Update libvpx snapshot to v0.9.7-p1 (Cayuga).

source/libvpx/vp8/common/x86/subpixel_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.
 ;
 
@@ skipping 21 matching lines @@
 ;    unsigned char *output_ptr,
 ;    unsigned int    output_pitch,
 ;    unsigned int    output_height,
 ;    unsigned int    vp8_filter_index
 ;)
 global sym(vp8_filter_block1d8_h6_ssse3)
 sym(vp8_filter_block1d8_h6_ssse3):
     push        rbp
     mov         rbp, rsp
     SHADOW_ARGS_TO_STACK 6
+    SAVE_XMM 7
     GET_GOT     rbx
     push        rsi
     push        rdi
     ; end prolog
 
     movsxd      rdx, DWORD PTR arg(5)   ;table index
     xor         rsi, rsi
     shl         rdx, 4
 
     movdqa      xmm7, [GLOBAL(rd)]
@@ skipping 48 matching lines @@
 
     packuswb    xmm0,   xmm0
 
     movq        MMWORD Ptr [rdi], xmm0
     jnz         filter_block1d8_h6_rowloop_ssse3
 
     ; begin epilog
     pop rdi
     pop rsi
     RESTORE_GOT
+    RESTORE_XMM
     UNSHADOW_ARGS
     pop         rbp
     ret
 
 vp8_filter_block1d8_h4_ssse3:
     movdqa      xmm5, XMMWORD PTR [rax+256]     ;k2_k4
     movdqa      xmm6, XMMWORD PTR [rax+128]     ;k1_k3
 
     movdqa      xmm3, XMMWORD PTR [GLOBAL(shuf2bfrom1)]
     movdqa      xmm4, XMMWORD PTR [GLOBAL(shuf3bfrom1)]
@@ skipping 35 matching lines @@
     packuswb    xmm0,   xmm0
 
     movq        MMWORD Ptr [rdi], xmm0
 
     jnz         filter_block1d8_h4_rowloop_ssse3
 
     ; begin epilog
     pop rdi
     pop rsi
     RESTORE_GOT
+    RESTORE_XMM
     UNSHADOW_ARGS
     pop         rbp
     ret
 ;void vp8_filter_block1d16_h6_ssse3
 ;(
 ;    unsigned char  *src_ptr,
 ;    unsigned int    src_pixels_per_line,
 ;    unsigned char  *output_ptr,
 ;    unsigned int    output_pitch,
 ;    unsigned int    output_height,
 ;    unsigned int    vp8_filter_index
 ;)
 global sym(vp8_filter_block1d16_h6_ssse3)
 sym(vp8_filter_block1d16_h6_ssse3):
     push        rbp
     mov         rbp, rsp
     SHADOW_ARGS_TO_STACK 6
-    SAVE_XMM
+    SAVE_XMM 7
     GET_GOT     rbx
     push        rsi
     push        rdi
     ; end prolog
 
     movsxd      rdx, DWORD PTR arg(5)           ;table index
     xor         rsi, rsi
     shl         rdx, 4      ;
 
     lea         rax, [GLOBAL(k0_k5)]
     add         rax, rdx
 
     mov         rdi, arg(2)                     ;output_ptr
 
-;;
-;;    cmp         esi, DWORD PTR [rax]
-;;    je          vp8_filter_block1d16_h4_ssse3
-
     mov         rsi, arg(0)                     ;src_ptr
 
     movdqa      xmm4, XMMWORD PTR [rax]         ;k0_k5
     movdqa      xmm5, XMMWORD PTR [rax+256]     ;k2_k4
     movdqa      xmm6, XMMWORD PTR [rax+128]     ;k1_k3
 
     movsxd      rax, dword ptr arg(1)           ;src_pixels_per_line
     movsxd      rcx, dword ptr arg(4)           ;output_height
     movsxd      rdx, dword ptr arg(3)           ;output_pitch
 
@@ skipping 53 matching lines @@
     movdqa      XMMWORD Ptr [rdi], xmm0
 
     lea         rdi,    [rdi + rdx]
     dec         rcx
     jnz         filter_block1d16_h6_rowloop_ssse3
 
