crypto/md5: fix md5block asm to work on big endian ppc64

crypto/md5: fix md5block asm to work on big endian ppc64

This handles a TODO in the md5block_ppc64le.s file to
make use of byte reverse loads so the function works for
big endian as well as little endian. File name is now
md5block_ppc64x.s.

name                 old time/op    new time/op    delta
Hash8Bytes              537ns ± 0%     299ns ± 0%  -44.32%  (p=0.000 n=8+1)
Hash1K                 4.74µs ± 0%    3.24µs ± 0%  -31.64%  (p=0.250 n=7+1)
Hash8K                 34.4µs ± 0%    23.6µs ± 0%  -31.56%  (p=0.222 n=8+1)
Hash8BytesUnaligned     537ns ± 0%     298ns ± 0%  -44.51%  (p=0.000 n=8+1)
Hash1KUnaligned        4.74µs ± 0%    3.24µs ± 0%  -31.48%  (p=0.222 n=8+1)
Hash8KUnaligned        34.4µs ± 0%    23.6µs ± 0%  -31.39%  (p=0.222 n=8+1)

name                 old speed      new speed      delta
Hash8Bytes           14.9MB/s ± 0%  26.8MB/s ± 0%  +79.76%  (p=0.222 n=8+1)
Hash1K                216MB/s ± 0%   316MB/s ± 0%  +46.29%  (p=0.250 n=7+1)
Hash8K                238MB/s ± 0%   348MB/s ± 0%  +46.11%  (p=0.222 n=8+1)
Hash8BytesUnaligned  14.9MB/s ± 0%  26.8MB/s ± 0%  +79.76%  (p=0.222 n=8+1)
Hash1KUnaligned       216MB/s ± 0%   316MB/s ± 0%  +45.95%  (p=0.222 n=8+1)
Hash8KUnaligned       238MB/s ± 0%   347MB/s ± 0%  +45.75%  (p=0.222 n=8+1)

Change-Id: I2e226bf7e69e0acd49db1af42e4fd8b87b155606
Reviewed-on: https://go-review.googlesource.com/c/144599
Run-TryBot: Lynn Boger <laboger@linux.vnet.ibm.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Carlos Eduardo Seo <cseo@linux.vnet.ibm.com>
Reviewed-by: Michael Munday <mike.munday@ibm.com>