runtime: adjust malloc race instrumentation for tiny allocs

Tiny alloc memory block is shared by different goroutines running on the same thread.
We call racemalloc after enabling preemption in mallocgc,
as the result another goroutine can act on not yet race-cleared tiny block.
Call racemalloc before enabling preemption.
Fixes #7224.

LGTM=dave
R=golang-codereviews, dave
CC=golang-codereviews
https://golang.org/cl/57730043

src/pkg/runtime/malloc.goc

@@ -182,6 +182,8 @@ runtime·mallocgc(uintptr size, uintptr typ, uint32 flag)
 	m->mallocing = 0;
 	if(UseSpanType && !(flag & FlagNoScan) && typ != 0 && m->settype_bufsize == nelem(m->settype_buf))
 		runtime·settype_flush(m);
+	if(raceenabled)
+		runtime·racemalloc(v, size);
 	m->locks--;
 	if(m->locks == 0 && g->preempt)  // restore the preemption request in case we've cleared it in newstack
 		g->stackguard0 = StackPreempt;
@@ -208,8 +210,6 @@ runtime·mallocgc(uintptr size, uintptr typ, uint32 flag)
 	if(!(flag & FlagNoInvokeGC) && mstats.heap_alloc >= mstats.next_gc)
 		runtime·gc(0);
 
-	if(raceenabled)
-		runtime·racemalloc(v, size);
 	return v;
 }
 

src/pkg/runtime/race/testdata/mop_test.go

@@ -1933,3 +1933,25 @@ func TestRaceMethodThunk4(t *testing.T) {
 	*(*int)(d.Base) = 42
 	<-done
 }
+
+func TestNoRaceTinyAlloc(t *testing.T) {
+	const P = 4
+	const N = 1e6
+	var tinySink *byte
+	done := make(chan bool)
+	for p := 0; p < P; p++ {
+		go func() {
+			for i := 0; i < N; i++ {
+				var b byte
+				if b != 0 {
+					tinySink = &b // make it heap allocated
+				}
+				b = 42
+			}
+			done <- true
+		}()
+	}
+	for p := 0; p < P; p++ {
+		<-done
+	}
+}