libretro/scummvm
1
0
Fork 0
mirror of https://github.com/libretro/scummvm.git synced 2025-04-01 06:14:07 +00:00
Code Issues Actions 2 Packages Projects Releases Wiki Activity

IMAGE: INDEO3: Make more use of aligned loads

Browse Source
This commit is contained in:
Cameron Cawley 2024-04-01 22:22:34 +01:00 committed by Eugene Sandulenko
parent f55925e58a
commit 67fc0b6b7d
1 changed files with 81 additions and 79 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

160
image/codecs/indeo3.cpp
View File

@ -382,6 +382,28 @@ typedef struct {
} \
lp2 = 4;
#define COPY_4_PIXELS(dst, src, count, pitch) \
if (IS_ALIGNED(src, 4)) { \
for (i = 0, j = 0; i < count; i++, j += pitch) \
dst[j] = src[j]; \
} else { \
for (i = 0, j = 0; i < count; i++, j += pitch) \
dst[j] = READ_UINT32(src+j); \
}
#define COPY_8_PIXELS(dst, src, count, pitch) \
if (IS_ALIGNED(src, 4)) { \
for (i = 0, j = 0; i < count; i++, j += pitch) {\
dst[j] = src[j]; \
dst[j+1] = src[j+1]; \
} \
} else { \
for (i = 0, j = 0; i < count; i++, j += pitch) { \
dst[j] = READ_UINT32(src+j); \
dst[j+1] = READ_UINT32(src+j+1); \
} \
}
void Indeo3Decoder::decodeChunk(byte *cur, byte *ref, int width, int height,
const byte *buf1, uint32 fflags2, const byte *hdr,
const byte *buf2, int min_width_160) {
@ -390,7 +412,8 @@ void Indeo3Decoder::decodeChunk(byte *cur, byte *ref, int width, int height,
uint32 bit_pos, lv, lv1, lv2;
int32 *width_tbl, width_tbl_arr[10];
const int8 *ref_vectors;
byte *cur_frm_pos, *ref_frm_pos, *cp, *cp2;
uint32 *cur_frm_pos, *ref_frm_pos;
byte *cp, *cp2;
uint32 *cur_lp, *ref_lp;
const uint32 *correction_lp[2], *correctionloworder_lp[2], *correctionhighorder_lp[2];
byte *correction_type_sp[2];
@ -402,7 +425,7 @@ void Indeo3Decoder::decodeChunk(byte *cur, byte *ref, int width, int height,
if ((width & 3) != 0) {
// This isn't a valid width according to http://wiki.multimedia.cx/index.php?title=Indeo_3
warning("Indeo3 file with width not divisible by 4. This will cause unaligned writes");
error("Indeo3 file with width not divisible by 4. This will cause unaligned writes");
}
bit_buf = 0;
@ -468,7 +491,7 @@ void Indeo3Decoder::decodeChunk(byte *cur, byte *ref, int width, int height,
}
}
cur_frm_pos = cur + width * strip->ypos + strip->xpos;
cur_frm_pos = (uint32 *)(cur + width * strip->ypos + strip->xpos);
if ((blks_width = strip->width) < 0)
blks_width += 3;
@ -476,10 +499,11 @@ void Indeo3Decoder::decodeChunk(byte *cur, byte *ref, int width, int height,
blks_height = strip->height;
if (ref_vectors != NULL) {
