yuzu-emu/mbedtls
1
0
Fork 0
mirror of https://github.com/yuzu-emu/mbedtls.git synced 2024-11-30 01:14:23 +01:00
Code Issues Packages Projects Releases Wiki Activity

Put local variables in a struct

Browse Source 
This way we can have a single call to mbedtls_zeroize, which
saves a few bytes of code size.

Additionally, on my PC, I notice a significant speed improvement
(x86_64 build with MBEDTLS_AESNI_C disabled, gcc 5.4.0 -O3). I don't
have an explanation for that (I expected no measurable difference).

Signed-off-by: Gilles Peskine <Gilles.Peskine@arm.com>
Signed-off-by: Ronald Cron <ronald.cron@arm.com>
This commit is contained in:
Gilles Peskine 2020-08-26 17:03:24 +02:00 committed by Ronald Cron
parent 5706e920a4
commit acbf9eccb5
1 changed files with 76 additions and 86 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

162
library/aes.c
View File

@ -740,61 +740,56 @@ int mbedtls_internal_aes_encrypt( mbedtls_aes_context *ctx,
unsigned char output[16] ) unsigned char output[16] )
{ {
int i; int i;
uint32_t *RK, X0, X1, X2, X3, Y0, Y1, Y2, Y3; uint32_t *RK = ctx->rk;
struct
{
uint32_t X[4];
uint32_t Y[4];
} t;
RK = ctx->rk; GET_UINT32_LE( t.X[0], input, 0 ); t.X[0] ^= *RK++;
GET_UINT32_LE( t.X[1], input, 4 ); t.X[1] ^= *RK++;
GET_UINT32_LE( X0, input, 0 ); X0 ^= *RK++; GET_UINT32_LE( t.X[2], input, 8 ); t.X[2] ^= *RK++;
GET_UINT32_LE( X1, input, 4 ); X1 ^= *RK++; GET_UINT32_LE( t.X[3], input, 12 ); t.X[3] ^= *RK++;
GET_UINT32_LE( X2, input, 8 ); X2 ^= *RK++;
GET_UINT32_LE( X3, input, 12 ); X3 ^= *RK++;
for( i = ( ctx->nr >> 1 ) - 1; i > 0; i-- ) for( i = ( ctx->nr >> 1 ) - 1; i > 0; i-- )
{ {
AES_FROUND( Y0, Y1, Y2, Y3, X0, X1, X2, X3 ); AES_FROUND( t.Y[0], t.Y[1], t.Y[2], t.Y[3], t.X[0], t.X[1], t.X[2], t.X[3] );
AES_FROUND( X0, X1, X2, X3, Y0, Y1, Y2, Y3 ); AES_FROUND( t.X[0], t.X[1], t.X[2], t.X[3], t.Y[0], t.Y[1], t.Y[2], t.Y[3] );
} }
AES_FROUND( Y0, Y1, Y2, Y3, X0, X1, X2, X3 ); AES_FROUND( t.Y[0], t.Y[1], t.Y[2], t.Y[3], t.X[0], t.X[1], t.X[2], t.X[3] );
X0 = *RK++ ^ \ t.X[0] = *RK++ ^ \
( (uint32_t) FSb[ ( Y0 ) & 0xFF ] ) ^ ( (uint32_t) FSb[ ( t.Y[0] ) & 0xFF ] ) ^
( (uint32_t) FSb[ ( Y1 >> 8 ) & 0xFF ] << 8 ) ^ ( (uint32_t) FSb[ ( t.Y[1] >> 8 ) & 0xFF ] << 8 ) ^
( (uint32_t) FSb[ ( Y2 >> 16 ) & 0xFF ] << 16 ) ^ ( (uint32_t) FSb[ ( t.Y[2] >> 16 ) & 0xFF ] << 16 ) ^
( (uint32_t) FSb[ ( Y3 >> 24 ) & 0xFF ] << 24 ); ( (uint32_t) FSb[ ( t.Y[3] >> 24 ) & 0xFF ] << 24 );
X1 = *RK++ ^ \ t.X[1] = *RK++ ^ \
( (uint32_t) FSb[ ( Y1 ) & 0xFF ] ) ^ ( (uint32_t) FSb[ ( t.Y[1] ) & 0xFF ] ) ^
( (uint32_t) FSb[ ( Y2 >> 8 ) & 0xFF ] << 8 ) ^ ( (uint32_t) FSb[ ( t.Y[2] >> 8 ) & 0xFF ] << 8 ) ^
