After review fixes

1. Formating
2. Check config added
3. Dependency description fixed

Signed-off-by: Shelly Liberman <shelly.liberman@arm.com>
This commit is contained in:
Shelly Liberman 2020-11-26 22:48:23 +02:00
parent 44b4229352
commit 11c64885a6
3 changed files with 240 additions and 179 deletions

View File

@ -78,6 +78,12 @@
#error "MBEDTLS_CTR_DRBG_C and MBEDTLS_AES_ONLY_128_BIT_KEY_LENGTH defined, but MBEDTLS_CTR_DRBG_USE_128_BIT_KEY is not defined" #error "MBEDTLS_CTR_DRBG_C and MBEDTLS_AES_ONLY_128_BIT_KEY_LENGTH defined, but MBEDTLS_CTR_DRBG_USE_128_BIT_KEY is not defined"
#endif #endif
#if defined(MBEDTLS_AES_128_BIT_MASKED) && ( !defined(MBEDTLS_AES_SCA_COUNTERMEASURES) || \
!defined(MBEDTLS_AES_ONLY_ENCRYPT) || \
!defined(MBEDTLS_AES_ONLY_128_BIT_KEY_LENGTH) )
#error "MBEDTLS_AES_128_BIT_MASKED defined, but not all prerequisites"
#endif
#if defined(MBEDTLS_DHM_C) && !defined(MBEDTLS_BIGNUM_C) #if defined(MBEDTLS_DHM_C) && !defined(MBEDTLS_BIGNUM_C)
#error "MBEDTLS_DHM_C defined, but not all prerequisites" #error "MBEDTLS_DHM_C defined, but not all prerequisites"
#endif #endif

View File

@ -658,18 +658,22 @@
* \def MBEDTLS_AES_128_BIT_MASKED * \def MBEDTLS_AES_128_BIT_MASKED
* *
* Requires MBEDTLS_AES_SCA_COUNTERMEASURES * Requires MBEDTLS_AES_SCA_COUNTERMEASURES
* MBEDTLS_AES_ONLY_128_BIT_KEY_LENGTH
* MBEDTLS_AES_ONLY_ENCRYPT
* *
* Add boolean masking against possible combined side-channel-attack fault injection attacks. * Add boolean masking against possible combined side-channel-attack
* fault injection attacks.
* *
* Uncommenting this macro adds data, key and Sbox masking additionally to dummy rounds * Uncommenting this macro adds data, key and Sbox masking additionally
* to dummy rounds.
* *
* Tradeoff: * Tradeoff:
* Uncommenting this macro does not increase codesize in MBEDTLS_AES_ROM_TABLES configuration. * Uncommenting this macro does not increase codesize in MBEDTLS_AES_ROM_TABLES
* Uncommenting this macro increases codesize in AES RAM tables configuration by ~600 byte. * configuration.
* Uncommenting this macro increases codesize in AES RAM tables configuration
* by ~600 bytes.
* The performance loss is ~50% with 128 bit AES encrypt. * The performance loss is ~50% with 128 bit AES encrypt.
* *
* This option is dependent of \c MBEDTLS_ENTROPY_HARDWARE_ALT.
