Skip to content

F
ffmpeg.wasm-core
Commit a6b4448c authored by Luca Barbato's avatar Luca Barbato
Browse files
Options

Cleanup _t types in libavcodec/ppc 

Originally committed as revision 16357 to svn://svn.ffmpeg.org/ffmpeg/trunk
parent 7f9b3266
Expand all Show whitespace changes
Inline Side-by-side
Showing with 330 additions and 338 deletions
libavcodec/ppc/h264_altivec.c
View file @ a6b4448c
This diff is collapsed.
libavcodec/ppc/h264_template_altivec.c
View file @ a6b4448c
This diff is collapsed.
libavcodec/ppc/idct_altivec.c
View file @ a6b4448c
...@@ -40,17 +40,9 @@ ...@@ -40,17 +40,9 @@
#include "libavcodec/dsputil.h" #include "libavcodec/dsputil.h"
#include "gcc_fixes.h" #include "gcc_fixes.h"
#include "types_altivec.h"
#include "dsputil_ppc.h" #include "dsputil_ppc.h"
#define vector_s16_t vector signed short
#define const_vector_s16_t const vector signed short
#define vector_u16_t vector unsigned short
#define vector_s8_t vector signed char
#define vector_u8_t vector unsigned char
#define vector_s32_t vector signed int
#define vector_u32_t vector unsigned int
#define IDCT_HALF \ #define IDCT_HALF \
/* 1st stage */ \ /* 1st stage */ \
t1 = vec_mradds (a1, vx7, vx1 ); \ t1 = vec_mradds (a1, vx7, vx1 ); \
...@@ -88,11 +80,11 @@ ...@@ -88,11 +80,11 @@
#define IDCT \ #define IDCT \
vector_s16_t vx0, vx1, vx2, vx3, vx4, vx5, vx6, vx7; \ vec_s16 vx0, vx1, vx2, vx3, vx4, vx5, vx6, vx7; \
vector_s16_t vy0, vy1, vy2, vy3, vy4, vy5, vy6, vy7; \ vec_s16 vy0, vy1, vy2, vy3, vy4, vy5, vy6, vy7; \
vector_s16_t a0, a1, a2, ma2, c4, mc4, zero, bias; \ vec_s16 a0, a1, a2, ma2, c4, mc4, zero, bias; \
vector_s16_t t0, t1, t2, t3, t4, t5, t6, t7, t8; \ vec_s16 t0, t1, t2, t3, t4, t5, t6, t7, t8; \
vector_u16_t shift; \ vec_u16 shift; \
\ \
c4 = vec_splat (constants[0], 0); \ c4 = vec_splat (constants[0], 0); \
a0 = vec_splat (constants[0], 1); \ a0 = vec_splat (constants[0], 1); \
...@@ -100,7 +92,7 @@ ...@@ -100,7 +92,7 @@
a2 = vec_splat (constants[0], 3); \ a2 = vec_splat (constants[0], 3); \
mc4 = vec_splat (constants[0], 4); \ mc4 = vec_splat (constants[0], 4); \
ma2 = vec_splat (constants[0], 5); \ ma2 = vec_splat (constants[0], 5); \
bias = (vector_s16_t)vec_splat ((vector_s32_t)constants[0], 3); \ bias = (vec_s16)vec_splat ((vec_s32)constants[0], 3); \
\ \
zero = vec_splat_s16 (0); \ zero = vec_splat_s16 (0); \
shift = vec_splat_u16 (4); \ shift = vec_splat_u16 (4); \
...@@ -156,7 +148,7 @@ ...@@ -156,7 +148,7 @@
vx7 = vec_sra (vy7, shift); vx7 = vec_sra (vy7, shift);
static const_vector_s16_t constants[5] = { static const vec_s16 constants[5] = {
{23170, 13573, 6518, 21895, -23170, -21895, 32, 31}, {23170, 13573, 6518, 21895, -23170, -21895, 32, 31},
{16384, 22725, 21407, 19266, 16384, 19266, 21407, 22725}, {16384, 22725, 21407, 19266, 16384, 19266, 21407, 22725},
{22725, 31521, 29692, 26722, 22725, 26722, 29692, 31521}, {22725, 31521, 29692, 26722, 22725, 26722, 29692, 31521},
...@@ -164,10 +156,10 @@ static const_vector_s16_t constants[5] = { ...@@ -164,10 +156,10 @@ static const_vector_s16_t constants[5] = {
