[ffvpx] Update ffvp9/ffvp8 to release 4.0.2

3 files changed, 79 insertions, 89 deletions

diff --git a/media/ffvpx/libavcodec/x86/constants.c b/media/ffvpx/libavcodec/x86/constants.c
index 11002ee61..4bfb78cc3 100644
--- a/media/ffvpx/libavcodec/x86/constants.c
+++ b/media/ffvpx/libavcodec/x86/constants.c
@@ -26,23 +26,23 @@ DECLARE_ALIGNED(32, const ymm_reg,  ff_pw_1)    = { 0x0001000100010001ULL, 0x000
                                                     0x0001000100010001ULL, 0x0001000100010001ULL };
 DECLARE_ALIGNED(32, const ymm_reg,  ff_pw_2)    = { 0x0002000200020002ULL, 0x0002000200020002ULL,
                                                     0x0002000200020002ULL, 0x0002000200020002ULL };
-DECLARE_ALIGNED(16, const xmm_reg,  ff_pw_3)    = { 0x0003000300030003ULL, 0x0003000300030003ULL };
-DECLARE_ALIGNED(32, const ymm_reg,  ff_pw_4)    = { 0x0004000400040004ULL, 0x0004000400040004ULL,
+DECLARE_ASM_ALIGNED(16, const xmm_reg,  ff_pw_3)    = { 0x0003000300030003ULL, 0x0003000300030003ULL };
+DECLARE_ASM_ALIGNED(32, const ymm_reg,  ff_pw_4)    = { 0x0004000400040004ULL, 0x0004000400040004ULL,
                                                     0x0004000400040004ULL, 0x0004000400040004ULL };
-DECLARE_ALIGNED(16, const xmm_reg,  ff_pw_5)    = { 0x0005000500050005ULL, 0x0005000500050005ULL };
+DECLARE_ASM_ALIGNED(16, const xmm_reg,  ff_pw_5)    = { 0x0005000500050005ULL, 0x0005000500050005ULL };
 DECLARE_ALIGNED(16, const xmm_reg,  ff_pw_8)    = { 0x0008000800080008ULL, 0x0008000800080008ULL };
-DECLARE_ALIGNED(16, const xmm_reg,  ff_pw_9)    = { 0x0009000900090009ULL, 0x0009000900090009ULL };
+DECLARE_ASM_ALIGNED(16, const xmm_reg,  ff_pw_9)    = { 0x0009000900090009ULL, 0x0009000900090009ULL };
 DECLARE_ALIGNED(8,  const uint64_t, ff_pw_15)   =   0x000F000F000F000FULL;
 DECLARE_ALIGNED(16, const xmm_reg,  ff_pw_16)   = { 0x0010001000100010ULL, 0x0010001000100010ULL };
 DECLARE_ALIGNED(16, const xmm_reg,  ff_pw_17)   = { 0x0011001100110011ULL, 0x0011001100110011ULL };
-DECLARE_ALIGNED(16, const xmm_reg,  ff_pw_18)   = { 0x0012001200120012ULL, 0x0012001200120012ULL };
+DECLARE_ASM_ALIGNED(16, const xmm_reg,  ff_pw_18)   = { 0x0012001200120012ULL, 0x0012001200120012ULL };
 DECLARE_ALIGNED(16, const xmm_reg,  ff_pw_20)   = { 0x0014001400140014ULL, 0x0014001400140014ULL };
 DECLARE_ALIGNED(16, const xmm_reg,  ff_pw_32)   = { 0x0020002000200020ULL, 0x0020002000200020ULL };
-DECLARE_ALIGNED(8,  const uint64_t, ff_pw_42)   =   0x002A002A002A002AULL;
-DECLARE_ALIGNED(8,  const uint64_t, ff_pw_53)   =   0x0035003500350035ULL;
-DECLARE_ALIGNED(16, const xmm_reg,  ff_pw_64)   = { 0x0040004000400040ULL, 0x0040004000400040ULL };
-DECLARE_ALIGNED(8,  const uint64_t, ff_pw_96)   =   0x0060006000600060ULL;
-DECLARE_ALIGNED(8,  const uint64_t, ff_pw_128)  =   0x0080008000800080ULL;
+DECLARE_ASM_ALIGNED(8,  const uint64_t, ff_pw_42)   =   0x002A002A002A002AULL;
+DECLARE_ASM_ALIGNED(8,  const uint64_t, ff_pw_53)   =   0x0035003500350035ULL;
+DECLARE_ASM_ALIGNED(16, const xmm_reg,  ff_pw_64)   = { 0x0040004000400040ULL, 0x0040004000400040ULL };
+DECLARE_ASM_ALIGNED(8,  const uint64_t, ff_pw_96)   =   0x0060006000600060ULL;
+DECLARE_ASM_ALIGNED(8,  const uint64_t, ff_pw_128)  =   0x0080008000800080ULL;
 DECLARE_ALIGNED(32, const ymm_reg,  ff_pw_255)  = { 0x00ff00ff00ff00ffULL, 0x00ff00ff00ff00ffULL,
                                                     0x00ff00ff00ff00ffULL, 0x00ff00ff00ff00ffULL };
 DECLARE_ALIGNED(32, const ymm_reg,  ff_pw_256)  = { 0x0100010001000100ULL, 0x0100010001000100ULL,
@@ -74,7 +74,8 @@ DECLARE_ALIGNED(32, const ymm_reg,  ff_pb_2)    = { 0x0202020202020202ULL, 0x020
 DECLARE_ALIGNED(32, const ymm_reg,  ff_pb_3)    = { 0x0303030303030303ULL, 0x0303030303030303ULL,
                                                     0x0303030303030303ULL, 0x0303030303030303ULL };
 DECLARE_ALIGNED(32, const xmm_reg,  ff_pb_15)   = { 0x0F0F0F0F0F0F0F0FULL, 0x0F0F0F0F0F0F0F0FULL };
-DECLARE_ALIGNED(16, const xmm_reg,  ff_pb_80)   = { 0x8080808080808080ULL, 0x8080808080808080ULL };
+DECLARE_ALIGNED(32, const ymm_reg,  ff_pb_80)   = { 0x8080808080808080ULL, 0x8080808080808080ULL,
+                                                    0x8080808080808080ULL, 0x8080808080808080ULL };
 DECLARE_ALIGNED(32, const ymm_reg,  ff_pb_FE)   = { 0xFEFEFEFEFEFEFEFEULL, 0xFEFEFEFEFEFEFEFEULL,
                                                     0xFEFEFEFEFEFEFEFEULL, 0xFEFEFEFEFEFEFEFEULL };
 DECLARE_ALIGNED(8,  const uint64_t, ff_pb_FC)   =   0xFCFCFCFCFCFCFCFCULL;
diff --git a/media/ffvpx/libavcodec/x86/constants.h b/media/ffvpx/libavcodec/x86/constants.h
index bbb0ef844..85da38b7b 100644
--- a/media/ffvpx/libavcodec/x86/constants.h
+++ b/media/ffvpx/libavcodec/x86/constants.h
@@ -57,7 +57,7 @@ extern const ymm_reg  ff_pb_0;
 extern const ymm_reg  ff_pb_1;
 extern const ymm_reg  ff_pb_2;
 extern const ymm_reg  ff_pb_3;
-extern const xmm_reg  ff_pb_80;
+extern const ymm_reg  ff_pb_80;
 extern const ymm_reg  ff_pb_FE;
 extern const uint64_t ff_pb_FC;
 
diff --git a/media/ffvpx/libavcodec/x86/vp8dsp.asm b/media/ffvpx/libavcodec/x86/vp8dsp.asm
index e303b8029..75de5690a 100644
--- a/media/ffvpx/libavcodec/x86/vp8dsp.asm
+++ b/media/ffvpx/libavcodec/x86/vp8dsp.asm
@@ -664,6 +664,37 @@ INIT_XMM sse2
 FILTER_V 8
 
 %macro FILTER_BILINEAR 1
+%if cpuflag(ssse3)
+cglobal put_vp8_bilinear%1_v, 7, 7, 5, dst, dststride, src, srcstride, height, picreg, my
+    shl      myd, 4
+%ifdef PIC
+    lea  picregq, [bilinear_filter_vb_m]
+%endif
+    pxor      m4, m4
+    mova      m3, [bilinear_filter_vb+myq-16]
+.nextrow:
+    movh      m0, [srcq+srcstrideq*0]
+    movh      m1, [srcq+srcstrideq*1]
+    movh      m2, [srcq+srcstrideq*2]
+    punpcklbw m0, m1
+    punpcklbw m1, m2
+    pmaddubsw m0, m3
+    pmaddubsw m1, m3
+    psraw     m0, 2
+    psraw     m1, 2
+    pavgw     m0, m4
+    pavgw     m1, m4
+%if mmsize==8
+    packuswb  m0, m0
+    packuswb  m1, m1
+    movh   [dstq+dststrideq*0], m0
+    movh   [dstq+dststrideq*1], m1
+%else
+    packuswb  m0, m1
+    movh   [dstq+dststrideq*0], m0
+    movhps [dstq+dststrideq*1], m0
+%endif
+%else ; cpuflag(ssse3)
 cglobal put_vp8_bilinear%1_v, 7, 7, 7, dst, dststride, src, srcstride, height, picreg, my
     shl      myd, 4
 %ifdef PIC
@@ -701,6 +732,7 @@ cglobal put_vp8_bilinear%1_v, 7, 7, 7, dst, dststride, src, srcstride, height, p
     movh   [dstq+dststrideq*0], m0
     movhps [dstq+dststrideq*1], m0
 %endif
+%endif ; cpuflag(ssse3)
 
     lea     dstq, [dstq+dststrideq*2]
     lea     srcq, [srcq+srcstrideq*2]
@@ -708,6 +740,37 @@ cglobal put_vp8_bilinear%1_v, 7, 7, 7, dst, dststride, src, srcstride, height, p
     jg .nextrow
     REP_RET
 
