runtime,runtime/pprof: get memory mappings on darwin.

Displaying assembly language has never worked for Apple Silicon
macs (see #50891). This change uses mach_vm_region to obtain the
necessary VM mappings to allow for locating assembly instructions
for a cpu profile.

Fixes #50891

Change-Id: Ib968c55a19b481b82f63337276b552f3b18f69d1
Reviewed-on: https://go-review.googlesource.com/c/go/+/503919
Run-TryBot: Cherry Mui <cherryyz@google.com>
Reviewed-by: Cherry Mui <cherryyz@google.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
Reviewed-by: David Chase <drchase@google.com>

9 files changed, 327 insertions, 16 deletions

diff --git a/src/runtime/pprof/defs_darwin.go b/src/runtime/pprof/defs_darwin.go
new file mode 100644
index 0000000000..2b2f68132c
--- /dev/null
+++ b/src/runtime/pprof/defs_darwin.go
@@ -0,0 +1,30 @@
+// Copyright 2023 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.
+
+// This file is used as input to cgo --godefs (GOOS=arm64 or amd64) to
+// generate the types used in viminfo_darwin_{arm64,amd64}.go which are
+// hand edited as appropriate, primarily to avoid exporting the types.
+
+//go:build ignore
+
+package pprof
+
+/*
+#include <sys/param.h>
+#include <mach/vm_prot.h>
+#include <mach/vm_region.h>
+*/
+import "C"
+
+type machVMRegionBasicInfoData C.vm_region_basic_info_data_64_t
+
+const (
+	_VM_PROT_READ    = C.VM_PROT_READ
+	_VM_PROT_WRITE   = C.VM_PROT_WRITE
+	_VM_PROT_EXECUTE = C.VM_PROT_EXECUTE
+
+	_MACH_SEND_INVALID_DEST = C.MACH_SEND_INVALID_DEST
+
+	_MAXPATHLEN = C.MAXPATHLEN
+)
diff --git a/src/runtime/pprof/defs_darwin_amd64.go b/src/runtime/pprof/defs_darwin_amd64.go
new file mode 100644
index 0000000000..14226495d8
--- /dev/null
+++ b/src/runtime/pprof/defs_darwin_amd64.go
@@ -0,0 +1,26 @@
+// Code generated by cmd/cgo -godefs; DO NOT EDIT.
+// cgo -godefs defs_darwin.go
+
+package pprof
+
+type machVMRegionBasicInfoData struct {
+	Protection       int32
+	Max_protection   int32
+	Inheritance      uint32
+	Shared           uint32
+	Reserved         uint32
+	Offset           uint64 // This is hand-edited since godefs generates: Pad_cgo_0 [8]byte
+	Behavior         int32
+	User_wired_count uint16
+	Pad_cgo_1        [2]byte
+}
+
+const (
+	_VM_PROT_READ    = 0x1
+	_VM_PROT_WRITE   = 0x2
+	_VM_PROT_EXECUTE = 0x4
+
+	_MACH_SEND_INVALID_DEST = 0x10000003
+
+	_MAXPATHLEN = 0x400
+)
diff --git a/src/runtime/pprof/defs_darwin_arm64.go b/src/runtime/pprof/defs_darwin_arm64.go
new file mode 100644
index 0000000000..2d34a80d80
--- /dev/null
+++ b/src/runtime/pprof/defs_darwin_arm64.go
@@ -0,0 +1,26 @@
+// Code generated by cmd/cgo -godefs; DO NOT EDIT.
+// cgo -godefs defs_darwin.go
+
+package pprof
+
+type machVMRegionBasicInfoData struct {
+	Protection       int32
+	Max_protection   int32
+	Inheritance      uint32
+	Shared           int32
+	Reserved         int32
+	Offset           uint64 // This is hand-edited since godefs generates: Pad_cgo_0 [8]byte
+	Behavior         int32
+	User_wired_count uint16
+	Pad_cgo_1        [2]byte
+}
+
+const (
+	_VM_PROT_READ    = 0x1
+	_VM_PROT_WRITE   = 0x2
+	_VM_PROT_EXECUTE = 0x4
+
+	_MACH_SEND_INVALID_DEST = 0x10000003
+
+	_MAXPATHLEN = 0x400
+)
diff --git a/src/runtime/pprof/proto_darwin.go b/src/runtime/pprof/proto_darwin.go
new file mode 100644
index 0000000000..8db9e1d2b3
--- /dev/null
+++ b/src/runtime/pprof/proto_darwin.go
@@ -0,0 +1,36 @@
+// Copyright 2023 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
+
+import (
+	"errors"
+)
+
+// readMapping adds a mapping entry for the text region of the running process.
+// It uses the mach_vm_region region system call to add mapping entries for the
+// text region of the running process. Note that currently no attempt is
+// made to obtain the buildID information.
+func (b *profileBuilder) readMapping() {
+	if !machVMInfo(b.addMapping) {
+		b.addMappingEntry(0, 0, 0, "", "", true)
+	}
+}
+
+func readMainModuleMapping() (start, end uint64, exe, buildID string, err error) {
+	first := true
+	ok := machVMInfo(func(lo, hi, off uint64, file, build string) {
+		if first {
+			start, end = lo, hi
+			exe, buildID = file, build
+		}
+		// May see multiple text segments if rosetta is used for running
+		// the go toolchain itself.
+		first = false
+	})
+	if !ok {
+		return 0, 0, "", "", errors.New("machVMInfo failed")
+	}
+	return start, end, exe, buildID, nil
+}
diff --git a/src/runtime/pprof/proto_other.go b/src/runtime/pprof/proto_other.go
index 4a7fe79501..7322e84803 100644
--- a/src/runtime/pprof/proto_other.go
+++ b/src/runtime/pprof/proto_other.go
@@ -2,7 +2,7 @@
 // Use of this source code is governed by a BSD-style
 // license that can be found in the LICENSE file.
 
