runtime: handle inlined calls in runtime.Callers

The `skip` argument passed to runtime.Caller and runtime.Callers should
be interpreted as the number of logical calls to skip (rather than the
number of physical stack frames to skip). This changes runtime.Callers
to skip inlined calls in addition to physical stack frames.

The result value of runtime.Callers is a slice of program counters
([]uintptr) representing physical stack frames. If the `skip` parameter
to runtime.Callers skips part-way into a physical frame, there is no
convenient way to encode that in the resulting slice. To avoid changing
the API in an incompatible way, our solution is to store the number of
skipped logical calls of the first frame in the _second_ uintptr
returned by runtime.Callers. Since this number is a small integer, we
encode it as a valid PC value into a small symbol called:

    runtime.skipPleaseUseCallersFrames

For example, if f() calls g(), g() calls `runtime.Callers(2, pcs)`, and
g() is inlined into f, then the frame for f will be partially skipped,
resulting in the following slice:

    pcs = []uintptr{pc_in_f, runtime.skipPleaseUseCallersFrames+1, ...}

We store the skip PC in pcs[1] instead of pcs[0] so that `pcs[i:]` will
truncate the captured stack trace rather than grow it for all i.

Updates #19348.

Change-Id: I1c56f89ac48c29e6f52a5d085567c6d77d499cf1
Reviewed-on: https://go-review.googlesource.com/37854
Run-TryBot: David Lazar <lazard@golang.org>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Austin Clements <austin@google.com>

src/runtime/asm.s

@@ -14,3 +14,24 @@ GLOBL runtime·no_pointers_stackmap(SB),RODATA, $8
 
 GLOBL runtime·mheap_(SB), NOPTR, $0
 GLOBL runtime·memstats(SB), NOPTR, $0
+
+// NaCl requires that these skips be verifiable machine code.
+#ifdef GOARCH_amd64
+#define SKIP4 BYTE $0x90; BYTE $0x90; BYTE $0x90; BYTE $0x90
+#endif
+#ifdef GOARCH_386
+#define SKIP4 BYTE $0x90; BYTE $0x90; BYTE $0x90; BYTE $0x90
+#endif
+#ifdef GOARCH_amd64p32
+#define SKIP4 BYTE $0x90; BYTE $0x90; BYTE $0x90; BYTE $0x90
+#endif
+#ifndef SKIP4
+#define SKIP4 WORD $0
+#endif
+
+#define SKIP16 SKIP4; SKIP4; SKIP4; SKIP4
+#define SKIP64 SKIP16; SKIP16; SKIP16; SKIP16
+
+// This function must be sizeofSkipFunction bytes.
+TEXT runtime·skipPleaseUseCallersFrames(SB),NOSPLIT,$0-0
+	SKIP64; SKIP64; SKIP64; SKIP64

src/runtime/traceback.go

@@ -114,11 +114,25 @@ func tracebackdefers(gp *g, callback func(*stkframe, unsafe.Pointer) bool, v uns
 	}
 }
 
+const sizeofSkipFunction = 256
+
+// This function is defined in asm.s to be sizeofSkipFunction bytes long.
+func skipPleaseUseCallersFrames()
+
 // Generic traceback. Handles runtime stack prints (pcbuf == nil),
 // the runtime.Callers function (pcbuf != nil), as well as the garbage
 // collector (callback != nil).  A little clunky to merge these, but avoids
 // duplicating the code and all its subtlety.
+//
+// The skip argument is only valid with pcbuf != nil and counts the number
+// of logical frames to skip rather than physical frames (with inlining, a
+// PC in pcbuf can represent multiple calls). If a PC is partially skipped
+// and max > 1, pcbuf[1] will be runtime.skipPleaseUseCallersFrames+N where
+// N indicates the number of logical frames to skip in pcbuf[0].
 func gentraceback(pc0, sp0, lr0 uintptr, gp *g, skip int, pcbuf *uintptr, max int, callback func(*stkframe, unsafe.Pointer) bool, v unsafe.Pointer, flags uint) int {
+	if skip > 0 && callback != nil {
+		throw("gentraceback callback cannot be used with non-zero skip")
+	}
 	if goexitPC == 0 {
 		throw("gentraceback before goexitPC initialization")
 	}
@@ -318,19 +332,57 @@ func gentraceback(pc0, sp0, lr0 uintptr, gp *g, skip int, pcbuf *uintptr, max in
 			_defer = _defer.link
 		}
 
