Merge commit '0d4d43513786f1df4d561e1fac924fb0722c6700'

* commit '0d4d43513786f1df4d561e1fac924fb0722c6700':
  hevc: Add NEON add_residual for bitdepth 8

See 03cecf45c134ebbaecb62505fe444ade423ea7dc

Merged-by: James Almer <jamrial@gmail.com>

2 files changed, 92 insertions, 75 deletions

diff --git a/libavcodec/arm/hevcdsp_idct_neon.S b/libavcodec/arm/hevcdsp_idct_neon.S
index e8f18d42a7..849304c350 100644
--- a/libavcodec/arm/hevcdsp_idct_neon.S
+++ b/libavcodec/arm/hevcdsp_idct_neon.S
@@ -29,75 +29,92 @@ const trans, align=4
         .short 57, 43, 25, 9
 endconst
 
-function ff_hevc_add_residual_4x4_neon_8, export=1
-        vldm        r1, {q0-q1}
-        vld1.32     d4[0], [r0], r2
-        vld1.32     d4[1], [r0], r2
-        vld1.32     d5[0], [r0], r2
-        vld1.32     d5[1], [r0], r2
-        sub         r0, r0, r2, lsl #2
-        vmovl.u8    q8, d4
-        vmovl.u8    q9, d5
-        vqadd.s16   q0, q0, q8
-        vqadd.s16   q1, q1, q9
-        vqmovun.s16 d0, q0
-        vqmovun.s16 d1, q1
-        vst1.32     d0[0], [r0], r2
-        vst1.32     d0[1], [r0], r2
-        vst1.32     d1[0], [r0], r2
-        vst1.32     d1[1], [r0], r2
-        bx          lr
+function ff_hevc_add_residual_4x4_8_neon, export=1
+        vld1.16         {q0-q1}, [r1, :128]
+        vld1.32         d4[0], [r0, :32], r2
+        vld1.32         d4[1], [r0, :32], r2
+        vld1.32         d5[0], [r0, :32], r2
+        vld1.32         d5[1], [r0, :32], r2
+        sub             r0, r0, r2, lsl #2
+        vmovl.u8        q8, d4
+        vmovl.u8        q9, d5
+        vqadd.s16       q0, q0, q8
+        vqadd.s16       q1, q1, q9
+        vqmovun.s16     d0, q0
+        vqmovun.s16     d1, q1
+        vst1.32         d0[0], [r0, :32], r2
+        vst1.32         d0[1], [r0, :32], r2
+        vst1.32         d1[0], [r0, :32], r2
+        vst1.32         d1[1], [r0, :32], r2
+        bx              lr
 endfunc
 
-function ff_hevc_add_residual_8x8_neon_8, export=1
-        mov         r3,   #8
-1:      subs        r3,   #1
-        vld1.16     {q0}, [r1]!
-        vld1.8      d16,  [r0]
-        vmovl.u8    q8,   d16
-        vqadd.s16   q0,   q8
-        vqmovun.s16 d0,   q0
-        vst1.32     d0,   [r0], r2
-        bne         1b
-        bx          lr
+function ff_hevc_add_residual_8x8_8_neon, export=1
+        mov             r3,   #8
+1:      subs            r3,   #2
+        vld1.16         {q0-q1}, [r1, :128]!
+        vld1.8          {d16},   [r0, :64]
+        add             r12, r0, r2
+        vld1.8          {d17},   [r12, :64]
+        vmovl.u8        q9,   d16
+        vmovl.u8        q8,   d17
+        vqadd.s16       q0,   q9
+        vqadd.s16       q1,   q8
+        vqmovun.s16     d0,   q0
+        vqmovun.s16     d1,   q1
+        vst1.8          d0,   [r0, :64], r2
+        vst1.8          d1,   [r0, :64], r2
+        bne             1b
+        bx              lr
 endfunc
 
