cmd/pprof: use DWARF info to lookup unknown PC addresses

Test to follow in a separate CL that arranges for the runtime package to
store non-Go addresses in a CPU profile.

Change-Id: I33ce1d66b77340b1e62b54505fc9b1abcec108a9
Reviewed-on: https://go-review.googlesource.com/21055
Reviewed-by: Austin Clements <austin@google.com>
Run-TryBot: Ian Lance Taylor <iant@golang.org>

7 files changed, 123 insertions, 7 deletions

diff --git a/src/cmd/internal/objfile/elf.go b/src/cmd/internal/objfile/elf.go
index 6369f7e259..3bad034097 100644
--- a/src/cmd/internal/objfile/elf.go
+++ b/src/cmd/internal/objfile/elf.go
@@ -7,6 +7,7 @@
 package objfile
 
 import (
+	"debug/dwarf"
 	"debug/elf"
 	"fmt"
 	"os"
@@ -104,3 +105,7 @@ func (f *elfFile) goarch() string {
 	}
 	return ""
 }
+
+func (f *elfFile) dwarf() (*dwarf.Data, error) {
+	return f.elf.DWARF()
+}
diff --git a/src/cmd/internal/objfile/goobj.go b/src/cmd/internal/objfile/goobj.go
index 46e5f4e866..5a084a94be 100644
--- a/src/cmd/internal/objfile/goobj.go
+++ b/src/cmd/internal/objfile/goobj.go
@@ -8,6 +8,8 @@ package objfile
 
 import (
 	"cmd/internal/goobj"
+	"debug/dwarf"
+	"errors"
 	"fmt"
 	"os"
 )
@@ -91,3 +93,7 @@ func (f *goobjFile) text() (textStart uint64, text []byte, err error) {
 func (f *goobjFile) goarch() string {
 	return "GOARCH unimplemented for debug/goobj files"
 }
+
+func (f *goobjFile) dwarf() (*dwarf.Data, error) {
+	return nil, errors.New("no DWARF data in go object file")
+}
diff --git a/src/cmd/internal/objfile/macho.go b/src/cmd/internal/objfile/macho.go
index c98d9520c2..754674d757 100644
--- a/src/cmd/internal/objfile/macho.go
+++ b/src/cmd/internal/objfile/macho.go
@@ -7,6 +7,7 @@
 package objfile
 
 import (
+	"debug/dwarf"
 	"debug/macho"
 	"fmt"
 	"os"
@@ -123,3 +124,7 @@ type uint64s []uint64
 func (x uint64s) Len() int           { return len(x) }
 func (x uint64s) Swap(i, j int)      { x[i], x[j] = x[j], x[i] }
 func (x uint64s) Less(i, j int) bool { return x[i] < x[j] }
+
+func (f *machoFile) dwarf() (*dwarf.Data, error) {
+	return f.macho.DWARF()
+}
diff --git a/src/cmd/internal/objfile/objfile.go b/src/cmd/internal/objfile/objfile.go
index cf58e9c6aa..48ed9ed489 100644
--- a/src/cmd/internal/objfile/objfile.go
+++ b/src/cmd/internal/objfile/objfile.go
@@ -6,6 +6,7 @@
 package objfile
 
 import (
+	"debug/dwarf"
 	"debug/gosym"
 	"fmt"
 	"os"
@@ -17,6 +18,7 @@ type rawFile interface {
 	pcln() (textStart uint64, symtab, pclntab []byte, err error)
 	text() (textStart uint64, text []byte, err error)
 	goarch() string
+	dwarf() (*dwarf.Data, error)
 }
 
 // A File is an opened executable file.
@@ -92,3 +94,9 @@ func (f *File) Text() (uint64, []byte, error) {
 func (f *File) GOARCH() string {
 	return f.raw.goarch()
 }
+
+// DWARF returns DWARF debug data for the file, if any.
+// This is for cmd/pprof to locate cgo functions.
+func (f *File) DWARF() (*dwarf.Data, error) {
+	return f.raw.dwarf()
+}
diff --git a/src/cmd/internal/objfile/pe.go b/src/cmd/internal/objfile/pe.go
index 503d657647..1b319941ac 100644
--- a/src/cmd/internal/objfile/pe.go
+++ b/src/cmd/internal/objfile/pe.go
@@ -7,6 +7,7 @@
 package objfile
 
