Singularity-Viewer/SingularityViewer
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Update openjpeg with alchemy changes

Browse Source
This commit is contained in:
Rye Mutt
2020-09-29 14:40:25 -04:00
parent f941bc4334
commit f0dfb0b080
12 changed files with 753 additions and 220 deletions
Download Patch File Download Diff File Expand all files Collapse all files

1
indra/libopenjpeg/CMakeLists.txt
View File

@@ -26,6 +26,7 @@ set(openjpeg_SOURCE_FILES
mct.c
mqc.c
openjpeg.c
opj_malloc.c
phix_manager.c
pi.c
ppix_manager.c

133
indra/libopenjpeg/dwt.c
View File

@@ -31,11 +31,16 @@
* POSSIBILITY OF SUCH DAMAGE.
*/
#define OPJ_SKIP_POISON
#include "opj_includes.h"
#ifdef __SSE__
#include <xmmintrin.h>
#endif
#include "opj_includes.h"
#if defined(__GNUC__)
#pragma GCC poison malloc calloc realloc free
#endif
/** @defgroup DWT DWT - Implementation of a discrete wavelet transform */
/*@{*/
@@ -499,7 +504,7 @@ void dwt_calc_explicit_stepsizes(opj_tccp_t * tccp, int prec) {
/* <summary> */
/* Determine maximum computed resolution level for inverse wavelet transform */
/* </summary> */
static int dwt_decode_max_resolution(opj_tcd_resolution_t* restrict r, int i) {
static int dwt_decode_max_resolution(opj_tcd_resolution_t* OPJ_RESTRICT r, int i) {
int mr = 1;
int w;
while( --i ) {
@@ -531,7 +536,7 @@ static void dwt_decode_tile(opj_tcd_tilecomp_t* tilec, int numres, DWT1DFN dwt_1
v.mem = h.mem;
while( --numres) {
int * restrict tiledp = tilec->data;
int * OPJ_RESTRICT tiledp = tilec->data;
int j;
++tr;
@@ -565,48 +570,49 @@ static void dwt_decode_tile(opj_tcd_tilecomp_t* tilec, int numres, DWT1DFN dwt_1
opj_aligned_free(h.mem);
}
static void v4dwt_interleave_h(v4dwt_t* restrict w, float* restrict a, int x, int size){
float* restrict bi = (float*) (w->wavelet + w->cas);
static void v4dwt_interleave_h(v4dwt_t* OPJ_RESTRICT w, float* OPJ_RESTRICT a, int x, int size) {
float* OPJ_RESTRICT bi = (float*)(w->wavelet + w->cas);
int count = w->sn;
int i, k;
for(k = 0; k < 2; ++k){
if (count + 3 * x < size && ((size_t) a & 0x0f) == 0 && ((size_t) bi & 0x0f) == 0 && (x & 0x0f) == 0) {
for (k = 0; k < 2; ++k) {
if (count + 3 * x < size && ((size_t)a & 0x0f) == 0 && ((size_t)bi & 0x0f) == 0 && (x & 0x0f) == 0) {
/* Fast code path */
for(i = 0; i < count; ++i){
for (i = 0; i < count; ++i) {
int j = i;
bi[i*8 ] = a[j];
bi[i * 8] = a[j];
j += x;
bi[i*8 + 1] = a[j];
bi[i * 8 + 1] = a[j];
j += x;
bi[i*8 + 2] = a[j];
bi[i * 8 + 2] = a[j];
j += x;
bi[i*8 + 3] = a[j];
bi[i * 8 + 3] = a[j];
}
} else {
}
else {
/* Slow code path */
for(i = 0; i < count; ++i){
int j = i;
bi[i*8 ] = a[j];
j += x;
if(j > size) continue;
bi[i*8 + 1] = a[j];
j += x;
if(j > size) continue;
bi[i*8 + 2] = a[j];
j += x;
if(j > size) continue;
bi[i*8 + 3] = a[j];
for (i = 0; i < count; ++i) {
int j = i;
bi[i * 8] = a[j];
j += x;
if (j > size) continue;
bi[i * 8 + 1] = a[j];
j += x;
if (j > size) continue;
bi[i * 8 + 2] = a[j];
j += x;
if (j > size) continue;
bi[i * 8 + 3] = a[j];
}
}
}
bi = (float*) (w->wavelet + 1 - w->cas);
bi = (float*)(w->wavelet + 1 - w->cas);
a += w->sn;
size -= w->sn;
count = w->dn;
}
}
static void v4dwt_interleave_v(v4dwt_t* restrict v , float* restrict a , int x){
v4* restrict bi = v->wavelet + v->cas;
static void v4dwt_interleave_v(v4dwt_t* OPJ_RESTRICT v , float* OPJ_RESTRICT a , int x){
v4* OPJ_RESTRICT bi = v->wavelet + v->cas;
int i;
for(i = 0; i < v->sn; ++i){
memcpy(&bi[i*2], &a[i*x], 4 * sizeof(float));
@@ -621,7 +627,7 @@ static void v4dwt_interleave_v(v4dwt_t* restrict v , float* restrict a , int x){
#ifdef __SSE__
static void v4dwt_decode_step1_sse(v4* w, int count, const __m128 c){
__m128* restrict vw = (__m128*) w;
__m128* OPJ_RESTRICT vw = (__m128*) w;
int i;
/* 4x unrolled loop */
for(i = 0; i < count >> 2; ++i){
@@ -642,22 +648,39 @@ static void v4dwt_decode_step1_sse(v4* w, int count, const __m128 c){
}
static void v4dwt_decode_step2_sse(v4* l, v4* w, int k, int m, __m128 c){
__m128* restrict vl = (__m128*) l;
__m128* restrict vw = (__m128*) w;
__m128* OPJ_RESTRICT vl = (__m128*) l;
__m128* OPJ_RESTRICT vw = (__m128*) w;
int i;
__m128 tmp1, tmp2, tmp3;
tmp1 = vl[0];
for(i = 0; i < m; ++i){
for (i = 0; i < m - 3; i += 4) {
__m128 tmp4, tmp5, tmp6, tmp7, tmp8, tmp9;
tmp2 = vw[-1];
tmp3 = vw[0];
tmp4 = vw[1];
tmp5 = vw[2];
tmp6 = vw[3];
tmp7 = vw[4];
tmp8 = vw[5];
tmp9 = vw[6];
vw[-1] = _mm_add_ps(tmp2, _mm_mul_ps(_mm_add_ps(tmp1, tmp3), c));