-		if skip > 0 {
-			skip--
-			goto skipped
-		}
-
-		if pcbuf != nil {
-			(*[1 << 20]uintptr)(unsafe.Pointer(pcbuf))[n] = frame.pc
-		}
 		if callback != nil {
 			if !callback((*stkframe)(noescape(unsafe.Pointer(&frame))), v) {
 				return n
 			}
 		}
+
+		if pcbuf != nil {
+			if skip == 0 {
+				(*[1 << 20]uintptr)(unsafe.Pointer(pcbuf))[n] = frame.pc
+			} else {
+				// backup to CALL instruction to read inlining info (same logic as below)
+				tracepc := frame.pc
+				if (n > 0 || flags&_TraceTrap == 0) && frame.pc > f.entry && !waspanic {
+					tracepc--
+				}
+				inldata := funcdata(f, _FUNCDATA_InlTree)
+
+				// no inlining info, skip the physical frame
+				if inldata == nil {
+					skip--
+					goto skipped
+				}
+
+				ix := pcdatavalue(f, _PCDATA_InlTreeIndex, tracepc, &cache)
+				inltree := (*[1 << 20]inlinedCall)(inldata)
+				// skip the logical (inlined) frames
+				logicalSkipped := 0
+				for ix >= 0 && skip > 0 {
+					skip--
+					logicalSkipped++
+					ix = inltree[ix].parent
+				}
+
+				// skip the physical frame if there's more to skip
+				if skip > 0 {
+					skip--
+					goto skipped
+				}
+
+				// now we have a partially skipped frame
+				(*[1 << 20]uintptr)(unsafe.Pointer(pcbuf))[n] = frame.pc
+
+				// if there's room, pcbuf[1] is a skip PC that encodes the number of skipped frames in pcbuf[0]
+				if n+1 < max {
+					n++
+					skipPC := funcPC(skipPleaseUseCallersFrames) + uintptr(logicalSkipped)
+					(*[1 << 20]uintptr)(unsafe.Pointer(pcbuf))[n] = skipPC
+				}
+			}
+		}
+
 		if printing {
 			// assume skip=0 for printing
 			if (flags&_TraceRuntimeFrames) != 0 || showframe(f, gp, nprint == 0) {

test/inline_callers.go

@@ -0,0 +1,72 @@
+// run -gcflags -l=4
+
+// Copyright 2017 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 main
+
+import (
+	"log"
+	"runtime"
+)
+
+var skip int
+var npcs int
+var pcs = make([]uintptr, 32)
+
+func f() {
+	g()
+}
+
+func g() {
+	h()
+}
+
+func h() {
+	npcs = runtime.Callers(skip, pcs)
+}
+
+func testCallers(skp int) (frames []string) {
+	skip = skp
+	f()
+	for i := 0; i < npcs; i++ {
+		fn := runtime.FuncForPC(pcs[i])
+		frames = append(frames, fn.Name())
+		if fn.Name() == "main.main" {
+			break
+		}
+	}
+	return
+}
+
+var expectedFrames [][]string = [][]string{
+	0: {"runtime.Callers", "main.testCallers", "main.main"},
+	1: {"main.testCallers", "main.main"},
+	2: {"main.testCallers", "runtime.skipPleaseUseCallersFrames", "main.main"},
+	3: {"main.testCallers", "runtime.skipPleaseUseCallersFrames", "main.main"},
+	4: {"main.testCallers", "runtime.skipPleaseUseCallersFrames", "main.main"},
+	5: {"main.main"},
+}
+
+func same(xs, ys []string) bool {
+	if len(xs) != len(ys) {
+		return false
+	}
+	for i := range xs {
+		if xs[i] != ys[i] {
+			return false
+		}
+	}
+	return true
+}
+
+func main() {
+	for i := 0; i <= 5; i++ {
+		frames := testCallers(i)
+		expected := expectedFrames[i]
+		if !same(frames, expected) {
+			log.Fatalf("testCallers(%d):\n got %v\n want %v", i, frames, expected)
+		}
+	}
+}