runtime: move stack shrinking until after sweepgen is incremented.

Before GC, we flush all the per-P allocation caches.  Doing
stack shrinking mid-GC causes these caches to fill up.  At the
end of gc, the sweepgen is incremented which causes all of the
data in these caches to be in a bad state (cached but not yet
swept).

Move the stack shrinking until after sweepgen is incremented,
so any caching that happens as part of shrinking is done with
already-swept data.

Reenable stack copying.

LGTM=bradfitz
R=golang-codereviews, bradfitz
CC=golang-codereviews
https://golang.org/cl/69620043

src/pkg/runtime/mgc0.c

@@ -1601,9 +1601,6 @@ addstackroots(G *gp, Workbuf **wbufp)
 	if((mp = gp->m) != nil && mp->helpgc)
 		runtime·throw("can't scan gchelper stack");
 
-	// Shrink stack if not much of it is being used.
-	runtime·shrinkstack(gp);
-
 	if(gp->syscallstack != (uintptr)nil) {
 		// Scanning another goroutine that is about to enter or might
 		// have just exited a system call. It may be executing code such
@@ -2426,6 +2423,11 @@ gc(struct gc_args *args)
 			gcstats.npausesweep++;
 	}
 
+	// Shrink a stack if not much of it is being used.
+	// TODO: do in a parfor
+	for(i = 0; i < runtime·allglen; i++)
+		runtime·shrinkstack(runtime·allg[i]);
+
 	runtime·MProf_GC();
 }
 

src/pkg/runtime/proc.c

@@ -174,7 +174,6 @@ runtime·schedinit(void)
 	procresize(procs);
 
 	runtime·copystack = runtime·precisestack;
-	runtime·copystack = false; // TODO: remove
 	p = runtime·getenv("GOCOPYSTACK");
 	if(p != nil && !runtime·strcmp(p, (byte*)"0"))
 		runtime·copystack = false;