vw[1] = _mm_add_ps(tmp4, _mm_mul_ps(_mm_add_ps(tmp3, tmp5), c));
vw[3] = _mm_add_ps(tmp6, _mm_mul_ps(_mm_add_ps(tmp5, tmp7), c));
vw[5] = _mm_add_ps(tmp8, _mm_mul_ps(_mm_add_ps(tmp7, tmp9), c));
tmp1 = tmp9;
vw += 8;
}
for ( ; i < m; ++i) {
tmp2 = vw[-1];
tmp3 = vw[ 0];
vw[-1] = _mm_add_ps(tmp2, _mm_mul_ps(_mm_add_ps(tmp1, tmp3), c));
tmp1 = tmp3;
vw += 2;
}
vl = vw - 2;
if(m >= k){
return;
}
vl = vw - 2;
c = _mm_add_ps(c, c);
c = _mm_mul_ps(c, vl[0]);
for(; m < k; ++m){
@@ -670,7 +693,7 @@ static void v4dwt_decode_step2_sse(v4* l, v4* w, int k, int m, __m128 c){
#else
static void v4dwt_decode_step1(v4* w, int count, const float c){
float* restrict fw = (float*) w;
float* OPJ_RESTRICT fw = (float*) w;
int i;
for(i = 0; i < count; ++i){
float tmp1 = fw[i*8 ];
@@ -685,8 +708,8 @@ static void v4dwt_decode_step1(v4* w, int count, const float c){
}
static void v4dwt_decode_step2(v4* l, v4* w, int k, int m, float c){
float* restrict fl = (float*) l;
float* restrict fw = (float*) w;
float* OPJ_RESTRICT fl = (float*) l;
float* OPJ_RESTRICT fw = (float*) w;
int i;
for(i = 0; i < m; ++i){
float tmp1_1 = fl[0];
@@ -737,42 +760,44 @@ static void v4dwt_decode_step2(v4* l, v4* w, int k, int m, float c){
/* <summary> */
/* Inverse 9-7 wavelet transform in 1-D. */
/* </summary> */
static void v4dwt_decode(v4dwt_t* restrict dwt){
static void v4dwt_decode(v4dwt_t* OPJ_RESTRICT dwt){
int a, b;
if(dwt->cas == 0) {
if(!((dwt->dn > 0) || (dwt->sn > 1))){
if (dwt->dn <= 0 && dwt->sn <= 1) {
return;
}
a = 0;
b = 1;
}else{
if(!((dwt->sn > 0) || (dwt->dn > 1))) {
if (dwt->sn <= 0 && dwt->dn <= 1) {
return;
}
a = 1;
b = 0;
}
v4* OPJ_RESTRICT waveleta = dwt->wavelet + a;
v4* OPJ_RESTRICT waveletb = dwt->wavelet + b;
#ifdef __SSE__
v4dwt_decode_step1_sse(dwt->wavelet+a, dwt->sn, _mm_set1_ps(K));
v4dwt_decode_step1_sse(dwt->wavelet+b, dwt->dn, _mm_set1_ps(c13318));
v4dwt_decode_step2_sse(dwt->wavelet+b, dwt->wavelet+a+1, dwt->sn, int_min(dwt->sn, dwt->dn-a), _mm_set1_ps(dwt_delta));
v4dwt_decode_step2_sse(dwt->wavelet+a, dwt->wavelet+b+1, dwt->dn, int_min(dwt->dn, dwt->sn-b), _mm_set1_ps(dwt_gamma));
v4dwt_decode_step2_sse(dwt->wavelet+b, dwt->wavelet+a+1, dwt->sn, int_min(dwt->sn, dwt->dn-a), _mm_set1_ps(dwt_beta));
v4dwt_decode_step2_sse(dwt->wavelet+a, dwt->wavelet+b+1, dwt->dn, int_min(dwt->dn, dwt->sn-b), _mm_set1_ps(dwt_alpha));
v4dwt_decode_step1_sse(waveleta, dwt->sn, _mm_set1_ps(K));
v4dwt_decode_step1_sse(waveletb, dwt->dn, _mm_set1_ps(c13318));
v4dwt_decode_step2_sse(waveletb, waveleta + 1, dwt->sn, int_min(dwt->sn, dwt->dn-a), _mm_set1_ps(dwt_delta));
v4dwt_decode_step2_sse(waveleta, waveletb + 1, dwt->dn, int_min(dwt->dn, dwt->sn-b), _mm_set1_ps(dwt_gamma));
v4dwt_decode_step2_sse(waveletb, waveleta + 1, dwt->sn, int_min(dwt->sn, dwt->dn-a), _mm_set1_ps(dwt_beta));
v4dwt_decode_step2_sse(waveleta, waveletb + 1, dwt->dn, int_min(dwt->dn, dwt->sn-b), _mm_set1_ps(dwt_alpha));
#else
v4dwt_decode_step1(dwt->wavelet+a, dwt->sn, K);
v4dwt_decode_step1(dwt->wavelet+b, dwt->dn, c13318);
v4dwt_decode_step2(dwt->wavelet+b, dwt->wavelet+a+1, dwt->sn, int_min(dwt->sn, dwt->dn-a), dwt_delta);
v4dwt_decode_step2(dwt->wavelet+a, dwt->wavelet+b+1, dwt->dn, int_min(dwt->dn, dwt->sn-b), dwt_gamma);
v4dwt_decode_step2(dwt->wavelet+b, dwt->wavelet+a+1, dwt->sn, int_min(dwt->sn, dwt->dn-a), dwt_beta);
v4dwt_decode_step2(dwt->wavelet+a, dwt->wavelet+b+1, dwt->dn, int_min(dwt->dn, dwt->sn-b), dwt_alpha);
v4dwt_decode_step1(waveleta, dwt->sn, K);
v4dwt_decode_step1(waveletb, dwt->dn, c13318);
v4dwt_decode_step2(waveletb, waveleta + 1, dwt->sn, int_min(dwt->sn, dwt->dn-a), dwt_delta);
v4dwt_decode_step2(waveleta, waveletb + 1, dwt->dn, int_min(dwt->dn, dwt->sn-b), dwt_gamma);
v4dwt_decode_step2(waveletb, waveleta + 1, dwt->sn, int_min(dwt->sn, dwt->dn-a), dwt_beta);
v4dwt_decode_step2(waveleta, waveletb + 1, dwt->dn, int_min(dwt->dn, dwt->sn-b), dwt_alpha);
#endif
}
/* <summary> */
/* Inverse 9-7 wavelet transform in 2-D. */
/* </summary> */
void dwt_decode_real(opj_tcd_tilecomp_t* restrict tilec, int numres){
void dwt_decode_real(opj_tcd_tilecomp_t* OPJ_RESTRICT tilec, int numres){
v4dwt_t h;
v4dwt_t v;
@@ -787,7 +812,7 @@ void dwt_decode_real(opj_tcd_tilecomp_t* restrict tilec, int numres){
v.wavelet = h.wavelet;
while( --numres) {
float * restrict aj = (float*) tilec->data;
float * OPJ_RESTRICT aj = (float*) tilec->data;
int bufsize = (tilec->x1 - tilec->x0) * (tilec->y1 - tilec->y0);
int j;