*
*/ */
//#define MBEDTLS_AES_128_BIT_MASKED //#define MBEDTLS_AES_128_BIT_MASKED

View File

@ -721,8 +721,6 @@ static void mbedtls_generate_fake_key( unsigned int keybits, mbedtls_aes_context
/* /*
* AES key schedule (encryption) * AES key schedule (encryption)
*/ */
#if !defined(MBEDTLS_AES_SETKEY_ENC_ALT) #if !defined(MBEDTLS_AES_SETKEY_ENC_ALT)
int mbedtls_aes_setkey_enc( mbedtls_aes_context *ctx, const unsigned char *key, int mbedtls_aes_setkey_enc( mbedtls_aes_context *ctx, const unsigned char *key,
unsigned int keybits ) unsigned int keybits )
@ -1061,207 +1059,226 @@ int mbedtls_aes_xts_setkey_dec( mbedtls_aes_xts_context *ctx,
#if defined(MBEDTLS_AES_128_BIT_MASKED) #if defined(MBEDTLS_AES_128_BIT_MASKED)
static uint8_t xtime(uint8_t x) static uint8_t xtime( uint8_t x )
{ {
return ((x << 1) ^ (((x >> 7) & 1) * 0x1b)); return ( ( x << 1 ) ^ ( ( ( x >> 7 ) & 1 ) * 0x1b ) );
} }
static int sub_bytes_masked(uint32_t *data, uint8_t sbox_masked[256]) static int sub_bytes_masked( uint32_t *data, uint8_t sbox_masked[256] )
{ {
volatile unsigned int i; volatile unsigned int i;
for (i = 0; i < 4; i++) { for ( i = 0; i < 4; i++ )
data[i] = ( (uint32_t) sbox_masked[ ( data[i] ) & 0xFF ] ) ^ {
( (uint32_t) sbox_masked[ ( data[i] >> 8 ) & 0xFF ] << 8 ) ^ data[i] = ( (uint32_t)sbox_masked[(data[i]) & 0xFF] ) ^
( (uint32_t) sbox_masked[ ( data[i] >> 16 ) & 0xFF ] << 16 ) ^ ( (uint32_t)sbox_masked[(data[i] >> 8 ) & 0xFF] << 8 ) ^
( (uint32_t) sbox_masked[ ( data[i] >> 24 ) & 0xFF ] << 24 ); ( (uint32_t)sbox_masked[(data[i] >> 16 ) & 0xFF] << 16 ) ^
( (uint32_t)sbox_masked[(data[i] >> 24 ) & 0xFF] << 24 );
} }
if (i == 4){ if ( i == 4 )
{
return 0; return 0;
} }
return MBEDTLS_ERR_PLATFORM_FAULT_DETECTED; return ( MBEDTLS_ERR_PLATFORM_FAULT_DETECTED );
} }
static int mix_columns(uint8_t *s) static int mix_columns( uint8_t *s )
{ {
masked_state_t *state = (masked_state_t *) s; masked_state_t *state = (masked_state_t *)s;
volatile unsigned int i = 0; volatile unsigned int i = 0;
uint8_t Tmp, Tm, t; uint8_t Tmp, Tm, t;
for (i = 0; i < 4; ++i) for ( i = 0; i < 4; ++i )
{ {
t = (*state)[i][0]; t = (*state)[i][0];
Tmp = (*state)[i][0] ^ (*state)[i][1] ^ (*state)[i][2] ^ (*state)[i][3]; Tmp = (*state)[i][0] ^ (*state)[i][1] ^ (*state)[i][2] ^ (*state)[i][3];
Tm = (*state)[i][0] ^ (*state)[i][1]; Tm = (*state)[i][0] ^ (*state)[i][1];
Tm = xtime(Tm); Tm = xtime(Tm);
(*state)[i][0] ^= Tm ^ Tmp; (*state)[i][0] ^= Tm ^ Tmp;
Tm = (*state)[i][1] ^ (*state)[i][2]; Tm = (*state)[i][1] ^ (*state)[i][2];
Tm = xtime(Tm); Tm = xtime(Tm);
(*state)[i][1] ^= Tm ^ Tmp; (*state)[i][1] ^= Tm ^ Tmp;
Tm = (*state)[i][2] ^ (*state)[i][3]; Tm = (*state)[i][2] ^ (*state)[i][3];
Tm = xtime(Tm); Tm = xtime(Tm);
(*state)[i][2] ^= Tm ^ Tmp; (*state)[i][2] ^= Tm ^ Tmp;
Tm = (*state)[i][3] ^ t; Tm = (*state)[i][3] ^ t;
Tm = xtime(Tm); Tm = xtime(Tm);
(*state)[i][3] ^= Tm ^ Tmp; (*state)[i][3] ^= Tm ^ Tmp;
} }
if (i == 4) { if ( i == 4 )
{
return 0; return 0;
} }
return MBEDTLS_ERR_PLATFORM_FAULT_DETECTED; return ( MBEDTLS_ERR_PLATFORM_FAULT_DETECTED );
} }
static void shift_rows(uint8_t *s) static void shift_rows( uint8_t *s )
