testing: detect a stopped timer in B.Loop

Currently, if the user stops the timer in a B.Loop benchmark loop, the
benchmark will run until it hits the timeout and fails.

Fix this by detecting that the timer is stopped and failing the
benchmark right away. We avoid making the fast path more expensive for
this check by "poisoning" the B.Loop iteration counter when the timer
is stopped so that it falls back to the slow path, which can check the
timer.

This causes b to escape from B.Loop, which is totally harmless because
it was already definitely heap-allocated. But it causes the
test/inline_testingbloop.go errorcheck test to fail. I don't think the
escape messages actually mattered to that test, they just had to be
matched. To fix this, we drop the debug level to -m=1, since -m=2
prints a lot of extra information for escaping parameters that we
don't want to deal with, and change one error check to allow b to
escape.

Fixes #72971.

Change-Id: I7d4abbb1ec1e096685514536f91ba0d581cca6b7
Reviewed-on: https://go-review.googlesource.com/c/go/+/659657
Auto-Submit: Austin Clements <[email protected]>
Reviewed-by: Junyang Shao <[email protected]>
LUCI-TryBot-Result: Go LUCI <[email protected]>

Author: aclements

src/cmd/compile/internal/inline/interleaved/interleaved.go

@@ -253,7 +253,7 @@ func (s *inlClosureState) mark(n ir.Node) ir.Node {
 
 	if isTestingBLoop(n) {
 		// No inlining nor devirtualization performed on b.Loop body
-		if base.Flag.LowerM > 1 {
+		if base.Flag.LowerM > 0 {
 			fmt.Printf("%v: skip inlining within testing.B.loop for %v\n", ir.Line(n), n)
 		}
 		// We still want to explore inlining opportunities in other parts of ForStmt.

src/cmd/compile/internal/test/inl_test.go

@@ -230,6 +230,9 @@ func TestIntendedInlining(t *testing.T) {
 			"(*Pointer[go.shape.int]).Store",
 			"(*Pointer[go.shape.int]).Swap",
 		},
+		"testing": {
+			"(*B).Loop",
+		},
 	}
 
 	if !goexperiment.SwissMap {

src/testing/benchmark.go

@@ -120,10 +120,13 @@ type B struct {
 		// n is the target number of iterations. It gets bumped up as we go.
 		// When the benchmark loop is done, we commit this to b.N so users can
 		// do reporting based on it, but we avoid exposing it until then.
-		n int
+		n uint64
 		// i is the current Loop iteration. It's strictly monotonically
 		// increasing toward n.
-		i int
+		//
+		// The high bit is used to poison the Loop fast path and fall back to
+		// the slow path.
+		i uint64
 
 		done bool // set when B.Loop return false
 	}
@@ -139,6 +142,7 @@ func (b *B) StartTimer() {
 		b.startBytes = memStats.TotalAlloc
 		b.start = highPrecisionTimeNow()
 		b.timerOn = true
+		b.loop.i &^= loopPoisonTimer
 	}
 }
 
@@ -151,6 +155,8 @@ func (b *B) StopTimer() {
 		b.netAllocs += memStats.Mallocs - b.startAllocs
 		b.netBytes += memStats.TotalAlloc - b.startBytes
 		b.timerOn = false
+		// If we hit B.Loop with the timer stopped, fail.
+		b.loop.i |= loopPoisonTimer
 	}
 }
 
@@ -388,19 +394,32 @@ func (b *B) stopOrScaleBLoop() bool {
 		// Stop the timer so we don't count cleanup time
 		b.StopTimer()
 		// Commit iteration count
-		b.N = b.loop.n
+		b.N = int(b.loop.n)
 		b.loop.done = true
 		return false
 	}
 	// Loop scaling
 	goalns := b.benchTime.d.Nanoseconds()
 	prevIters := int64(b.loop.n)
-	b.loop.n = predictN(goalns, prevIters, t.Nanoseconds(), prevIters)
+	b.loop.n = uint64(predictN(goalns, prevIters, t.Nanoseconds(), prevIters))
+	if b.loop.n&loopPoisonMask != 0 {
+		// The iteration count should never get this high, but if it did we'd be
+		// in big trouble.
+		panic("loop iteration target overflow")
+	}
 	b.loop.i++
 	return true
 }
 
