generic-library/vpx
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

Merge changes from topic 'missing-proto'

Browse Source 
* changes:
  fwd_txfm_msa.c: add missing vpx_dsp_rtcd.h
  vpx_convolve_*_msa.c: add missing vpx_dsp_rtcd.h
  loopfilter_*_msa.c: add missing vpx_dsp_rtcd.h
This commit is contained in:
James Zern 2017-07-05 20:00:25 +00:00 committed by Gerrit Code Review
commit a6531cbc54
6 changed files with 53 additions and 44 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

29
vpx_dsp/mips/fwd_txfm_msa.c
View File

@ -8,8 +8,23 @@
* be found in the AUTHORS file in the root of the source tree.
*/
#include "./vpx_dsp_rtcd.h"
#include "vpx_dsp/mips/fwd_txfm_msa.h"
void vpx_fdct8x8_1_msa(const int16_t *input, tran_low_t *out, int32_t stride) {
v8i16 in0, in1, in2, in3, in4, in5, in6, in7;
v4i32 vec_w;
LD_SH8(input, stride, in0, in1, in2, in3, in4, in5, in6, in7);
ADD4(in0, in1, in2, in3, in4, in5, in6, in7, in0, in2, in4, in6);
ADD2(in0, in2, in4, in6, in0, in4);
vec_w = __msa_hadd_s_w(in0, in0);
vec_w += __msa_hadd_s_w(in4, in4);
out[0] = HADD_SW_S32(vec_w);
out[1] = 0;
}
#if !CONFIG_VP9_HIGHBITDEPTH
void fdct8x16_1d_column(const int16_t *input, int16_t *tmp_ptr,
int32_t src_stride) {
v8i16 tmp0, tmp1, tmp2, tmp3, tmp4, tmp5, tmp6, tmp7;
@ -215,19 +230,6 @@ void vpx_fdct8x8_msa(const int16_t *input, int16_t *output,
ST_SH8(in0, in1, in2, in3, in4, in5, in6, in7, output, 8);
}
void vpx_fdct8x8_1_msa(const int16_t *input, tran_low_t *out, int32_t stride) {
v8i16 in0, in1, in2, in3, in4, in5, in6, in7;
v4i32 vec_w;
LD_SH8(input, stride, in0, in1, in2, in3, in4, in5, in6, in7);
ADD4(in0, in1, in2, in3, in4, in5, in6, in7, in0, in2, in4, in6);
ADD2(in0, in2, in4, in6, in0, in4);
vec_w = __msa_hadd_s_w(in0, in0);
vec_w += __msa_hadd_s_w(in4, in4);
out[0] = HADD_SW_S32(vec_w);
out[1] = 0;
}
void vpx_fdct16x16_msa(const int16_t *input, int16_t *output,
int32_t src_stride) {
int32_t i;
@ -267,3 +269,4 @@ void vpx_fdct16x16_1_msa(const int16_t *input, int16_t *out, int32_t stride) {
sum = HADD_SW_S32(vec_w);
out[0] = (int16_t)(sum >> 1);
}
#endif // !CONFIG_VP9_HIGHBITDEPTH

64
vpx_dsp/mips/loopfilter_16_msa.c
View File

@ -8,13 +8,15 @@
* be found in the AUTHORS file in the root of the source tree.
*/
#include "vpx_ports/mem.h"
#include "./vpx_dsp_rtcd.h"
#include "vpx_dsp/mips/loopfilter_msa.h"
#include "vpx_ports/mem.h"
int32_t vpx_hz_lpf_t4_and_t8_16w(uint8_t *src, int32_t pitch, uint8_t *filter48,
const uint8_t *b_limit_ptr,
const uint8_t *limit_ptr,
const uint8_t *thresh_ptr) {
static int32_t hz_lpf_t4_and_t8_16w(uint8_t *src, int32_t pitch,
uint8_t *filter48,
const uint8_t *b_limit_ptr,
const uint8_t *limit_ptr,
const uint8_t *thresh_ptr) {
v16u8 p3, p2, p1, p0, q3, q2, q1, q0;
v16u8 p2_out, p1_out, p0_out, q0_out, q1_out, q2_out;
v16u8 flat, mask, hev, thresh, b_limit, limit;
@ -77,7 +79,7 @@ int32_t vpx_hz_lpf_t4_and_t8_16w(uint8_t *src, int32_t pitch, uint8_t *filter48,
}
}
void vpx_hz_lpf_t16_16w(uint8_t *src, int32_t pitch, uint8_t *filter48) {
static void hz_lpf_t16_16w(uint8_t *src, int32_t pitch, uint8_t *filter48) {
v16u8 flat, flat2, filter8;
v16i8 zero = { 0 };
v16u8 p7, p6, p5, p4, p3, p2, p1, p0, q0, q1, q2, q3, q4, q5, q6, q7;
@ -413,11 +415,11 @@ static void mb_lpf_horizontal_edge_dual(uint8_t *src, int32_t pitch,
(void)count;
early_exit = vpx_hz_lpf_t4_and_t8_16w(src, pitch, &filter48[0], b_limit_ptr,
limit_ptr, thresh_ptr);
early_exit = hz_lpf_t4_and_t8_16w(src, pitch, &filter48[0], b_limit_ptr,
limit_ptr, thresh_ptr);
if (0 == early_exit) {
vpx_hz_lpf_t16_16w(src, pitch, filter48);
hz_lpf_t16_16w(src, pitch, filter48);
}
}
@ -753,11 +755,11 @@ static void transpose_16x16(uint8_t *input, int32_t in_pitch, uint8_t *output,
ST_UB8(q0, q1, q2, q3, q4, q5, q6, q7, output, out_pitch);
}
int32_t vpx_vt_lpf_t4_and_t8_8w(uint8_t *src, uint8_t *filter48,
uint8_t *src_org, int32_t pitch_org,
const uint8_t *b_limit_ptr,
const uint8_t *limit_ptr,
const uint8_t *thresh_ptr) {
static int32_t vt_lpf_t4_and_t8_8w(uint8_t *src, uint8_t *filter48,
uint8_t *src_org, int32_t pitch_org,
const uint8_t *b_limit_ptr,
const uint8_t *limit_ptr,
const uint8_t *thresh_ptr) {
v16u8 p3, p2, p1, p0, q3, q2, q1, q0;
v16u8 p2_out, p1_out, p0_out, q0_out, q1_out, q2_out;
v16u8 flat, mask, hev, thresh, b_limit, limit;
@ -820,8 +822,8 @@ int32_t vpx_vt_lpf_t4_and_t8_8w(uint8_t *src, uint8_t *filter48,
}
}
int32_t vpx_vt_lpf_t16_8w(uint8_t *src, uint8_t *src_org, int32_t pitch,
uint8_t *filter48) {
static int32_t vt_lpf_t16_8w(uint8_t *src, uint8_t *src_org, int32_t pitch,
uint8_t *filter48) {
v16i8 zero = { 0 };
v16u8 filter8, flat, flat2;
v16u8 p7, p6, p5, p4, p3, p2, p1, p0, q0, q1, q2, q3, q4, q5, q6, q7;
@ -1051,12 +1053,12 @@ void vpx_lpf_vertical_16_msa(uint8_t *src, int32_t pitch,
transpose_16x8_to_8x16(src - 8, pitch, transposed_input, 16);
early_exit =
vpx_vt_lpf_t4_and_t8_8w((transposed_input + 16 * 8), &filter48[0], src,
pitch, b_limit_ptr, limit_ptr, thresh_ptr);
vt_lpf_t4_and_t8_8w((transposed_input + 16 * 8), &filter48[0], src, pitch,
b_limit_ptr, limit_ptr, thresh_ptr);
if (0 == early_exit) {
early_exit = vpx_vt_lpf_t16_8w((transposed_input + 16 * 8), src, pitch,
&filter48[0]);
early_exit =
vt_lpf_t16_8w((transposed_input + 16 * 8), src, pitch, &filter48[0]);
if (0 == early_exit) {
transpose_8x16_to_16x8(transposed_input, 16, src - 8, pitch);
@ -1064,11 +1066,11 @@ void vpx_lpf_vertical_16_msa(uint8_t *src, int32_t pitch,
}
}
int32_t vpx_vt_lpf_t4_and_t8_16w(uint8_t *src, uint8_t *filter48,
uint8_t *src_org, int32_t pitch,
const uint8_t *b_limit_ptr,
const uint8_t *limit_ptr,
const uint8_t *thresh_ptr) {
static int32_t vt_lpf_t4_and_t8_16w(uint8_t *src, uint8_t *filter48,
uint8_t *src_org, int32_t pitch,
const uint8_t *b_limit_ptr,
const uint8_t *limit_ptr,
const uint8_t *thresh_ptr) {
v16u8 p3, p2, p1, p0, q3, q2, q1, q0;
v16u8 p2_out, p1_out, p0_out, q0_out, q1_out, q2_out;
v16u8 flat, mask, hev, thresh, b_limit, limit;
@ -1141,8 +1143,8 @@ int32_t vpx_vt_lpf_t4_and_t8_16w(uint8_t *src, uint8_t *filter48,
}
}
int32_t vpx_vt_lpf_t16_16w(uint8_t *src, uint8_t *src_org, int32_t pitch,
uint8_t *filter48) {
static int32_t vt_lpf_t16_16w(uint8_t *src, uint8_t *src_org, int32_t pitch,
uint8_t *filter48) {
v16u8 flat, flat2, filter8;
v16i8 zero = { 0 };
v16u8 p7, p6, p5, p4, p3, p2, p1, p0, q0, q1, q2, q3, q4, q5, q6, q7;
@ -1473,12 +1475,12 @@ void vpx_lpf_vertical_16_dual_msa(uint8_t *src, int32_t pitch,
transpose_16x16((src - 8), pitch, &transposed_input[0], 16);
early_exit =
vpx_vt_lpf_t4_and_t8_16w((transposed_input + 16 * 8), &filter48[0], src,
pitch, b_limit_ptr, limit_ptr, thresh_ptr);
vt_lpf_t4_and_t8_16w((transposed_input + 16 * 8), &filter48[0], src,
pitch, b_limit_ptr, limit_ptr, thresh_ptr);
if (0 == early_exit) {
early_exit = vpx_vt_lpf_t16_16w((transposed_input + 16 * 8), src, pitch,
&filter48[0]);
early_exit =
vt_lpf_t16_16w((transposed_input + 16 * 8), src, pitch, &filter48[0]);
if (0 == early_exit) {
transpose_16x16(transposed_input, 16, (src - 8), pitch);

