Request 2278 (accepted)

avidemux3.changes Changed

avidemux3.spec Changed

ffmpeg-2.8.1.tar.bz2/Changelog -> ffmpeg-2.8.3.tar.bz2/Changelog Changed

@@ -1,6 +1,87 @@
 Entries are sorted chronologically from oldest to youngest within each release,
 releases are sorted from youngest to oldest.
 
+version 2.8.3
+- avcodec/cabac: Check initial cabac decoder state
+- avcodec/cabac_functions: Fix "left shift of negative value -31767"
+- avcodec/h264_slice: Limit max_contexts when slice_context_count is initialized
+- rtmpcrypt: Do the xtea decryption in little endian mode
+- avformat/matroskadec: Check subtitle stream before dereferencing
+- avcodec/pngdec: Replace assert by request for sample for unsupported TRNS cases
+- avformat/utils: Do not init parser if probing is unfinished
+- avcodec/jpeg2000dec: Fix potential integer overflow with tile dimensions
+- avcodec/jpeg2000: Use av_image_check_size() in ff_jpeg2000_init_component()
+- avcodec/wmaprodec: Check for overread in decode_packet()
+- avcodec/smacker: Check that the data size is a multiple of a sample vector
+- avcodec/takdec: Skip last p2 sample (which is unused)
+- avcodec/dxtory: Fix input size check in dxtory_decode_v1_410()
+- avcodec/dxtory: Fix input size check in dxtory_decode_v1_420()
+- avcodec/error_resilience: avoid accessing previous or next frames tables beyond height
+- avcodec/dpx: Move need_align to act per line
+- avcodec/flashsv: Check size before updating it
+- avcodec/ivi: Check image dimensions
+- avcodec/utils: Better check for channels in av_get_audio_frame_duration()
+- avcodec/jpeg2000dec: Check for duplicate SIZ marker
+- aacsbr: don't call sbr_dequant twice without intermediate read_sbr_data
+- hqx: correct type and size check of info_offset
+- mxfdec: check edit_rate also for physical_track
+- avcodec/jpeg2000: Change coord to 32bit to support larger than 32k width or height
+- avcodec/jpeg2000dec: Check SIZ dimensions to be within the supported range
+- avcodec/jpeg2000: Check comp coords to be within the supported size
+- mpegvideo: clear overread in clear_context
+- avcodec/avrndec: Use the AVFrame format instead of the context
+- dds: disable palette flag for compressed images
+- dds: validate compressed source buffer size
+- dds: validate source buffer size before copying
+- dvdsubdec: validate offset2 similar to offset1
+- brstm: reject negative sample rate
+- aacps: avoid division by zero in stereo_processing
+- softfloat: assert when the argument of av_sqrt_sf is negative
+
+version 2.8.2
+- various fixes in the aac_fixed decoder
+- various fixes in softfloat
+- swresample/resample: increase precision for compensation
+- lavf/mov: add support for sidx fragment indexes
+- avformat/mxfenc: Only store user comment related tags when needed
+- tests/fate/avformat: Fix fate-lavf
+- doc/ffmpeg: Clarify that the sdp_file option requires an rtp output.
+- ffmpeg: Don't try and write sdp info if none of the outputs had an rtp format.
+- apng: use correct size for output buffer
+- jvdec: avoid unsigned overflow in comparison
+- avcodec/jpeg2000dec: Clip all tile coordinates
+- avcodec/microdvddec: Check for string end in 'P' case
+- avcodec/dirac_parser: Fix undefined memcpy() use
+- avformat/xmv: Discard remainder of packet on error
+- avformat/xmv: factor return check out of if/else
+- avcodec/mpeg12dec: Do not call show_bits() with invalid bits
+- avcodec/faxcompr: Add missing runs check in decode_uncompressed()
+- libavutil/channel_layout: Check strtol*() for failure
+- avformat/mpegts: Only start probing data streams within probe_packets
+- avcodec/hevc_ps: Check chroma_format_idc
+- avcodec/ffv1dec: Check for 0 quant tables
+- avcodec/mjpegdec: Reinitialize IDCT on BPP changes
+- avcodec/mjpegdec: Check index in ljpeg_decode_yuv_scan() before using it
+- avutil/file_open: avoid file handle inheritance on Windows
+- avcodec/h264_slice: Disable slice threads if there are multiple access units in a packet
+- avformat/hls: update cookies on setcookie response
+- opusdec: Don't run vector_fmul_scalar on zero length arrays
+- avcodec/opusdec: Fix extra samples read index
+- avcodec/ffv1: Initialize vlc_state on allocation
+- avcodec/ffv1dec: update progress in case of broken pointer chains
+- avcodec/ffv1dec: Clear slice coordinates if they are invalid or slice header decoding fails for other reasons
+- rtsp: Allow $ as interleaved packet indicator before a complete response header
+- videodsp: don't overread edges in vfix3 emu_edge.
+- avformat/mp3dec: improve junk skipping heuristic
+- concatdec: fix file_start_time calculation regression
+- avcodec: loongson optimize h264dsp idct and loop filter with mmi
+- avcodec/jpeg2000dec: Clear properties in jpeg2000_dec_cleanup() too
+- avformat/hls: add support for EXT-X-MAP
+- avformat/hls: fix segment selection regression on track changes of live streams
+- configure: Require libkvazaar < 0.7.
+- avcodec/vp8: Do not use num_coeff_partitions in thread/buffer setup
+
+
 version 2.8.1:
 - swscale: fix ticket #4881
 - doc: fix spelling errors
@@ -31,7 +112,7 @@
 - dnxhddec: decode and use interlace mb flag
 - swscale: fix ticket #4877
 - avcodec/rangecoder: Check e
-- avcodec/ffv1: seperate slice_count from max_slice_count
+- avcodec/ffv1: separate slice_count from max_slice_count
 - swscale: fix ticket 4850
 - cmdutils: Filter dst/srcw/h
 - avutil/log: fix zero length gnu_printf format string warning

ffmpeg-2.8.1.tar.bz2/RELEASE -> ffmpeg-2.8.3.tar.bz2/RELEASE Changed

ffmpeg-2.8.1.tar.bz2/VERSION -> ffmpeg-2.8.3.tar.bz2/VERSION Changed

ffmpeg-2.8.1.tar.bz2/configure -> ffmpeg-2.8.3.tar.bz2/configure Changed

ffmpeg-2.8.1.tar.bz2/doc/Doxyfile -> ffmpeg-2.8.3.tar.bz2/doc/Doxyfile Changed

ffmpeg-2.8.1.tar.bz2/doc/ffmpeg.texi -> ffmpeg-2.8.3.tar.bz2/doc/ffmpeg.texi Changed

ffmpeg-2.8.1.tar.bz2/doc/muxers.texi -> ffmpeg-2.8.3.tar.bz2/doc/muxers.texi Changed

ffmpeg-2.8.1.tar.bz2/ffmpeg.c -> ffmpeg-2.8.3.tar.bz2/ffmpeg.c Changed

ffmpeg-2.8.1.tar.bz2/libavcodec/aacdec_template.c -> ffmpeg-2.8.3.tar.bz2/libavcodec/aacdec_template.c Changed

ffmpeg-2.8.1.tar.bz2/libavcodec/aacps.c -> ffmpeg-2.8.3.tar.bz2/libavcodec/aacps.c Changed

ffmpeg-2.8.1.tar.bz2/libavcodec/aacsbr_fixed.c -> ffmpeg-2.8.3.tar.bz2/libavcodec/aacsbr_fixed.c Changed

@@ -169,6 +169,10 @@
                 else
                   temp1.mant = 0x20000000;
                 temp1.exp = (temp1.exp >> 1) + 1;
+                if (temp1.exp > 66) { // temp1 > 1E20
+                    av_log(NULL, AV_LOG_ERROR, "envelope scalefactor overflow in dequant\n");
+                    temp1 = FLOAT_1;
+                }
 
                 temp2.exp = (pan_offset - sbr->data[1].env_facs[e][k].mant) * alpha;
                 if (temp2.exp & 1)
@@ -188,6 +192,10 @@
                 temp1.exp = NOISE_FLOOR_OFFSET - \
                     sbr->data[0].noise_facs[e][k].mant + 2;
                 temp1.mant = 0x20000000;
+                if (temp1.exp > 66) { // temp1 > 1E20
+                    av_log(NULL, AV_LOG_ERROR, "envelope scalefactor overflow in dequant\n");
+                    temp1 = FLOAT_1;
+                }
                 temp2.exp = 12 - sbr->data[1].noise_facs[e][k].mant + 1;
                 temp2.mant = 0x20000000;
                 fac   = av_div_sf(temp1, av_add_sf(FLOAT_1, temp2));
@@ -208,7 +216,10 @@
                     else
                         temp1.mant = 0x20000000;
                     temp1.exp = (temp1.exp >> 1) + 1;
-
+                    if (temp1.exp > 66) { // temp1 > 1E20
+                        av_log(NULL, AV_LOG_ERROR, "envelope scalefactor overflow in dequant\n");
+                        temp1 = FLOAT_1;
+                    }
                     sbr->data[ch].env_facs[e][k] = temp1;
                 }
             for (e = 1; e <= sbr->data[ch].bs_num_noise; e++)
@@ -397,7 +408,8 @@
         int delta = !((e == e_a[1]) || (e == e_a[0]));
         for (k = 0; k < sbr->n_lim; k++) {
             SoftFloat gain_boost, gain_max;
-            SoftFloat sum[2] = { { 0, 0}, { 0, 0 } };
+            SoftFloat sum[2];
+            sum[0] = sum[1] = FLOAT_0;
             for (m = sbr->f_tablelim[k] - sbr->kx[1]; m < sbr->f_tablelim[k + 1] - sbr->kx[1]; m++) {
                 const SoftFloat temp = av_div_sf(sbr->e_origmapped[e][m],
                                             av_add_sf(FLOAT_1, sbr->q_mapped[e][m]));

ffmpeg-2.8.1.tar.bz2/libavcodec/aacsbr_template.c -> ffmpeg-2.8.3.tar.bz2/libavcodec/aacsbr_template.c Changed

@@ -70,6 +70,7 @@
 /** Places SBR in pure upsampling mode. */
 static void sbr_turnoff(SpectralBandReplication *sbr) {
     sbr->start = 0;
+    sbr->ready_for_dequant = 0;
     // Init defults used in pure upsampling mode
     sbr->kx[1] = 32; //Typo in spec, kx' inits to 32
     sbr->m[1] = 0;
@@ -177,6 +178,7 @@
     SpectrumParameters old_spectrum_params;
 
     sbr->start = 1;
+    sbr->ready_for_dequant = 0;
 
     // Save last spectrum parameters variables to compare to new ones
     memcpy(&old_spectrum_params, &sbr->spectrum_params, sizeof(SpectrumParameters));
@@ -1032,6 +1034,7 @@
     unsigned int cnt = get_bits_count(gb);
 
     sbr->id_aac = id_aac;
+    sbr->ready_for_dequant = 1;
 
     if (id_aac == TYPE_SCE || id_aac == TYPE_CCE) {
         if (read_sbr_single_channel_element(ac, sbr, gb)) {
@@ -1412,7 +1415,7 @@
 
             for (p = 0; p < sbr->n[ch_data->bs_freq_res[e + 1]]; p++) {
 #if USE_FIXED
-                SoftFloat sum = { 0, 0 };
+                SoftFloat sum = FLOAT_0;
                 const SoftFloat den = av_int2sf(0x20000000 / (env_size * (table[p + 1] - table[p])), 29);
                 for (k = table[p]; k < table[p + 1]; k++) {
                     sum = av_add_sf(sum, sbr->dsp.sum_square(X_high[k] + ilb, iub - ilb));
@@ -1449,6 +1452,12 @@
         sbr_turnoff(sbr);
     }
 
+    if (sbr->start && !sbr->ready_for_dequant) {
+        av_log(ac->avctx, AV_LOG_ERROR,
+               "No quantized data read for sbr_dequant.\n");
+        sbr_turnoff(sbr);
+    }
+
     if (!sbr->kx_and_m_pushed) {
         sbr->kx[0] = sbr->kx[1];
         sbr->m[0] = sbr->m[1];
@@ -1458,6 +1467,7 @@
 
     if (sbr->start) {
         sbr_dequant(sbr, id_aac);
+        sbr->ready_for_dequant = 0;
     }
     for (ch = 0; ch < nch; ch++) {
         /* decode channel */

ffmpeg-2.8.1.tar.bz2/libavcodec/avrndec.c -> ffmpeg-2.8.3.tar.bz2/libavcodec/avrndec.c Changed

ffmpeg-2.8.1.tar.bz2/libavcodec/cabac.c -> ffmpeg-2.8.3.tar.bz2/libavcodec/cabac.c Changed

ffmpeg-2.8.1.tar.bz2/libavcodec/cabac.h -> ffmpeg-2.8.3.tar.bz2/libavcodec/cabac.h Changed

ffmpeg-2.8.1.tar.bz2/libavcodec/cabac_functions.h -> ffmpeg-2.8.3.tar.bz2/libavcodec/cabac_functions.h Changed

ffmpeg-2.8.1.tar.bz2/libavcodec/dds.c -> ffmpeg-2.8.3.tar.bz2/libavcodec/dds.c Changed

@@ -141,6 +141,12 @@
     normal_map      = flags & DDPF_NORMALMAP;
     fourcc = bytestream2_get_le32(gbc);
 
+    if (ctx->compressed && ctx->paletted) {
+        av_log(avctx, AV_LOG_WARNING,
+               "Disabling invalid palette flag for compressed dds.\n");
+        ctx->paletted = 0;
+    }
+
     bpp = bytestream2_get_le32(gbc); // rgbbitcount
     r   = bytestream2_get_le32(gbc); // rbitmask
     g   = bytestream2_get_le32(gbc); // gbitmask
@@ -642,9 +648,18 @@
         return ret;
 
     if (ctx->compressed) {
+        int size = (avctx->coded_height / TEXTURE_BLOCK_H) *
+                   (avctx->coded_width / TEXTURE_BLOCK_W) * ctx->tex_ratio;
         ctx->slice_count = av_clip(avctx->thread_count, 1,
                                    avctx->coded_height / TEXTURE_BLOCK_H);
 
+        if (bytestream2_get_bytes_left(gbc) < size) {
+            av_log(avctx, AV_LOG_ERROR,
+                   "Compressed Buffer is too small (%d < %d).\n",
+                   bytestream2_get_bytes_left(gbc), size);
+            return AVERROR_INVALIDDATA;
+        }
+
         /* Use the decompress function on the texture, one block per thread. */
         ctx->tex_data = gbc->buffer;
         avctx->execute2(avctx, decompress_texture_thread, frame, NULL, ctx->slice_count);
@@ -666,6 +681,12 @@
             frame->palette_has_changed = 1;
         }
 
+        if (bytestream2_get_bytes_left(gbc) < frame->height * linesize) {
+            av_log(avctx, AV_LOG_ERROR, "Buffer is too small (%d < %d).\n",
+                   bytestream2_get_bytes_left(gbc), frame->height * linesize);
+            return AVERROR_INVALIDDATA;
+        }
+
         av_image_copy_plane(frame->data[0], frame->linesize[0],
                             gbc->buffer, linesize,
                             linesize, frame->height);

ffmpeg-2.8.1.tar.bz2/libavcodec/dirac_parser.c -> ffmpeg-2.8.3.tar.bz2/libavcodec/dirac_parser.c Changed

ffmpeg-2.8.1.tar.bz2/libavcodec/dpx.c -> ffmpeg-2.8.3.tar.bz2/libavcodec/dpx.c Changed

ffmpeg-2.8.1.tar.bz2/libavcodec/dvdsubdec.c -> ffmpeg-2.8.3.tar.bz2/libavcodec/dvdsubdec.c Changed

ffmpeg-2.8.1.tar.bz2/libavcodec/dxtory.c -> ffmpeg-2.8.3.tar.bz2/libavcodec/dxtory.c Changed

ffmpeg-2.8.1.tar.bz2/libavcodec/error_resilience.c -> ffmpeg-2.8.3.tar.bz2/libavcodec/error_resilience.c Changed

@@ -381,14 +381,19 @@
 #define MV_UNCHANGED 1
     const int mb_stride = s->mb_stride;
     const int mb_width  = s->mb_width;
-    const int mb_height = s->mb_height;
+    int mb_height = s->mb_height;
     int i, depth, num_avail;
     int mb_x, mb_y, mot_step, mot_stride;
 
+    if (s->last_pic.f && s->last_pic.f->data[0])
+        mb_height = FFMIN(mb_height, (s->last_pic.f->height+15)>>4);
+    if (s->next_pic.f && s->next_pic.f->data[0])
+        mb_height = FFMIN(mb_height, (s->next_pic.f->height+15)>>4);
+
     set_mv_strides(s, &mot_step, &mot_stride);
 
     num_avail = 0;
-    for (i = 0; i < s->mb_num; i++) {
+    for (i = 0; i < mb_width * mb_height; i++) {
         const int mb_xy = s->mb_index2xy[i];
         int f = 0;
         int error = s->error_status_table[mb_xy];
@@ -413,7 +418,7 @@
 
     if ((!(s->avctx->error_concealment&FF_EC_GUESS_MVS)) ||
         num_avail <= mb_width / 2) {
-        for (mb_y = 0; mb_y < s->mb_height; mb_y++) {
+        for (mb_y = 0; mb_y < mb_height; mb_y++) {
             for (mb_x = 0; mb_x < s->mb_width; mb_x++) {
                 const int mb_xy = mb_x + mb_y * s->mb_stride;
                 int mv_dir = (s->last_pic.f && s->last_pic.f->data[0]) ? MV_DIR_FORWARD : MV_DIR_BACKWARD;
@@ -442,7 +447,7 @@
             int score_sum = 0;
 
             changed = 0;
-            for (mb_y = 0; mb_y < s->mb_height; mb_y++) {
+            for (mb_y = 0; mb_y < mb_height; mb_y++) {
                 for (mb_x = 0; mb_x < s->mb_width; mb_x++) {
                     const int mb_xy        = mb_x + mb_y * s->mb_stride;
                     int mv_predictor[8][2] = { { 0 } };
@@ -675,7 +680,7 @@
         if (none_left)
             return;
 
