delve/proc/variables.go
aarzilli b839eda2a9 proc/variables: prefetch of target process memory
Prefetch the entire memory of structs and arrays and cache it instead
of issuing readMemory calls only when we get down to primitive types.
This reduces the number of system calls to ptrace that variables makes.

Improves performance in general, greatly improving it in some
particular cases (involving large structs).

Benchmarks without prefetching:
	BenchmarkArray-4         	      10	 132189944 ns/op	   0.06 MB/s
	BenchmarkArrayPointer-4  	       5	 202538503 ns/op	   0.04 MB/s
	BenchmarkMap-4           	     500	   3804336 ns/op	   0.27 MB/s
	BenchmarkGoroutinesInfo-4	      10	 126397104 ns/op
	BenchmarkLocalVariables-4	     500	   2494846 ns/op

Benchmarks with prefetching:
	BenchmarkArray-4         	     200	  10719087 ns/op	   0.76 MB/s
	BenchmarkArrayPointer-4  	     100	  11931326 ns/op	   0.73 MB/s
	BenchmarkMap-4           	    1000	   1466479 ns/op	   0.70 MB/s
	BenchmarkGoroutinesInfo-4	      10	 103407004 ns/op
	BenchmarkLocalVariables-4	    1000	   1530395 ns/op

Improvement factors:
	BenchmarkArray				12.33x
	BenchmarkArrayPointer		16.97x
	BenchmarkMap					 2.59x
	BenchmarkGoroutinesInfo		 1.22x
	BenchmarkLocalVariables		 1.63x
2016-01-10 13:49:03 +01:00

1550 lines
38 KiB
Go

package proc
import (
"bytes"
"debug/dwarf"
"encoding/binary"
"fmt"
"go/constant"
"go/parser"
"go/token"
"reflect"
"strings"
"unsafe"
"github.com/derekparker/delve/dwarf/op"
"github.com/derekparker/delve/dwarf/reader"
)
const (
maxVariableRecurse = 1 // How far to recurse when evaluating nested types.
maxArrayValues = 64 // Max value for reading large arrays.
maxErrCount = 3 // Max number of read errors to accept while evaluating slices, arrays and structs
maxArrayStridePrefetch = 1024 // Maximum size of array stride for which we will prefetch the array contents
chanRecv = "chan receive"
chanSend = "chan send"
hashTophashEmpty = 0 // used by map reading code, indicates an empty bucket
hashMinTopHash = 4 // used by map reading code, indicates minimum value of tophash that isn't empty or evacuated
)
// Variable represents a variable. It contains the address, name,
// type and other information parsed from both the Dwarf information
// and the memory of the debugged process.
// If OnlyAddr is true, the variables value has not been loaded.
type Variable struct {
Addr uintptr
OnlyAddr bool
Name string
DwarfType dwarf.Type
RealType dwarf.Type
Kind reflect.Kind
mem memoryReadWriter
dbp *Process
Value constant.Value
Len int64
Cap int64
// base address of arrays, base address of the backing array for slices (0 for nil slices)
// base address of the backing byte array for strings
// address of the struct backing chan and map variables
// address of the function entry point for function variables (0 for nil function pointers)
base uintptr
stride int64
fieldType dwarf.Type
// number of elements to skip when loading a map
mapSkip int
Children []Variable
loaded bool
Unreadable error
}
// M represents a runtime M (OS thread) structure.
type M struct {
procid int // Thread ID or port.
spinning uint8 // Busy looping.
blocked uint8 // Waiting on futex / semaphore.
curg uintptr // Current G running on this thread.
}
// G status, from: src/runtime/runtime2.go
const (
Gidle uint64 = iota // 0
Grunnable // 1 runnable and on a run queue
Grunning // 2
Gsyscall // 3
Gwaiting // 4
GmoribundUnused // 5 currently unused, but hardcoded in gdb scripts
Gdead // 6
Genqueue // 7 Only the Gscanenqueue is used.
Gcopystack // 8 in this state when newstack is moving the stack
)
// G represents a runtime G (goroutine) structure (at least the
// fields that Delve is interested in).
type G struct {
ID int // Goroutine ID
PC uint64 // PC of goroutine when it was parked.
SP uint64 // SP of goroutine when it was parked.
GoPC uint64 // PC of 'go' statement that created this goroutine.
WaitReason string // Reason for goroutine being parked.
Status uint64
// Information on goroutine location
CurrentLoc Location
// PC of entry to top-most deferred function.
DeferPC uint64
// Thread that this goroutine is currently allocated to
thread *Thread
dbp *Process
}
// EvalScope is the scope for variable evaluation. Contains the thread,
// current location (PC), and canonical frame address.
type EvalScope struct {
Thread *Thread
PC uint64
CFA int64
}
// IsNilErr is returned when a variable is nil.
