alpha: simplify and optimize sched_find_first_bit

Search only the first 100 bits instead of 140, saving a couple
instructions. The resulting code is about 1/3 faster (40K ticks/1000
iterations down to 30K ticks/1000 iterations).

Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Ivan Kokshaysky <ink@jurassic.park.msu.ru>
Cc: linux-alpha@vger.kernel.org
Acked-by: Richard Henderson <rth@twiddle.net>
Signed-off-by: Matt Turner <mattst88@gmail.com>

arch/alpha/include/asm/bitops.h

@@ -438,22 +438,20 @@ static inline unsigned int __arch_hweight8(unsigned int w)
 
 /*
  * Every architecture must define this function. It's the fastest
- * way of searching a 140-bit bitmap where the first 100 bits are
+ * way of searching a 100-bit bitmap.  It's guaranteed that at least
- * unlikely to be set. It's guaranteed that at least one of the 140
+ * one of the 100 bits is cleared.
- * bits is set.
  */
 static inline unsigned long
-sched_find_first_bit(unsigned long b[3])
+sched_find_first_bit(const unsigned long b[2])
 {
-	unsigned long b0 = b[0], b1 = b[1], b2 = b[2];
+	unsigned long b0, b1, ofs, tmp;
-	unsigned long ofs;
 
-	ofs = (b1 ? 64 : 128);
+	b0 = b[0];
-	b1 = (b1 ? b1 : b2);
+	b1 = b[1];
-	ofs = (b0 ? 0 : ofs);
+	ofs = (b0 ? 0 : 64);
-	b0 = (b0 ? b0 : b1);
+	tmp = (b0 ? b0 : b1);
 
-	return __ffs(b0) + ofs;
+	return __ffs(tmp) + ofs;
 }
 
 #include <asm-generic/bitops/ext2-non-atomic.h>