libvpx: Pull from upstream

source/libvpx/vp9/common/x86/vp9_subpixel_8t_ssse3.asm

 ;
 ;  Copyright (c) 2010 The WebM project authors. All Rights Reserved.
 ;
 ;  Use of this source code is governed by a BSD-style license
 ;  that can be found in the LICENSE file in the root of the source
 ;  tree. An additional intellectual property rights grant can be found
 ;  in the file PATENTS.  All contributing project authors may
 ;  be found in the AUTHORS file in the root of the source tree.
 ;
 
 
 %include "vpx_ports/x86_abi_support.asm"
 
 %macro VERTx4 1
     mov         rdx, arg(5)                 ;filter ptr
     mov         rsi, arg(0)                 ;src_ptr
     mov         rdi, arg(2)                 ;output_ptr
     mov         rcx, 0x0400040
 
     movdqa      xmm4, [rdx]                 ;load filters
-    movd        xmm5, rcx
+    movq        xmm5, rcx
     packsswb    xmm4, xmm4
     pshuflw     xmm0, xmm4, 0b              ;k0_k1
     pshuflw     xmm1, xmm4, 01010101b       ;k2_k3
     pshuflw     xmm2, xmm4, 10101010b       ;k4_k5
     pshuflw     xmm3, xmm4, 11111111b       ;k6_k7
 
     punpcklqdq  xmm0, xmm0
     punpcklqdq  xmm1, xmm1
     punpcklqdq  xmm2, xmm2
     punpcklqdq  xmm3, xmm3
@@ skipping 622 matching lines @@
     packuswb    %1,   %1
 %endm
 
 %macro HORIZx8 1
     mov         rdx, arg(5)                 ;filter ptr
     mov         rsi, arg(0)                 ;src_ptr
     mov         rdi, arg(2)                 ;output_ptr
     mov         rcx, 0x0400040
 
     movdqa      xmm4, [rdx]                 ;load filters
-    movd        xmm5, rcx
+    movq        xmm5, rcx
     packsswb    xmm4, xmm4
     pshuflw     xmm0, xmm4, 0b              ;k0_k1
     pshuflw     xmm1, xmm4, 01010101b       ;k2_k3
     pshuflw     xmm2, xmm4, 10101010b       ;k4_k5
     pshuflw     xmm3, xmm4, 11111111b       ;k6_k7
 
     punpcklqdq  xmm0, xmm0
     punpcklqdq  xmm1, xmm1
     punpcklqdq  xmm2, xmm2
     punpcklqdq  xmm3, xmm3
@@ skipping 83 matching lines @@
 
     movsxd      rax, dword ptr arg(1)       ;src_pixels_per_line
     movsxd      rdx, dword ptr arg(3)       ;output_pitch
     movsxd      rcx, dword ptr arg(4)       ;output_height
 
 .loop:
     prefetcht0  [rsi + 2 * rax -3]
 
     movq        xmm0,   [rsi - 3]           ;load src data
     movq        xmm4,   [rsi + 5]
-    movq        xmm7,   [rsi + 13]
+    movq        xmm6,   [rsi + 13]
     punpcklqdq  xmm0,   xmm4
-    punpcklqdq  xmm4,   xmm7
+    punpcklqdq  xmm4,   xmm6
 
+    movdqa      xmm7,   xmm0
+
+    punpcklbw   xmm7,   xmm7
+    punpckhbw   xmm0,   xmm0
     movdqa      xmm1,   xmm0
     movdqa      xmm2,   xmm0
     movdqa      xmm3,   xmm0
+
+    palignr     xmm0,   xmm7, 1
+    palignr     xmm1,   xmm7, 5
+    pmaddubsw   xmm0,   k0k1
+    palignr     xmm2,   xmm7, 9
+    pmaddubsw   xmm1,   k2k3
+    palignr     xmm3,   xmm7, 13
+
+    pmaddubsw   xmm2,   k4k5
+    pmaddubsw   xmm3,   k6k7
+    paddsw      xmm0,   xmm3
+
+    movdqa      xmm3,   xmm4
+    punpcklbw   xmm3,   xmm3
+    punpckhbw   xmm4,   xmm4
+
     movdqa      xmm5,   xmm4
     movdqa      xmm6,   xmm4
     movdqa      xmm7,   xmm4
 
-    pshufb      xmm0,   [GLOBAL(shuf_t0t1)]
-    pshufb      xmm1,   [GLOBAL(shuf_t2t3)]
-    pshufb      xmm2,   [GLOBAL(shuf_t4t5)]
-    pshufb      xmm3,   [GLOBAL(shuf_t6t7)]
-    pshufb      xmm4,   [GLOBAL(shuf_t0t1)]
-    pshufb      xmm5,   [GLOBAL(shuf_t2t3)]
-    pshufb      xmm6,   [GLOBAL(shuf_t4t5)]
-    pshufb      xmm7,   [GLOBAL(shuf_t6t7)]
+    palignr     xmm4,   xmm3, 1
+    palignr     xmm5,   xmm3, 5
+    palignr     xmm6,   xmm3, 9
+    palignr     xmm7,   xmm3, 13
 
-    pmaddubsw   xmm0,   k0k1
-    pmaddubsw   xmm1,   k2k3
-    pmaddubsw   xmm2,   k4k5
-    pmaddubsw   xmm3,   k6k7
+    movdqa      xmm3,   xmm1
     pmaddubsw   xmm4,   k0k1
+    pmaxsw      xmm1,   xmm2
     pmaddubsw   xmm5,   k2k3
+    pminsw      xmm2,   xmm3
     pmaddubsw   xmm6,   k4k5
+    paddsw      xmm0,   xmm2
     pmaddubsw   xmm7,   k6k7
-
-    paddsw      xmm0,   xmm3
-    movdqa      xmm3,   xmm1
-    pmaxsw      xmm1,   xmm2
-    pminsw      xmm2,   xmm3
-    paddsw      xmm0,   xmm2
     paddsw      xmm0,   xmm1
 
     paddsw      xmm4,   xmm7
     movdqa      xmm7,   xmm5
     pmaxsw      xmm5,   xmm6
     pminsw      xmm6,   xmm7
     paddsw      xmm4,   xmm6
     paddsw      xmm4,   xmm5
 
     paddsw      xmm0,   krd
@@ skipping 240 matching lines @@
     db  0, 1, 1, 2, 2, 3, 3, 4, 4, 5, 5, 6, 6, 7, 7, 8
 align 16
 shuf_t2t3:
     db  2, 3, 3, 4, 4, 5, 5, 6, 6, 7, 7, 8, 8, 9, 9, 10
 align 16
 shuf_t4t5:
     db  4, 5, 5, 6, 6, 7, 7, 8, 8, 9, 9, 10, 10, 11, 11, 12
 align 16
 shuf_t6t7:
     db  6, 7, 7, 8, 8, 9, 9, 10, 10, 11, 11, 12, 12, 13, 13, 14