Fix large allocation to search optimal size class heap.

Fix arena_run_alloc_large_helper() to not convert size to usize when
searching for the first best fit via arena_run_first_best_fit().  This
allows the search to consider the optimal quantized size class, so that
e.g. allocating and deallocating 40 KiB in a tight loop can reuse the
same memory.

This regression was nominally caused by
5707d6f (Quantize szad trees by size
class.), but it did not commonly cause problems until
8a03cf0 (Implement cache index
randomization for large allocations.).  These regressions were first
released in 4.0.0.

This resolves jemalloc#487.

Author: jasone

src/arena.c

@@ -1059,7 +1059,7 @@ arena_run_first_best_fit(arena_t *arena, size_t size)
 static arena_run_t *
 arena_run_alloc_large_helper(arena_t *arena, size_t size, bool zero)
 {
-	arena_run_t *run = arena_run_first_best_fit(arena, s2u(size));
+	arena_run_t *run = arena_run_first_best_fit(arena, size);
 	if (run != NULL) {
 		if (arena_run_split_large(arena, run, size, zero))
 			run = NULL;