type IsNilErr struct {
name string
}
func (err *IsNilErr) Error() string {
return fmt.Sprintf("%s is nil", err.name)
}
func ptrTypeKind(t *dwarf.PtrType) reflect.Kind {
structtyp, isstruct := t.Type.(*dwarf.StructType)
_, isvoid := t.Type.(*dwarf.VoidType)
if isstruct && strings.HasPrefix(structtyp.StructName, "hchan<") {
return reflect.Chan
} else if isstruct && strings.HasPrefix(structtyp.StructName, "hash<") {
return reflect.Map
} else if isvoid {
return reflect.UnsafePointer
}
return reflect.Ptr
}
func (scope *EvalScope) newVariable(name string, addr uintptr, dwarfType dwarf.Type) *Variable {
return newVariable(name, addr, dwarfType, scope.Thread.dbp, scope.Thread)
}
func (v *Variable) newVariable(name string, addr uintptr, dwarfType dwarf.Type) *Variable {
return newVariable(name, addr, dwarfType, v.dbp, v.mem)
}
func newVariable(name string, addr uintptr, dwarfType dwarf.Type, dbp *Process, mem memoryReadWriter) *Variable {
v := &Variable{
Name: name,
Addr: addr,
DwarfType: dwarfType,
mem: mem,
dbp: dbp,
}
v.RealType = resolveTypedef(v.DwarfType)
switch t := v.RealType.(type) {
case *dwarf.PtrType:
v.Kind = ptrTypeKind(t)
case *dwarf.StructType:
switch {
case t.StructName == "string":
v.Kind = reflect.String
v.stride = 1
v.fieldType = &dwarf.UintType{BasicType: dwarf.BasicType{CommonType: dwarf.CommonType{ByteSize: 1, Name: "byte"}, BitSize: 8, BitOffset: 0}}
if v.Addr != 0 {
v.base, v.Len, v.Unreadable = readStringInfo(v.mem, v.dbp.arch, v.Addr)
}
case t.StructName == "runtime.iface" || t.StructName == "runtime.eface":
v.Kind = reflect.Interface
case strings.HasPrefix(t.StructName, "[]"):
v.Kind = reflect.Slice
if v.Addr != 0 {
v.loadSliceInfo(t)
}
default:
v.Kind = reflect.Struct
}
case *dwarf.ArrayType:
v.Kind = reflect.Array
v.base = v.Addr
v.Len = t.Count
v.Cap = -1
v.fieldType = t.Type
v.stride = 0
if t.Count > 0 {
v.stride = t.ByteSize / t.Count
}
case *dwarf.ComplexType:
switch t.ByteSize {
case 8:
v.Kind = reflect.Complex64
case 16:
v.Kind = reflect.Complex128
}
case *dwarf.IntType:
v.Kind = reflect.Int
case *dwarf.UintType:
v.Kind = reflect.Uint
case *dwarf.FloatType:
switch t.ByteSize {
case 4:
v.Kind = reflect.Float32
case 8:
v.Kind = reflect.Float64
}
case *dwarf.BoolType:
v.Kind = reflect.Bool
case *dwarf.FuncType:
v.Kind = reflect.Func
case *dwarf.VoidType:
v.Kind = reflect.Invalid
case *dwarf.UnspecifiedType:
v.Kind = reflect.Invalid
default:
v.Unreadable = fmt.Errorf("Unknown type: %T", t)
}
return v
}
func resolveTypedef(typ dwarf.Type) dwarf.Type {
for {
if tt, ok := typ.(*dwarf.TypedefType); ok {
typ = tt.Type
} else {
return typ
}
}
}
func newConstant(val constant.Value, mem memoryReadWriter) *Variable {
v := &Variable{Value: val, mem: mem, loaded: true}
switch val.Kind() {
case constant.Int:
v.Kind = reflect.Int
case constant.Float:
v.Kind = reflect.Float64
case constant.Bool:
v.Kind = reflect.Bool
case constant.Complex:
v.Kind = reflect.Complex128
case constant.String:
v.Kind = reflect.String
v.Len = int64(len(constant.StringVal(val)))
}
return v
}
var nilVariable = &Variable{
Addr: 0,
base: 0,
Kind: reflect.Ptr,
Children: []Variable{{Addr: 0, OnlyAddr: true}},
}
func (v *Variable) clone() *Variable {
r := *v
return &r
}
// TypeString returns the string representation
// of the type of this variable.
func (v *Variable) TypeString() string {
if v == nilVariable {
return "nil"
}
if v.DwarfType != nil {
return v.DwarfType.String()
}
return v.Kind.String()
}
func (v *Variable) toField(field *dwarf.StructField) (*Variable, error) {
if v.Unreadable != nil {
return v.clone(), nil
}
if v.Addr == 0 {
return nil, &IsNilErr{v.Name}
}
name := ""
if v.Name != "" {
parts := strings.Split(field.Name, ".")
if len(parts) > 1 {
name = fmt.Sprintf("%s.%s", v.Name, parts[1])
} else {
name = fmt.Sprintf("%s.%s", v.Name, field.Name)
}
}
return v.newVariable(name, uintptr(int64(v.Addr)+field.ByteOffset), field.Type), nil
}
// DwarfReader returns the DwarfReader containing the
// Dwarf information for the target process.
func (scope *EvalScope) DwarfReader() *reader.Reader {
return scope.Thread.dbp.DwarfReader()
}
// Type returns the Dwarf type entry at `offset`.
func (scope *EvalScope) Type(offset dwarf.Offset) (dwarf.Type, error) {
return scope.Thread.dbp.dwarf.Type(offset)
}
// PtrSize returns the size of a pointer.
func (scope *EvalScope) PtrSize() int {
return scope.Thread.dbp.arch.PtrSize()
}
// ChanRecvBlocked returns whether the goroutine is blocked on
// a channel read operation.
func (g *G) ChanRecvBlocked() bool {
return g.WaitReason == chanRecv
}
// chanRecvReturnAddr returns the address of the return from a channel read.
func (g *G) chanRecvReturnAddr(dbp *Process) (uint64, error) {
locs, err := dbp.GoroutineStacktrace(g, 4)
if err != nil {
return 0, err
}
topLoc := locs[len(locs)-1]
return topLoc.Current.PC, nil
}
// NoGError returned when a G could not be found
// for a specific thread.
