virtualenv/FFmpeg
1
0
Fork 0
mirror of https://github.com/FFmpeg/FFmpeg.git synced 2025-08-15 14:13:16 +02:00
Code Issues Releases Activity

Revert "lavc/aarch64: add hevc sao band 8x8 tiling"

Browse Source 
This reverts commit f63f9be37c, as
it breaks fate-hevc.

Signed-off-by: Martin Storsjö <martin@martin.st>
This commit is contained in:
Martin Storsjö
2022-01-05 10:22:06 +02:00
parent 0c9c9bbd01
commit cafed377eb
2 changed files with 5 additions and 12 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
libavcodec/aarch64/hevcdsp_init_aarch64.c
View File

@@ -77,11 +77,7 @@ av_cold void ff_hevc_dsp_init_aarch64(HEVCDSPContext *c, const int bit_depth)
c->idct_dc[1] = ff_hevc_idct_8x8_dc_8_neon; c->idct_dc[1] = ff_hevc_idct_8x8_dc_8_neon;
c->idct_dc[2] = ff_hevc_idct_16x16_dc_8_neon; c->idct_dc[2] = ff_hevc_idct_16x16_dc_8_neon;
c->idct_dc[3] = ff_hevc_idct_32x32_dc_8_neon; c->idct_dc[3] = ff_hevc_idct_32x32_dc_8_neon;
c->sao_band_filter[0] = c->sao_band_filter[0] = ff_hevc_sao_band_filter_8x8_8_neon;
c->sao_band_filter[1] =
c->sao_band_filter[2] =
c->sao_band_filter[3] =
c->sao_band_filter[4] = ff_hevc_sao_band_filter_8x8_8_neon;
c->sao_edge_filter[0] = ff_hevc_sao_edge_filter_8x8_8_neon; c->sao_edge_filter[0] = ff_hevc_sao_edge_filter_8x8_8_neon;
c->sao_edge_filter[1] = c->sao_edge_filter[1] =
c->sao_edge_filter[2] = c->sao_edge_filter[2] =

11
libavcodec/aarch64/hevcdsp_sao_neon.S
View File

@@ -3,7 +3,7 @@
* *
* AArch64 NEON optimised SAO functions for HEVC decoding * AArch64 NEON optimised SAO functions for HEVC decoding
* *
* Copyright (c) 2020-2021 J. Dekker <jdek@itanimul.li> * Copyright (c) 2020 Josh Dekker <josh@itanimul.li>
* *
* This file is part of FFmpeg. * This file is part of FFmpeg.
* *
@@ -35,7 +35,6 @@ function ff_hevc_sao_band_filter_8x8_8_neon, export=1
stp xzr, xzr, [sp, #32] stp xzr, xzr, [sp, #32]
stp xzr, xzr, [sp, #48] stp xzr, xzr, [sp, #48]
mov w8, #4 mov w8, #4
sxtw x6, w6
0: ldrsh x9, [x4, x8, lsl #1] // sao_offset_val[k+1] 0: ldrsh x9, [x4, x8, lsl #1] // sao_offset_val[k+1]
subs w8, w8, #1 subs w8, w8, #1
add w10, w8, w5 // k + sao_left_class add w10, w8, w5 // k + sao_left_class
@@ -44,9 +43,7 @@ function ff_hevc_sao_band_filter_8x8_8_neon, export=1
bne 0b bne 0b
ld1 {v16.16b-v19.16b}, [sp], #64 ld1 {v16.16b-v19.16b}, [sp], #64
movi v20.8h, #1 movi v20.8h, #1
sub x2, x2, x6 // stride_dst - width 1: mov w8, w6 // beginning of line
sub x3, x3, x6 // stride_src - width
1: mov x8, x6 // beginning of line
2: // Simple layout for accessing 16bit values 2: // Simple layout for accessing 16bit values
// with 8bit LUT. // with 8bit LUT.
// //
@@ -55,7 +52,7 @@ function ff_hevc_sao_band_filter_8x8_8_neon, export=1
// |xDE#xAD|xCA#xFE|xBE#xEF|xFE#xED|.... // |xDE#xAD|xCA#xFE|xBE#xEF|xFE#xED|....
// +-----------------------------------> // +----------------------------------->
// i-0 i-1 i-2 i-3 // i-0 i-1 i-2 i-3
ld1 {v2.8b}, [x1], #8 // dst[x] = av_clip_pixel(src[x] + offset_table[src[x] >> shift]); ld1 {v2.8b}, [x1] // dst[x] = av_clip_pixel(src[x] + offset_table[src[x] >> shift]);
uxtl v0.8h, v2.8b // load src[x] uxtl v0.8h, v2.8b // load src[x]
ushr v2.8h, v0.8h, #3 // >> BIT_DEPTH - 3 ushr v2.8h, v0.8h, #3 // >> BIT_DEPTH - 3
shl v1.8h, v2.8h, #1 // low (x2, accessing short) shl v1.8h, v2.8h, #1 // low (x2, accessing short)
@@ -64,7 +61,7 @@ function ff_hevc_sao_band_filter_8x8_8_neon, export=1
tbx v2.16b, {v16.16b-v19.16b}, v1.16b // table tbx v2.16b, {v16.16b-v19.16b}, v1.16b // table
add v1.8h, v0.8h, v2.8h // src[x] + table add v1.8h, v0.8h, v2.8h // src[x] + table
sqxtun v4.8b, v1.8h // clip + narrow sqxtun v4.8b, v1.8h // clip + narrow
st1 {v4.8b}, [x0], #8 // store st1 {v4.8b}, [x0] // store
subs w8, w8, #8 // done 8 pixels subs w8, w8, #8 // done 8 pixels
bne 2b bne 2b
subs w7, w7, #1 // finished line, prep. new subs w7, w7, #1 // finished line, prep. new