ref: 37225744dbf30d79711fa9ef182d2007a51b11bd
parent: 47dee375db30e803a9ff8c0694040b1eb819add7
author: James Zern <jzern@google.com>
date: Thu Feb 11 14:43:36 EST 2016
vpx_lpf_vertical_8: remove unused count param Change-Id: Ic69406da00afb0f06588e8c0deb2b043952b078c
--- a/test/lpf_8_test.cc
+++ b/test/lpf_8_test.cc
@@ -522,7 +522,8 @@
make_tuple(&vpx_lpf_horizontal_8_sse2, &vpx_lpf_horizontal_8_c, 8, 1),
make_tuple(&vpx_lpf_horizontal_16_sse2, &vpx_lpf_horizontal_16_c, 8, 1),
make_tuple(&vpx_lpf_horizontal_16_sse2, &vpx_lpf_horizontal_16_c, 8, 2),
- make_tuple(&vpx_lpf_vertical_8_sse2, &vpx_lpf_vertical_8_c, 8, 1),
+ make_tuple(&wrapper_nc<vpx_lpf_vertical_8_sse2>,
+ &wrapper_nc<vpx_lpf_vertical_8_c>, 8, 1),
make_tuple(&wrapper_nc<vpx_lpf_vertical_16_sse2>,
&wrapper_nc<vpx_lpf_vertical_16_c>, 8, 1),
make_tuple(&wrapper_nc<vpx_lpf_vertical_16_dual_sse2>,
@@ -604,8 +605,8 @@
#endif // HAVE_NEON_ASM
make_tuple(&vpx_lpf_horizontal_8_neon,
&vpx_lpf_horizontal_8_c, 8, 1),
- make_tuple(&vpx_lpf_vertical_8_neon,
- &vpx_lpf_vertical_8_c, 8, 1),
+ make_tuple(&wrapper_nc<vpx_lpf_vertical_8_neon>,
+ &wrapper_nc<vpx_lpf_vertical_8_c>, 8, 1),
make_tuple(&vpx_lpf_horizontal_4_neon,
&vpx_lpf_horizontal_4_c, 8, 1),
make_tuple(&vpx_lpf_vertical_4_neon,
@@ -637,7 +638,8 @@
make_tuple(&vpx_lpf_horizontal_16_dspr2,
&vpx_lpf_horizontal_16_c, 8, 2),
make_tuple(&vpx_lpf_vertical_4_dspr2, &vpx_lpf_vertical_4_c, 8, 1),
- make_tuple(&vpx_lpf_vertical_8_dspr2, &vpx_lpf_vertical_8_c, 8, 1),
+ make_tuple(&wrapper_nc<vpx_lpf_vertical_8_dspr2>,
+ &wrapper_nc<vpx_lpf_vertical_8_c>, 8, 1),
make_tuple(&wrapper_nc<vpx_lpf_vertical_16_dspr2>,
&wrapper_nc<vpx_lpf_vertical_16_c>, 8, 1),
make_tuple(&wrapper_nc<vpx_lpf_vertical_16_dual_dspr2>,
@@ -665,7 +667,8 @@
make_tuple(&vpx_lpf_horizontal_16_msa, &vpx_lpf_horizontal_16_c, 8, 1),
make_tuple(&vpx_lpf_horizontal_16_msa, &vpx_lpf_horizontal_16_c, 8, 2),
make_tuple(&vpx_lpf_vertical_4_msa, &vpx_lpf_vertical_4_c, 8, 1),
- make_tuple(&vpx_lpf_vertical_8_msa, &vpx_lpf_vertical_8_c, 8, 1),
+ make_tuple(&wrapper_nc<vpx_lpf_vertical_8_msa>,
+ &wrapper_nc<vpx_lpf_vertical_8_c>, 8, 1),
make_tuple(&wrapper_nc<vpx_lpf_vertical_16_msa>,
&wrapper_nc<vpx_lpf_vertical_16_c>, 8, 1)));
--- a/vp10/common/loopfilter.c
+++ b/vp10/common/loopfilter.c
@@ -345,11 +345,10 @@
lfi0->hev_thr, lfi1->mblim, lfi1->lim,
lfi1->hev_thr);
} else if (mask_8x8_0 & 1) {- vpx_lpf_vertical_8(s, pitch, lfi0->mblim, lfi0->lim, lfi0->hev_thr,
- 1);
+ vpx_lpf_vertical_8(s, pitch, lfi0->mblim, lfi0->lim, lfi0->hev_thr);
} else {vpx_lpf_vertical_8(s + 8 * pitch, pitch, lfi1->mblim, lfi1->lim,
- lfi1->hev_thr, 1);
+ lfi1->hev_thr);
}
}
@@ -1127,7 +1126,7 @@
if (mask_16x16 & 1) {vpx_lpf_vertical_16(s, pitch, lfi->mblim, lfi->lim, lfi->hev_thr);
} else if (mask_8x8 & 1) {- vpx_lpf_vertical_8(s, pitch, lfi->mblim, lfi->lim, lfi->hev_thr, 1);
+ vpx_lpf_vertical_8(s, pitch, lfi->mblim, lfi->lim, lfi->hev_thr);
} else if (mask_4x4 & 1) {vpx_lpf_vertical_4(s, pitch, lfi->mblim, lfi->lim, lfi->hev_thr, 1);
}
--- a/vp9/common/vp9_loopfilter.c
+++ b/vp9/common/vp9_loopfilter.c
@@ -345,11 +345,10 @@
lfi0->hev_thr, lfi1->mblim, lfi1->lim,
lfi1->hev_thr);
} else if (mask_8x8_0 & 1) {- vpx_lpf_vertical_8(s, pitch, lfi0->mblim, lfi0->lim, lfi0->hev_thr,
- 1);
+ vpx_lpf_vertical_8(s, pitch, lfi0->mblim, lfi0->lim, lfi0->hev_thr);
} else {vpx_lpf_vertical_8(s + 8 * pitch, pitch, lfi1->mblim, lfi1->lim,
- lfi1->hev_thr, 1);
+ lfi1->hev_thr);
}
}
@@ -1102,7 +1101,7 @@
if (mask_16x16 & 1) {vpx_lpf_vertical_16(s, pitch, lfi->mblim, lfi->lim, lfi->hev_thr);
} else if (mask_8x8 & 1) {- vpx_lpf_vertical_8(s, pitch, lfi->mblim, lfi->lim, lfi->hev_thr, 1);
+ vpx_lpf_vertical_8(s, pitch, lfi->mblim, lfi->lim, lfi->hev_thr);
} else if (mask_4x4 & 1) {vpx_lpf_vertical_4(s, pitch, lfi->mblim, lfi->lim, lfi->hev_thr, 1);
}
--- a/vpx_dsp/arm/loopfilter_8_neon.asm
+++ b/vpx_dsp/arm/loopfilter_8_neon.asm
@@ -82,8 +82,7 @@
; int pitch,
; const uint8_t *blimit,
; const uint8_t *limit,
-; const uint8_t *thresh,
-; int count)
+; const uint8_t *thresh)
;
; r0 uint8_t *s,
; r1 int pitch,
@@ -90,22 +89,17 @@
; r2 const uint8_t *blimit,
; r3 const uint8_t *limit,
; sp const uint8_t *thresh,
-; sp+4 int count
|vpx_lpf_vertical_8_neon| PROC
push {r4-r5, lr} vld1.8 {d0[]}, [r2] ; duplicate *blimit- ldr r12, [sp, #16] ; load count
vld1.8 {d1[]}, [r3] ; duplicate *limitldr r3, [sp, #12] ; load thresh
sub r2, r0, #4 ; move s pointer down by 4 columns
- cmp r12, #0
- beq end_vpx_mblf_v_edge
vld1.8 {d2[]}, [r3] ; duplicate *thresh-count_mblf_v_loop
vld1.u8 {d3}, [r2], r1 ; load s data vld1.u8 {d4}, [r2], r1 vld1.u8 {d5}, [r2], r1@@ -156,12 +150,6 @@
vst2.8 {d4[6], d5[6]}, [r3], r1 vst2.8 {d4[7], d5[7]}, [r3]- add r0, r0, r1, lsl #3 ; s += pitch * 8
- subs r12, r12, #1
