Port latest x264 deblock asm (before they moved to using NV12 as internal

format), LGPL'ed with permission from Jason and Loren. This includes mmx2
code, so remove inline asm from h264dsp_mmx.c accordingly.

Originally committed as revision 25031 to svn://svn.ffmpeg.org/ffmpeg/trunk

libavcodec/x86/Makefile

@@ -9,7 +9,7 @@ YASM-OBJS-$(CONFIG_FFT)                += x86/fft_mmx.o                 \
 YASM-OBJS-$(CONFIG_GPL)                += x86/h264_idct_sse2.o          \
 
 MMX-OBJS-$(CONFIG_H264DSP)             += x86/h264dsp_mmx.o
-YASM-OBJS-$(CONFIG_H264DSP)            += x86/h264_deblock_sse2.o       \
+YASM-OBJS-$(CONFIG_H264DSP)            += x86/h264_deblock.o            \
                                           x86/h264_weight.o             \
 
 YASM-OBJS-$(CONFIG_H264PRED)           += x86/h264_intrapred.o

libavcodec/x86/dsputil_mmx.c

@@ -60,6 +60,7 @@ DECLARE_ALIGNED(8,  const uint64_t, ff_pw_96 ) = 0x0060006000600060ULL;
 DECLARE_ALIGNED(8,  const uint64_t, ff_pw_128) = 0x0080008000800080ULL;
 DECLARE_ALIGNED(8,  const uint64_t, ff_pw_255) = 0x00ff00ff00ff00ffULL;
 
+DECLARE_ALIGNED(16, const xmm_reg,  ff_pb_0  ) = {0x0000000000000000ULL, 0x0000000000000000ULL};
 DECLARE_ALIGNED(16, const xmm_reg,  ff_pb_1  ) = {0x0101010101010101ULL, 0x0101010101010101ULL};
 DECLARE_ALIGNED(16, const xmm_reg,  ff_pb_3  ) = {0x0303030303030303ULL, 0x0303030303030303ULL};
 DECLARE_ALIGNED(16, const xmm_reg,  ff_pb_4  ) = {0x0404040404040404ULL, 0x0404040404040404ULL};
@@ -68,7 +69,7 @@ DECLARE_ALIGNED(8,  const uint64_t, ff_pb_1F ) = 0x1F1F1F1F1F1F1F1FULL;
 DECLARE_ALIGNED(8,  const uint64_t, ff_pb_3F ) = 0x3F3F3F3F3F3F3F3FULL;
 DECLARE_ALIGNED(16, const xmm_reg,  ff_pb_80 ) = {0x8080808080808080ULL, 0x8080808080808080ULL};
 DECLARE_ALIGNED(8,  const uint64_t, ff_pb_81 ) = 0x8181818181818181ULL;
-DECLARE_ALIGNED(8,  const uint64_t, ff_pb_A1 ) = 0xA1A1A1A1A1A1A1A1ULL;
+DECLARE_ALIGNED(16, const xmm_reg,  ff_pb_A1 ) = {0xA1A1A1A1A1A1A1A1ULL, 0xA1A1A1A1A1A1A1A1ULL};
 DECLARE_ALIGNED(16, const xmm_reg,  ff_pb_F8 ) = {0xF8F8F8F8F8F8F8F8ULL, 0xF8F8F8F8F8F8F8F8ULL};
 DECLARE_ALIGNED(8,  const uint64_t, ff_pb_FC ) = 0xFCFCFCFCFCFCFCFCULL;
 DECLARE_ALIGNED(16, const xmm_reg,  ff_pb_FE ) = {0xFEFEFEFEFEFEFEFEULL, 0xFEFEFEFEFEFEFEFEULL};

libavcodec/x86/dsputil_mmx.h

@@ -57,7 +57,7 @@ extern const uint64_t ff_pb_7;
 extern const uint64_t ff_pb_1F;
 extern const uint64_t ff_pb_3F;
 extern const uint64_t ff_pb_81;
-extern const uint64_t ff_pb_A1;
+extern const xmm_reg  ff_pb_A1;
 extern const xmm_reg  ff_pb_F8;
 extern const uint64_t ff_pb_FC;
 extern const xmm_reg  ff_pb_FE;