1
vpx_dsp/mips/loopfilter_4_msa.c
View File

@ -8,6 +8,7 @@
* be found in the AUTHORS file in the root of the source tree.
*/
#include "./vpx_dsp_rtcd.h"
#include "vpx_dsp/mips/loopfilter_msa.h"
void vpx_lpf_horizontal_4_msa(uint8_t *src, int32_t pitch,

1
vpx_dsp/mips/loopfilter_8_msa.c
View File

@ -8,6 +8,7 @@
* be found in the AUTHORS file in the root of the source tree.
*/
#include "./vpx_dsp_rtcd.h"
#include "vpx_dsp/mips/loopfilter_msa.h"
void vpx_lpf_horizontal_8_msa(uint8_t *src, int32_t pitch,

1
vpx_dsp/mips/vpx_convolve_avg_msa.c
View File

@ -8,6 +8,7 @@
* be found in the AUTHORS file in the root of the source tree.
*/
#include "./vpx_dsp_rtcd.h"
#include "vpx_dsp/mips/macros_msa.h"
static void avg_width4_msa(const uint8_t *src, int32_t src_stride, uint8_t *dst,

1
vpx_dsp/mips/vpx_convolve_copy_msa.c
View File

@ -9,6 +9,7 @@
*/
#include <string.h>
#include "./vpx_dsp_rtcd.h"
#include "vpx_dsp/mips/macros_msa.h"
static void copy_width8_msa(const uint8_t *src, int32_t src_stride,