ref_frm_pos = ref + (ref_vectors[0] + strip->ypos) * width +
ref_vectors[1] + strip->xpos;
/* This may not be aligned, so is accessed using functions like READ_UINT32 */
ref_frm_pos = (uint32 *)(ref + (ref_vectors[0] + strip->ypos) * width +
ref_vectors[1] + strip->xpos);
} else
ref_frm_pos = cur_frm_pos - width_tbl[4];
ref_frm_pos = (uint32 *)(((byte *)cur_frm_pos) - width_tbl[4]);
if (cmd == 2) {
if (bit_pos <= 0) {
@ -492,10 +516,9 @@ void Indeo3Decoder::decodeChunk(byte *cur, byte *ref, int width, int height,
if (cmd == 0 || ref_vectors != NULL) {
for (lp1 = 0; lp1 < blks_width; lp1++) {
for (i = 0, j = 0; i < blks_height; i++, j += width_tbl[1])
((uint32 *)cur_frm_pos)[j] = READ_UINT32(((uint32 *)ref_frm_pos)+j);
cur_frm_pos += 4;
ref_frm_pos += 4;
COPY_4_PIXELS(cur_frm_pos, ref_frm_pos, blks_height, width_tbl[1]);
cur_frm_pos++;
ref_frm_pos++;
}
} else if (cmd != 1)
return;
@ -507,7 +530,7 @@ void Indeo3Decoder::decodeChunk(byte *cur, byte *ref, int width, int height,
if ((lv - 8) <= 7 && (k == 0 || k == 3 || k == 10)) {
cp2 = _ModPred + ((lv - 8) << 7);
cp = ref_frm_pos;
cp = (byte *)ref_frm_pos;
for (i = 0; i < blks_width << 2; i++) {
int v = *cp >> 1;
*(cp++) = cp2[v];
@ -535,8 +558,8 @@ void Indeo3Decoder::decodeChunk(byte *cur, byte *ref, int width, int height,
for (lp1 = 0; lp1 < blks_width; lp1++) {
for (lp2 = 0; lp2 < 4; ) {
k = *buf1++;
cur_lp = ((uint32 *)cur_frm_pos) + width_tbl[lp2];
ref_lp = ((uint32 *)ref_frm_pos) + width_tbl[lp2];
cur_lp = cur_frm_pos + width_tbl[lp2];
ref_lp = ref_frm_pos + width_tbl[lp2];
switch (correction_type_sp[0][k]) {
case 0:
@ -565,16 +588,14 @@ void Indeo3Decoder::decodeChunk(byte *cur, byte *ref, int width, int height,
case 2:
if (lp2 == 0) {
for (i = 0, j = 0; i < 2; i++, j += width_tbl[1])
cur_lp[j] = READ_UINT32(ref_lp+j);
COPY_4_PIXELS(cur_lp, ref_lp, 2, width_tbl[1]);
lp2 += 2;
}
break;
case 3:
if (lp2 < 2) {
for (i = 0, j = 0; i < (3 - lp2); i++, j += width_tbl[1])
cur_lp[j] = READ_UINT32(ref_lp+j);
COPY_4_PIXELS(cur_lp, ref_lp, (3 - lp2), width_tbl[1]);
lp2 = 3;
}
break;
@ -584,8 +605,7 @@ void Indeo3Decoder::decodeChunk(byte *cur, byte *ref, int width, int height,
RLE_V3_CHECK(buf1,rle_v1,rle_v2,rle_v3)
if (rle_v1 == 1 || ref_vectors != NULL) {
for (i = 0, j = 0; i < 4; i++, j += width_tbl[1])
cur_lp[j] = READ_UINT32(ref_lp+j);
COPY_4_PIXELS(cur_lp, ref_lp, 4, width_tbl[1]);
}
RLE_V2_CHECK(buf1,rle_v2, rle_v3,lp2)
@ -599,8 +619,7 @@ void Indeo3Decoder::decodeChunk(byte *cur, byte *ref, int width, int height,
LP2_CHECK(buf1,rle_v3,lp2)
// fall through
case 4:
for (i = 0, j = 0; i < (4 - lp2); i++, j += width_tbl[1])
cur_lp[j] = READ_UINT32(ref_lp+j);
COPY_4_PIXELS(cur_lp, ref_lp, (4 - lp2), width_tbl[1]);
lp2 = 4;
break;
@ -620,8 +639,7 @@ void Indeo3Decoder::decodeChunk(byte *cur, byte *ref, int width, int height,
return;
}
if (ref_vectors != NULL) {
for (i = 0, j = 0; i < 4; i++, j += width_tbl[1])
cur_lp[j] = READ_UINT32(ref_lp+j);
COPY_4_PIXELS(cur_lp, ref_lp, 4, width_tbl[1]);
}
lp2 = 4;
break;
@ -642,12 +660,12 @@ void Indeo3Decoder::decodeChunk(byte *cur, byte *ref, int width, int height,
}
}
cur_frm_pos += 4;
ref_frm_pos += 4;
cur_frm_pos++;
ref_frm_pos++;
}
cur_frm_pos += ((width - blks_width) * 4);
ref_frm_pos += ((width - blks_width) * 4);
cur_frm_pos += (width - blks_width);
ref_frm_pos += (width - blks_width);
}
break;
@ -662,8 +680,8 @@ void Indeo3Decoder::decodeChunk(byte *cur, byte *ref, int width, int height,
for (lp2 = 0; lp2 < 4; ) {
k = *buf1++;
cur_lp = ((uint32 *)cur_frm_pos) + width_tbl[lp2 * 2];
ref_lp = ((uint32 *)cur_frm_pos) + width_tbl[(lp2 * 2) - 1];
cur_lp = cur_frm_pos + width_tbl[lp2 * 2];
ref_lp = cur_frm_pos + width_tbl[(lp2 * 2) - 1];
switch (correction_type_sp[lp2 & 0x01][k]) {
case 0:
@ -691,16 +709,18 @@ void Indeo3Decoder::decodeChunk(byte *cur, byte *ref, int width, int height,
case 2:
if (lp2 == 0) {
lv = READ_UINT32(ref_lp);
for (i = 0, j = 0; i < 4; i++, j += width_tbl[1])
cur_lp[j] = READ_UINT32(ref_lp);
cur_lp[j] = lv;
lp2 += 2;
}
break;
case 3:
if (lp2 < 2) {
lv = READ_UINT32(ref_lp);
for (i = 0, j = 0; i < 6 - (lp2 * 2); i++, j += width_tbl[1])
cur_lp[j] = READ_UINT32(ref_lp);
cur_lp[j] = lv;
lp2 = 3;
}
break;
@ -724,8 +744,7 @@ void Indeo3Decoder::decodeChunk(byte *cur, byte *ref, int width, int height,
RLE_V3_CHECK(buf1,rle_v1,rle_v2,rle_v3)
if (rle_v1 == 1) {
for (i = 0, j = 0; i < 8; i++, j += width_tbl[1])
cur_lp[j] = READ_UINT32(ref_lp+j);
COPY_4_PIXELS(cur_lp, ref_lp, 8, width_tbl[1]);
}
RLE_V2_CHECK(buf1,rle_v2, rle_v3,lp2)
@ -739,8 +758,9 @@ void Indeo3Decoder::decodeChunk(byte *cur, byte *ref, int width, int height,
LP2_CHECK(buf1,rle_v3,lp2)
// fall through
case 4:
lv = READ_UINT32(ref_lp);
for (i = 0, j = 0; i < 8 - (lp2 * 2); i++, j += width_tbl[1])
cur_lp[j] = READ_UINT32(ref_lp);
cur_lp[j] = lv;
lp2 = 4;
break;
@ -762,10 +782,10 @@ void Indeo3Decoder::decodeChunk(byte *cur, byte *ref, int width, int height,
}
}
cur_frm_pos += 4;
cur_frm_pos++;
}
cur_frm_pos += (((width * 2) - blks_width) * 4);
cur_frm_pos += ((width * 2) - blks_width);
flag1 = 0;
}
break;
@ -778,8 +798,8 @@ void Indeo3Decoder::decodeChunk(byte *cur, byte *ref, int width, int height,
