Memory Alignment issues on ARM processors

Unaligned data on ARM architectures do not perform efficiently unaligned
memory access, and in the case of ARMv7 and iOS it completely breaks. The OCB
algorithm dereferences a uint64x2_t pointer, and is replaced by a
memcpy to avoid penalties when trying to align it.

More info https://brewx.qualcomm.com/bws/content/gi/common/appseng/en/knowledgebase/docs/kb95.html

src/crypto/ocb.cc

@@ -271,8 +271,9 @@
 	static block gen_offset(uint64_t KtopStr[3], unsigned bot) {
 		const union { unsigned x; unsigned char endian; } little = { 1 };
 		const int64x2_t k64 = {-64,-64};
-		uint64x2_t hi = *(uint64x2_t *)(KtopStr+0);   /* hi = A B */
-		uint64x2_t lo = *(uint64x2_t *)(KtopStr+1);   /* hi = B C */
+		uint64x2_t hi, lo;
+		memcpy(&hi, KtopStr, sizeof(hi));
+		memcpy(&lo, KtopStr+1, sizeof(lo));
 		int64x2_t ls = vdupq_n_s64(bot);
 		int64x2_t rs = vqaddq_s64(k64,ls);
 		block rval = (block)veorq_u64(vshlq_u64(hi,ls),vshlq_u64(lo,rs));