( (uint32_t) FSb[ ( Y3 >> 16 ) & 0xFF ] << 16 ) ^ ( (uint32_t) FSb[ ( t.Y[3] >> 16 ) & 0xFF ] << 16 ) ^
( (uint32_t) FSb[ ( Y0 >> 24 ) & 0xFF ] << 24 ); ( (uint32_t) FSb[ ( t.Y[0] >> 24 ) & 0xFF ] << 24 );
X2 = *RK++ ^ \ t.X[2] = *RK++ ^ \
( (uint32_t) FSb[ ( Y2 ) & 0xFF ] ) ^ ( (uint32_t) FSb[ ( t.Y[2] ) & 0xFF ] ) ^
( (uint32_t) FSb[ ( Y3 >> 8 ) & 0xFF ] << 8 ) ^ ( (uint32_t) FSb[ ( t.Y[3] >> 8 ) & 0xFF ] << 8 ) ^
( (uint32_t) FSb[ ( Y0 >> 16 ) & 0xFF ] << 16 ) ^ ( (uint32_t) FSb[ ( t.Y[0] >> 16 ) & 0xFF ] << 16 ) ^
( (uint32_t) FSb[ ( Y1 >> 24 ) & 0xFF ] << 24 ); ( (uint32_t) FSb[ ( t.Y[1] >> 24 ) & 0xFF ] << 24 );
X3 = *RK++ ^ \ t.X[3] = *RK++ ^ \
( (uint32_t) FSb[ ( Y3 ) & 0xFF ] ) ^ ( (uint32_t) FSb[ ( t.Y[3] ) & 0xFF ] ) ^
( (uint32_t) FSb[ ( Y0 >> 8 ) & 0xFF ] << 8 ) ^ ( (uint32_t) FSb[ ( t.Y[0] >> 8 ) & 0xFF ] << 8 ) ^
( (uint32_t) FSb[ ( Y1 >> 16 ) & 0xFF ] << 16 ) ^ ( (uint32_t) FSb[ ( t.Y[1] >> 16 ) & 0xFF ] << 16 ) ^
( (uint32_t) FSb[ ( Y2 >> 24 ) & 0xFF ] << 24 ); ( (uint32_t) FSb[ ( t.Y[2] >> 24 ) & 0xFF ] << 24 );
PUT_UINT32_LE( X0, output, 0 ); PUT_UINT32_LE( t.X[0], output, 0 );
PUT_UINT32_LE( X1, output, 4 ); PUT_UINT32_LE( t.X[1], output, 4 );
PUT_UINT32_LE( X2, output, 8 ); PUT_UINT32_LE( t.X[2], output, 8 );
PUT_UINT32_LE( X3, output, 12 ); PUT_UINT32_LE( t.X[3], output, 12 );
mbedtls_zeroize( &X0, sizeof( X0 ) ); mbedtls_zeroize( &t, sizeof( t ) );
mbedtls_zeroize( &X1, sizeof( X1 ) );
mbedtls_zeroize( &X2, sizeof( X2 ) );
mbedtls_zeroize( &X3, sizeof( X3 ) );
mbedtls_zeroize( &Y0, sizeof( Y0 ) );
mbedtls_zeroize( &Y1, sizeof( Y1 ) );
mbedtls_zeroize( &Y2, sizeof( Y2 ) );
mbedtls_zeroize( &Y3, sizeof( Y3 ) );
return( 0 ); return( 0 );
} }
@ -818,61 +813,56 @@ int mbedtls_internal_aes_decrypt( mbedtls_aes_context *ctx,
unsigned char output[16] ) unsigned char output[16] )
{ {
int i; int i;
uint32_t *RK, X0, X1, X2, X3, Y0, Y1, Y2, Y3; uint32_t *RK = ctx->rk;
struct
{
uint32_t X[4];
uint32_t Y[4];
} t;
RK = ctx->rk; GET_UINT32_LE( t.X[0], input, 0 ); t.X[0] ^= *RK++;
GET_UINT32_LE( t.X[1], input, 4 ); t.X[1] ^= *RK++;
GET_UINT32_LE( X0, input, 0 ); X0 ^= *RK++; GET_UINT32_LE( t.X[2], input, 8 ); t.X[2] ^= *RK++;
GET_UINT32_LE( X1, input, 4 ); X1 ^= *RK++; GET_UINT32_LE( t.X[3], input, 12 ); t.X[3] ^= *RK++;
GET_UINT32_LE( X2, input, 8 ); X2 ^= *RK++;
GET_UINT32_LE( X3, input, 12 ); X3 ^= *RK++;
for( i = ( ctx->nr >> 1 ) - 1; i > 0; i-- ) for( i = ( ctx->nr >> 1 ) - 1; i > 0; i-- )
{ {
AES_RROUND( Y0, Y1, Y2, Y3, X0, X1, X2, X3 ); AES_RROUND( t.Y[0], t.Y[1], t.Y[2], t.Y[3], t.X[0], t.X[1], t.X[2], t.X[3] );
