diff --git a/src/runtime/export_test.go b/src/runtime/export_test.go
index 25758972f13..31d32eabebd 100644
--- a/src/runtime/export_test.go
+++ b/src/runtime/export_test.go
@@ -483,7 +483,9 @@ func LockOSCounts() (external, internal uint32) {
 func TracebackSystemstack(stk []uintptr, i int) int {
 	if i == 0 {
 		pc, sp := getcallerpc(), getcallersp()
-		return gentraceback(pc, sp, 0, getg(), 0, &stk[0], len(stk), nil, nil, _TraceJumpStack)
+		var u unwinder
+		u.initAt(pc, sp, 0, getg(), unwindJumpStack) // Don't ignore errors, for testing
+		return tracebackPCs(&u, 0, stk)
 	}
 	n := 0
 	systemstack(func() {
diff --git a/src/runtime/mprof.go b/src/runtime/mprof.go
index cf9e0329816..dfaa369740b 100644
--- a/src/runtime/mprof.go
+++ b/src/runtime/mprof.go
@@ -1173,7 +1173,9 @@ func GoroutineProfile(p []StackRecord) (n int, ok bool) {
 }
 
 func saveg(pc, sp uintptr, gp *g, r *StackRecord) {
-	n := gentraceback(pc, sp, 0, gp, 0, &r.Stack0[0], len(r.Stack0), nil, nil, 0)
+	var u unwinder
+	u.initAt(pc, sp, 0, gp, unwindSilentErrors)
+	n := tracebackPCs(&u, 0, r.Stack0[:])
 	if n < len(r.Stack0) {
 		r.Stack0[n] = 0
 	}
diff --git a/src/runtime/proc.go b/src/runtime/proc.go
index ee13debf545..533cf208380 100644
--- a/src/runtime/proc.go
+++ b/src/runtime/proc.go
@@ -4687,6 +4687,7 @@ func sigprof(pc, sp, lr uintptr, gp *g, mp *m) {
 	// See golang.org/issue/17165.
 	getg().m.mallocing++
 
+	var u unwinder
 	var stk [maxCPUProfStack]uintptr
 	n := 0
 	if mp.ncgo > 0 && mp.curg != nil && mp.curg.syscallpc != 0 && mp.curg.syscallsp != 0 {
@@ -4700,26 +4701,24 @@ func sigprof(pc, sp, lr uintptr, gp *g, mp *m) {
 			for cgoOff < len(mp.cgoCallers) && mp.cgoCallers[cgoOff] != 0 {
 				cgoOff++
 			}
-			copy(stk[:], mp.cgoCallers[:cgoOff])
+			n += copy(stk[:], mp.cgoCallers[:cgoOff])
 			mp.cgoCallers[0] = 0
 		}
 
 		// Collect Go stack that leads to the cgo call.
-		n = gentraceback(mp.curg.syscallpc, mp.curg.syscallsp, 0, mp.curg, 0, &stk[cgoOff], len(stk)-cgoOff, nil, nil, 0)
-		if n > 0 {
-			n += cgoOff
-		}
+		u.initAt(mp.curg.syscallpc, mp.curg.syscallsp, 0, mp.curg, unwindSilentErrors)
 	} else if usesLibcall() && mp.libcallg != 0 && mp.libcallpc != 0 && mp.libcallsp != 0 {
 		// Libcall, i.e. runtime syscall on windows.
 		// Collect Go stack that leads to the call.
-		n = gentraceback(mp.libcallpc, mp.libcallsp, 0, mp.libcallg.ptr(), 0, &stk[n], len(stk[n:]), nil, nil, 0)
+		u.initAt(mp.libcallpc, mp.libcallsp, 0, mp.libcallg.ptr(), unwindSilentErrors)
 	} else if mp != nil && mp.vdsoSP != 0 {
 		// VDSO call, e.g. nanotime1 on Linux.
 		// Collect Go stack that leads to the call.
-		n = gentraceback(mp.vdsoPC, mp.vdsoSP, 0, gp, 0, &stk[n], len(stk[n:]), nil, nil, _TraceJumpStack)
+		u.initAt(mp.vdsoPC, mp.vdsoSP, 0, gp, unwindSilentErrors|unwindJumpStack)
 	} else {
-		n = gentraceback(pc, sp, lr, gp, 0, &stk[0], len(stk), nil, nil, _TraceTrap|_TraceJumpStack)
+		u.initAt(pc, sp, lr, gp, unwindSilentErrors|unwindTrap|unwindJumpStack)
 	}
+	n += tracebackPCs(&u, 0, stk[n:])
 
 	if n <= 0 {
 		// Normal traceback is impossible or has failed.
diff --git a/src/runtime/traceback.go b/src/runtime/traceback.go
index 968823316e4..b100a3c3b21 100644
--- a/src/runtime/traceback.go
+++ b/src/runtime/traceback.go
@@ -591,27 +591,64 @@ func (u *unwinder) cgoCallers(pcBuf []uintptr) int {
 	return len(pcBuf)
 }
 
