mirrors/gstreamer
mirrors/gstreamer
1
1
Fork 0
mirror of https://gitlab.freedesktop.org/gstreamer/gstreamer.git synced 2024-11-18 07:47:17 +00:00
Code Issues 1 Projects Releases Wiki Activity

audioresample: Fix out of bounds memory accesses

Browse source
This commit is contained in:
Sebastian Dröge 2014-04-15 19:16:18 +02:00
parent 3e2a693ca9
commit 5826f79980
1 changed files with 115 additions and 66 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

85
gst/audioresample/resample_sse.h
View file

@ -41,10 +41,13 @@
#define OVERRIDE_INNER_PRODUCT_SINGLE
static inline float inner_product_single(const float *a, const float *b, unsigned int len)
{
int i;
float ret;
int i = 0;
float ret = 0;
__m128 sum = _mm_setzero_ps();
for (i=0;i<len;i+=8)
if (len > 7)
{
for (;i<len-7;i+=8)
{
sum = _mm_add_ps(sum, _mm_mul_ps(_mm_loadu_ps(a+i), _mm_loadu_ps(b+i)));
sum = _mm_add_ps(sum, _mm_mul_ps(_mm_loadu_ps(a+i+4), _mm_loadu_ps(b+i+4)));
@ -52,24 +55,38 @@ static inline float inner_product_single(const float *a, const float *b, unsigne
sum = _mm_add_ps(sum, _mm_movehl_ps(sum, sum));
sum = _mm_add_ss(sum, _mm_shuffle_ps(sum, sum, 0x55));
_mm_store_ss(&ret, sum);
}
for (; i < len; i++)
ret += a[i] * b[i];
return ret;
}
#define OVERRIDE_INTERPOLATE_PRODUCT_SINGLE
static inline float interpolate_product_single(const float *a, const float *b, unsigned int len, const spx_uint32_t oversample, float *frac) {
int i;
float ret;
int i = 0;
float ret = 0;
__m128 sum = _mm_setzero_ps();
__m128 f = _mm_loadu_ps(frac);
for(i=0;i<len;i+=2)
if (len > 1)
{
for(;i<len-1;i+=2)
{
sum = _mm_add_ps(sum, _mm_mul_ps(_mm_load1_ps(a+i), _mm_loadu_ps(b+i*oversample)));
sum = _mm_add_ps(sum, _mm_mul_ps(_mm_load1_ps(a+i+1), _mm_loadu_ps(b+(i+1)*oversample)));
}
sum = _mm_mul_ps(f, sum);
sum = _mm_add_ps(sum, _mm_movehl_ps(sum, sum));
sum = _mm_add_ss(sum, _mm_shuffle_ps(sum, sum, 0x55));
_mm_store_ss(&ret, sum);
}
if (i == len-1)
ret += a[i] * (frac[0]*b[i*oversample] + frac[1]*b[i*oversample + 1] + frac[2]*b[i*oversample + 2] + frac[3]*b[i*oversample + 3]);
return ret;
}
@ -82,26 +99,37 @@ static inline float interpolate_product_single(const float *a, const float *b, u
#ifdef DOUBLE_PRECISION
static inline double inner_product_double(const double *a, const double *b, unsigned int len)
{
int i;
double ret;
int i = 0;
double ret = 0;
__m128d sum = _mm_setzero_pd();
for (i=0;i<len;i+=4)
if (len > 3)
{
for (;i<len-3;i+=4)
{
sum = _mm_add_pd(sum, _mm_mul_pd(_mm_loadu_pd(a+i), _mm_loadu_pd(b+i)));
sum = _mm_add_pd(sum, _mm_mul_pd(_mm_loadu_pd(a+i+2), _mm_loadu_pd(b+i+2)));
}
sum = _mm_add_sd(sum, _mm_unpackhi_pd(sum, sum));
_mm_store_sd(&ret, sum);
}
for (; i < len; i++)
ret += a[i] * b[i];
return ret;
}
#else
static inline double inner_product_double(const float *a, const float *b, unsigned int len)
{
int i;
double ret;
int i = 0;
double ret = 0;
__m128d sum = _mm_setzero_pd();
__m128 t;
for (i=0;i<len;i+=8)
if (len > 7)
{
for (;i<len-7;i+=8)
{
t = _mm_mul_ps(_mm_loadu_ps(a+i), _mm_loadu_ps(b+i));
sum = _mm_add_pd(sum, _mm_cvtps_pd(t));
@ -113,6 +141,11 @@ static inline double inner_product_double(const float *a, const float *b, unsign
}
sum = _mm_add_sd(sum, _mm_unpackhi_pd(sum, sum));
_mm_store_sd(&ret, sum);
}
for (; i < len; i++)
ret += a[i] * b[i];
return ret;
}
#endif
@ -122,15 +155,18 @@ static inline double inner_product_double(const float *a, const float *b, unsign
#ifdef DOUBLE_PRECISION
static inline double interpolate_product_double(const double *a, const double *b, unsigned int len, const spx_uint32_t oversample, double *frac) {
int i;
double ret;
int i = 0;
double ret = 0;
__m128d sum;
__m128d sum1 = _mm_setzero_pd();
__m128d sum2 = _mm_setzero_pd();
__m128d f1 = _mm_loadu_pd(frac);
__m128d f2 = _mm_loadu_pd(frac+2);
__m128d t;
for(i=0;i<len;i+=2)
if (len > 1)
{
for(;i<len-1;i+=2)
{
t = _mm_mul_pd(_mm_load1_pd(a+i), _mm_loadu_pd(b+i*oversample));
sum1 = _mm_add_pd(sum1, t);
@ -145,12 +181,17 @@ static inline double interpolate_product_double(const double *a, const double *b
sum = _mm_add_pd(sum1, sum2);
sum = _mm_add_sd(sum, _mm_unpackhi_pd(sum, sum));
_mm_store_sd(&ret, sum);
}
if (i == len-1)
ret += a[i] * (frac[0]*b[i*oversample] + frac[1]*b[i*oversample + 1] + frac[2]*b[i*oversample + 2] + frac[3]*b[i*oversample + 3]);
return ret;
}
#else
static inline double interpolate_product_double(const float *a, const float *b, unsigned int len, const spx_uint32_t oversample, float *frac) {
int i;
double ret;
int i = 0;
double ret = 0;
__m128d sum;
__m128d sum1 = _mm_setzero_pd();
__m128d sum2 = _mm_setzero_pd();
@ -158,7 +199,10 @@ static inline double interpolate_product_double(const float *a, const float *b,
__m128d f1 = _mm_cvtps_pd(f);
__m128d f2 = _mm_cvtps_pd(_mm_movehl_ps(f,f));
__m128 t;
for(i=0;i<len;i+=2)
if (len > 1)
{
for(;i<len-1;i+=2)
{
t = _mm_mul_ps(_mm_load1_ps(a+i), _mm_loadu_ps(b+i*oversample));
sum1 = _mm_add_pd(sum1, _mm_cvtps_pd(t));
@ -173,6 +217,11 @@ static inline double interpolate_product_double(const float *a, const float *b,
sum = _mm_add_pd(sum1, sum2);
sum = _mm_add_sd(sum, _mm_unpackhi_pd(sum, sum));
_mm_store_sd(&ret, sum);
}
if (i == len-1)
ret += a[i] * (frac[0]*b[i*oversample] + frac[1]*b[i*oversample + 1] + frac[2]*b[i*oversample + 2] + frac[3]*b[i*oversample + 3]);
return ret;
}
#endif