runtime: implement MemStats.NumGC

Reading runtime.MemStats.NumGC is common enough in dependency code that its
absence is a compile error for programs that never look at the value. Rather
than add the field as a constant zero, track it:

- gc_blocks: count completed cycles in runGC, so collections triggered by an
  allocation are counted as well as explicit runtime.GC() calls. The counter is
  read and written under gcLock, like the other counters beside it.
- gc_boehm: report bdwgc's own gc_no from the prof_stats struct.
- gc_leaking: always 0, since that collector never completes a cycle.

src/runtime/gc_blocks.go

@@ -56,6 +56,7 @@ var (
 	endBlock      gcBlock        // the block just past the end of the available space
 	gcTotalAlloc  uint64         // total number of bytes allocated
 	gcMallocs     uint64         // total number of allocations
+	gcNumGC       uint32         // total number of completed collection cycles
 	gcLock        task.PMutex    // lock to avoid race conditions on multicore systems
 )
 
@@ -610,6 +611,11 @@ func runGC() (freeBytes uintptr) {
 		dumpHeap()
 	}
 
+	// The cycle is complete. Counted here rather than in GC() so that
+	// collections triggered by an allocation are counted too. Every caller
+	// holds gcLock, the same lock ReadMemStats reads it under.
+	gcNumGC++
+
 	return
 }
 
@@ -850,6 +856,9 @@ func ReadMemStats(m *MemStats) {
 	// Record the total allocated bytes.
 	m.TotalAlloc = gcTotalAlloc
 
+	// Record the number of completed collection cycles.
+	m.NumGC = gcNumGC
+
 	gcLock.Unlock()
 }