-//go:build !windows
+//go:build !windows && !darwin
 
 package pprof
 
@@ -25,6 +25,6 @@ func (b *profileBuilder) readMapping() {
 	}
 }
 
-func readMainModuleMapping() (start, end uint64, err error) {
-	return 0, 0, errors.New("not implemented")
+func readMainModuleMapping() (start, end uint64, exe, buildID string, err error) {
+	return 0, 0, "", "", errors.New("not implemented")
 }
diff --git a/src/runtime/pprof/proto_test.go b/src/runtime/pprof/proto_test.go
index 8ec9c9109a..eb43816b7c 100644
--- a/src/runtime/pprof/proto_test.go
+++ b/src/runtime/pprof/proto_test.go
@@ -101,16 +101,11 @@ func testPCs(t *testing.T) (addr1, addr2 uint64, map1, map2 *profile.Mapping) {
 		addr2 = mprof.Mapping[1].Start
 		map2 = mprof.Mapping[1]
 		map2.BuildID, _ = elfBuildID(map2.File)
-	case "windows":
+	case "windows", "darwin":
 		addr1 = uint64(abi.FuncPCABIInternal(f1))
 		addr2 = uint64(abi.FuncPCABIInternal(f2))
 
-		exe, err := os.Executable()
-		if err != nil {
-			t.Fatal(err)
-		}
-
-		start, end, err := readMainModuleMapping()
+		start, end, exe, buildID, err := readMainModuleMapping()
 		if err != nil {
 			t.Fatal(err)
 		}
@@ -120,7 +115,7 @@ func testPCs(t *testing.T) (addr1, addr2 uint64, map1, map2 *profile.Mapping) {
 			Start:        start,
 			Limit:        end,
 			File:         exe,
-			BuildID:      peBuildID(exe),
+			BuildID:      buildID,
 			HasFunctions: true,
 		}
 		map2 = &profile.Mapping{
@@ -128,7 +123,7 @@ func testPCs(t *testing.T) (addr1, addr2 uint64, map1, map2 *profile.Mapping) {
 			Start:        start,
 			Limit:        end,
 			File:         exe,
-			BuildID:      peBuildID(exe),
+			BuildID:      buildID,
 			HasFunctions: true,
 		}
 	case "js", "wasip1":
diff --git a/src/runtime/pprof/proto_windows.go b/src/runtime/pprof/proto_windows.go
index d5ae4a5eec..f4dc44bd07 100644
--- a/src/runtime/pprof/proto_windows.go
+++ b/src/runtime/pprof/proto_windows.go
@@ -7,6 +7,7 @@ package pprof
 import (
 	"errors"
 	"internal/syscall/windows"
+	"os"
 	"syscall"
 )
 
@@ -42,10 +43,14 @@ func (b *profileBuilder) readMapping() {
 	}
 }
 
-func readMainModuleMapping() (start, end uint64, err error) {
+func readMainModuleMapping() (start, end uint64, exe, buildID string, err error) {
+	exe, err = os.Executable()
+	if err != nil {
+		return 0, 0, "", "", err
+	}
 	snap, err := createModuleSnapshot()
 	if err != nil {
-		return 0, 0, err
+		return 0, 0, "", "", err
 	}
 	defer func() { _ = syscall.CloseHandle(snap) }()
 