     ; begin epilog
     pop rdi
     pop rsi
     RESTORE_GOT
-    UNSHADOW_ARGS
-    pop         rbp
-    ret
-
-vp8_filter_block1d16_h4_ssse3:
-    movdqa      xmm5, XMMWORD PTR [rax+256]     ;k2_k4
-    movdqa      xmm6, XMMWORD PTR [rax+128]     ;k1_k3
-
-    mov         rsi, arg(0)             ;src_ptr
-    movsxd      rax, dword ptr arg(1)   ;src_pixels_per_line
-    movsxd      rcx, dword ptr arg(4)   ;output_height
-    movsxd      rdx, dword ptr arg(3)   ;output_pitch
-
-filter_block1d16_h4_rowloop_ssse3:
-    movdqu      xmm1,   XMMWORD PTR [rsi - 2]
-
-    movdqa      xmm2, xmm1
-    pshufb      xmm1, [GLOBAL(shuf2b)]
-    pshufb      xmm2, [GLOBAL(shuf3b)]
-    pmaddubsw   xmm1, xmm5
-
-    movdqu      xmm3,   XMMWORD PTR [rsi + 6]
-
-    pmaddubsw   xmm2, xmm6
-    movdqa      xmm0, xmm3
-    pshufb      xmm3, [GLOBAL(shuf3b)]
-    pshufb      xmm0, [GLOBAL(shuf2b)]
-
-    paddsw      xmm1, [GLOBAL(rd)]
-    paddsw      xmm1, xmm2
-
-    pmaddubsw   xmm0, xmm5
-    pmaddubsw   xmm3, xmm6
-
-    psraw       xmm1, 7
-    packuswb    xmm1, xmm1
-    lea         rsi,    [rsi + rax]
-    paddsw      xmm3, xmm0
-    paddsw      xmm3, [GLOBAL(rd)]
-    psraw       xmm3, 7
-    packuswb    xmm3, xmm3
-
-    punpcklqdq  xmm1, xmm3
-
-    movdqa      XMMWORD Ptr [rdi], xmm1
-
-    add         rdi, rdx
-    dec         rcx
-    jnz         filter_block1d16_h4_rowloop_ssse3
-
-
-    ; begin epilog
-    pop rdi
-    pop rsi
-    RESTORE_GOT
+    RESTORE_XMM
     UNSHADOW_ARGS
     pop         rbp
     ret
 
 ;void vp8_filter_block1d4_h6_ssse3
 ;(
 ;    unsigned char  *src_ptr,
 ;    unsigned int    src_pixels_per_line,
 ;    unsigned char  *output_ptr,
 ;    unsigned int    output_pitch,
 ;    unsigned int    output_height,
 ;    unsigned int    vp8_filter_index
 ;)
 global sym(vp8_filter_block1d4_h6_ssse3)
 sym(vp8_filter_block1d4_h6_ssse3):
     push        rbp
     mov         rbp, rsp
     SHADOW_ARGS_TO_STACK 6
+    SAVE_XMM 7
     GET_GOT     rbx
     push        rsi
     push        rdi
     ; end prolog
 
     movsxd      rdx, DWORD PTR arg(5)   ;table index
     xor         rsi, rsi
     shl         rdx, 4      ;
 
     lea         rax, [GLOBAL(k0_k5)]
@@ skipping 87 matching lines @@
     movd        DWORD PTR [rdi], xmm1
 
     add         rdi, rdx
     dec         rcx
     jnz         filter_block1d4_h4_rowloop_ssse3
 
     ; begin epilog
     pop rdi
     pop rsi
     RESTORE_GOT
+    RESTORE_XMM
     UNSHADOW_ARGS
     pop         rbp
     ret
 
 
 
 ;void vp8_filter_block1d16_v6_ssse3
 ;(
 ;    unsigned char *src_ptr,
 ;    unsigned int   src_pitch,
 ;    unsigned char *output_ptr,
 ;    unsigned int   out_pitch,
 ;    unsigned int   output_height,
 ;    unsigned int   vp8_filter_index
 ;)
 global sym(vp8_filter_block1d16_v6_ssse3)
 sym(vp8_filter_block1d16_v6_ssse3):
     push        rbp
     mov         rbp, rsp
     SHADOW_ARGS_TO_STACK 6
+    SAVE_XMM 7
     GET_GOT     rbx
     push        rsi
     push        rdi
     ; end prolog
 
     movsxd      rdx, DWORD PTR arg(5)   ;table index
     xor         rsi, rsi
     shl         rdx, 4      ;
 
     lea         rax, [GLOBAL(k0_k5)]
@@ skipping 75 matching lines @@
 %else
     add         rdi,        r8
 %endif
     dec         rcx
     jnz         vp8_filter_block1d16_v6_ssse3_loop
 
     ; begin epilog
     pop rdi
     pop rsi
     RESTORE_GOT
+    RESTORE_XMM
     UNSHADOW_ARGS
     pop         rbp
     ret
 
 vp8_filter_block1d16_v4_ssse3:
     movdqa      xmm6, XMMWORD PTR [rax+256]     ;k2_k4
     movdqa      xmm7, XMMWORD PTR [rax+128]     ;k1_k3
 
     mov         rsi, arg(0)             ;src_ptr
     movsxd      rdx, DWORD PTR arg(1)   ;pixels_per_line
@@ skipping 52 matching lines @@
 %else
     add         rdi,        r8
 %endif
     dec         rcx
     jnz         vp8_filter_block1d16_v4_ssse3_loop
 