- subne r2, r0, #4 ; move s pointer down by 4 columns
- bne count_mblf_v_loop
-
-end_vpx_mblf_v_edge
pop {r4-r5, pc}ENDP ; |vpx_lpf_vertical_8_neon|
--- a/vpx_dsp/arm/loopfilter_8_neon.c
+++ b/vpx_dsp/arm/loopfilter_8_neon.c
@@ -328,8 +328,7 @@
int pitch,
const uint8_t *blimit,
const uint8_t *limit,
- const uint8_t *thresh,
- int count) {+ const uint8_t *thresh) {int i;
uint8_t *s;
uint8x8_t dblimit, dlimit, dthresh;
@@ -341,14 +340,11 @@
uint8x8x4_t d4Result;
uint8x8x2_t d2Result;
- if (count == 0)
- return;
-
dblimit = vld1_u8(blimit);
dlimit = vld1_u8(limit);
dthresh = vld1_u8(thresh);
- for (i = 0; i < count; i++) {+ for (i = 0; i < 1; i++) {s = src + (i * (pitch << 3)) - 4;
d3u8 = vld1_u8(s);
--- a/vpx_dsp/arm/loopfilter_neon.c
+++ b/vpx_dsp/arm/loopfilter_neon.c
@@ -44,8 +44,8 @@
const uint8_t *blimit1,
const uint8_t *limit1,
const uint8_t *thresh1) {- vpx_lpf_vertical_8_neon(s, p, blimit0, limit0, thresh0, 1);
- vpx_lpf_vertical_8_neon(s + 8 * p, p, blimit1, limit1, thresh1, 1);
+ vpx_lpf_vertical_8_neon(s, p, blimit0, limit0, thresh0);
+ vpx_lpf_vertical_8_neon(s + 8 * p, p, blimit1, limit1, thresh1);
}
void vpx_lpf_vertical_16_dual_neon(uint8_t *s, int p,
--- a/vpx_dsp/loopfilter.c
+++ b/vpx_dsp/loopfilter.c
@@ -218,11 +218,10 @@
}
void vpx_lpf_vertical_8_c(uint8_t *s, int pitch, const uint8_t *blimit,
- const uint8_t *limit, const uint8_t *thresh,
- int count) {+ const uint8_t *limit, const uint8_t *thresh) {int i;
- for (i = 0; i < 8 * count; ++i) {+ for (i = 0; i < 8; ++i) {const uint8_t p3 = s[-4], p2 = s[-3], p1 = s[-2], p0 = s[-1];
const uint8_t q0 = s[0], q1 = s[1], q2 = s[2], q3 = s[3];
const int8_t mask = filter_mask(*limit, *blimit,
@@ -238,9 +237,8 @@
const uint8_t *limit0, const uint8_t *thresh0,
const uint8_t *blimit1, const uint8_t *limit1,
const uint8_t *thresh1) {- vpx_lpf_vertical_8_c(s, pitch, blimit0, limit0, thresh0, 1);
- vpx_lpf_vertical_8_c(s + 8 * pitch, pitch, blimit1, limit1,
- thresh1, 1);
+ vpx_lpf_vertical_8_c(s, pitch, blimit0, limit0, thresh0);
+ vpx_lpf_vertical_8_c(s + 8 * pitch, pitch, blimit1, limit1, thresh1);
}
static INLINE void filter16(int8_t mask, uint8_t thresh,
--- a/vpx_dsp/mips/loopfilter_8_msa.c
+++ b/vpx_dsp/mips/loopfilter_8_msa.c
@@ -161,8 +161,7 @@
void vpx_lpf_vertical_8_msa(uint8_t *src, int32_t pitch,
const uint8_t *b_limit_ptr,
const uint8_t *limit_ptr,
- const uint8_t *thresh_ptr,