 func (b *B) loopSlowPath() bool {
+	// Consistency checks
+	if !b.timerOn {
+		b.Fatal("B.Loop called with timer stopped")
+	}
+	if b.loop.i&loopPoisonMask != 0 {
+		panic(fmt.Sprintf("unknown loop stop condition: %#x", b.loop.i))
+	}
+
 	if b.loop.n == 0 {
 		// If it's the first call to b.Loop() in the benchmark function.
 		// Allows more precise measurement of benchmark loop cost counts.
@@ -414,14 +433,14 @@ func (b *B) loopSlowPath() bool {
 	}
 	// Handles fixed iterations case
 	if b.benchTime.n > 0 {
-		if b.loop.n < b.benchTime.n {
-			b.loop.n = b.benchTime.n
+		if b.loop.n < uint64(b.benchTime.n) {
+			b.loop.n = uint64(b.benchTime.n)
 			b.loop.i++
 			return true
 		}
 		b.StopTimer()
 		// Commit iteration count
-		b.N = b.loop.n
+		b.N = int(b.loop.n)
 		b.loop.done = true
 		return false
 	}
@@ -463,15 +482,38 @@ func (b *B) loopSlowPath() bool {
 // whereas b.N-based benchmarks must run the benchmark function (and any
 // associated setup and cleanup) several times.
 func (b *B) Loop() bool {
-	// On the first call, both i and n are 0, so we'll fall through to the slow
-	// path in that case, too.
+	// This is written such that the fast path is as fast as possible and can be
+	// inlined.
+	//
+	// There are three cases where we'll fall out of the fast path:
+	//
+	// - On the first call, both i and n are 0.
+	//
+	// - If the loop reaches the n'th iteration, then i == n and we need
+	//   to figure out the new target iteration count or if we're done.
+	//
+	// - If the timer is stopped, it poisons the top bit of i so the slow
+	//   path can do consistency checks and fail.
 	if b.loop.i < b.loop.n {
 		b.loop.i++
 		return true
 	}
 	return b.loopSlowPath()
 }
 
+// The loopPoison constants can be OR'd into B.loop.i to cause it to fall back
+// to the slow path.
+const (
+	loopPoisonTimer = uint64(1 << (63 - iota))
+	// If necessary, add more poison bits here.
+
+	// loopPoisonMask is the set of all loop poison bits. (iota-1) is the index
+	// of the bit we just set, from which we recreate that bit mask. We subtract
+	// 1 to set all of the bits below that bit, then complement the result to
+	// get the mask. Sorry, not sorry.
+	loopPoisonMask = ^uint64((1 << (63 - (iota - 1))) - 1)
+)
+
 // BenchmarkResult contains the results of a benchmark run.
 type BenchmarkResult struct {
 	N         int           // The number of iterations.

src/testing/loop_test.go

@@ -129,3 +129,26 @@ func TestBenchmarkBLoopError(t *T) {
 		t.Errorf("want N == 0, got %d", bRet.N)
 	}
 }
+
+func TestBenchmarkBLoopStop(t *T) {
+	var bState *B
+	var bLog bytes.Buffer
+	bRet := Benchmark(func(b *B) {
+		bState = b
+		b.common.w = &bLog
+
+		for i := 0; b.Loop(); i++ {
+			b.StopTimer()
+		}
+	})
+	if !bState.failed {
+		t.Errorf("benchmark should have failed")
+	}
+	const wantLog = "B.Loop called with timer stopped"
+	if log := bLog.String(); !strings.Contains(log, wantLog) {
+		t.Errorf("missing error %q in output:\n%s", wantLog, log)
+	}
+	if bRet.N != 0 {
+		t.Errorf("want N == 0, got %d", bRet.N)
+	}
+}

test/inline_testingbloop.go

@@ -1,4 +1,4 @@
-// errorcheck -0 -m=2
+// errorcheck -0 -m
 
 // Copyright 2024 The Go Authors. All rights reserved.
 // Use of this source code is governed by a BSD-style
@@ -15,7 +15,7 @@ func caninline(x int) int { // ERROR "can inline caninline"
 	return x
 }
 
-func cannotinline(b *testing.B) { // ERROR "b does not escape" "cannot inline cannotinline.*"
+func test(b *testing.B) { // ERROR "leaking param: b"
 	for i := 0; i < b.N; i++ {
 		caninline(1) // ERROR "inlining call to caninline"
 	}