-function ff_hevc_add_residual_16x16_neon_8, export=1
-        mov         r3,   #16
-1:      subs        r3,   #1
-        vld1.16     {q0, q1}, [r1]!
-        vld1.8      {q8},  [r0]
-        vmovl.u8    q9,  d16
-        vmovl.u8    q10, d17
-        vqadd.s16   q0,  q9
-        vqadd.s16   q1,  q10
-        vqmovun.s16 d0,  q0
-        vqmovun.s16 d1,  q1
-        vst1.8      {q0},   [r0], r2
-        bne         1b
-        bx          lr
+function ff_hevc_add_residual_16x16_8_neon, export=1
+        mov             r3,  #16
+        add             r12, r0, r2
+        add             r2,  r2, r2
+1:      subs            r3,  #2
+        vld1.8          {q8},     [r0, :128]
+        vld1.16         {q0, q1}, [r1, :128]!
+        vld1.8          {q11},    [r12, :128]
+        vld1.16         {q2, q3}, [r1, :128]!
+        vmovl.u8        q9,  d16
+        vmovl.u8        q10, d17
+        vmovl.u8        q12, d22
+        vmovl.u8        q13, d23
+        vqadd.s16       q0,  q9
+        vqadd.s16       q1,  q10
+        vqadd.s16       q2,  q12
+        vqadd.s16       q3,  q13
+        vqmovun.s16     d0,  q0
+        vqmovun.s16     d1,  q1
+        vqmovun.s16     d2,  q2
+        vqmovun.s16     d3,  q3
+        vst1.8          {q0},     [r0, :128], r2
+        vst1.8          {q1},     [r12, :128], r2
+        bne             1b
+        bx              lr
 endfunc
 
-function ff_hevc_add_residual_32x32_neon_8, export=1
-        mov         r3,   #32
-1:      subs        r3,   #1
-        vldm        r1!, {q0-q3}
-        vld1.8      {q8, q9},  [r0]
-        vmovl.u8    q10, d16
-        vmovl.u8    q11, d17
-        vmovl.u8    q12, d18
-        vmovl.u8    q13, d19
-        vqadd.s16   q0,  q10
-        vqadd.s16   q1,  q11
-        vqadd.s16   q2,  q12
-        vqadd.s16   q3,  q13
-        vqmovun.s16 d0,  q0
-        vqmovun.s16 d1,  q1
-        vqmovun.s16 d2,  q2
-        vqmovun.s16 d3,  q3
-        vst1.8     {q0, q1},   [r0], r2
-        bne         1b
-        bx          lr
+function ff_hevc_add_residual_32x32_8_neon, export=1
+        mov             r3,  #32
+1:      subs            r3,  #1
+        vldm            r1!, {q0-q3}
+        vld1.8          {q8, q9}, [r0, :128]
+        vmovl.u8        q10, d16
+        vmovl.u8        q11, d17
+        vmovl.u8        q12, d18
+        vmovl.u8        q13, d19
+        vqadd.s16       q0,  q10
+        vqadd.s16       q1,  q11
+        vqadd.s16       q2,  q12
+        vqadd.s16       q3,  q13
+        vqmovun.s16     d0,  q0
+        vqmovun.s16     d1,  q1
+        vqmovun.s16     d2,  q2
+        vqmovun.s16     d3,  q3
+        vst1.8          {q0, q1}, [r0, :128], r2
+        bne             1b
+        bx              lr
 endfunc
 
