Commit 13a09979 authored by Ronald S. Bultje's avatar Ronald S. Bultje
Browse files

swscale: change prototypes of scaled YUV output functions.

Remove unused variables "flags" and "dstFormat" in yuv2packed1,
merge source rows per plane for yuv2packed[12], and make every
source argument int16_t (some where invalidly set to uint16_t).
This prevents stack pollution and is part of the Great Evil Plan
to simplify swscale.
parent dff5a835
......@@ -98,10 +98,9 @@ yuv2yuvX_altivec_real(SwsContext *c,
int lumFilterSize, const int16_t *chrFilter,
const int16_t **chrUSrc, const int16_t **chrVSrc,
int chrFilterSize, const int16_t **alpSrc,
uint8_t *dest, uint8_t *uDest,
uint8_t *vDest, uint8_t *aDest,
int dstW, int chrDstW)
uint8_t *dest[4], int dstW, int chrDstW)
{
uint8_t *yDest = dest[0], *uDest = dest[1], *vDest = dest[2];
const vector signed int vini = {(1 << 18), (1 << 18), (1 << 18), (1 << 18)};
register int i, j;
{
......@@ -150,7 +149,7 @@ yuv2yuvX_altivec_real(SwsContext *c,
val[i] += lumSrc[j][i] * lumFilter[j];
}
}
altivec_packIntArrayToCharArray(val, dest, dstW);
altivec_packIntArrayToCharArray(val, yDest, dstW);
}
if (uDest != 0) {
DECLARE_ALIGNED(16, int, u)[chrDstW];
......
This diff is collapsed.
......@@ -60,37 +60,27 @@ typedef int (*SwsFunc)(struct SwsContext *context, const uint8_t* src[],
typedef void (*yuv2planar1_fn) (struct SwsContext *c,
const int16_t *lumSrc, const int16_t *chrUSrc,
const int16_t *chrVSrc, const int16_t *alpSrc,
uint8_t *dest,
uint8_t *uDest, uint8_t *vDest, uint8_t *aDest,
int dstW, int chrDstW);
typedef void (*yuv2planarX_fn) (struct SwsContext *c,
const int16_t *lumFilter, const int16_t **lumSrc, int lumFilterSize,
uint8_t *dest[4], int dstW, int chrDstW);
typedef void (*yuv2planarX_fn) (struct SwsContext *c, const int16_t *lumFilter,
const int16_t **lumSrc, int lumFilterSize,
const int16_t *chrFilter, const int16_t **chrUSrc,
const int16_t **chrVSrc, int chrFilterSize,
const int16_t **alpSrc,
uint8_t *dest,
uint8_t *uDest, uint8_t *vDest, uint8_t *aDest,
const int16_t **chrVSrc, int chrFilterSize,
const int16_t **alpSrc, uint8_t *dest[4],
int dstW, int chrDstW);
typedef void (*yuv2packed1_fn) (struct SwsContext *c,
const uint16_t *buf0,
const uint16_t *ubuf0, const uint16_t *ubuf1,
const uint16_t *vbuf0, const uint16_t *vbuf1,
const uint16_t *abuf0,
uint8_t *dest,
int dstW, int uvalpha, int dstFormat, int flags, int y);
typedef void (*yuv2packed2_fn) (struct SwsContext *c,
const uint16_t *buf0, const uint16_t *buf1,
const uint16_t *ubuf0, const uint16_t *ubuf1,
const uint16_t *vbuf0, const uint16_t *vbuf1,
const uint16_t *abuf0, const uint16_t *abuf1,
uint8_t *dest,
typedef void (*yuv2packed1_fn) (struct SwsContext *c, const int16_t *lumSrc,
const int16_t *chrUSrc[2], const int16_t *chrVSrc[2],
const int16_t *alpSrc, uint8_t *dest,
int dstW, int uvalpha, int y);
typedef void (*yuv2packed2_fn) (struct SwsContext *c, const int16_t *lumSrc[2],
const int16_t *chrUSrc[2], const int16_t *chrVSrc[2],
const int16_t *alpSrc[2], uint8_t *dest,
int dstW, int yalpha, int uvalpha, int y);
typedef void (*yuv2packedX_fn) (struct SwsContext *c,
const int16_t *lumFilter, const int16_t **lumSrc, int lumFilterSize,
typedef void (*yuv2packedX_fn) (struct SwsContext *c, const int16_t *lumFilter,
const int16_t **lumSrc, int lumFilterSize,
const int16_t *chrFilter, const int16_t **chrUSrc,
const int16_t **chrVSrc, int chrFilterSize,
const int16_t **alpSrc, uint8_t *dest,
int dstW, int dstY);
int dstW, int y);
/* This struct should be aligned on at least a 32-byte boundary. */
typedef struct SwsContext {
......