197
indra/libopenjpeg/mct.c
View File

@@ -29,11 +29,16 @@
* POSSIBILITY OF SUCH DAMAGE.
*/
#define OPJ_SKIP_POISON
#include "opj_includes.h"
#ifdef __SSE__
#include <xmmintrin.h>
#endif
#include "opj_includes.h"
#if defined(__GNUC__)
#pragma GCC poison malloc calloc realloc free
#endif
/* <summary> */
/* This table contains the norms of the basis function of the reversible MCT. */
@@ -49,17 +54,38 @@ static const double mct_norms_real[3] = { 1.732, 1.805, 1.573 };
/* Foward reversible MCT. */
/* </summary> */
void mct_encode(
int* restrict c0,
int* restrict c1,
int* restrict c2,
int* OPJ_RESTRICT c0,
int* OPJ_RESTRICT c1,
int* OPJ_RESTRICT c2,
int n)
{
int i;
for(i = 0; i < n; ++i) {
int i = 0;
#ifdef __SSE2__
/* Buffers are normally aligned on 16 bytes... */
if (((size_t)c0 & 0xf) == 0 && ((size_t)c1 & 0xf) == 0 && ((size_t)c2 & 0xf) == 0) {
const int cnt = n & ~3U;
for (; i < cnt; i += 4) {
__m128i y, u, v;
__m128i r = _mm_load_si128((const __m128i*) & (c0[i]));
__m128i g = _mm_load_si128((const __m128i*) & (c1[i]));
__m128i b = _mm_load_si128((const __m128i*) & (c2[i]));
y = _mm_add_epi32(g, g);
y = _mm_add_epi32(y, b);
y = _mm_add_epi32(y, r);
y = _mm_srai_epi32(y, 2);
u = _mm_sub_epi32(b, g);
v = _mm_sub_epi32(r, g);
_mm_store_si128((__m128i*) & (c0[i]), y);
_mm_store_si128((__m128i*) & (c1[i]), u);
_mm_store_si128((__m128i*) & (c2[i]), v);
}
}
#endif
for (; i < n; ++i) {
int r = c0[i];
int g = c1[i];
int b = c2[i];
int y = (r + (g * 2) + b) >> 2;
int y = (r + g + g + b) >> 2;
int u = b - g;
int v = r - g;
c0[i] = y;
@@ -72,13 +98,32 @@ void mct_encode(
/* Inverse reversible MCT. */
/* </summary> */
void mct_decode(
int* restrict c0,
int* restrict c1,
int* restrict c2,
int* OPJ_RESTRICT c0,
int* OPJ_RESTRICT c1,
int* OPJ_RESTRICT c2,
int n)
{
int i;
for (i = 0; i < n; ++i) {
int i = 0;
#ifdef __SSE2__
/* Buffers are normally aligned on 16 bytes... */
if (((size_t)c0 & 0xf) == 0 && ((size_t)c1 & 0xf) == 0 && ((size_t)c2 & 0xf) == 0) {
const int cnt = n & ~3U;
for (; i < cnt; i += 4) {
__m128i r, g, b;
__m128i y = _mm_load_si128((const __m128i*) & (c0[i]));
__m128i u = _mm_load_si128((const __m128i*) & (c1[i]));
__m128i v = _mm_load_si128((const __m128i*) & (c2[i]));
g = y;
g = _mm_sub_epi32(g, _mm_srai_epi32(_mm_add_epi32(u, v), 2));
r = _mm_add_epi32(v, g);
b = _mm_add_epi32(u, g);
_mm_store_si128((__m128i*) & (c0[i]), r);
_mm_store_si128((__m128i*) & (c1[i]), g);
_mm_store_si128((__m128i*) & (c2[i]), b);
}
}
#endif
for (; i < n; ++i) {
int y = c0[i];
int u = c1[i];
int v = c2[i];
@@ -102,13 +147,119 @@ double mct_getnorm(int compno) {
/* Foward irreversible MCT. */
/* </summary> */
void mct_encode_real(
int* restrict c0,
int* restrict c1,
int* restrict c2,
int* OPJ_RESTRICT c0,
int* OPJ_RESTRICT c1,
int* OPJ_RESTRICT c2,
int n)
{
int i;
for(i = 0; i < n; ++i) {
int i = 0;
#ifdef __SSE4_1__
/* Buffers are normally aligned on 16 bytes... */
if (((size_t)c0 & 0xf) == 0 && ((size_t)c1 & 0xf) == 0 && ((size_t)c2 & 0xf) == 0) {
const int cnt = n & ~3U;
const __m128i ry = _mm_set1_epi32(2449);
const __m128i gy = _mm_set1_epi32(4809);
const __m128i by = _mm_set1_epi32(934);
const __m128i ru = _mm_set1_epi32(1382);
const __m128i gu = _mm_set1_epi32(2714);
const __m128i gv = _mm_set1_epi32(3430);
const __m128i bv = _mm_set1_epi32(666);
const __m128i mulround = _mm_shuffle_epi32(_mm_cvtsi32_si128(4096), _MM_SHUFFLE(1, 0, 1, 0));
for (; i < cnt; i += 4) {
__m128i lo, hi, y, u, v;
__m128i r = _mm_load_si128((const __m128i*) & (c0[i]));
__m128i g = _mm_load_si128((const __m128i*) & (c1[i]));
__m128i b = _mm_load_si128((const __m128i*) & (c2[i]));
hi = _mm_shuffle_epi32(r, _MM_SHUFFLE(3, 3, 1, 1));
lo = _mm_mul_epi32(r, ry);