type NoGError struct {
tid int
}
func (ng NoGError) Error() string {
return fmt.Sprintf("no G executing on thread %d", ng.tid)
}
func parseG(thread *Thread, gaddr uint64, deref bool) (*G, error) {
initialInstructions := make([]byte, thread.dbp.arch.PtrSize()+1)
initialInstructions[0] = op.DW_OP_addr
binary.LittleEndian.PutUint64(initialInstructions[1:], gaddr)
if deref {
gaddrbytes, err := thread.readMemory(uintptr(gaddr), thread.dbp.arch.PtrSize())
if err != nil {
return nil, fmt.Errorf("error derefing *G %s", err)
}
initialInstructions = append([]byte{op.DW_OP_addr}, gaddrbytes...)
gaddr = binary.LittleEndian.Uint64(gaddrbytes)
if gaddr == 0 {
return nil, NoGError{tid: thread.ID}
}
}
rdr := thread.dbp.DwarfReader()
rdr.Seek(0)
entry, err := rdr.SeekToTypeNamed("runtime.g")
if err != nil {
return nil, err
}
var mem memoryReadWriter = thread
if gtype, err := thread.dbp.dwarf.Type(entry.Offset); err == nil {
mem = cacheMemory(thread, uintptr(gaddr), int(gtype.Size()))
}
// Parse defer
deferAddr, err := rdr.AddrForMember("_defer", initialInstructions)
if err != nil {
return nil, err
}
var deferPC uint64
// Dereference *defer pointer
deferAddrBytes, err := mem.readMemory(uintptr(deferAddr), thread.dbp.arch.PtrSize())
if err != nil {
return nil, fmt.Errorf("error derefing defer %s", err)
}
if binary.LittleEndian.Uint64(deferAddrBytes) != 0 {
initialDeferInstructions := append([]byte{op.DW_OP_addr}, deferAddrBytes...)
_, err = rdr.SeekToTypeNamed("runtime._defer")
if err != nil {
return nil, err
}
deferPCAddr, err := rdr.AddrForMember("fn", initialDeferInstructions)
deferPC, err = readUintRaw(mem, uintptr(deferPCAddr), 8)
if err != nil {
return nil, err
}
deferPC, err = readUintRaw(mem, uintptr(deferPC), 8)
if err != nil {
return nil, err
}
}
// Let's parse all of the members we care about in order so that
// we don't have to spend any extra time seeking.
err = rdr.SeekToEntry(entry)
if err != nil {
return nil, err
}
// Parse sched
schedAddr, err := rdr.AddrForMember("sched", initialInstructions)
if err != nil {
return nil, err
}
// From sched, let's parse PC and SP.
sp, err := readUintRaw(mem, uintptr(schedAddr), 8)
if err != nil {
return nil, err
}
pc, err := readUintRaw(mem, uintptr(schedAddr+uint64(thread.dbp.arch.PtrSize())), 8)
if err != nil {
return nil, err
}
// Parse atomicstatus
atomicStatusAddr, err := rdr.AddrForMember("atomicstatus", initialInstructions)
if err != nil {
return nil, err
}
atomicStatus, err := readUintRaw(mem, uintptr(atomicStatusAddr), 4)
// Parse goid
goidAddr, err := rdr.AddrForMember("goid", initialInstructions)
if err != nil {
return nil, err
}
goid, err := readIntRaw(mem, uintptr(goidAddr), 8)
if err != nil {
return nil, err
}
// Parse waitreason
waitReasonAddr, err := rdr.AddrForMember("waitreason", initialInstructions)
if err != nil {
return nil, err
}
waitreason, _, err := readString(mem, thread.dbp.arch, uintptr(waitReasonAddr))
if err != nil {
return nil, err
}
// Parse gopc
gopcAddr, err := rdr.AddrForMember("gopc", initialInstructions)
if err != nil {
return nil, err
}
gopc, err := readUintRaw(mem, uintptr(gopcAddr), 8)
if err != nil {
return nil, err
}
f, l, fn := thread.dbp.goSymTable.PCToLine(pc)
g := &G{
ID: int(goid),
GoPC: gopc,
PC: pc,
SP: sp,
CurrentLoc: Location{PC: pc, File: f, Line: l, Fn: fn},
WaitReason: waitreason,
DeferPC: deferPC,
Status: atomicStatus,
dbp: thread.dbp,
}
return g, nil
}
// From $GOROOT/src/runtime/traceback.go:597
// isExportedRuntime reports whether name is an exported runtime function.
// It is only for runtime functions, so ASCII A-Z is fine.
func isExportedRuntime(name string) bool {
const n = len("runtime.")
return len(name) > n && name[:n] == "runtime." && 'A' <= name[n] && name[n] <= 'Z'
}
// UserCurrent returns the location the users code is at,
// or was at before entering a runtime function.
func (g *G) UserCurrent() Location {
pc, sp := g.PC, g.SP
if g.thread != nil {
regs, err := g.thread.Registers()
if err != nil {
return g.CurrentLoc
}
pc, sp = regs.PC(), regs.SP()
}
it := newStackIterator(g.dbp, pc, sp)
for it.Next() {
frame := it.Frame()
name := frame.Call.Fn.Name
if (strings.Index(name, ".") >= 0) && (!strings.HasPrefix(name, "runtime.") || isExportedRuntime(name)) {
return frame.Call
}
}
return g.CurrentLoc
}
// Go returns the location of the 'go' statement
// that spawned this goroutine.
func (g *G) Go() Location {
f, l, fn := g.dbp.goSymTable.PCToLine(g.GoPC)
return Location{PC: g.GoPC, File: f, Line: l, Fn: fn}
}
// EvalVariable returns the value of the given expression (backwards compatibility).