-        for (i = 0; i < s->mb_num; i++) {
+        for (i = 0; i < mb_width * mb_height; i++) {
             int mb_xy = s->mb_index2xy[i];
             if (fixed[mb_xy])
                 fixed[mb_xy] = MV_FROZEN;

ffmpeg-2.8.1.tar.bz2/libavcodec/faxcompr.c -> ffmpeg-2.8.3.tar.bz2/libavcodec/faxcompr.c Changed

ffmpeg-2.8.1.tar.bz2/libavcodec/ffv1.c -> ffmpeg-2.8.3.tar.bz2/libavcodec/ffv1.c Changed

ffmpeg-2.8.1.tar.bz2/libavcodec/ffv1dec.c -> ffmpeg-2.8.3.tar.bz2/libavcodec/ffv1dec.c Changed

ffmpeg-2.8.1.tar.bz2/libavcodec/flashsv.c -> ffmpeg-2.8.3.tar.bz2/libavcodec/flashsv.c Changed

ffmpeg-2.8.1.tar.bz2/libavcodec/h264_cabac.c -> ffmpeg-2.8.3.tar.bz2/libavcodec/h264_cabac.c Changed

ffmpeg-2.8.1.tar.bz2/libavcodec/h264_slice.c -> ffmpeg-2.8.3.tar.bz2/libavcodec/h264_slice.c Changed

@@ -1097,6 +1097,7 @@
         nb_slices = max_slices;
     }
     h->slice_context_count = nb_slices;
+    h->max_contexts = FFMIN(h->max_contexts, nb_slices);
 
     if (!HAVE_THREADS || !(h->avctx->active_thread_type & FF_THREAD_SLICE)) {
         ret = ff_h264_slice_context_init(h, &h->slice_ctx[0]);
@@ -1177,6 +1178,15 @@
                 av_log(h->avctx, AV_LOG_ERROR, "Too many fields\n");
                 return AVERROR_INVALIDDATA;
             }
+            if (h->max_contexts > 1) {
+                if (!h->single_decode_warning) {
+                    av_log(h->avctx, AV_LOG_WARNING, "Cannot decode multiple access units as slice threads\n");
+                    h->single_decode_warning = 1;
+                }
+                h->max_contexts = 1;
+                return SLICE_SINGLETHREAD;
+            }
+
             if (h->cur_pic_ptr && FIELD_PICTURE(h) && h->first_field) {
                 ret = ff_h264_field_end(h, h->slice_ctx, 1);
                 h->current_slice = 0;
@@ -2362,9 +2372,11 @@
         align_get_bits(&sl->gb);
 
         /* init cabac */
-        ff_init_cabac_decoder(&sl->cabac,
+        ret = ff_init_cabac_decoder(&sl->cabac,
                               sl->gb.buffer + get_bits_count(&sl->gb) / 8,
                               (get_bits_left(&sl->gb) + 7) / 8);
+        if (ret < 0)
+            return ret;
 
         ff_h264_init_cabac_states(h, sl);

ffmpeg-2.8.1.tar.bz2/libavcodec/hevc_ps.c -> ffmpeg-2.8.3.tar.bz2/libavcodec/hevc_ps.c Changed

ffmpeg-2.8.1.tar.bz2/libavcodec/hqx.c -> ffmpeg-2.8.3.tar.bz2/libavcodec/hqx.c Changed

ffmpeg-2.8.1.tar.bz2/libavcodec/ivi.c -> ffmpeg-2.8.3.tar.bz2/libavcodec/ivi.c Changed

ffmpeg-2.8.1.tar.bz2/libavcodec/jpeg2000.c -> ffmpeg-2.8.3.tar.bz2/libavcodec/jpeg2000.c Changed

@@ -28,6 +28,7 @@
 #include "libavutil/attributes.h"
 #include "libavutil/avassert.h"
 #include "libavutil/common.h"
+#include "libavutil/imgutils.h"
 #include "libavutil/mem.h"
 #include "avcodec.h"
 #include "jpeg2000.h"
@@ -210,9 +211,17 @@
                                    codsty->nreslevels2decode - 1,
                                    codsty->transform))
         return ret;
-    // component size comp->coord is uint16_t so ir cannot overflow
+
+    if (av_image_check_size(comp->coord[0][1] - comp->coord[0][0],
+                            comp->coord[1][1] - comp->coord[1][0], 0, avctx))
+        return AVERROR_INVALIDDATA;
     csize = (comp->coord[0][1] - comp->coord[0][0]) *
             (comp->coord[1][1] - comp->coord[1][0]);
+    if (comp->coord[0][1] - comp->coord[0][0] > 32768 ||
+        comp->coord[1][1] - comp->coord[1][0] > 32768) {
+        av_log(avctx, AV_LOG_ERROR, "component size too large\n");
+        return AVERROR_PATCHWELCOME;
+    }
 
     if (codsty->transform == FF_DWT97) {
         csize += AV_INPUT_BUFFER_PADDING_SIZE / sizeof(*comp->f_data);
@@ -450,7 +459,7 @@
                     return AVERROR(ENOMEM);
                 for (cblkno = 0; cblkno < nb_codeblocks; cblkno++) {
                     Jpeg2000Cblk *cblk = prec->cblk + cblkno;
-                    uint16_t Cx0, Cy0;
+                    int Cx0, Cy0;
 
                     /* Compute coordinates of codeblocks */
                     /* Compute Cx0*/

ffmpeg-2.8.1.tar.bz2/libavcodec/jpeg2000.h -> ffmpeg-2.8.3.tar.bz2/libavcodec/jpeg2000.h Changed

@@ -174,21 +174,21 @@
     int nb_terminationsinc;
     int data_start[JPEG2000_MAX_PASSES];
     Jpeg2000Pass passes[JPEG2000_MAX_PASSES];
-    uint16_t coord[2][2]; // border coordinates {{x0, x1}, {y0, y1}}
+    int coord[2][2]; // border coordinates {{x0, x1}, {y0, y1}}
 } Jpeg2000Cblk; // code block
 
 typedef struct Jpeg2000Prec {
-    uint16_t nb_codeblocks_width;
-    uint16_t nb_codeblocks_height;
+    int nb_codeblocks_width;
+    int nb_codeblocks_height;
     Jpeg2000TgtNode *zerobits;
     Jpeg2000TgtNode *cblkincl;
     Jpeg2000Cblk *cblk;
     int decoded_layers;
-    uint16_t coord[2][2]; // border coordinates {{x0, x1}, {y0, y1}}
+    int coord[2][2]; // border coordinates {{x0, x1}, {y0, y1}}
 } Jpeg2000Prec; // precinct
 
 typedef struct Jpeg2000Band {
-    uint16_t coord[2][2]; // border coordinates {{x0, x1}, {y0, y1}}
+    int coord[2][2]; // border coordinates {{x0, x1}, {y0, y1}}
     uint16_t log2_cblk_width, log2_cblk_height;
     int i_stepsize; // quantization stepsize
     float f_stepsize; // quantization stepsize
@@ -197,8 +197,8 @@
 
 typedef struct Jpeg2000ResLevel {
     uint8_t nbands;
-    uint16_t coord[2][2]; // border coordinates {{x0, x1}, {y0, y1}}
-    uint16_t num_precincts_x, num_precincts_y; // number of precincts in x/y direction
+    int coord[2][2]; // border coordinates {{x0, x1}, {y0, y1}}
+    int num_precincts_x, num_precincts_y; // number of precincts in x/y direction
     uint8_t log2_prec_width, log2_prec_height; // exponent of precinct size
     Jpeg2000Band *band;
 } Jpeg2000ResLevel; // resolution level
@@ -208,8 +208,8 @@
     DWTContext dwt;
     float *f_data;
     int *i_data;
-    uint16_t coord[2][2];   // border coordinates {{x0, x1}, {y0, y1}} -- can be reduced with lowres option
-    uint16_t coord_o[2][2]; // border coordinates {{x0, x1}, {y0, y1}} -- original values from jpeg2000 headers
+    int coord[2][2];   // border coordinates {{x0, x1}, {y0, y1}} -- can be reduced with lowres option
+    int coord_o[2][2]; // border coordinates {{x0, x1}, {y0, y1}} -- original values from jpeg2000 headers
 } Jpeg2000Component;
 
 /* misc tools */

ffmpeg-2.8.1.tar.bz2/libavcodec/jpeg2000dec.c -> ffmpeg-2.8.3.tar.bz2/libavcodec/jpeg2000dec.c Changed

@@ -30,6 +30,7 @@
 #include "libavutil/attributes.h"
 #include "libavutil/avassert.h"
 #include "libavutil/common.h"
+#include "libavutil/imgutils.h"
 #include "libavutil/opt.h"
 #include "libavutil/pixdesc.h"
 #include "avcodec.h"
@@ -279,6 +280,10 @@
         avpriv_request_sample(s->avctx, "Support for image offsets");
         return AVERROR_PATCHWELCOME;
     }
+    if (av_image_check_size(s->width, s->height, 0, s->avctx)) {
+        avpriv_request_sample(s->avctx, "Large Dimensions");
+        return AVERROR_PATCHWELCOME;
+    }
 
     if (ncomponents <= 0) {
         av_log(s->avctx, AV_LOG_ERROR, "Invalid number of components: %d\n",
@@ -826,10 +831,10 @@
     if (!tile->comp)
         return AVERROR(ENOMEM);
 
-    tile->coord[0][0] = FFMAX(tilex       * s->tile_width  + s->tile_offset_x, s->image_offset_x);
-    tile->coord[0][1] = FFMIN((tilex + 1) * s->tile_width  + s->tile_offset_x, s->width);
-    tile->coord[1][0] = FFMAX(tiley       * s->tile_height + s->tile_offset_y, s->image_offset_y);
-    tile->coord[1][1] = FFMIN((tiley + 1) * s->tile_height + s->tile_offset_y, s->height);
+    tile->coord[0][0] = av_clip(tilex       * (int64_t)s->tile_width  + s->tile_offset_x, s->image_offset_x, s->width);
+    tile->coord[0][1] = av_clip((tilex + 1) * (int64_t)s->tile_width  + s->tile_offset_x, s->image_offset_x, s->width);
+    tile->coord[1][0] = av_clip(tiley       * (int64_t)s->tile_height + s->tile_offset_y, s->image_offset_y, s->height);
+    tile->coord[1][1] = av_clip((tiley + 1) * (int64_t)s->tile_height + s->tile_offset_y, s->image_offset_y, s->height);
 
     for (compno = 0; compno < s->ncomponents; compno++) {
         Jpeg2000Component *comp = tile->comp + compno;
@@ -1809,8 +1814,10 @@
     av_freep(&s->tile);
     memset(s->codsty, 0, sizeof(s->codsty));
     memset(s->qntsty, 0, sizeof(s->qntsty));
+    memset(s->properties, 0, sizeof(s->properties));
     memset(&s->poc  , 0, sizeof(s->poc));
     s->numXtiles = s->numYtiles = 0;
+    s->ncomponents = 0;
 }
 
 static int jpeg2000_read_main_headers(Jpeg2000DecoderContext *s)
@@ -1868,6 +1875,10 @@
 
         switch (marker) {
         case JPEG2000_SIZ:
+            if (s->ncomponents) {
+                av_log(s->avctx, AV_LOG_ERROR, "Duplicate SIZ\n");
+                return AVERROR_INVALIDDATA;
+            }
             ret = get_siz(s);
             if (!s->tile)
                 s->numXtiles = s->numYtiles = 0;

ffmpeg-2.8.1.tar.bz2/libavcodec/jpeg2000dwt.c -> ffmpeg-2.8.3.tar.bz2/libavcodec/jpeg2000dwt.c Changed

@@ -534,7 +534,7 @@
         data[i] = (data[i] + ((1<<I_PRESHIFT)>>1)) >> I_PRESHIFT;
 }
 
-int ff_jpeg2000_dwt_init(DWTContext *s, uint16_t border[2][2],
+int ff_jpeg2000_dwt_init(DWTContext *s, int border[2][2],
                          int decomp_levels, int type)
 {
     int i, j, lev = decomp_levels, maxlen,
@@ -623,7 +623,7 @@
 
 #define MAX_W 256
 
-static int test_dwt(int *array, int *ref, uint16_t border[2][2], int decomp_levels, int type, int max_diff) {
+static int test_dwt(int *array, int *ref, int border[2][2], int decomp_levels, int type, int max_diff) {
     int ret, j;
     DWTContext s1={{{0}}}, *s= &s1;
     int64_t err2 = 0;
@@ -662,7 +662,7 @@
     return 0;
 }
 
-static int test_dwtf(float *array, float *ref, uint16_t border[2][2], int decomp_levels, float max_diff) {
+static int test_dwtf(float *array, float *ref, int border[2][2], int decomp_levels, float max_diff) {
     int ret, j;
     DWTContext s1={{{0}}}, *s= &s1;
     double err2 = 0;
@@ -708,7 +708,7 @@
 int main(void) {
     AVLFG prng;
     int i,j;
-    uint16_t border[2][2];
+    int border[2][2];
     int ret, decomp_levels;
 
     av_lfg_init(&prng, 1);

ffmpeg-2.8.1.tar.bz2/libavcodec/jpeg2000dwt.h -> ffmpeg-2.8.3.tar.bz2/libavcodec/jpeg2000dwt.h Changed

ffmpeg-2.8.1.tar.bz2/libavcodec/microdvddec.c -> ffmpeg-2.8.3.tar.bz2/libavcodec/microdvddec.c Changed

ffmpeg-2.8.1.tar.bz2/libavcodec/mips/h264dsp_init_mips.c -> ffmpeg-2.8.3.tar.bz2/libavcodec/mips/h264dsp_init_mips.c Changed

@@ -82,11 +82,33 @@
 #endif  // #if HAVE_MSA
 
 #if HAVE_MMI
-static av_cold void h264dsp_init_mmi(H264DSPContext * c,
-                                     const int bit_depth,
-                                     const int chroma_format_idc)
+static av_cold void h264dsp_init_mmi(H264DSPContext * c, const int bit_depth,
+        const int chroma_format_idc)
 {
     if (bit_depth == 8) {
+        c->h264_add_pixels4_clear = ff_h264_add_pixels4_8_mmi;
+        c->h264_idct_add = ff_h264_idct_add_8_mmi;
+        c->h264_idct8_add = ff_h264_idct8_add_8_mmi;
+        c->h264_idct_dc_add = ff_h264_idct_dc_add_8_mmi;
+        c->h264_idct8_dc_add = ff_h264_idct8_dc_add_8_mmi;
+        c->h264_idct_add16 = ff_h264_idct_add16_8_mmi;
+        c->h264_idct_add16intra = ff_h264_idct_add16intra_8_mmi;
+        c->h264_idct8_add4 = ff_h264_idct8_add4_8_mmi;
+
+        if (chroma_format_idc <= 1)
+            c->h264_idct_add8 = ff_h264_idct_add8_8_mmi;
+        else
+            c->h264_idct_add8 = ff_h264_idct_add8_422_8_mmi;
+
+        c->h264_luma_dc_dequant_idct = ff_h264_luma_dc_dequant_idct_8_mmi;
+
+        if (chroma_format_idc <= 1)
+            c->h264_chroma_dc_dequant_idct =
+                ff_h264_chroma_dc_dequant_idct_8_mmi;
+        else
+            c->h264_chroma_dc_dequant_idct =
+                ff_h264_chroma422_dc_dequant_idct_8_mmi;
+
         c->weight_h264_pixels_tab[0] = ff_h264_weight_pixels16_8_mmi;
         c->weight_h264_pixels_tab[1] = ff_h264_weight_pixels8_8_mmi;
         c->weight_h264_pixels_tab[2] = ff_h264_weight_pixels4_8_mmi;
@@ -94,6 +116,21 @@
         c->biweight_h264_pixels_tab[0] = ff_h264_biweight_pixels16_8_mmi;
         c->biweight_h264_pixels_tab[1] = ff_h264_biweight_pixels8_8_mmi;
         c->biweight_h264_pixels_tab[2] = ff_h264_biweight_pixels4_8_mmi;
+
+        c->h264_v_loop_filter_chroma       = ff_deblock_v_chroma_8_mmi;
+        c->h264_v_loop_filter_chroma_intra = ff_deblock_v_chroma_intra_8_mmi;
+
+        if (chroma_format_idc <= 1) {
+            c->h264_h_loop_filter_chroma =
+                ff_deblock_h_chroma_8_mmi;
+            c->h264_h_loop_filter_chroma_intra =
+                ff_deblock_h_chroma_intra_8_mmi;
+        }
+
+        c->h264_v_loop_filter_luma = ff_deblock_v_luma_8_mmi;
+        c->h264_v_loop_filter_luma_intra = ff_deblock_v_luma_intra_8_mmi;
+        c->h264_h_loop_filter_luma = ff_deblock_h_luma_8_mmi;
+        c->h264_h_loop_filter_luma_intra = ff_deblock_h_luma_intra_8_mmi;
     }
 }
 #endif /* HAVE_MMI */

ffmpeg-2.8.1.tar.bz2/libavcodec/mips/h264dsp_mips.h -> ffmpeg-2.8.3.tar.bz2/libavcodec/mips/h264dsp_mips.h Changed

@@ -319,6 +319,26 @@
 void ff_vp8_pred16x16_127_dc_8_msa(uint8_t *src, ptrdiff_t stride);
 void ff_vp8_pred16x16_129_dc_8_msa(uint8_t *src, ptrdiff_t stride);
 
