sl@0: /*
sl@0:  * LIBOIL - Library of Optimized Inner Loops
sl@0:  * Copyright (c) 2003,2004 David A. Schleef <ds@schleef.org>
sl@0:  * All rights reserved.
sl@0:  *
sl@0:  * Redistribution and use in source and binary forms, with or without
sl@0:  * modification, are permitted provided that the following conditions
sl@0:  * are met:
sl@0:  * 1. Redistributions of source code must retain the above copyright
sl@0:  *    notice, this list of conditions and the following disclaimer.
sl@0:  * 2. Redistributions in binary form must reproduce the above copyright
sl@0:  *    notice, this list of conditions and the following disclaimer in the
sl@0:  *    documentation and/or other materials provided with the distribution.
sl@0:  * 
sl@0:  * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR
sl@0:  * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
sl@0:  * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
sl@0:  * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT,
sl@0:  * INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
sl@0:  * (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
sl@0:  * SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
sl@0:  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
sl@0:  * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING
sl@0:  * IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
sl@0:  * POSSIBILITY OF SUCH DAMAGE.
sl@0:  */
sl@0: //Portions Copyright (c)  2008-2009 Nokia Corporation and/or its subsidiary(-ies). All rights reserved. 
sl@0: 
sl@0: #ifdef HAVE_CONFIG_H
sl@0: #include "config.h"
sl@0: #endif
sl@0: 
sl@0: #include <liboil/liboil.h>
sl@0: #include <liboil/liboilfunction.h>
sl@0: #include <liboil/liboilclasses.h>
sl@0: 
sl@0: 
sl@0: #if 0
sl@0: static int16_t matrix[][4] = {
sl@0:   { 0,      0,      -8192,   -8192  },
sl@0:   { 16384,  0,      0,       0      },
sl@0:   { 0,      16384,  16384,   16384  },
sl@0:   { 0,      0,      -5638,   29032  },
sl@0:   { 0,      22970,  -11700,  0      },
sl@0:   { 0, 0, 0, 0 } 
sl@0: #if 0
sl@0:   { 0, 0, 0, 0 },
sl@0:   { 16384,  0,      0,       0      },
sl@0:   { 0, 16384,  0,      0,        },
sl@0:   { 0, 0, 16384,  0,         },
sl@0:   { 0, 0, 0, 16384,    },
sl@0:   { 0, 0, 0, 0 } 
sl@0: #endif
sl@0: };
sl@0: #endif
sl@0: 
sl@0: static void
sl@0: colorspace_argb_mmx (uint32_t *argb, uint32_t *ayuv, int16_t *matrix, int n)
sl@0: {
sl@0: #if !defined(__WINSCW__) && !defined(__WINS__)      
sl@0:   asm volatile (
sl@0:       "  sub $48, %%esp\n"
sl@0:       "  pshufw $0x1b, 0(%3), %%mm0\n"
sl@0:       "  movq %%mm0, 0(%%esp)\n"
sl@0:       "  pshufw $0x1b, 8(%3), %%mm0\n"
sl@0:       "  movq %%mm0, 8(%%esp)\n"
sl@0:       "  pshufw $0x1b, 16(%3), %%mm0\n"
sl@0:       "  movq %%mm0, 16(%%esp)\n"
sl@0:       "  pshufw $0x1b, 16(%3), %%mm0\n"
sl@0:       "  movq %%mm0, 16(%%esp)\n"
sl@0:       "  pshufw $0x1b, 24(%3), %%mm0\n"
sl@0:       "  movq %%mm0, 24(%%esp)\n"
sl@0:       "  pshufw $0x1b, 32(%3), %%mm0\n"
sl@0:       "  movq %%mm0, 32(%%esp)\n"
sl@0:       "  pshufw $0x1b, 40(%3), %%mm7\n"
sl@0:       "  pxor %%mm0, %%mm0\n"
sl@0:       "  mov $0x08080808, %%eax\n"
sl@0:       "  movd %%eax, %%mm1\n"
sl@0:       "  punpcklbw %%mm0, %%mm1\n"
sl@0:       "  paddw %%mm1, %%mm7\n"
sl@0:       "1:\n"
sl@0:       "  movd (%1), %%mm1\n"
sl@0:       "  add $4, %1\n"
sl@0:       "  punpcklbw %%mm0, %%mm1\n"
sl@0:       "  psllw $6, %%mm1\n"
sl@0:       "  paddw 0(%%esp), %%mm1\n"
sl@0:       "  pshufw $0x00, %%mm1, %%mm4\n"
sl@0:       "  pshufw $0x55, %%mm1, %%mm3\n"
sl@0:       "  pshufw $0xaa, %%mm1, %%mm2\n"
sl@0:       "  pshufw $0xff, %%mm1, %%mm1\n"
sl@0:       "  pmulhw 8(%%esp), %%mm1\n"
sl@0:       "  pmulhw 16(%%esp), %%mm2\n"
sl@0:       "  pmulhw 24(%%esp), %%mm3\n"
sl@0:       "  pmulhw 32(%%esp), %%mm4\n"
sl@0:       "  paddsw %%mm2, %%mm1\n"
sl@0:       "  paddsw %%mm4, %%mm3\n"
sl@0:       "  paddsw %%mm3, %%mm1\n"
sl@0:       "  paddsw %%mm7, %%mm1\n"
sl@0:       "  psraw $4, %%mm1\n"
sl@0:       "  packuswb %%mm0, %%mm1\n"
sl@0:       "  movd %%mm1, (%0)\n"
sl@0:       "  add $4, %0\n"
sl@0:       "  decl %2\n"
sl@0:       "  jne 1b\n"
sl@0:       "  emms\n"
sl@0:       "  add $48, %%esp\n"
sl@0:       : "+r" (argb), "+r" (ayuv), "+r" (n)
sl@0:       : "r" (matrix)
sl@0:       : "eax" );
sl@0: #endif
sl@0: }
sl@0: OIL_DEFINE_IMPL_FULL (colorspace_argb_mmx, colorspace_argb, OIL_IMPL_FLAG_MMX|OIL_IMPL_FLAG_MMXEXT);
sl@0: 
sl@0: 
sl@0: 
sl@0: #ifdef	__SYMBIAN32__
sl@0:  
sl@0: OilFunctionImpl* __oil_function_impl_colorspace_argb_mmx, colorspace_argb() {
sl@0: 		return &_oil_function_impl_colorspace_argb_mmx, colorspace_argb;
sl@0: }
sl@0: #endif
sl@0: