swscale/arm/yuv2rgb: add ff_yuv420p_to_{argb,rgba,abgr,bgra}_neon_{16,32}

2 files changed, 118 insertions, 11 deletions

diff --git a/libswscale/arm/swscale_unscaled.c b/libswscale/arm/swscale_unscaled.c
index 4c121228ab..1b50acd7e2 100644
--- a/libswscale/arm/swscale_unscaled.c
+++ b/libswscale/arm/swscale_unscaled.c
@@ -63,6 +63,50 @@ static int rgbx_to_nv12_neon_16_wrapper(SwsContext *context, const uint8_t *src[
 }
 #endif
 
+#define YUV_TO_RGB_TABLE(precision)                                                         \
+        c->yuv2rgb_v2r_coeff / ((precision) == 16 ? 1 << 7 : 1),                            \
+        c->yuv2rgb_u2g_coeff / ((precision) == 16 ? 1 << 7 : 1),                            \
+        c->yuv2rgb_v2g_coeff / ((precision) == 16 ? 1 << 7 : 1),                            \
+        c->yuv2rgb_u2b_coeff / ((precision) == 16 ? 1 << 7 : 1),                            \
+
+#define DECLARE_FF_YUV420P_TO_RGBX_FUNCS(ofmt, precision)                                   \
+int ff_yuv420p_to_##ofmt##_neon_##precision(int w, int h,                                   \
+                                 uint8_t *dst, int linesize,                                \
+                                 const uint8_t *srcY, int linesizeY,                        \
+                                 const uint8_t *srcU, int linesizeU,                        \
+                                 const uint8_t *srcV, int linesizeV,                        \
+                                 const int16_t *table,                                      \
+                                 int y_offset,                                              \
+                                 int y_coeff);                                              \
+                                                                                            \
+static int yuv420p_to_##ofmt##_neon_wrapper_##precision(SwsContext *c, const uint8_t *src[],\
+                                           int srcStride[], int srcSliceY, int srcSliceH,   \
+                                           uint8_t *dst[], int dstStride[]) {               \
+    const int16_t yuv2rgb_table[] = { YUV_TO_RGB_TABLE(precision) };                        \
+                                                                                            \
+    ff_yuv420p_to_##ofmt##_neon_##precision(c->srcW, srcSliceH,                             \
+                                 dst[0] + srcSliceY * dstStride[0], dstStride[0],           \
+                                 src[0], srcStride[0],                                      \
+                                 src[1], srcStride[1],                                      \
+                                 src[2], srcStride[2],                                      \
+                                 yuv2rgb_table,                                             \
+                                 c->yuv2rgb_y_offset >> 9,                                  \
+                                 c->yuv2rgb_y_coeff / ((precision) == 16 ? 1 << 7 : 1));    \
+                                                                                            \
+    return 0;                                                                               \
+}                                                                                           \
+
+#define DECLARE_FF_YUV420P_TO_ALL_RGBX_FUNCS(precision)                                     \
+DECLARE_FF_YUV420P_TO_RGBX_FUNCS(argb, precision)                                           \
+DECLARE_FF_YUV420P_TO_RGBX_FUNCS(rgba, precision)                                           \
+DECLARE_FF_YUV420P_TO_RGBX_FUNCS(abgr, precision)                                           \
+DECLARE_FF_YUV420P_TO_RGBX_FUNCS(bgra, precision)                                           \
+
+#define DECLARE_FF_YUV420P_TO_ALL_RGBX_ALL_PRECISION_FUNCS                                  \
+DECLARE_FF_YUV420P_TO_ALL_RGBX_FUNCS(16)                                                    \
+
+DECLARE_FF_YUV420P_TO_ALL_RGBX_ALL_PRECISION_FUNCS
+
 #define DECLARE_FF_NVX_TO_RGBX_FUNCS(ifmt, ofmt, precision)                                 \
 int ff_##ifmt##_to_##ofmt##_neon_##precision(int w, int h,                                  \
                                  uint8_t *dst, int linesize,                                \
@@ -75,12 +119,7 @@ int ff_##ifmt##_to_##ofmt##_neon_##precision(int w, int h,
 static int ifmt##_to_##ofmt##_neon_wrapper_##precision(SwsContext *c, const uint8_t *src[], \
                                            int srcStride[], int srcSliceY, int srcSliceH,   \
                                            uint8_t *dst[], int dstStride[]) {               \
-    const int16_t yuv2rgb_table[] = {                                                       \
-        c->yuv2rgb_v2r_coeff / ((precision) == 16 ? 1 << 7 : 1),                            \
-        c->yuv2rgb_u2g_coeff / ((precision) == 16 ? 1 << 7 : 1),                            \
-        c->yuv2rgb_v2g_coeff / ((precision) == 16 ? 1 << 7 : 1),                            \
-        c->yuv2rgb_u2b_coeff / ((precision) == 16 ? 1 << 7 : 1),                            \
-    };                                                                                      \
+    const int16_t yuv2rgb_table[] = { YUV_TO_RGB_TABLE(precision) };                        \
                                                                                             \
     ff_##ifmt##_to_##ofmt##_neon_##precision(c->srcW, srcSliceH,                            \
                                  dst[0] + srcSliceY * dstStride[0], dstStride[0],           \
@@ -138,6 +177,7 @@ static void get_unscaled_swscale_neon(SwsContext *c) {
 