{ {
uint8_t temp; uint8_t temp;
masked_state_t *state = (masked_state_t *) s; masked_state_t *state = (masked_state_t *)s;
// Rotate first row 1 columns to left // Rotate first row 1 columns to left
temp = (*state)[0][1]; temp = (*state)[0][1];
(*state)[0][1] = (*state)[1][1]; (*state)[0][1] = (*state)[1][1];
(*state)[1][1] = (*state)[2][1]; (*state)[1][1] = (*state)[2][1];
(*state)[2][1] = (*state)[3][1]; (*state)[2][1] = (*state)[3][1];
(*state)[3][1] = temp; (*state)[3][1] = temp;
// Rotate second row 2 columns to left // Rotate second row 2 columns to left
temp = (*state)[0][2]; temp = (*state)[0][2];
(*state)[0][2] = (*state)[2][2]; (*state)[0][2] = (*state)[2][2];
(*state)[2][2] = temp; (*state)[2][2] = temp;
temp = (*state)[1][2]; temp = (*state)[1][2];
(*state)[1][2] = (*state)[3][2]; (*state)[1][2] = (*state)[3][2];
(*state)[3][2] = temp; (*state)[3][2] = temp;
// Rotate third row 3 columns to left
temp = (*state)[0][3];
(*state)[0][3] = (*state)[3][3];
(*state)[3][3] = (*state)[2][3];
(*state)[2][3] = (*state)[1][3];
(*state)[1][3] = temp;
// Rotate third row 3 columns to left
temp = (*state)[0][3];
(*state)[0][3] = (*state)[3][3];
(*state)[3][3] = (*state)[2][3];
(*state)[2][3] = (*state)[1][3];
(*state)[1][3] = temp;
} }
#define mul_02(num) ( (num << 1) ^ (0x11b & -(num >> 7)) ) #define mul_02( num ) ( ( num << 1 ) ^ ( 0x11b & - ( num >> 7 ) ) )
#define mul_03(num) ( mul_02(num) ^ num ) #define mul_03( num ) ( mul_02( num ) ^ num )
static void calcMixColmask(uint32_t mask[10]) static void calc_mix_colmn_mask( uint32_t mask[10] )
{ {
mask[6] = mul_02(mask[0]) ^ mul_03(mask[1]) ^ mask[2] ^ mask[3]; mask[6] = mul_02( mask[0] ) ^ mul_03( mask[1] ) ^ mask[2] ^ mask[3];
mask[7] = mask[0] ^ mul_02(mask[1]) ^ mul_03(mask[2]) ^ mask[3]; mask[7] = mask[0] ^ mul_02( mask[1] ) ^ mul_03( mask[2] ) ^ mask[3];
mask[8] = mask[0] ^ mask[1] ^ mul_02(mask[2]) ^ mul_03(mask[3]); mask[8] = mask[0] ^ mask[1] ^ mul_02( mask[2] ) ^ mul_03( mask[3] );
mask[9] = mul_03(mask[0]) ^ mask[1] ^ mask[2] ^ mul_02(mask[3]); mask[9] = mul_03( mask[0] ) ^ mask[1] ^ mask[2] ^ mul_02( mask[3] );
} }
//Calculate the the invSbox to change from Mask m to Mask m' //Calculate the the invSbox to change from Mask m to Mask m'
static int calcSboxMasked(uint32_t mask[10], uint8_t sbox_masked[256]) static int calc_sbox_masked( uint32_t mask[10], uint8_t sbox_masked[256] )
{ {
volatile unsigned int i = 0; volatile unsigned int i = 0;
for ( i = 0; i < 256; i++ ) for ( i = 0; i < 256; i++ )
{ {
sbox_masked[i ^ mask[4]] = FSb[i] ^ mask[5]; sbox_masked[i ^ mask[4]] = FSb[i] ^ mask[5];
} }
if (i == 256) { if ( i == 256 )
{
return 0; return 0;
} }
return MBEDTLS_ERR_PLATFORM_FAULT_DETECTED; return ( MBEDTLS_ERR_PLATFORM_FAULT_DETECTED );
} }
static int remask(uint32_t *data, uint32_t m1, uint32_t m2, uint32_t m3, uint32_t m4, uint32_t m5, uint32_t m6, uint32_t m7, uint32_t m8) static int remask( uint32_t *data, uint32_t m1, uint32_t m2,
uint32_t m3, uint32_t m4, uint32_t m5,
uint32_t m6, uint32_t m7, uint32_t m8 )
{ {
volatile unsigned int i = 0;
volatile unsigned int i = 0; for ( i = 0; i < 4; i++ )
