@@ -115,8 +115,7 @@ config.h: .config

 SUBDIR_VARS := CLEANFILES EXAMPLES FFLIBS HOSTPROGS TESTPROGS TOOLS      \

                HEADERS ARCH_HEADERS BUILT_HEADERS SKIPHEADERS            \

                ARMV5TE-OBJS ARMV6-OBJS VFP-OBJS NEON-OBJS                \

-               ALTIVEC-OBJS VIS-OBJS                                     \

-               MMX-OBJS YASM-OBJS                                        \

+               ALTIVEC-OBJS MMX-OBJS YASM-OBJS                           \

                OBJS HOSTOBJS TESTOBJS

 define RESET

@@ -5,7 +5,5 @@ OBJS-$(HAVE_NEON)    += $(NEON-OBJS)    $(NEON-OBJS-yes)

 OBJS-$(HAVE_ALTIVEC) += $(ALTIVEC-OBJS) $(ALTIVEC-OBJS-yes)

-OBJS-$(HAVE_VIS)     += $(VIS-OBJS)     $(VIS-OBJS-yes)

-

 OBJS-$(HAVE_MMX)     += $(MMX-OBJS)     $(MMX-OBJS-yes)

 OBJS-$(HAVE_YASM)    += $(YASM-OBJS)    $(YASM-OBJS-yes)

@@ -284,7 +284,6 @@ Optimization options (experts only):

   --disable-armv6t2        disable armv6t2 optimizations

   --disable-vfp            disable VFP optimizations

   --disable-neon           disable NEON optimizations

-  --disable-vis            disable VIS optimizations

   --disable-inline-asm     disable use of inline assembler

   --disable-yasm           disable use of yasm assembler

@@ -1294,7 +1293,6 @@ ARCH_EXT_LIST="

     $ARCH_EXT_LIST_X86

     altivec

     ppc4xx

-    vis

 "

 HAVE_LIST_CMDLINE="

@@ -1584,8 +1582,6 @@ map 'eval ${v}_inline_deps=inline_asm' $ARCH_EXT_LIST_ARM

 altivec_deps="ppc"

 ppc4xx_deps="ppc"

-vis_deps="sparc"

-

 cpunop_deps="i686"

 x86_64_select="i686"

 x86_64_suggest="fast_cmov"

@@ -3151,7 +3147,6 @@ elif enabled sparc; then

     case $cpu in

         cypress|f93[04]|tsc701|sparcl*|supersparc|hypersparc|niagara|v[789])

             cpuflags="-mcpu=$cpu"

-            disable vis

         ;;

         ultrasparc*|niagara[234])

             cpuflags="-mcpu=$cpu"

@@ -3740,10 +3735,6 @@ EOF

         enabled altivec || warn "Altivec disabled, possibly missing --cpu flag"

     fi

-elif enabled sparc; then

-

-    enabled vis && check_inline_asm vis '"pdist %f0, %f0, %f0"'

-

 elif enabled x86; then

     check_builtin rdtsc    intrin.h   "__rdtsc()"

@@ -4351,9 +4342,6 @@ if enabled ppc; then

     echo "PPC 4xx optimizations     ${ppc4xx-no}"

     echo "dcbzl available           ${dcbzl-no}"

 fi

-if enabled sparc; then

-    echo "VIS enabled               ${vis-no}"

-fi

 echo "debug symbols             ${debug-no}"

 echo "optimize for size         ${small-no}"

 echo "optimizations             ${optimizations-no}"

@@ -268,17 +268,6 @@ CELL/SPU:

 http://www-01.ibm.com/chips/techlib/techlib.nsf/techdocs/30B3520C93F437AB87257060006FFE5E/$file/Language_Extensions_for_CBEA_2.4.pdf

 http://www-01.ibm.com/chips/techlib/techlib.nsf/techdocs/9F820A5FFA3ECE8C8725716A0062585F/$file/CBE_Handbook_v1.1_24APR2007_pub.pdf

-SPARC-specific:

-SPARC Joint Programming Specification (JPS1): Commonality

-http://www.fujitsu.com/downloads/PRMPWR/JPS1-R1.0.4-Common-pub.pdf

-

-UltraSPARC III Processor User's Manual (contains instruction timings)

-http://www.sun.com/processors/manuals/USIIIv2.pdf

-

-VIS Whitepaper (contains optimization guidelines)

-http://www.sun.com/processors/vis/download/vis/vis_whitepaper.pdf

-

 GCC asm links:

 --------------

 official doc but quite ugly

@@ -2468,7 +2468,9 @@ typedef struct AVCodecContext {

 #define FF_IDCT_XVIDMMX       14

 #define FF_IDCT_SIMPLEARMV5TE 16

 #define FF_IDCT_SIMPLEARMV6   17

+#if FF_API_ARCH_SPARC

 #define FF_IDCT_SIMPLEVIS     18

+#endif

 #define FF_IDCT_FAAN          20

 #define FF_IDCT_SIMPLENEON    22

 #if FF_API_ARCH_ALPHA

@@ -2653,8 +2653,6 @@ av_cold void ff_dsputil_init(DSPContext* c, AVCodecContext *avctx)

         ff_dsputil_init_bfin(c, avctx);

     if (ARCH_PPC)

         ff_dsputil_init_ppc(c, avctx);

-    if (HAVE_VIS)

-        ff_dsputil_init_vis(c, avctx);

     if (ARCH_X86)

         ff_dsputil_init_x86(c, avctx);

@@ -300,7 +300,6 @@ void ff_set_cmp(DSPContext* c, me_cmp_func *cmp, int type);

 void ff_dsputil_init_arm(DSPContext* c, AVCodecContext *avctx);

 void ff_dsputil_init_bfin(DSPContext* c, AVCodecContext *avctx);

 void ff_dsputil_init_ppc(DSPContext* c, AVCodecContext *avctx);

-void ff_dsputil_init_vis(DSPContext* c, AVCodecContext *avctx);

 void ff_dsputil_init_x86(DSPContext* c, AVCodecContext *avctx);

 #endif /* AVCODEC_DSPUTIL_H */

@@ -62,8 +62,6 @@ av_cold void ff_hpeldsp_init(HpelDSPContext *c, int flags)

         ff_hpeldsp_init_bfin(c, flags);

     if (ARCH_PPC)

         ff_hpeldsp_init_ppc(c, flags);

-    if (HAVE_VIS)

-        ff_hpeldsp_init_vis(c, flags);

     if (ARCH_X86)

         ff_hpeldsp_init_x86(c, flags);

 }

@@ -98,7 +98,6 @@ void ff_hpeldsp_init_aarch64(HpelDSPContext *c, int flags);

 void ff_hpeldsp_init_arm(HpelDSPContext *c, int flags);

 void ff_hpeldsp_init_bfin(HpelDSPContext *c, int flags);

 void ff_hpeldsp_init_ppc(HpelDSPContext *c, int flags);

-void ff_hpeldsp_init_vis(HpelDSPContext *c, int flags);

 void ff_hpeldsp_init_x86(HpelDSPContext *c, int flags);

 #endif /* AVCODEC_HPELDSP_H */

deleted file mode 100644

@@ -1,4 +0,0 @@

-VIS-OBJS += sparc/dsputil_vis.o                                         \

-            sparc/simple_idct_vis.o                                     \

-

-VIS-OBJS-$(CONFIG_HPELDSP) += sparc/hpeldsp_vis.o

deleted file mode 100644

@@ -1,40 +0,0 @@

-/*

- * Copyright (C) 2003 David S. Miller <davem@redhat.com>

- *

- * This file is part of Libav.

- *

- * Libav is free software; you can redistribute it and/or

- * modify it under the terms of the GNU Lesser General Public

- * License as published by the Free Software Foundation; either

- * version 2.1 of the License, or (at your option) any later version.

- *

- * Libav is distributed in the hope that it will be useful,

- * but WITHOUT ANY WARRANTY; without even the implied warranty of

- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU

- * Lesser General Public License for more details.

- *

- * You should have received a copy of the GNU Lesser General Public

- * License along with Libav; if not, write to the Free Software

- * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA

- */

-

-#include "libavutil/attributes.h"

-#include "libavcodec/dsputil.h"

-#include "dsputil_vis.h"

-#include "vis.h"

-

-av_cold void ff_dsputil_init_vis(DSPContext *c, AVCodecContext *avctx)

-{

-  /* VIS-specific optimizations */

-  int accel = vis_level ();

-  const int high_bit_depth = avctx->bits_per_raw_sample > 8;

-

-  if (accel & ACCEL_SPARC_VIS && !high_bit_depth) {

-      if (avctx->idct_algo == FF_IDCT_SIMPLEVIS) {

-          c->idct_put = ff_simple_idct_put_vis;

-          c->idct_add = ff_simple_idct_add_vis;

-          c->idct     = ff_simple_idct_vis;

-          c->idct_permutation_type = FF_TRANSPOSE_IDCT_PERM;

-      }

-  }

-}

deleted file mode 100644

@@ -1,28 +0,0 @@

-/*

- * This file is part of Libav.

- *

- * Libav is free software; you can redistribute it and/or

- * modify it under the terms of the GNU Lesser General Public

- * License as published by the Free Software Foundation; either

- * version 2.1 of the License, or (at your option) any later version.

- *

- * Libav is distributed in the hope that it will be useful,

- * but WITHOUT ANY WARRANTY; without even the implied warranty of

- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU

- * Lesser General Public License for more details.

- *

- * You should have received a copy of the GNU Lesser General Public

- * License along with Libav; if not, write to the Free Software

- * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA

- */

-

-#ifndef AVCODEC_SPARC_DSPUTIL_VIS_H

-#define AVCODEC_SPARC_DSPUTIL_VIS_H

-

-#include <stdint.h>

-

-void ff_simple_idct_put_vis(uint8_t *dest, int line_size, int16_t *data);

-void ff_simple_idct_add_vis(uint8_t *dest, int line_size, int16_t *data);

-void ff_simple_idct_vis(int16_t *data);

-

-#endif /* AVCODEC_SPARC_DSPUTIL_VIS_H */

deleted file mode 100644

@@ -1,3524 +0,0 @@

-/*

- * Copyright (C) 2003 David S. Miller <davem@redhat.com>

- *

- * This file is part of Libav.

- *

- * Libav is free software; you can redistribute it and/or

- * modify it under the terms of the GNU Lesser General Public

- * License as published by the Free Software Foundation; either

- * version 2.1 of the License, or (at your option) any later version.

- *

- * Libav is distributed in the hope that it will be useful,

- * but WITHOUT ANY WARRANTY; without even the implied warranty of

- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU

- * Lesser General Public License for more details.

- *

- * You should have received a copy of the GNU Lesser General Public

- * License along with Libav; if not, write to the Free Software

- * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA

- */

-

-/* The *no_round* functions have been added by James A. Morrison, 2003,2004.

-   The vis code from libmpeg2 was adapted for libavcodec by James A. Morrison.

- */

-

-#include <stddef.h>

-#include <stdint.h>

-

-#include "libavutil/attributes.h"

-#include "libavutil/mem.h"

-#include "libavcodec/hpeldsp.h"

-#include "vis.h"

-

-/* The trick used in some of this file is the formula from the MMX

- * motion comp code, which is:

- *

- * (x+y+1)>>1 == (x|y)-((x^y)>>1)

- *

- * This allows us to average 8 bytes at a time in a 64-bit FPU reg.

- * We avoid overflows by masking before we do the shift, and we

- * implement the shift by multiplying by 1/2 using mul8x16.  So in

- * VIS this is (assume 'x' is in f0, 'y' is in f2, a repeating mask

- * of '0xfe' is in f4, a repeating mask of '0x7f' is in f6, and

- * the value 0x80808080 is in f8):

- *

- *      fxor            f0,   f2, f10

- *      fand            f10,  f4, f10

- *      fmul8x16        f8,  f10, f10

- *      fand            f10,  f6, f10

- *      for             f0,   f2, f12

- *      fpsub16         f12, f10, f10

- */

-

-#define DUP4(x) {x, x, x, x}

-#define DUP8(x) {x, x, x, x, x, x, x, x}

-DECLARE_ALIGNED(8, static const int16_t, constants1)[] = DUP4 (1);

-DECLARE_ALIGNED(8, static const int16_t, constants2)[] = DUP4 (2);

-DECLARE_ALIGNED(8, static const int16_t, constants3)[] = DUP4 (3);

-DECLARE_ALIGNED(8, static const int16_t, constants6)[] = DUP4 (6);

-DECLARE_ALIGNED(8, static const int8_t, constants_fe)[] = DUP8 (0xfe);

-DECLARE_ALIGNED(8, static const int8_t, constants_7f)[] = DUP8 (0x7f);

-DECLARE_ALIGNED(8, static const int8_t, constants128)[] = DUP8 (128);

-DECLARE_ALIGNED(8, static const int16_t, constants256_512)[] =

-        {256, 512, 256, 512};

-DECLARE_ALIGNED(8, static const int16_t, constants256_1024)[] =

-        {256, 1024, 256, 1024};

-

-#define REF_0           0

-#define REF_0_1         1

-#define REF_2           2

-#define REF_2_1         3

-#define REF_4           4

-#define REF_4_1         5

-#define REF_6           6

-#define REF_6_1         7

-#define REF_S0          8

-#define REF_S0_1        9

-#define REF_S2          10

-#define REF_S2_1        11

-#define REF_S4          12

-#define REF_S4_1        13

-#define REF_S6          14

-#define REF_S6_1        15

-#define DST_0           16

-#define DST_1           17

-#define DST_2           18

-#define DST_3           19

-#define CONST_1         20

-#define CONST_2         20

-#define CONST_3         20

-#define CONST_6         20

-#define MASK_fe         20

-#define CONST_128       22

-#define CONST_256       22

-#define CONST_512       22

-#define CONST_1024      22

-#define TMP0            24

-#define TMP1            25

-#define TMP2            26

-#define TMP3            27

-#define TMP4            28

-#define TMP5            29

-#define ZERO            30

-#define MASK_7f         30

-

-#define TMP6            32

-#define TMP8            34

-#define TMP10           36

-#define TMP12           38

-#define TMP14           40

-#define TMP16           42

-#define TMP18           44

-#define TMP20           46

-#define TMP22           48

-#define TMP24           50

-#define TMP26           52

-#define TMP28           54

-#define TMP30           56

-#define TMP32           58

-

-static void MC_put_o_16_vis (uint8_t * dest, const uint8_t * ref,

-                             const ptrdiff_t stride, int height)

-{

-        ref = vis_alignaddr(ref);

-        do {    /* 5 cycles */

-                vis_ld64(ref[0], TMP0);

-

-                vis_ld64_2(ref, 8, TMP2);

-

-                vis_ld64_2(ref, 16, TMP4);

-                ref += stride;

-

-                vis_faligndata(TMP0, TMP2, REF_0);

-                vis_st64(REF_0, dest[0]);

-

-                vis_faligndata(TMP2, TMP4, REF_2);

-                vis_st64_2(REF_2, dest, 8);

-                dest += stride;

-        } while (--height);

-}

-

-static void MC_put_o_8_vis (uint8_t * dest, const uint8_t * ref,

-                            const ptrdiff_t stride, int height)

-{

-        ref = vis_alignaddr(ref);

-        do {    /* 4 cycles */

-                vis_ld64(ref[0], TMP0);

-

-                vis_ld64(ref[8], TMP2);

-                ref += stride;

-

-                /* stall */

-

-                vis_faligndata(TMP0, TMP2, REF_0);

-                vis_st64(REF_0, dest[0]);

-                dest += stride;

-        } while (--height);

-}

-

-

-static void MC_avg_o_16_vis (uint8_t * dest, const uint8_t * ref,

-                             const ptrdiff_t stride, int height)

-{

-        int stride_8 = stride + 8;

-

-        ref = vis_alignaddr(ref);

-

-        vis_ld64(ref[0], TMP0);

-

-        vis_ld64(ref[8], TMP2);

-

-        vis_ld64(ref[16], TMP4);

-

-        vis_ld64(dest[0], DST_0);

-

-        vis_ld64(dest[8], DST_2);

-

-        vis_ld64(constants_fe[0], MASK_fe);

-        vis_faligndata(TMP0, TMP2, REF_0);

-

-        vis_ld64(constants_7f[0], MASK_7f);

-        vis_faligndata(TMP2, TMP4, REF_2);

-

-        vis_ld64(constants128[0], CONST_128);

-

-        ref += stride;

-        height = (height >> 1) - 1;

-

-        do {    /* 24 cycles */

-                vis_ld64(ref[0], TMP0);

-                vis_xor(DST_0, REF_0, TMP6);

-

-                vis_ld64_2(ref, 8, TMP2);

-                vis_and(TMP6, MASK_fe, TMP6);

-

-                vis_ld64_2(ref, 16, TMP4);

-                ref += stride;

-                vis_mul8x16(CONST_128, TMP6, TMP6);

-                vis_xor(DST_2, REF_2, TMP8);

-

-                vis_and(TMP8, MASK_fe, TMP8);

-

-                vis_or(DST_0, REF_0, TMP10);

-                vis_ld64_2(dest, stride, DST_0);

-                vis_mul8x16(CONST_128, TMP8, TMP8);

-

-                vis_or(DST_2, REF_2, TMP12);

-                vis_ld64_2(dest, stride_8, DST_2);

-

-                vis_ld64(ref[0], TMP14);

-                vis_and(TMP6, MASK_7f, TMP6);

-

-                vis_and(TMP8, MASK_7f, TMP8);

-

-                vis_psub16(TMP10, TMP6, TMP6);

-                vis_st64(TMP6, dest[0]);

-

-                vis_psub16(TMP12, TMP8, TMP8);

-                vis_st64_2(TMP8, dest, 8);

-

-                dest += stride;

-                vis_ld64_2(ref, 8, TMP16);

-                vis_faligndata(TMP0, TMP2, REF_0);

-

-                vis_ld64_2(ref, 16, TMP18);

-                vis_faligndata(TMP2, TMP4, REF_2);

-                ref += stride;

-

-                vis_xor(DST_0, REF_0, TMP20);

-

-                vis_and(TMP20, MASK_fe, TMP20);

-

-                vis_xor(DST_2, REF_2, TMP22);

-                vis_mul8x16(CONST_128, TMP20, TMP20);

-

-                vis_and(TMP22, MASK_fe, TMP22);

-

-                vis_or(DST_0, REF_0, TMP24);

-                vis_mul8x16(CONST_128, TMP22, TMP22);

-

-                vis_or(DST_2, REF_2, TMP26);

-

-                vis_ld64_2(dest, stride, DST_0);

-                vis_faligndata(TMP14, TMP16, REF_0);

-

-                vis_ld64_2(dest, stride_8, DST_2);

-                vis_faligndata(TMP16, TMP18, REF_2);

-

-                vis_and(TMP20, MASK_7f, TMP20);

-

-                vis_and(TMP22, MASK_7f, TMP22);

-

-                vis_psub16(TMP24, TMP20, TMP20);

-                vis_st64(TMP20, dest[0]);

-

-                vis_psub16(TMP26, TMP22, TMP22);

-                vis_st64_2(TMP22, dest, 8);

-                dest += stride;

-        } while (--height);

-

-        vis_ld64(ref[0], TMP0);

-        vis_xor(DST_0, REF_0, TMP6);

-

-        vis_ld64_2(ref, 8, TMP2);

-        vis_and(TMP6, MASK_fe, TMP6);

-

-        vis_ld64_2(ref, 16, TMP4);

-        vis_mul8x16(CONST_128, TMP6, TMP6);

-        vis_xor(DST_2, REF_2, TMP8);

-

-        vis_and(TMP8, MASK_fe, TMP8);

-

-        vis_or(DST_0, REF_0, TMP10);

-        vis_ld64_2(dest, stride, DST_0);

-        vis_mul8x16(CONST_128, TMP8, TMP8);

-

-        vis_or(DST_2, REF_2, TMP12);

-        vis_ld64_2(dest, stride_8, DST_2);

-

-        vis_ld64(ref[0], TMP14);

-        vis_and(TMP6, MASK_7f, TMP6);

-

-        vis_and(TMP8, MASK_7f, TMP8);

-

-        vis_psub16(TMP10, TMP6, TMP6);

-        vis_st64(TMP6, dest[0]);

-

-        vis_psub16(TMP12, TMP8, TMP8);

-        vis_st64_2(TMP8, dest, 8);

-

-        dest += stride;

-        vis_faligndata(TMP0, TMP2, REF_0);

-

-        vis_faligndata(TMP2, TMP4, REF_2);

-

-        vis_xor(DST_0, REF_0, TMP20);

-

-        vis_and(TMP20, MASK_fe, TMP20);

-

-        vis_xor(DST_2, REF_2, TMP22);

-        vis_mul8x16(CONST_128, TMP20, TMP20);

-

-        vis_and(TMP22, MASK_fe, TMP22);

-

-        vis_or(DST_0, REF_0, TMP24);

-        vis_mul8x16(CONST_128, TMP22, TMP22);

-

-        vis_or(DST_2, REF_2, TMP26);

-

-        vis_and(TMP20, MASK_7f, TMP20);

-

-        vis_and(TMP22, MASK_7f, TMP22);

-

-        vis_psub16(TMP24, TMP20, TMP20);

-        vis_st64(TMP20, dest[0]);

-

-        vis_psub16(TMP26, TMP22, TMP22);

-        vis_st64_2(TMP22, dest, 8);

-}

-

-static void MC_avg_o_8_vis (uint8_t * dest, const uint8_t * ref,

-                            const ptrdiff_t stride, int height)

-{

-        ref = vis_alignaddr(ref);

-

-        vis_ld64(ref[0], TMP0);

-

-        vis_ld64(ref[8], TMP2);

-

-        vis_ld64(dest[0], DST_0);

-

-        vis_ld64(constants_fe[0], MASK_fe);

-

-        vis_ld64(constants_7f[0], MASK_7f);

-        vis_faligndata(TMP0, TMP2, REF_0);

-

-        vis_ld64(constants128[0], CONST_128);

-

-        ref += stride;

-        height = (height >> 1) - 1;

-

-        do {    /* 12 cycles */

-                vis_ld64(ref[0], TMP0);

-                vis_xor(DST_0, REF_0, TMP4);

-

-                vis_ld64(ref[8], TMP2);

-                vis_and(TMP4, MASK_fe, TMP4);

-

-                vis_or(DST_0, REF_0, TMP6);

-                vis_ld64_2(dest, stride, DST_0);

-                ref += stride;

-                vis_mul8x16(CONST_128, TMP4, TMP4);

-

-                vis_ld64(ref[0], TMP12);

-                vis_faligndata(TMP0, TMP2, REF_0);

-

-                vis_ld64(ref[8], TMP2);

-                vis_xor(DST_0, REF_0, TMP0);

-                ref += stride;

-

-                vis_and(TMP0, MASK_fe, TMP0);

-

-                vis_and(TMP4, MASK_7f, TMP4);

-

-                vis_psub16(TMP6, TMP4, TMP4);

-                vis_st64(TMP4, dest[0]);

-                dest += stride;

-                vis_mul8x16(CONST_128, TMP0, TMP0);

-

-                vis_or(DST_0, REF_0, TMP6);

-                vis_ld64_2(dest, stride, DST_0);

-

-                vis_faligndata(TMP12, TMP2, REF_0);

-

-                vis_and(TMP0, MASK_7f, TMP0);

-

-                vis_psub16(TMP6, TMP0, TMP4);

-                vis_st64(TMP4, dest[0]);

-                dest += stride;

-        } while (--height);

-

-        vis_ld64(ref[0], TMP0);

-        vis_xor(DST_0, REF_0, TMP4);

-

-        vis_ld64(ref[8], TMP2);

-        vis_and(TMP4, MASK_fe, TMP4);

-

-        vis_or(DST_0, REF_0, TMP6);

-        vis_ld64_2(dest, stride, DST_0);

-        vis_mul8x16(CONST_128, TMP4, TMP4);

-

-        vis_faligndata(TMP0, TMP2, REF_0);

-

-        vis_xor(DST_0, REF_0, TMP0);

-

-        vis_and(TMP0, MASK_fe, TMP0);

-

-        vis_and(TMP4, MASK_7f, TMP4);

-

-        vis_psub16(TMP6, TMP4, TMP4);

-        vis_st64(TMP4, dest[0]);

-        dest += stride;

-        vis_mul8x16(CONST_128, TMP0, TMP0);

-

-        vis_or(DST_0, REF_0, TMP6);

-

-        vis_and(TMP0, MASK_7f, TMP0);

-

-        vis_psub16(TMP6, TMP0, TMP4);

-        vis_st64(TMP4, dest[0]);

-}

-

-static void MC_put_x_16_vis (uint8_t * dest, const uint8_t * ref,

-                             const ptrdiff_t stride, int height)

-{

-        unsigned long off = (unsigned long) ref & 0x7;

-        unsigned long off_plus_1 = off + 1;

-

-        ref = vis_alignaddr(ref);

-

-        vis_ld64(ref[0],    TMP0);

-

-        vis_ld64_2(ref, 8,  TMP2);

-

-        vis_ld64_2(ref, 16, TMP4);

-

-        vis_ld64(constants_fe[0], MASK_fe);

-

-        vis_ld64(constants_7f[0], MASK_7f);

-        vis_faligndata(TMP0, TMP2, REF_0);

-

-        vis_ld64(constants128[0], CONST_128);

-        vis_faligndata(TMP2, TMP4, REF_4);

-

-        if (off != 0x7) {

-                vis_alignaddr_g0((void *)off_plus_1);

-                vis_faligndata(TMP0, TMP2, REF_2);

-                vis_faligndata(TMP2, TMP4, REF_6);

-        } else {

-                vis_src1(TMP2, REF_2);

-                vis_src1(TMP4, REF_6);

-        }

-

-        ref += stride;

-        height = (height >> 1) - 1;

-

-        do {    /* 34 cycles */

-                vis_ld64(ref[0],    TMP0);

-                vis_xor(REF_0, REF_2, TMP6);

-

-                vis_ld64_2(ref, 8,  TMP2);

-                vis_xor(REF_4, REF_6, TMP8);

-

-                vis_ld64_2(ref, 16, TMP4);

-                vis_and(TMP6, MASK_fe, TMP6);

-                ref += stride;

-

-                vis_ld64(ref[0],    TMP14);

-                vis_mul8x16(CONST_128, TMP6, TMP6);

-                vis_and(TMP8, MASK_fe, TMP8);

-

-                vis_ld64_2(ref, 8,  TMP16);

-                vis_mul8x16(CONST_128, TMP8, TMP8);

-                vis_or(REF_0, REF_2, TMP10);

-

-                vis_ld64_2(ref, 16, TMP18);

-                ref += stride;

-                vis_or(REF_4, REF_6, TMP12);

-

-                vis_alignaddr_g0((void *)off);

-

-                vis_faligndata(TMP0, TMP2, REF_0);

-

-                vis_faligndata(TMP2, TMP4, REF_4);

-

-                if (off != 0x7) {

-                        vis_alignaddr_g0((void *)off_plus_1);

-                        vis_faligndata(TMP0, TMP2, REF_2);

-                        vis_faligndata(TMP2, TMP4, REF_6);

-                } else {

-                        vis_src1(TMP2, REF_2);

-                        vis_src1(TMP4, REF_6);

-                }

-

-                vis_and(TMP6, MASK_7f, TMP6);

-

-                vis_and(TMP8, MASK_7f, TMP8);

-

-                vis_psub16(TMP10, TMP6, TMP6);

-                vis_st64(TMP6, dest[0]);

-

-                vis_psub16(TMP12, TMP8, TMP8);

-                vis_st64_2(TMP8, dest, 8);

-                dest += stride;

-

-                vis_xor(REF_0, REF_2, TMP6);

-

-                vis_xor(REF_4, REF_6, TMP8);

-

-                vis_and(TMP6, MASK_fe, TMP6);

-

-                vis_mul8x16(CONST_128, TMP6, TMP6);

-                vis_and(TMP8, MASK_fe, TMP8);

-

-                vis_mul8x16(CONST_128, TMP8, TMP8);

-                vis_or(REF_0, REF_2, TMP10);

-

-                vis_or(REF_4, REF_6, TMP12);

-

-                vis_alignaddr_g0((void *)off);

-

-                vis_faligndata(TMP14, TMP16, REF_0);

-

-                vis_faligndata(TMP16, TMP18, REF_4);

-

-                if (off != 0x7) {

-                        vis_alignaddr_g0((void *)off_plus_1);

-                        vis_faligndata(TMP14, TMP16, REF_2);

-                        vis_faligndata(TMP16, TMP18, REF_6);

-                } else {

-                        vis_src1(TMP16, REF_2);

-                        vis_src1(TMP18, REF_6);

-                }

-

-                vis_and(TMP6, MASK_7f, TMP6);

-

-                vis_and(TMP8, MASK_7f, TMP8);

-

-                vis_psub16(TMP10, TMP6, TMP6);

-                vis_st64(TMP6, dest[0]);

-

-                vis_psub16(TMP12, TMP8, TMP8);

-                vis_st64_2(TMP8, dest, 8);

-                dest += stride;

-        } while (--height);

-

-        vis_ld64(ref[0],    TMP0);

-        vis_xor(REF_0, REF_2, TMP6);

-

-        vis_ld64_2(ref, 8,  TMP2);

-        vis_xor(REF_4, REF_6, TMP8);

-

-        vis_ld64_2(ref, 16, TMP4);

-        vis_and(TMP6, MASK_fe, TMP6);

-

-        vis_mul8x16(CONST_128, TMP6, TMP6);

-        vis_and(TMP8, MASK_fe, TMP8);

-

-        vis_mul8x16(CONST_128, TMP8, TMP8);

-        vis_or(REF_0, REF_2, TMP10);

-

-        vis_or(REF_4, REF_6, TMP12);

-

-        vis_alignaddr_g0((void *)off);

-

-        vis_faligndata(TMP0, TMP2, REF_0);

-

-        vis_faligndata(TMP2, TMP4, REF_4);

-

-        if (off != 0x7) {

-                vis_alignaddr_g0((void *)off_plus_1);

-                vis_faligndata(TMP0, TMP2, REF_2);

-                vis_faligndata(TMP2, TMP4, REF_6);

-        } else {

-                vis_src1(TMP2, REF_2);

-                vis_src1(TMP4, REF_6);

-        }

-

-        vis_and(TMP6, MASK_7f, TMP6);

-

-        vis_and(TMP8, MASK_7f, TMP8);

-

-        vis_psub16(TMP10, TMP6, TMP6);

-        vis_st64(TMP6, dest[0]);

-

-        vis_psub16(TMP12, TMP8, TMP8);

-        vis_st64_2(TMP8, dest, 8);

-        dest += stride;

-

-        vis_xor(REF_0, REF_2, TMP6);

-

-        vis_xor(REF_4, REF_6, TMP8);

-

-        vis_and(TMP6, MASK_fe, TMP6);

-

-        vis_mul8x16(CONST_128, TMP6, TMP6);

-        vis_and(TMP8, MASK_fe, TMP8);

-

-        vis_mul8x16(CONST_128, TMP8, TMP8);

-        vis_or(REF_0, REF_2, TMP10);

-

-        vis_or(REF_4, REF_6, TMP12);

-

-        vis_and(TMP6, MASK_7f, TMP6);

-

-        vis_and(TMP8, MASK_7f, TMP8);

-

-        vis_psub16(TMP10, TMP6, TMP6);

-        vis_st64(TMP6, dest[0]);

-

-        vis_psub16(TMP12, TMP8, TMP8);

-        vis_st64_2(TMP8, dest, 8);

-}

-

-static void MC_put_x_8_vis (uint8_t * dest, const uint8_t * ref,

-                            const ptrdiff_t stride, int height)

-{

-        unsigned long off = (unsigned long) ref & 0x7;

-        unsigned long off_plus_1 = off + 1;

-

-        ref = vis_alignaddr(ref);

-

-        vis_ld64(ref[0], TMP0);

-

-        vis_ld64(ref[8], TMP2);

-

-        vis_ld64(constants_fe[0], MASK_fe);

-

-        vis_ld64(constants_7f[0], MASK_7f);

-

-        vis_ld64(constants128[0], CONST_128);

-        vis_faligndata(TMP0, TMP2, REF_0);

-

-        if (off != 0x7) {

-                vis_alignaddr_g0((void *)off_plus_1);

-                vis_faligndata(TMP0, TMP2, REF_2);

-        } else {

-                vis_src1(TMP2, REF_2);

-        }

-

-        ref += stride;

-        height = (height >> 1) - 1;

-

-        do {    /* 20 cycles */

-                vis_ld64(ref[0], TMP0);

-                vis_xor(REF_0, REF_2, TMP4);

-

-                vis_ld64_2(ref, 8, TMP2);

-                vis_and(TMP4, MASK_fe, TMP4);

-                ref += stride;

-

-                vis_ld64(ref[0], TMP8);

-                vis_or(REF_0, REF_2, TMP6);

-                vis_mul8x16(CONST_128, TMP4, TMP4);

-

-                vis_alignaddr_g0((void *)off);

-

-                vis_ld64_2(ref, 8, TMP10);

-                ref += stride;

-                vis_faligndata(TMP0, TMP2, REF_0);

-

-                if (off != 0x7) {