     ; begin epilog
     pop rdi
     pop rsi
     RESTORE_GOT
+    RESTORE_XMM
     UNSHADOW_ARGS
     pop         rbp
     ret
 
 ;void vp8_filter_block1d8_v6_ssse3
 ;(
 ;    unsigned char *src_ptr,
 ;    unsigned int   src_pitch,
 ;    unsigned char *output_ptr,
 ;    unsigned int   out_pitch,
 ;    unsigned int   output_height,
 ;    unsigned int   vp8_filter_index
 ;)
 global sym(vp8_filter_block1d8_v6_ssse3)
 sym(vp8_filter_block1d8_v6_ssse3):
     push        rbp
     mov         rbp, rsp
     SHADOW_ARGS_TO_STACK 6
+    SAVE_XMM 7
     GET_GOT     rbx
     push        rsi
     push        rdi
     ; end prolog
 
     movsxd      rdx, DWORD PTR arg(5)   ;table index
     xor         rsi, rsi
     shl         rdx, 4      ;
 
     lea         rax, [GLOBAL(k0_k5)]
@@ skipping 52 matching lines @@
 %else
     add         rdi,        r8
 %endif
     dec         rcx
     jnz         vp8_filter_block1d8_v6_ssse3_loop
 
     ; begin epilog
     pop rdi
     pop rsi
     RESTORE_GOT
+    RESTORE_XMM
     UNSHADOW_ARGS
     pop         rbp
     ret
 
 vp8_filter_block1d8_v4_ssse3:
     movdqa      xmm6, XMMWORD PTR [rax+256]     ;k2_k4
     movdqa      xmm7, XMMWORD PTR [rax+128]     ;k1_k3
     movdqa      xmm5, [GLOBAL(rd)]
 
     mov         rsi, arg(0)             ;src_ptr
@@ skipping 28 matching lines @@
 %else
     add         rdi,        r8
 %endif
     dec         rcx
     jnz         vp8_filter_block1d8_v4_ssse3_loop
 
     ; begin epilog
     pop rdi
     pop rsi
     RESTORE_GOT
+    RESTORE_XMM
     UNSHADOW_ARGS
     pop         rbp
     ret
 ;void vp8_filter_block1d4_v6_ssse3
 ;(
 ;    unsigned char *src_ptr,
 ;    unsigned int   src_pitch,
 ;    unsigned char *output_ptr,
 ;    unsigned int   out_pitch,
 ;    unsigned int   output_height,
@@ skipping 136 matching lines @@
 ;    int  xoffset,
 ;    int  yoffset,
 ;    unsigned char *dst_ptr,
 ;    int dst_pitch
 ;)
 global sym(vp8_bilinear_predict16x16_ssse3)
 sym(vp8_bilinear_predict16x16_ssse3):
     push        rbp
     mov         rbp, rsp
     SHADOW_ARGS_TO_STACK 6
-    SAVE_XMM
+    SAVE_XMM 7
     GET_GOT     rbx
     push        rsi
     push        rdi
     ; end prolog
 
         lea         rcx,        [GLOBAL(vp8_bilinear_filters_ssse3)]
         movsxd      rax,        dword ptr arg(2)    ; xoffset
 
         cmp         rax,        0                   ; skip first_pass filter if xoffset=0
         je          b16x16_sp_only
@@ skipping 242 matching lines @@
 ;    int  xoffset,
 ;    int  yoffset,
 ;    unsigned char *dst_ptr,
 ;    int dst_pitch
 ;)
 global sym(vp8_bilinear_predict8x8_ssse3)
 sym(vp8_bilinear_predict8x8_ssse3):
     push        rbp
     mov         rbp, rsp
     SHADOW_ARGS_TO_STACK 6
-    SAVE_XMM
+    SAVE_XMM 7
     GET_GOT     rbx
     push        rsi
     push        rdi
     ; end prolog
 
     ALIGN_STACK 16, rax
     sub         rsp, 144                         ; reserve 144 bytes
 
         lea         rcx,        [GLOBAL(vp8_bilinear_filters_ssse3)]
 
@@ skipping 336 matching lines @@
 vp8_bilinear_filters_ssse3:
     times 8 db 128, 0
     times 8 db 112, 16
     times 8 db 96,  32
     times 8 db 80,  48
     times 8 db 64,  64
     times 8 db 48,  80
     times 8 db 32,  96
     times 8 db 16,  112