+void ff_h264_add_pixels4_8_mmi(uint8_t *_dst, int16_t *_src, int stride);
+void ff_h264_idct_add_8_mmi(uint8_t *dst, int16_t *block, int stride);
+void ff_h264_idct8_add_8_mmi(uint8_t *dst, int16_t *block, int stride);
+void ff_h264_idct_dc_add_8_mmi(uint8_t *dst, int16_t *block, int stride);
+void ff_h264_idct8_dc_add_8_mmi(uint8_t *dst, int16_t *block, int stride);
+void ff_h264_idct_add16_8_mmi(uint8_t *dst, const int *block_offset,
+        int16_t *block, int stride, const uint8_t nnzc[15*8]);
+void ff_h264_idct_add16intra_8_mmi(uint8_t *dst, const int *block_offset,
+        int16_t *block, int stride, const uint8_t nnzc[15*8]);
+void ff_h264_idct8_add4_8_mmi(uint8_t *dst, const int *block_offset,
+        int16_t *block, int stride, const uint8_t nnzc[15*8]);
+void ff_h264_idct_add8_8_mmi(uint8_t **dest, const int *block_offset,
+        int16_t *block, int stride, const uint8_t nnzc[15*8]);
+void ff_h264_idct_add8_422_8_mmi(uint8_t **dest, const int *block_offset,
+        int16_t *block, int stride, const uint8_t nnzc[15*8]);
+void ff_h264_luma_dc_dequant_idct_8_mmi(int16_t *output, int16_t *input,
+        int qmul);
+void ff_h264_chroma_dc_dequant_idct_8_mmi(int16_t *block, int qmul);
+void ff_h264_chroma422_dc_dequant_idct_8_mmi(int16_t *block, int qmul);
+
 void ff_h264_weight_pixels16_8_mmi(uint8_t *block, int stride, int height,
         int log2_denom, int weight, int offset);
 void ff_h264_biweight_pixels16_8_mmi(uint8_t *dst, uint8_t *src,
@@ -335,6 +355,27 @@
         int stride, int height, int log2_denom, int weightd, int weights,
         int offset);
 
+void ff_deblock_v_chroma_8_mmi(uint8_t *pix, int stride, int alpha, int beta,
+        int8_t *tc0);
+void ff_deblock_v_chroma_intra_8_mmi(uint8_t *pix, int stride, int alpha,
+        int beta);
+void ff_deblock_h_chroma_8_mmi(uint8_t *pix, int stride, int alpha, int beta,
+        int8_t *tc0);
+void ff_deblock_h_chroma_intra_8_mmi(uint8_t *pix, int stride, int alpha,
+        int beta);
+void ff_deblock_v_luma_8_mmi(uint8_t *pix, int stride, int alpha, int beta,
+        int8_t *tc0);
+void ff_deblock_v_luma_intra_8_mmi(uint8_t *pix, int stride, int alpha,
+        int beta);
+void ff_deblock_h_luma_8_mmi(uint8_t *pix, int stride, int alpha, int beta,
+        int8_t *tc0);
+void ff_deblock_h_luma_intra_8_mmi(uint8_t *pix, int stride, int alpha,
+        int beta);
+void ff_deblock_v8_luma_8_mmi(uint8_t *pix, int stride, int alpha, int beta,
+        int8_t *tc0);
+void ff_deblock_v8_luma_intra_8_mmi(uint8_t *pix, int stride, int alpha,
+        int beta);
+
 void ff_put_h264_qpel16_mc00_mmi(uint8_t *dst, const uint8_t *src,
         ptrdiff_t dst_stride);
 void ff_put_h264_qpel16_mc10_mmi(uint8_t *dst, const uint8_t *src,

ffmpeg-2.8.1.tar.bz2/libavcodec/mips/h264dsp_mmi.c -> ffmpeg-2.8.3.tar.bz2/libavcodec/mips/h264dsp_mmi.c Changed

@@ -4,6 +4,7 @@
  * Copyright (c) 2015 Loongson Technology Corporation Limited
  * Copyright (c) 2015 Zhou Xiaoyong <zhouxiaoyong@loongson.cn>
  *                    Zhang Shuangshuang <zhangshuangshuang@ict.ac.cn>
+ *                    Heiher <r@hev.cc>
  *
  * This file is part of FFmpeg.
  *
@@ -25,6 +26,1011 @@
 #include "libavcodec/bit_depth_template.c"
 #include "h264dsp_mips.h"
 
+void ff_h264_add_pixels4_8_mmi(uint8_t *dst, int16_t *src, int stride)
+{
+    __asm__ volatile (
+        "xor $f0, $f0, $f0              \r\n"
+        "ldc1 $f2, 0(%[src])            \r\n"
+        "ldc1 $f4, 8(%[src])            \r\n"
+        "ldc1 $f6, 16(%[src])           \r\n"
+        "ldc1 $f8, 24(%[src])           \r\n"
+        "lwc1 $f10, 0(%[dst0])          \r\n"
+        "lwc1 $f12, 0(%[dst1])          \r\n"
+        "lwc1 $f14, 0(%[dst2])          \r\n"
+        "lwc1 $f16, 0(%[dst3])          \r\n"
+        "punpcklbh $f10, $f10, $f0      \r\n"
+        "punpcklbh $f12, $f12, $f0      \r\n"
+        "punpcklbh $f14, $f14, $f0      \r\n"
+        "punpcklbh $f16, $f16, $f0      \r\n"
+        "paddh $f2, $f2, $f10           \r\n"
+        "paddh $f4, $f4, $f12           \r\n"
+        "paddh $f6, $f6, $f14           \r\n"
+        "paddh $f8, $f8, $f16           \r\n"
+        "packushb $f2, $f2, $f0         \r\n"
+        "packushb $f4, $f4, $f0         \r\n"
+        "packushb $f6, $f6, $f0         \r\n"
+        "packushb $f8, $f8, $f0         \r\n"
+        "swc1 $f2, 0(%[dst0])           \r\n"
+        "swc1 $f4, 0(%[dst1])           \r\n"
+        "swc1 $f6, 0(%[dst2])           \r\n"
+        "swc1 $f8, 0(%[dst3])           \r\n"
+        ::[dst0]"r"(dst),[dst1]"r"(dst+stride),[dst2]"r"(dst+2*stride),
+          [dst3]"r"(dst+3*stride),[src]"r"(src)
+        : "$f0","$f2","$f4","$f6","$f8","$f10","$f12","$f14","$f16"
+    );
+
+    memset(src, 0, 32);
+}
+
+void ff_h264_idct_add_8_mmi(uint8_t *dst, int16_t *block, int stride)
+{
+    __asm__ volatile (
+        "dli $8, 1                              \r\n"
+        "ldc1 $f0, 0(%[block])                  \r\n"
+        "dmtc1 $8, $f16                         \r\n"
+        "ldc1 $f2, 8(%[block])                  \r\n"
+        "dli $8, 6                              \r\n"
+        "ldc1 $f4, 16(%[block])                 \r\n"
+        "dmtc1 $8, $f18                         \r\n"
+        "psrah $f8, $f2, $f16                   \r\n"
+        "ldc1 $f6, 24(%[block])                 \r\n"
+        "psrah $f10, $f6, $f16                  \r\n"
+        "psubh $f8, $f8, $f6                    \r\n"
+        "paddh $f10, $f10, $f2                  \r\n"
+        "paddh $f20, $f4, $f0                   \r\n"
+        "psubh $f0, $f0, $f4                    \r\n"
+        "paddh $f22, $f10, $f20                 \r\n"
+        "psubh $f4, $f20, $f10                  \r\n"
+        "paddh $f20, $f8, $f0                   \r\n"
+        "psubh $f0, $f0, $f8                    \r\n"
+        "punpckhhw $f2, $f22, $f20              \r\n"
+        "punpcklhw $f10, $f22, $f20             \r\n"
+        "punpckhhw $f8, $f0, $f4                \r\n"
+        "punpcklhw $f0, $f0, $f4                \r\n"
+        "punpckhwd $f4, $f10, $f0               \r\n"
+        "punpcklwd $f10, $f10, $f0              \r\n"
+        "punpcklwd $f20, $f2, $f8               \r\n"
+        "punpckhwd $f0, $f2, $f8                \r\n"
+        "paddh $f10, $f10, %[ff_pw_32]          \r\n"
+        "psrah $f8, $f4, $f16                   \r\n"
+        "psrah $f6, $f0, $f16                   \r\n"
+        "psubh $f8, $f8, $f0                    \r\n"
+        "paddh $f6, $f6, $f4                    \r\n"
+        "paddh $f2, $f20, $f10                  \r\n"
+        "psubh $f10, $f10, $f20                 \r\n"
+        "paddh $f20, $f6, $f2                   \r\n"
+        "psubh $f2, $f2, $f6                    \r\n"
+        "paddh $f22, $f8, $f10                  \r\n"
+        "xor $f14, $f14, $f14                   \r\n"
+        "psubh $f10, $f10, $f8                  \r\n"
+        "sdc1 $f14, 0(%[block])                 \r\n"
+        "sdc1 $f14, 8(%[block])                 \r\n"
+        "sdc1 $f14, 16(%[block])                \r\n"
+        "sdc1 $f14, 24(%[block])                \r\n"
+        "lwc1 $f4, 0(%[dst])                    \r\n"
+        "psrah $f6, $f20, $f18                  \r\n"
+        "gslwxc1 $f0, 0(%[dst], %[stride])      \r\n"
+        "psrah $f8, $f22, $f18                  \r\n"
+        "punpcklbh $f4, $f4, $f14               \r\n"
+        "punpcklbh $f0, $f0, $f14               \r\n"
+        "paddh $f4, $f4, $f6                    \r\n"
+        "paddh $f0, $f0, $f8                    \r\n"
+        "packushb $f4, $f4, $f14                \r\n"
+        "packushb $f0, $f0, $f14                \r\n"
+        "swc1 $f4, 0(%[dst])                    \r\n"
+        "gsswxc1 $f0, 0(%[dst], %[stride])      \r\n"
+        "daddu %[dst], %[dst], %[stride]        \r\n"
+        "daddu %[dst], %[dst], %[stride]        \r\n"
+        "lwc1 $f4, 0(%[dst])                    \r\n"
+        "psrah $f10, $f10, $f18                 \r\n"
+        "gslwxc1 $f0, 0(%[dst], %[stride])      \r\n"
+        "psrah $f2, $f2, $f18                   \r\n"
+        "punpcklbh $f4, $f4, $f14               \r\n"
+        "punpcklbh $f0, $f0, $f14               \r\n"
+        "paddh $f4, $f4, $f10                   \r\n"
+        "paddh $f0, $f0, $f2                    \r\n"
+        "packushb $f4, $f4, $f14                \r\n"
+        "swc1 $f4, 0(%[dst])                    \r\n"
+        "packushb $f0, $f0, $f14                \r\n"
+        "gsswxc1 $f0, 0(%[dst], %[stride])      \r\n"
+        ::[dst]"r"(dst),[block]"r"(block),[stride]"r"((uint64_t)stride),
+          [ff_pw_32]"f"(ff_pw_32)
+        : "$8","$f0","$f2","$f4","$f6","$f8","$f10","$f12","$f14","$f16",
+          "$f18","$f20","$f22"
+    );
+
+    memset(block, 0, 32);
+}
+
+void ff_h264_idct8_add_8_mmi(uint8_t *dst, int16_t *block, int stride)
+{
+    __asm__ volatile (
+        "lhu $10, 0x0(%[block])                     \r\n"
+        "daddiu $29, $29, -0x20                     \r\n"
+        "daddiu $10, $10, 0x20                      \r\n"
+        "ldc1 $f2, 0x10(%[block])                   \r\n"
+        "sh $10, 0x0(%[block])                      \r\n"
+        "ldc1 $f4, 0x20(%[block])                   \r\n"
+        "dli $10, 0x1                               \r\n"
+        "ldc1 $f6, 0x30(%[block])                   \r\n"
+        "dmtc1 $10, $f16                            \r\n"
+        "ldc1 $f10, 0x50(%[block])                  \r\n"
+        "ldc1 $f12, 0x60(%[block])                  \r\n"
+        "ldc1 $f14, 0x70(%[block])                  \r\n"
+        "mov.d $f0, $f2                             \r\n"
+        "psrah $f2, $f2, $f16                       \r\n"
+        "psrah $f8, $f10, $f16                      \r\n"
+        "paddh $f2, $f2, $f0                        \r\n"
+        "paddh $f8, $f8, $f10                       \r\n"
+        "paddh $f2, $f2, $f10                       \r\n"
+        "paddh $f8, $f8, $f14                       \r\n"
+        "paddh $f2, $f2, $f6                        \r\n"
+        "psubh $f8, $f8, $f0                        \r\n"
+        "psubh $f0, $f0, $f6                        \r\n"
+        "psubh $f10, $f10, $f6                      \r\n"
+        "psrah $f6, $f6, $f16                       \r\n"
+        "paddh $f0, $f0, $f14                       \r\n"
+        "psubh $f10, $f10, $f14                     \r\n"
+        "psrah $f14, $f14, $f16                     \r\n"
+        "psubh $f0, $f0, $f6                        \r\n"
+        "dli $10, 0x2                               \r\n"
+        "psubh $f10, $f10, $f14                     \r\n"
+        "dmtc1 $10, $f18                            \r\n"
+        "mov.d $f14, $f2                            \r\n"
+        "psrah $f2, $f2, $f18                       \r\n"
+        "psrah $f6, $f8, $f18                       \r\n"
+        "paddh $f6, $f6, $f0                        \r\n"
+        "psrah $f0, $f0, $f18                       \r\n"
+        "paddh $f2, $f2, $f10                       \r\n"
+        "psrah $f10, $f10, $f18                     \r\n"
+        "psubh $f0, $f0, $f8                        \r\n"
+        "psubh $f14, $f14, $f10                     \r\n"
+        "mov.d $f10, $f12                           \r\n"
+        "psrah $f12, $f12, $f16                     \r\n"
+        "psrah $f8, $f4, $f16                       \r\n"
+        "paddh $f12, $f12, $f4                      \r\n"
+        "psubh $f8, $f8, $f10                       \r\n"
+        "ldc1 $f4, 0x0(%[block])                    \r\n"
+        "ldc1 $f10, 0x40(%[block])                  \r\n"
+        "paddh $f10, $f10, $f4                      \r\n"
+        "paddh $f4, $f4, $f4                        \r\n"
+        "paddh $f12, $f12, $f10                     \r\n"
+        "psubh $f4, $f4, $f10                       \r\n"
+        "paddh $f10, $f10, $f10                     \r\n"
+        "paddh $f8, $f8, $f4                        \r\n"
+        "psubh $f10, $f10, $f12                     \r\n"
+        "paddh $f4, $f4, $f4                        \r\n"
+        "paddh $f14, $f14, $f12                     \r\n"
+        "psubh $f4, $f4, $f8                        \r\n"
+        "paddh $f12, $f12, $f12                     \r\n"
+        "paddh $f0, $f0, $f8                        \r\n"
+        "psubh $f12, $f12, $f14                     \r\n"
+        "paddh $f8, $f8, $f8                        \r\n"
+        "paddh $f6, $f6, $f4                        \r\n"
+        "psubh $f8, $f8, $f0                        \r\n"
+        "paddh $f4, $f4, $f4                        \r\n"
+        "paddh $f2, $f2, $f10                       \r\n"
+        "psubh $f4, $f4, $f6                        \r\n"
+        "paddh $f10, $f10, $f10                     \r\n"
+        "sdc1 $f12, 0x0(%[block])                   \r\n"
+        "psubh $f10, $f10, $f2                      \r\n"
+        "punpckhhw $f12, $f14, $f0                  \r\n"
+        "punpcklhw $f14, $f14, $f0                  \r\n"
+        "punpckhhw $f0, $f6, $f2                    \r\n"
+        "punpcklhw $f6, $f6, $f2                    \r\n"
+        "punpckhwd $f2, $f14, $f6                   \r\n"
+        "punpcklwd $f14, $f14, $f6                  \r\n"
+        "punpckhwd $f6, $f12, $f0                   \r\n"
+        "punpcklwd $f12, $f12, $f0                  \r\n"
+        "ldc1 $f0, 0x0(%[block])                    \r\n"
+        "sdc1 $f14, 0x0($29)                        \r\n"
+        "sdc1 $f2, 0x10($29)                        \r\n"
+        "dmfc1 $8, $f12                             \r\n"
+        "dmfc1 $11, $f6                             \r\n"
+        "punpckhhw $f6, $f10, $f4                   \r\n"
+        "punpcklhw $f10, $f10, $f4                  \r\n"
+        "punpckhhw $f4, $f8, $f0                    \r\n"
+        "punpcklhw $f8, $f8, $f0                    \r\n"
+        "punpckhwd $f0, $f10, $f8                   \r\n"
+        "punpcklwd $f10, $f10, $f8                  \r\n"
+        "punpckhwd $f8, $f6, $f4                    \r\n"
+        "punpcklwd $f6, $f6, $f4                    \r\n"
+        "sdc1 $f10, 0x8($29)                        \r\n"
+        "sdc1 $f0, 0x18($29)                        \r\n"
+        "dmfc1 $9, $f6                              \r\n"
+        "dmfc1 $12, $f8                             \r\n"
+        "ldc1 $f2, 0x18(%[block])                   \r\n"
+        "ldc1 $f12, 0x28(%[block])                  \r\n"
+        "ldc1 $f4, 0x38(%[block])                   \r\n"
+        "ldc1 $f0, 0x58(%[block])                   \r\n"
+        "ldc1 $f6, 0x68(%[block])                   \r\n"
+        "ldc1 $f8, 0x78(%[block])                   \r\n"
+        "mov.d $f14, $f2                            \r\n"
+        "psrah $f10, $f0, $f16                      \r\n"
+        "psrah $f2, $f2, $f16                       \r\n"
+        "paddh $f10, $f10, $f0                      \r\n"
+        "paddh $f2, $f2, $f14                       \r\n"
+        "paddh $f10, $f10, $f8                      \r\n"
+        "paddh $f2, $f2, $f0                        \r\n"
+        "psubh $f10, $f10, $f14                     \r\n"
+        "paddh $f2, $f2, $f4                        \r\n"
+        "psubh $f14, $f14, $f4                      \r\n"
+        "psubh $f0, $f0, $f4                        \r\n"
+        "psrah $f4, $f4, $f16                       \r\n"
+        "paddh $f14, $f14, $f8                      \r\n"
+        "psubh $f0, $f0, $f8                        \r\n"
+        "psrah $f8, $f8, $f16                       \r\n"
+        "psubh $f14, $f14, $f4                      \r\n"
+        "psubh $f0, $f0, $f8                        \r\n"
+        "mov.d $f8, $f2                             \r\n"
+        "psrah $f4, $f10, $f18                      \r\n"
+        "psrah $f2, $f2, $f18                       \r\n"
+        "paddh $f4, $f4, $f14                       \r\n"
+        "psrah $f14, $f14, $f18                     \r\n"
+        "paddh $f2, $f2, $f0                        \r\n"
+        "psrah $f0, $f0, $f18                       \r\n"
+        "psubh $f14, $f14, $f10                     \r\n"
+        "psubh $f8, $f8, $f0                        \r\n"
+        "mov.d $f0, $f6                             \r\n"
+        "psrah $f6, $f6, $f16                       \r\n"
+        "psrah $f10, $f12, $f16                     \r\n"
+        "paddh $f6, $f6, $f12                       \r\n"
+        "psubh $f10, $f10, $f0                      \r\n"
+        "ldc1 $f12, 0x8(%[block])                   \r\n"
+        "ldc1 $f0, 0x48(%[block])                   \r\n"
+        "paddh $f0, $f0, $f12                       \r\n"
+        "paddh $f12, $f12, $f12                     \r\n"
+        "paddh $f6, $f6, $f0                        \r\n"
+        "psubh $f12, $f12, $f0                      \r\n"
+        "paddh $f0, $f0, $f0                        \r\n"
+        "paddh $f10, $f10, $f12                     \r\n"
+        "psubh $f0, $f0, $f6                        \r\n"
+        "paddh $f12, $f12, $f12                     \r\n"
+        "paddh $f8, $f8, $f6                        \r\n"
+        "psubh $f12, $f12, $f10                     \r\n"
+        "paddh $f6, $f6, $f6                        \r\n"
+        "paddh $f14, $f14, $f10                     \r\n"
+        "psubh $f6, $f6, $f8                        \r\n"
+        "paddh $f10, $f10, $f10                     \r\n"
+        "paddh $f4, $f4, $f12                       \r\n"
+        "psubh $f10, $f10, $f14                     \r\n"
+        "paddh $f12, $f12, $f12                     \r\n"
+        "paddh $f2, $f2, $f0                        \r\n"
+        "psubh $f12, $f12, $f4                      \r\n"
+        "paddh $f0, $f0, $f0                        \r\n"
+        "sdc1 $f6, 0x8(%[block])                    \r\n"
+        "psubh $f0, $f0, $f2                        \r\n"
+        "punpckhhw $f6, $f8, $f14                   \r\n"
+        "punpcklhw $f8, $f8, $f14                   \r\n"
+        "punpckhhw $f14, $f4, $f2                   \r\n"
+        "punpcklhw $f4, $f4, $f2                    \r\n"
+        "punpckhwd $f2, $f8, $f4                    \r\n"
+        "punpcklwd $f8, $f8, $f4                    \r\n"
+        "punpckhwd $f4, $f6, $f14                   \r\n"
+        "punpcklwd $f6, $f6, $f14                   \r\n"
+        "ldc1 $f14, 0x8(%[block])                   \r\n"
+        "dmfc1 $13, $f8                             \r\n"
+        "dmfc1 $15, $f2                             \r\n"
+        "mov.d $f24, $f6                            \r\n"
+        "mov.d $f28, $f4                            \r\n"
+        "punpckhhw $f4, $f0, $f12                   \r\n"
+        "punpcklhw $f0, $f0, $f12                   \r\n"
+        "punpckhhw $f12, $f10, $f14                 \r\n"
+        "punpcklhw $f10, $f10, $f14                 \r\n"
+        "punpckhwd $f14, $f0, $f10                  \r\n"
+        "punpcklwd $f0, $f0, $f10                   \r\n"
+        "punpckhwd $f10, $f4, $f12                  \r\n"
+        "punpcklwd $f4, $f4, $f12                   \r\n"
+        "dmfc1 $14, $f0                             \r\n"
+        "mov.d $f22, $f14                           \r\n"
+        "mov.d $f26, $f4                            \r\n"
+        "mov.d $f30, $f10                           \r\n"
+        "daddiu $10, %[dst], 0x4                    \r\n"
+        "dmtc1 $15, $f14                            \r\n"
+        "dmtc1 $11, $f12                            \r\n"
+        "ldc1 $f2, 0x10($29)                        \r\n"
+        "dmtc1 $8, $f6                              \r\n"
+        "mov.d $f8, $f2                             \r\n"
+        "psrah $f2, $f2, $f16                       \r\n"
+        "psrah $f0, $f14, $f16                      \r\n"
+        "paddh $f2, $f2, $f8                        \r\n"
+        "paddh $f0, $f0, $f14                       \r\n"
+        "paddh $f2, $f2, $f14                       \r\n"
+        "paddh $f0, $f0, $f28                       \r\n"
+        "paddh $f2, $f2, $f12                       \r\n"
+        "psubh $f0, $f0, $f8                        \r\n"
+        "psubh $f8, $f8, $f12                       \r\n"
+        "psubh $f14, $f14, $f12                     \r\n"
+        "psrah $f12, $f12, $f16                     \r\n"
+        "paddh $f8, $f8, $f28                       \r\n"
+        "psubh $f14, $f14, $f28                     \r\n"
+        "psrah $f10, $f28, $f16                     \r\n"
+        "psubh $f8, $f8, $f12                       \r\n"
+        "psubh $f14, $f14, $f10                     \r\n"
+        "mov.d $f10, $f2                            \r\n"
+        "psrah $f2, $f2, $f18                       \r\n"
+        "psrah $f12, $f0, $f18                      \r\n"
+        "paddh $f2, $f2, $f14                       \r\n"
+        "paddh $f12, $f12, $f8                      \r\n"
+        "psrah $f8, $f8, $f18                       \r\n"
+        "psrah $f14, $f14, $f18                     \r\n"
+        "psubh $f8, $f8, $f0                        \r\n"
+        "psubh $f10, $f10, $f14                     \r\n"
+        "mov.d $f14, $f24                           \r\n"
+        "psrah $f4, $f24, $f16                      \r\n"
+        "psrah $f0, $f6, $f16                       \r\n"
+        "paddh $f4, $f4, $f6                        \r\n"
+        "psubh $f0, $f0, $f14                       \r\n"
+        "ldc1 $f6, 0x0($29)                         \r\n"
+        "dmtc1 $13, $f14                            \r\n"
+        "paddh $f14, $f14, $f6                      \r\n"
+        "paddh $f6, $f6, $f6                        \r\n"
+        "paddh $f4, $f4, $f14                       \r\n"
+        "psubh $f6, $f6, $f14                       \r\n"
+        "paddh $f14, $f14, $f14                     \r\n"
+        "paddh $f0, $f0, $f6                        \r\n"
+        "psubh $f14, $f14, $f4                      \r\n"
+        "paddh $f6, $f6, $f6                        \r\n"
+        "paddh $f10, $f10, $f4                      \r\n"
+        "psubh $f6, $f6, $f0                        \r\n"
+        "paddh $f4, $f4, $f4                        \r\n"
+        "paddh $f8, $f8, $f0                        \r\n"
+        "psubh $f4, $f4, $f10                       \r\n"
+        "paddh $f0, $f0, $f0                        \r\n"
+        "paddh $f12, $f12, $f6                      \r\n"
+        "psubh $f0, $f0, $f8                        \r\n"
+        "paddh $f6, $f6, $f6                        \r\n"
+        "paddh $f2, $f2, $f14                       \r\n"
+        "psubh $f6, $f6, $f12                       \r\n"
+        "paddh $f14, $f14, $f14                     \r\n"
+        "sdc1 $f6, 0x0($29)                         \r\n"
+        "psubh $f14, $f14, $f2                      \r\n"
+        "sdc1 $f0, 0x10($29)                        \r\n"
+        "dmfc1 $8, $f4                              \r\n"
+        "xor $f4, $f4, $f4                          \r\n"
+        "sdc1 $f4, 0x0(%[block])                    \r\n"
+        "sdc1 $f4, 0x8(%[block])                    \r\n"
+        "sdc1 $f4, 0x10(%[block])                   \r\n"
+        "sdc1 $f4, 0x18(%[block])                   \r\n"
+        "sdc1 $f4, 0x20(%[block])                   \r\n"
+        "sdc1 $f4, 0x28(%[block])                   \r\n"
+        "sdc1 $f4, 0x30(%[block])                   \r\n"
+        "sdc1 $f4, 0x38(%[block])                   \r\n"
+        "sdc1 $f4, 0x40(%[block])                   \r\n"
+        "sdc1 $f4, 0x48(%[block])                   \r\n"
+        "sdc1 $f4, 0x50(%[block])                   \r\n"
+        "sdc1 $f4, 0x58(%[block])                   \r\n"
+        "sdc1 $f4, 0x60(%[block])                   \r\n"
+        "sdc1 $f4, 0x68(%[block])                   \r\n"
+        "sdc1 $f4, 0x70(%[block])                   \r\n"
+        "sdc1 $f4, 0x78(%[block])                   \r\n"
+        "dli $11, 0x6                               \r\n"
+        "lwc1 $f6, 0x0(%[dst])                      \r\n"
+        "dmtc1 $11, $f20                            \r\n"
+        "gslwxc1 $f0, 0x0(%[dst], %[stride])        \r\n"
+        "psrah $f10, $f10, $f20                     \r\n"
+        "psrah $f8, $f8, $f20                       \r\n"
+        "punpcklbh $f6, $f6, $f4                    \r\n"
+        "punpcklbh $f0, $f0, $f4                    \r\n"
+        "paddh $f6, $f6, $f10                       \r\n"
+        "paddh $f0, $f0, $f8                        \r\n"
+        "packushb $f6, $f6, $f4                     \r\n"
+        "packushb $f0, $f0, $f4                     \r\n"
+        "swc1 $f6, 0x0(%[dst])                      \r\n"
+        "gsswxc1 $f0, 0x0(%[dst], %[stride])        \r\n"
+        "daddu %[dst], %[dst], %[stride]            \r\n"
+        "daddu %[dst], %[dst], %[stride]            \r\n"
+        "lwc1 $f6, 0x0(%[dst])                      \r\n"
+        "gslwxc1 $f0, 0x0(%[dst], %[stride])        \r\n"
+        "psrah $f12, $f12, $f20                     \r\n"
+        "psrah $f2, $f2, $f20                       \r\n"
+        "punpcklbh $f6, $f6, $f4                    \r\n"
+        "punpcklbh $f0, $f0, $f4                    \r\n"
+        "paddh $f6, $f6, $f12                       \r\n"
+        "paddh $f0, $f0, $f2                        \r\n"
+        "packushb $f6, $f6, $f4                     \r\n"
+        "packushb $f0, $f0, $f4                     \r\n"
+        "swc1 $f6, 0x0(%[dst])                      \r\n"
+        "gsswxc1 $f0, 0x0(%[dst], %[stride])        \r\n"
+        "ldc1 $f10, 0x0($29)                        \r\n"
+        "ldc1 $f8, 0x10($29)                        \r\n"
+        "dmtc1 $8, $f12                             \r\n"
+        "daddu %[dst], %[dst], %[stride]            \r\n"
+        "daddu %[dst], %[dst], %[stride]            \r\n"
+        "lwc1 $f6, 0x0(%[dst])                      \r\n"
+        "gslwxc1 $f0, 0x0(%[dst], %[stride])        \r\n"
+        "psrah $f14, $f14, $f20                     \r\n"
+        "psrah $f10, $f10, $f20                     \r\n"
+        "punpcklbh $f6, $f6, $f4                    \r\n"
+        "punpcklbh $f0, $f0, $f4                    \r\n"
+        "paddh $f6, $f6, $f14                       \r\n"
+        "paddh $f0, $f0, $f10                       \r\n"
+        "packushb $f6, $f6, $f4                     \r\n"
+        "packushb $f0, $f0, $f4                     \r\n"
+        "swc1 $f6, 0x0(%[dst])                      \r\n"
+        "gsswxc1 $f0, 0x0(%[dst], %[stride])        \r\n"
+        "daddu %[dst], %[dst], %[stride]            \r\n"
+        "daddu %[dst], %[dst], %[stride]            \r\n"
+        "lwc1 $f6, 0x0(%[dst])                      \r\n"
+        "gslwxc1 $f0, 0x0(%[dst], %[stride])        \r\n"
+        "psrah $f8, $f8, $f20                       \r\n"
+        "psrah $f12, $f12, $f20                     \r\n"
+        "punpcklbh $f6, $f6, $f4                    \r\n"
+        "punpcklbh $f0, $f0, $f4                    \r\n"
+        "paddh $f6, $f6, $f8                        \r\n"
+        "paddh $f0, $f0, $f12                       \r\n"
+        "packushb $f6, $f6, $f4                     \r\n"
+        "packushb $f0, $f0, $f4                     \r\n"
+        "swc1 $f6, 0x0(%[dst])                      \r\n"
+        "gsswxc1 $f0, 0x0(%[dst], %[stride])        \r\n"
+        "dmtc1 $12, $f2                             \r\n"
+        "dmtc1 $9, $f12                             \r\n"
+        "ldc1 $f8, 0x18($29)                        \r\n"
+        "mov.d $f10, $f8                            \r\n"
+        "psrah $f8, $f8, $f16                       \r\n"
+        "psrah $f14, $f22, $f16                     \r\n"
+        "paddh $f14, $f14, $f22                     \r\n"
+        "paddh $f8, $f8, $f10                       \r\n"
+        "paddh $f14, $f14, $f30                     \r\n"
+        "paddh $f8, $f8, $f22                       \r\n"
+        "psubh $f14, $f14, $f10                     \r\n"
+        "paddh $f8, $f8, $f2                        \r\n"
+        "psubh $f10, $f10, $f2                      \r\n"
+        "psubh $f6, $f22, $f2                       \r\n"
+        "psrah $f2, $f2, $f16                       \r\n"
+        "paddh $f10, $f10, $f30                     \r\n"
+        "psubh $f6, $f6, $f30                       \r\n"
+        "psrah $f4, $f30, $f16                      \r\n"
+        "psubh $f10, $f10, $f2                      \r\n"
+        "psubh $f6, $f6, $f4                        \r\n"
+        "mov.d $f4, $f8                             \r\n"
+        "psrah $f8, $f8, $f18                       \r\n"
+        "psrah $f2, $f14, $f18                      \r\n"
+        "paddh $f8, $f8, $f6                        \r\n"
+        "paddh $f2, $f2, $f10                       \r\n"
+        "psrah $f10, $f10, $f18                     \r\n"
+        "psrah $f6, $f6, $f18                       \r\n"
+        "psubh $f10, $f10, $f14                     \r\n"
+        "psubh $f4, $f4, $f6                        \r\n"
+        "mov.d $f6, $f26                            \r\n"
+        "psrah $f0, $f26, $f16                      \r\n"
+        "psrah $f14, $f12, $f16                     \r\n"
+        "paddh $f0, $f0, $f12                       \r\n"
+        "psubh $f14, $f14, $f6                      \r\n"
+        "ldc1 $f12, 0x8($29)                        \r\n"
+        "dmtc1 $14, $f6                             \r\n"
+        "paddh $f6, $f6, $f12                       \r\n"
+        "paddh $f12, $f12, $f12                     \r\n"
+        "paddh $f0, $f0, $f6                        \r\n"
+        "psubh $f12, $f12, $f6                      \r\n"
+        "paddh $f6, $f6, $f6                        \r\n"
+        "paddh $f14, $f14, $f12                     \r\n"
+        "psubh $f6, $f6, $f0                        \r\n"
+        "paddh $f12, $f12, $f12                     \r\n"
+        "paddh $f4, $f4, $f0                        \r\n"
+        "psubh $f12, $f12, $f14                     \r\n"
+        "paddh $f0, $f0, $f0                        \r\n"
+        "paddh $f10, $f10, $f14                     \r\n"
+        "psubh $f0, $f0, $f4                        \r\n"
+        "paddh $f14, $f14, $f14                     \r\n"
+        "paddh $f2, $f2, $f12                       \r\n"
+        "psubh $f14, $f14, $f10                     \r\n"
+        "paddh $f12, $f12, $f12                     \r\n"
+        "paddh $f8, $f8, $f6                        \r\n"
+        "psubh $f12, $f12, $f2                      \r\n"
+        "paddh $f6, $f6, $f6                        \r\n"
+        "sdc1 $f12, 0x8($29)                        \r\n"
+        "psubh $f6, $f6, $f8                        \r\n"
+        "sdc1 $f14, 0x18($29)                       \r\n"
+        "dmfc1 $9, $f0                              \r\n"
+        "xor $f0, $f0, $f0                          \r\n"
+        "lwc1 $f12, 0x0($10)                        \r\n"
+        "gslwxc1 $f14, 0x0($10, %[stride])          \r\n"
+        "psrah $f4, $f4, $f20                       \r\n"
+        "psrah $f10, $f10, $f20                     \r\n"
+        "punpcklbh $f12, $f12, $f0                  \r\n"
+        "punpcklbh $f14, $f14, $f0                  \r\n"
+        "paddh $f12, $f12, $f4                      \r\n"
+        "paddh $f14, $f14, $f10                     \r\n"
+        "packushb $f12, $f12, $f0                   \r\n"
+        "packushb $f14, $f14, $f0                   \r\n"
+        "swc1 $f12, 0x0($10)                        \r\n"
+        "gsswxc1 $f14, 0x0($10, %[stride])          \r\n"
+        "daddu $10, $10, %[stride]                  \r\n"
+        "daddu $10, $10, %[stride]                  \r\n"
+        "lwc1 $f12, 0x0($10)                        \r\n"
+        "gslwxc1 $f14, 0x0($10, %[stride])          \r\n"
+        "psrah $f2, $f2, $f20                       \r\n"
+        "psrah $f8, $f8, $f20                       \r\n"
+        "punpcklbh $f12, $f12, $f0                  \r\n"
+        "punpcklbh $f14, $f14, $f0                  \r\n"
+        "paddh $f12, $f12, $f2                      \r\n"
+        "paddh $f14, $f14, $f8                      \r\n"
+        "packushb $f12, $f12, $f0                   \r\n"
+        "packushb $f14, $f14, $f0                   \r\n"
+        "swc1 $f12, 0x0($10)                        \r\n"
+        "gsswxc1 $f14, 0x0($10, %[stride])          \r\n"
+        "ldc1 $f4, 0x8($29)                         \r\n"
+        "ldc1 $f10, 0x18($29)                       \r\n"
+        "daddu $10, $10, %[stride]                  \r\n"
+        "dmtc1 $9, $f2                              \r\n"
+        "daddu $10, $10, %[stride]                  \r\n"
+        "lwc1 $f12, 0x0($10)                        \r\n"
+        "gslwxc1 $f14, 0x0($10, %[stride])          \r\n"
+        "psrah $f6, $f6, $f20                       \r\n"
+        "psrah $f4, $f4, $f20                       \r\n"
+        "punpcklbh $f12, $f12, $f0                  \r\n"
+        "punpcklbh $f14, $f14, $f0                  \r\n"
+        "paddh $f12, $f12, $f6                      \r\n"
+        "paddh $f14, $f14, $f4                      \r\n"
+        "packushb $f12, $f12, $f0                   \r\n"
+        "packushb $f14, $f14, $f0                   \r\n"
+        "swc1 $f12, 0x0($10)                        \r\n"
+        "gsswxc1 $f14, 0x0($10, %[stride])          \r\n"
+        "daddu $10, $10, %[stride]                  \r\n"
+        "daddu $10, $10, %[stride]                  \r\n"
+        "lwc1 $f12, 0x0($10)                        \r\n"
+        "gslwxc1 $f14, 0x0($10, %[stride])          \r\n"
+        "psrah $f10, $f10, $f20                     \r\n"
+        "psrah $f2, $f2, $f20                       \r\n"
+        "punpcklbh $f12, $f12, $f0                  \r\n"
+        "punpcklbh $f14, $f14, $f0                  \r\n"
+        "paddh $f12, $f12, $f10                     \r\n"
+        "paddh $f14, $f14, $f2                      \r\n"
+        "packushb $f12, $f12, $f0                   \r\n"
+        "packushb $f14, $f14, $f0                   \r\n"
+        "swc1 $f12, 0x0($10)                        \r\n"
+        "gsswxc1 $f14, 0x0($10, %[stride])          \r\n"
+        "daddiu $29, $29, 0x20                      \r\n"
+        ::[dst]"r"(dst),[block]"r"(block),[stride]"r"((uint64_t)stride)
+        :"$8","$9","$10","$11","$12","$13","$14","$15","$29","$f0","$f2","$f4",
+         "$f8","$f10","$f12","$f14","$f16","$f18","$f20","$f22","$f24","$f26",
+         "$f28","$f30"
+    );
+
+    memset(block, 0, 128);
+}
+
+void ff_h264_idct_dc_add_8_mmi(uint8_t *dst, int16_t *block, int stride)
+{
+    __asm__ volatile (
+        "lh $8, 0x0(%[block])                       \r\n"
+        "sd $0, 0x0(%[block])                       \r\n"
+        "daddiu $8, $8, 0x20                        \r\n"
+        "daddu $10, %[stride], %[stride]            \r\n"
+        "dsra $8, $8, 0x6                           \r\n"
+        "xor $f2, $f2, $f2                          \r\n"
+        "mtc1 $8, $f0                               \r\n"
+        "pshufh $f0, $f0, $f2                       \r\n"
+        "daddu $8, $10, %[stride]                   \r\n"
+        "psubh $f2, $f2, $f0                        \r\n"
+        "packushb $f0, $f0, $f0                     \r\n"
+        "packushb $f2, $f2, $f2                     \r\n"
+        "lwc1 $f4, 0x0(%[dst])                      \r\n"
+        "gslwxc1 $f6, 0x0(%[dst], %[stride])        \r\n"
+        "gslwxc1 $f8, 0x0(%[dst], $10)              \r\n"
+        "gslwxc1 $f10, 0x0(%[dst], $8)              \r\n"
+        "paddusb $f4, $f4, $f0                      \r\n"
+        "paddusb $f6, $f6, $f0                      \r\n"
+        "paddusb $f8, $f8, $f0                      \r\n"
+        "paddusb $f10, $f10, $f0                    \r\n"
+        "psubusb $f4, $f4, $f2                      \r\n"
+        "psubusb $f6, $f6, $f2                      \r\n"
+        "psubusb $f8, $f8, $f2                      \r\n"
+        "psubusb $f10, $f10, $f2                    \r\n"
+        "swc1 $f4, 0x0(%[dst])                      \r\n"
+        "gsswxc1 $f6, 0x0(%[dst], %[stride])        \r\n"
+        "gsswxc1 $f8, 0x0(%[dst], $10)              \r\n"
+        "gsswxc1 $f10, 0x0(%[dst], $8)              \r\n"
+        ::[dst]"r"(dst),[block]"r"(block),[stride]"r"((uint64_t)stride)
+        : "$8","$10","$f0","$f2","$f4","$f6","$f8","$f10"
+    );
+}
+
+void ff_h264_idct8_dc_add_8_mmi(uint8_t *dst, int16_t *block, int stride)
+{
+    __asm__ volatile (
+        "lh $8, 0x0(%[block])                       \r\n"
+        "sd $0, 0x0(%[block])                       \r\n"
+        "daddiu $8, $8, 0x20                        \r\n"
+        "daddu $10, %[stride], %[stride]            \r\n"
+        "dsra $8, $8, 0x6                           \r\n"
+        "xor $f2, $f2, $f2                          \r\n"
+        "mtc1 $8, $f0                               \r\n"
+        "pshufh $f0, $f0, $f2                       \r\n"
+        "daddu $8, $10, %[stride]                   \r\n"
+        "psubh $f2, $f2, $f0                        \r\n"
+        "packushb $f0, $f0, $f0                     \r\n"
+        "packushb $f2, $f2, $f2                     \r\n"
+        "ldc1 $f4, 0x0(%[dst])                      \r\n"
+        "gsldxc1 $f6, 0x0(%[dst], %[stride])        \r\n"
+        "gsldxc1 $f8, 0x0(%[dst], $10)              \r\n"
+        "gsldxc1 $f10, 0x0(%[dst], $8)              \r\n"
+        "paddusb $f4, $f4, $f0                      \r\n"
+        "paddusb $f6, $f6, $f0                      \r\n"
+        "paddusb $f8, $f8, $f0                      \r\n"
+        "paddusb $f10, $f10, $f0                    \r\n"
+        "psubusb $f4, $f4, $f2                      \r\n"
+        "psubusb $f6, $f6, $f2                      \r\n"
+        "psubusb $f8, $f8, $f2                      \r\n"
+        "psubusb $f10, $f10, $f2                    \r\n"
+        "sdc1 $f4, 0x0(%[dst])                      \r\n"
+        "gssdxc1 $f6, 0x0(%[dst], %[stride])        \r\n"
+        "gssdxc1 $f8, 0x0(%[dst], $10)              \r\n"
+        "daddu $9, $10, $10                         \r\n"
+        "gssdxc1 $f10, 0x0(%[dst], $8)              \r\n"
+        "daddu %[dst], %[dst], $9                   \r\n"
+        "ldc1 $f4, 0x0(%[dst])                      \r\n"
+        "gsldxc1 $f6, 0x0(%[dst], %[stride])        \r\n"
+        "gsldxc1 $f8, 0x0(%[dst], $10)              \r\n"
+        "gsldxc1 $f10, 0x0(%[dst], $8)              \r\n"
+        "paddusb $f4, $f4, $f0                      \r\n"
+        "paddusb $f6, $f6, $f0                      \r\n"
+        "paddusb $f8, $f8, $f0                      \r\n"
+        "paddusb $f10, $f10, $f0                    \r\n"
+        "psubusb $f4, $f4, $f2                      \r\n"
+        "psubusb $f6, $f6, $f2                      \r\n"
+        "psubusb $f8, $f8, $f2                      \r\n"
+        "psubusb $f10, $f10, $f2                    \r\n"
+        "sdc1 $f4, 0x0(%[dst])                      \r\n"
+        "gssdxc1 $f6, 0x0(%[dst], %[stride])        \r\n"
+        "gssdxc1 $f8, 0x0(%[dst], $10)              \r\n"
+        "gssdxc1 $f10, 0x0(%[dst], $8)              \r\n"
+        ::[dst]"r"(dst),[block]"r"(block),[stride]"r"((uint64_t)stride)
+        : "$8","$9","$10","$f0","$f2","$f4","$f6","$f8","$f10"
+    );
+}
+
+void ff_h264_idct_add16_8_mmi(uint8_t *dst, const int *block_offset,
+        int16_t *block, int stride, const uint8_t nnzc[15*8])
+{
+    int i;
+    for(i=0; i<16; i++){
+        int nnz = nnzc[ scan8[i] ];
+        if(nnz){
+            if(nnz==1 && ((int16_t*)block)[i*16])
+                ff_h264_idct_dc_add_8_mmi(dst + block_offset[i], block + i*16,
+                        stride);
+            else
+                ff_h264_idct_add_8_mmi(dst + block_offset[i], block + i*16,
+                        stride);
+        }
+    }
+}
+
+void ff_h264_idct_add16intra_8_mmi(uint8_t *dst, const int *block_offset,
+        int16_t *block, int stride, const uint8_t nnzc[15*8])
+{
+    int i;
+    for(i=0; i<16; i++){
+        if(nnzc[ scan8[i] ])
+            ff_h264_idct_add_8_mmi(dst + block_offset[i], block + i*16, stride);
+        else if(((int16_t*)block)[i*16])
+            ff_h264_idct_dc_add_8_mmi(dst + block_offset[i], block + i*16,
+                    stride);
+    }
+}
+
+void ff_h264_idct8_add4_8_mmi(uint8_t *dst, const int *block_offset,
+        int16_t *block, int stride, const uint8_t nnzc[15*8])
+{
+    int i;
+    for(i=0; i<16; i+=4){
+        int nnz = nnzc[ scan8[i] ];
+        if(nnz){
+            if(nnz==1 && ((int16_t*)block)[i*16])
+                ff_h264_idct8_dc_add_8_mmi(dst + block_offset[i],
+                        block + i*16, stride);
+            else
+                ff_h264_idct8_add_8_mmi(dst + block_offset[i], block + i*16,
+                        stride);
+        }
+    }
+}
+
+void ff_h264_idct_add8_8_mmi(uint8_t **dest, const int *block_offset,
+        int16_t *block, int stride, const uint8_t nnzc[15*8])
