Merge remote-tracking branch 'qatar/master'

* qatar/master:
  pcm-mpeg: convert to bytestream2 API
  Revert "h264: clear trailing bits in partially parsed NAL units"
  remove iwmmxt optimizations
  mimic: do not continue if swap_buf_size is 0
  mimic: convert to bytestream2 API
  frwu: use MKTAG to check marker instead of AV_RL32
  txd: port to bytestream2 API
  c93: convert to bytestream2 API
  iff: make .long_name more descriptive
  FATE: add test for cdxl demuxer
  rtsp: Fix a typo

Conflicts:
	libavcodec/arm/dsputil_iwmmxt.c
	libavcodec/arm/dsputil_iwmmxt_rnd_template.c
	libavcodec/arm/mpegvideo_iwmmxt.c
	libavcodec/c93.c
	libavcodec/txd.c
	libavutil/arm/cpu.c
	tests/fate/demux.mak

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

8 files changed, 0 insertions, 1438 deletions

diff --git a/libavcodec/arm/Makefile b/libavcodec/arm/Makefile
index 0e45bc0a52..39852c4f4a 100644
--- a/libavcodec/arm/Makefile
+++ b/libavcodec/arm/Makefile
@@ -44,9 +44,6 @@ OBJS-$(HAVE_ARMVFP)                    += arm/dsputil_vfp.o             \
                                           arm/dsputil_init_vfp.o        \
                                           $(VFP-OBJS-yes)
 
-OBJS-$(HAVE_IWMMXT)                    += arm/dsputil_iwmmxt.o          \
-                                          arm/mpegvideo_iwmmxt.o        \
-
 NEON-OBJS-$(CONFIG_FFT)                += arm/fft_neon.o                \
                                           arm/fft_fixed_neon.o          \
 
diff --git a/libavcodec/arm/dsputil_arm.h b/libavcodec/arm/dsputil_arm.h
index b333c70226..b7b5bdc0e6 100644
--- a/libavcodec/arm/dsputil_arm.h
+++ b/libavcodec/arm/dsputil_arm.h
@@ -28,6 +28,5 @@ void ff_dsputil_init_armv5te(DSPContext* c, AVCodecContext *avctx);
 void ff_dsputil_init_armv6(DSPContext* c, AVCodecContext *avctx);
 void ff_dsputil_init_vfp(DSPContext* c, AVCodecContext *avctx);
 void ff_dsputil_init_neon(DSPContext *c, AVCodecContext *avctx);
-void ff_dsputil_init_iwmmxt(DSPContext* c, AVCodecContext *avctx);
 
 #endif /* AVCODEC_ARM_DSPUTIL_H */
diff --git a/libavcodec/arm/dsputil_init_arm.c b/libavcodec/arm/dsputil_init_arm.c
index 58577dde9a..82af718b66 100644
--- a/libavcodec/arm/dsputil_init_arm.c
+++ b/libavcodec/arm/dsputil_init_arm.c
@@ -119,7 +119,6 @@ void ff_dsputil_init_arm(DSPContext* c, AVCodecContext *avctx)
 
     if (HAVE_ARMV5TE) ff_dsputil_init_armv5te(c, avctx);
     if (HAVE_ARMV6)   ff_dsputil_init_armv6(c, avctx);
-    if (HAVE_IWMMXT)  ff_dsputil_init_iwmmxt(c, avctx);
     if (HAVE_ARMVFP)  ff_dsputil_init_vfp(c, avctx);
     if (HAVE_NEON)    ff_dsputil_init_neon(c, avctx);
 }
diff --git a/libavcodec/arm/dsputil_iwmmxt.c b/libavcodec/arm/dsputil_iwmmxt.c
deleted file mode 100644
index 2837af119f..0000000000
--- a/libavcodec/arm/dsputil_iwmmxt.c
+++ /dev/null
@@ -1,210 +0,0 @@
-/*
- * iWMMXt optimized DSP utils
- * Copyright (c) 2004 AGAWA Koji
- *
- * This file is part of FFmpeg.
- *
- * FFmpeg 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.
- *
- * FFmpeg 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 FFmpeg; if not, write to the Free Software
- * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
- */
-
-#include "libavutil/cpu.h"
-#include "libavcodec/dsputil.h"
-
-#define DEF(x, y) x ## _no_rnd_ ## y ##_iwmmxt
-#define SET_RND(regd)  __asm__ volatile ("mov r12, #1 \n\t tbcsth " #regd ", r12":::"r12");
-#define WAVG2B "wavg2b"
-#include "dsputil_iwmmxt_rnd_template.c"
-#undef DEF
-#undef SET_RND
-#undef WAVG2B
-
-#define DEF(x, y) x ## _ ## y ##_iwmmxt
-#define SET_RND(regd)  __asm__ volatile ("mov r12, #2 \n\t tbcsth " #regd ", r12":::"r12");
-#define WAVG2B "wavg2br"
-#include "dsputil_iwmmxt_rnd_template.c"
-#undef DEF
-#undef SET_RND
-#undef WAVG2BR
-
-// need scheduling
-#define OP(AVG)                                         \
-    __asm__ volatile (                                      \
-        /* alignment */                                 \
-        "and r12, %[pixels], #7 \n\t"                   \
-        "bic %[pixels], %[pixels], #7 \n\t"             \
-        "tmcr wcgr1, r12 \n\t"                          \
-                                                        \
-        "wldrd wr0, [%[pixels]] \n\t"                   \
-        "wldrd wr1, [%[pixels], #8] \n\t"               \
-        "add %[pixels], %[pixels], %[line_size] \n\t"   \
-        "walignr1 wr4, wr0, wr1 \n\t"                   \
-                                                        \
-        "1: \n\t"                                       \
-                                                        \
-        "wldrd wr2, [%[pixels]] \n\t"                   \
-        "wldrd wr3, [%[pixels], #8] \n\t"               \
-        "add %[pixels], %[pixels], %[line_size] \n\t"   \
-        "pld [%[pixels]] \n\t"                          \
-        "walignr1 wr5, wr2, wr3 \n\t"                   \
-        AVG " wr6, wr4, wr5 \n\t"                       \
-        "wstrd wr6, [%[block]] \n\t"                    \
-        "add %[block], %[block], %[line_size] \n\t"     \
-                                                        \
-        "wldrd wr0, [%[pixels]] \n\t"                   \
-        "wldrd wr1, [%[pixels], #8] \n\t"               \
-        "add %[pixels], %[pixels], %[line_size] \n\t"   \
-        "walignr1 wr4, wr0, wr1 \n\t"                   \
-        "pld [%[pixels]] \n\t"                          \
-        AVG " wr6, wr4, wr5 \n\t"                       \
-        "wstrd wr6, [%[block]] \n\t"                    \
-        "add %[block], %[block], %[line_size] \n\t"     \
-                                                        \
-        "subs %[h], %[h], #2 \n\t"                      \
-        "bne 1b \n\t"                                   \
-        : [block]"+r"(block), [pixels]"+r"(pixels), [h]"+r"(h)  \
-        : [line_size]"r"(line_size) \
-        : "memory", "r12");
-void put_pixels8_y2_iwmmxt(uint8_t *block, const uint8_t *pixels, const int line_size, int h)
-{
-    OP("wavg2br");
-}
-void put_no_rnd_pixels8_y2_iwmmxt(uint8_t *block, const uint8_t *pixels, const int line_size, int h)
-{
-    OP("wavg2b");
-}
-#undef OP
-
-void add_pixels_clamped_iwmmxt(const DCTELEM *block, uint8_t *pixels, int line_size)
-{
-    uint8_t *pixels2 = pixels + line_size;
-
-    __asm__ volatile (
-        "mov            r12, #4                 \n\t"
-        "1:                                     \n\t"
-        "pld            [%[pixels], %[line_size2]]              \n\t"
-        "pld            [%[pixels2], %[line_size2]]             \n\t"
-        "wldrd          wr4, [%[pixels]]        \n\t"
-        "wldrd          wr5, [%[pixels2]]       \n\t"
-        "pld            [%[block], #32]         \n\t"
-        "wunpckelub     wr6, wr4                \n\t"
-        "wldrd          wr0, [%[block]]         \n\t"
-        "wunpckehub     wr7, wr4                \n\t"
-        "wldrd          wr1, [%[block], #8]     \n\t"
-        "wunpckelub     wr8, wr5                \n\t"
-        "wldrd          wr2, [%[block], #16]    \n\t"
-        "wunpckehub     wr9, wr5                \n\t"
-        "wldrd          wr3, [%[block], #24]    \n\t"
-        "add            %[block], %[block], #32 \n\t"
-        "waddhss        wr10, wr0, wr6          \n\t"
-        "waddhss        wr11, wr1, wr7          \n\t"
-        "waddhss        wr12, wr2, wr8          \n\t"
-        "waddhss        wr13, wr3, wr9          \n\t"
-        "wpackhus       wr14, wr10, wr11        \n\t"
-        "wpackhus       wr15, wr12, wr13        \n\t"
-        "wstrd          wr14, [%[pixels]]       \n\t"
-        "add            %[pixels], %[pixels], %[line_size2]     \n\t"
-        "subs           r12, r12, #1            \n\t"
-        "wstrd          wr15, [%[pixels2]]      \n\t"
-        "add            %[pixels2], %[pixels2], %[line_size2]   \n\t"
-        "bne            1b                      \n\t"
-        : [block]"+r"(block), [pixels]"+r"(pixels), [pixels2]"+r"(pixels2)
-        : [line_size2]"r"(line_size << 1)
-        : "cc", "memory", "r12");
-}
-
-static void clear_blocks_iwmmxt(DCTELEM *blocks)
-{
-    __asm__ volatile(
-                "wzero wr0                      \n\t"
-                "mov r1, #(128 * 6 / 32)        \n\t"
-                "1:                             \n\t"
-                "wstrd wr0, [%0]                \n\t"
-                "wstrd wr0, [%0, #8]            \n\t"
-                "wstrd wr0, [%0, #16]           \n\t"
-                "wstrd wr0, [%0, #24]           \n\t"
-                "subs r1, r1, #1                \n\t"
-                "add %0, %0, #32                \n\t"
-                "bne 1b                         \n\t"
-                : "+r"(blocks)
-                :
-                : "r1"
-        );
-}
-
-static void nop(uint8_t *block, const uint8_t *pixels, int line_size, int h)
-{
-    return;
-}
-
-/* A run time test is not simple. If this file is compiled in
- * then we should install the functions
- */
-
-void ff_dsputil_init_iwmmxt(DSPContext* c, AVCodecContext *avctx)
-{
-    int mm_flags = AV_CPU_FLAG_IWMMXT; /* multimedia extension flags */
-    const int high_bit_depth = avctx->bits_per_raw_sample > 8;
-
-    if (avctx->dsp_mask) {
-        if (avctx->dsp_mask & AV_CPU_FLAG_FORCE)
-            mm_flags |= (avctx->dsp_mask & 0xffff);
-        else
-            mm_flags &= ~(avctx->dsp_mask & 0xffff);
-    }
-
-    if (!(mm_flags & AV_CPU_FLAG_IWMMXT)) return;
-
-    c->add_pixels_clamped = add_pixels_clamped_iwmmxt;
-
-    if (!high_bit_depth) {
-    c->clear_blocks = clear_blocks_iwmmxt;
-
-    c->put_pixels_tab[0][0] = put_pixels16_iwmmxt;
-    c->put_pixels_tab[0][1] = put_pixels16_x2_iwmmxt;
-    c->put_pixels_tab[0][2] = put_pixels16_y2_iwmmxt;
-    c->put_pixels_tab[0][3] = put_pixels16_xy2_iwmmxt;
-    c->put_no_rnd_pixels_tab[0][0] = put_pixels16_iwmmxt;
-    c->put_no_rnd_pixels_tab[0][1] = put_no_rnd_pixels16_x2_iwmmxt;
-    c->put_no_rnd_pixels_tab[0][2] = put_no_rnd_pixels16_y2_iwmmxt;
-    c->put_no_rnd_pixels_tab[0][3] = put_no_rnd_pixels16_xy2_iwmmxt;
-
-    c->put_pixels_tab[1][0] = put_pixels8_iwmmxt;
-    c->put_pixels_tab[1][1] = put_pixels8_x2_iwmmxt;
-    c->put_pixels_tab[1][2] = put_pixels8_y2_iwmmxt;
-    c->put_pixels_tab[1][3] = put_pixels8_xy2_iwmmxt;
-    c->put_no_rnd_pixels_tab[1][0] = put_pixels8_iwmmxt;
-    c->put_no_rnd_pixels_tab[1][1] = put_no_rnd_pixels8_x2_iwmmxt;
-    c->put_no_rnd_pixels_tab[1][2] = put_no_rnd_pixels8_y2_iwmmxt;
-    c->put_no_rnd_pixels_tab[1][3] = put_no_rnd_pixels8_xy2_iwmmxt;
-
-    c->avg_pixels_tab[0][0] = avg_pixels16_iwmmxt;
-    c->avg_pixels_tab[0][1] = avg_pixels16_x2_iwmmxt;
-    c->avg_pixels_tab[0][2] = avg_pixels16_y2_iwmmxt;
-    c->avg_pixels_tab[0][3] = avg_pixels16_xy2_iwmmxt;
-    c->avg_no_rnd_pixels_tab[0][0] = avg_pixels16_iwmmxt;
-    c->avg_no_rnd_pixels_tab[0][1] = avg_no_rnd_pixels16_x2_iwmmxt;
-    c->avg_no_rnd_pixels_tab[0][2] = avg_no_rnd_pixels16_y2_iwmmxt;
-    c->avg_no_rnd_pixels_tab[0][3] = avg_no_rnd_pixels16_xy2_iwmmxt;
-
-    c->avg_pixels_tab[1][0] = avg_pixels8_iwmmxt;
-    c->avg_pixels_tab[1][1] = avg_pixels8_x2_iwmmxt;
-    c->avg_pixels_tab[1][2] = avg_pixels8_y2_iwmmxt;
-    c->avg_pixels_tab[1][3] = avg_pixels8_xy2_iwmmxt;
-    c->avg_no_rnd_pixels_tab[1][0] = avg_no_rnd_pixels8_iwmmxt;
-    c->avg_no_rnd_pixels_tab[1][1] = avg_no_rnd_pixels8_x2_iwmmxt;
-    c->avg_no_rnd_pixels_tab[1][2] = avg_no_rnd_pixels8_y2_iwmmxt;
-    c->avg_no_rnd_pixels_tab[1][3] = avg_no_rnd_pixels8_xy2_iwmmxt;
-    }
-}
diff --git a/libavcodec/arm/dsputil_iwmmxt_rnd_template.c b/libavcodec/arm/dsputil_iwmmxt_rnd_template.c
deleted file mode 100644
index 35a5a9b8b4..0000000000
--- a/libavcodec/arm/dsputil_iwmmxt_rnd_template.c
+++ /dev/null
@@ -1,1114 +0,0 @@
-/*
- * iWMMXt optimized DSP utils
- * copyright (c) 2004 AGAWA Koji
- *
- * This file is part of FFmpeg.
- *
- * FFmpeg 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.
- *
- * FFmpeg 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 FFmpeg; if not, write to the Free Software
- * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
- */
-
-void DEF(put, pixels8)(uint8_t *block, const uint8_t *pixels, const int line_size, int h)
-{
-    int stride = line_size;
-    __asm__ volatile (
-        "and r12, %[pixels], #7 \n\t"
-        "bic %[pixels], %[pixels], #7 \n\t"
-        "tmcr wcgr1, r12 \n\t"
-        "add r4, %[pixels], %[line_size] \n\t"
-        "add r5, %[block], %[line_size] \n\t"
-        "mov %[line_size], %[line_size], lsl #1 \n\t"
-        "1: \n\t"
-        "wldrd wr0, [%[pixels]] \n\t"
-        "subs %[h], %[h], #2 \n\t"
-        "wldrd wr1, [%[pixels], #8] \n\t"
-        "add %[pixels], %[pixels], %[line_size] \n\t"
-        "wldrd wr3, [r4] \n\t"
-        "pld [%[pixels]] \n\t"
-        "pld [%[pixels], #32] \n\t"
-        "wldrd wr4, [r4, #8] \n\t"
-        "add r4, r4, %[line_size] \n\t"
-        "walignr1 wr8, wr0, wr1 \n\t"
-        "pld [r4] \n\t"
-        "pld [r4, #32] \n\t"
-        "walignr1 wr10, wr3, wr4 \n\t"
-        "wstrd wr8, [%[block]] \n\t"
-        "add %[block], %[block], %[line_size] \n\t"
-        "wstrd wr10, [r5] \n\t"
-        "add r5, r5, %[line_size] \n\t"
-        "bne 1b \n\t"
-        : [block]"+r"(block), [pixels]"+r"(pixels), [line_size]"+r"(stride), [h]"+r"(h)
-        :
-        : "memory", "r4", "r5", "r12");
-}
-
-void DEF(avg, pixels8)(uint8_t *block, const uint8_t *pixels, const int line_size, int h)
-{
-    int stride = line_size;
-    __asm__ volatile (
-        "and r12, %[pixels], #7 \n\t"
-        "bic %[pixels], %[pixels], #7 \n\t"
-        "tmcr wcgr1, r12 \n\t"
-        "add r4, %[pixels], %[line_size] \n\t"
-        "add r5, %[block], %[line_size] \n\t"
-        "mov %[line_size], %[line_size], lsl #1 \n\t"
-        "1: \n\t"
-        "wldrd wr0, [%[pixels]] \n\t"
-        "subs %[h], %[h], #2 \n\t"
-        "wldrd wr1, [%[pixels], #8] \n\t"
-        "add %[pixels], %[pixels], %[line_size] \n\t"
-        "wldrd wr3, [r4] \n\t"
-        "pld [%[pixels]] \n\t"
-        "pld [%[pixels], #32] \n\t"
-        "wldrd wr4, [r4, #8] \n\t"
-        "add r4, r4, %[line_size] \n\t"
-        "walignr1 wr8, wr0, wr1 \n\t"
-        "wldrd wr0, [%[block]] \n\t"
-        "wldrd wr2, [r5] \n\t"
-        "pld [r4] \n\t"
-        "pld [r4, #32] \n\t"
-        "walignr1 wr10, wr3, wr4 \n\t"
-        WAVG2B" wr8, wr8, wr0 \n\t"
-        WAVG2B" wr10, wr10, wr2 \n\t"
-        "wstrd wr8, [%[block]] \n\t"
-        "add %[block], %[block], %[line_size] \n\t"
-        "wstrd wr10, [r5] \n\t"
-        "pld [%[block]] \n\t"
-        "pld [%[block], #32] \n\t"
-        "add r5, r5, %[line_size] \n\t"
-        "pld [r5] \n\t"
-        "pld [r5, #32] \n\t"
-        "bne 1b \n\t"
-        : [block]"+r"(block), [pixels]"+r"(pixels), [line_size]"+r"(stride), [h]"+r"(h)
-        :
-        : "memory", "r4", "r5", "r12");
-}
-
-void DEF(put, pixels16)(uint8_t *block, const uint8_t *pixels, const int line_size, int h)
-{
-    int stride = line_size;
-    __asm__ volatile (
-        "and r12, %[pixels], #7 \n\t"
-        "bic %[pixels], %[pixels], #7 \n\t"
-        "tmcr wcgr1, r12 \n\t"
-        "add r4, %[pixels], %[line_size] \n\t"
-        "add r5, %[block], %[line_size] \n\t"
-        "mov %[line_size], %[line_size], lsl #1 \n\t"
-        "1: \n\t"
-        "wldrd wr0, [%[pixels]] \n\t"
-        "wldrd wr1, [%[pixels], #8] \n\t"
-        "subs %[h], %[h], #2 \n\t"
-        "wldrd wr2, [%[pixels], #16] \n\t"
-        "add %[pixels], %[pixels], %[line_size] \n\t"
-        "wldrd wr3, [r4] \n\t"
-        "pld [%[pixels]] \n\t"
-        "pld [%[pixels], #32] \n\t"
-        "walignr1 wr8, wr0, wr1 \n\t"
-        "wldrd wr4, [r4, #8] \n\t"
-        "walignr1 wr9, wr1, wr2 \n\t"
-        "wldrd wr5, [r4, #16] \n\t"
-        "add r4, r4, %[line_size] \n\t"
-        "pld [r4] \n\t"
-        "pld [r4, #32] \n\t"
-        "walignr1 wr10, wr3, wr4 \n\t"
-        "wstrd wr8, [%[block]] \n\t"
-        "walignr1 wr11, wr4, wr5 \n\t"
-        "wstrd wr9, [%[block], #8] \n\t"
-        "add %[block], %[block], %[line_size] \n\t"
-        "wstrd wr10, [r5] \n\t"
-        "wstrd wr11, [r5, #8] \n\t"
-        "add r5, r5, %[line_size] \n\t"
-        "bne 1b \n\t"
-        : [block]"+r"(block), [pixels]"+r"(pixels), [line_size]"+r"(stride), [h]"+r"(h)
-        :
-        : "memory", "r4", "r5", "r12");
-}
-
-void DEF(avg, pixels16)(uint8_t *block, const uint8_t *pixels, const int line_size, int h)
-{
-    int stride = line_size;
-    __asm__ volatile (
-        "pld [%[pixels]]                \n\t"
-        "pld [%[pixels], #32]           \n\t"
-        "pld [%[block]]                 \n\t"
-        "pld [%[block], #32]            \n\t"
-        "and r12, %[pixels], #7         \n\t"
-        "bic %[pixels], %[pixels], #7   \n\t"
-        "tmcr wcgr1, r12                \n\t"
-        "add r4, %[pixels], %[line_size]\n\t"
-        "add r5, %[block], %[line_size] \n\t"
-        "mov %[line_size], %[line_size], lsl #1 \n\t"
-        "1:                             \n\t"
-        "wldrd wr0, [%[pixels]]         \n\t"
-        "wldrd wr1, [%[pixels], #8]     \n\t"
-        "subs %[h], %[h], #2            \n\t"
-        "wldrd wr2, [%[pixels], #16]    \n\t"
-        "add %[pixels], %[pixels], %[line_size] \n\t"
-        "wldrd wr3, [r4]                \n\t"
-        "pld [%[pixels]]                \n\t"
-        "pld [%[pixels], #32]           \n\t"
-        "walignr1 wr8, wr0, wr1         \n\t"
-        "wldrd wr4, [r4, #8]            \n\t"
-        "walignr1 wr9, wr1, wr2         \n\t"
-        "wldrd wr5, [r4, #16]           \n\t"
-        "add r4, r4, %[line_size]       \n\t"
-        "wldrd wr0, [%[block]]          \n\t"
-        "pld [r4]                       \n\t"
-        "wldrd wr1, [%[block], #8]      \n\t"
-        "pld [r4, #32]                  \n\t"
-        "wldrd wr2, [r5]                \n\t"
-        "walignr1 wr10, wr3, wr4        \n\t"
-        "wldrd wr3, [r5, #8]            \n\t"
-        WAVG2B" wr8, wr8, wr0           \n\t"
-        WAVG2B" wr9, wr9, wr1           \n\t"
-        WAVG2B" wr10, wr10, wr2         \n\t"
-        "wstrd wr8, [%[block]]          \n\t"
-        "walignr1 wr11, wr4, wr5        \n\t"
-        WAVG2B" wr11, wr11, wr3         \n\t"
-        "wstrd wr9, [%[block], #8]      \n\t"
-        "add %[block], %[block], %[line_size] \n\t"
-        "wstrd wr10, [r5]               \n\t"
-        "pld [%[block]]                 \n\t"
-        "pld [%[block], #32]            \n\t"
-        "wstrd wr11, [r5, #8]           \n\t"
-        "add r5, r5, %[line_size]       \n\t"
-        "pld [r5]                       \n\t"
-        "pld [r5, #32]                  \n\t"
-        "bne 1b \n\t"
-        : [block]"+r"(block), [pixels]"+r"(pixels), [line_size]"+r"(stride), [h]"+r"(h)
-        :
-        : "memory", "r4", "r5", "r12");
-}
-
-void DEF(put, pixels8_x2)(uint8_t *block, const uint8_t *pixels, const int line_size, int h)
-{
-    int stride = line_size;
-    // [wr0 wr1 wr2 wr3] for previous line
-    // [wr4 wr5 wr6 wr7] for current line
-    SET_RND(wr15); // =2 for rnd  and  =1 for no_rnd version
-    __asm__ volatile(
-        "pld [%[pixels]]                \n\t"
-        "pld [%[pixels], #32]           \n\t"
-        "and r12, %[pixels], #7         \n\t"
-        "bic %[pixels], %[pixels], #7   \n\t"
-        "tmcr wcgr1, r12                \n\t"
-        "add r12, r12, #1               \n\t"
-        "add r4, %[pixels], %[line_size]\n\t"
-        "tmcr wcgr2, r12                \n\t"
-        "add r5, %[block], %[line_size] \n\t"
-        "mov %[line_size], %[line_size], lsl #1 \n\t"
-
-        "1:                             \n\t"
-        "wldrd wr10, [%[pixels]]        \n\t"
-        "cmp r12, #8                    \n\t"
-        "wldrd wr11, [%[pixels], #8]    \n\t"
-        "add %[pixels], %[pixels], %[line_size] \n\t"
-        "wldrd wr13, [r4]               \n\t"
-        "pld [%[pixels]]                \n\t"
-        "wldrd wr14, [r4, #8]           \n\t"
-        "pld [%[pixels], #32]           \n\t"
-        "add r4, r4, %[line_size]       \n\t"
-        "walignr1 wr0, wr10, wr11       \n\t"
-        "pld [r4]                       \n\t"
-        "pld [r4, #32]                  \n\t"
-        "walignr1 wr2, wr13, wr14       \n\t"
-        "wmoveq wr4, wr11               \n\t"
-        "wmoveq wr6, wr14               \n\t"
-        "walignr2ne wr4, wr10, wr11     \n\t"
-        "walignr2ne wr6, wr13, wr14     \n\t"
-        WAVG2B" wr0, wr0, wr4           \n\t"
-        WAVG2B" wr2, wr2, wr6           \n\t"
-        "wstrd wr0, [%[block]]          \n\t"
-        "subs %[h], %[h], #2            \n\t"
-        "wstrd wr2, [r5]                \n\t"
-        "add %[block], %[block], %[line_size]   \n\t"
-        "add r5, r5, %[line_size]       \n\t"
-        "bne 1b                         \n\t"
-        : [h]"+r"(h), [pixels]"+r"(pixels), [block]"+r"(block), [line_size]"+r"(stride)
-        :
-        : "r4", "r5", "r12", "memory");
-}
-
-void DEF(put, pixels16_x2)(uint8_t *block, const uint8_t *pixels, const int line_size, int h)
-{
-    int stride = line_size;
-    // [wr0 wr1 wr2 wr3] for previous line
-    // [wr4 wr5 wr6 wr7] for current line
-    SET_RND(wr15); // =2 for rnd  and  =1 for no_rnd version
-    __asm__ volatile(
-        "pld [%[pixels]]                \n\t"
-        "pld [%[pixels], #32]           \n\t"
-        "and r12, %[pixels], #7         \n\t"
-        "bic %[pixels], %[pixels], #7   \n\t"
-        "tmcr wcgr1, r12                \n\t"
-        "add r12, r12, #1               \n\t"
-        "add r4, %[pixels], %[line_size]\n\t"
-        "tmcr wcgr2, r12                \n\t"
-        "add r5, %[block], %[line_size] \n\t"
-        "mov %[line_size], %[line_size], lsl #1 \n\t"
-
-        "1:                             \n\t"
-        "wldrd wr10, [%[pixels]]        \n\t"
-        "cmp r12, #8                    \n\t"
-        "wldrd wr11, [%[pixels], #8]    \n\t"
-        "wldrd wr12, [%[pixels], #16]   \n\t"
-        "add %[pixels], %[pixels], %[line_size] \n\t"
-        "wldrd wr13, [r4]               \n\t"
-        "pld [%[pixels]]                \n\t"
-        "wldrd wr14, [r4, #8]           \n\t"
-        "pld [%[pixels], #32]           \n\t"
-        "wldrd wr15, [r4, #16]          \n\t"
-        "add r4, r4, %[line_size]       \n\t"
-        "walignr1 wr0, wr10, wr11       \n\t"
-        "pld [r4]                       \n\t"
-        "pld [r4, #32]                  \n\t"
-        "walignr1 wr1, wr11, wr12       \n\t"
-        "walignr1 wr2, wr13, wr14       \n\t"
-        "walignr1 wr3, wr14, wr15       \n\t"
-        "wmoveq wr4, wr11               \n\t"
-        "wmoveq wr5, wr12               \n\t"
-        "wmoveq wr6, wr14               \n\t"
-        "wmoveq wr7, wr15               \n\t"
-        "walignr2ne wr4, wr10, wr11     \n\t"
-        "walignr2ne wr5, wr11, wr12     \n\t"
-        "walignr2ne wr6, wr13, wr14     \n\t"
-        "walignr2ne wr7, wr14, wr15     \n\t"
-        WAVG2B" wr0, wr0, wr4           \n\t"
-        WAVG2B" wr1, wr1, wr5           \n\t"
-        "wstrd wr0, [%[block]]          \n\t"
-        WAVG2B" wr2, wr2, wr6           \n\t"
-        "wstrd wr1, [%[block], #8]      \n\t"
-        WAVG2B" wr3, wr3, wr7           \n\t"
-        "add %[block], %[block], %[line_size]   \n\t"
-        "wstrd wr2, [r5]                \n\t"
-        "subs %[h], %[h], #2            \n\t"
-        "wstrd wr3, [r5, #8]            \n\t"
-        "add r5, r5, %[line_size]       \n\t"
-        "bne 1b                         \n\t"
-        : [h]"+r"(h), [pixels]"+r"(pixels), [block]"+r"(block), [line_size]"+r"(stride)
-        :
-        : "r4", "r5", "r12", "memory");
-}
-
-void DEF(avg, pixels8_x2)(uint8_t *block, const uint8_t *pixels, const int line_size, int h)
-{
-    int stride = line_size;
-    // [wr0 wr1 wr2 wr3] for previous line
-    // [wr4 wr5 wr6 wr7] for current line
-    SET_RND(wr15); // =2 for rnd  and  =1 for no_rnd version
-    __asm__ volatile(
-        "pld [%[pixels]]                \n\t"
-        "pld [%[pixels], #32]           \n\t"
-        "pld [%[block]]                 \n\t"
-        "pld [%[block], #32]            \n\t"
-        "and r12, %[pixels], #7         \n\t"
-        "bic %[pixels], %[pixels], #7   \n\t"
-        "tmcr wcgr1, r12                \n\t"
-        "add r12, r12, #1               \n\t"
-        "add r4, %[pixels], %[line_size]\n\t"
-        "tmcr wcgr2, r12                \n\t"
-        "add r5, %[block], %[line_size] \n\t"
-        "mov %[line_size], %[line_size], lsl #1 \n\t"
-        "pld [r5]                       \n\t"
-        "pld [r5, #32]                  \n\t"
-
-        "1:                             \n\t"
-        "wldrd wr10, [%[pixels]]        \n\t"
-        "cmp r12, #8                    \n\t"
-        "wldrd wr11, [%[pixels], #8]    \n\t"
-        "add %[pixels], %[pixels], %[line_size] \n\t"
-        "wldrd wr13, [r4]               \n\t"
-        "pld [%[pixels]]                \n\t"
-        "wldrd wr14, [r4, #8]           \n\t"
-        "pld [%[pixels], #32]           \n\t"
-        "add r4, r4, %[line_size]       \n\t"
-        "walignr1 wr0, wr10, wr11       \n\t"
-        "pld [r4]                       \n\t"
-        "pld [r4, #32]                  \n\t"
-        "walignr1 wr2, wr13, wr14       \n\t"
-        "wmoveq wr4, wr11               \n\t"
-        "wmoveq wr6, wr14               \n\t"
-        "walignr2ne wr4, wr10, wr11     \n\t"
-        "wldrd wr10, [%[block]]         \n\t"
-        "walignr2ne wr6, wr13, wr14     \n\t"
-        "wldrd wr12, [r5]               \n\t"
-        WAVG2B" wr0, wr0, wr4           \n\t"
-        WAVG2B" wr2, wr2, wr6           \n\t"
-        WAVG2B" wr0, wr0, wr10          \n\t"
-        WAVG2B" wr2, wr2, wr12          \n\t"
-        "wstrd wr0, [%[block]]          \n\t"
-        "subs %[h], %[h], #2            \n\t"
-        "wstrd wr2, [r5]                \n\t"
-        "add %[block], %[block], %[line_size]   \n\t"
-        "add r5, r5, %[line_size]       \n\t"
-        "pld [%[block]]                 \n\t"
-        "pld [%[block], #32]            \n\t"
-        "pld [r5]                       \n\t"
-        "pld [r5, #32]                  \n\t"
-        "bne 1b                         \n\t"
-        : [h]"+r"(h), [pixels]"+r"(pixels), [block]"+r"(block), [line_size]"+r"(stride)
-        :
-        : "r4", "r5", "r12", "memory");
-}
-
-void DEF(avg, pixels16_x2)(uint8_t *block, const uint8_t *pixels, const int line_size, int h)
-{
-    int stride = line_size;
-    // [wr0 wr1 wr2 wr3] for previous line
-    // [wr4 wr5 wr6 wr7] for current line
-    SET_RND(wr15); // =2 for rnd  and  =1 for no_rnd version
-    __asm__ volatile(
-        "pld [%[pixels]]                \n\t"
-        "pld [%[pixels], #32]           \n\t"
-        "pld [%[block]]                 \n\t"
-        "pld [%[block], #32]            \n\t"
-        "and r12, %[pixels], #7         \n\t"
-        "bic %[pixels], %[pixels], #7   \n\t"
-        "tmcr wcgr1, r12                \n\t"
-        "add r12, r12, #1               \n\t"
-        "add r4, %[pixels], %[line_size]\n\t"
-        "tmcr wcgr2, r12                \n\t"
-        "add r5, %[block], %[line_size] \n\t"
-        "mov %[line_size], %[line_size], lsl #1 \n\t"
-        "pld [r5]                       \n\t"
-        "pld [r5, #32]                  \n\t"
-
-        "1:                             \n\t"
-        "wldrd wr10, [%[pixels]]        \n\t"
-        "cmp r12, #8                    \n\t"
-        "wldrd wr11, [%[pixels], #8]    \n\t"
-        "wldrd wr12, [%[pixels], #16]   \n\t"
-        "add %[pixels], %[pixels], %[line_size] \n\t"
-        "wldrd wr13, [r4]               \n\t"
-        "pld [%[pixels]]                \n\t"
-        "wldrd wr14, [r4, #8]           \n\t"
-        "pld [%[pixels], #32]           \n\t"
-        "wldrd wr15, [r4, #16]          \n\t"
-        "add r4, r4, %[line_size]       \n\t"
-        "walignr1 wr0, wr10, wr11       \n\t"
-        "pld [r4]                       \n\t"
-        "pld [r4, #32]                  \n\t"
-        "walignr1 wr1, wr11, wr12       \n\t"
-        "walignr1 wr2, wr13, wr14       \n\t"
-        "walignr1 wr3, wr14, wr15       \n\t"
-        "wmoveq wr4, wr11               \n\t"
-        "wmoveq wr5, wr12               \n\t"
-        "wmoveq wr6, wr14               \n\t"
-        "wmoveq wr7, wr15               \n\t"
-        "walignr2ne wr4, wr10, wr11     \n\t"
-        "walignr2ne wr5, wr11, wr12     \n\t"
-        "walignr2ne wr6, wr13, wr14     \n\t"
-        "walignr2ne wr7, wr14, wr15     \n\t"
-        "wldrd wr10, [%[block]]         \n\t"
-        WAVG2B" wr0, wr0, wr4           \n\t"
-        "wldrd wr11, [%[block], #8]     \n\t"
-        WAVG2B" wr1, wr1, wr5           \n\t"
-        "wldrd wr12, [r5]               \n\t"
-        WAVG2B" wr2, wr2, wr6           \n\t"
-        "wldrd wr13, [r5, #8]           \n\t"
-        WAVG2B" wr3, wr3, wr7           \n\t"
-        WAVG2B" wr0, wr0, wr10          \n\t"
-        WAVG2B" wr1, wr1, wr11          \n\t"
-        WAVG2B" wr2, wr2, wr12          \n\t"
-        WAVG2B" wr3, wr3, wr13          \n\t"
-        "wstrd wr0, [%[block]]          \n\t"
-        "subs %[h], %[h], #2            \n\t"
-        "wstrd wr1, [%[block], #8]      \n\t"
-        "add %[block], %[block], %[line_size]   \n\t"
-        "wstrd wr2, [r5]                \n\t"
-        "pld [%[block]]                 \n\t"
-        "wstrd wr3, [r5, #8]            \n\t"
-        "add r5, r5, %[line_size]       \n\t"
-        "pld [%[block], #32]            \n\t"
-        "pld [r5]                       \n\t"
-        "pld [r5, #32]                  \n\t"
-        "bne 1b                         \n\t"
-        : [h]"+r"(h), [pixels]"+r"(pixels), [block]"+r"(block), [line_size]"+r"(stride)
-        :
-        :"r4", "r5", "r12", "memory");
-}
-
-void DEF(avg, pixels8_y2)(uint8_t *block, const uint8_t *pixels, const int line_size, int h)
-{
-    int stride = line_size;
-    // [wr0 wr1 wr2 wr3] for previous line
-    // [wr4 wr5 wr6 wr7] for current line
-    __asm__ volatile(
-        "pld            [%[pixels]]                             \n\t"
-        "pld            [%[pixels], #32]                        \n\t"
-        "and            r12, %[pixels], #7                      \n\t"
-        "tmcr           wcgr1, r12                              \n\t"
-        "bic            %[pixels], %[pixels], #7                \n\t"
-
-        "wldrd          wr10, [%[pixels]]                       \n\t"
-        "wldrd          wr11, [%[pixels], #8]                   \n\t"
-        "pld            [%[block]]                              \n\t"
-        "add            %[pixels], %[pixels], %[line_size]      \n\t"
-        "walignr1       wr0, wr10, wr11                         \n\t"
-        "pld            [%[pixels]]                             \n\t"
-        "pld            [%[pixels], #32]                        \n\t"
-
-      "1:                                                       \n\t"
-        "wldrd          wr10, [%[pixels]]                       \n\t"
-        "wldrd          wr11, [%[pixels], #8]                   \n\t"
-        "add            %[pixels], %[pixels], %[line_size]      \n\t"
-        "pld            [%[pixels]]                             \n\t"
-        "pld            [%[pixels], #32]                        \n\t"
-        "walignr1       wr4, wr10, wr11                         \n\t"
-        "wldrd          wr10, [%[block]]                        \n\t"
-         WAVG2B"        wr8, wr0, wr4                           \n\t"
-         WAVG2B"        wr8, wr8, wr10                          \n\t"
-        "wstrd          wr8, [%[block]]                         \n\t"
-        "add            %[block], %[block], %[line_size]        \n\t"
-
-        "wldrd          wr10, [%[pixels]]                       \n\t"
-        "wldrd          wr11, [%[pixels], #8]                   \n\t"
-        "pld            [%[block]]                              \n\t"
-        "add            %[pixels], %[pixels], %[line_size]      \n\t"
-        "pld            [%[pixels]]                             \n\t"
-        "pld            [%[pixels], #32]                        \n\t"
-        "walignr1       wr0, wr10, wr11                         \n\t"
-        "wldrd          wr10, [%[block]]                        \n\t"
-         WAVG2B"        wr8, wr0, wr4                           \n\t"
-         WAVG2B"        wr8, wr8, wr10                          \n\t"
-        "wstrd          wr8, [%[block]]                         \n\t"
-        "add            %[block], %[block], %[line_size]        \n\t"
-
-        "subs           %[h], %[h], #2                          \n\t"
-        "pld            [%[block]]                              \n\t"
-        "bne            1b                                      \n\t"
-        : [h]"+r"(h), [pixels]"+r"(pixels), [block]"+r"(block), [line_size]"+r"(stride)
-        :
-        : "cc", "memory", "r12");
-}
-
-void DEF(put, pixels16_y2)(uint8_t *block, const uint8_t *pixels, const int line_size, int h)
-{
-    int stride = line_size;
-    // [wr0 wr1 wr2 wr3] for previous line
-    // [wr4 wr5 wr6 wr7] for current line
-    __asm__ volatile(
-        "pld [%[pixels]]                \n\t"
-        "pld [%[pixels], #32]           \n\t"
-        "and r12, %[pixels], #7         \n\t"
-        "tmcr wcgr1, r12                \n\t"
-        "bic %[pixels], %[pixels], #7   \n\t"
-
-        "wldrd wr10, [%[pixels]]        \n\t"
-        "wldrd wr11, [%[pixels], #8]    \n\t"
-        "wldrd wr12, [%[pixels], #16]   \n\t"
-        "add %[pixels], %[pixels], %[line_size] \n\t"
-        "pld [%[pixels]]                \n\t"
-        "pld [%[pixels], #32]           \n\t"
-        "walignr1 wr0, wr10, wr11       \n\t"
-        "walignr1 wr1, wr11, wr12       \n\t"
-
-        "1:                             \n\t"
-        "wldrd wr10, [%[pixels]]        \n\t"
-        "wldrd wr11, [%[pixels], #8]    \n\t"
-        "wldrd wr12, [%[pixels], #16]   \n\t"
-        "add %[pixels], %[pixels], %[line_size] \n\t"
-        "pld [%[pixels]]                \n\t"
-        "pld [%[pixels], #32]           \n\t"
-        "walignr1 wr4, wr10, wr11       \n\t"
-        "walignr1 wr5, wr11, wr12       \n\t"
-        WAVG2B" wr8, wr0, wr4           \n\t"
-        WAVG2B" wr9, wr1, wr5           \n\t"
-        "wstrd wr8, [%[block]]          \n\t"
-        "wstrd wr9, [%[block], #8]      \n\t"
-        "add %[block], %[block], %[line_size]   \n\t"
-
-        "wldrd wr10, [%[pixels]]        \n\t"
-        "wldrd wr11, [%[pixels], #8]    \n\t"
-        "wldrd wr12, [%[pixels], #16]   \n\t"
-        "add %[pixels], %[pixels], %[line_size] \n\t"
-        "pld [%[pixels]]                \n\t"
-        "pld [%[pixels], #32]           \n\t"
-        "walignr1 wr0, wr10, wr11       \n\t"
-        "walignr1 wr1, wr11, wr12       \n\t"
-        WAVG2B" wr8, wr0, wr4           \n\t"
-        WAVG2B" wr9, wr1, wr5           \n\t"
-        "wstrd wr8, [%[block]]          \n\t"
-        "wstrd wr9, [%[block], #8]      \n\t"
-        "add %[block], %[block], %[line_size]   \n\t"
-
-        "subs %[h], %[h], #2            \n\t"
-        "bne 1b                         \n\t"
-        : [h]"+r"(h), [pixels]"+r"(pixels), [block]"+r"(block), [line_size]"+r"(stride)
-        :
-        : "r4", "r5", "r12", "memory");
-}
-
-void DEF(avg, pixels16_y2)(uint8_t *block, const uint8_t *pixels, const int line_size, int h)
-{
-    int stride = line_size;
-    // [wr0 wr1 wr2 wr3] for previous line
-    // [wr4 wr5 wr6 wr7] for current line
-    __asm__ volatile(
-        "pld [%[pixels]]                \n\t"
-        "pld [%[pixels], #32]           \n\t"
-        "and r12, %[pixels], #7         \n\t"
-        "tmcr wcgr1, r12                \n\t"
-        "bic %[pixels], %[pixels], #7   \n\t"
-
-        "wldrd wr10, [%[pixels]]        \n\t"
-        "wldrd wr11, [%[pixels], #8]    \n\t"
-        "pld [%[block]]                 \n\t"
-        "wldrd wr12, [%[pixels], #16]   \n\t"
-        "add %[pixels], %[pixels], %[line_size] \n\t"
-        "pld [%[pixels]]                \n\t"
-        "pld [%[pixels], #32]           \n\t"
-        "walignr1 wr0, wr10, wr11       \n\t"
-        "walignr1 wr1, wr11, wr12       \n\t"
-
-        "1:                             \n\t"
-        "wldrd wr10, [%[pixels]]        \n\t"
-        "wldrd wr11, [%[pixels], #8]    \n\t"
-        "wldrd wr12, [%[pixels], #16]   \n\t"
-        "add %[pixels], %[pixels], %[line_size] \n\t"
-        "pld [%[pixels]]                \n\t"
-        "pld [%[pixels], #32]           \n\t"
-        "walignr1 wr4, wr10, wr11       \n\t"
-        "walignr1 wr5, wr11, wr12       \n\t"
-        "wldrd wr10, [%[block]]         \n\t"
-        "wldrd wr11, [%[block], #8]     \n\t"
-        WAVG2B" wr8, wr0, wr4           \n\t"
-        WAVG2B" wr9, wr1, wr5           \n\t"
-        WAVG2B" wr8, wr8, wr10          \n\t"
-        WAVG2B" wr9, wr9, wr11          \n\t"
-        "wstrd wr8, [%[block]]          \n\t"
-        "wstrd wr9, [%[block], #8]      \n\t"
-        "add %[block], %[block], %[line_size]   \n\t"
-
-        "wldrd wr10, [%[pixels]]        \n\t"
-        "wldrd wr11, [%[pixels], #8]    \n\t"
-        "pld [%[block]]                 \n\t"
-        "wldrd wr12, [%[pixels], #16]   \n\t"
-        "add %[pixels], %[pixels], %[line_size] \n\t"
-        "pld [%[pixels]]                \n\t"
-        "pld [%[pixels], #32]           \n\t"
-        "walignr1 wr0, wr10, wr11       \n\t"
-        "walignr1 wr1, wr11, wr12       \n\t"
-        "wldrd wr10, [%[block]]         \n\t"
-        "wldrd wr11, [%[block], #8]     \n\t"
-        WAVG2B" wr8, wr0, wr4           \n\t"
-        WAVG2B" wr9, wr1, wr5           \n\t"
-        WAVG2B" wr8, wr8, wr10          \n\t"
-        WAVG2B" wr9, wr9, wr11          \n\t"
-        "wstrd wr8, [%[block]]          \n\t"
-        "wstrd wr9, [%[block], #8]      \n\t"
-        "add %[block], %[block], %[line_size]   \n\t"
-
-        "subs %[h], %[h], #2            \n\t"
-        "pld [%[block]]                 \n\t"
-        "bne 1b                         \n\t"
-        : [h]"+r"(h), [pixels]"+r"(pixels), [block]"+r"(block), [line_size]"+r"(stride)
-        :
-        : "r4", "r5", "r12", "memory");
-}
-
-void DEF(put, pixels8_xy2)(uint8_t *block, const uint8_t *pixels, const int line_size, int h)
-{
-    // [wr0 wr1 wr2 wr3] for previous line
-    // [wr4 wr5 wr6 wr7] for current line
-    SET_RND(wr15); // =2 for rnd  and  =1 for no_rnd version
-    __asm__ volatile(
-        "pld [%[pixels]]                \n\t"
-        "mov r12, #2                    \n\t"
-        "pld [%[pixels], #32]           \n\t"
-        "tmcr wcgr0, r12                \n\t" /* for shift value */
-        "and r12, %[pixels], #7         \n\t"
-        "bic %[pixels], %[pixels], #7   \n\t"
-        "tmcr wcgr1, r12                \n\t"
-
-        // [wr0 wr1 wr2 wr3] <= *
-        // [wr4 wr5 wr6 wr7]
-        "wldrd wr12, [%[pixels]]        \n\t"
-        "add r12, r12, #1               \n\t"
-        "wldrd wr13, [%[pixels], #8]    \n\t"
-        "tmcr wcgr2, r12                \n\t"
-        "add %[pixels], %[pixels], %[line_size] \n\t"
-        "cmp r12, #8                    \n\t"
-        "pld [%[pixels]]                \n\t"
-        "pld [%[pixels], #32]           \n\t"
-        "walignr1 wr2, wr12, wr13       \n\t"
-        "wmoveq wr10, wr13              \n\t"
-        "walignr2ne wr10, wr12, wr13    \n\t"
-        "wunpckelub wr0, wr2            \n\t"
-        "wunpckehub wr1, wr2            \n\t"
-        "wunpckelub wr8, wr10           \n\t"
-        "wunpckehub wr9, wr10           \n\t"
-        "waddhus wr0, wr0, wr8          \n\t"
-        "waddhus wr1, wr1, wr9          \n\t"
-
-        "1:                             \n\t"
-        // [wr0 wr1 wr2 wr3]
-        // [wr4 wr5 wr6 wr7] <= *
-        "wldrd wr12, [%[pixels]]        \n\t"
-        "cmp r12, #8                    \n\t"
-        "wldrd wr13, [%[pixels], #8]    \n\t"
-        "add %[pixels], %[pixels], %[line_size] \n\t"
-        "walignr1 wr6, wr12, wr13       \n\t"
-        "pld [%[pixels]]                \n\t"
-        "pld [%[pixels], #32]           \n\t"
-        "wmoveq wr10, wr13              \n\t"
-        "walignr2ne wr10, wr12, wr13    \n\t"
-        "wunpckelub wr4, wr6            \n\t"
-        "wunpckehub wr5, wr6            \n\t"
-        "wunpckelub wr8, wr10           \n\t"
-        "wunpckehub wr9, wr10           \n\t"
-        "waddhus wr4, wr4, wr8          \n\t"
-        "waddhus wr5, wr5, wr9          \n\t"
-        "waddhus wr8, wr0, wr4          \n\t"
-        "waddhus wr9, wr1, wr5          \n\t"
-        "waddhus wr8, wr8, wr15         \n\t"
-        "waddhus wr9, wr9, wr15         \n\t"
-        "wsrlhg wr8, wr8, wcgr0         \n\t"
-        "wsrlhg wr9, wr9, wcgr0         \n\t"
-        "wpackhus wr8, wr8, wr9         \n\t"
-        "wstrd wr8, [%[block]]          \n\t"
-        "add %[block], %[block], %[line_size]   \n\t"
-
-        // [wr0 wr1 wr2 wr3] <= *
-        // [wr4 wr5 wr6 wr7]
-        "wldrd wr12, [%[pixels]]        \n\t"
-        "wldrd wr13, [%[pixels], #8]    \n\t"
-        "add %[pixels], %[pixels], %[line_size] \n\t"
-        "walignr1 wr2, wr12, wr13       \n\t"
-        "pld [%[pixels]]                \n\t"
-        "pld [%[pixels], #32]           \n\t"
-        "wmoveq wr10, wr13              \n\t"
-        "walignr2ne wr10, wr12, wr13    \n\t"
-        "wunpckelub wr0, wr2            \n\t"
-        "wunpckehub wr1, wr2            \n\t"
-        "wunpckelub wr8, wr10           \n\t"
-        "wunpckehub wr9, wr10           \n\t"
-        "waddhus wr0, wr0, wr8          \n\t"
-        "waddhus wr1, wr1, wr9          \n\t"
-        "waddhus wr8, wr0, wr4          \n\t"
-        "waddhus wr9, wr1, wr5          \n\t"
-        "waddhus wr8, wr8, wr15         \n\t"
-        "waddhus wr9, wr9, wr15         \n\t"
-        "wsrlhg wr8, wr8, wcgr0         \n\t"
-        "wsrlhg wr9, wr9, wcgr0         \n\t"
-        "wpackhus wr8, wr8, wr9         \n\t"
-        "subs %[h], %[h], #2            \n\t"
-        "wstrd wr8, [%[block]]          \n\t"
-        "add %[block], %[block], %[line_size]   \n\t"
-        "bne 1b                         \n\t"
-        : [h]"+r"(h), [pixels]"+r"(pixels), [block]"+r"(block)
-        : [line_size]"r"(line_size)
-        : "r12", "memory");
-}
-
-void DEF(put, pixels16_xy2)(uint8_t *block, const uint8_t *pixels, const int line_size, int h)
-{
-    // [wr0 wr1 wr2 wr3] for previous line
-    // [wr4 wr5 wr6 wr7] for current line
-    SET_RND(wr15); // =2 for rnd  and  =1 for no_rnd version
-    __asm__ volatile(
-        "pld [%[pixels]]                \n\t"
-        "mov r12, #2                    \n\t"
-        "pld [%[pixels], #32]           \n\t"
-        "tmcr wcgr0, r12                \n\t" /* for shift value */
-        /* alignment */
-        "and r12, %[pixels], #7         \n\t"
-        "bic %[pixels], %[pixels], #7   \n\t"
-        "tmcr wcgr1, r12                \n\t"
-        "add r12, r12, #1               \n\t"
-        "tmcr wcgr2, r12                \n\t"
-
-        // [wr0 wr1 wr2 wr3] <= *
-        // [wr4 wr5 wr6 wr7]
-        "wldrd wr12, [%[pixels]]        \n\t"
-        "cmp r12, #8                    \n\t"
-        "wldrd wr13, [%[pixels], #8]    \n\t"
-        "wldrd wr14, [%[pixels], #16]   \n\t"
-        "add %[pixels], %[pixels], %[line_size] \n\t"
-        "pld [%[pixels]]                \n\t"
-        "walignr1 wr2, wr12, wr13       \n\t"
-        "pld [%[pixels], #32]           \n\t"
-        "walignr1 wr3, wr13, wr14       \n\t"
-        "wmoveq wr10, wr13              \n\t"
-        "wmoveq wr11, wr14              \n\t"
-        "walignr2ne wr10, wr12, wr13    \n\t"
-        "walignr2ne wr11, wr13, wr14    \n\t"
-        "wunpckelub wr0, wr2            \n\t"
-        "wunpckehub wr1, wr2            \n\t"
-        "wunpckelub wr2, wr3            \n\t"
-        "wunpckehub wr3, wr3            \n\t"
-        "wunpckelub wr8, wr10           \n\t"
-        "wunpckehub wr9, wr10           \n\t"
-        "wunpckelub wr10, wr11          \n\t"
-        "wunpckehub wr11, wr11          \n\t"
-        "waddhus wr0, wr0, wr8          \n\t"
-        "waddhus wr1, wr1, wr9          \n\t"
-        "waddhus wr2, wr2, wr10         \n\t"
-        "waddhus wr3, wr3, wr11         \n\t"
-
-        "1:                             \n\t"
-        // [wr0 wr1 wr2 wr3]
-        // [wr4 wr5 wr6 wr7] <= *
-        "wldrd wr12, [%[pixels]]        \n\t"
-        "cmp r12, #8                    \n\t"
-        "wldrd wr13, [%[pixels], #8]    \n\t"
-        "wldrd wr14, [%[pixels], #16]   \n\t"
-        "add %[pixels], %[pixels], %[line_size] \n\t"
-        "walignr1 wr6, wr12, wr13       \n\t"
-        "pld [%[pixels]]                \n\t"
-        "pld [%[pixels], #32]           \n\t"
-        "walignr1 wr7, wr13, wr14       \n\t"
-        "wmoveq wr10, wr13              \n\t"
-        "wmoveq wr11, wr14              \n\t"
-        "walignr2ne wr10, wr12, wr13    \n\t"
-        "walignr2ne wr11, wr13, wr14    \n\t"
-        "wunpckelub wr4, wr6            \n\t"
-        "wunpckehub wr5, wr6            \n\t"
-        "wunpckelub wr6, wr7            \n\t"
-        "wunpckehub wr7, wr7            \n\t"
-        "wunpckelub wr8, wr10           \n\t"
-        "wunpckehub wr9, wr10           \n\t"
-        "wunpckelub wr10, wr11          \n\t"
-        "wunpckehub wr11, wr11          \n\t"
-        "waddhus wr4, wr4, wr8          \n\t"
-        "waddhus wr5, wr5, wr9          \n\t"
-        "waddhus wr6, wr6, wr10         \n\t"
-        "waddhus wr7, wr7, wr11         \n\t"
-        "waddhus wr8, wr0, wr4          \n\t"
-        "waddhus wr9, wr1, wr5          \n\t"
-        "waddhus wr10, wr2, wr6         \n\t"
-        "waddhus wr11, wr3, wr7         \n\t"
-        "waddhus wr8, wr8, wr15         \n\t"
-        "waddhus wr9, wr9, wr15         \n\t"
-        "waddhus wr10, wr10, wr15       \n\t"
-        "waddhus wr11, wr11, wr15       \n\t"
-        "wsrlhg wr8, wr8, wcgr0         \n\t"
-        "wsrlhg wr9, wr9, wcgr0         \n\t"
-        "wsrlhg wr10, wr10, wcgr0       \n\t"
-        "wsrlhg wr11, wr11, wcgr0       \n\t"
-        "wpackhus wr8, wr8, wr9         \n\t"
-        "wpackhus wr9, wr10, wr11       \n\t"
-        "wstrd wr8, [%[block]]          \n\t"
-        "wstrd wr9, [%[block], #8]      \n\t"
-        "add %[block], %[block], %[line_size]   \n\t"
-
-        // [wr0 wr1 wr2 wr3] <= *
-        // [wr4 wr5 wr6 wr7]
-        "wldrd wr12, [%[pixels]]        \n\t"
-        "wldrd wr13, [%[pixels], #8]    \n\t"
-        "wldrd wr14, [%[pixels], #16]   \n\t"
-        "add %[pixels], %[pixels], %[line_size] \n\t"
-        "walignr1 wr2, wr12, wr13       \n\t"
-        "pld [%[pixels]]                \n\t"
-        "pld [%[pixels], #32]           \n\t"
-        "walignr1 wr3, wr13, wr14       \n\t"
-        "wmoveq wr10, wr13              \n\t"
-        "wmoveq wr11, wr14              \n\t"
-        "walignr2ne wr10, wr12, wr13    \n\t"
-        "walignr2ne wr11, wr13, wr14    \n\t"
-        "wunpckelub wr0, wr2            \n\t"
-        "wunpckehub wr1, wr2            \n\t"
-        "wunpckelub wr2, wr3            \n\t"
-        "wunpckehub wr3, wr3            \n\t"
-        "wunpckelub wr8, wr10           \n\t"
-        "wunpckehub wr9, wr10           \n\t"
-        "wunpckelub wr10, wr11          \n\t"
-        "wunpckehub wr11, wr11          \n\t"
-        "waddhus wr0, wr0, wr8          \n\t"
-        "waddhus wr1, wr1, wr9          \n\t"
-        "waddhus wr2, wr2, wr10         \n\t"
-        "waddhus wr3, wr3, wr11         \n\t"
-        "waddhus wr8, wr0, wr4          \n\t"
-        "waddhus wr9, wr1, wr5          \n\t"
-        "waddhus wr10, wr2, wr6         \n\t"
-        "waddhus wr11, wr3, wr7         \n\t"
-        "waddhus wr8, wr8, wr15         \n\t"
-        "waddhus wr9, wr9, wr15         \n\t"
-        "waddhus wr10, wr10, wr15       \n\t"
-        "waddhus wr11, wr11, wr15       \n\t"
-        "wsrlhg wr8, wr8, wcgr0         \n\t"
-        "wsrlhg wr9, wr9, wcgr0         \n\t"
-        "wsrlhg wr10, wr10, wcgr0       \n\t"
-        "wsrlhg wr11, wr11, wcgr0       \n\t"
-        "wpackhus wr8, wr8, wr9         \n\t"
-        "wpackhus wr9, wr10, wr11       \n\t"
-        "wstrd wr8, [%[block]]          \n\t"
-        "wstrd wr9, [%[block], #8]      \n\t"
-        "add %[block], %[block], %[line_size]   \n\t"
-
-        "subs %[h], %[h], #2            \n\t"
-        "bne 1b                         \n\t"
-        : [h]"+r"(h), [pixels]"+r"(pixels), [block]"+r"(block)
-        : [line_size]"r"(line_size)
-        : "r12", "memory");
-}
-
-void DEF(avg, pixels8_xy2)(uint8_t *block, const uint8_t *pixels, const int line_size, int h)
-{
-    // [wr0 wr1 wr2 wr3] for previous line
-    // [wr4 wr5 wr6 wr7] for current line
-    SET_RND(wr15); // =2 for rnd  and  =1 for no_rnd version
-    __asm__ volatile(
-        "pld [%[block]]                 \n\t"
-        "pld [%[block], #32]            \n\t"
-        "pld [%[pixels]]                \n\t"
-        "mov r12, #2                    \n\t"
-        "pld [%[pixels], #32]           \n\t"
-        "tmcr wcgr0, r12                \n\t" /* for shift value */
-        "and r12, %[pixels], #7         \n\t"
-        "bic %[pixels], %[pixels], #7   \n\t"
-        "tmcr wcgr1, r12                \n\t"
-
-        // [wr0 wr1 wr2 wr3] <= *
-        // [wr4 wr5 wr6 wr7]
-        "wldrd wr12, [%[pixels]]        \n\t"
-        "add r12, r12, #1               \n\t"
-        "wldrd wr13, [%[pixels], #8]    \n\t"
-        "tmcr wcgr2, r12                \n\t"
-        "add %[pixels], %[pixels], %[line_size] \n\t"
-        "cmp r12, #8                    \n\t"
-        "pld [%[pixels]]                \n\t"
-        "pld [%[pixels], #32]           \n\t"
-        "walignr1 wr2, wr12, wr13       \n\t"
-        "wmoveq wr10, wr13              \n\t"
-        "walignr2ne wr10, wr12, wr13    \n\t"
-        "wunpckelub wr0, wr2            \n\t"
-        "wunpckehub wr1, wr2            \n\t"
-        "wunpckelub wr8, wr10           \n\t"
-        "wunpckehub wr9, wr10           \n\t"
-        "waddhus wr0, wr0, wr8          \n\t"
-        "waddhus wr1, wr1, wr9          \n\t"
-
-        "1:                             \n\t"
-        // [wr0 wr1 wr2 wr3]
-        // [wr4 wr5 wr6 wr7] <= *
-        "wldrd wr12, [%[pixels]]        \n\t"
-        "cmp r12, #8                    \n\t"
-        "wldrd wr13, [%[pixels], #8]    \n\t"
-        "add %[pixels], %[pixels], %[line_size] \n\t"
-        "walignr1 wr6, wr12, wr13       \n\t"
-        "pld [%[pixels]]                \n\t"
-        "pld [%[pixels], #32]           \n\t"
-        "wmoveq wr10, wr13              \n\t"
-        "walignr2ne wr10, wr12, wr13    \n\t"
-        "wunpckelub wr4, wr6            \n\t"
-        "wunpckehub wr5, wr6            \n\t"
-        "wunpckelub wr8, wr10           \n\t"
-        "wunpckehub wr9, wr10           \n\t"
-        "waddhus wr4, wr4, wr8          \n\t"
-        "waddhus wr5, wr5, wr9          \n\t"
-        "waddhus wr8, wr0, wr4          \n\t"
-        "waddhus wr9, wr1, wr5          \n\t"
-        "waddhus wr8, wr8, wr15         \n\t"
-        "waddhus wr9, wr9, wr15         \n\t"
-        "wldrd wr12, [%[block]]         \n\t"
-        "wsrlhg wr8, wr8, wcgr0         \n\t"
-        "wsrlhg wr9, wr9, wcgr0         \n\t"
-        "wpackhus wr8, wr8, wr9         \n\t"
-        WAVG2B" wr8, wr8, wr12          \n\t"
-        "wstrd wr8, [%[block]]          \n\t"
-        "add %[block], %[block], %[line_size]   \n\t"
-        "wldrd wr12, [%[pixels]]        \n\t"
-        "pld [%[block]]                 \n\t"
-        "pld [%[block], #32]            \n\t"
-
-        // [wr0 wr1 wr2 wr3] <= *
-        // [wr4 wr5 wr6 wr7]
-        "wldrd wr13, [%[pixels], #8]    \n\t"
-        "add %[pixels], %[pixels], %[line_size] \n\t"
-        "walignr1 wr2, wr12, wr13       \n\t"
-        "pld [%[pixels]]                \n\t"
-        "pld [%[pixels], #32]           \n\t"
-        "wmoveq wr10, wr13              \n\t"
-        "walignr2ne wr10, wr12, wr13    \n\t"
-        "wunpckelub wr0, wr2            \n\t"
-        "wunpckehub wr1, wr2            \n\t"
-        "wunpckelub wr8, wr10           \n\t"
-        "wunpckehub wr9, wr10           \n\t"
-        "waddhus wr0, wr0, wr8          \n\t"
-        "waddhus wr1, wr1, wr9          \n\t"
-        "waddhus wr8, wr0, wr4          \n\t"
-        "waddhus wr9, wr1, wr5          \n\t"
-        "waddhus wr8, wr8, wr15         \n\t"
-        "waddhus wr9, wr9, wr15         \n\t"
-        "wldrd wr12, [%[block]]         \n\t"
-        "wsrlhg wr8, wr8, wcgr0         \n\t"
-        "wsrlhg wr9, wr9, wcgr0         \n\t"
-        "wpackhus wr8, wr8, wr9         \n\t"
-        "subs %[h], %[h], #2            \n\t"
-        WAVG2B" wr8, wr8, wr12          \n\t"
-        "wstrd wr8, [%[block]]          \n\t"
-        "add %[block], %[block], %[line_size]   \n\t"
-        "pld [%[block]]                 \n\t"
-        "pld [%[block], #32]            \n\t"
-        "bne 1b                         \n\t"
-        : [h]"+r"(h), [pixels]"+r"(pixels), [block]"+r"(block)
-        : [line_size]"r"(line_size)
-        : "r12", "memory");
-}
-
-void DEF(avg, pixels16_xy2)(uint8_t *block, const uint8_t *pixels, const int line_size, int h)
-{
-    // [wr0 wr1 wr2 wr3] for previous line
-    // [wr4 wr5 wr6 wr7] for current line
-    SET_RND(wr15); // =2 for rnd  and  =1 for no_rnd version
-    __asm__ volatile(
-        "pld [%[block]]                 \n\t"
-        "pld [%[block], #32]            \n\t"
-        "pld [%[pixels]]                \n\t"
-        "mov r12, #2                    \n\t"
-        "pld [%[pixels], #32]           \n\t"
-        "tmcr wcgr0, r12                \n\t" /* for shift value */
-        /* alignment */
-        "and r12, %[pixels], #7         \n\t"
-        "bic %[pixels], %[pixels], #7           \n\t"
-        "tmcr wcgr1, r12                \n\t"
-        "add r12, r12, #1               \n\t"
-        "tmcr wcgr2, r12                \n\t"
-
-        // [wr0 wr1 wr2 wr3] <= *
-        // [wr4 wr5 wr6 wr7]
-        "wldrd wr12, [%[pixels]]        \n\t"
-        "cmp r12, #8                    \n\t"
-        "wldrd wr13, [%[pixels], #8]    \n\t"
-        "wldrd wr14, [%[pixels], #16]   \n\t"
-        "add %[pixels], %[pixels], %[line_size] \n\t"
-        "pld [%[pixels]]                \n\t"
-        "walignr1 wr2, wr12, wr13       \n\t"
-        "pld [%[pixels], #32]           \n\t"
-        "walignr1 wr3, wr13, wr14       \n\t"
-        "wmoveq wr10, wr13              \n\t"
-        "wmoveq wr11, wr14              \n\t"
-        "walignr2ne wr10, wr12, wr13    \n\t"
-        "walignr2ne wr11, wr13, wr14    \n\t"
-        "wunpckelub wr0, wr2            \n\t"
-        "wunpckehub wr1, wr2            \n\t"
-        "wunpckelub wr2, wr3            \n\t"
-        "wunpckehub wr3, wr3            \n\t"
-        "wunpckelub wr8, wr10           \n\t"
-        "wunpckehub wr9, wr10           \n\t"
-        "wunpckelub wr10, wr11          \n\t"
-        "wunpckehub wr11, wr11          \n\t"
-        "waddhus wr0, wr0, wr8          \n\t"
-        "waddhus wr1, wr1, wr9          \n\t"
-        "waddhus wr2, wr2, wr10         \n\t"
-        "waddhus wr3, wr3, wr11         \n\t"
-
-        "1:                             \n\t"
-        // [wr0 wr1 wr2 wr3]
-        // [wr4 wr5 wr6 wr7] <= *
-        "wldrd wr12, [%[pixels]]        \n\t"
-        "cmp r12, #8                    \n\t"
-        "wldrd wr13, [%[pixels], #8]    \n\t"
-        "wldrd wr14, [%[pixels], #16]   \n\t"
-        "add %[pixels], %[pixels], %[line_size] \n\t"
-        "walignr1 wr6, wr12, wr13       \n\t"
-        "pld [%[pixels]]                \n\t"
-        "pld [%[pixels], #32]           \n\t"
-        "walignr1 wr7, wr13, wr14       \n\t"
-        "wmoveq wr10, wr13              \n\t"
-        "wmoveq wr11, wr14              \n\t"
-        "walignr2ne wr10, wr12, wr13    \n\t"
-        "walignr2ne wr11, wr13, wr14    \n\t"
-        "wunpckelub wr4, wr6            \n\t"
-        "wunpckehub wr5, wr6            \n\t"
-        "wunpckelub wr6, wr7            \n\t"
-        "wunpckehub wr7, wr7            \n\t"
-        "wunpckelub wr8, wr10           \n\t"
-        "wunpckehub wr9, wr10           \n\t"
-        "wunpckelub wr10, wr11          \n\t"
-        "wunpckehub wr11, wr11          \n\t"
-        "waddhus wr4, wr4, wr8          \n\t"
-        "waddhus wr5, wr5, wr9          \n\t"
-        "waddhus wr6, wr6, wr10         \n\t"
-        "waddhus wr7, wr7, wr11         \n\t"
-        "waddhus wr8, wr0, wr4          \n\t"
-        "waddhus wr9, wr1, wr5          \n\t"
-        "waddhus wr10, wr2, wr6         \n\t"
-        "waddhus wr11, wr3, wr7         \n\t"
-        "waddhus wr8, wr8, wr15         \n\t"
-        "waddhus wr9, wr9, wr15         \n\t"
-        "waddhus wr10, wr10, wr15       \n\t"
-        "waddhus wr11, wr11, wr15       \n\t"
-        "wsrlhg wr8, wr8, wcgr0         \n\t"
-        "wsrlhg wr9, wr9, wcgr0         \n\t"
-        "wldrd wr12, [%[block]]         \n\t"
-        "wldrd wr13, [%[block], #8]     \n\t"
-        "wsrlhg wr10, wr10, wcgr0       \n\t"
-        "wsrlhg wr11, wr11, wcgr0       \n\t"
-        "wpackhus wr8, wr8, wr9         \n\t"
-        "wpackhus wr9, wr10, wr11       \n\t"
-        WAVG2B" wr8, wr8, wr12          \n\t"
-        WAVG2B" wr9, wr9, wr13          \n\t"
-        "wstrd wr8, [%[block]]          \n\t"
-        "wstrd wr9, [%[block], #8]      \n\t"
-        "add %[block], %[block], %[line_size]   \n\t"
-
-        // [wr0 wr1 wr2 wr3] <= *
-        // [wr4 wr5 wr6 wr7]
-        "wldrd wr12, [%[pixels]]        \n\t"
-        "pld [%[block]]                 \n\t"
-        "wldrd wr13, [%[pixels], #8]    \n\t"
-        "pld [%[block], #32]            \n\t"
-        "wldrd wr14, [%[pixels], #16]   \n\t"
-        "add %[pixels], %[pixels], %[line_size] \n\t"
-        "walignr1 wr2, wr12, wr13       \n\t"
-        "pld [%[pixels]]                \n\t"
-        "pld [%[pixels], #32]           \n\t"
-        "walignr1 wr3, wr13, wr14       \n\t"
-        "wmoveq wr10, wr13              \n\t"
-        "wmoveq wr11, wr14              \n\t"
-        "walignr2ne wr10, wr12, wr13    \n\t"
-        "walignr2ne wr11, wr13, wr14    \n\t"
-        "wunpckelub wr0, wr2            \n\t"
-        "wunpckehub wr1, wr2            \n\t"
-        "wunpckelub wr2, wr3            \n\t"
-        "wunpckehub wr3, wr3            \n\t"
-        "wunpckelub wr8, wr10           \n\t"
-        "wunpckehub wr9, wr10           \n\t"
-        "wunpckelub wr10, wr11          \n\t"
-        "wunpckehub wr11, wr11          \n\t"
-        "waddhus wr0, wr0, wr8          \n\t"
-        "waddhus wr1, wr1, wr9          \n\t"
-        "waddhus wr2, wr2, wr10         \n\t"
-        "waddhus wr3, wr3, wr11         \n\t"
-        "waddhus wr8, wr0, wr4          \n\t"
-        "waddhus wr9, wr1, wr5          \n\t"
-        "waddhus wr10, wr2, wr6         \n\t"
-        "waddhus wr11, wr3, wr7         \n\t"
-        "waddhus wr8, wr8, wr15         \n\t"
-        "waddhus wr9, wr9, wr15         \n\t"
-        "waddhus wr10, wr10, wr15       \n\t"
-        "waddhus wr11, wr11, wr15       \n\t"
-        "wsrlhg wr8, wr8, wcgr0         \n\t"
-        "wsrlhg wr9, wr9, wcgr0         \n\t"
-        "wldrd wr12, [%[block]]         \n\t"
-        "wldrd wr13, [%[block], #8]     \n\t"
-        "wsrlhg wr10, wr10, wcgr0       \n\t"
-        "wsrlhg wr11, wr11, wcgr0       \n\t"
-        "wpackhus wr8, wr8, wr9         \n\t"
-        "wpackhus wr9, wr10, wr11       \n\t"
-        WAVG2B" wr8, wr8, wr12          \n\t"
-        WAVG2B" wr9, wr9, wr13          \n\t"
-        "wstrd wr8, [%[block]]          \n\t"
-        "wstrd wr9, [%[block], #8]      \n\t"
-        "add %[block], %[block], %[line_size]   \n\t"
-        "subs %[h], %[h], #2            \n\t"
-        "pld [%[block]]                 \n\t"
-        "pld [%[block], #32]            \n\t"
-        "bne 1b                         \n\t"
-        : [h]"+r"(h), [pixels]"+r"(pixels), [block]"+r"(block)
-        : [line_size]"r"(line_size)
-        : "r12", "memory");
-}
diff --git a/libavcodec/arm/mpegvideo_arm.c b/libavcodec/arm/mpegvideo_arm.c
index b2b254e2d3..73d03c1a5e 100644
--- a/libavcodec/arm/mpegvideo_arm.c
+++ b/libavcodec/arm/mpegvideo_arm.c
@@ -40,16 +40,9 @@ void ff_dct_unquantize_h263_intra_neon(MpegEncContext *s, DCTELEM *block,
 
 void ff_MPV_common_init_arm(MpegEncContext *s)
 {
-    /* IWMMXT support is a superset of armv5te, so
-     * allow optimized functions for armv5te unless
-     * a better iwmmxt function exists
-     */
 #if HAVE_ARMV5TE
     ff_MPV_common_init_armv5te(s);
 #endif
-#if HAVE_IWMMXT
-    ff_MPV_common_init_iwmmxt(s);
-#endif
 
     if (HAVE_NEON) {
         s->dct_unquantize_h263_intra = ff_dct_unquantize_h263_intra_neon;
diff --git a/libavcodec/arm/mpegvideo_arm.h b/libavcodec/arm/mpegvideo_arm.h
index e9de979300..4ff93b76b9 100644
--- a/libavcodec/arm/mpegvideo_arm.h
+++ b/libavcodec/arm/mpegvideo_arm.h
@@ -21,7 +21,6 @@
 
 #include "libavcodec/mpegvideo.h"
 
-void ff_MPV_common_init_iwmmxt(MpegEncContext *s);
 void ff_MPV_common_init_armv5te(MpegEncContext *s);
 
 #endif /* AVCODEC_ARM_MPEGVIDEO_H */
diff --git a/libavcodec/arm/mpegvideo_iwmmxt.c b/libavcodec/arm/mpegvideo_iwmmxt.c
deleted file mode 100644
index c85e6dc8b1..0000000000
--- a/libavcodec/arm/mpegvideo_iwmmxt.c
+++ /dev/null
@@ -1,101 +0,0 @@
-/*
- * copyright (c) 2004 AGAWA Koji
- *
- * This file is part of FFmpeg.
- *
- * FFmpeg 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.
- *
- * FFmpeg 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 FFmpeg; if not, write to the Free Software
- * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
- */
-
-#include "libavutil/cpu.h"
-#include "libavcodec/avcodec.h"
-#include "libavcodec/dsputil.h"
-#include "libavcodec/mpegvideo.h"
-#include "mpegvideo_arm.h"
-
-static void dct_unquantize_h263_intra_iwmmxt(MpegEncContext *s,
-                                             DCTELEM *block, int n, int qscale)
-{
-    int level, qmul, qadd;
-    int nCoeffs;
-    DCTELEM *block_orig = block;
-
-    assert(s->block_last_index[n]>=0);
-
-    qmul = qscale << 1;
-
-    if (!s->h263_aic) {
-        if (n < 4)
-            level = block[0] * s->y_dc_scale;
-        else
-            level = block[0] * s->c_dc_scale;
-        qadd = (qscale - 1) | 1;
-    }else{
-        qadd = 0;
-        level = block[0];
-    }
-    if(s->ac_pred)
-        nCoeffs=63;
-    else
-        nCoeffs= s->inter_scantable.raster_end[ s->block_last_index[n] ];
-
-    __asm__ volatile (
-/*      "movd %1, %%mm6                 \n\t" //qmul */
-/*      "packssdw %%mm6, %%mm6          \n\t" */
-/*      "packssdw %%mm6, %%mm6          \n\t" */
-        "tbcsth wr6, %[qmul]            \n\t"
-/*      "movd %2, %%mm5                 \n\t" //qadd */
-/*      "packssdw %%mm5, %%mm5          \n\t" */
-/*      "packssdw %%mm5, %%mm5          \n\t" */
-        "tbcsth wr5, %[qadd]            \n\t"
-        "wzero wr7                      \n\t" /* "pxor %%mm7, %%mm7             \n\t" */
-        "wzero wr4                      \n\t" /* "pxor %%mm4, %%mm4             \n\t" */
-        "wsubh wr7, wr5, wr7            \n\t" /* "psubw %%mm5, %%mm7            \n\t" */
-        "1:                             \n\t"
-        "wldrd wr2, [%[block]]          \n\t" /* "movq (%0, %3), %%mm0          \n\t" */
-        "wldrd wr3, [%[block], #8]      \n\t" /* "movq 8(%0, %3), %%mm1         \n\t" */
-        "wmulsl wr0, wr6, wr2           \n\t" /* "pmullw %%mm6, %%mm0           \n\t" */
-        "wmulsl wr1, wr6, wr3           \n\t" /* "pmullw %%mm6, %%mm1           \n\t" */
-/*      "movq (%0, %3), %%mm2           \n\t" */
-/*      "movq 8(%0, %3), %%mm3          \n\t" */
-        "wcmpgtsh wr2, wr4, wr2         \n\t" /* "pcmpgtw %%mm4, %%mm2          \n\t" // block[i] < 0 ? -1 : 0 */
-        "wcmpgtsh wr3, wr4, wr2         \n\t" /* "pcmpgtw %%mm4, %%mm3          \n\t" // block[i] < 0 ? -1 : 0 */
-        "wxor wr0, wr2, wr0             \n\t" /* "pxor %%mm2, %%mm0             \n\t" */
-        "wxor wr1, wr3, wr1             \n\t" /* "pxor %%mm3, %%mm1             \n\t" */
-        "waddh wr0, wr7, wr0            \n\t" /* "paddw %%mm7, %%mm0            \n\t" */
-        "waddh wr1, wr7, wr1            \n\t" /* "paddw %%mm7, %%mm1            \n\t" */
-        "wxor wr2, wr0, wr2             \n\t" /* "pxor %%mm0, %%mm2             \n\t" */
-        "wxor wr3, wr1, wr3             \n\t" /* "pxor %%mm1, %%mm3             \n\t" */
-        "wcmpeqh wr0, wr7, wr0          \n\t" /* "pcmpeqw %%mm7, %%mm0          \n\t" // block[i] == 0 ? -1 : 0 */
-        "wcmpeqh wr1, wr7, wr1          \n\t" /* "pcmpeqw %%mm7, %%mm1          \n\t" // block[i] == 0 ? -1 : 0 */
-        "wandn wr0, wr2, wr0            \n\t" /* "pandn %%mm2, %%mm0            \n\t" */
-        "wandn wr1, wr3, wr1            \n\t" /* "pandn %%mm3, %%mm1            \n\t" */
-        "wstrd wr0, [%[block]]          \n\t" /* "movq %%mm0, (%0, %3)          \n\t" */
-        "wstrd wr1, [%[block], #8]      \n\t" /* "movq %%mm1, 8(%0, %3)         \n\t" */
-        "add %[block], %[block], #16    \n\t" /* "addl $16, %3                  \n\t" */
-        "subs %[i], %[i], #1            \n\t"
-        "bne 1b                         \n\t" /* "jng 1b                                \n\t" */
-        :[block]"+r"(block)
-        :[i]"r"((nCoeffs + 8) / 8), [qmul]"r"(qmul), [qadd]"r"(qadd)
-        :"memory");
-
-    block_orig[0] = level;
-}
-
-void ff_MPV_common_init_iwmmxt(MpegEncContext *s)
-{
-    if (!(mm_flags & AV_CPU_FLAG_IWMMXT)) return;
-
-    s->dct_unquantize_h263_intra = dct_unquantize_h263_intra_iwmmxt;
-}