{19266, 26722, 25172, 22654, 19266, 22654, 25172, 26722} {19266, 26722, 25172, 22654, 19266, 22654, 25172, 26722}
}; };
void idct_put_altivec(uint8_t* dest, int stride, vector_s16_t* block) void idct_put_altivec(uint8_t* dest, int stride, vec_s16* block)
{ {
POWERPC_PERF_DECLARE(altivec_idct_put_num, 1); POWERPC_PERF_DECLARE(altivec_idct_put_num, 1);
vector_u8_t tmp; vec_u8 tmp;
#ifdef CONFIG_POWERPC_PERF #ifdef CONFIG_POWERPC_PERF
POWERPC_PERF_START_COUNT(altivec_idct_put_num, 1); POWERPC_PERF_START_COUNT(altivec_idct_put_num, 1);
...@@ -176,8 +168,8 @@ POWERPC_PERF_START_COUNT(altivec_idct_put_num, 1); ...@@ -176,8 +168,8 @@ POWERPC_PERF_START_COUNT(altivec_idct_put_num, 1);
#define COPY(dest,src) \ #define COPY(dest,src) \
tmp = vec_packsu (src, src); \ tmp = vec_packsu (src, src); \
vec_ste ((vector_u32_t)tmp, 0, (unsigned int *)dest); \ vec_ste ((vec_u32)tmp, 0, (unsigned int *)dest); \
vec_ste ((vector_u32_t)tmp, 4, (unsigned int *)dest); vec_ste ((vec_u32)tmp, 4, (unsigned int *)dest);
COPY (dest, vx0) dest += stride; COPY (dest, vx0) dest += stride;
COPY (dest, vx1) dest += stride; COPY (dest, vx1) dest += stride;
...@@ -191,14 +183,14 @@ POWERPC_PERF_START_COUNT(altivec_idct_put_num, 1); ...@@ -191,14 +183,14 @@ POWERPC_PERF_START_COUNT(altivec_idct_put_num, 1);
POWERPC_PERF_STOP_COUNT(altivec_idct_put_num, 1); POWERPC_PERF_STOP_COUNT(altivec_idct_put_num, 1);
} }
void idct_add_altivec(uint8_t* dest, int stride, vector_s16_t* block) void idct_add_altivec(uint8_t* dest, int stride, vec_s16* block)
{ {
POWERPC_PERF_DECLARE(altivec_idct_add_num, 1); POWERPC_PERF_DECLARE(altivec_idct_add_num, 1);
vector_u8_t tmp; vec_u8 tmp;
vector_s16_t tmp2, tmp3; vec_s16 tmp2, tmp3;
vector_u8_t perm0; vec_u8 perm0;
vector_u8_t perm1; vec_u8 perm1;
vector_u8_t p0, p1, p; vec_u8 p0, p1, p;
#ifdef CONFIG_POWERPC_PERF #ifdef CONFIG_POWERPC_PERF
POWERPC_PERF_START_COUNT(altivec_idct_add_num, 1); POWERPC_PERF_START_COUNT(altivec_idct_add_num, 1);
...@@ -215,11 +207,11 @@ POWERPC_PERF_START_COUNT(altivec_idct_add_num, 1); ...@@ -215,11 +207,11 @@ POWERPC_PERF_START_COUNT(altivec_idct_add_num, 1);
#define ADD(dest,src,perm) \ #define ADD(dest,src,perm) \
/* *(uint64_t *)&tmp = *(uint64_t *)dest; */ \ /* *(uint64_t *)&tmp = *(uint64_t *)dest; */ \
tmp = vec_ld (0, dest); \ tmp = vec_ld (0, dest); \
tmp2 = (vector_s16_t)vec_perm (tmp, (vector_u8_t)zero, perm); \ tmp2 = (vec_s16)vec_perm (tmp, (vec_u8)zero, perm); \
tmp3 = vec_adds (tmp2, src); \ tmp3 = vec_adds (tmp2, src); \
tmp = vec_packsu (tmp3, tmp3); \ tmp = vec_packsu (tmp3, tmp3); \
vec_ste ((vector_u32_t)tmp, 0, (unsigned int *)dest); \ vec_ste ((vec_u32)tmp, 0, (unsigned int *)dest); \
vec_ste ((vector_u32_t)tmp, 4, (unsigned int *)dest); vec_ste ((vec_u32)tmp, 4, (unsigned int *)dest);
ADD (dest, vx0, perm0) dest += stride; ADD (dest, vx0, perm0) dest += stride;
ADD (dest, vx1, perm1) dest += stride; ADD (dest, vx1, perm1) dest += stride;
......