{
data[i] = data[i] ^ ( ( m1 ^ m5 ) );
data[i] = data[i] ^ ( ( m2 ^ m6 ) << 8 );
data[i] = data[i] ^ ( ( m3 ^ m7 ) << 16 );
data[i] = data[i] ^ ( ( m4 ^ m8 ) << 24 );
}
for ( i = 0; i < 4; i++) if ( i == 4 )
{ {
data[i] = data[i] ^ ( (m1^m5) );
data[i] = data[i] ^ ( (m2^m6) << 8 );
data[i] = data[i] ^ ( (m3^m7) << 16 );
data[i] = data[i] ^ ( (m4^m8) << 24 );
}
if (i == 4) {
return 0; return 0;
} }
return MBEDTLS_ERR_PLATFORM_FAULT_DETECTED; return ( MBEDTLS_ERR_PLATFORM_FAULT_DETECTED );
} }
static int init_masking_encrypt(const uint8_t *rk, uint8_t *rk_masked, uint32_t mask[10], uint8_t sbox_masked[256] )
{
volatile int flow_control = 0;
unsigned int i = 0;
mbedtls_platform_memcpy(rk_masked, rk, MBEDTLS_AES_128_EXPANDED_KEY_SIZE_IN_WORDS*4);
//Randomly generate the masks: m1 m2 m3 m4 m m'
for (i = 0; i < 6; i++)
{
mask[i] = mbedtls_platform_random_in_range( 0xFF );
flow_control++;
}
//Calculate m1',m2',m3',m4'
calcMixColmask(mask);
flow_control++;
//Calculate the masked Sbox
if (calcSboxMasked(mask, sbox_masked) == 0){
flow_control++;
}
#define MASK_INIT_CONTROL 19 #define MASK_INIT_CONTROL 19
//Init masked key
if (remask( (uint32_t *)&rk_masked[(Nr * Nb * 4)], 0, 0, 0, 0, mask[5], mask[5], mask[5], mask[5]) == 0) {
flow_control++;
}
// Mask change from M1',M2',M3',M4' to M static int init_masking_encrypt( const uint8_t *rk, uint8_t *rk_masked,
for (i = 0; i < Nr; i++) uint32_t mask[10], uint8_t sbox_masked[256] )
{
if ( remask( (uint32_t *)&rk_masked[( i * Nb * 4 )], mask[6], mask[7], mask[8], mask[9], mask[4], mask[4], mask[4], mask[4]) == 0 )
flow_control++;
}
if( flow_control == MASK_INIT_CONTROL ) {
mbedtls_platform_random_delay();
if( flow_control == MASK_INIT_CONTROL ) {
return MASK_INIT_CONTROL;
}
}
return MBEDTLS_ERR_PLATFORM_FAULT_DETECTED;
}
static int add_rk_masked(uint32_t round, uint32_t *data, const uint32_t * rk_masked)
{ {
volatile unsigned int i; volatile int flow_control = 0;
unsigned int offset = round*4; unsigned int i = 0;
for( i = 0; i < 4; i++ )
{
data[i] ^= rk_masked[offset + i] ;
}
if (i == 4) { mbedtls_platform_memcpy( rk_masked, rk,
return 0; MBEDTLS_AES_128_EXPANDED_KEY_SIZE_IN_WORDS * 4 );
}
return MBEDTLS_ERR_PLATFORM_FAULT_DETECTED; //Randomly generate the masks: m1 m2 m3 m4 m m'
for ( i = 0; i < 6; i++ )
{
mask[i] = mbedtls_platform_random_in_range( 0xFF );
flow_control++;
}
//Calculate m1',m2',m3',m4'
calc_mix_colmn_mask( mask );
flow_control++;
//Calculate the masked Sbox
if ( calc_sbox_masked( mask, sbox_masked ) == 0 )
{
flow_control++;
}
//Init masked key
if ( remask( (uint32_t *)&rk_masked[(Nr * Nb * 4)], 0, 0, 0, 0,
mask[5], mask[5], mask[5], mask[5]) == 0 )
{
flow_control++;
}
// Mask change from M1',M2',M3',M4' to M
for ( i = 0; i < Nr; i++ )
{
if ( remask( (uint32_t *)&rk_masked[( i * Nb * 4 )], mask[6],
mask[7], mask[8], mask[9], mask[4], mask[4], mask[4], mask[4]) == 0 )
{
flow_control++;
}
}
if ( flow_control == MASK_INIT_CONTROL )
{
mbedtls_platform_random_delay();
if (flow_control == MASK_INIT_CONTROL)
{
return MASK_INIT_CONTROL;
}
}
return ( MBEDTLS_ERR_PLATFORM_FAULT_DETECTED );
} }