hi = _mm_mul_epi32(hi, ry);
lo = _mm_add_epi64(lo, mulround);
hi = _mm_add_epi64(hi, mulround);
lo = _mm_srli_epi64(lo, 13);
hi = _mm_slli_epi64(hi, 32 - 13);
y = _mm_blend_epi16(lo, hi, 0xCC);
hi = _mm_shuffle_epi32(g, _MM_SHUFFLE(3, 3, 1, 1));
lo = _mm_mul_epi32(g, gy);
hi = _mm_mul_epi32(hi, gy);
lo = _mm_add_epi64(lo, mulround);
hi = _mm_add_epi64(hi, mulround);
lo = _mm_srli_epi64(lo, 13);
hi = _mm_slli_epi64(hi, 32 - 13);
y = _mm_add_epi32(y, _mm_blend_epi16(lo, hi, 0xCC));
hi = _mm_shuffle_epi32(b, _MM_SHUFFLE(3, 3, 1, 1));
lo = _mm_mul_epi32(b, by);
hi = _mm_mul_epi32(hi, by);
lo = _mm_add_epi64(lo, mulround);
hi = _mm_add_epi64(hi, mulround);
lo = _mm_srli_epi64(lo, 13);
hi = _mm_slli_epi64(hi, 32 - 13);
y = _mm_add_epi32(y, _mm_blend_epi16(lo, hi, 0xCC));
_mm_store_si128((__m128i*) & (c0[i]), y);
lo = _mm_cvtepi32_epi64(_mm_shuffle_epi32(b, _MM_SHUFFLE(3, 2, 2, 0)));
hi = _mm_cvtepi32_epi64(_mm_shuffle_epi32(b, _MM_SHUFFLE(3, 2, 3, 1)));
lo = _mm_slli_epi64(lo, 12);
hi = _mm_slli_epi64(hi, 12);
lo = _mm_add_epi64(lo, mulround);
hi = _mm_add_epi64(hi, mulround);
lo = _mm_srli_epi64(lo, 13);
hi = _mm_slli_epi64(hi, 32 - 13);
u = _mm_blend_epi16(lo, hi, 0xCC);
hi = _mm_shuffle_epi32(r, _MM_SHUFFLE(3, 3, 1, 1));
lo = _mm_mul_epi32(r, ru);
hi = _mm_mul_epi32(hi, ru);
lo = _mm_add_epi64(lo, mulround);
hi = _mm_add_epi64(hi, mulround);
lo = _mm_srli_epi64(lo, 13);
hi = _mm_slli_epi64(hi, 32 - 13);
u = _mm_sub_epi32(u, _mm_blend_epi16(lo, hi, 0xCC));
hi = _mm_shuffle_epi32(g, _MM_SHUFFLE(3, 3, 1, 1));
lo = _mm_mul_epi32(g, gu);
hi = _mm_mul_epi32(hi, gu);
lo = _mm_add_epi64(lo, mulround);
hi = _mm_add_epi64(hi, mulround);
lo = _mm_srli_epi64(lo, 13);
hi = _mm_slli_epi64(hi, 32 - 13);
u = _mm_sub_epi32(u, _mm_blend_epi16(lo, hi, 0xCC));
_mm_store_si128((__m128i*) & (c1[i]), u);
lo = _mm_cvtepi32_epi64(_mm_shuffle_epi32(r, _MM_SHUFFLE(3, 2, 2, 0)));
hi = _mm_cvtepi32_epi64(_mm_shuffle_epi32(r, _MM_SHUFFLE(3, 2, 3, 1)));
lo = _mm_slli_epi64(lo, 12);
hi = _mm_slli_epi64(hi, 12);
lo = _mm_add_epi64(lo, mulround);
hi = _mm_add_epi64(hi, mulround);
lo = _mm_srli_epi64(lo, 13);
hi = _mm_slli_epi64(hi, 32 - 13);
v = _mm_blend_epi16(lo, hi, 0xCC);
hi = _mm_shuffle_epi32(g, _MM_SHUFFLE(3, 3, 1, 1));
lo = _mm_mul_epi32(g, gv);
hi = _mm_mul_epi32(hi, gv);
lo = _mm_add_epi64(lo, mulround);
hi = _mm_add_epi64(hi, mulround);
lo = _mm_srli_epi64(lo, 13);
hi = _mm_slli_epi64(hi, 32 - 13);
v = _mm_sub_epi32(v, _mm_blend_epi16(lo, hi, 0xCC));
hi = _mm_shuffle_epi32(b, _MM_SHUFFLE(3, 3, 1, 1));
lo = _mm_mul_epi32(b, bv);
hi = _mm_mul_epi32(hi, bv);
lo = _mm_add_epi64(lo, mulround);
hi = _mm_add_epi64(hi, mulround);
lo = _mm_srli_epi64(lo, 13);
hi = _mm_slli_epi64(hi, 32 - 13);
v = _mm_sub_epi32(v, _mm_blend_epi16(lo, hi, 0xCC));
_mm_store_si128((__m128i*) & (c2[i]), v);
}
}
#endif
for (; i < n; ++i) {
int r = c0[i];
int g = c1[i];
int b = c2[i];
@@ -125,19 +276,21 @@ void mct_encode_real(
/* Inverse irreversible MCT. */
/* </summary> */
void mct_decode_real(
float* restrict c0,
float* restrict c1,
float* restrict c2,
float* OPJ_RESTRICT c0,
float* OPJ_RESTRICT c1,
float* OPJ_RESTRICT c2,
int n)
{
int i;
#ifdef __SSE__
int count;
__m128 vrv, vgu, vgv, vbu;
vrv = _mm_set1_ps(1.402f);
vgu = _mm_set1_ps(0.34413f);
vgv = _mm_set1_ps(0.71414f);
vbu = _mm_set1_ps(1.772f);
for (i = 0; i < (n >> 3); ++i) {
count = n >> 3;
for (i = 0; i < count; ++i) {
__m128 vy, vu, vv;
__m128 vr, vg, vb;
@@ -174,7 +327,7 @@ void mct_decode_real(
float u = c1[i];
float v = c2[i];
float r = y + (v * 1.402f);
float g = y - (u * 0.34413f) - (v * (0.71414f));
float g = y - (u * 0.34413f) - (v * 0.71414f);
float b = y + (u * 1.772f);
c0[i] = r;
c1[i] = g;

56
indra/libopenjpeg/openjpeg.h
View File

@@ -40,33 +40,71 @@
==========================================================
*/
/*
The inline keyword is supported by C99 but not by C90.
Most compilers implement their own version of this keyword ...
*/
#ifndef INLINE
#if defined(_MSC_VER)
#define INLINE __forceinline
#elif defined(__GNUC__)
#define INLINE __inline__
#elif defined(__MWERKS__)
#define INLINE inline
#else
/* add other compilers here ... */
#define INLINE
#endif /* defined(<Compiler>) */
#endif /* INLINE */
#if defined(OPJ_STATIC) || !defined(_WIN32)
#define OPJ_API
#define OPJ_CALLCONV
#else
#define OPJ_CALLCONV __stdcall
/*
The following ifdef block is the standard way of creating macros which make exporting
The following ifdef block is the standard way of creating macros which make exporting
from a DLL simpler. All files within this DLL are compiled with the OPJ_EXPORTS
symbol defined on the command line. this symbol should not be defined on any project
that uses this DLL. This way any other project whose source files include this file see
OPJ_API functions as being imported from a DLL, wheras this DLL sees symbols
that uses this DLL. This way any other project whose source files include this file see
OPJ_API functions as being imported from a DLL, whereas this DLL sees symbols
defined with this macro as being exported.
*/
#if defined(OPJ_EXPORTS) || defined(DLL_EXPORT)
#define OPJ_API __declspec(dllexport)
#else
#define OPJ_API __declspec(dllimport)
#endif /* OPJ_EXPORTS */
# if defined(OPJ_EXPORTS) || defined(DLL_EXPORT)
# define OPJ_API __declspec(dllexport)
# else
# define OPJ_API __declspec(dllimport)
# endif /* OPJ_EXPORTS */
#endif /* !OPJ_STATIC || !_WIN32 */
typedef int opj_bool;
#define OPJ_TRUE 1
#define OPJ_FALSE 0
typedef char OPJ_CHAR;
typedef float OPJ_FLOAT32;
typedef double OPJ_FLOAT64;
typedef unsigned char OPJ_BYTE;
#include "opj_stdint.h"
typedef int8_t OPJ_INT8;
typedef uint8_t OPJ_UINT8;
typedef int16_t OPJ_INT16;
typedef uint16_t OPJ_UINT16;
typedef int32_t OPJ_INT32;
typedef uint32_t OPJ_UINT32;
typedef int64_t OPJ_INT64;
typedef uint64_t OPJ_UINT64;
typedef int64_t OPJ_OFF_T; /* 64-bit file offset type */
#include <stdio.h>
typedef size_t OPJ_SIZE_T;
/* Avoid compile-time warning because parameter is not used */
#define OPJ_ARG_NOT_USED(x) (void)(x)
/*
/*
==========================================================
Useful constant definitions
==========================================================

127
indra/libopenjpeg/opj_includes.h
View File

@@ -40,6 +40,8 @@
#include <stdio.h>
#include <stdarg.h>
#include <ctype.h>
#include <assert.h>
#include <limits.h>
/*
==========================================================
@@ -54,56 +56,115 @@
==========================================================
*/
/* Are restricted pointers available? (C99) */
#if (__STDC_VERSION__ >= 199901L)
#define OPJ_RESTRICT restrict
#else
/* Not a C99 compiler */
#if defined(__GNUC__)
#define OPJ_RESTRICT __restrict__
#elif defined(_MSC_VER) && (_MSC_VER >= 1400)
#define OPJ_RESTRICT __restrict
#else
#define OPJ_RESTRICT /* restrict */
#endif
#endif
/* Ignore GCC attributes if this is not GCC */
#ifndef __GNUC__
#define __attribute__(x) /* __attribute__(x) */
#endif
/*
The inline keyword is supported by C99 but not by C90.
Most compilers implement their own version of this keyword ...
*/
#ifndef INLINE
#if defined(_MSC_VER)
#define INLINE __forceinline
#elif defined(__GNUC__)
#define INLINE __inline__
#elif defined(__MWERKS__)
#define INLINE inline
#else
/* add other compilers here ... */
#define INLINE
#endif /* defined(<Compiler>) */
#endif /* INLINE */
/* Are restricted pointers available? (C99) */
#if (__STDC_VERSION__ != 199901L)
/* Not a C99 compiler */
#ifdef __GNUC__
#define restrict __restrict__
#else
#define restrict /* restrict */
#endif
#endif
/* MSVC and Borland C do not have lrintf */
#if defined(_MSC_VER) || defined(__BORLANDC__)
static INLINE long lrintf(float f){
/* MSVC before 2013 and Borland C do not have lrintf */
#if defined(_MSC_VER)
#include <intrin.h>
static INLINE long opj_lrintf(float f)
{
#ifdef _M_X64
return (long)((f>0.0f) ? (f + 0.5f):(f -0.5f));
#else
return _mm_cvt_ss2si(_mm_load_ss(&f));
/* commented out line breaks many tests */
/* return (long)((f>0.0f) ? (f + 0.5f):(f -0.5f)); */
#elif defined(_M_IX86)
int i;
_asm{
fld f
fistp i
};
return i;
#else
return (long)((f>0.0f) ? (f + 0.5f) : (f - 0.5f));
#endif
}
#elif defined(__BORLANDC__)
static INLINE long opj_lrintf(float f)
{
#ifdef _M_X64
return (long)((f > 0.0f) ? (f + 0.5f) : (f - 0.5f));
#else
int i;
_asm {
fld f
fistp i
};
return i;
#endif
}
#else
static INLINE long opj_lrintf(float f)
{
return lrintf(f);
}
#endif
#if defined(_MSC_VER) && (_MSC_VER < 1400)
#define vsnprintf _vsnprintf
#endif
/* MSVC x86 is really bad at doing int64 = int32 * int32 on its own. Use intrinsic. */
#if defined(_MSC_VER) && (_MSC_VER >= 1400) && !defined(__INTEL_COMPILER) && defined(_M_IX86)
# include <intrin.h>
# pragma intrinsic(__emul)
#endif
/* Apparently Visual Studio doesn't define __SSE__ / __SSE2__ macros */
#if defined(_M_X64)
/* Intel 64bit support SSE and SSE2 */
# ifndef __SSE__
# define __SSE__ 1
# endif
# ifndef __SSE2__
# define __SSE2__ 1
# endif
# if !defined(__SSE4_1__) && defined(__AVX__)
# define __SSE4_1__ 1
# endif
#endif
/* For x86, test the value of the _M_IX86_FP macro. */
/* See https://msdn.microsoft.com/en-us/library/b0084kay.aspx */
#if defined(_M_IX86_FP)
# if _M_IX86_FP >= 1
# ifndef __SSE__
# define __SSE__ 1
# endif
# endif
# if _M_IX86_FP >= 2
# ifndef __SSE2__
# define __SSE2__ 1
# endif
# endif
#endif
/* Type to use for bit-fields in internal headers */
typedef unsigned int OPJ_BITFIELD;
#define OPJ_UNUSED(x) (void)x
#include "j2k_lib.h"
#include "opj_malloc.h"
#include "event.h"