libavcodec/ppc/int_altivec.c
View file @ a6b4448c
...@@ -79,10 +79,10 @@ static int ssd_int8_vs_int16_altivec(const int8_t *pix1, const int16_t *pix2, ...@@ -79,10 +79,10 @@ static int ssd_int8_vs_int16_altivec(const int8_t *pix1, const int16_t *pix2,
static void add_int16_altivec(int16_t * v1, int16_t * v2, int order) static void add_int16_altivec(int16_t * v1, int16_t * v2, int order)
{ {
int i; int i;
register vec_s16_t vec, *pv; register vec_s16 vec, *pv;
for(i = 0; i < order; i += 8){ for(i = 0; i < order; i += 8){
pv = (vec_s16_t*)v2; pv = (vec_s16*)v2;
vec = vec_perm(pv[0], pv[1], vec_lvsl(0, v2)); vec = vec_perm(pv[0], pv[1], vec_lvsl(0, v2));
vec_st(vec_add(vec_ld(0, v1), vec), 0, v1); vec_st(vec_add(vec_ld(0, v1), vec), 0, v1);
v1 += 8; v1 += 8;
...@@ -93,10 +93,10 @@ static void add_int16_altivec(int16_t * v1, int16_t * v2, int order) ...@@ -93,10 +93,10 @@ static void add_int16_altivec(int16_t * v1, int16_t * v2, int order)
static void sub_int16_altivec(int16_t * v1, int16_t * v2, int order) static void sub_int16_altivec(int16_t * v1, int16_t * v2, int order)
{ {
int i; int i;
register vec_s16_t vec, *pv; register vec_s16 vec, *pv;
for(i = 0; i < order; i += 8){ for(i = 0; i < order; i += 8){
pv = (vec_s16_t*)v2; pv = (vec_s16*)v2;
vec = vec_perm(pv[0], pv[1], vec_lvsl(0, v2)); vec = vec_perm(pv[0], pv[1], vec_lvsl(0, v2));
vec_st(vec_sub(vec_ld(0, v1), vec), 0, v1); vec_st(vec_sub(vec_ld(0, v1), vec), 0, v1);
v1 += 8; v1 += 8;
...@@ -108,9 +108,9 @@ static int32_t scalarproduct_int16_altivec(int16_t * v1, int16_t * v2, int order ...@@ -108,9 +108,9 @@ static int32_t scalarproduct_int16_altivec(int16_t * v1, int16_t * v2, int order
{ {
int i; int i;
LOAD_ZERO; LOAD_ZERO;
register vec_s16_t vec1, *pv; register vec_s16 vec1, *pv;
register vec_s32_t res = vec_splat_s32(0), t; register vec_s32 res = vec_splat_s32(0), t;
register vec_u32_t shifts; register vec_u32 shifts;
DECLARE_ALIGNED_16(int32_t, ires); DECLARE_ALIGNED_16(int32_t, ires);
shifts = zero_u32v; shifts = zero_u32v;
...@@ -121,7 +121,7 @@ static int32_t scalarproduct_int16_altivec(int16_t * v1, int16_t * v2, int order ...@@ -121,7 +121,7 @@ static int32_t scalarproduct_int16_altivec(int16_t * v1, int16_t * v2, int order
if(shift & 0x01) shifts = vec_add(shifts, vec_splat_u32(0x01)); if(shift & 0x01) shifts = vec_add(shifts, vec_splat_u32(0x01));
for(i = 0; i < order; i += 8){ for(i = 0; i < order; i += 8){
pv = (vec_s16_t*)v1; pv = (vec_s16*)v1;
vec1 = vec_perm(pv[0], pv[1], vec_lvsl(0, v1)); vec1 = vec_perm(pv[0], pv[1], vec_lvsl(0, v1));
t = vec_msum(vec1, vec_ld(0, v2), zero_s32v); t = vec_msum(vec1, vec_ld(0, v2), zero_s32v);
t = vec_sr(t, shifts); t = vec_sr(t, shifts);
......
libavcodec/ppc/types_altivec.h
View file @ a6b4448c
...@@ -24,23 +24,23 @@ ...@@ -24,23 +24,23 @@
/*********************************************************************** /***********************************************************************
* Vector types * Vector types
**********************************************************************/ **********************************************************************/
#define vec_u8_t vector unsigned char #define vec_u8 vector unsigned char
#define vec_s8_t vector signed char #define vec_s8 vector signed char
#define vec_u16_t vector unsigned short #define vec_u16 vector unsigned short
#define vec_s16_t vector signed short #define vec_s16 vector signed short
#define vec_u32_t vector unsigned int #define vec_u32 vector unsigned int
#define vec_s32_t vector signed int #define vec_s32 vector signed int
/*********************************************************************** /***********************************************************************
* Null vector * Null vector
**********************************************************************/ **********************************************************************/
#define LOAD_ZERO const vec_u8_t zerov = vec_splat_u8( 0 ) #define LOAD_ZERO const vec_u8 zerov = vec_splat_u8( 0 )
#define zero_u8v (vec_u8_t) zerov #define zero_u8v (vec_u8) zerov
#define zero_s8v (vec_s8_t) zerov #define zero_s8v (vec_s8) zerov
#define zero_u16v (vec_u16_t) zerov #define zero_u16v (vec_u16) zerov
#define zero_s16v (vec_s16_t) zerov #define zero_s16v (vec_s16) zerov
#define zero_u32v (vec_u32_t) zerov #define zero_u32v (vec_u32) zerov
#define zero_s32v (vec_s32_t) zerov #define zero_s32v (vec_s32) zerov
#endif /* AVCODEC_PPC_TYPES_ALTIVEC_H */ #endif /* AVCODEC_PPC_TYPES_ALTIVEC_H */
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