 import (
+	"debug/dwarf"
 	"debug/pe"
 	"fmt"
 	"os"
@@ -199,3 +200,7 @@ func (f *peFile) goarch() string {
 	}
 	return ""
 }
+
+func (f *peFile) dwarf() (*dwarf.Data, error) {
+	return f.pe.DWARF()
+}
diff --git a/src/cmd/internal/objfile/plan9obj.go b/src/cmd/internal/objfile/plan9obj.go
index 45a6d02748..1d808f77eb 100644
--- a/src/cmd/internal/objfile/plan9obj.go
+++ b/src/cmd/internal/objfile/plan9obj.go
@@ -7,7 +7,9 @@
 package objfile
 
 import (
+	"debug/dwarf"
 	"debug/plan9obj"
+	"errors"
 	"fmt"
 	"os"
 	"sort"
@@ -144,3 +146,7 @@ func (f *plan9File) goarch() string {
 	}
 	return ""
 }
+
+func (f *plan9File) dwarf() (*dwarf.Data, error) {
+	return nil, errors.New("no DWARF data in Plan 9 file")
+}
diff --git a/src/cmd/pprof/pprof.go b/src/cmd/pprof/pprof.go
index 2b20f1da77..1c55d05d5d 100644
--- a/src/cmd/pprof/pprof.go
+++ b/src/cmd/pprof/pprof.go
@@ -5,6 +5,7 @@
 package main
 
 import (
+	"debug/dwarf"
 	"debug/gosym"
 	"flag"
 	"fmt"
@@ -172,6 +173,9 @@ type file struct {
 	sym  []objfile.Sym
 	file *objfile.File
 	pcln *gosym.Table
+
+	triedDwarf bool
+	dwarf      *dwarf.Data
 }
 
 func (f *file) Name() string {
@@ -197,17 +201,94 @@ func (f *file) SourceLine(addr uint64) ([]plugin.Frame, error) {
 		f.pcln = pcln
 	}
 	file, line, fn := f.pcln.PCToLine(addr)
-	if fn == nil {
-		return nil, fmt.Errorf("no line information for PC=%#x", addr)
+	if fn != nil {
+		frame := []plugin.Frame{
+			{
+				Func: fn.Name,
+				File: file,
+				Line: line,
+			},
+		}
+		return frame, nil
+	}
+
+	frames := f.dwarfSourceLine(addr)
+	if frames != nil {
+		return frames, nil
+	}
+
+	return nil, fmt.Errorf("no line information for PC=%#x", addr)
+}
+
+// dwarfSourceLine tries to get file/line information using DWARF.
+// This is for C functions that appear in the profile.
+// Returns nil if there is no information available.
+func (f *file) dwarfSourceLine(addr uint64) []plugin.Frame {
+	if f.dwarf == nil && !f.triedDwarf {
+		// Ignore any error--we don't care exactly why there
+		// is no DWARF info.
+		f.dwarf, _ = f.file.DWARF()
+		f.triedDwarf = true
+	}
+
+	if f.dwarf != nil {
+		r := f.dwarf.Reader()
+		unit, err := r.SeekPC(addr)
+		if err == nil {
+			if frames := f.dwarfSourceLineEntry(r, unit, addr); frames != nil {
+				return frames
+			}
+		}
+	}
+
+	return nil
+}
+
+// dwarfSourceLineEntry tries to get file/line information from a
+// DWARF compilation unit. Returns nil if it doesn't find anything.
+func (f *file) dwarfSourceLineEntry(r *dwarf.Reader, entry *dwarf.Entry, addr uint64) []plugin.Frame {
+	lines, err := f.dwarf.LineReader(entry)
+	if err != nil {
+		return nil
+	}
+	var lentry dwarf.LineEntry
+	if err := lines.SeekPC(addr, &lentry); err != nil {
+		return nil
 	}
-	frame := []plugin.Frame{
+
+	// Try to find the function name.
+	name := ""
+FindName:
+	for entry, err := r.Next(); entry != nil && err == nil; entry, err = r.Next() {
+		if entry.Tag == dwarf.TagSubprogram {
+			ranges, err := f.dwarf.Ranges(entry)
+			if err != nil {
+				return nil
+			}
+			for _, pcs := range ranges {
+				if pcs[0] <= addr && addr < pcs[1] {
+					var ok bool
+					// TODO: AT_linkage_name, AT_MIPS_linkage_name.
+					name, ok = entry.Val(dwarf.AttrName).(string)
+					if ok {
+						break FindName
+					}
+				}
+			}
+		}
+	}
+
+	// TODO: Report inlined functions.
+
+	frames := []plugin.Frame{
 		{
-			Func: fn.Name,
-			File: file,
-			Line: line,
+			Func: name,
+			File: lentry.File.Name,
+			Line: lentry.Line,
 		},
 	}
-	return frame, nil
+
+	return frames
 }
 
 func (f *file) Symbols(r *regexp.Regexp, addr uint64) ([]*plugin.Sym, error) {