AES_RROUND( X0, X1, X2, X3, Y0, Y1, Y2, Y3 ); AES_RROUND( t.X[0], t.X[1], t.X[2], t.X[3], t.Y[0], t.Y[1], t.Y[2], t.Y[3] );
} }
AES_RROUND( Y0, Y1, Y2, Y3, X0, X1, X2, X3 ); AES_RROUND( t.Y[0], t.Y[1], t.Y[2], t.Y[3], t.X[0], t.X[1], t.X[2], t.X[3] );
X0 = *RK++ ^ \ t.X[0] = *RK++ ^ \
( (uint32_t) RSb[ ( Y0 ) & 0xFF ] ) ^ ( (uint32_t) RSb[ ( t.Y[0] ) & 0xFF ] ) ^
( (uint32_t) RSb[ ( Y3 >> 8 ) & 0xFF ] << 8 ) ^ ( (uint32_t) RSb[ ( t.Y[3] >> 8 ) & 0xFF ] << 8 ) ^
( (uint32_t) RSb[ ( Y2 >> 16 ) & 0xFF ] << 16 ) ^ ( (uint32_t) RSb[ ( t.Y[2] >> 16 ) & 0xFF ] << 16 ) ^
( (uint32_t) RSb[ ( Y1 >> 24 ) & 0xFF ] << 24 ); ( (uint32_t) RSb[ ( t.Y[1] >> 24 ) & 0xFF ] << 24 );
X1 = *RK++ ^ \ t.X[1] = *RK++ ^ \
( (uint32_t) RSb[ ( Y1 ) & 0xFF ] ) ^ ( (uint32_t) RSb[ ( t.Y[1] ) & 0xFF ] ) ^
( (uint32_t) RSb[ ( Y0 >> 8 ) & 0xFF ] << 8 ) ^ ( (uint32_t) RSb[ ( t.Y[0] >> 8 ) & 0xFF ] << 8 ) ^
( (uint32_t) RSb[ ( Y3 >> 16 ) & 0xFF ] << 16 ) ^ ( (uint32_t) RSb[ ( t.Y[3] >> 16 ) & 0xFF ] << 16 ) ^
( (uint32_t) RSb[ ( Y2 >> 24 ) & 0xFF ] << 24 ); ( (uint32_t) RSb[ ( t.Y[2] >> 24 ) & 0xFF ] << 24 );
X2 = *RK++ ^ \ t.X[2] = *RK++ ^ \
( (uint32_t) RSb[ ( Y2 ) & 0xFF ] ) ^ ( (uint32_t) RSb[ ( t.Y[2] ) & 0xFF ] ) ^
( (uint32_t) RSb[ ( Y1 >> 8 ) & 0xFF ] << 8 ) ^ ( (uint32_t) RSb[ ( t.Y[1] >> 8 ) & 0xFF ] << 8 ) ^
( (uint32_t) RSb[ ( Y0 >> 16 ) & 0xFF ] << 16 ) ^ ( (uint32_t) RSb[ ( t.Y[0] >> 16 ) & 0xFF ] << 16 ) ^
( (uint32_t) RSb[ ( Y3 >> 24 ) & 0xFF ] << 24 ); ( (uint32_t) RSb[ ( t.Y[3] >> 24 ) & 0xFF ] << 24 );
X3 = *RK++ ^ \ t.X[3] = *RK++ ^ \
( (uint32_t) RSb[ ( Y3 ) & 0xFF ] ) ^ ( (uint32_t) RSb[ ( t.Y[3] ) & 0xFF ] ) ^
( (uint32_t) RSb[ ( Y2 >> 8 ) & 0xFF ] << 8 ) ^ ( (uint32_t) RSb[ ( t.Y[2] >> 8 ) & 0xFF ] << 8 ) ^
( (uint32_t) RSb[ ( Y1 >> 16 ) & 0xFF ] << 16 ) ^ ( (uint32_t) RSb[ ( t.Y[1] >> 16 ) & 0xFF ] << 16 ) ^
( (uint32_t) RSb[ ( Y0 >> 24 ) & 0xFF ] << 24 ); ( (uint32_t) RSb[ ( t.Y[0] >> 24 ) & 0xFF ] << 24 );
PUT_UINT32_LE( X0, output, 0 ); PUT_UINT32_LE( t.X[0], output, 0 );
PUT_UINT32_LE( X1, output, 4 ); PUT_UINT32_LE( t.X[1], output, 4 );
PUT_UINT32_LE( X2, output, 8 ); PUT_UINT32_LE( t.X[2], output, 8 );
PUT_UINT32_LE( X3, output, 12 ); PUT_UINT32_LE( t.X[3], output, 12 );
mbedtls_zeroize( &X0, sizeof( X0 ) ); mbedtls_zeroize( &t, sizeof( t ) );
mbedtls_zeroize( &X1, sizeof( X1 ) );
mbedtls_zeroize( &X2, sizeof( X2 ) );
mbedtls_zeroize( &X3, sizeof( X3 ) );
mbedtls_zeroize( &Y0, sizeof( Y0 ) );
mbedtls_zeroize( &Y1, sizeof( Y1 ) );
mbedtls_zeroize( &Y2, sizeof( Y2 ) );
mbedtls_zeroize( &Y3, sizeof( Y3 ) );
return( 0 ); return( 0 );
} }