......@@ -75,9 +75,11 @@ static void RENAME(yuv2yuvX)(SwsContext *c, const int16_t *lumFilter,
const int16_t *chrFilter, const int16_t **chrUSrc,
const int16_t **chrVSrc,
int chrFilterSize, const int16_t **alpSrc,
uint8_t *dest, uint8_t *uDest, uint8_t *vDest,
uint8_t *aDest, int dstW, int chrDstW)
uint8_t *dest[4], int dstW, int chrDstW)
{
uint8_t *yDest = dest[0], *uDest = dest[1], *vDest = dest[2],
*aDest = CONFIG_SWSCALE_ALPHA ? dest[3] : NULL;
if (uDest) {
x86_reg uv_off = c->uv_off;
YSCALEYUV2YV12X(CHR_MMX_FILTER_OFFSET, uDest, chrDstW, 0)
......@@ -87,7 +89,7 @@ static void RENAME(yuv2yuvX)(SwsContext *c, const int16_t *lumFilter,
YSCALEYUV2YV12X(ALP_MMX_FILTER_OFFSET, aDest, dstW, 0)
}
YSCALEYUV2YV12X(LUM_MMX_FILTER_OFFSET, dest, dstW, 0)
YSCALEYUV2YV12X(LUM_MMX_FILTER_OFFSET, yDest, dstW, 0)
}
#define YSCALEYUV2YV12X_ACCURATE(offset, dest, end, pos) \
......@@ -156,9 +158,11 @@ static void RENAME(yuv2yuvX_ar)(SwsContext *c, const int16_t *lumFilter,
const int16_t *chrFilter, const int16_t **chrUSrc,
const int16_t **chrVSrc,
int chrFilterSize, const int16_t **alpSrc,
uint8_t *dest, uint8_t *uDest, uint8_t *vDest,
uint8_t *aDest, int dstW, int chrDstW)
uint8_t *dest[4], int dstW, int chrDstW)
{
uint8_t *yDest = dest[0], *uDest = dest[1], *vDest = dest[2],
*aDest = CONFIG_SWSCALE_ALPHA ? dest[3] : NULL;
if (uDest) {
x86_reg uv_off = c->uv_off;
YSCALEYUV2YV12X_ACCURATE(CHR_MMX_FILTER_OFFSET, uDest, chrDstW, 0)
......@@ -168,19 +172,20 @@ static void RENAME(yuv2yuvX_ar)(SwsContext *c, const int16_t *lumFilter,
YSCALEYUV2YV12X_ACCURATE(ALP_MMX_FILTER_OFFSET, aDest, dstW, 0)
}
YSCALEYUV2YV12X_ACCURATE(LUM_MMX_FILTER_OFFSET, dest, dstW, 0)
YSCALEYUV2YV12X_ACCURATE(LUM_MMX_FILTER_OFFSET, yDest, dstW, 0)
}
static void RENAME(yuv2yuv1)(SwsContext *c, const int16_t *lumSrc,
const int16_t *chrUSrc, const int16_t *chrVSrc,
const int16_t *alpSrc,
uint8_t *dest, uint8_t *uDest, uint8_t *vDest,
uint8_t *aDest, int dstW, int chrDstW)
uint8_t *dst[4], int dstW, int chrDstW)
{
int p= 4;
const int16_t *src[4]= { alpSrc + dstW, lumSrc + dstW, chrUSrc + chrDstW, chrVSrc + chrDstW };
uint8_t *dst[4]= { aDest, dest, uDest, vDest };
x86_reg counter[4]= { dstW, dstW, chrDstW, chrDstW };
const int16_t *src[4]= {
lumSrc + dstW, chrUSrc + chrDstW,
chrVSrc + chrDstW, alpSrc + dstW
};
x86_reg counter[4]= { dstW, chrDstW, chrDstW, dstW };
while (p--) {
if (dst[p]) {
......@@ -207,13 +212,14 @@ static void RENAME(yuv2yuv1)(SwsContext *c, const int16_t *lumSrc,
static void RENAME(yuv2yuv1_ar)(SwsContext *c, const int16_t *lumSrc,
const int16_t *chrUSrc, const int16_t *chrVSrc,
const int16_t *alpSrc,
uint8_t *dest, uint8_t *uDest, uint8_t *vDest,
uint8_t *aDest, int dstW, int chrDstW)
uint8_t *dst[4], int dstW, int chrDstW)
{
int p= 4;
const int16_t *src[4]= { alpSrc + dstW, lumSrc + dstW, chrUSrc + chrDstW, chrVSrc + chrDstW };
uint8_t *dst[4]= { aDest, dest, uDest, vDest };