+{
+    int i, j;
+    for(j=1; j<3; j++){
+        for(i=j*16; i<j*16+4; i++){
+            if(nnzc[ scan8[i] ])
+                ff_h264_idct_add_8_mmi(dest[j-1] + block_offset[i],
+                        block + i*16, stride);
+            else if(((int16_t*)block)[i*16])
+                ff_h264_idct_dc_add_8_mmi(dest[j-1] + block_offset[i],
+                        block + i*16, stride);
+        }
+    }
+}
+
+void ff_h264_idct_add8_422_8_mmi(uint8_t **dest, const int *block_offset,
+        int16_t *block, int stride, const uint8_t nnzc[15*8])
+{
+    int i, j;
+
+    for(j=1; j<3; j++){
+        for(i=j*16; i<j*16+4; i++){
+            if(nnzc[ scan8[i] ])
+                ff_h264_idct_add_8_mmi(dest[j-1] + block_offset[i],
+                        block + i*16, stride);
+            else if(((int16_t*)block)[i*16])
+                ff_h264_idct_dc_add_8_mmi(dest[j-1] + block_offset[i],
+                        block + i*16, stride);
+        }
+    }
+
+    for(j=1; j<3; j++){
+        for(i=j*16+4; i<j*16+8; i++){
+            if(nnzc[ scan8[i+4] ])
+                ff_h264_idct_add_8_mmi(dest[j-1] + block_offset[i+4],
+                        block + i*16, stride);
+            else if(((int16_t*)block)[i*16])
+                ff_h264_idct_dc_add_8_mmi(dest[j-1] + block_offset[i+4],
+                        block + i*16, stride);
+        }
+    }
+}
+
+void ff_h264_luma_dc_dequant_idct_8_mmi(int16_t *output, int16_t *input,
+        int qmul)
+{
+    __asm__ volatile (
+        ".set noreorder                                 \r\n"
+        "dli $10, 0x8                                   \r\n"
+        "ldc1 $f6, 0x18(%[input])                       \r\n"
+        "dmtc1 $10, $f16                                \r\n"
+        "ldc1 $f4, 0x10(%[input])                       \r\n"
+        "dli $10, 0x20                                  \r\n"
+        "ldc1 $f2, 0x8(%[input])                        \r\n"
+        "dmtc1 $10, $f18                                \r\n"
+        "ldc1 $f0, 0x0(%[input])                        \r\n"
+        "mov.d $f8, $f6                                 \r\n"
+        "paddh $f6, $f6, $f4                            \r\n"
+        "psubh $f4, $f4, $f8                            \r\n"
+        "mov.d $f8, $f2                                 \r\n"
+        "paddh $f2, $f2, $f0                            \r\n"
+        "psubh $f0, $f0, $f8                            \r\n"
+        "mov.d $f8, $f6                                 \r\n"
+        "paddh $f6, $f6, $f2                            \r\n"
+        "psubh $f2, $f2, $f8                            \r\n"
+        "mov.d $f8, $f4                                 \r\n"
+        "paddh $f4, $f4, $f0                            \r\n"
+        "psubh $f0, $f0, $f8                            \r\n"
+        "mov.d $f8, $f6                                 \r\n"
+        "punpcklhw $f6, $f6, $f2                        \r\n"
+        "punpckhhw $f8, $f8, $f2                        \r\n"
+        "punpckhhw $f2, $f0, $f4                        \r\n"
+        "punpcklhw $f0, $f0, $f4                        \r\n"
+        "punpckhwd $f4, $f6, $f0                        \r\n"
+        "punpcklwd $f6, $f6, $f0                        \r\n"
+        "mov.d $f0, $f8                                 \r\n"
+        "punpcklwd $f8, $f8, $f2                        \r\n"
+        "punpckhwd $f0, $f0, $f2                        \r\n"
+        "mov.d $f2, $f0                                 \r\n"
+        "paddh $f0, $f0, $f8                            \r\n"
+        "psubh $f8, $f8, $f2                            \r\n"
+        "mov.d $f2, $f4                                 \r\n"
+        "paddh $f4, $f4, $f6                            \r\n"
+        "psubh $f6, $f6, $f2                            \r\n"
+        "mov.d $f2, $f0                                 \r\n"
+        "paddh $f0, $f0, $f4                            \r\n"
+        "psubh $f4, $f4, $f2                            \r\n"
+        "mov.d $f2, $f8                                 \r\n"
+        "daddiu $10, %[qmul], -0x7fff                   \r\n"
+        "paddh $f8, $f8, $f6                            \r\n"
+        "bgtz $10, 1f                                   \r\n"
+        "psubh $f6, $f6, $f2                            \r\n"
+        "ori $10, $0, 0x80                              \r\n"
+        "dsll $10, $10, 0x10                            \r\n"
+        "punpckhhw $f2, $f0, %[ff_pw_1]                 \r\n"
+        "daddu %[qmul], %[qmul], $10                    \r\n"
+        "punpcklhw $f0, $f0, %[ff_pw_1]                 \r\n"
+        "punpckhhw $f10, $f4, %[ff_pw_1]                \r\n"
+        "punpcklhw $f4, $f4, %[ff_pw_1]                 \r\n"
+        "mtc1 %[qmul], $f14                             \r\n"
+        "punpcklwd $f14, $f14, $f14                     \r\n"
+        "pmaddhw $f0, $f0, $f14                         \r\n"
+        "pmaddhw $f4, $f4, $f14                         \r\n"
+        "pmaddhw $f2, $f2, $f14                         \r\n"
+        "pmaddhw $f10, $f10, $f14                       \r\n"
+        "psraw $f0, $f0, $f16                           \r\n"
+        "psraw $f4, $f4, $f16                           \r\n"
+        "psraw $f2, $f2, $f16                           \r\n"
+        "psraw $f10, $f10, $f16                         \r\n"
+        "packsswh $f0, $f0, $f2                         \r\n"
+        "packsswh $f4, $f4, $f10                        \r\n"
+        "mfc1 $9, $f0                                   \r\n"
+        "dsrl $f0, $f0, $f18                            \r\n"
+        "mfc1 %[input], $f0                             \r\n"
+        "sh $9, 0x0(%[output])                          \r\n"
+        "sh %[input], 0x80(%[output])                   \r\n"
+        "dsrl $9, $9, 0x10                              \r\n"
+        "dsrl %[input], %[input], 0x10                  \r\n"
+        "sh $9, 0x20(%[output])                         \r\n"
+        "sh %[input], 0xa0(%[output])                   \r\n"
+        "mfc1 $9, $f4                                   \r\n"
+        "dsrl $f4, $f4, $f18                            \r\n"
+        "mfc1 %[input], $f4                             \r\n"
+        "sh $9, 0x40(%[output])                         \r\n"
+        "sh %[input], 0xc0(%[output])                   \r\n"
+        "dsrl $9, $9, 0x10                              \r\n"
+        "dsrl %[input], %[input], 0x10                  \r\n"
+        "sh $9, 0x60(%[output])                         \r\n"
+        "sh %[input], 0xe0(%[output])                   \r\n"
+        "punpckhhw $f2, $f6, %[ff_pw_1]                 \r\n"
+        "punpcklhw $f6, $f6, %[ff_pw_1]                 \r\n"
+        "punpckhhw $f10, $f8, %[ff_pw_1]                \r\n"
+        "punpcklhw $f8, $f8, %[ff_pw_1]                 \r\n"
+        "mtc1 %[qmul], $f14                             \r\n"
+        "punpcklwd $f14, $f14, $f14                     \r\n"
+        "pmaddhw $f6, $f6, $f14                         \r\n"
+        "pmaddhw $f8, $f8, $f14                         \r\n"
+        "pmaddhw $f2, $f2, $f14                         \r\n"
+        "pmaddhw $f10, $f10, $f14                       \r\n"
+        "psraw $f6, $f6, $f16                           \r\n"
+        "psraw $f8, $f8, $f16                           \r\n"
+        "psraw $f2, $f2, $f16                           \r\n"
+        "psraw $f10, $f10, $f16                         \r\n"
+        "packsswh $f6, $f6, $f2                         \r\n"
+        "packsswh $f8, $f8, $f10                        \r\n"
+        "mfc1 $9, $f6                                   \r\n"
+        "dsrl $f6, $f6, $f18                            \r\n"
+        "mfc1 %[input], $f6                             \r\n"
+        "sh $9, 0x100(%[output])                        \r\n"
+        "sh %[input], 0x180(%[output])                  \r\n"
+        "dsrl $9, $9, 0x10                              \r\n"
+        "dsrl %[input], %[input], 0x10                  \r\n"
+        "sh $9, 0x120(%[output])                        \r\n"
+        "sh %[input], 0x1a0(%[output])                  \r\n"
+        "mfc1 $9, $f8                                   \r\n"
+        "dsrl $f8, $f8, $f18                            \r\n"
+        "mfc1 %[input], $f8                             \r\n"
+        "sh $9, 0x140(%[output])                        \r\n"
+        "sh %[input], 0x1c0(%[output])                  \r\n"
+        "dsrl $9, $9, 0x10                              \r\n"
+        "dsrl %[input], %[input], 0x10                  \r\n"
+        "sh $9, 0x160(%[output])                        \r\n"
+        "jr $31                                         \r\n"
+        "sh %[input], 0x1e0(%[output])                  \r\n"
+        "1:                                             \r\n"
+        "ori $10, $0, 0x1f                              \r\n"
+        "clz $9, %[qmul]                                \r\n"
+        "ori %[input], $0, 0x7                          \r\n"
+        "dsubu $9, $10, $9                              \r\n"
+        "ori $10, $0, 0x80                              \r\n"
+        "dsll $10, $10, 0x10                            \r\n"
+        "daddu %[qmul], %[qmul], $10                    \r\n"
+        "dsubu $10, $9, %[input]                        \r\n"
+        "movn $9, %[input], $10                         \r\n"
+        "daddiu %[input], %[input], 0x1                 \r\n"
+        "andi $10, $9, 0xff                             \r\n"
+        "dsrlv %[qmul], %[qmul], $10                    \r\n"
+        "dsubu %[input], %[input], $9                   \r\n"
+        "mtc1 %[input], $f12                            \r\n"
+        "punpckhhw $f2, $f0, %[ff_pw_1]                 \r\n"
+        "punpcklhw $f0, $f0, %[ff_pw_1]                 \r\n"
+        "punpckhhw $f10, $f4, %[ff_pw_1]                \r\n"
+        "punpcklhw $f4, $f4, %[ff_pw_1]                 \r\n"
+        "mtc1 %[qmul], $f14                             \r\n"
+        "punpcklwd $f14, $f14, $f14                     \r\n"
+        "pmaddhw $f0, $f0, $f14                         \r\n"
+        "pmaddhw $f4, $f4, $f14                         \r\n"
+        "pmaddhw $f2, $f2, $f14                         \r\n"
+        "pmaddhw $f10, $f10, $f14                       \r\n"
+        "psraw $f0, $f0, $f12                           \r\n"
+        "psraw $f4, $f4, $f12                           \r\n"
+        "psraw $f2, $f2, $f12                           \r\n"
+        "psraw $f10, $f10, $f12                         \r\n"
+        "packsswh $f0, $f0, $f2                         \r\n"
+        "packsswh $f4, $f4, $f10                        \r\n"
+        "mfc1 $9, $f0                                   \r\n"
+        "dsrl $f0, $f0, $f18                            \r\n"
+        "sh $9, 0x0(%[output])                          \r\n"
+        "mfc1 %[input], $f0                             \r\n"
+        "dsrl $9, $9, 0x10                              \r\n"
+        "sh %[input], 0x80(%[output])                   \r\n"
+        "sh $9, 0x20(%[output])                         \r\n"
+        "dsrl %[input], %[input], 0x10                  \r\n"
+        "mfc1 $9, $f4                                   \r\n"
+        "sh %[input], 0xa0(%[output])                   \r\n"
+        "dsrl $f4, $f4, $f18                            \r\n"
+        "sh $9, 0x40(%[output])                         \r\n"
+        "mfc1 %[input], $f4                             \r\n"
+        "dsrl $9, $9, 0x10                              \r\n"
+        "sh %[input], 0xc0(%[output])                   \r\n"
+        "sh $9, 0x60(%[output])                         \r\n"
+        "dsrl %[input], %[input], 0x10                  \r\n"
+        "sh %[input], 0xe0(%[output])                   \r\n"
+        "punpckhhw $f2, $f6, %[ff_pw_1]                 \r\n"
+        "punpcklhw $f6, $f6, %[ff_pw_1]                 \r\n"
+        "punpckhhw $f10, $f8, %[ff_pw_1]                \r\n"
+        "punpcklhw $f8, $f8, %[ff_pw_1]                 \r\n"
+        "mtc1 %[qmul], $f14                             \r\n"
+        "punpcklwd $f14, $f14, $f14                     \r\n"
+        "pmaddhw $f6, $f6, $f14                         \r\n"
+        "pmaddhw $f8, $f8, $f14                         \r\n"
+        "pmaddhw $f2, $f2, $f14                         \r\n"
+        "pmaddhw $f10, $f10, $f14                       \r\n"
+        "psraw $f6, $f6, $f12                           \r\n"
+        "psraw $f8, $f8, $f12                           \r\n"
+        "psraw $f2, $f2, $f12                           \r\n"
+        "psraw $f10, $f10, $f12                         \r\n"
+        "packsswh $f6, $f6, $f2                         \r\n"
+        "packsswh $f8, $f8, $f10                        \r\n"
+        "mfc1 $9, $f6                                   \r\n"
+        "dsrl $f6, $f6, $f18                            \r\n"
+        "mfc1 %[input], $f6                             \r\n"
+        "sh $9, 0x100(%[output])                        \r\n"
+        "sh %[input], 0x180(%[output])                  \r\n"
+        "dsrl $9, $9, 0x10                              \r\n"
+        "dsrl %[input], %[input], 0x10                  \r\n"
+        "sh $9, 0x120(%[output])                        \r\n"
+        "sh %[input], 0x1a0(%[output])                  \r\n"
+        "mfc1 $9, $f8                                   \r\n"
+        "dsrl $f8, $f8, $f18                            \r\n"
+        "mfc1 %[input], $f8                             \r\n"
+        "sh $9, 0x140(%[output])                        \r\n"
+        "sh %[input], 0x1c0(%[output])                  \r\n"
+        "dsrl $9, $9, 0x10                              \r\n"
+        "dsrl %[input], %[input], 0x10                  \r\n"
+        "sh $9, 0x160(%[output])                        \r\n"
+        "sh %[input], 0x1e0(%[output])                  \r\n"
+        ".set reorder                                   \r\n"
+        ::[output]"r"(output),[input]"r"(input),[qmul]"r"((uint64_t)qmul),
+          [ff_pw_1]"f"(ff_pw_1)
+        : "$9","$10","$f0","$f2","$f4","$f6","$f8","$f10","$f12","$f14","$f16",
+          "$f18"
+    );
+}
+
+void ff_h264_chroma422_dc_dequant_idct_8_mmi(int16_t *block, int qmul)
+{
+    int temp[8];
+    int t[8];
+
+    temp[0] = block[0] + block[16];
+    temp[1] = block[0] - block[16];
+    temp[2] = block[32] + block[48];
+    temp[3] = block[32] - block[48];
+    temp[4] = block[64] + block[80];
+    temp[5] = block[64] - block[80];
+    temp[6] = block[96] + block[112];
+    temp[7] = block[96] - block[112];
+
+    t[0] = temp[0] + temp[4] + temp[2] + temp[6];
+    t[1] = temp[0] - temp[4] + temp[2] - temp[6];
+    t[2] = temp[0] - temp[4] - temp[2] + temp[6];
+    t[3] = temp[0] + temp[4] - temp[2] - temp[6];
+    t[4] = temp[1] + temp[5] + temp[3] + temp[7];
+    t[5] = temp[1] - temp[5] + temp[3] - temp[7];
+    t[6] = temp[1] - temp[5] - temp[3] + temp[7];
+    t[7] = temp[1] + temp[5] - temp[3] - temp[7];
+
+    block[  0]= (t[0]*qmul + 128) >> 8;
+    block[ 32]= (t[1]*qmul + 128) >> 8;
+    block[ 64]= (t[2]*qmul + 128) >> 8;
+    block[ 96]= (t[3]*qmul + 128) >> 8;
+    block[ 16]= (t[4]*qmul + 128) >> 8;
+    block[ 48]= (t[5]*qmul + 128) >> 8;
+    block[ 80]= (t[6]*qmul + 128) >> 8;
+    block[112]= (t[7]*qmul + 128) >> 8;
+}
+
+void ff_h264_chroma_dc_dequant_idct_8_mmi(int16_t *block, int qmul)
+{
+    int a,b,c,d;
+
+    d = block[0] - block[16];
+    a = block[0] + block[16];
+    b = block[32] - block[48];
+    c = block[32] + block[48];
+    block[0] = ((a+c)*qmul) >> 7;
+    block[16]= ((d+b)*qmul) >> 7;
+    block[32]= ((a-c)*qmul) >> 7;
+    block[48]= ((d-b)*qmul) >> 7;
+}
+
 void ff_h264_weight_pixels16_8_mmi(uint8_t *block, int stride,
         int height, int log2_denom, int weight, int offset)
 {
@@ -276,3 +1282,1219 @@
         );
     }
 }
+
+static void inline chroma_inter_body_mmi(uint8_t *pix, int stride,
+        int alpha, int beta, int8_t *tc0)
+{
+    __asm__ volatile (
+        "xor $f16, $f16, $f16                           \r\n"
+        "mtc1 %[alpha], $f8                             \r\n"
+        "mtc1 %[beta], $f10                             \r\n"
+        "pshufh $f8, $f8, $f16                          \r\n"
+        "pshufh $f10, $f10, $f16                        \r\n"
+        "packushb $f8, $f8, $f8                         \r\n"
+        "packushb $f10, $f10, $f10                      \r\n"
+        "psubusb $f12, $f4, $f2                         \r\n"
+        "psubusb $f14, $f2, $f4                         \r\n"
+        "or $f14, $f14, $f12                            \r\n"
+        "psubusb $f14, $f14, $f8                        \r\n"
+        "psubusb $f12, $f2, $f0                         \r\n"
+        "psubusb $f8, $f0, $f2                          \r\n"
+        "or $f8, $f8, $f12                              \r\n"
+        "psubusb $f8, $f8, $f10                         \r\n"
+        "or $f14, $f14, $f8                             \r\n"
+        "psubusb $f12, $f4, $f6                         \r\n"
+        "psubusb $f8, $f6, $f4                          \r\n"
+        "or $f8, $f8, $f12                              \r\n"
+        "psubusb $f8, $f8, $f10                         \r\n"
+        "or $f14, $f14, $f8                             \r\n"
+        "xor $f12, $f12, $f12                           \r\n"
+        "pcmpeqb $f14, $f14, $f12                       \r\n"
+        "lwc1 $f12, 0x0(%[tc0])                         \r\n"
+        "punpcklbh $f12, $f12, $f12                     \r\n"
+        "and $f14, $f14, $f12                           \r\n"
+        "pcmpeqb $f8, $f8, $f8                          \r\n"
+        "xor $f10, $f2, $f4                             \r\n"
+        "xor $f6, $f6, $f8                              \r\n"
+        "and $f10, $f10, %[ff_pb_1]                     \r\n"
+        "pavgb $f6, $f6, $f0                            \r\n"
+        "xor $f8, $f8, $f2                              \r\n"
+        "pavgb $f6, $f6, %[ff_pb_3]                     \r\n"
+        "pavgb $f8, $f8, $f4                            \r\n"
+        "pavgb $f6, $f6, $f10                           \r\n"
+        "paddusb $f6, $f6, $f8                          \r\n"
+        "psubusb $f12, %[ff_pb_A1], $f6                 \r\n"
+        "psubusb $f6, $f6, %[ff_pb_A1]                  \r\n"
+        "pminub $f12, $f12, $f14                        \r\n"
+        "pminub $f6, $f6, $f14                          \r\n"
+        "psubusb $f2, $f2, $f12                         \r\n"
+        "psubusb $f4, $f4, $f6                          \r\n"
+        "paddusb $f2, $f2, $f6                          \r\n"
+        "paddusb $f4, $f4, $f12                         \r\n"
+        ::[pix]"r"(pix),[stride]"r"((int64_t)stride),
+          [alpha]"r"((int64_t)alpha),[beta]"r"((int64_t)beta),[tc0]"r"(tc0),
