Found-by: ubitux
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

The asm code is not valid for older compilers as it uses too many
operands, ICC on x86_32 seems affected by this.
This patch disables the affected code for ICC on x86_32 and should
make it compileable again.
A better fix would be to use fewer operands or to change this code
to yasm, later is being worked on AFAIK so this is a temporary
solution.

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

These are all part of splited out dsp utils from FFmpeg

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Code originates from:
910b9f30 libavcodec/dsputil.c            (David Conrad        2010-05-27 04:39:27 +0000  334) void ff_emulated_edge_mc(uint8_t *buf, const uint8_t *src, int linesize, int block_w, int block_h,
93a21abd libavcodec/mpegvideo.c          (Michael Niedermayer 2002-07-14 18:37:35 +0000  335)                                     int src_x, int src_y, int w, int h){
93a21abd libavcodec/mpegvideo.c          (Michael Niedermayer 2002-07-14 18:37:35 +0000  336)     int x, y;
93a21abd libavcodec/mpegvideo.c          (Michael Niedermayer 2002-07-14 18:37:35 +0000  337)     int start_y, start_x, end_y, end_x;
b5a093b3 libavcodec/mpegvideo.c          (Michael Niedermayer 2002-07-25 20:22:36 +0000  338)
93a21abd libavcodec/mpegvideo.c          (Michael Niedermayer 2002-07-14 18:37:35 +0000  339)     if(src_y>= h){
93a21abd libavcodec/mpegvideo.c          (Michael Niedermayer 2002-07-14 18:37:35 +0000  340)         src+= (h-1-src_y)*linesize;
93a21abd libavcodec/mpegvideo.c          (Michael Niedermayer 2002-07-14 18:37:35 +0000  341)         src_y=h-1;
225f9c44 libavcodec/mpegvideo.c          (Michael Niedermayer 2002-07-15 00:25:53 +0000  342)     }else if(src_y<=-block_h){
225f9c44 libavcodec/mpegvideo.c          (Michael Niedermayer 2002-07-15 00:25:53 +0000  343)         src+= (1-block_h-src_y)*linesize;
225f9c44 libavcodec/mpegvideo.c          (Michael Niedermayer 2002-07-15 00:25:53 +0000  344)         src_y=1-block_h;
93a21abd libavcodec/mpegvideo.c          (Michael Niedermayer 2002-07-14 18:37:35 +0000  345)     }
93a21abd libavcodec/mpegvideo.c          (Michael Niedermayer 2002-07-14 18:37:35 +0000  346)     if(src_x>= w){
93a21abd libavcodec/mpegvideo.c          (Michael Niedermayer 2002-07-14 18:37:35 +0000  347)         src+= (w-1-src_x);
93a21abd libavcodec/mpegvideo.c          (Michael Niedermayer 2002-07-14 18:37:35 +0000  348)         src_x=w-1;
225f9c44 libavcodec/mpegvideo.c          (Michael Niedermayer 2002-07-15 00:25:53 +0000  349)     }else if(src_x<=-block_w){
225f9c44 libavcodec/mpegvideo.c          (Michael Niedermayer 2002-07-15 00:25:53 +0000  350)         src+= (1-block_w-src_x);
225f9c44 libavcodec/mpegvideo.c          (Michael Niedermayer 2002-07-15 00:25:53 +0000  351)         src_x=1-block_w;
93a21abd libavcodec/mpegvideo.c          (Michael Niedermayer 2002-07-14 18:37:35 +0000  352)     }
93a21abd libavcodec/mpegvideo.c          (Michael Niedermayer 2002-07-14 18:37:35 +0000  353)
b8a78f41 libavcodec/mpegvideo.c          (Michael Niedermayer 2002-11-10 11:46:59 +0000  354)     start_y= FFMAX(0, -src_y);
b8a78f41 libavcodec/mpegvideo.c          (Michael Niedermayer 2002-11-10 11:46:59 +0000  355)     start_x= FFMAX(0, -src_x);
b8a78f41 libavcodec/mpegvideo.c          (Michael Niedermayer 2002-11-10 11:46:59 +0000  356)     end_y= FFMIN(block_h, h-src_y);
b8a78f41 libavcodec/mpegvideo.c          (Michael Niedermayer 2002-11-10 11:46:59 +0000  357)     end_x= FFMIN(block_w, w-src_x);

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Move some functions from dsputil. The idea is that videodsp contains
functions that are useful for a large and varied set of video decoders.
Currently, it contains emulated_edge_mc() and prefetch().

Signed-off-by: Luca Barbato <lu_zero@gentoo.org>

Use this in VP8/H264-8bit loopfilter functions so they can be used if
there is no aligned stack (e.g. MSVC 32bit or ICC 10.x).

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Use this in VP8/H264-8bit loopfilter functions so they can be used if
there is no aligned stack (e.g. MSVC 32bit or ICC 10.x).

Signed-off-by: Luca Barbato <lu_zero@gentoo.org>

According to lavc/proresdsp.c, both prores and prores-lgpl decoders need
lavc/x86/proresdsp_init.c:ff_proresdsp_x86_init().

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Core I7 (Sandy Bridge) 135 to 107 cycles
Core i5 (Arrandale) 162 to 142 (Thanks to Christophe Gisquet for testing)

Reviewed-by: Christophe Gisquet <christophe.gisquet@gmail.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Start and end index are multiple of 2, therefore guaranteeing aligned access.
Also, this allows to generate 4 floats per loop, keeping the alignment all
along.

Timing:
- 32 bits: 326c -> 172c
- 64 bits: 323c -> 156c

Signed-off-by: Diego Biurrun <diego@biurrun.de>

This code never did anything as far as i can remember

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Fixes assertion failure

Found-by: Mateusz "j00ru" Jurczyk and Gynvael Coldwind
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Some optimized functions reference optimized symbols, so the functions
must be explicitly disabled when those symbols are unavailable.

Fix compile failure without yasm

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

The qpel functions referenced here are not related to h264 and should
thus never have been under CONFIG_H264QPEL.

Signed-off-by: Mans Rullgard <mans@mansr.com>
Signed-off-by: Diego Biurrun <diego@biurrun.de>

Fixes xmm register clobbering on win64.

Signed-off-by: Diego Biurrun <diego@biurrun.de>

The only CPUs that have 3dnow and don't have mmxext are 12 years old.
Moreover, AMD has dropped 3dnow extensions from newer CPUs.

Signed-off-by: Diego Biurrun <diego@biurrun.de>

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Function names changed after switching to declaration with
PRED4x4/8x8/8x8L/16x16 macros in the C code.

Found-by: Mateusz "j00ru" Jurczyk and Gynvael Coldwind
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

An assembler able to cope with AVX instructions is now required.

Fixes out of array accesses

Found-by: Mateusz "j00ru" Jurczyk and Gynvael Coldwind
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

This fixes compilation with --disable-everything and components enabled.

It is only used in one place so there is no need for the abstraction.

This simplifies cpuflags porting.