)
}
return r
}
// FindThread attempts to find the thread with the specified ID.
func (dbp *Process) FindThread(threadID int) (proc.Thread, bool) {
th, ok := dbp.threads[threadID]
return th, ok
}
// CurrentThread returns the current selected, active thread.
func (dbp *Process) CurrentThread() proc.Thread {
return dbp.currentThread
}
// SetCurrentThread is used internally by proc.Target to change the current thread.
func (p *Process) SetCurrentThread(th proc.Thread) {
p.currentThread = th.(*Thread)
}
// Breakpoints returns a list of breakpoints currently set.
func (dbp *Process) Breakpoints() *proc.BreakpointMap {
return &dbp.breakpoints
}
// RequestManualStop sets the `halt` flag and
// sends SIGSTOP to all threads.
func (dbp *Process) RequestManualStop() error {
if dbp.exited {
return &proc.ErrProcessExited{Pid: dbp.Pid()}
}
dbp.stopMu.Lock()
defer dbp.stopMu.Unlock()
dbp.manualStopRequested = true
return dbp.requestManualStop()
}
// CheckAndClearManualStopRequest checks if a manual stop has
// been requested, and then clears that state.
func (dbp *Process) CheckAndClearManualStopRequest() bool {
dbp.stopMu.Lock()
defer dbp.stopMu.Unlock()
msr := dbp.manualStopRequested
dbp.manualStopRequested = false
return msr
}
func (dbp *Process) writeBreakpoint(addr uint64) (string, int, *proc.Function, []byte, error) {
f, l, fn := dbp.bi.PCToLine(uint64(addr))
originalData := make([]byte, dbp.bi.Arch.BreakpointSize())
_, err := dbp.currentThread.ReadMemory(originalData, uintptr(addr))
if err != nil {
return "", 0, nil, nil, err
}
if err := dbp.writeSoftwareBreakpoint(dbp.currentThread, addr); err != nil {
return "", 0, nil, nil, err
}
return f, l, fn, originalData, nil
}
// SetBreakpoint sets a breakpoint at addr, and stores it in the process wide
// break point table.
func (dbp *Process) SetBreakpoint(addr uint64, kind proc.BreakpointKind, cond ast.Expr) (*proc.Breakpoint, error) {
return dbp.breakpoints.Set(addr, kind, cond, dbp.writeBreakpoint)
}
// ClearBreakpoint clears the breakpoint at addr.
func (dbp *Process) ClearBreakpoint(addr uint64) (*proc.Breakpoint, error) {
if dbp.exited {
return nil, &proc.ErrProcessExited{Pid: dbp.Pid()}
}
return dbp.breakpoints.Clear(addr, dbp.currentThread.ClearBreakpoint)
}
// ContinueOnce will continue the target until it stops.
// This could be the result of a breakpoint or signal.
func (dbp *Process) ContinueOnce() (proc.Thread, proc.StopReason, error) {
if dbp.exited {
return nil, proc.StopExited, &proc.ErrProcessExited{Pid: dbp.Pid()}
}
if err := dbp.resume(); err != nil {
return nil, proc.StopUnknown, err
}
for _, th := range dbp.threads {
th.CurrentBreakpoint.Clear()
}
if dbp.resumeChan != nil {
close(dbp.resumeChan)
dbp.resumeChan = nil
}
trapthread, err := dbp.trapWait(-1)
if err != nil {
return nil, proc.StopUnknown, err
}
if err := dbp.stop(trapthread); err != nil {
return nil, proc.StopUnknown, err
}
return trapthread, proc.StopUnknown, err
}
// FindBreakpoint finds the breakpoint for the given pc.
func (dbp *Process) FindBreakpoint(pc uint64, adjustPC bool) (*proc.Breakpoint, bool) {
if adjustPC {
// Check to see if address is past the breakpoint, (i.e. breakpoint was hit).
if bp, ok := dbp.breakpoints.M[pc-uint64(dbp.bi.Arch.BreakpointSize())]; ok {
return bp, true
}
}
// Directly use addr to lookup breakpoint.
if bp, ok := dbp.breakpoints.M[pc]; ok {
return bp, true
}
return nil, false
}
// initialize will ensure that all relevant information is loaded
// so the process is ready to be debugged.
func (dbp *Process) initialize(path string, debugInfoDirs []string) (*proc.Target, error) {
if err := initialize(dbp); err != nil {
return nil, err
}
if err := dbp.updateThreadList(); err != nil {
return nil, err
}
stopReason := proc.StopLaunched
if !dbp.childProcess {
stopReason = proc.StopAttached
}
return proc.NewTarget(dbp, proc.NewTargetConfig{
Path: path,
DebugInfoDirs: debugInfoDirs,
WriteBreakpoint: dbp.writeBreakpoint,
DisableAsyncPreempt: runtime.GOOS == "windows",
StopReason: stopReason})
}
// ClearInternalBreakpoints will clear all non-user set breakpoints. These
// breakpoints are set for internal operations such as 'next'.
func (dbp *Process) ClearInternalBreakpoints() error {
return dbp.breakpoints.ClearInternalBreakpoints(func(bp *proc.Breakpoint) error {
if err := dbp.currentThread.ClearBreakpoint(bp); err != nil {
return err
}
for _, thread := range dbp.threads {
if thread.CurrentBreakpoint.Breakpoint == bp {
thread.CurrentBreakpoint.Clear()
}
}
return nil
})
}
func (dbp *Process) handlePtraceFuncs() {
// We must ensure here that we are running on the same thread during
// while invoking the ptrace(2) syscall. This is due to the fact that ptrace(2) expects
// all commands after PTRACE_ATTACH to come from the same thread.
runtime.LockOSThread()
for fn := range dbp.ptraceChan {
fn()
dbp.ptraceDoneChan <- nil
}
}
func (dbp *Process) execPtraceFunc(fn func()) {
dbp.ptraceChan <- fn
<-dbp.ptraceDoneChan
}
func (dbp *Process) postExit() {
dbp.exited = true
close(dbp.ptraceChan)
close(dbp.ptraceDoneChan)
dbp.bi.Close()
}
func (dbp *Process) writeSoftwareBreakpoint(thread *Thread, addr uint64) error {
_, err := thread.WriteMemory(uintptr(addr), dbp.bi.Arch.BreakpointInstruction())
return err
}