249
indra/libopenjpeg/opj_malloc.c Normal file
View File

@@ -0,0 +1,249 @@
/*
* The copyright in this software is being made available under the 2-clauses
* BSD License, included below. This software may be subject to other third
* party and contributor rights, including patent rights, and no such rights
* are granted under this license.
*
* Copyright (c) 2015, Mathieu Malaterre <mathieu.malaterre@gmail.com>
* Copyright (c) 2015, Matthieu Darbois
* All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS `AS IS'
* AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
* LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
* CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
* SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
* INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
* CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
* ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
* POSSIBILITY OF SUCH DAMAGE.
*/
#define OPJ_SKIP_POISON
#include "opj_includes.h"
#if defined(OPJ_HAVE_MALLOC_H) && defined(OPJ_HAVE_MEMALIGN)
# include <malloc.h>
#endif
#ifndef SIZE_MAX
# define SIZE_MAX ((size_t) -1)
#endif
static INLINE void *opj_aligned_alloc_n(size_t alignment, size_t size)
{
void* ptr;
/* alignment shall be power of 2 */
assert((alignment != 0U) && ((alignment & (alignment - 1U)) == 0U));
/* alignment shall be at least sizeof(void*) */
assert(alignment >= sizeof(void*));
if (size == 0U) { /* prevent implementation defined behavior of realloc */
return NULL;
}
#if defined(OPJ_HAVE_POSIX_MEMALIGN)
/* aligned_alloc requires c11, restrict to posix_memalign for now. Quote:
* This function was introduced in POSIX 1003.1d. Although this function is
* superseded by aligned_alloc, it is more portable to older POSIX systems
* that do not support ISO C11. */
if (posix_memalign(&ptr, alignment, size)) {
ptr = NULL;
}
/* older linux */
#elif defined(OPJ_HAVE_MEMALIGN)
ptr = memalign(alignment, size);
/* _MSC_VER */
#elif defined(OPJ_HAVE__ALIGNED_MALLOC)
ptr = _aligned_malloc(size, alignment);
#else
/*
* Generic aligned malloc implementation.
* Uses size_t offset for the integer manipulation of the pointer,
* as uintptr_t is not available in C89 to do
* bitwise operations on the pointer itself.
*/
alignment--;
{
size_t offset;
OPJ_UINT8 *mem;
/* Room for padding and extra pointer stored in front of allocated area */
size_t overhead = alignment + sizeof(void *);
/* let's be extra careful */
assert(alignment <= (SIZE_MAX - sizeof(void *)));
/* Avoid integer overflow */
if (size > (SIZE_MAX - overhead)) {
return NULL;
}
mem = (OPJ_UINT8*)malloc(size + overhead);
if (mem == NULL) {
return mem;
}
/* offset = ((alignment + 1U) - ((size_t)(mem + sizeof(void*)) & alignment)) & alignment; */
/* Use the fact that alignment + 1U is a power of 2 */
offset = ((alignment ^ ((size_t)(mem + sizeof(void*)) & alignment)) + 1U) &
alignment;
ptr = (void *)(mem + sizeof(void*) + offset);
((void**) ptr)[-1] = mem;
}
#endif
return ptr;
}
static INLINE void *opj_aligned_realloc_n(void *ptr, size_t alignment,
size_t new_size)
{
void *r_ptr;
/* alignment shall be power of 2 */
assert((alignment != 0U) && ((alignment & (alignment - 1U)) == 0U));
/* alignment shall be at least sizeof(void*) */
assert(alignment >= sizeof(void*));
if (new_size == 0U) { /* prevent implementation defined behavior of realloc */
return NULL;
}
/* no portable aligned realloc */
#if defined(OPJ_HAVE_POSIX_MEMALIGN) || defined(OPJ_HAVE_MEMALIGN)
/* glibc doc states one can mix aligned malloc with realloc */
r_ptr = realloc(ptr, new_size); /* fast path */
/* we simply use `size_t` to cast, since we are only interest in binary AND
* operator */
if (((size_t)r_ptr & (alignment - 1U)) != 0U) {
/* this is non-trivial to implement a portable aligned realloc, so use a
* simple approach where we do not need a function that return the size of an
* allocated array (eg. _msize on Windows, malloc_size on MacOS,
* malloc_usable_size on systems with glibc) */
void *a_ptr = opj_aligned_alloc_n(alignment, new_size);
if (a_ptr != NULL) {
memcpy(a_ptr, r_ptr, new_size);
}
free(r_ptr);
r_ptr = a_ptr;
}
/* _MSC_VER */
#elif defined(OPJ_HAVE__ALIGNED_MALLOC)
r_ptr = _aligned_realloc(ptr, new_size, alignment);
#else
if (ptr == NULL) {
return opj_aligned_alloc_n(alignment, new_size);
}
alignment--;
{
void *oldmem;
OPJ_UINT8 *newmem;
size_t overhead = alignment + sizeof(void *);
/* let's be extra careful */
assert(alignment <= (SIZE_MAX - sizeof(void *)));
/* Avoid integer overflow */
if (new_size > SIZE_MAX - overhead) {
return NULL;
}
oldmem = ((void**) ptr)[-1];
newmem = (OPJ_UINT8*)realloc(oldmem, new_size + overhead);
if (newmem == NULL) {
return newmem;
}
if (newmem == oldmem) {
r_ptr = ptr;
} else {
size_t old_offset;
size_t new_offset;
/* realloc created a new copy, realign the copied memory block */
old_offset = (size_t)((OPJ_UINT8*)ptr - (OPJ_UINT8*)oldmem);
/* offset = ((alignment + 1U) - ((size_t)(mem + sizeof(void*)) & alignment)) & alignment; */
/* Use the fact that alignment + 1U is a power of 2 */
new_offset = ((alignment ^ ((size_t)(newmem + sizeof(void*)) & alignment)) +
1U) & alignment;
new_offset += sizeof(void*);
r_ptr = (void *)(newmem + new_offset);
if (new_offset != old_offset) {
memmove(newmem + new_offset, newmem + old_offset, new_size);
}
((void**) r_ptr)[-1] = newmem;
}
}
#endif
return r_ptr;
}
void * opj_malloc(size_t size)
{
if (size == 0U) { /* prevent implementation defined behavior of realloc */
return NULL;
}
return malloc(size);
}
void * opj_calloc(size_t num, size_t size)
{
if (num == 0 || size == 0) {
/* prevent implementation defined behavior of realloc */
return NULL;
}
return calloc(num, size);
}
void *opj_aligned_malloc(size_t size)
{
return opj_aligned_alloc_n(16U, size);
}
void * opj_aligned_realloc(void *ptr, size_t size)
{
return opj_aligned_realloc_n(ptr, 16U, size);
}
void *opj_aligned_32_malloc(size_t size)
{
return opj_aligned_alloc_n(32U, size);
}
void * opj_aligned_32_realloc(void *ptr, size_t size)
{
return opj_aligned_realloc_n(ptr, 32U, size);
}
void opj_aligned_free(void* ptr)
{
#if defined(OPJ_HAVE_POSIX_MEMALIGN) || defined(OPJ_HAVE_MEMALIGN)
free(ptr);
#elif defined(OPJ_HAVE__ALIGNED_MALLOC)
_aligned_free(ptr);
#else
/* Generic implementation has malloced pointer stored in front of used area */
if (ptr != NULL) {
free(((void**) ptr)[-1]);
}
#endif
}
void * opj_realloc(void *ptr, size_t new_size)
{
if (new_size == 0U) { /* prevent implementation defined behavior of realloc */
return NULL;
}
return realloc(ptr, new_size);
}
void opj_free(void *ptr)
{
free(ptr);
}