- int32_t count) {+ const uint8_t *thresh_ptr) {v16u8 p3, p2, p1, p0, q3, q2, q1, q0;
v16u8 p1_out, p0_out, q0_out, q1_out;
v16u8 flat, mask, hev, thresh, b_limit, limit;
@@ -170,8 +169,6 @@
v8i16 p2_filt8_r, p1_filt8_r, p0_filt8_r, q0_filt8_r, q1_filt8_r, q2_filt8_r;
v16u8 zero = { 0 };v8i16 vec0, vec1, vec2, vec3, vec4;
-
- (void)count;
/* load vector elements */
LD_UB8(src - 4, pitch, p3, p2, p1, p0, q0, q1, q2, q3);
--- a/vpx_dsp/mips/loopfilter_filters_dspr2.c
+++ b/vpx_dsp/mips/loopfilter_filters_dspr2.c
@@ -346,9 +346,8 @@
const uint8_t *blimit1,
const uint8_t *limit1,
const uint8_t *thresh1) {- vpx_lpf_vertical_8_dspr2(s, p, blimit0, limit0, thresh0, 1);
- vpx_lpf_vertical_8_dspr2(s + 8 * p, p, blimit1, limit1, thresh1,
- 1);
+ vpx_lpf_vertical_8_dspr2(s, p, blimit0, limit0, thresh0);
+ vpx_lpf_vertical_8_dspr2(s + 8 * p, p, blimit1, limit1, thresh1);
}
void vpx_lpf_vertical_16_dual_dspr2(uint8_t *s, int p,
--- a/vpx_dsp/mips/loopfilter_mb_dspr2.c
+++ b/vpx_dsp/mips/loopfilter_mb_dspr2.c
@@ -322,8 +322,7 @@
int pitch,
const uint8_t *blimit,
const uint8_t *limit,
- const uint8_t *thresh,
- int count) {+ const uint8_t *thresh) {uint8_t i;
uint32_t mask, hev, flat;
uint8_t *s1, *s2, *s3, *s4;
--- a/vpx_dsp/vpx_dsp_rtcd_defs.pl
+++ b/vpx_dsp/vpx_dsp_rtcd_defs.pl
@@ -535,7 +535,7 @@
specialize qw/vpx_lpf_vertical_16_dual sse2 neon_asm dspr2 msa/;
$vpx_lpf_vertical_16_dual_neon_asm=vpx_lpf_vertical_16_dual_neon;
-add_proto qw/void vpx_lpf_vertical_8/, "uint8_t *s, int pitch, const uint8_t *blimit, const uint8_t *limit, const uint8_t *thresh, int count";
+add_proto qw/void vpx_lpf_vertical_8/, "uint8_t *s, int pitch, const uint8_t *blimit, const uint8_t *limit, const uint8_t *thresh";
specialize qw/vpx_lpf_vertical_8 sse2 neon dspr2 msa/;
add_proto qw/void vpx_lpf_vertical_8_dual/, "uint8_t *s, int pitch, const uint8_t *blimit0, const uint8_t *limit0, const uint8_t *thresh0, const uint8_t *blimit1, const uint8_t *limit1, const uint8_t *thresh1";
--- a/vpx_dsp/x86/loopfilter_sse2.c
+++ b/vpx_dsp/x86/loopfilter_sse2.c
@@ -1492,11 +1492,10 @@
void vpx_lpf_vertical_8_sse2(unsigned char *s, int p,
const unsigned char *blimit,
const unsigned char *limit,
- const unsigned char *thresh, int count) {+ const unsigned char *thresh) {DECLARE_ALIGNED(8, unsigned char, t_dst[8 * 8]);
unsigned char *src[1];
unsigned char *dst[1];
- (void)count;
// Transpose 8x8
src[0] = s - 4;
--
⑨