for (lp1 = 0; lp1 < blks_width; lp1 += 2) {
for (lp2 = 0; lp2 < 4; ) {
k = *buf1++;
cur_lp = ((uint32 *)cur_frm_pos) + width_tbl[lp2 * 2];
ref_lp = ((uint32 *)cur_frm_pos) + width_tbl[(lp2 * 2) - 1];
cur_lp = cur_frm_pos + width_tbl[lp2 * 2];
ref_lp = cur_frm_pos + width_tbl[(lp2 * 2) - 1];
lv1 = READ_UINT32(ref_lp);
lv2 = READ_UINT32(ref_lp+1);
if (lp2 == 0 && flag1 != 0) {
@ -944,10 +964,10 @@ void Indeo3Decoder::decodeChunk(byte *cur, byte *ref, int width, int height,
}
}
cur_frm_pos += 8;
cur_frm_pos += 2;
}
cur_frm_pos += (((width * 2) - blks_width) * 4);
cur_frm_pos += ((width * 2) - blks_width);
flag1 = 0;
}
} else {
@ -955,8 +975,8 @@ void Indeo3Decoder::decodeChunk(byte *cur, byte *ref, int width, int height,
for (lp1 = 0; lp1 < blks_width; lp1 += 2) {
for (lp2 = 0; lp2 < 4; ) {
k = *buf1++;
cur_lp = ((uint32 *)cur_frm_pos) + width_tbl[lp2 * 2];
ref_lp = ((uint32 *)ref_frm_pos) + width_tbl[lp2 * 2];
cur_lp = cur_frm_pos + width_tbl[lp2 * 2];
ref_lp = ref_frm_pos + width_tbl[lp2 * 2];
switch (correction_type_sp[lp2 & 0x01][k]) {
case 0:
@ -981,20 +1001,14 @@ void Indeo3Decoder::decodeChunk(byte *cur, byte *ref, int width, int height,
case 2:
if (lp2 == 0) {
for (i = 0, j = 0; i < 4; i++, j += width_tbl[1]) {
cur_lp[j] = READ_UINT32(ref_lp+j);
cur_lp[j+1] = READ_UINT32(ref_lp+j+1);
}
COPY_8_PIXELS(cur_lp, ref_lp, 4, width_tbl[1]);
lp2 += 2;
}
break;
case 3:
if (lp2 < 2) {
for (i = 0, j = 0; i < 6 - (lp2 * 2); i++, j += width_tbl[1]) {
cur_lp[j] = READ_UINT32(ref_lp+j);
cur_lp[j+1] = READ_UINT32(ref_lp+j+1);
}
COPY_8_PIXELS(cur_lp, ref_lp, 6 - (lp2 * 2), width_tbl[1]);
lp2 = 3;
}
break;
@ -1002,10 +1016,7 @@ void Indeo3Decoder::decodeChunk(byte *cur, byte *ref, int width, int height,
case 8:
if (lp2 == 0) {
RLE_V3_CHECK(buf1,rle_v1,rle_v2,rle_v3)
for (i = 0, j = 0; i < 8; i++, j += width_tbl[1]) {
((uint32 *)cur_frm_pos)[j] = READ_UINT32(((uint32 *)ref_frm_pos)+j);
((uint32 *)cur_frm_pos)[j+1] = READ_UINT32(((uint32 *)ref_frm_pos)+j+1);
}
COPY_8_PIXELS(cur_frm_pos, ref_frm_pos, 8, width_tbl[1]);
RLE_V2_CHECK(buf1,rle_v2, rle_v3,lp2)
break;
} else {
@ -1019,10 +1030,7 @@ void Indeo3Decoder::decodeChunk(byte *cur, byte *ref, int width, int height,
// fall through
case 6:
case 4:
for (i = 0, j = 0; i < 8 - (lp2 * 2); i++, j += width_tbl[1]) {
cur_lp[j] = READ_UINT32(ref_lp+j);
cur_lp[j+1] = READ_UINT32(ref_lp+j+1);
}
COPY_8_PIXELS(cur_lp, ref_lp, 8 - (lp2 * 2), width_tbl[1]);
lp2 = 4;
break;
@ -1033,7 +1041,7 @@ void Indeo3Decoder::decodeChunk(byte *cur, byte *ref, int width, int height,
lv += (lv << 8);
lv += (lv << 16);
for (i = 0, j = 0; i < 8; i++, j += width_tbl[1])
((uint32 *)cur_frm_pos)[j] = ((uint32 *)cur_frm_pos)[j+1] = lv;
cur_frm_pos[j] = cur_frm_pos[j+1] = lv;
LV1_CHECK(buf1,rle_v3,lv1,lp2)
break;
@ -1042,12 +1050,12 @@ void Indeo3Decoder::decodeChunk(byte *cur, byte *ref, int width, int height,
}
}
cur_frm_pos += 8;
ref_frm_pos += 8;
cur_frm_pos += 2;
ref_frm_pos += 2;
}
cur_frm_pos += (((width * 2) - blks_width) * 4);
ref_frm_pos += (((width * 2) - blks_width) * 4);
cur_frm_pos += ((width * 2) - blks_width);
ref_frm_pos += ((width * 2) - blks_width);
}
}
break;
@ -1060,8 +1068,8 @@ void Indeo3Decoder::decodeChunk(byte *cur, byte *ref, int width, int height,
for (lp1 = 0; lp1 < blks_width; lp1++) {
for (lp2 = 0; lp2 < 4; ) {
k = *buf1++;
cur_lp = ((uint32 *)cur_frm_pos) + width_tbl[lp2 * 2];
ref_lp = ((uint32 *)ref_frm_pos) + width_tbl[lp2 * 2];
cur_lp = cur_frm_pos + width_tbl[lp2 * 2];
ref_lp = ref_frm_pos + width_tbl[lp2 * 2];
switch (correction_type_sp[lp2 & 0x01][k]) {
case 0:
@ -1092,16 +1100,14 @@ void Indeo3Decoder::decodeChunk(byte *cur, byte *ref, int width, int height,
case 2:
if (lp2 == 0) {
for (i = 0, j = 0; i < 4; i++, j += width_tbl[1])
cur_lp[j] = READ_UINT32(ref_lp+j);
COPY_4_PIXELS(cur_lp, ref_lp, 4, width_tbl[1]);
lp2 += 2;
}
break;
case 3:
if (lp2 < 2) {
for (i = 0, j = 0; i < 6 - (lp2 * 2); i++, j += width_tbl[1])
cur_lp[j] = READ_UINT32(ref_lp+j);
COPY_4_PIXELS(cur_lp, ref_lp, 6 - (lp2 * 2), width_tbl[1]);
lp2 = 3;
}
break;
@ -1109,10 +1115,7 @@ void Indeo3Decoder::decodeChunk(byte *cur, byte *ref, int width, int height,
case 8:
if (lp2 == 0) {
RLE_V3_CHECK(buf1,rle_v1,rle_v2,rle_v3)
for (i = 0, j = 0; i < 8; i++, j += width_tbl[1])
cur_lp[j] = READ_UINT32(ref_lp+j);
COPY_4_PIXELS(cur_lp, ref_lp, 8, width_tbl[1]);
RLE_V2_CHECK(buf1,rle_v2, rle_v3,lp2)
break;
} else {
@ -1126,8 +1129,7 @@ void Indeo3Decoder::decodeChunk(byte *cur, byte *ref, int width, int height,
// fall through
case 4:
case 6:
for (i = 0, j = 0; i < 8 - (lp2 * 2); i++, j += width_tbl[1])
cur_lp[j] = READ_UINT32(ref_lp+j);
COPY_4_PIXELS(cur_lp, ref_lp, 8 - (lp2 * 2), width_tbl[1]);
lp2 = 4;
break;
@ -1147,12 +1149,12 @@ void Indeo3Decoder::decodeChunk(byte *cur, byte *ref, int width, int height,
}
}
cur_frm_pos += 4;
ref_frm_pos += 4;
cur_frm_pos++;
ref_frm_pos++;
}
cur_frm_pos += (((width * 2) - blks_width) * 4);
ref_frm_pos += (((width * 2) - blks_width) * 4);
cur_frm_pos += ((width * 2) - blks_width);
ref_frm_pos += ((width * 2) - blks_width);
}
break;