+%if cpuflag(ssse3)
+cglobal put_vp8_bilinear%1_h, 6, 6 + npicregs, 5, dst, dststride, src, srcstride, height, mx, picreg
+    shl      mxd, 4
+%ifdef PIC
+    lea  picregq, [bilinear_filter_vb_m]
+%endif
+    pxor      m4, m4
+    mova      m2, [filter_h2_shuf]
+    mova      m3, [bilinear_filter_vb+mxq-16]
+.nextrow:
+    movu      m0, [srcq+srcstrideq*0]
+    movu      m1, [srcq+srcstrideq*1]
+    pshufb    m0, m2
+    pshufb    m1, m2
+    pmaddubsw m0, m3
+    pmaddubsw m1, m3
+    psraw     m0, 2
+    psraw     m1, 2
+    pavgw     m0, m4
+    pavgw     m1, m4
+%if mmsize==8
+    packuswb  m0, m0
+    packuswb  m1, m1
+    movh   [dstq+dststrideq*0], m0
+    movh   [dstq+dststrideq*1], m1
+%else
+    packuswb  m0, m1
+    movh   [dstq+dststrideq*0], m0
+    movhps [dstq+dststrideq*1], m0
+%endif
+%else ; cpuflag(ssse3)
 cglobal put_vp8_bilinear%1_h, 6, 6 + npicregs, 7, dst, dststride, src, srcstride, height, mx, picreg
     shl      mxd, 4
 %ifdef PIC
@@ -746,6 +809,7 @@ cglobal put_vp8_bilinear%1_h, 6, 6 + npicregs, 7, dst, dststride, src, srcstride
     movh   [dstq+dststrideq*0], m0
     movhps [dstq+dststrideq*1], m0
 %endif
+%endif ; cpuflag(ssse3)
 
     lea     dstq, [dstq+dststrideq*2]
     lea     srcq, [srcq+srcstrideq*2]
@@ -758,85 +822,10 @@ INIT_MMX mmxext
 FILTER_BILINEAR 4
 INIT_XMM sse2
 FILTER_BILINEAR 8
-
-%macro FILTER_BILINEAR_SSSE3 1
-cglobal put_vp8_bilinear%1_v, 7, 7, 5, dst, dststride, src, srcstride, height, picreg, my
-    shl      myd, 4
-%ifdef PIC
-    lea  picregq, [bilinear_filter_vb_m]
-%endif
-    pxor      m4, m4
-    mova      m3, [bilinear_filter_vb+myq-16]
-.nextrow:
-    movh      m0, [srcq+srcstrideq*0]
-    movh      m1, [srcq+srcstrideq*1]
-    movh      m2, [srcq+srcstrideq*2]
-    punpcklbw m0, m1
-    punpcklbw m1, m2
-    pmaddubsw m0, m3
-    pmaddubsw m1, m3
-    psraw     m0, 2
-    psraw     m1, 2
-    pavgw     m0, m4
-    pavgw     m1, m4
-%if mmsize==8
-    packuswb  m0, m0
-    packuswb  m1, m1
-    movh   [dstq+dststrideq*0], m0
-    movh   [dstq+dststrideq*1], m1
-%else
-    packuswb  m0, m1
-    movh   [dstq+dststrideq*0], m0
-    movhps [dstq+dststrideq*1], m0
-%endif
-
-    lea     dstq, [dstq+dststrideq*2]
-    lea     srcq, [srcq+srcstrideq*2]
-    sub  heightd, 2
-    jg .nextrow
-    REP_RET
-
-cglobal put_vp8_bilinear%1_h, 6, 6 + npicregs, 5, dst, dststride, src, srcstride, height, mx, picreg
-    shl      mxd, 4
-%ifdef PIC
-    lea  picregq, [bilinear_filter_vb_m]
-%endif
-    pxor      m4, m4
-    mova      m2, [filter_h2_shuf]
-    mova      m3, [bilinear_filter_vb+mxq-16]
-.nextrow:
-    movu      m0, [srcq+srcstrideq*0]
-    movu      m1, [srcq+srcstrideq*1]
-    pshufb    m0, m2
-    pshufb    m1, m2
-    pmaddubsw m0, m3
-    pmaddubsw m1, m3
-    psraw     m0, 2
-    psraw     m1, 2
-    pavgw     m0, m4
-    pavgw     m1, m4
-%if mmsize==8
-    packuswb  m0, m0
-    packuswb  m1, m1
-    movh   [dstq+dststrideq*0], m0
-    movh   [dstq+dststrideq*1], m1
-%else
-    packuswb  m0, m1
-    movh   [dstq+dststrideq*0], m0
-    movhps [dstq+dststrideq*1], m0
-%endif
-
-    lea     dstq, [dstq+dststrideq*2]
-    lea     srcq, [srcq+srcstrideq*2]
-    sub  heightd, 2
-    jg .nextrow
-    REP_RET
-%endmacro
-
 INIT_MMX ssse3
-FILTER_BILINEAR_SSSE3 4
+FILTER_BILINEAR 4
 INIT_XMM ssse3
-FILTER_BILINEAR_SSSE3 8
+FILTER_BILINEAR 8
 
 INIT_MMX mmx
 cglobal put_vp8_pixels8, 5, 5, 0, dst, dststride, src, srcstride, height