runtime: set MSpan.limit properly for large spans.

Then use the limit to make sure MHeap_LookupMaybe & inlined
copies don't return a span if the pointer is beyond the limit.
Use this fact to optimize all call sites.

R=golang-dev, dvyukov
CC=golang-dev
https://golang.org/cl/9869045

src/pkg/runtime/malloc.goc

@@ -84,6 +84,7 @@ runtime·mallocgc(uintptr size, uint32 flag, int32 dogc, int32 zeroed)
 		s = runtime·MHeap_Alloc(&runtime·mheap, npages, 0, 1, zeroed);
 		if(s == nil)
 			runtime·throw("out of memory");
+		s->limit = (byte*)(s->start<<PageShift) + size;
 		size = npages<<PageShift;
 		c->local_alloc += size;
 		c->local_total_alloc += size;
@@ -238,11 +239,6 @@ runtime·mlookup(void *v, byte **base, uintptr *size, MSpan **sp)
 		return 1;
 	}
 
-	if((byte*)v >= (byte*)s->limit) {
-		// pointers past the last block do not count as pointers.
-		return 0;
-	}
-
 	n = s->elemsize;
 	if(base) {
 		i = ((byte*)v - p)/n;

src/pkg/runtime/mgc0.c

@@ -231,14 +231,12 @@ markonly(void *obj)
 	if(sizeof(void*) == 8)
 		x -= (uintptr)runtime·mheap.arena_start>>PageShift;
 	s = runtime·mheap.spans[x];
-	if(s == nil || k < s->start || k - s->start >= s->npages || s->state != MSpanInUse)
+	if(s == nil || k < s->start || obj >= s->limit || s->state != MSpanInUse)
 		return false;
 	p = (byte*)((uintptr)s->start<<PageShift);
 	if(s->sizeclass == 0) {
 		obj = p;
 	} else {
-		if((byte*)obj >= (byte*)s->limit)
-			return false;
 		uintptr size = s->elemsize;
 		int32 i = ((byte*)obj - p)/size;
 		obj = p+i*size;
@@ -411,14 +409,12 @@ flushptrbuf(PtrTarget *ptrbuf, PtrTarget **ptrbufpos, Obj **_wp, Workbuf **_wbuf
 			if(sizeof(void*) == 8)
 				x -= (uintptr)arena_start>>PageShift;
 			s = runtime·mheap.spans[x];
-			if(s == nil || k < s->start || k - s->start >= s->npages || s->state != MSpanInUse)
+			if(s == nil || k < s->start || obj >= s->limit || s->state != MSpanInUse)
 				continue;
 			p = (byte*)((uintptr)s->start<<PageShift);
 			if(s->sizeclass == 0) {
 				obj = p;
 			} else {
-				if((byte*)obj >= (byte*)s->limit)
-					continue;
 				size = s->elemsize;
 				int32 i = ((byte*)obj - p)/size;
 				obj = p+i*size;
@@ -1173,8 +1169,6 @@ debug_scanblock(byte *b, uintptr n)
 		if(s->sizeclass == 0) {
 			obj = p;
 		} else {
-			if((byte*)obj >= (byte*)s->limit)
-				continue;
 			int32 i = ((byte*)obj - p)/size;
 			obj = p+i*size;
 		}

src/pkg/runtime/mheap.c

@@ -303,9 +303,7 @@ runtime·MHeap_LookupMaybe(MHeap *h, void *v)
 	if(sizeof(void*) == 8)
 		q -= (uintptr)h->arena_start >> PageShift;
 	s = h->spans[q];
-	if(s == nil || p < s->start || p - s->start >= s->npages)
-		return nil;
-	if(s->state != MSpanInUse)
+	if(s == nil || p < s->start || v >= s->limit || s->state != MSpanInUse)
 		return nil;
 	return s;
 }