x86_reg counter[4]= { dstW, dstW, chrDstW, chrDstW };
const int16_t *src[4]= {
lumSrc + dstW, chrUSrc + chrDstW,
chrVSrc + chrDstW, alpSrc + dstW
};
x86_reg counter[4]= { dstW, chrDstW, chrDstW, dstW };
while (p--) {
if (dst[p]) {
......@@ -969,14 +975,16 @@ static void RENAME(yuv2yuyv422_X)(SwsContext *c, const int16_t *lumFilter,
/**
* vertical bilinear scale YV12 to RGB
*/
static void RENAME(yuv2rgb32_2)(SwsContext *c, const uint16_t *buf0,
const uint16_t *buf1, const uint16_t *ubuf0,
const uint16_t *ubuf1, const uint16_t *vbuf0,
const uint16_t *vbuf1, const uint16_t *abuf0,
const uint16_t *abuf1, uint8_t *dest,
static void RENAME(yuv2rgb32_2)(SwsContext *c, const int16_t *buf[2],
const int16_t *ubuf[2], const int16_t *vbuf[2],
const int16_t *abuf[2], uint8_t *dest,
int dstW, int yalpha, int uvalpha, int y)
{
const int16_t *buf0 = buf[0], *buf1 = buf[1],
*ubuf0 = ubuf[0], *ubuf1 = ubuf[1];
if (CONFIG_SWSCALE_ALPHA && c->alpPixBuf) {
const int16_t *abuf0 = abuf[0], *abuf1 = abuf[1];
#if ARCH_X86_64
__asm__ volatile(
YSCALEYUV2RGB(%%r8, %5)
......@@ -1031,13 +1039,14 @@ static void RENAME(yuv2rgb32_2)(SwsContext *c, const uint16_t *buf0,
}
}
static void RENAME(yuv2bgr24_2)(SwsContext *c, const uint16_t *buf0,
const uint16_t *buf1, const uint16_t *ubuf0,
const uint16_t *ubuf1, const uint16_t *vbuf0,
const uint16_t *vbuf1, const uint16_t *abuf0,
const uint16_t *abuf1, uint8_t *dest,
static void RENAME(yuv2bgr24_2)(SwsContext *c, const int16_t *buf[2],
const int16_t *ubuf[2], const int16_t *vbuf[2],
const int16_t *abuf[2], uint8_t *dest,
int dstW, int yalpha, int uvalpha, int y)
{
const int16_t *buf0 = buf[0], *buf1 = buf[1],
*ubuf0 = ubuf[0], *ubuf1 = ubuf[1];
//Note 8280 == DSTW_OFFSET but the preprocessor can't handle that there :(
__asm__ volatile(
"mov %%"REG_b", "ESP_OFFSET"(%5) \n\t"
......@@ -1053,13 +1062,14 @@ static void RENAME(yuv2bgr24_2)(SwsContext *c, const uint16_t *buf0,
);
}
static void RENAME(yuv2rgb555_2)(SwsContext *c, const uint16_t *buf0,
const uint16_t *buf1, const uint16_t *ubuf0,
const uint16_t *ubuf1, const uint16_t *vbuf0,
const uint16_t *vbuf1, const uint16_t *abuf0,
const uint16_t *abuf1, uint8_t *dest,
static void RENAME(yuv2rgb555_2)(SwsContext *c, const int16_t *buf[2],
const int16_t *ubuf[2], const int16_t *vbuf[2],
const int16_t *abuf[2], uint8_t *dest,
int dstW, int yalpha, int uvalpha, int y)
{
const int16_t *buf0 = buf[0], *buf1 = buf[1],
*ubuf0 = ubuf[0], *ubuf1 = ubuf[1];
//Note 8280 == DSTW_OFFSET but the preprocessor can't handle that there :(
__asm__ volatile(
"mov %%"REG_b", "ESP_OFFSET"(%5) \n\t"
......@@ -1081,13 +1091,14 @@ static void RENAME(yuv2rgb555_2)(SwsContext *c, const uint16_t *buf0,
);
}
static void RENAME(yuv2rgb565_2)(SwsContext *c, const uint16_t *buf0,
const uint16_t *buf1, const uint16_t *ubuf0,
const uint16_t *ubuf1, const uint16_t *vbuf0,