func (scope *EvalScope) EvalVariable(name string) (*Variable, error) {
return scope.EvalExpression(name)
}
// SetVariable sets the value of the named variable
func (scope *EvalScope) SetVariable(name, value string) error {
t, err := parser.ParseExpr(name)
if err != nil {
return err
}
xv, err := scope.evalAST(t)
if err != nil {
return err
}
if xv.Addr == 0 {
return fmt.Errorf("Can not assign to \"%s\"", name)
}
if xv.Unreadable != nil {
return fmt.Errorf("Expression \"%s\" is unreadable: %v", name, xv.Unreadable)
}
t, err = parser.ParseExpr(value)
if err != nil {
return err
}
yv, err := scope.evalAST(t)
if err != nil {
return err
}
yv.loadValue()
if err := yv.isType(xv.RealType, xv.Kind); err != nil {
return err
}
if yv.Unreadable != nil {
return fmt.Errorf("Expression \"%s\" is unreadable: %v", value, yv.Unreadable)
}
return xv.setValue(yv)
}
func (scope *EvalScope) extractVariableFromEntry(entry *dwarf.Entry) (*Variable, error) {
rdr := scope.DwarfReader()
v, err := scope.extractVarInfoFromEntry(entry, rdr)
if err != nil {
return nil, err
}
v.loadValue()
return v, nil
}
func (scope *EvalScope) extractVarInfo(varName string) (*Variable, error) {
reader := scope.DwarfReader()
_, err := reader.SeekToFunction(scope.PC)
if err != nil {
return nil, err
}
for entry, err := reader.NextScopeVariable(); entry != nil; entry, err = reader.NextScopeVariable() {
if err != nil {
return nil, err
}
n, ok := entry.Val(dwarf.AttrName).(string)
if !ok {
continue
}
if n == varName {
return scope.extractVarInfoFromEntry(entry, reader)
}
}
return nil, fmt.Errorf("could not find symbol value for %s", varName)
}
// LocalVariables returns all local variables from the current function scope.
func (scope *EvalScope) LocalVariables() ([]*Variable, error) {
return scope.variablesByTag(dwarf.TagVariable)
}
// FunctionArguments returns the name, value, and type of all current function arguments.
func (scope *EvalScope) FunctionArguments() ([]*Variable, error) {
return scope.variablesByTag(dwarf.TagFormalParameter)
}
// PackageVariables returns the name, value, and type of all package variables in the application.
func (scope *EvalScope) PackageVariables() ([]*Variable, error) {
var vars []*Variable
reader := scope.DwarfReader()
for entry, err := reader.NextPackageVariable(); entry != nil; entry, err = reader.NextPackageVariable() {
if err != nil {
return nil, err
}
// Ignore errors trying to extract values
val, err := scope.extractVariableFromEntry(entry)
if err != nil {
continue
}
vars = append(vars, val)
}
return vars, nil
}
// EvalPackageVariable will evaluate the package level variable
// specified by 'name'.
func (dbp *Process) EvalPackageVariable(name string) (*Variable, error) {
scope := &EvalScope{Thread: dbp.CurrentThread, PC: 0, CFA: 0}
v, err := scope.packageVarAddr(name)
if err != nil {
return nil, err
}
v.loadValue()
return v, nil
}
func (scope *EvalScope) packageVarAddr(name string) (*Variable, error) {
reader := scope.DwarfReader()
for entry, err := reader.NextPackageVariable(); entry != nil; entry, err = reader.NextPackageVariable() {
if err != nil {
return nil, err
}
n, ok := entry.Val(dwarf.AttrName).(string)
if !ok {
continue
}
if n == name {
return scope.extractVarInfoFromEntry(entry, reader)
}
}
return nil, fmt.Errorf("could not find symbol value for %s", name)
}
func (v *Variable) structMember(memberName string) (*Variable, error) {
if v.Unreadable != nil {
return v.clone(), nil
}
structVar := v.maybeDereference()
structVar.Name = v.Name
if structVar.Unreadable != nil {
return structVar, nil
}
switch t := structVar.RealType.(type) {
case *dwarf.StructType:
for _, field := range t.Field {
if field.Name != memberName {
continue
}
return structVar.toField(field)
}
// Check for embedded field only if field was
// not a regular struct member
for _, field := range t.Field {
isEmbeddedStructMember :=
(field.Type.String() == ("struct " + field.Name)) ||
(len(field.Name) > 1 &&
field.Name[0] == '*' &&
field.Type.String()[1:] == ("struct "+field.Name[1:]))
if !isEmbeddedStructMember {
continue
}
// Check for embedded field referenced by type name
parts := strings.Split(field.Name, ".")
if len(parts) > 1 && parts[1] == memberName {
embeddedVar, err := structVar.toField(field)
if err != nil {
return nil, err
}
return embeddedVar, nil
}
// Recursively check for promoted fields on the embedded field
embeddedVar, err := structVar.toField(field)
if err != nil {
return nil, err
}
embeddedVar.Name = structVar.Name
embeddedField, err := embeddedVar.structMember(memberName)
if embeddedField != nil {
return embeddedField, nil
}
}
return nil, fmt.Errorf("%s has no member %s", v.Name, memberName)
default:
if v.Name == "" {
return nil, fmt.Errorf("type %s is not a struct", structVar.TypeString())
}
return nil, fmt.Errorf("%s (type %s) is not a struct", v.Name, structVar.TypeString())
}
}
// Extracts the name and type of a variable from a dwarf entry
// then executes the instructions given in the DW_AT_location attribute to grab the variable's address
func (scope *EvalScope) extractVarInfoFromEntry(entry *dwarf.Entry, rdr *reader.Reader) (*Variable, error) {
if entry == nil {
return nil, fmt.Errorf("invalid entry")
}
if entry.Tag != dwarf.TagFormalParameter && entry.Tag != dwarf.TagVariable {
return nil, fmt.Errorf("invalid entry tag, only supports FormalParameter and Variable, got %s", entry.Tag.String())
}
n, ok := entry.Val(dwarf.AttrName).(string)
if !ok {
return nil, fmt.Errorf("type assertion failed")
}
offset, ok := entry.Val(dwarf.AttrType).(dwarf.Offset)
if !ok {
return nil, fmt.Errorf("type assertion failed")
}
t, err := scope.Type(offset)
if err != nil {
return nil, err
}
instructions, ok := entry.Val(dwarf.AttrLocation).([]byte)
if !ok {
return nil, fmt.Errorf("type assertion failed")
}
addr, err := op.ExecuteStackProgram(scope.CFA, instructions)
if err != nil {
return nil, err
}
return scope.newVariable(n, uintptr(addr), t), nil
}
// If v is a pointer a new variable is returned containing the value pointed by v.
