The performance gain is tiny, but measurable.
if (plain_text != 0) {
FOR (j, 0, 16) {
u32 plain = load32_le(plain_text);
- store32_le(cipher_text + j * 4, ctx->pool[j] ^ plain);
- plain_text += 4;
+ store32_le(cipher_text, ctx->pool[j] ^ plain);
+ plain_text += 4;
+ cipher_text += 4;
}
} else {
FOR (j, 0, 16) {
- store32_le(cipher_text + j * 4, ctx->pool[j]);
+ store32_le(cipher_text, ctx->pool[j]);
+ cipher_text += 4;
}
}
- cipher_text += 64;
ctx->pool_idx = 64;
}