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

source/libvpx/vp8/common/x86/subpixel_sse2.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 19 matching lines @@
 ;    unsigned int    pixel_step,
 ;    unsigned int    output_height,
 ;    unsigned int    output_width,
 ;    short           *vp8_filter
 ;)
 global sym(vp8_filter_block1d8_h6_sse2)
 sym(vp8_filter_block1d8_h6_sse2):
     push        rbp
     mov         rbp, rsp
     SHADOW_ARGS_TO_STACK 7
-    SAVE_XMM
+    SAVE_XMM 7
     GET_GOT     rbx
     push        rsi
     push        rdi
     ; end prolog
 
         mov         rdx,        arg(6) ;vp8_filter
         mov         rsi,        arg(0) ;src_ptr
 
         mov         rdi,        arg(1) ;output_ptr
 
@@ skipping 99 matching lines @@
 ; Notes: filter_block1d_h6 applies a 6 tap filter horizontally to the input pixels. The
 ; input pixel array has output_height rows. This routine assumes that output_height is an
 ; even number. This function handles 8 pixels in horizontal direction, calculating ONE
 ; rows each iteration to take advantage of the 128 bits operations.
 ;*************************************************************************************/
 global sym(vp8_filter_block1d16_h6_sse2)
 sym(vp8_filter_block1d16_h6_sse2):
     push        rbp
     mov         rbp, rsp
     SHADOW_ARGS_TO_STACK 7
-    SAVE_XMM
+    SAVE_XMM 7
     GET_GOT     rbx
     push        rsi
     push        rdi
     ; end prolog
 
         mov         rdx,        arg(6) ;vp8_filter
         mov         rsi,        arg(0) ;src_ptr
 
         mov         rdi,        arg(1) ;output_ptr
 
@@ skipping 155 matching lines @@
 ;)
 ;/************************************************************************************
 ; Notes: filter_block1d8_v6 applies a 6 tap filter vertically to the input pixels. The
 ; input pixel array has output_height rows.
 ;*************************************************************************************/
 global sym(vp8_filter_block1d8_v6_sse2)
 sym(vp8_filter_block1d8_v6_sse2):
     push        rbp
     mov         rbp, rsp
     SHADOW_ARGS_TO_STACK 8
-    SAVE_XMM
+    SAVE_XMM 7
     GET_GOT     rbx
     push        rsi
     push        rdi
     ; end prolog
 
         mov         rax,        arg(7) ;vp8_filter
         movsxd      rdx,        dword ptr arg(3) ;pixels_per_line
 
         mov         rdi,        arg(1) ;output_ptr
         mov         rsi,        arg(0) ;src_ptr
@@ skipping 74 matching lines @@
 ;)
 ;/************************************************************************************
 ; Notes: filter_block1d16_v6 applies a 6 tap filter vertically to the input pixels. The
 ; input pixel array has output_height rows.
 ;*************************************************************************************/
 global sym(vp8_filter_block1d16_v6_sse2)
 sym(vp8_filter_block1d16_v6_sse2):
     push        rbp
     mov         rbp, rsp
     SHADOW_ARGS_TO_STACK 8
-    SAVE_XMM
+    SAVE_XMM 7
     GET_GOT     rbx
     push        rsi
     push        rdi
     ; end prolog
 
         mov         rax,        arg(7) ;vp8_filter
         movsxd      rdx,        dword ptr arg(3) ;pixels_per_line
 
         mov         rdi,        arg(1) ;output_ptr
         mov         rsi,        arg(0) ;src_ptr
@@ skipping 89 matching lines @@
 ;    int dst_ptich,
 ;    unsigned int    output_height,
 ;    const short    *vp8_filter
 ;)
 ; First-pass filter only when yoffset==0
 global sym(vp8_filter_block1d8_h6_only_sse2)
 sym(vp8_filter_block1d8_h6_only_sse2):
     push        rbp
     mov         rbp, rsp
     SHADOW_ARGS_TO_STACK 6
-    SAVE_XMM
+    SAVE_XMM 7
     GET_GOT     rbx
     push        rsi
     push        rdi
     ; end prolog
 
         mov         rdx,        arg(5) ;vp8_filter
         mov         rsi,        arg(0) ;src_ptr
 
         mov         rdi,        arg(2) ;output_ptr
 
@@ skipping 92 matching lines @@
 ;    int dst_ptich,
 ;    unsigned int    output_height,
 ;    const short    *vp8_filter
 ;)
 ; First-pass filter only when yoffset==0
 global sym(vp8_filter_block1d16_h6_only_sse2)
 sym(vp8_filter_block1d16_h6_only_sse2):
     push        rbp
     mov         rbp, rsp
     SHADOW_ARGS_TO_STACK 6
-    SAVE_XMM
+    SAVE_XMM 7
     GET_GOT     rbx
     push        rsi
     push        rdi
     ; end prolog
 
         mov         rdx,        arg(5) ;vp8_filter
         mov         rsi,        arg(0) ;src_ptr
 
         mov         rdi,        arg(2) ;output_ptr
 
@@ skipping 144 matching lines @@
 ;    int dst_ptich,
 ;    unsigned int output_height,
 ;    const short    *vp8_filter
 ;)
 ; Second-pass filter only when xoffset==0
 global sym(vp8_filter_block1d8_v6_only_sse2)
 sym(vp8_filter_block1d8_v6_only_sse2):
     push        rbp
     mov         rbp, rsp
     SHADOW_ARGS_TO_STACK 6
-    SAVE_XMM
+    SAVE_XMM 7
     GET_GOT     rbx
     push        rsi
     push        rdi
     ; end prolog
 
         mov         rsi,        arg(0) ;src_ptr
         mov         rdi,        arg(2) ;output_ptr
 
         movsxd      rcx,        dword ptr arg(4) ;output_height
         movsxd      rdx,        dword ptr arg(1) ;src_pixels_per_line
@@ skipping 71 matching lines @@
 ;    unsigned short *output_ptr,
 ;    unsigned int    src_pixels_per_line,
 ;    unsigned int    output_height,
 ;    unsigned int    output_width
 ;)
 global sym(vp8_unpack_block1d16_h6_sse2)
 sym(vp8_unpack_block1d16_h6_sse2):
     push        rbp
     mov         rbp, rsp
     SHADOW_ARGS_TO_STACK 5
-    ;SAVE_XMM                          ;xmm6, xmm7 are not used here.
     GET_GOT     rbx
     push        rsi
     push        rdi
     ; end prolog
 
         mov         rsi,        arg(0) ;src_ptr
         mov         rdi,        arg(1) ;output_ptr
 
         movsxd      rcx,        dword ptr arg(3) ;output_height
         movsxd      rax,        dword ptr arg(2) ;src_pixels_per_line            ; Pitch for Source
@@ skipping 19 matching lines @@
 %else
         add         rdi,        r8
 %endif
         dec         rcx
         jnz         unpack_block1d16_h6_sse2_rowloop                ; next row
 
     ; begin epilog
     pop rdi
     pop rsi
     RESTORE_GOT
-    ;RESTORE_XMM
     UNSHADOW_ARGS
     pop         rbp
     ret
 
 
 ;void vp8_bilinear_predict16x16_sse2
 ;(
 ;    unsigned char  *src_ptr,
 ;    int   src_pixels_per_line,
 ;    int  xoffset,
 ;    int  yoffset,
 ;    unsigned char *dst_ptr,
 ;    int dst_pitch
 ;)
 extern sym(vp8_bilinear_filters_mmx)
 global sym(vp8_bilinear_predict16x16_sse2)
 sym(vp8_bilinear_predict16x16_sse2):
     push        rbp
     mov         rbp, rsp
     SHADOW_ARGS_TO_STACK 6
-    SAVE_XMM
+    SAVE_XMM 7
     GET_GOT     rbx
     push        rsi
     push        rdi
     ; end prolog
 
     ;const short *HFilter = bilinear_filters_mmx[xoffset]
     ;const short *VFilter = bilinear_filters_mmx[yoffset]
 
         lea         rcx,        [GLOBAL(sym(vp8_bilinear_filters_mmx))]
         movsxd      rax,        dword ptr arg(2) ;xoffset
@@ skipping 248 matching lines @@
 ;    int  yoffset,
 ;    unsigned char *dst_ptr,
 ;    int dst_pitch
 ;)
 extern sym(vp8_bilinear_filters_mmx)
 global sym(vp8_bilinear_predict8x8_sse2)
 sym(vp8_bilinear_predict8x8_sse2):
     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
 
     ;const short *HFilter = bilinear_filters_mmx[xoffset]
     ;const short *VFilter = bilinear_filters_mmx[yoffset]
@@ skipping 113 matching lines @@
     RESTORE_XMM
     UNSHADOW_ARGS
     pop         rbp
     ret
 
 
 SECTION_RODATA
 align 16
 rd:
     times 8 dw 0x40