static int add_rk_masked( uint32_t round, uint32_t *data,
const uint32_t * rk_masked )
{
volatile unsigned int i;
unsigned int offset = round * 4;
static int aes_masked_round(uint32_t *data, uint32_t *key, uint32_t round, uint32_t mask[10], uint8_t sbox_masked[256]) for ( i = 0; i < 4; i++ )
{
data[i] ^= rk_masked[offset + i];
}
if ( i == 4 )
{
return 0;
}
return ( MBEDTLS_ERR_PLATFORM_FAULT_DETECTED );
}
static int aes_masked_round( uint32_t *data, uint32_t *key, uint32_t round,
uint32_t mask[10], uint8_t sbox_masked[256] )
{ {
volatile uint32_t flow_control = 0; volatile uint32_t flow_control = 0;
// Mask changes from M to M' // Mask changes from M to M'
if ( sub_bytes_masked(data, sbox_masked) == 0 ) if ( sub_bytes_masked( data, sbox_masked ) == 0 )
{
flow_control++; flow_control++;
}
//No impact on mask //No impact on mask
shift_rows((uint8_t *)data); shift_rows((uint8_t *)data);
@ -1271,44 +1288,62 @@ static int aes_masked_round(uint32_t *data, uint32_t *key, uint32_t round, uint3
// M2 for second row // M2 for second row
// M3 for third row // M3 for third row
// M4 for fourth row // M4 for fourth row
if ( remask(data, mask[0], mask[1], mask[2], mask[3], mask[5], mask[5], mask[5], mask[5]) == 0) if ( remask( data, mask[0], mask[1], mask[2], mask[3],
mask[5], mask[5], mask[5], mask[5]) == 0 )
{
flow_control++; flow_control++;
}
// Masks change from M1,M2,M3,M4 to M1',M2',M3',M4' // Masks change from M1,M2,M3,M4 to M1',M2',M3',M4'
if ( mix_columns((uint8_t *)data) == 0) if ( mix_columns( (uint8_t *)data ) == 0 )
{
flow_control++; flow_control++;
}
// Add the First round key to the state before starting the rounds. // Add the First round key to the state before starting the rounds.
// Masks change from M1',M2',M3',M4' to M // Masks change from M1',M2',M3',M4' to M
if ( add_rk_masked(round,data, key) == 0 ) if ( add_rk_masked( round, data, key ) == 0 )
{
flow_control++; flow_control++;
}
if ( flow_control == 4 ) if ( flow_control == 4 )
{
return 0; return 0;
}
return MBEDTLS_ERR_PLATFORM_FAULT_DETECTED; return ( MBEDTLS_ERR_PLATFORM_FAULT_DETECTED );
} }
static int aes_masked_round_final( uint32_t *data, uint32_t *key, uint8_t sbox_masked[256] ) static int aes_masked_round_final( uint32_t *data, uint32_t *key,
uint8_t sbox_masked[256] )
{ {
volatile uint32_t flow_control = 0; volatile uint32_t flow_control = 0;
if ( sub_bytes_masked(data, sbox_masked) == 0 ) if ( sub_bytes_masked(data, sbox_masked) == 0 )
{
flow_control++; flow_control++;
}
shift_rows((uint8_t *)data); shift_rows( (uint8_t *)data );
// Mask are removed by the last addroundkey // Mask are removed by the last addroundkey
// From M' to 0 // From M' to 0
if( add_rk_masked(Nr, data, key) == 0) if( add_rk_masked( Nr, data, key ) == 0 )
{
flow_control++; flow_control++;
}
if ( flow_control == 2 ) if ( flow_control == 2 )
{
return 0; return 0;
}
return MBEDTLS_ERR_PLATFORM_FAULT_DETECTED; return ( MBEDTLS_ERR_PLATFORM_FAULT_DETECTED );
} }
#define MASKING_FLOW_CONTORL (MASK_INIT_CONTROL + 2) //2 comes from initial data remask of real and fake data
//2 comes from initial data remask of real and fake data
