diff options
author | Martin Storsjö <martin@martin.st> | 2020-02-10 11:02:53 +0300 |
---|---|---|
committer | Martin Storsjö <martin@martin.st> | 2020-02-11 11:43:43 +0300 |
commit | 32e265a86e535b5fad47bcac9b54f83e1e5eab33 (patch) | |
tree | bd352f335abeef50306704186301fee610df94d0 /src/arm/64 | |
parent | 96da9cc232b4b238bcd215025d075039a5b0a86f (diff) |
arm: looprestoration: Add 8bpc to existing function names, add HIGHBD_*_SUFFIX
Don't add it to dav1d_sgr_calc_ab1/2_neon and box3/5_v, as the same
concrete function implementations can be shared for both 8 and 16 bpc
for those functions.
Diffstat (limited to 'src/arm/64')
-rw-r--r-- | src/arm/64/looprestoration.S | 66 |
1 files changed, 33 insertions, 33 deletions
diff --git a/src/arm/64/looprestoration.S b/src/arm/64/looprestoration.S index af0ec1b..8a6bfab 100644 --- a/src/arm/64/looprestoration.S +++ b/src/arm/64/looprestoration.S @@ -617,12 +617,12 @@ endfunc #define SUM_STRIDE (384+16) -// void dav1d_sgr_box3_h_neon(int32_t *sumsq, int16_t *sum, -// const pixel (*left)[4], -// const pixel *src, const ptrdiff_t stride, -// const int w, const int h, -// const enum LrEdgeFlags edges); -function sgr_box3_h_neon, export=1 +// void dav1d_sgr_box3_h_8bpc_neon(int32_t *sumsq, int16_t *sum, +// const pixel (*left)[4], +// const pixel *src, const ptrdiff_t stride, +// const int w, const int h, +// const enum LrEdgeFlags edges); +function sgr_box3_h_8bpc_neon, export=1 add w5, w5, #2 // w += 2 // Set up pointers for reading/writing alternate rows @@ -879,12 +879,12 @@ L(box3_variable_shift_tbl): .purgem add3 endfunc -// void dav1d_sgr_box5_h_neon(int32_t *sumsq, int16_t *sum, -// const pixel (*left)[4], -// const pixel *src, const ptrdiff_t stride, -// const int w, const int h, -// const enum LrEdgeFlags edges); -function sgr_box5_h_neon, export=1 +// void dav1d_sgr_box5_h_8bpc_neon(int32_t *sumsq, int16_t *sum, +// const pixel (*left)[4], +// const pixel *src, const ptrdiff_t stride, +// const int w, const int h, +// const enum LrEdgeFlags edges); +function sgr_box5_h_8bpc_neon, export=1 add w5, w5, #2 // w += 2 // Set up pointers for reading/writing alternate rows @@ -1540,11 +1540,11 @@ endfunc #define FILTER_OUT_STRIDE 384 -// void dav1d_sgr_finish_filter1_neon(int16_t *tmp, -// const pixel *src, const ptrdiff_t stride, -// const int32_t *a, const int16_t *b, -// const int w, const int h); -function sgr_finish_filter1_neon, export=1 +// void dav1d_sgr_finish_filter1_8bpc_neon(int16_t *tmp, +// const pixel *src, const ptrdiff_t stride, +// const int32_t *a, const int16_t *b, +// const int w, const int h); +function sgr_finish_filter1_8bpc_neon, export=1 sub x7, x3, #(4*SUM_STRIDE) add x8, x3, #(4*SUM_STRIDE) sub x9, x4, #(2*SUM_STRIDE) @@ -1657,11 +1657,11 @@ function sgr_finish_filter1_neon, export=1 ret endfunc -// void dav1d_sgr_finish_filter2_neon(int16_t *tmp, -// const pixel *src, const ptrdiff_t stride, -// const int32_t *a, const int16_t *b, -// const int w, const int h); -function sgr_finish_filter2_neon, export=1 +// void dav1d_sgr_finish_filter2_8bpc_neon(int16_t *tmp, +// const pixel *src, const ptrdiff_t stride, +// const int32_t *a, const int16_t *b, +// const int w, const int h); +function sgr_finish_filter2_8bpc_neon, export=1 add x7, x3, #(4*(SUM_STRIDE)) sub x3, x3, #(4*(SUM_STRIDE)) add x8, x4, #(2*(SUM_STRIDE)) @@ -1807,11 +1807,11 @@ function sgr_finish_filter2_neon, export=1 ret endfunc -// void dav1d_sgr_weighted1_neon(pixel *dst, const ptrdiff_t dst_stride, -// const pixel *src, const ptrdiff_t src_stride, -// const int16_t *t1, const int w, const int h, -// const int wt); -function sgr_weighted1_neon, export=1 +// void dav1d_sgr_weighted1_8bpc_neon(pixel *dst, const ptrdiff_t dst_stride, +// const pixel *src, const ptrdiff_t src_stride, +// const int16_t *t1, const int w, const int h, +// const int wt); +function sgr_weighted1_8bpc_neon, export=1 dup v31.8h, w7 cmp x6, #2 add x9, x0, x1 @@ -1887,12 +1887,12 @@ function sgr_weighted1_neon, export=1 ret endfunc -// void dav1d_sgr_weighted2_neon(pixel *dst, const ptrdiff_t stride, -// const pixel *src, const ptrdiff_t src_stride, -// const int16_t *t1, const int16_t *t2, -// const int w, const int h, -// const int16_t wt[2]); -function sgr_weighted2_neon, export=1 +// void dav1d_sgr_weighted2_8bpc_neon(pixel *dst, const ptrdiff_t stride, +// const pixel *src, const ptrdiff_t src_stride, +// const int16_t *t1, const int16_t *t2, +// const int w, const int h, +// const int16_t wt[2]); +function sgr_weighted2_8bpc_neon, export=1 ldr x8, [sp] cmp x7, #2 add x10, x0, x1 |