-// Generic traceback. Handles runtime stack prints (pcbuf == nil),
-// and the runtime.Callers function (pcbuf != nil).
-// A little clunky to merge these, but avoids
-// duplicating the code and all its subtlety.
+// tracebackPCs populates pcBuf with the return addresses for each frame from u
+// and returns the number of PCs written to pcBuf. The returned PCs correspond
+// to "logical frames" rather than "physical frames"; that is if A is inlined
+// into B, this will still return a PCs for both A and B. This also includes PCs
+// generated by the cgo unwinder, if one is registered.
+//
+// If skip != 0, this skips this many logical frames.
+//
+// Callers should set the unwindSilentErrors flag on u.
+func tracebackPCs(u *unwinder, skip int, pcBuf []uintptr) int {
+	var cgoBuf [32]uintptr
+	n := 0
+	for ; n < len(pcBuf) && u.valid(); u.next() {
+		f := u.frame.fn
+		cgoN := u.cgoCallers(cgoBuf[:])
+
+		// TODO: Why does &u.cache cause u to escape?
+		for iu, uf := newInlineUnwinder(f, u.symPC(), noEscapePtr(&u.cache)); n < len(pcBuf) && uf.valid(); uf = iu.next(uf) {
+			sf := iu.srcFunc(uf)
+			if sf.funcID == funcID_wrapper && elideWrapperCalling(u.calleeFuncID) {
+				// ignore wrappers
+			} else if skip > 0 {
+				skip--
+			} else {
+				// Callers expect the pc buffer to contain return addresses
+				// and do the -1 themselves, so we add 1 to the call PC to
+				// create a return PC.
+				pcBuf[n] = uf.pc + 1
+				n++
+			}
+			u.calleeFuncID = sf.funcID
+		}
+		// Add cgo frames (if we're done skipping over the requested number of
+		// Go frames).
+		if skip == 0 {
+			n += copy(pcBuf[n:], cgoBuf[:cgoN])
+		}
+	}
+	return n
+}
+
+// Generic traceback. Handles runtime stack prints (pcbuf == nil).
 //
 // 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).
 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 pcbuf != nil {
+		throw("pcbuf argument no longer supported")
+	}
 	if callback != nil {
 		throw("callback argument no longer supported")
 	}
 
 	// Translate flags
 	var uflags unwindFlags
-	printing := pcbuf == nil && callback == nil
-	if printing {
-		uflags |= unwindPrintErrors
-	} else {
-		uflags |= unwindSilentErrors
-	}
+	printing := true
+	uflags |= unwindPrintErrors
 	if flags&_TraceTrap != 0 {
 		uflags |= unwindTrap
 	}
@@ -634,33 +671,6 @@ func gentraceback(pc0, sp0, lr0 uintptr, gp *g, skip int, pcbuf *uintptr, max in
 
 		cgoN := u.cgoCallers(cgoBuf[:])
 
-		if pcbuf != nil {
-			// TODO: Why does cache escape? (Same below)
-			for iu, uf := newInlineUnwinder(f, u.symPC(), noEscapePtr(&u.cache)); uf.valid(); uf = iu.next(uf) {
-				sf := iu.srcFunc(uf)
-				if sf.funcID == funcID_wrapper && elideWrapperCalling(u.calleeFuncID) {
-					// ignore wrappers
-				} else if skip > 0 {
-					skip--
-				} else if n < max {
-					// Callers expect the pc buffer to contain return addresses
-					// and do the -1 themselves, so we add 1 to the call PC to
-					// create a return PC.
-					(*[1 << 20]uintptr)(unsafe.Pointer(pcbuf))[n] = uf.pc + 1
-					n++
-				}
-				u.calleeFuncID = sf.funcID
-			}
-			// Add cgo frames
-			if skip == 0 { // skip only applies to Go frames
-				for i := 0; i < cgoN && n < max; i++ {
-					(*[1 << 20]uintptr)(unsafe.Pointer(pcbuf))[n] = cgoBuf[i]
-					n++
-				}
-			}
-			n-- // offset n++ below
-		}
-
 		if printing {
 			// assume skip=0 for printing.
 			//
@@ -981,13 +991,17 @@ func callers(skip int, pcbuf []uintptr) int {
 	gp := getg()
 	var n int
 	systemstack(func() {
-		n = gentraceback(pc, sp, 0, gp, skip, &pcbuf[0], len(pcbuf), nil, nil, 0)
+		var u unwinder
+		u.initAt(pc, sp, 0, gp, unwindSilentErrors)
+		n = tracebackPCs(&u, skip, pcbuf)
 	})
 	return n
 }
 
 func gcallers(gp *g, skip int, pcbuf []uintptr) int {
-	return gentraceback(^uintptr(0), ^uintptr(0), 0, gp, skip, &pcbuf[0], len(pcbuf), nil, nil, 0)
+	var u unwinder
+	u.init(gp, unwindSilentErrors)
+	return tracebackPCs(&u, skip, pcbuf)
 }
 
 // showframe reports whether the frame with the given characteristics should