     SET_FF_NVX_TO_ALL_RGBX_FUNC(nv12, NV12, accurate_rnd);
     SET_FF_NVX_TO_ALL_RGBX_FUNC(nv21, NV21, accurate_rnd);
+    SET_FF_NVX_TO_ALL_RGBX_FUNC(yuv420p, YUV420P, accurate_rnd);
 }
 
 void ff_get_unscaled_swscale_arm(SwsContext *c)
diff --git a/libswscale/arm/yuv2rgb_neon.S b/libswscale/arm/yuv2rgb_neon.S
index 9f9dd2aaa1..dd00246ef3 100644
--- a/libswscale/arm/yuv2rgb_neon.S
+++ b/libswscale/arm/yuv2rgb_neon.S
@@ -103,7 +103,8 @@
     vmovl.u8            q15, \y1                                       @ 8px of y
 
     vdup.16             q5, r9                                         @ q5  = y_offset
-    vdup.16             q7, r10                                        @ q7  = y_coeff
+    vmov                d14, d0                                        @ q7  = y_coeff
+    vmov                d15, d0                                        @ q7  = y_coeff
 
     vsub.s16            q14, q5
     vsub.s16            q15, q5
@@ -184,7 +185,7 @@
     compute_8px_32      r11, d30, \ofmt
 .endm
 
-.macro load_args
+.macro load_args_nvx
     push                {r4-r12, lr}
     vpush               {q4-q7}
     ldr                 r4, [sp, #104]                                 @ r4  = srcY
@@ -206,9 +207,42 @@
     sub                 r7, r7, r0                                     @ r7 = linesizeC     - width     (paddingC)
 .endm
 
