blob: fdeceb4e8dc6747e612615f92c38cceccc188b8b [file] [log] [blame]
// Copyright 2010 The Go Authors. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
// Package pprof writes runtime profiling data in the format expected
// by the pprof visualization tool.
// For more information about pprof, see
// http://code.google.com/p/google-perftools/.
package pprof
import (
"bufio"
"fmt"
"io"
"os"
"runtime"
"sync"
)
// WriteHeapProfile writes a pprof-formatted heap profile to w.
// If a write to w returns an error, WriteHeapProfile returns that error.
// Otherwise, WriteHeapProfile returns nil.
func WriteHeapProfile(w io.Writer) os.Error {
// Find out how many records there are (MemProfile(nil, false)),
// allocate that many records, and get the data.
// There's a race—more records might be added between
// the two calls—so allocate a few extra records for safety
// and also try again if we're very unlucky.
// The loop should only execute one iteration in the common case.
var p []runtime.MemProfileRecord
n, ok := runtime.MemProfile(nil, false)
for {
// Allocate room for a slightly bigger profile,
// in case a few more entries have been added
// since the call to MemProfile.
p = make([]runtime.MemProfileRecord, n+50)
n, ok = runtime.MemProfile(p, false)
if ok {
p = p[0:n]
break
}
// Profile grew; try again.
}
var total runtime.MemProfileRecord
for i := range p {
r := &p[i]
total.AllocBytes += r.AllocBytes
total.AllocObjects += r.AllocObjects
total.FreeBytes += r.FreeBytes
total.FreeObjects += r.FreeObjects
}
// Technically the rate is MemProfileRate not 2*MemProfileRate,
// but early versions of the C++ heap profiler reported 2*MemProfileRate,
// so that's what pprof has come to expect.
b := bufio.NewWriter(w)
fmt.Fprintf(b, "heap profile: %d: %d [%d: %d] @ heap/%d\n",
total.InUseObjects(), total.InUseBytes(),
total.AllocObjects, total.AllocBytes,
2*runtime.MemProfileRate)
for i := range p {
r := &p[i]
fmt.Fprintf(b, "%d: %d [%d: %d] @",
r.InUseObjects(), r.InUseBytes(),
r.AllocObjects, r.AllocBytes)
for _, pc := range r.Stack() {
fmt.Fprintf(b, " %#x", pc)
}
fmt.Fprintf(b, "\n")
}
// Print memstats information too.
// Pprof will ignore, but useful for people.
s := &runtime.MemStats
fmt.Fprintf(b, "\n# runtime.MemStats\n")
fmt.Fprintf(b, "# Alloc = %d\n", s.Alloc)
fmt.Fprintf(b, "# TotalAlloc = %d\n", s.TotalAlloc)
fmt.Fprintf(b, "# Sys = %d\n", s.Sys)
fmt.Fprintf(b, "# Lookups = %d\n", s.Lookups)
fmt.Fprintf(b, "# Mallocs = %d\n", s.Mallocs)
fmt.Fprintf(b, "# HeapAlloc = %d\n", s.HeapAlloc)
fmt.Fprintf(b, "# HeapSys = %d\n", s.HeapSys)
fmt.Fprintf(b, "# HeapIdle = %d\n", s.HeapIdle)
fmt.Fprintf(b, "# HeapInuse = %d\n", s.HeapInuse)
fmt.Fprintf(b, "# Stack = %d / %d\n", s.StackInuse, s.StackSys)
fmt.Fprintf(b, "# MSpan = %d / %d\n", s.MSpanInuse, s.MSpanSys)
fmt.Fprintf(b, "# MCache = %d / %d\n", s.MCacheInuse, s.MCacheSys)
fmt.Fprintf(b, "# BuckHashSys = %d\n", s.BuckHashSys)
fmt.Fprintf(b, "# NextGC = %d\n", s.NextGC)
fmt.Fprintf(b, "# PauseNs = %d\n", s.PauseNs)
fmt.Fprintf(b, "# NumGC = %d\n", s.NumGC)
fmt.Fprintf(b, "# EnableGC = %v\n", s.EnableGC)
fmt.Fprintf(b, "# DebugGC = %v\n", s.DebugGC)
fmt.Fprintf(b, "# BySize = Size * (Active = Mallocs - Frees)\n")
fmt.Fprintf(b, "# (Excluding large blocks.)\n")
for _, t := range s.BySize {
if t.Mallocs > 0 {
fmt.Fprintf(b, "# %d * (%d = %d - %d)\n", t.Size, t.Mallocs-t.Frees, t.Mallocs, t.Frees)
}
}
return b.Flush()
}
var cpu struct {
sync.Mutex
profiling bool
done chan bool
}
// StartCPUProfile enables CPU profiling for the current process.
// While profiling, the profile will be buffered and written to w.
// StartCPUProfile returns an error if profiling is already enabled.
func StartCPUProfile(w io.Writer) os.Error {
// The runtime routines allow a variable profiling rate,
// but in practice operating systems cannot trigger signals
// at more than about 500 Hz, and our processing of the
// signal is not cheap (mostly getting the stack trace).
// 100 Hz is a reasonable choice: it is frequent enough to
// produce useful data, rare enough not to bog down the
// system, and a nice round number to make it easy to
// convert sample counts to seconds. Instead of requiring
// each client to specify the frequency, we hard code it.
const hz = 100
// Avoid queueing behind StopCPUProfile.
// Could use TryLock instead if we had it.
if cpu.profiling {
return fmt.Errorf("cpu profiling already in use")
}
cpu.Lock()
defer cpu.Unlock()
if cpu.done == nil {
cpu.done = make(chan bool)
}
// Double-check.
if cpu.profiling {
return fmt.Errorf("cpu profiling already in use")
}
cpu.profiling = true
runtime.SetCPUProfileRate(hz)
go profileWriter(w)
return nil
}
func profileWriter(w io.Writer) {
for {
data := runtime.CPUProfile()
if data == nil {
break
}
w.Write(data)
}
cpu.done <- true
}
// StopCPUProfile stops the current CPU profile, if any.
// StopCPUProfile only returns after all the writes for the
// profile have completed.
func StopCPUProfile() {
cpu.Lock()
defer cpu.Unlock()
if !cpu.profiling {
return
}
cpu.profiling = false
runtime.SetCPUProfileRate(0)
<-cpu.done
}