func (v *Variable) maybeDereference() *Variable {
if v.Unreadable != nil {
return v
}
switch t := v.RealType.(type) {
case *dwarf.PtrType:
ptrval, err := readUintRaw(v.mem, uintptr(v.Addr), t.ByteSize)
r := v.newVariable("", uintptr(ptrval), t.Type)
if err != nil {
r.Unreadable = err
}
return r
default:
return v
}
}
// Extracts the value of the variable at the given address.
func (v *Variable) loadValue() {
v.loadValueInternal(0)
}
func (v *Variable) loadValueInternal(recurseLevel int) {
if v.Unreadable != nil || v.loaded || (v.Addr == 0 && v.base == 0) {
return
}
v.loaded = true
switch v.Kind {
case reflect.Ptr, reflect.UnsafePointer:
v.Len = 1
v.Children = []Variable{*v.maybeDereference()}
// Don't increase the recursion level when dereferencing pointers
v.Children[0].loadValueInternal(recurseLevel)
case reflect.Chan:
sv := v.maybeDereference()
sv.loadValueInternal(recurseLevel)
v.Children = sv.Children
v.Len = sv.Len
v.base = sv.Addr
case reflect.Map:
v.loadMap(recurseLevel)
case reflect.String:
var val string
val, v.Unreadable = readStringValue(v.mem, v.base, v.Len)
v.Value = constant.MakeString(val)
case reflect.Slice, reflect.Array:
v.loadArrayValues(recurseLevel)
case reflect.Struct:
v.mem = cacheMemory(v.mem, v.Addr, int(v.RealType.Size()))
t := v.RealType.(*dwarf.StructType)
v.Len = int64(len(t.Field))
// Recursively call extractValue to grab
// the value of all the members of the struct.
if recurseLevel <= maxVariableRecurse {
v.Children = make([]Variable, 0, len(t.Field))
for i, field := range t.Field {
f, _ := v.toField(field)
v.Children = append(v.Children, *f)
v.Children[i].Name = field.Name
v.Children[i].loadValueInternal(recurseLevel + 1)
}
}
case reflect.Interface:
v.loadInterface(recurseLevel, true)
case reflect.Complex64, reflect.Complex128:
v.readComplex(v.RealType.(*dwarf.ComplexType).ByteSize)
case reflect.Int, reflect.Int8, reflect.Int16, reflect.Int32, reflect.Int64:
var val int64
val, v.Unreadable = readIntRaw(v.mem, v.Addr, v.RealType.(*dwarf.IntType).ByteSize)
v.Value = constant.MakeInt64(val)
case reflect.Uint, reflect.Uint8, reflect.Uint16, reflect.Uint32, reflect.Uint64, reflect.Uintptr:
var val uint64
val, v.Unreadable = readUintRaw(v.mem, v.Addr, v.RealType.(*dwarf.UintType).ByteSize)
v.Value = constant.MakeUint64(val)
case reflect.Bool:
val, err := v.mem.readMemory(v.Addr, 1)
v.Unreadable = err
if err == nil {
v.Value = constant.MakeBool(val[0] != 0)
}
case reflect.Float32, reflect.Float64:
var val float64
val, v.Unreadable = v.readFloatRaw(v.RealType.(*dwarf.FloatType).ByteSize)
v.Value = constant.MakeFloat64(val)
case reflect.Func:
v.readFunctionPtr()
default:
v.Unreadable = fmt.Errorf("unknown or unsupported kind: \"%s\"", v.Kind.String())
}
}
func (v *Variable) setValue(y *Variable) error {
var err error
switch v.Kind {
case reflect.Float32, reflect.Float64:
f, _ := constant.Float64Val(y.Value)
err = v.writeFloatRaw(f, v.RealType.Size())
case reflect.Int, reflect.Int8, reflect.Int16, reflect.Int32, reflect.Int64:
n, _ := constant.Int64Val(y.Value)
err = v.writeUint(uint64(n), v.RealType.Size())
case reflect.Uint, reflect.Uint8, reflect.Uint16, reflect.Uint32, reflect.Uint64:
n, _ := constant.Uint64Val(y.Value)
err = v.writeUint(n, v.RealType.Size())
case reflect.Bool:
err = v.writeBool(constant.BoolVal(y.Value))
case reflect.Complex64, reflect.Complex128:
real, _ := constant.Float64Val(constant.Real(y.Value))
imag, _ := constant.Float64Val(constant.Imag(y.Value))
err = v.writeComplex(real, imag, v.RealType.Size())
default:
fmt.Printf("default\n")
if t, isptr := v.RealType.(*dwarf.PtrType); isptr {
err = v.writeUint(uint64(y.Children[0].Addr), int64(t.ByteSize))
} else {
return fmt.Errorf("can not set variables of type %s (not implemented)", v.Kind.String())
}
}
return err
}
func readStringInfo(mem memoryReadWriter, arch Arch, addr uintptr) (uintptr, int64, error) {
// string data structure is always two ptrs in size. Addr, followed by len
// http://research.swtch.com/godata
mem = cacheMemory(mem, addr, arch.PtrSize()*2)
// read len
val, err := mem.readMemory(addr+uintptr(arch.PtrSize()), arch.PtrSize())