#define MASKING_FLOW_CONTORL ( MASK_INIT_CONTROL + 2 )
#else // end of MBEDTLS_AES_128_BIT_MASKED #else // end of MBEDTLS_AES_128_BIT_MASKED
@ -1407,8 +1442,11 @@ int mbedtls_internal_aes_encrypt( mbedtls_aes_context *ctx,
#endif #endif
#if defined (MBEDTLS_AES_128_BIT_MASKED) #if defined (MBEDTLS_AES_128_BIT_MASKED)
//Flow control should be MASK_INIT_CONTROL and it will be checked as a part last flow control verification /* Flow control should be MASK_INIT_CONTROL and it will be checked as
flow_control = init_masking_encrypt((uint8_t*)ctx->rk, (uint8_t*)rk_masked, mask, sbox_masked); a part last flow control verification */
flow_control = init_masking_encrypt( (uint8_t *)ctx->rk,
(uint8_t *)rk_masked, mask, sbox_masked );
aes_data_real.rk_ptr = &rk_masked[0]; aes_data_real.rk_ptr = &rk_masked[0];
#else #else
aes_data_real.rk_ptr = ctx->rk; aes_data_real.rk_ptr = ctx->rk;
@ -1441,11 +1479,17 @@ int mbedtls_internal_aes_encrypt( mbedtls_aes_context *ctx,
#if defined (MBEDTLS_AES_128_BIT_MASKED) #if defined (MBEDTLS_AES_128_BIT_MASKED)
//Plain text masked with m1',m2',m3',m4' //Plain text masked with m1',m2',m3',m4'
if (remask( &aes_data_real.xy_values[0], mask[6], mask[7], mask[8], mask[9], 0, 0, 0, 0) == 0) if (remask( &aes_data_real.xy_values[0], mask[6],
mask[7], mask[8], mask[9], 0, 0, 0, 0) == 0 )
{
flow_control++; flow_control++;
}
if (remask( &aes_data_fake.xy_values[0], mask[6], mask[7], mask[8], mask[9], 0, 0, 0, 0) == 0) if (remask( &aes_data_fake.xy_values[0], mask[6],
mask[7], mask[8], mask[9], 0, 0, 0, 0) == 0 )
{
flow_control++; flow_control++;
}
#endif #endif
tindex = 0; tindex = 0;
@ -1457,8 +1501,11 @@ int mbedtls_internal_aes_encrypt( mbedtls_aes_context *ctx,
// initial round key addition // initial round key addition
#if defined (MBEDTLS_AES_128_BIT_MASKED) #if defined (MBEDTLS_AES_128_BIT_MASKED)
if ( add_rk_masked(0, &aes_data_ptr->xy_values[0], aes_data_ptr->rk_ptr) == 0) if ( add_rk_masked( 0, &aes_data_ptr->xy_values[0],
aes_data_ptr->rk_ptr ) == 0 )
{
flow_control++; flow_control++;
}
aes_data_ptr->round = 1; aes_data_ptr->round = 1;
#else #else
for( i = 0; i < 4; i++ ) for( i = 0; i < 4; i++ )
@ -1471,7 +1518,6 @@ int mbedtls_internal_aes_encrypt( mbedtls_aes_context *ctx,
tindex++; tindex++;
} while( stop_mark == 0 ); } while( stop_mark == 0 );
// Calculate AES rounds (9, 11 or 13 rounds) + dummy rounds // Calculate AES rounds (9, 11 or 13 rounds) + dummy rounds
do do
{ {
@ -1480,9 +1526,12 @@ int mbedtls_internal_aes_encrypt( mbedtls_aes_context *ctx,
offset = round_ctrl_table[tindex] & 0x04; offset = round_ctrl_table[tindex] & 0x04;
stop_mark = round_ctrl_table[tindex] & 0x03; stop_mark = round_ctrl_table[tindex] & 0x03;
#if defined (MBEDTLS_AES_128_BIT_MASKED) #if defined (MBEDTLS_AES_128_BIT_MASKED)
if (aes_masked_round( &aes_data_ptr->xy_values[0], aes_data_ptr->rk_ptr, if ( aes_masked_round( &aes_data_ptr->xy_values[0],
aes_data_ptr->round, mask, sbox_masked) == 0) aes_data_ptr->rk_ptr,
aes_data_ptr->round, mask, sbox_masked ) == 0 )
{
flow_control++; flow_control++;
}
aes_data_ptr->round ++; aes_data_ptr->round ++;
#else #else