126
indra/libopenjpeg/opj_malloc.h
View File

@@ -1,4 +1,9 @@
/*
* The copyright in this software is being made available under the 2-clauses
* BSD License, included below. This software may be subject to other third
* party and contributor rights, including patent rights, and no such rights
* are granted under this license.
*
* Copyright (c) 2005, Herve Drolon, FreeImage Team
* Copyright (c) 2007, Callum Lerwick <seg@haxxed.com>
* All rights reserved.
@@ -24,8 +29,10 @@
* ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
* POSSIBILITY OF SUCH DAMAGE.
*/
#ifndef __OPJ_MALLOC_H
#define __OPJ_MALLOC_H
#ifndef OPJ_MALLOC_H
#define OPJ_MALLOC_H
#include <stddef.h>
/**
@file opj_malloc.h
@brief Internal functions
@@ -36,6 +43,17 @@ The functions in opj_malloc.h are internal utilities used for memory management.
/** @defgroup MISC MISC - Miscellaneous internal functions */
/*@{*/
/* FIXME: These should be set with cmake tests, but we're currently not requiring use of cmake */
#ifdef _WIN32
#define OPJ_HAVE__ALIGNED_MALLOC
#else /* Not _WIN32 */
#if defined(__sun)
#define OPJ_HAVE_MEMALIGN
#elif defined(__linux__) || defined(__APPLE__) || defined(__FreeBSD__)
#define OPJ_HAVE_POSIX_MEMALIGN
#endif
#endif
/** @name Exported functions */
/*@{*/
/* ----------------------------------------------------------------------- */
@@ -45,90 +63,32 @@ Allocate an uninitialized memory block
@param size Bytes to allocate
@return Returns a void pointer to the allocated space, or NULL if there is insufficient memory available
*/
#ifdef ALLOC_PERF_OPT
void * OPJ_CALLCONV opj_malloc(size_t size);
#else
#define opj_malloc(size) malloc(size)
#endif
void * opj_malloc(size_t size);
/**
Allocate a memory block with elements initialized to 0
@param num Blocks to allocate
@param size Bytes per block to allocate
@param numOfElements Blocks to allocate
@param sizeOfElements Bytes per block to allocate
@return Returns a void pointer to the allocated space, or NULL if there is insufficient memory available
*/
#ifdef ALLOC_PERF_OPT
void * OPJ_CALLCONV opj_calloc(size_t _NumOfElements, size_t _SizeOfElements);
#else
#define opj_calloc(num, size) calloc(num, size)
#endif
void * opj_calloc(size_t numOfElements, size_t sizeOfElements);
/**
Allocate memory aligned to a 16 byte boundry
Allocate memory aligned to a 16 byte boundary
@param size Bytes to allocate
@return Returns a void pointer to the allocated space, or NULL if there is insufficient memory available
*/
/* FIXME: These should be set with cmake tests, but we're currently not requiring use of cmake */
#ifdef _WIN32
/* Someone should tell the mingw people that their malloc.h ought to provide _mm_malloc() */
#ifdef __GNUC__
#include <mm_malloc.h>
#define HAVE_MM_MALLOC
#else /* MSVC, Intel C++ */
#include <malloc.h>
#ifdef _mm_malloc
#define HAVE_MM_MALLOC
#endif
#endif
#else /* Not _WIN32 */
#if defined(__sun)
#define HAVE_MEMALIGN
#elif defined(__FreeBSD__)
#define HAVE_POSIX_MEMALIGN
/* Linux x86_64 and OSX always align allocations to 16 bytes */
#elif !defined(__amd64__) && !defined(__APPLE__) && !defined(_AIX)
#define HAVE_MEMALIGN
#include <malloc.h>
#endif
#endif
void * opj_aligned_malloc(size_t size);
void * opj_aligned_realloc(void *ptr, size_t size);
void opj_aligned_free(void* ptr);
#define opj_aligned_malloc(size) malloc(size)
#define opj_aligned_free(m) free(m)
#ifdef HAVE_MM_MALLOC
#undef opj_aligned_malloc
#define opj_aligned_malloc(size) _mm_malloc(size, 16)
#undef opj_aligned_free
#define opj_aligned_free(m) _mm_free(m)
#endif
#ifdef HAVE_MEMALIGN
extern void* memalign(size_t, size_t);
#undef opj_aligned_malloc
#define opj_aligned_malloc(size) memalign(16, (size))
#undef opj_aligned_free
#define opj_aligned_free(m) free(m)
#endif
#ifdef HAVE_POSIX_MEMALIGN
#undef opj_aligned_malloc
extern int posix_memalign(void**, size_t, size_t);
static INLINE void* __attribute__ ((malloc)) opj_aligned_malloc(size_t size){
void* mem = NULL;
posix_memalign(&mem, 16, size);
return mem;
}
#undef opj_aligned_free
#define opj_aligned_free(m) free(m)
#endif
#ifdef ALLOC_PERF_OPT
#undef opj_aligned_malloc
#define opj_aligned_malloc(size) opj_malloc(size)
#undef opj_aligned_free
#define opj_aligned_free(m) opj_free(m)
#endif
/**
Allocate memory aligned to a 32 byte boundary
@param size Bytes to allocate
@return Returns a void pointer to the allocated space, or NULL if there is insufficient memory available
*/
void * opj_aligned_32_malloc(size_t size);
void * opj_aligned_32_realloc(void *ptr, size_t size);
/**
Reallocate memory blocks.
@@ -136,23 +96,15 @@ Reallocate memory blocks.
@param s New size in bytes
@return Returns a void pointer to the reallocated (and possibly moved) memory block
*/
#ifdef ALLOC_PERF_OPT
void * OPJ_CALLCONV opj_realloc(void * m, size_t s);
#else
#define opj_realloc(m, s) realloc(m, s)
#endif
void * opj_realloc(void * m, size_t s);
/**
Deallocates or frees a memory block.
@param m Previously allocated memory block to be freed
*/
#ifdef ALLOC_PERF_OPT
void OPJ_CALLCONV opj_free(void * m);
#else
#define opj_free(m) free(m)
#endif
void opj_free(void * m);
#ifdef __GNUC__
#if defined(__GNUC__) && !defined(OPJ_SKIP_POISON)
#pragma GCC poison malloc calloc realloc free
#endif
@@ -161,5 +113,5 @@ void OPJ_CALLCONV opj_free(void * m);
/*@}*/
#endif /* __OPJ_MALLOC_H */
#endif /* OPJ_MALLOC_H */