+.macro load_args_yuv420p
+    push                {r4-r12, lr}
+    vpush               {q4-q7}
+    ldr                 r4, [sp, #104]                                 @ r4  = srcY
+    ldr                 r5, [sp, #108]                                 @ r5  = linesizeY
+    ldr                 r6, [sp, #112]                                 @ r6  = srcU
+    ldr                 r8, [sp, #128]                                 @ r8  = table
+    ldr                 r9, [sp, #132]                                 @ r9  = y_offset
+    ldr                 r10,[sp, #136]                                 @ r10 = y_coeff
+    vdup.16             d0, r10                                        @ d0  = y_coeff
+    vld1.16             {d1}, [r8]                                     @ d1  = *table
+    add                 r11, r2, r3                                    @ r11 = dst + linesize (dst2)
+    add                 r12, r4, r5                                    @ r12 = srcY + linesizeY (srcY2)
+    lsl                 r3, r3, #1
+    lsl                 r5, r5, #1
+    lsl                 r8, r0, #2
+    sub                 r3, r3, r8                                     @ r3 = linesize  * 2 - width * 4 (padding)
+    sub                 r5, r5, r0                                     @ r5 = linesizeY * 2 - width     (paddingY)
+    ldr                 r10,[sp, #120]                                 @ r10 = srcV
+.endm
+
 .macro declare_func ifmt ofmt precision
 function ff_\ifmt\()_to_\ofmt\()_neon_\precision\(), export=1
-    load_args
+
+.ifc \ifmt,nv12
+    load_args_nvx
+.endif
+
+.ifc \ifmt,nv21
+    load_args_nvx
+.endif
+
+.ifc \ifmt,yuv420p
+    load_args_yuv420p
+.endif
+
 1:
     mov                 r8, r0                                         @ r8 = width
 2:
@@ -216,16 +250,30 @@ function ff_\ifmt\()_to_\ofmt\()_neon_\precision\(), export=1
     pld [r4, #64*3]
     pld [r12, #64*3]
 
-    vld2.8              {d2, d3}, [r6]!                                @ q1: interleaved chroma line
     vmov.i8             d10, #128
+
 .ifc \ifmt,nv12
+    vld2.8              {d2, d3}, [r6]!                                @ q1: interleaved chroma line
     vsubl.u8            q14, d2, d10                                   @ q14 = U - 128
     vsubl.u8            q15, d3, d10                                   @ q15 = V - 128
-.else
+.endif
+
+.ifc \ifmt,nv21
+    vld2.8              {d2, d3}, [r6]!                                @ q1: interleaved chroma line
     vsubl.u8            q14, d3, d10                                   @ q14 = U - 128
     vsubl.u8            q15, d2, d10                                   @ q15 = V - 128
 .endif
 
+.ifc \ifmt,yuv420p
+    pld [r10, #64*3]
+
+    vld1.8              d2, [r6]!                                      @ d2: chroma red line
+    vld1.8              d3, [r10]!                                     @ d3: chroma blue line
+    vsubl.u8            q14, d2, d10                                   @ q14 = U - 128
+    vsubl.u8            q15, d3, d10                                   @ q15 = V - 128
+.endif
+
+
     process_16px_\precision \ofmt
 
     subs                r8, r8, #16                                    @ width -= 16
@@ -235,7 +283,24 @@ function ff_\ifmt\()_to_\ofmt\()_neon_\precision\(), export=1
     add                 r4, r4, r5                                     @ srcY  += paddingY
     add                 r11, r11, r3                                   @ dst2  += padding
     add                 r12, r12, r5                                   @ srcY2 += paddingY
+
+.ifc \ifmt,nv12
     add                 r6, r6, r7                                     @ srcC  += paddingC
+.endif
+
+.ifc \ifmt,nv21
+    add                 r6, r6, r7                                     @ srcC  += paddingC
+.endif
+
+.ifc \ifmt,yuv420p
+    ldr                 r7, [sp, #116]                                 @ r7     = linesizeU
+    sub                 r7, r7, r0, lsr #1                             @ r7     = linesizeU - width / 2 (paddingU)
+    add                 r6, r6, r7                                     @ srcU  += paddingU
+
+    ldr                 r7, [sp, #124]                                 @ r7     = linesizeV
+    sub                 r7, r7, r0, lsr #1                             @ r7     = linesizeV - width / 2 (paddingV)
+    add                 r10, r10, r7                                   @ srcU  += paddingV
+.endif
 
     subs                r1, r1, #2                                     @ height -= 2
     bgt                 1b
@@ -257,3 +322,5 @@ declare_rgb_funcs nv12, 16
 declare_rgb_funcs nv21, 16
 declare_rgb_funcs nv12, 32
 declare_rgb_funcs nv21, 32
+declare_rgb_funcs yuv420p, 16
+declare_rgb_funcs yuv420p, 32