aes_data_ptr->rk_ptr = aes_fround( aes_data_ptr->rk_ptr, aes_data_ptr->rk_ptr = aes_fround( aes_data_ptr->rk_ptr,
@ -1507,10 +1556,12 @@ int mbedtls_internal_aes_encrypt( mbedtls_aes_context *ctx,
stop_mark = round_ctrl_table[tindex] & 0x03; stop_mark = round_ctrl_table[tindex] & 0x03;
#if defined (MBEDTLS_AES_128_BIT_MASKED) #if defined (MBEDTLS_AES_128_BIT_MASKED)
if ( aes_masked_round_final( &aes_data_ptr->xy_values[0], if ( aes_masked_round_final( &aes_data_ptr->xy_values[0],
aes_data_ptr->rk_ptr, sbox_masked ) == 0) aes_data_ptr->rk_ptr, sbox_masked ) == 0 )
{
flow_control++; flow_control++;
}
//Cleanup the masked key //Cleanup the masked key
mbedtls_platform_memset(rk_masked, 0, sizeof(rk_masked)); mbedtls_platform_memset( rk_masked, 0, sizeof(rk_masked) );
#else #else
aes_fround_final( aes_data_ptr->rk_ptr, aes_fround_final( aes_data_ptr->rk_ptr,
&aes_data_ptr->xy_values[0], &aes_data_ptr->xy_values[0],
@ -1526,7 +1577,6 @@ int mbedtls_internal_aes_encrypt( mbedtls_aes_context *ctx,
tindex++; tindex++;
} while( stop_mark == 0 ); } while( stop_mark == 0 );
// SCA countermeasure, safely clear the output // SCA countermeasure, safely clear the output
mbedtls_platform_memset( output, 0, 16 ); mbedtls_platform_memset( output, 0, 16 );
@ -1541,7 +1591,7 @@ int mbedtls_internal_aes_encrypt( mbedtls_aes_context *ctx,
} while( ( i = ( i + 1 ) % 4 ) != offset ); } while( ( i = ( i + 1 ) % 4 ) != offset );
#if defined (MBEDTLS_AES_128_BIT_MASKED) #if defined (MBEDTLS_AES_128_BIT_MASKED)
mbedtls_platform_memset(rk_masked, 0, sizeof(rk_masked)); mbedtls_platform_memset( rk_masked, 0, sizeof(rk_masked) );
#endif #endif
/* Double negation is used to silence an "extraneous parentheses" warning */ /* Double negation is used to silence an "extraneous parentheses" warning */
if( ! ( flow_control != tindex + dummy_rounds + MASKING_FLOW_CONTORL + 8 ) if( ! ( flow_control != tindex + dummy_rounds + MASKING_FLOW_CONTORL + 8 )
@ -1561,13 +1611,14 @@ int mbedtls_internal_aes_encrypt( mbedtls_aes_context *ctx,
// Clear the output in case of a FI // Clear the output in case of a FI
mbedtls_platform_memset( output, 0, 16 ); mbedtls_platform_memset( output, 0, 16 );
mbedtls_platform_memset( (uint8_t*)&aes_data_real, 0, sizeof(aes_data_real) ); mbedtls_platform_memset( (uint8_t*)&aes_data_real, 0,
mbedtls_platform_memset (aes_data_table, 0, sizeof(aes_data_table)); sizeof(aes_data_real) );
mbedtls_platform_memset ( aes_data_table, 0, sizeof(aes_data_table) );
#if defined (MBEDTLS_AES_128_BIT_MASKED) #if defined (MBEDTLS_AES_128_BIT_MASKED)
//Clear masked key, masked sbox and mask in case of a FI //Clear masked key, masked sbox and mask in case of a FI
mbedtls_platform_memset(rk_masked, 0, sizeof(rk_masked)); mbedtls_platform_memset( rk_masked, 0, sizeof(rk_masked) );
mbedtls_platform_memset(mask, 0, sizeof(mask)); mbedtls_platform_memset( mask, 0, sizeof(mask) );
mbedtls_platform_memset(sbox_masked, 0, sizeof(sbox_masked)); mbedtls_platform_memset( sbox_masked, 0, sizeof(sbox_masked) );
#endif #endif
return( MBEDTLS_ERR_PLATFORM_FAULT_DETECTED ); return( MBEDTLS_ERR_PLATFORM_FAULT_DETECTED );
} }