UPSTREAM: crypto: arm/blake2s - fix for big endian

The new ARM BLAKE2s code doesn't work correctly (fails the self-tests)
in big endian kernel builds because it doesn't swap the endianness of
the message words when loading them.  Fix this.

Fixes: 5172d322d34c ("crypto: arm/blake2s - add ARM scalar optimized BLAKE2s")
Signed-off-by: Eric Biggers <ebiggers@google.com>
Acked-by: Ard Biesheuvel <ardb@kernel.org>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>

(cherry picked from commit d2f2516a3882c0c6463e33c9b112b39bd483f821)
Bug: 178411248
Change-Id: Iff419963407eeb0435b9550ef126db4bf137a8a1
Signed-off-by: Eric Biggers <ebiggers@google.com>