Move arena_slab_data_t's nfree into extent_t's e_bits.

Compact extent_t to 128 bytes on 64-bit systems by moving
arena_slab_data_t's nfree into extent_t's e_bits.

Cacheline-align extent_t structures so that they always cross the
minimum number of cacheline boundaries.

Re-order extent_t fields such that all fields except the slab bitmap
(and overlaid heap profiling context pointer) are in the first
cacheline.

This resolves #461.

include/jemalloc/internal/arena_structs_a.h

@@ -2,9 +2,6 @@
 #define JEMALLOC_INTERNAL_ARENA_STRUCTS_A_H
 
 struct arena_slab_data_s {
-	/* Number of free regions in slab. */
-	unsigned	nfree;
-
 	/* Per region allocated/deallocated bitmap. */
 	bitmap_t	bitmap[BITMAP_GROUPS_MAX];
 };