@@ -53,10 +58,10 @@ func readMainModuleMapping() (start, end uint64, err error) {
 	module.Size = uint32(windows.SizeofModuleEntry32)
 	err = windows.Module32First(snap, &module)
 	if err != nil {
-		return 0, 0, err
+		return 0, 0, "", "", err
 	}
 
-	return uint64(module.ModBaseAddr), uint64(module.ModBaseAddr) + uint64(module.ModBaseSize), nil
+	return uint64(module.ModBaseAddr), uint64(module.ModBaseAddr) + uint64(module.ModBaseSize), exe, peBuildID(exe), nil
 }
 
 func createModuleSnapshot() (syscall.Handle, error) {
diff --git a/src/runtime/pprof/vminfo_darwin.go b/src/runtime/pprof/vminfo_darwin.go
new file mode 100644
index 0000000000..7ddb0d1c68
--- /dev/null
+++ b/src/runtime/pprof/vminfo_darwin.go
@@ -0,0 +1,71 @@
+// Copyright 2023 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
+
+import (
+	"os"
+	"unsafe"
+)
+
+func isExecutable(protection int32) bool {
+	return (protection&_VM_PROT_EXECUTE) != 0 && (protection&_VM_PROT_READ) != 0
+}
+
+// machVMInfo uses the mach_vm_region region system call to add mapping entries
+// for the text region of the running process.
+func machVMInfo(addMapping func(lo, hi, offset uint64, file, buildID string)) bool {
+	added := false
+	var addr uint64 = 0x1
+	for {
+		var memRegionSize uint64
+		var info machVMRegionBasicInfoData
+		// Get the first address and page size.
+		kr := mach_vm_region(
+			&addr,
+			&memRegionSize,
+			unsafe.Pointer(&info))
+		if kr != 0 {
+			if kr == _MACH_SEND_INVALID_DEST {
+				// No more memory regions.
+				return true
+			}
+			return added // return true if at least one mapping was added
+		}
+		if isExecutable(info.Protection) {
+			// NOTE: the meaning/value of Offset is unclear. However,
+			// this likely doesn't matter as the text segment's file
+			// offset is usually 0.
+			addMapping(addr,
+				addr+memRegionSize,
+				uint64(info.Offset),
+				regionFilename(addr),
+				"")
+			added = true
+		}
+		addr += memRegionSize
+	}
+}
+
+func regionFilename(address uint64) string {
+	buf := make([]byte, _MAXPATHLEN)
+	r := proc_regionfilename(
+		os.Getpid(),
+		address,
+		unsafe.SliceData(buf),
+		int64(cap(buf)))
+	if r == 0 {
+		return ""
+	}
+	return string(buf[:r])
+}
+
+// mach_vm_region and proc_regionfilename are implemented by
+// the runtime package (runtime/sys_darwin.go).
+//
+//go:noescape
+func mach_vm_region(address, region_size *uint64, info unsafe.Pointer) int32
+
+//go:noescape
+func proc_regionfilename(pid int, address uint64, buf *byte, buflen int64) int32
diff --git a/src/runtime/pprof/vminfo_darwin_test.go b/src/runtime/pprof/vminfo_darwin_test.go
new file mode 100644
index 0000000000..3023878fde
--- /dev/null
+++ b/src/runtime/pprof/vminfo_darwin_test.go
@@ -0,0 +1,122 @@
+// Copyright 2023 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
+
+import (
+	"bufio"
+	"bytes"
+	"internal/abi"
+	"os"
+	"os/exec"
+	"strconv"
+	"strings"
+	"testing"
+)
+
+func TestVMInfo(t *testing.T) {
+	var begin, end, offset uint64
+	var filename string
+	first := true
+	machVMInfo(func(lo, hi, off uint64, file, buildID string) {
+		if first {
+			begin = lo
+			end = hi
+			offset = off
+			filename = file
+		}
+		// May see multiple text segments if rosetta is used for running
+		// the go toolchain itself.
+		first = false
+	})
+	lo, hi := useVMMap(t)
+	if got, want := begin, lo; got != want {
+		t.Errorf("got %x, want %x", got, want)
+	}
+	if got, want := end, hi; got != want {
+		t.Errorf("got %x, want %x", got, want)
+	}