const uint16_t *vbuf1, const uint16_t *abuf0,
const uint16_t *abuf1, uint8_t *dest,
static void RENAME(yuv2rgb565_2)(SwsContext *c, const int16_t *buf[2],
const int16_t *ubuf[2], const int16_t *vbuf[2],
const int16_t *abuf[2], uint8_t *dest,
int dstW, int yalpha, int uvalpha, int y)
{
const int16_t *buf0 = buf[0], *buf1 = buf[1],
*ubuf0 = ubuf[0], *ubuf1 = ubuf[1];
//Note 8280 == DSTW_OFFSET but the preprocessor can't handle that there :(
__asm__ volatile(
"mov %%"REG_b", "ESP_OFFSET"(%5) \n\t"
......@@ -1149,13 +1160,14 @@ static void RENAME(yuv2rgb565_2)(SwsContext *c, const uint16_t *buf0,
#define YSCALEYUV2PACKED(index, c) REAL_YSCALEYUV2PACKED(index, c)
static void RENAME(yuv2yuyv422_2)(SwsContext *c, const uint16_t *buf0,
const uint16_t *buf1, const uint16_t *ubuf0,
const uint16_t *ubuf1, const uint16_t *vbuf0,
const uint16_t *vbuf1, const uint16_t *abuf0,
const uint16_t *abuf1, uint8_t *dest,
static void RENAME(yuv2yuyv422_2)(SwsContext *c, const int16_t *buf[2],
const int16_t *ubuf[2], const int16_t *vbuf[2],
const int16_t *abuf[2], uint8_t *dest,
int dstW, int yalpha, int uvalpha, int y)
{
const int16_t *buf0 = buf[0], *buf1 = buf[1],
*ubuf0 = ubuf[0], *ubuf1 = ubuf[1];
//Note 8280 == DSTW_OFFSET but the preprocessor can't handle that there :(
__asm__ volatile(
"mov %%"REG_b", "ESP_OFFSET"(%5) \n\t"
......@@ -1288,14 +1300,13 @@ static void RENAME(yuv2yuyv422_2)(SwsContext *c, const uint16_t *buf0,
/**
* YV12 to RGB without scaling or interpolating
*/
static void RENAME(yuv2rgb32_1)(SwsContext *c, const uint16_t *buf0,
const uint16_t *ubuf0, const uint16_t *ubuf1,
const uint16_t *vbuf0, const uint16_t *vbuf1,
const uint16_t *abuf0, uint8_t *dest,
int dstW, int uvalpha, enum PixelFormat dstFormat,
int flags, int y)
static void RENAME(yuv2rgb32_1)(SwsContext *c, const int16_t *buf0,
const int16_t *ubuf[2], const int16_t *bguf[2],
const int16_t *abuf0, uint8_t *dest,
int dstW, int uvalpha, int y)
{
const uint16_t *buf1= buf0; //FIXME needed for RGB1/BGR1
const int16_t *ubuf0 = ubuf[0], *ubuf1 = ubuf[1];
const int16_t *buf1= buf0; //FIXME needed for RGB1/BGR1
if (uvalpha < 2048) { // note this is not correct (shifts chrominance by 0.5 pixels) but it is a bit faster
if (CONFIG_SWSCALE_ALPHA && c->alpPixBuf) {
......@@ -1356,14 +1367,13 @@ static void RENAME(yuv2rgb32_1)(SwsContext *c, const uint16_t *buf0,
}
}
static void RENAME(yuv2bgr24_1)(SwsContext *c, const uint16_t *buf0,
const uint16_t *ubuf0, const uint16_t *ubuf1,
const uint16_t *vbuf0, const uint16_t *vbuf1,
const uint16_t *abuf0, uint8_t *dest,
int dstW, int uvalpha, enum PixelFormat dstFormat,
int flags, int y)
static void RENAME(yuv2bgr24_1)(SwsContext *c, const int16_t *buf0,
const int16_t *ubuf[2], const int16_t *bguf[2],
const int16_t *abuf0, uint8_t *dest,
int dstW, int uvalpha, int y)
{
const uint16_t *buf1= buf0; //FIXME needed for RGB1/BGR1
const int16_t *ubuf0 = ubuf[0], *ubuf1 = ubuf[1];