if err != nil {
return 0, 0, fmt.Errorf("could not read string len %s", err)
}
strlen := int64(binary.LittleEndian.Uint64(val))
if strlen < 0 {
return 0, 0, fmt.Errorf("invalid length: %d", strlen)
}
// read addr
val, err = mem.readMemory(addr, arch.PtrSize())
if err != nil {
return 0, 0, fmt.Errorf("could not read string pointer %s", err)
}
addr = uintptr(binary.LittleEndian.Uint64(val))
if addr == 0 {
return 0, 0, nil
}
return addr, strlen, nil
}
func readStringValue(mem memoryReadWriter, addr uintptr, strlen int64) (string, error) {
count := strlen
if count > maxArrayValues {
count = maxArrayValues
}
val, err := mem.readMemory(addr, int(count))
if err != nil {
return "", fmt.Errorf("could not read string at %#v due to %s", addr, err)
}
retstr := *(*string)(unsafe.Pointer(&val))
return retstr, nil
}
func readString(mem memoryReadWriter, arch Arch, addr uintptr) (string, int64, error) {
addr, strlen, err := readStringInfo(mem, arch, addr)
if err != nil {
return "", 0, err
}
retstr, err := readStringValue(mem, addr, strlen)
return retstr, strlen, err
}
func (v *Variable) loadSliceInfo(t *dwarf.StructType) {
v.mem = cacheMemory(v.mem, v.Addr, int(t.Size()))
var err error
for _, f := range t.Field {
switch f.Name {
case "array":
var base uint64
base, err = readUintRaw(v.mem, uintptr(int64(v.Addr)+f.ByteOffset), f.Type.Size())
if err == nil {
v.base = uintptr(base)
// Dereference array type to get value type
ptrType, ok := f.Type.(*dwarf.PtrType)
if !ok {
v.Unreadable = fmt.Errorf("Invalid type %s in slice array", f.Type)
return
}
v.fieldType = ptrType.Type
}
case "len":
lstrAddr, _ := v.toField(f)
lstrAddr.loadValue()
err = lstrAddr.Unreadable
if err == nil {
v.Len, _ = constant.Int64Val(lstrAddr.Value)
}
case "cap":
cstrAddr, _ := v.toField(f)
cstrAddr.loadValue()
err = cstrAddr.Unreadable
if err == nil {
v.Cap, _ = constant.Int64Val(cstrAddr.Value)
}
}
if err != nil {
v.Unreadable = err
return
}
}
v.stride = v.fieldType.Size()
if t, ok := v.fieldType.(*dwarf.PtrType); ok {
v.stride = t.ByteSize
}
return
}
func (v *Variable) loadArrayValues(recurseLevel int) {
if v.Unreadable != nil {
return
}
count := v.Len
// Cap number of elements
if count > maxArrayValues {
count = maxArrayValues
}
if v.stride < maxArrayStridePrefetch {
v.mem = cacheMemory(v.mem, v.base, int(v.stride*count))
}
errcount := 0
for i := int64(0); i < count; i++ {
fieldvar := v.newVariable("", uintptr(int64(v.base)+(i*v.stride)), v.fieldType)
fieldvar.loadValueInternal(recurseLevel + 1)
if fieldvar.Unreadable != nil {
errcount++
}
v.Children = append(v.Children, *fieldvar)
if errcount > maxErrCount {
break
}
}
}
func (v *Variable) readComplex(size int64) {
var fs int64
switch size {
case 8:
fs = 4
case 16:
fs = 8
default:
v.Unreadable = fmt.Errorf("invalid size (%d) for complex type", size)
return
}
ftyp := &dwarf.FloatType{BasicType: dwarf.BasicType{CommonType: dwarf.CommonType{ByteSize: fs, Name: fmt.Sprintf("float%d", fs)}, BitSize: fs * 8, BitOffset: 0}}
realvar := v.newVariable("real", v.Addr, ftyp)
imagvar := v.newVariable("imaginary", v.Addr+uintptr(fs), ftyp)
realvar.loadValue()
imagvar.loadValue()
v.Value = constant.BinaryOp(realvar.Value, token.ADD, constant.MakeImag(imagvar.Value))
}
func (v *Variable) writeComplex(real, imag float64, size int64) error {
err := v.writeFloatRaw(real, int64(size/2))
if err != nil {
return err
}
imagaddr := *v
imagaddr.Addr += uintptr(size / 2)
return imagaddr.writeFloatRaw(imag, int64(size/2))
}
func readIntRaw(mem memoryReadWriter, addr uintptr, size int64) (int64, error) {
var n int64
val, err := mem.readMemory(addr, int(size))
if err != nil {
return 0, err
}
switch size {
case 1:
n = int64(val[0])
case 2:
n = int64(binary.LittleEndian.Uint16(val))
case 4:
n = int64(binary.LittleEndian.Uint32(val))
case 8:
n = int64(binary.LittleEndian.Uint64(val))
}
return n, nil
}
func (v *Variable) writeUint(value uint64, size int64) error {
val := make([]byte, size)
switch size {
case 1:
val[0] = byte(value)
case 2:
binary.LittleEndian.PutUint16(val, uint16(value))
case 4:
binary.LittleEndian.PutUint32(val, uint32(value))
