lib/mbedtls
1
0
Fork 0
mirror of https://github.com/Mbed-TLS/mbedtls.git synced 2025-08-01 10:06:53 +03:00
Code Activity

Unroll aesce_decrypt_block

Browse Source 
Signed-off-by: Dave Rodgman <dave.rodgman@arm.com>
This commit is contained in:
Dave Rodgman
2023-06-15 16:28:00 +01:00
parent 96fdfb8e62
commit 1c4451d089
1 changed files with 65 additions and 19 deletions
Download Patch File Download Diff File Expand all files Collapse all files

56
library/aesce.c
View File

@ -171,10 +171,16 @@ static uint8x16_t aesce_decrypt_block(uint8x16_t block,
unsigned char *keys, unsigned char *keys,
int rounds) int rounds)
{ {
/* Assume either 10, 12 or 14 rounds */
if (rounds == 10) {
goto rounds_10;
}
if (rounds == 12) {
goto rounds_12;
}
for (int i = 0; i < rounds - 1; i++) {
/* AES AddRoundKey, SubBytes, ShiftRows */ /* AES AddRoundKey, SubBytes, ShiftRows */
block = vaesdq_u8(block, vld1q_u8(keys + i * 16)); block = vaesdq_u8(block, vld1q_u8(keys));
/* AES inverse MixColumns for the next round. /* AES inverse MixColumns for the next round.
* *
* This means that we switch the order of the inverse AddRoundKey and * This means that we switch the order of the inverse AddRoundKey and
@ -188,14 +194,54 @@ static uint8x16_t aesce_decrypt_block(uint8x16_t block,
* affected round keys separately which has been done when the * affected round keys separately which has been done when the
* decryption round keys were calculated.) */ * decryption round keys were calculated.) */
block = vaesimcq_u8(block); block = vaesimcq_u8(block);
} keys += 16;
block = vaesdq_u8(block, vld1q_u8(keys));
block = vaesimcq_u8(block);
keys += 16;
rounds_12:
block = vaesdq_u8(block, vld1q_u8(keys));
block = vaesimcq_u8(block);
keys += 16;
block = vaesdq_u8(block, vld1q_u8(keys));
block = vaesimcq_u8(block);
keys += 16;
rounds_10:
block = vaesdq_u8(block, vld1q_u8(keys));
block = vaesimcq_u8(block);
keys += 16;
block = vaesdq_u8(block, vld1q_u8(keys));
block = vaesimcq_u8(block);
keys += 16;
block = vaesdq_u8(block, vld1q_u8(keys));
block = vaesimcq_u8(block);
keys += 16;
block = vaesdq_u8(block, vld1q_u8(keys));
block = vaesimcq_u8(block);
keys += 16;
block = vaesdq_u8(block, vld1q_u8(keys));
block = vaesimcq_u8(block);
keys += 16;
block = vaesdq_u8(block, vld1q_u8(keys));
block = vaesimcq_u8(block);
keys += 16;
block = vaesdq_u8(block, vld1q_u8(keys));
block = vaesimcq_u8(block);
keys += 16;
block = vaesdq_u8(block, vld1q_u8(keys));
block = vaesimcq_u8(block);
keys += 16;
block = vaesdq_u8(block, vld1q_u8(keys));
block = vaesimcq_u8(block);
keys += 16;
/* The inverses of AES AddRoundKey, SubBytes, ShiftRows finishing up the /* The inverses of AES AddRoundKey, SubBytes, ShiftRows finishing up the
* last full round. */ * last full round. */
block = vaesdq_u8(block, vld1q_u8(keys + (rounds - 1) * 16)); block = vaesdq_u8(block, vld1q_u8(keys));
keys += 16;
/* Inverse AddRoundKey for inverting the initial round key addition. */ /* Inverse AddRoundKey for inverting the initial round key addition. */
block = veorq_u8(block, vld1q_u8(keys + rounds * 16)); block = veorq_u8(block, vld1q_u8(keys));
return block; return block;
} }