+          [ff_pb_1]"f"(ff_pb_1),[ff_pb_3]"f"(ff_pb_3),[ff_pb_A1]"f"(ff_pb_A1)
+        : "$f0","$f2","$f4","$f6","$f8","$f10","$f12","$f14","$f16"
+    );
+}
+
+static void inline chroma_intra_body_mmi(uint8_t *pix, int stride,
+        int alpha, int beta)
+{
+    __asm__ volatile (
+        "xor $f16, $f16, $f16                           \r\n"
+        "mtc1 %[alpha], $f8                             \r\n"
+        "mtc1 %[beta], $f10                             \r\n"
+        "pshufh $f8, $f8, $f16                          \r\n"
+        "pshufh $f10, $f10, $f16                        \r\n"
+        "packushb $f8, $f8, $f8                         \r\n"
+        "packushb $f10, $f10, $f10                      \r\n"
+        "psubusb $f12, $f4, $f2                         \r\n"
+        "psubusb $f14, $f2, $f4                         \r\n"
+        "or $f14, $f14, $f12                            \r\n"
+        "psubusb $f14, $f14, $f8                        \r\n"
+        "psubusb $f12, $f2, $f0                         \r\n"
+        "psubusb $f8, $f0, $f2                          \r\n"
+        "or $f8, $f8, $f12                              \r\n"
+        "psubusb $f8, $f8, $f10                         \r\n"
+        "or $f14, $f14, $f8                             \r\n"
+        "psubusb $f12, $f4, $f6                         \r\n"
+        "psubusb $f8, $f6, $f4                          \r\n"
+        "or $f8, $f8, $f12                              \r\n"
+        "psubusb $f8, $f8, $f10                         \r\n"
+        "or $f14, $f14, $f8                             \r\n"
+        "xor $f12, $f12, $f12                           \r\n"
+        "pcmpeqb $f14, $f14, $f12                       \r\n"
+        "mov.d $f10, $f2                                \r\n"
+        "mov.d $f12, $f4                                \r\n"
+        "xor $f8, $f2, $f6                              \r\n"
+        "and $f8, $f8, %[ff_pb_1]                       \r\n"
+        "pavgb $f2, $f2, $f6                            \r\n"
+        "psubusb $f2, $f2, $f8                          \r\n"
+        "pavgb $f2, $f2, $f0                            \r\n"
+        "xor $f8, $f4, $f0                              \r\n"
+        "and $f8, $f8, %[ff_pb_1]                       \r\n"
+        "pavgb $f4, $f4, $f0                            \r\n"
+        "psubusb $f4, $f4, $f8                          \r\n"
+        "pavgb $f4, $f4, $f6                            \r\n"
+        "psubb $f2, $f2, $f10                           \r\n"
+        "psubb $f4, $f4, $f12                           \r\n"
+        "and $f2, $f2, $f14                             \r\n"
+        "and $f4, $f4, $f14                             \r\n"
+        "paddb $f2, $f2, $f10                           \r\n"
+        "paddb $f4, $f4, $f12                           \r\n"
+        ::[pix]"r"(pix),[stride]"r"((int64_t)stride),
+          [alpha]"r"((int64_t)alpha),[beta]"r"((int64_t)beta),
+          [ff_pb_1]"f"(ff_pb_1)
+        : "$f0","$f2","$f4","$f8","$f10","$f12","$f14","$f16"
+    );
+}
+
+void ff_deblock_v8_luma_8_mmi(uint8_t *pix, int stride, int alpha, int beta,
+        int8_t *tc0)
+{
+    __asm__ volatile (
+        "daddu $8, %[stride], %[stride]                 \r\n"
+        "xor $f16, $f16, $f16                           \r\n"
+        "daddu $9, %[stride], $8                        \r\n"
+        "daddiu %[alpha], %[alpha], -0x1                \r\n"
+        "dsubu $9, $0, $9                               \r\n"
+        "daddiu %[beta], %[beta], -0x1                  \r\n"
+        "daddu $9, $9, %[pix]                           \r\n"
+        "ldc1 $f4, 0x0(%[pix])                          \r\n"
+        "gsldxc1 $f0, 0x0($9, %[stride])                \r\n"
+        "gsldxc1 $f2, 0x0($9, $8)                       \r\n"
+        "gsldxc1 $f6, 0x0(%[pix], %[stride])            \r\n"
+        "mtc1 %[alpha], $f8                             \r\n"
+        "mtc1 %[beta], $f10                             \r\n"
+        "pshufh $f8, $f8, $f16                          \r\n"
+        "pshufh $f10, $f10, $f16                        \r\n"
+        "packushb $f8, $f8, $f8                         \r\n"
+        "packushb $f10, $f10, $f10                      \r\n"
+        "psubusb $f12, $f4, $f2                         \r\n"
+        "psubusb $f14, $f2, $f4                         \r\n"
+        "or $f14, $f14, $f12                            \r\n"
+        "psubusb $f12, $f2, $f0                         \r\n"
+        "psubusb $f14, $f14, $f8                        \r\n"
+        "psubusb $f8, $f0, $f2                          \r\n"
+        "or $f8, $f8, $f12                              \r\n"
+        "psubusb $f12, $f4, $f6                         \r\n"
+        "psubusb $f8, $f8, $f10                         \r\n"
+        "or $f14, $f14, $f8                             \r\n"
+        "psubusb $f8, $f6, $f4                          \r\n"
+        "or $f8, $f8, $f12                              \r\n"
+        "psubusb $f8, $f8, $f10                         \r\n"
+        "or $f14, $f14, $f8                             \r\n"
+        "pcmpeqb $f14, $f14, $f16                       \r\n"
+        "pcmpeqb $f6, $f6, $f6                          \r\n"
+        "gslwlc1 $f8, 0x3(%[tc0])                       \r\n"
+        "gslwrc1 $f8, 0x0(%[tc0])                       \r\n"
+        "punpcklbh $f8, $f8, $f8                        \r\n"
+        "punpcklbh $f18, $f8, $f8                       \r\n"
+        "pcmpgtb $f8, $f18, $f6                         \r\n"
+        "ldc1 $f6, 0x0($9)                              \r\n"
+        "and $f20, $f8, $f14                            \r\n"
+        "psubusb $f14, $f6, $f2                         \r\n"
+        "psubusb $f12, $f2, $f6                         \r\n"
+        "psubusb $f14, $f14, $f10                       \r\n"
+        "psubusb $f12, $f12, $f10                       \r\n"
+        "pcmpeqb $f12, $f12, $f14                       \r\n"
+        "and $f12, $f12, $f20                           \r\n"
+        "and $f8, $f20, $f18                            \r\n"
+        "psubb $f14, $f8, $f12                          \r\n"
+        "and $f12, $f12, $f8                            \r\n"
+        "pavgb $f8, $f2, $f4                            \r\n"
+        "ldc1 $f22, 0x0($9)                             \r\n"
+        "pavgb $f6, $f6, $f8                            \r\n"
+        "xor $f8, $f8, $f22                             \r\n"
+        "and $f8, $f8, %[ff_pb_1]                       \r\n"
+        "psubusb $f6, $f6, $f8                          \r\n"
+        "psubusb $f8, $f0, $f12                         \r\n"
+        "paddusb $f12, $f12, $f0                        \r\n"
+        "pmaxub $f6, $f6, $f8                           \r\n"
+        "pminub $f6, $f6, $f12                          \r\n"
+        "gssdxc1 $f6, 0x0($9, %[stride])                \r\n"
+        "gsldxc1 $f8, 0x0(%[pix], $8)                   \r\n"
+        "psubusb $f6, $f8, $f4                          \r\n"
+        "psubusb $f12, $f4, $f8                         \r\n"
+        "psubusb $f6, $f6, $f10                         \r\n"
+        "psubusb $f12, $f12, $f10                       \r\n"
+        "pcmpeqb $f12, $f12, $f6                        \r\n"
+        "and $f12, $f12, $f20                           \r\n"
+        "psubb $f14, $f14, $f12                         \r\n"
+        "and $f10, $f18, $f12                           \r\n"
+        "gsldxc1 $f6, 0x0(%[pix], %[stride])            \r\n"
+        "pavgb $f12, $f2, $f4                           \r\n"
+        "gsldxc1 $f22, 0x0(%[pix], $8)                  \r\n"
+        "pavgb $f8, $f8, $f12                           \r\n"
+        "xor $f12, $f12, $f22                           \r\n"
+        "and $f12, $f12, %[ff_pb_1]                     \r\n"
+        "psubusb $f8, $f8, $f12                         \r\n"
+        "psubusb $f12, $f6, $f10                        \r\n"
+        "paddusb $f10, $f10, $f6                        \r\n"
+        "pmaxub $f8, $f8, $f12                          \r\n"
+        "pminub $f8, $f8, $f10                          \r\n"
+        "gssdxc1 $f8, 0x0(%[pix], %[stride])            \r\n"
+        "xor $f10, $f2, $f4                             \r\n"
+        "pcmpeqb $f8, $f8, $f8                          \r\n"
+        "and $f10, $f10, %[ff_pb_1]                     \r\n"
+        "xor $f6, $f6, $f8                              \r\n"
+        "xor $f8, $f8, $f2                              \r\n"
+        "pavgb $f6, $f6, $f0                            \r\n"
+        "pavgb $f6, $f6, %[ff_pb_3]                     \r\n"
+        "pavgb $f8, $f8, $f4                            \r\n"
+        "pavgb $f6, $f6, $f10                           \r\n"
+        "paddusb $f6, $f6, $f8                          \r\n"
+        "psubusb $f12, %[ff_pb_A1], $f6                 \r\n"
+        "psubusb $f6, $f6, %[ff_pb_A1]                  \r\n"
+        "pminub $f12, $f12, $f14                        \r\n"
+        "pminub $f6, $f6, $f14                          \r\n"
+        "psubusb $f2, $f2, $f12                         \r\n"
+        "psubusb $f4, $f4, $f6                          \r\n"
+        "paddusb $f2, $f2, $f6                          \r\n"
+        "paddusb $f4, $f4, $f12                         \r\n"
+        "gssdxc1 $f2, 0x0($9, $8)                       \r\n"
+        "sdc1 $f4, 0x0(%[pix])                          \r\n"
+        ::[pix]"r"(pix),[stride]"r"((int64_t)stride),
+          [alpha]"r"((int64_t)alpha),[beta]"r"((int64_t)beta),[tc0]"r"(tc0),
+          [ff_pb_1]"f"(ff_pb_1),[ff_pb_3]"f"(ff_pb_3),[ff_pb_A1]"f"(ff_pb_A1)
+        : "$8","$9","$f0","$f2","$f4","$f6","$f8","$f10","$f12","$f14","$f16",
+          "$f18","$f20","$f22"
+    );
+}
+
+void ff_deblock_v8_luma_intra_8_mmi(uint8_t *pix, int stride, int alpha,
+        int beta)
+{
+    uint64_t stack[0xa];
+
+    __asm__ volatile (
+        "ori $8, $0, 0x1                                \r\n"
+        "xor $f30, $f30, $f30                           \r\n"
+        "dmtc1 $8, $f16                                 \r\n"
+        "dsll $8, %[stride], 2                          \r\n"
+        "daddu $10, %[stride], %[stride]                \r\n"
+        "daddiu %[alpha], %[alpha], -0x1                \r\n"
+        "dsll $f20, $f16, $f16                          \r\n"
+        "bltz %[alpha], 1f                              \r\n"
+        "daddu $9, $10, %[stride]                       \r\n"
+        "daddiu %[beta], %[beta], -0x1                  \r\n"
+        "bltz %[beta], 1f                               \r\n"
+        "dsubu $8, $0, $8                               \r\n"
+        "daddu $8, $8, %[pix]                           \r\n"
+        "ldc1 $f4, 0x0(%[pix])                          \r\n"
+        "gsldxc1 $f0, 0x0($8, $10)                      \r\n"
+        "gsldxc1 $f2, 0x0($8, $9)                       \r\n"
+        "gsldxc1 $f6, 0x0(%[pix], %[stride])            \r\n"
+        "mtc1 %[alpha], $f8                             \r\n"
+        "mtc1 %[beta], $f10                             \r\n"
+        "pshufh $f8, $f8, $f30                          \r\n"
+        "pshufh $f10, $f10, $f30                        \r\n"
+        "packushb $f8, $f8, $f8                         \r\n"
+        "psubusb $f12, $f4, $f2                         \r\n"
+        "psubusb $f14, $f2, $f4                         \r\n"
+        "packushb $f10, $f10, $f10                      \r\n"
+        "or $f14, $f14, $f12                            \r\n"
+        "sdc1 $f8, 0x10+%[stack]                        \r\n"
+        "psubusb $f14, $f14, $f8                        \r\n"
+        "psubusb $f12, $f2, $f0                         \r\n"
+        "psubusb $f8, $f0, $f2                          \r\n"
+        "or $f8, $f8, $f12                              \r\n"
+        "psubusb $f8, $f8, $f10                         \r\n"
+        "or $f14, $f14, $f8                             \r\n"
+        "psubusb $f12, $f4, $f6                         \r\n"
+        "psubusb $f8, $f6, $f4                          \r\n"
+        "or $f8, $f8, $f12                              \r\n"
+        "psubusb $f8, $f8, $f10                         \r\n"
+        "or $f14, $f14, $f8                             \r\n"
+        "xor $f12, $f12, $f12                           \r\n"
+        "ldc1 $f8, 0x10+%[stack]                        \r\n"
+        "pcmpeqb $f14, $f14, $f12                       \r\n"
+        "sdc1 $f14, 0x20+%[stack]                       \r\n"
+        "pavgb $f8, $f8, $f30                           \r\n"
+        "psubusb $f14, $f4, $f2                         \r\n"
+        "pavgb $f8, $f8, %[ff_pb_1]                     \r\n"
+        "psubusb $f12, $f2, $f4                         \r\n"
+        "psubusb $f14, $f14, $f8                        \r\n"
+        "psubusb $f12, $f12, $f8                        \r\n"
+        "ldc1 $f28, 0x20+%[stack]                       \r\n"
+        "pcmpeqb $f12, $f12, $f14                       \r\n"
+        "and $f12, $f12, $f28                           \r\n"
+        "gsldxc1 $f28, 0x0($8, %[stride])               \r\n"
+        "psubusb $f14, $f28, $f2                        \r\n"
+        "psubusb $f8, $f2, $f28                         \r\n"
+        "psubusb $f14, $f14, $f10                       \r\n"
+        "psubusb $f8, $f8, $f10                         \r\n"
+        "pcmpeqb $f8, $f8, $f14                         \r\n"
+        "and $f8, $f8, $f12                             \r\n"
+        "gsldxc1 $f26, 0x0(%[pix], $10)                 \r\n"
+        "sdc1 $f8, 0x30+%[stack]                        \r\n"
+        "psubusb $f14, $f26, $f4                        \r\n"
+        "psubusb $f8, $f4, $f26                         \r\n"
+        "psubusb $f14, $f14, $f10                       \r\n"
+        "psubusb $f8, $f8, $f10                         \r\n"
+        "pcmpeqb $f8, $f8, $f14                         \r\n"
+        "and $f8, $f8, $f12                             \r\n"
+        "sdc1 $f8, 0x40+%[stack]                        \r\n"
+        "pavgb $f8, $f28, $f0                           \r\n"
+        "pavgb $f10, $f2, $f4                           \r\n"
+        "pavgb $f8, $f8, $f10                           \r\n"
+        "sdc1 $f10, 0x10+%[stack]                       \r\n"
+        "paddb $f12, $f28, $f0                          \r\n"
+        "paddb $f14, $f2, $f4                           \r\n"
+        "paddb $f12, $f12, $f14                         \r\n"
+        "mov.d $f14, $f12                               \r\n"
+        "sdc1 $f12, 0x0+%[stack]                        \r\n"
+        "psrlh $f12, $f12, $f16                         \r\n"
+        "pavgb $f12, $f12, $f30                         \r\n"
+        "xor $f12, $f12, $f8                            \r\n"
+        "and $f12, $f12, %[ff_pb_1]                     \r\n"
+        "psubb $f8, $f8, $f12                           \r\n"
+        "pavgb $f10, $f28, $f6                          \r\n"
+        "psubb $f12, $f28, $f6                          \r\n"
+        "paddb $f14, $f14, $f14                         \r\n"
+        "psubb $f14, $f14, $f12                         \r\n"
+        "and $f12, $f12, %[ff_pb_1]                     \r\n"
+        "psubb $f10, $f10, $f12                         \r\n"
+        "ldc1 $f24, 0x10+%[stack]                       \r\n"
+        "pavgb $f10, $f10, $f0                          \r\n"
+        "psrlh $f14, $f14, $f20                         \r\n"
+        "pavgb $f10, $f10, $f24                         \r\n"
+        "pavgb $f14, $f14, $f30                         \r\n"
+        "xor $f14, $f14, $f10                           \r\n"
+        "and $f14, $f14, %[ff_pb_1]                     \r\n"
+        "psubb $f10, $f10, $f14                         \r\n"
+        "xor $f14, $f2, $f6                             \r\n"
+        "pavgb $f12, $f2, $f6                           \r\n"
+        "and $f14, $f14, %[ff_pb_1]                     \r\n"
+        "psubb $f12, $f12, $f14                         \r\n"
+        "ldc1 $f24, 0x30+%[stack]                       \r\n"
+        "pavgb $f12, $f12, $f0                          \r\n"
+        "ldc1 $f22, 0x20+%[stack]                       \r\n"
+        "xor $f10, $f10, $f12                           \r\n"
+        "xor $f12, $f12, $f2                            \r\n"
+        "and $f10, $f10, $f24                           \r\n"
+        "and $f12, $f12, $f22                           \r\n"
+        "xor $f10, $f10, $f12                           \r\n"
+        "xor $f10, $f10, $f2                            \r\n"
+        "gssdxc1 $f10, 0x0($8, $9)                      \r\n"
+        "ldc1 $f10, 0x0($8)                             \r\n"
+        "paddb $f12, $f28, $f10                         \r\n"
+        "pavgb $f10, $f10, $f28                         \r\n"
+        "ldc1 $f22, 0x0+%[stack]                        \r\n"
+        "pavgb $f10, $f10, $f8                          \r\n"
+        "paddb $f12, $f12, $f12                         \r\n"
+        "paddb $f12, $f12, $f22                         \r\n"
+        "psrlh $f12, $f12, $f20                         \r\n"
+        "pavgb $f12, $f12, $f30                         \r\n"
+        "xor $f12, $f12, $f10                           \r\n"
+        "and $f12, $f12, %[ff_pb_1]                     \r\n"
+        "ldc1 $f22, 0x30+%[stack]                       \r\n"
+        "psubb $f10, $f10, $f12                         \r\n"
+        "xor $f8, $f8, $f0                              \r\n"
+        "xor $f10, $f10, $f28                           \r\n"
+        "and $f8, $f8, $f22                             \r\n"
+        "and $f10, $f10, $f22                           \r\n"
+        "xor $f8, $f8, $f0                              \r\n"
+        "xor $f10, $f10, $f28                           \r\n"
+        "gssdxc1 $f8, 0x0($8, $10)                      \r\n"
+        "gssdxc1 $f10, 0x0($8, %[stride])               \r\n"
+        "pavgb $f8, $f26, $f6                           \r\n"
+        "pavgb $f10, $f4, $f2                           \r\n"
+        "pavgb $f8, $f8, $f10                           \r\n"
+        "sdc1 $f10, 0x10+%[stack]                       \r\n"
+        "paddb $f12, $f26, $f6                          \r\n"
+        "paddb $f14, $f4, $f2                           \r\n"
+        "paddb $f12, $f12, $f14                         \r\n"
+        "mov.d $f14, $f12                               \r\n"
+        "sdc1 $f12, 0x0+%[stack]                        \r\n"
+        "psrlh $f12, $f12, $f16                         \r\n"
+        "pavgb $f12, $f12, $f30                         \r\n"
+        "xor $f12, $f12, $f8                            \r\n"
+        "and $f12, $f12, %[ff_pb_1]                     \r\n"
+        "psubb $f8, $f8, $f12                           \r\n"
+        "pavgb $f10, $f26, $f0                          \r\n"