case 8:
binary.LittleEndian.PutUint64(val, uint64(value))
}
_, err := v.mem.writeMemory(v.Addr, val)
return err
}
func readUintRaw(mem memoryReadWriter, addr uintptr, size int64) (uint64, error) {
var n uint64
val, err := mem.readMemory(addr, int(size))
if err != nil {
return 0, err
}
switch size {
case 1:
n = uint64(val[0])
case 2:
n = uint64(binary.LittleEndian.Uint16(val))
case 4:
n = uint64(binary.LittleEndian.Uint32(val))
case 8:
n = uint64(binary.LittleEndian.Uint64(val))
}
return n, nil
}
func (v *Variable) readFloatRaw(size int64) (float64, error) {
val, err := v.mem.readMemory(v.Addr, int(size))
if err != nil {
return 0.0, err
}
buf := bytes.NewBuffer(val)
switch size {
case 4:
n := float32(0)
binary.Read(buf, binary.LittleEndian, &n)
return float64(n), nil
case 8:
n := float64(0)
binary.Read(buf, binary.LittleEndian, &n)
return n, nil
}
return 0.0, fmt.Errorf("could not read float")
}
func (v *Variable) writeFloatRaw(f float64, size int64) error {
buf := bytes.NewBuffer(make([]byte, 0, size))
switch size {
case 4:
n := float32(f)
binary.Write(buf, binary.LittleEndian, n)
case 8:
n := float64(f)
binary.Write(buf, binary.LittleEndian, n)
}
_, err := v.mem.writeMemory(v.Addr, buf.Bytes())
return err
}
func (v *Variable) writeBool(value bool) error {
val := []byte{0}
val[0] = *(*byte)(unsafe.Pointer(&value))
_, err := v.mem.writeMemory(v.Addr, val)
return err
}
func (v *Variable) readFunctionPtr() {
val, err := v.mem.readMemory(v.Addr, v.dbp.arch.PtrSize())
if err != nil {
v.Unreadable = err
return
}
// dereference pointer to find function pc
fnaddr := uintptr(binary.LittleEndian.Uint64(val))
if fnaddr == 0 {
v.base = 0
v.Value = constant.MakeString("")
return
}
val, err = v.mem.readMemory(fnaddr, v.dbp.arch.PtrSize())
if err != nil {
v.Unreadable = err
return
}
v.base = uintptr(binary.LittleEndian.Uint64(val))
fn := v.dbp.goSymTable.PCToFunc(uint64(v.base))
if fn == nil {
v.Unreadable = fmt.Errorf("could not find function for %#v", v.base)
return
}
v.Value = constant.MakeString(fn.Name)
}
func (v *Variable) loadMap(recurseLevel int) {
it := v.mapIterator()
if it == nil {
return
}
for skip := 0; skip < v.mapSkip; skip++ {
if ok := it.next(); !ok {
v.Unreadable = fmt.Errorf("map index out of bounds")
return
}
}
count := 0
errcount := 0
for it.next() {
if count >= maxArrayValues {
break
}
key := it.key()
val := it.value()
if recurseLevel <= maxVariableRecurse {
key.loadValueInternal(recurseLevel + 1)
val.loadValueInternal(recurseLevel + 1)
}
if key.Unreadable != nil || val.Unreadable != nil {
errcount++
}
v.Children = append(v.Children, *key)
v.Children = append(v.Children, *val)
count++
if errcount > maxErrCount {
break
}
}
}
type mapIterator struct {
v *Variable
numbuckets uint64
oldmask uint64
buckets *Variable
oldbuckets *Variable
b *Variable
bidx uint64
tophashes *Variable
keys *Variable
values *Variable
overflow *Variable
idx int64
}
// Code derived from go/src/runtime/hashmap.go
func (v *Variable) mapIterator() *mapIterator {
sv := v.maybeDereference()
v.base = sv.Addr
maptype, ok := sv.RealType.(*dwarf.StructType)
if !ok {
v.Unreadable = fmt.Errorf("wrong real type for map")
return nil
}
it := &mapIterator{v: v, bidx: 0, b: nil, idx: 0}
if sv.Addr == 0 {
it.numbuckets = 0
return it
}
v.mem = cacheMemory(v.mem, v.base, int(v.RealType.Size()))
for _, f := range maptype.Field {
var err error
field, _ := sv.toField(f)
switch f.Name {
case "count":
v.Len, err = field.asInt()
case "B":
var b uint64
b, err = field.asUint()
it.numbuckets = 1 << b
it.oldmask = (1 << (b - 1)) - 1
case "buckets":
it.buckets = field.maybeDereference()
case "oldbuckets":
it.oldbuckets = field.maybeDereference()
}
if err != nil {
v.Unreadable = err
return nil
}
}
return it
}
func (it *mapIterator) nextBucket() bool {
if it.overflow != nil && it.overflow.Addr > 0 {
it.b = it.overflow
} else {
it.b = nil
for it.bidx < it.numbuckets {
it.b = it.buckets.clone()
it.b.Addr += uintptr(uint64(it.buckets.DwarfType.Size()) * it.bidx)
if it.oldbuckets.Addr <= 0 {
break
}
// if oldbuckets is not nil we are iterating through a map that is in
// the middle of a grow.