+	if got, want := offset, uint64(0); got != want {
+		t.Errorf("got %x, want %x", got, want)
+	}
+	if !strings.HasSuffix(filename, "pprof.test") {
+		t.Errorf("got %s, want pprof.test", filename)
+	}
+	addr := uint64(abi.FuncPCABIInternal(TestVMInfo))
+	if addr < lo || addr > hi {
+		t.Errorf("%x..%x does not contain function %p (%x)", lo, hi, TestVMInfo, addr)
+	}
+}
+
+func useVMMap(t *testing.T) (hi, lo uint64) {
+	pid := strconv.Itoa(os.Getpid())
+	out, err := exec.Command("vmmap", pid).Output()
+	if err != nil {
+		t.Fatal(err)
+	}
+	return parseVmmap(t, out)
+}
+
+// parseVmmap parses the output of vmmap and calls addMapping for the first r-x TEXT segment in the output.
+func parseVmmap(t *testing.T, data []byte) (hi, lo uint64) {
+	// vmmap 53799
+	// Process:         gopls [53799]
+	// Path:            /Users/USER/*/gopls
+	// Load Address:    0x1029a0000
+	// Identifier:      gopls
+	// Version:         ???
+	// Code Type:       ARM64
+	// Platform:        macOS
+	// Parent Process:  Code Helper (Plugin) [53753]
+	//
+	// Date/Time:       2023-05-25 09:45:49.331 -0700
+	// Launch Time:     2023-05-23 09:35:37.514 -0700
+	// OS Version:      macOS 13.3.1 (22E261)
+	// Report Version:  7
+	// Analysis Tool:   /Applications/Xcode.app/Contents/Developer/usr/bin/vmmap
+	// Analysis Tool Version:  Xcode 14.3 (14E222b)
+	//
+	// Physical footprint:         1.2G
+	// Physical footprint (peak):  1.2G
+	// Idle exit:                  untracked
+	// ----
+	//
+	// Virtual Memory Map of process 53799 (gopls)
+	// Output report format:  2.4  -64-bit process
+	// VM page size:  16384 bytes
+	//
+	// ==== Non-writable regions for process 53799
+	// REGION TYPE                    START END         [ VSIZE  RSDNT  DIRTY   SWAP] PRT/MAX SHRMOD PURGE    REGION DETAIL
+	// __TEXT                      1029a0000-1033bc000    [ 10.1M  7360K     0K     0K] r-x/rwx SM=COW          /Users/USER/*/gopls
+	// __DATA_CONST                1033bc000-1035bc000    [ 2048K  2000K     0K     0K] r--/rwSM=COW          /Users/USER/*/gopls
+	// __DATA_CONST                1035bc000-103a48000    [ 4656K  3824K     0K     0K] r--/rwSM=COW          /Users/USER/*/gopls
+	// __LINKEDIT                  103b00000-103c98000    [ 1632K  1616K     0K     0K] r--/r-SM=COW          /Users/USER/*/gopls
+	// dyld private memory         103cd8000-103cdc000    [   16K     0K     0K     0K] ---/--SM=NUL
+	// shared memory               103ce4000-103ce8000    [   16K    16K    16K     0K] r--/r-SM=SHM
+	// MALLOC metadata             103ce8000-103cec000    [   16K    16K    16K     0K] r--/rwx SM=COW          DefaultMallocZone_0x103ce8000 zone structure
+	// MALLOC guard page           103cf0000-103cf4000    [   16K     0K     0K     0K] ---/rwx SM=COW
+	// MALLOC guard page           103cfc000-103d00000    [   16K     0K     0K     0K] ---/rwx SM=COW
+	// MALLOC guard page           103d00000-103d04000    [   16K     0K     0K     0K] ---/rwx SM=NUL
+
+	banner := "==== Non-writable regions for process"
+	grabbing := false
+	sc := bufio.NewScanner(bytes.NewReader(data))
+	for sc.Scan() {
+		l := sc.Text()
+		if grabbing {
+			p := strings.Fields(l)
+			if len(p) > 7 && p[0] == "__TEXT" && p[7] == "r-x/rwx" {
+				locs := strings.Split(p[1], "-")
+				start, _ := strconv.ParseUint(locs[0], 16, 64)
+				end, _ := strconv.ParseUint(locs[1], 16, 64)
+				return start, end
+			}
+		}
+		if strings.HasPrefix(l, banner) {
+			grabbing = true
+		}
+	}
+	t.Fatal("vmmap no text segment found")
+	return 0, 0
+}