runtime: cache unrolled GC bitmask

Cache unrolled GC bitmask for types up to 64/32K on 64/32-bit systems,
this corresponds to up to 4K cached bitmask.
Perf builders say that 2% of time is spent in unrollgcproginplace_m/unrollgcprog1
on http benchmark:
http://goperfd.appspot.com/log/f42045f45bf61a0da53b724a7c8567824a0ad6c9

LGTM=rsc
R=golang-codereviews, rsc
CC=golang-codereviews, khr, rlh
https://golang.org/cl/122680043

src/pkg/runtime/malloc.go

@@ -28,7 +28,7 @@ const (
 	bitsMask           = 1<<bitsPerPointer - 1
 	pointersPerByte    = 8 / bitsPerPointer
 	bitPtrMask         = bitsMask << 2
-	maxGCMask          = 0 // disabled because wastes several bytes of memory
+	maxGCMask          = 64
 	bitsDead           = 0
 	bitsPointer        = 2
 

src/pkg/runtime/mgc0.h

@@ -55,7 +55,8 @@ enum {
 	BitsIface	= 2,
 	BitsEface	= 3,
 
-	MaxGCMask	= 0,	// disabled because wastes several bytes of memory
+	// 64 bytes cover objects of size 1024/512 on 64/32 bits, respectively.
+	MaxGCMask	= 64,
 };
 
 // Bits in per-word bitmap.