// if the bucket we are looking at hasn't been filled in we iterate
// instead through its corresponding "oldbucket" (i.e. the bucket the
// elements of this bucket are coming from) but only if this is the first
// of the two buckets being created from the same oldbucket (otherwise we
// would print some keys twice)
oldbidx := it.bidx & it.oldmask
oldb := it.oldbuckets.clone()
oldb.Addr += uintptr(uint64(it.oldbuckets.DwarfType.Size()) * oldbidx)
if mapEvacuated(oldb) {
break
}
if oldbidx == it.bidx {
it.b = oldb
break
}
// oldbucket origin for current bucket has not been evacuated but we have already
// iterated over it so we should just skip it
it.b = nil
it.bidx++
}
if it.b == nil {
return false
}
it.bidx++
}
if it.b.Addr <= 0 {
return false
}
it.b.mem = cacheMemory(it.b.mem, it.b.Addr, int(it.b.RealType.Size()))
it.tophashes = nil
it.keys = nil
it.values = nil
it.overflow = nil
for _, f := range it.b.DwarfType.(*dwarf.StructType).Field {
field, err := it.b.toField(f)
if err != nil {
it.v.Unreadable = err
return false
}
if field.Unreadable != nil {
it.v.Unreadable = field.Unreadable
return false
}
switch f.Name {
case "tophash":
it.tophashes = field
case "keys":
it.keys = field
case "values":
it.values = field
case "overflow":
it.overflow = field.maybeDereference()
}
}
// sanity checks
if it.tophashes == nil || it.keys == nil || it.values == nil {
it.v.Unreadable = fmt.Errorf("malformed map type")
return false
}
if it.tophashes.Kind != reflect.Array || it.keys.Kind != reflect.Array || it.values.Kind != reflect.Array {
it.v.Unreadable = fmt.Errorf("malformed map type: keys, values or tophash of a bucket is not an array")
return false
}
if it.tophashes.Len != it.keys.Len || it.tophashes.Len != it.values.Len {
it.v.Unreadable = fmt.Errorf("malformed map type: inconsistent array length in bucket")
return false
}
return true
}
func (it *mapIterator) next() bool {
for {
if it.b == nil || it.idx >= it.tophashes.Len {
r := it.nextBucket()
if !r {
return false
}
it.idx = 0
}
tophash, _ := it.tophashes.sliceAccess(int(it.idx))
h, err := tophash.asUint()
if err != nil {
it.v.Unreadable = fmt.Errorf("unreadable tophash: %v", err)
return false
}
it.idx++
if h != hashTophashEmpty {
return true
}
}
}
func (it *mapIterator) key() *Variable {
k, _ := it.keys.sliceAccess(int(it.idx - 1))
return k
}
func (it *mapIterator) value() *Variable {
v, _ := it.values.sliceAccess(int(it.idx - 1))
return v
}
func mapEvacuated(b *Variable) bool {
if b.Addr == 0 {
return true
}
for _, f := range b.DwarfType.(*dwarf.StructType).Field {
if f.Name != "tophash" {
continue
}
tophashes, _ := b.toField(f)
tophash0var, _ := tophashes.sliceAccess(0)
tophash0, err := tophash0var.asUint()
if err != nil {
return true
}
return tophash0 > hashTophashEmpty && tophash0 < hashMinTopHash
}
return true
}
func (v *Variable) loadInterface(recurseLevel int, loadData bool) {
var typestring, data *Variable
isnil := false
v.mem = cacheMemory(v.mem, v.Addr, int(v.RealType.Size()))
for _, f := range v.RealType.(*dwarf.StructType).Field {
switch f.Name {
case "tab": // for runtime.iface
tab, _ := v.toField(f)
_type, err := tab.structMember("_type")
if err != nil {
_, isnil = err.(*IsNilErr)
if !isnil {
v.Unreadable = fmt.Errorf("invalid interface type: %v", err)
return
}
} else {
typestring, err = _type.structMember("_string")
if err != nil {
v.Unreadable = fmt.Errorf("invalid interface type: %v", err)
return
}
typestring = typestring.maybeDereference()
}
case "_type": // for runtime.eface
var err error
_type, _ := v.toField(f)
typestring, err = _type.structMember("_string")
if err != nil {
_, isnil = err.(*IsNilErr)
if !isnil {
v.Unreadable = fmt.Errorf("invalid interface type: %v", err)
return
}
} else {
typestring = typestring.maybeDereference()
}
case "data":
data, _ = v.toField(f)
}
}
if isnil {
// interface to nil
data = data.maybeDereference()
v.Children = []Variable{*data}
v.Children[0].loadValueInternal(recurseLevel)
return
}
if typestring == nil || data == nil || typestring.Addr == 0 || typestring.Kind != reflect.String {
v.Unreadable = fmt.Errorf("invalid interface type")
return
}
typestring.loadValue()
if typestring.Unreadable != nil {
v.Unreadable = fmt.Errorf("invalid interface type: %v", typestring.Unreadable)
return
}
t, err := parser.ParseExpr(constant.StringVal(typestring.Value))
if err != nil {
v.Unreadable = fmt.Errorf("invalid interface type, unparsable data type: %v", err)
return
}
typ, err := v.dbp.findTypeExpr(t)
if err != nil {
v.Unreadable = fmt.Errorf("interface type \"%s\" not found for 0x%x: %v", constant.StringVal(typestring.Value), data.Addr, err)
return
}
realtyp := resolveTypedef(typ)
if _, isptr := realtyp.(*dwarf.PtrType); !isptr {
// interface to non-pointer types are pointers even if the type says otherwise
typ = v.dbp.pointerTo(typ)
}
data = data.newVariable("data", data.Addr, typ)
v.Children = []Variable{*data}
if loadData {
v.Children[0].loadValueInternal(recurseLevel)
}
return
}
// Fetches all variables of a specific type in the current function scope
func (scope *EvalScope) variablesByTag(tag dwarf.Tag) ([]*Variable, error) {
reader := scope.DwarfReader()
_, err := reader.SeekToFunction(scope.PC)
if err != nil {
return nil, err
}
var vars []*Variable
for entry, err := reader.NextScopeVariable(); entry != nil; entry, err = reader.NextScopeVariable() {
if err != nil {
return nil, err
}
if entry.Tag == tag {
val, err := scope.extractVariableFromEntry(entry)
if err != nil {
// skip variables that we can't parse yet
continue
}
vars = append(vars, val)
}
}
return vars, nil
}