 /* uses registers q2 - q9 for temp values */
diff --git a/libavcodec/arm/hevcdsp_init_neon.c b/libavcodec/arm/hevcdsp_init_neon.c
index 740de86bb2..841030e9bb 100644
--- a/libavcodec/arm/hevcdsp_init_neon.c
+++ b/libavcodec/arm/hevcdsp_init_neon.c
@@ -27,6 +27,14 @@ void ff_hevc_v_loop_filter_luma_neon(uint8_t *_pix, ptrdiff_t _stride, int _beta
 void ff_hevc_h_loop_filter_luma_neon(uint8_t *_pix, ptrdiff_t _stride, int _beta, int *_tc, uint8_t *_no_p, uint8_t *_no_q);
 void ff_hevc_v_loop_filter_chroma_neon(uint8_t *_pix, ptrdiff_t _stride, int *_tc, uint8_t *_no_p, uint8_t *_no_q);
 void ff_hevc_h_loop_filter_chroma_neon(uint8_t *_pix, ptrdiff_t _stride, int *_tc, uint8_t *_no_p, uint8_t *_no_q);
+void ff_hevc_add_residual_4x4_8_neon(uint8_t *_dst, int16_t *coeffs,
+                                     ptrdiff_t stride);
+void ff_hevc_add_residual_8x8_8_neon(uint8_t *_dst, int16_t *coeffs,
+                                     ptrdiff_t stride);
+void ff_hevc_add_residual_16x16_8_neon(uint8_t *_dst, int16_t *coeffs,
+                                       ptrdiff_t stride);
+void ff_hevc_add_residual_32x32_8_neon(uint8_t *_dst, int16_t *coeffs,
+                                       ptrdiff_t stride);
 void ff_hevc_idct_4x4_dc_8_neon(int16_t *coeffs);
 void ff_hevc_idct_8x8_dc_8_neon(int16_t *coeffs);
 void ff_hevc_idct_16x16_dc_8_neon(int16_t *coeffs);
@@ -42,14 +50,6 @@ void ff_hevc_idct_4x4_10_neon(int16_t *coeffs, int col_limit);
 void ff_hevc_idct_8x8_10_neon(int16_t *coeffs, int col_limit);
 void ff_hevc_idct_16x16_10_neon(int16_t *coeffs, int col_limit);
 void ff_hevc_transform_luma_4x4_neon_8(int16_t *coeffs);
-void ff_hevc_add_residual_4x4_neon_8(uint8_t *_dst, int16_t *coeffs,
-                                     ptrdiff_t stride);
-void ff_hevc_add_residual_8x8_neon_8(uint8_t *_dst, int16_t *coeffs,
-                                     ptrdiff_t stride);
-void ff_hevc_add_residual_16x16_neon_8(uint8_t *_dst, int16_t *coeffs,
-                                       ptrdiff_t stride);
-void ff_hevc_add_residual_32x32_neon_8(uint8_t *_dst, int16_t *coeffs,
-                                       ptrdiff_t stride);
 
 #define PUT_PIXELS(name) \
     void name(int16_t *dst, uint8_t *src, \
@@ -158,6 +158,10 @@ av_cold void ff_hevc_dsp_init_neon(HEVCDSPContext *c, const int bit_depth)
         c->hevc_h_loop_filter_luma     = ff_hevc_h_loop_filter_luma_neon;
         c->hevc_v_loop_filter_chroma   = ff_hevc_v_loop_filter_chroma_neon;
         c->hevc_h_loop_filter_chroma   = ff_hevc_h_loop_filter_chroma_neon;
+        c->add_residual[0]             = ff_hevc_add_residual_4x4_8_neon;
+        c->add_residual[1]             = ff_hevc_add_residual_8x8_8_neon;
+        c->add_residual[2]             = ff_hevc_add_residual_16x16_8_neon;
+        c->add_residual[3]             = ff_hevc_add_residual_32x32_8_neon;
         c->idct_dc[0]                  = ff_hevc_idct_4x4_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;
@@ -165,10 +169,6 @@ av_cold void ff_hevc_dsp_init_neon(HEVCDSPContext *c, const int bit_depth)
         c->idct[0]                     = ff_hevc_idct_4x4_8_neon;
         c->idct[1]                     = ff_hevc_idct_8x8_8_neon;
         c->idct[2]                     = ff_hevc_idct_16x16_8_neon;
-        c->add_residual[0]             = ff_hevc_add_residual_4x4_neon_8;
-        c->add_residual[1]             = ff_hevc_add_residual_8x8_neon_8;
-        c->add_residual[2]             = ff_hevc_add_residual_16x16_neon_8;
-        c->add_residual[3]             = ff_hevc_add_residual_32x32_neon_8;
         c->transform_4x4_luma          = ff_hevc_transform_luma_4x4_neon_8;
         put_hevc_qpel_neon[1][0]       = ff_hevc_put_qpel_v1_neon_8;
         put_hevc_qpel_neon[2][0]       = ff_hevc_put_qpel_v2_neon_8;