51
indra/libopenjpeg/opj_stdint.h Normal file
View File

@@ -0,0 +1,51 @@
/*
* The copyright in this software is being made available under the 2-clauses
* BSD License, included below. This software may be subject to other third
* party and contributor rights, including patent rights, and no such rights
* are granted under this license.
*
* Copyright (c) 2012, Mathieu Malaterre <mathieu.malaterre@gmail.com>
* All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS `AS IS'
* AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
* LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
* CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
* SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
* INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
* CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
* ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
* POSSIBILITY OF SUCH DAMAGE.
*/
#ifndef OPJ_STDINT_H
#define OPJ_STDINT_H
#if defined(__linux__) || defined(__APPLE__) || defined(__FreeBSD__) || _MSC_VER >= 1900
#include <stdint.h>
#else
#if defined(_WIN32)
typedef signed __int8 int8_t;
typedef unsigned __int8 uint8_t;
typedef signed __int16 int16_t;
typedef unsigned __int16 uint16_t;
typedef signed __int32 int32_t;
typedef unsigned __int32 uint32_t;
typedef signed __int64 int64_t;
typedef unsigned __int64 uint64_t;
#else
#error unsupported platform
#endif
#endif
#endif /* OPJ_STDINT_H */

16
indra/libopenjpeg/t1.c
View File

@@ -1427,7 +1427,7 @@ void t1_encode_cblks(
opj_tcd_resolution_t *res = &tilec->resolutions[resno];
for (bandno = 0; bandno < res->numbands; ++bandno) {
opj_tcd_band_t* restrict band = &res->bands[bandno];
opj_tcd_band_t* OPJ_RESTRICT band = &res->bands[bandno];
int bandconst = 8192 * 8192 / ((int) floor(band->stepsize * 8192));
for (precno = 0; precno < res->pw * res->ph; ++precno) {
@@ -1435,8 +1435,8 @@ void t1_encode_cblks(
for (cblkno = 0; cblkno < prc->cw * prc->ch; ++cblkno) {
opj_tcd_cblk_enc_t* cblk = &prc->cblks.enc[cblkno];
int* restrict datap;
int* restrict tiledp;
int* OPJ_RESTRICT datap;
int* OPJ_RESTRICT tiledp;
int cblk_w;
int cblk_h;
int i, j;
@@ -1517,14 +1517,14 @@ void t1_decode_cblks(
opj_tcd_resolution_t* res = &tilec->resolutions[resno];
for (bandno = 0; bandno < res->numbands; ++bandno) {
opj_tcd_band_t* restrict band = &res->bands[bandno];
opj_tcd_band_t* OPJ_RESTRICT band = &res->bands[bandno];
for (precno = 0; precno < res->pw * res->ph; ++precno) {
opj_tcd_precinct_t* precinct = &band->precincts[precno];
for (cblkno = 0; cblkno < precinct->cw * precinct->ch; ++cblkno) {
opj_tcd_cblk_dec_t* cblk = &precinct->cblks.dec[cblkno];
int* restrict datap;
int* OPJ_RESTRICT datap;
int cblk_w, cblk_h;
int x, y;
int i, j;
@@ -1566,7 +1566,7 @@ void t1_decode_cblks(
}
if (tccp->qmfbid == 1) {
int* restrict tiledp = &tilec->data[(y * tile_w) + x];
int* OPJ_RESTRICT tiledp = &tilec->data[(y * tile_w) + x];
for (j = 0; j < cblk_h; ++j) {
for (i = 0; i < cblk_w; ++i) {
int tmp = datap[(j * cblk_w) + i];
@@ -1574,9 +1574,9 @@ void t1_decode_cblks(
}
}
} else { /* if (tccp->qmfbid == 0) */
float* restrict tiledp = (float*) &tilec->data[(y * tile_w) + x];
float* OPJ_RESTRICT tiledp = (float*) &tilec->data[(y * tile_w) + x];
for (j = 0; j < cblk_h; ++j) {
float* restrict tiledp2 = tiledp;
float* OPJ_RESTRICT tiledp2 = tiledp;
for (i = 0; i < cblk_w; ++i) {
float tmp = *datap * band->stepsize;
*tiledp2 = tmp;

6
indra/libopenjpeg/t1_generate_luts.c
View File

@@ -194,7 +194,7 @@ int main(){
printf("/* This file was automatically generated by t1_generate_luts.c */\n\n");
// lut_ctxno_zc
/* lut_ctxno_zc */
for (j = 0; j < 4; ++j) {
for (i = 0; i < 256; ++i) {
int orient = j;
@@ -215,7 +215,7 @@ int main(){
}
printf("%i\n};\n\n", lut_ctxno_zc[1023]);
// lut_ctxno_sc
/* lut_ctxno_sc */
printf("static char lut_ctxno_sc[256] = {\n ");
for (i = 0; i < 255; ++i) {
printf("0x%x, ", t1_init_ctxno_sc(i << 4));
@@ -224,7 +224,7 @@ int main(){
}
printf("0x%x\n};\n\n", t1_init_ctxno_sc(255 << 4));
// lut_spb
/* lut_spb */
printf("static char lut_spb[256] = {\n ");
for (i = 0; i < 255; ++i) {
printf("%i, ", t1_init_spb(i << 4));

9
indra/libopenjpeg/t2.c
View File

@@ -30,6 +30,7 @@
*/
#include "opj_includes.h"
#include <assert.h>
/** @defgroup T2 T2 - Implementation of a tier-2 coding */
/*@{*/
@@ -340,13 +341,15 @@ static int t2_decode_packet(opj_t2_t* t2, unsigned char *src, int len, opj_tcd_t
int precno = pi->precno; /* precinct value */
int layno = pi->layno; /* quality layer value */
opj_tcd_resolution_t* res = &tile->comps[compno].resolutions[resno];
unsigned char *hd = NULL;
int present;
opj_bio_t *bio = NULL; /* BIO component */
opj_tcd_resolution_t* res;
assert(&tile->comps[compno] != NULL);
res = &tile->comps[compno].resolutions[resno];
if (layno == 0) {
for (bandno = 0; bandno < res->numbands; bandno++) {
opj_tcd_band_t *band = &res->bands[bandno];

2
indra/libopenjpeg/tcd.c
View File

@@ -1507,7 +1507,7 @@ opj_bool tcd_decode_tile(opj_tcd_t *tcd, unsigned char *src, int len, int tileno
for(j = res->y0; j < res->y1; ++j) {
for(i = res->x0; i < res->x1; ++i) {
float tmp = ((float*)tilec->data)[i - res->x0 + (j - res->y0) * tw];
int v = lrintf(tmp);
int v = opj_lrintf(tmp);
v += adjust;
imagec->data[(i - offset_x) + (j - offset_y) * w] = int_clamp(v, min, max);
}