New pow implementation

The new implementation is provided under !__OBSOLETE_MATH, it uses
ISO C99 code.  With default settings the worst case error in nearest
rounding mode is 0.54 ULP with inlined fma and fma contraction.  It uses
a 4 KB lookup table in addition to the table in exp_data.c, on aarch64
.text+.rodata size of libm.a is increased by 2295 bytes.

Improvements on Cortex-A72:
latency: 3.3x
thruput: 4.9x

newlib/libm/math/w_pow.c

@@ -52,6 +52,7 @@ PORTABILITY
  */
 
 #include "fdlibm.h"
+#if __OBSOLETE_MATH
 #include <errno.h>
 
 #ifndef _DOUBLE_IS_32BITS
@@ -207,7 +208,7 @@ PORTABILITY
 }
 
 #endif /* defined(_DOUBLE_IS_32BITS) */
-
+#endif /* __OBSOLETE_MATH */