const int16_t *buf1= buf0; //FIXME needed for RGB1/BGR1
if (uvalpha < 2048) { // note this is not correct (shifts chrominance by 0.5 pixels) but it is a bit faster
__asm__ volatile(
......@@ -1394,14 +1404,13 @@ static void RENAME(yuv2bgr24_1)(SwsContext *c, const uint16_t *buf0,
}
}
static void RENAME(yuv2rgb555_1)(SwsContext *c, const uint16_t *buf0,
const uint16_t *ubuf0, const uint16_t *ubuf1,
const uint16_t *vbuf0, const uint16_t *vbuf1,
const uint16_t *abuf0, uint8_t *dest,
int dstW, int uvalpha, enum PixelFormat dstFormat,
int flags, int y)
static void RENAME(yuv2rgb555_1)(SwsContext *c, const int16_t *buf0,
const int16_t *ubuf[2], const int16_t *bguf[2],
const int16_t *abuf0, uint8_t *dest,
int dstW, int uvalpha, int y)
{
const uint16_t *buf1= buf0; //FIXME needed for RGB1/BGR1
const int16_t *ubuf0 = ubuf[0], *ubuf1 = ubuf[1];
const int16_t *buf1= buf0; //FIXME needed for RGB1/BGR1
if (uvalpha < 2048) { // note this is not correct (shifts chrominance by 0.5 pixels) but it is a bit faster
__asm__ volatile(
......@@ -1444,14 +1453,13 @@ static void RENAME(yuv2rgb555_1)(SwsContext *c, const uint16_t *buf0,
}
}
static void RENAME(yuv2rgb565_1)(SwsContext *c, const uint16_t *buf0,
const uint16_t *ubuf0, const uint16_t *ubuf1,
const uint16_t *vbuf0, const uint16_t *vbuf1,
const uint16_t *abuf0, uint8_t *dest,
int dstW, int uvalpha, enum PixelFormat dstFormat,
int flags, int y)
static void RENAME(yuv2rgb565_1)(SwsContext *c, const int16_t *buf0,
const int16_t *ubuf[2], const int16_t *bguf[2],
const int16_t *abuf0, uint8_t *dest,
int dstW, int uvalpha, int y)
{
const uint16_t *buf1= buf0; //FIXME needed for RGB1/BGR1
const int16_t *ubuf0 = ubuf[0], *ubuf1 = ubuf[1];
const int16_t *buf1= buf0; //FIXME needed for RGB1/BGR1
if (uvalpha < 2048) { // note this is not correct (shifts chrominance by 0.5 pixels) but it is a bit faster
__asm__ volatile(
......@@ -1531,14 +1539,13 @@ static void RENAME(yuv2rgb565_1)(SwsContext *c, const uint16_t *buf0,
"psraw $7, %%mm7 \n\t"
#define YSCALEYUV2PACKED1b(index, c) REAL_YSCALEYUV2PACKED1b(index, c)
static void RENAME(yuv2yuyv422_1)(SwsContext *c, const uint16_t *buf0,
const uint16_t *ubuf0, const uint16_t *ubuf1,
const uint16_t *vbuf0, const uint16_t *vbuf1,
const uint16_t *abuf0, uint8_t *dest,
int dstW, int uvalpha, enum PixelFormat dstFormat,
int flags, int y)
static void RENAME(yuv2yuyv422_1)(SwsContext *c, const int16_t *buf0,
const int16_t *ubuf[2], const int16_t *bguf[2],
const int16_t *abuf0, uint8_t *dest,
int dstW, int uvalpha, int y)
{
const uint16_t *buf1= buf0; //FIXME needed for RGB1/BGR1
const int16_t *ubuf0 = ubuf[0], *ubuf1 = ubuf[1];
const int16_t *buf1= buf0; //FIXME needed for RGB1/BGR1
if (uvalpha < 2048) { // note this is not correct (shifts chrominance by 0.5 pixels) but it is a bit faster
__asm__ volatile(
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment