crypto: x86/glue_helper - use le128 instead of u128 for CTR mode

'u128' currently used for CTR mode is on little-endian 'long long' swapped
and would require extra swap operations by SSE/AVX code. Use of le128
instead of u128 allows IV calculations to be done with vector registers
easier.

Signed-off-by: Jussi Kivilinna <jussi.kivilinna@mbnet.fi>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>

arch/x86/crypto/cast6_avx_glue.c

@@ -78,19 +78,19 @@ static void cast6_decrypt_cbc_xway(void *ctx, u128 *dst, const u128 *src)
 		u128_xor(dst + (j + 1), dst + (j + 1), ivs + j);
 }
 
-static void cast6_crypt_ctr(void *ctx, u128 *dst, const u128 *src, u128 *iv)
+static void cast6_crypt_ctr(void *ctx, u128 *dst, const u128 *src, le128 *iv)
 {
 	be128 ctrblk;
 
-	u128_to_be128(&ctrblk, iv);
-	u128_inc(iv);
+	le128_to_be128(&ctrblk, iv);
+	le128_inc(iv);
 
 	__cast6_encrypt(ctx, (u8 *)&ctrblk, (u8 *)&ctrblk);
 	u128_xor(dst, src, (u128 *)&ctrblk);
 }
 
 static void cast6_crypt_ctr_xway(void *ctx, u128 *dst, const u128 *src,
-				   u128 *iv)
+				 le128 *iv)
 {
 	be128 ctrblks[CAST6_PARALLEL_BLOCKS];
 	unsigned int i;
@@ -99,8 +99,8 @@ static void cast6_crypt_ctr_xway(void *ctx, u128 *dst, const u128 *src,
 		if (dst != src)
 			dst[i] = src[i];
 
-		u128_to_be128(&ctrblks[i], iv);
-		u128_inc(iv);
+		le128_to_be128(&ctrblks[i], iv);
+		le128_inc(iv);
 	}
 
 	cast6_enc_blk_xway_xor(ctx, (u8 *)dst, (u8 *)ctrblks);