+        "paddb $f14, $f14, $f14                         \r\n"
+        "psubb $f12, $f26, $f0                          \r\n"
+        "psubb $f14, $f14, $f12                         \r\n"
+        "and $f12, $f12, %[ff_pb_1]                     \r\n"
+        "psubb $f10, $f10, $f12                         \r\n"
+        "ldc1 $f22, 0x10+%[stack]                       \r\n"
+        "pavgb $f10, $f10, $f6                          \r\n"
+        "pavgb $f10, $f10, $f22                         \r\n"
+        "psrlh $f14, $f14, $f20                         \r\n"
+        "pavgb $f14, $f14, $f30                         \r\n"
+        "xor $f14, $f14, $f10                           \r\n"
+        "and $f14, $f14, %[ff_pb_1]                     \r\n"
+        "psubb $f10, $f10, $f14                         \r\n"
+        "xor $f14, $f4, $f0                             \r\n"
+        "pavgb $f12, $f4, $f0                           \r\n"
+        "and $f14, $f14, %[ff_pb_1]                     \r\n"
+        "ldc1 $f22, 0x40+%[stack]                       \r\n"
+        "psubb $f12, $f12, $f14                         \r\n"
+        "ldc1 $f24, 0x20+%[stack]                       \r\n"
+        "pavgb $f12, $f12, $f6                          \r\n"
+        "xor $f10, $f10, $f12                           \r\n"
+        "xor $f12, $f12, $f4                            \r\n"
+        "and $f10, $f10, $f22                           \r\n"
+        "and $f12, $f12, $f24                           \r\n"
+        "xor $f10, $f10, $f12                           \r\n"
+        "xor $f10, $f10, $f4                            \r\n"
+        "sdc1 $f10, 0x0(%[pix])                         \r\n"
+        "gsldxc1 $f10, 0x0(%[pix], $9)                  \r\n"
+        "paddb $f12, $f26, $f10                         \r\n"
+        "pavgb $f10, $f10, $f26                         \r\n"
+        "ldc1 $f22, 0x0+%[stack]                        \r\n"
+        "pavgb $f10, $f10, $f8                          \r\n"
+        "paddb $f12, $f12, $f12                         \r\n"
+        "paddb $f12, $f12, $f22                         \r\n"
+        "psrlh $f12, $f12, $f20                         \r\n"
+        "pavgb $f12, $f12, $f30                         \r\n"
+        "xor $f12, $f12, $f10                           \r\n"
+        "and $f12, $f12, %[ff_pb_1]                     \r\n"
+        "ldc1 $f22, 0x40+%[stack]                       \r\n"
+        "psubb $f10, $f10, $f12                         \r\n"
+        "xor $f8, $f8, $f6                              \r\n"
+        "xor $f10, $f10, $f26                           \r\n"
+        "and $f8, $f8, $f22                             \r\n"
+        "and $f10, $f10, $f22                           \r\n"
+        "xor $f8, $f8, $f6                              \r\n"
+        "xor $f10, $f10, $f26                           \r\n"
+        "gssdxc1 $f8, 0x0(%[pix], %[stride])            \r\n"
+        "gssdxc1 $f10, 0x0(%[pix], $10)                 \r\n"
+        "1:                                             \r\n"
+        ::[pix]"r"(pix),[stride]"r"((int64_t)stride),
+          [alpha]"r"((int64_t)alpha),[beta]"r"((int64_t)beta),
+          [stack]"m"(stack[0]),[ff_pb_1]"f"(ff_pb_1)
+        : "$8","$9","$10","$f0","$f2","$f4","$f6","$f8","$f10","$f12","$f14",
+          "$f16","$f18","$f20","$f22","$f24","$f26","$f28","$f30"
+    );
+}
+
+void ff_deblock_v_chroma_8_mmi(uint8_t *pix, int stride, int alpha, int beta,
+        int8_t *tc0)
+{
+    __asm__ volatile (
+        "daddiu %[alpha], %[alpha], -0x1                \r\n"
+        "daddiu %[beta], %[beta], -0x1                  \r\n"
+        "or $16, $0, %[pix]                             \r\n"
+        "dsubu $16, $16, %[stride]                      \r\n"
+        "dsubu $16, $16, %[stride]                      \r\n"
+        "ldc1 $f0, 0x0($16)                             \r\n"
+        "gsldxc1 $f2, 0x0($16, %[stride])               \r\n"
+        "ldc1 $f4, 0x0(%[pix])                          \r\n"
+        "gsldxc1 $f6, 0x0(%[pix], %[stride])            \r\n"
+        : [pix]"+r"(pix),[stride]"+r"(stride),[alpha]"+r"(alpha),
+          [beta]"+r"(beta)
+        : [tc0]"r"(tc0)
+        : "$16","$f2","$f4"
+    );
+
+    chroma_inter_body_mmi(pix, stride, alpha, beta, tc0);
+
+    __asm__ volatile (
+        "gssdxc1 $f2, 0x0($16, %[stride])               \r\n"
+        "sdc1 $f4, 0x0(%[pix])                          \r\n"
+        ::[pix]"r"(pix),[stride]"r"((int64_t)stride)
+        : "$16","$f2","$f4"
+    );
+}
+
+void ff_deblock_v_chroma_intra_8_mmi(uint8_t *pix, int stride, int alpha,
+        int beta)
+{
+    __asm__ volatile (
+        "daddiu %[alpha], %[alpha], -0x1                \r\n"
+        "daddiu %[beta], %[beta], -0x1                  \r\n"
+        "or $16, $0, %[pix]                             \r\n"
+        "dsubu $16, $16, %[stride]                      \r\n"
+        "dsubu $16, $16, %[stride]                      \r\n"
+        "ldc1 $f0, 0x0($16)                             \r\n"
+        "gsldxc1 $f2, 0x0($16, %[stride])               \r\n"
+        "ldc1 $f4, 0x0(%[pix])                          \r\n"
+        "gsldxc1 $f6, 0x0(%[pix], %[stride])            \r\n"
+        : [pix]"+r"(pix),[stride]"+r"(stride),[alpha]"+r"(alpha),
+          [beta]"+r"(beta)
+        ::"$16","$f0","$f2","$f4","$f6"
+    );
+
+    chroma_intra_body_mmi(pix, stride, alpha, beta);
+
+    __asm__ volatile (
+        "gssdxc1 $f2, 0x0($16, %[stride])               \r\n"
+        "sdc1 $f4, 0x0(%[pix])                          \r\n"
+        ::[pix]"r"(pix),[stride]"r"((int64_t)stride)
+        : "$16","$f2","$f4"
+    );
+}
+
+void ff_deblock_h_chroma_8_mmi(uint8_t *pix, int stride, int alpha, int beta,
+        int8_t *tc0)
+{
+    __asm__ volatile (
+        "daddiu %[alpha], %[alpha], -0x1                \r\n"
+        "daddiu %[beta], %[beta], -0x1                  \r\n"
+        "daddu $16, %[stride], %[stride]                \r\n"
+        "daddiu %[pix], %[pix], -0x2                    \r\n"
+        "daddu $17, $16, %[stride]                      \r\n"
+        "daddu $19, $16, $16                            \r\n"
+        "or $18, $0, %[pix]                             \r\n"
+        "daddu %[pix], %[pix], $17                      \r\n"
+        "gslwlc1 $f0, 0x3($18)                          \r\n"
+        "daddu $12, $18, %[stride]                      \r\n"
+        "gslwrc1 $f0, 0x0($18)                          \r\n"
+        "gslwlc1 $f4, 0x3($12)                          \r\n"
+        "daddu $13, $18, $16                            \r\n"
+        "gslwrc1 $f4, 0x0($12)                          \r\n"
+        "gslwlc1 $f2, 0x3($13)                          \r\n"
+        "gslwrc1 $f2, 0x0($13)                          \r\n"
+        "gslwlc1 $f6, 0x3(%[pix])                       \r\n"
+        "gslwrc1 $f6, 0x0(%[pix])                       \r\n"
+        "punpcklbh $f0, $f0, $f4                        \r\n"
+        "punpcklbh $f2, $f2, $f6                        \r\n"
+        "daddu $12, %[pix], %[stride]                   \r\n"
+        "punpckhhw $f4, $f0, $f2                        \r\n"
+        "punpcklhw $f0, $f0, $f2                        \r\n"
+        "gslwlc1 $f8, 0x3($12)                          \r\n"
+        "daddu $13, %[pix], $16                         \r\n"
+        "gslwrc1 $f8, 0x0($12)                          \r\n"
+        "gslwlc1 $f12, 0x3($13)                         \r\n"
+        "daddu $12, %[pix], $17                         \r\n"
+        "gslwrc1 $f12, 0x0($13)                         \r\n"
+        "gslwlc1 $f10, 0x3($12)                         \r\n"
+        "daddu $13, %[pix], $19                         \r\n"
+        "gslwrc1 $f10, 0x0($12)                         \r\n"
+        "gslwlc1 $f14, 0x3($13)                         \r\n"
+        "gslwrc1 $f14, 0x0($13)                         \r\n"
+        "punpcklbh $f8, $f8, $f12                       \r\n"
+        "punpcklbh $f10, $f10, $f14                     \r\n"
+        "mov.d $f12, $f8                                \r\n"
+        "punpcklhw $f8, $f8, $f10                       \r\n"
+        "punpckhhw $f12, $f12, $f10                     \r\n"
+        "punpckhwd $f2, $f0, $f8                        \r\n"
+        "punpckhwd $f6, $f4, $f12                       \r\n"
+        "punpcklwd $f0, $f0, $f8                        \r\n"
+        "punpcklwd $f4, $f4, $f12                       \r\n"
+        "mov.d $f20, $f0                                \r\n"
+        "mov.d $f22, $f6                                \r\n"
+        : [pix]"+r"(pix),[stride]"+r"(stride),[alpha]"+r"(alpha),
+          [beta]"+r"(beta)
+        ::"$12","$13","$16","$17","$18","$19","$f0","$f2","$f4","$f6","$f8",
+          "$f10","$f12","$f14","$f20","$f22"
+    );
+
+    chroma_inter_body_mmi(pix, stride, alpha, beta, tc0);
+
+    __asm__ volatile (
+        "punpckhwd $f8, $f20, $f20                      \r\n"
+        "punpckhwd $f10, $f2, $f2                       \r\n"
+        "punpckhwd $f12, $f4, $f4                       \r\n"
+        "punpcklbh $f0, $f20, $f2                       \r\n"
+        "punpcklbh $f4, $f4, $f22                       \r\n"
+        "punpcklhw $f2, $f0, $f4                        \r\n"
+        "punpckhhw $f0, $f0, $f4                        \r\n"
+        "gsswlc1 $f2, 0x3($18)                          \r\n"
+        "gsswrc1 $f2, 0x0($18)                          \r\n"
+        "daddu $12, $18, %[stride]                      \r\n"
+        "punpckhwd $f2, $f2, $f2                        \r\n"
+        "gsswlc1 $f2, 0x3($12)                          \r\n"
+        "daddu $13, $18, $16                            \r\n"
+        "gsswrc1 $f2, 0x0($12)                          \r\n"
+        "gsswlc1 $f0, 0x3($13)                          \r\n"
+        "gsswrc1 $f0, 0x0($13)                          \r\n"
+        "punpckhwd $f0, $f0, $f0                        \r\n"
+        "punpckhwd $f6, $f22, $f22                      \r\n"
+        "gsswlc1 $f0, 0x3(%[pix])                       \r\n"
+        "gsswrc1 $f0, 0x0(%[pix])                       \r\n"
+        "punpcklbh $f8, $f8, $f10                       \r\n"
+        "punpcklbh $f12, $f12, $f6                      \r\n"
+        "daddu $12, %[pix], %[stride]                   \r\n"
+        "punpcklhw $f10, $f8, $f12                      \r\n"
+        "punpckhhw $f8, $f8, $f12                       \r\n"
+        "gsswlc1 $f10, 0x3($12)                         \r\n"
+        "gsswrc1 $f10, 0x0($12)                         \r\n"
+        "punpckhwd $f10, $f10, $f10                     \r\n"
+        "daddu $12, %[pix], $16                         \r\n"
+        "daddu $13, %[pix], $17                         \r\n"
+        "gsswlc1 $f10, 0x3($12)                         \r\n"
+        "gsswrc1 $f10, 0x0($12)                         \r\n"
+        "gsswlc1 $f8, 0x3($13)                          \r\n"
+        "daddu $12, %[pix], $19                         \r\n"
+        "punpckhwd $f20, $f8, $f8                       \r\n"
+        "gsswrc1 $f8, 0x0($13)                          \r\n"
+        "gsswlc1 $f20, 0x3($12)                         \r\n"
+        "gsswrc1 $f20, 0x0($12)                         \r\n"
+        ::[pix]"r"(pix),[stride]"r"((int64_t)stride)
+        : "$12","$13","$16","$17","$18","$19","$f0","$f2","$f4","$f6","$f8",
+          "$f10","$f12","$f20"
+    );
+}
+
+void ff_deblock_h_chroma_intra_8_mmi(uint8_t *pix, int stride, int alpha,
+        int beta)
+{
+    __asm__ volatile (
+        "daddiu %[alpha], %[alpha], -0x1                \r\n"
+        "daddiu %[beta], %[beta], -0x1                  \r\n"
+        "daddu $16, %[stride], %[stride]                \r\n"
+        "daddiu %[pix], %[pix], -0x2                    \r\n"
+        "daddu $17, $16, %[stride]                      \r\n"
+        "daddu $19, $16, $16                            \r\n"
+        "or $18, $0, %[pix]                             \r\n"
+        "daddu %[pix], %[pix], $17                      \r\n"
+        "gslwlc1 $f0, 0x3($18)                          \r\n"
+        "daddu $12, $18, %[stride]                      \r\n"
+        "gslwrc1 $f0, 0x0($18)                          \r\n"
+        "gslwlc1 $f4, 0x3($12)                          \r\n"
+        "daddu $13, $18, $16                            \r\n"
+        "gslwrc1 $f4, 0x0($12)                          \r\n"
+        "gslwlc1 $f2, 0x3($13)                          \r\n"
+        "gslwrc1 $f2, 0x0($13)                          \r\n"
+        "gslwlc1 $f6, 0x3(%[pix])                       \r\n"
+        "gslwrc1 $f6, 0x0(%[pix])                       \r\n"
+        "punpcklbh $f0, $f0, $f4                        \r\n"
+        "punpcklbh $f2, $f2, $f6                        \r\n"
+        "daddu $12, %[pix], %[stride]                   \r\n"
+        "punpckhhw $f4, $f0, $f2                        \r\n"
+        "punpcklhw $f0, $f0, $f2                        \r\n"
+        "gslwlc1 $f8, 0x3($12)                          \r\n"
+        "daddu $13, %[pix], $16                         \r\n"
+        "gslwrc1 $f8, 0x0($12)                          \r\n"
+        "gslwlc1 $f12, 0x3($13)                         \r\n"
+        "daddu $12, %[pix], $17                         \r\n"
+        "gslwrc1 $f12, 0x0($13)                         \r\n"
+        "gslwlc1 $f10, 0x3($12)                         \r\n"
+        "daddu $13, %[pix], $19                         \r\n"
+        "gslwrc1 $f10, 0x0($12)                         \r\n"
+        "gslwlc1 $f14, 0x3($13)                         \r\n"
+        "gslwrc1 $f14, 0x0($13)                         \r\n"
+        "punpcklbh $f8, $f8, $f12                       \r\n"
+        "punpcklbh $f10, $f10, $f14                     \r\n"
+        "mov.d $f12, $f8                                \r\n"
+        "punpcklhw $f8, $f8, $f10                       \r\n"
+        "punpckhhw $f12, $f12, $f10                     \r\n"
+        "punpckhwd $f2, $f0, $f8                        \r\n"
+        "punpckhwd $f6, $f4, $f12                       \r\n"
+        "punpcklwd $f0, $f0, $f8                        \r\n"
+        "punpcklwd $f4, $f4, $f12                       \r\n"
+        : [pix]"+r"(pix),[stride]"+r"(stride),[alpha]"+r"(alpha),
+          [beta]"+r"(beta)
+        ::"$12","$13","$16","$17","$18","$19","$f0","$f2","$f4","$f6","$f8",
+          "$f10","$f12","$f14","$f20","$f22"
+    );
+
+    chroma_intra_body_mmi(pix, stride, alpha, beta);
+
+    __asm__ volatile (
+        "punpckhwd $f8, $f0, $f0                        \r\n"
+        "punpckhwd $f10, $f2, $f2                       \r\n"
+        "punpckhwd $f12, $f4, $f4                       \r\n"
+        "punpcklbh $f0, $f0, $f2                        \r\n"
+        "punpcklbh $f4, $f4, $f6                        \r\n"
+        "punpcklhw $f2, $f0, $f4                        \r\n"
+        "punpckhhw $f0, $f0, $f4                        \r\n"
+        "gsswlc1 $f2, 0x3($18)                          \r\n"
+        "gsswrc1 $f2, 0x0($18)                          \r\n"
+        "daddu $12, $18, %[stride]                      \r\n"
+        "punpckhwd $f2, $f2, $f2                        \r\n"
+        "gsswlc1 $f2, 0x3($12)                          \r\n"
+        "daddu $13, $18, $16                            \r\n"
+        "gsswrc1 $f2, 0x0($12)                          \r\n"
+        "gsswlc1 $f0, 0x3($13)                          \r\n"
+        "gsswrc1 $f0, 0x0($13)                          \r\n"
+        "punpckhwd $f0, $f0, $f0                        \r\n"
+        "punpckhwd $f6, $f6, $f6                        \r\n"
+        "gsswlc1 $f0, 0x3(%[pix])                       \r\n"
+        "gsswrc1 $f0, 0x0(%[pix])                       \r\n"
+        "punpcklbh $f8, $f8, $f10                       \r\n"
+        "punpcklbh $f12, $f12, $f6                      \r\n"
+        "daddu $12, %[pix], %[stride]                   \r\n"
+        "punpcklhw $f10, $f8, $f12                      \r\n"
+        "punpckhhw $f8, $f8, $f12                       \r\n"
+        "gsswlc1 $f10, 0x3($12)                         \r\n"
+        "gsswrc1 $f10, 0x0($12)                         \r\n"
+        "punpckhwd $f10, $f10, $f10                     \r\n"
+        "daddu $12, %[pix], $16                         \r\n"
+        "daddu $13, %[pix], $17                         \r\n"
+        "gsswlc1 $f10, 0x3($12)                         \r\n"
+        "gsswrc1 $f10, 0x0($12)                         \r\n"
+        "gsswlc1 $f8, 0x3($13)                          \r\n"
+        "daddu $12, %[pix], $19                         \r\n"
+        "punpckhwd $f20, $f8, $f8                       \r\n"
+        "gsswrc1 $f8, 0x0($13)                          \r\n"
+        "gsswlc1 $f20, 0x3($12)                         \r\n"
+        "gsswrc1 $f20, 0x0($12)                         \r\n"
+        ::[pix]"r"(pix),[stride]"r"((int64_t)stride)
+        : "$12","$13","$16","$17","$18","$19","$f0","$f2","$f4","$f6","$f8",
+          "$f10","$f12","$f20"
+    );
+}
+
+void ff_deblock_v_luma_8_mmi(uint8_t *pix, int stride, int alpha, int beta,
+        int8_t *tc0)
+{
+    if ((tc0[0] & tc0[1]) >= 0)
+        ff_deblock_v8_luma_8_mmi(pix + 0, stride, alpha, beta, tc0);
+    if ((tc0[2] & tc0[3]) >= 0)
+        ff_deblock_v8_luma_8_mmi(pix + 8, stride, alpha, beta, tc0 + 2);
+}
+
+void ff_deblock_v_luma_intra_8_mmi(uint8_t *pix, int stride, int alpha,
+        int beta)
+{
+    ff_deblock_v8_luma_intra_8_mmi(pix + 0, stride, alpha, beta);
+    ff_deblock_v8_luma_intra_8_mmi(pix + 8, stride, alpha, beta);
+}
+
+void ff_deblock_h_luma_8_mmi(uint8_t *pix, int stride, int alpha, int beta,
+        int8_t *tc0)
+{
+    uint64_t stack[0xd];
+
+    __asm__ volatile (
+        "daddu $15, %[stride], %[stride]                \r\n"
+        "daddiu $8, %[pix], -0x4                        \r\n"
+        "daddu $9, %[stride], $15                       \r\n"
+        "gsldlc1 $f0, 0x7($8)                           \r\n"
+        "gsldrc1 $f0, 0x0($8)                           \r\n"
+        "daddu $12, $8, %[stride]                       \r\n"
+        "daddu $10, $8, $9                              \r\n"
+        "gsldlc1 $f2, 0x7($12)                          \r\n"
+        "daddu $11, $8, $15                             \r\n"
+        "gsldrc1 $f2, 0x0($12)                          \r\n"
+        "gsldlc1 $f4, 0x7($11)                          \r\n"
+        "gsldrc1 $f4, 0x0($11)                          \r\n"
+        "gsldlc1 $f6, 0x7($10)                          \r\n"
+        "daddu $12, $10, %[stride]                      \r\n"
+        "gsldrc1 $f6, 0x0($10)                          \r\n"
+        "gsldlc1 $f8, 0x7($12)                          \r\n"
+        "daddu $11, $10, $15                            \r\n"
+        "gsldrc1 $f8, 0x0($12)                          \r\n"
+        "gsldlc1 $f10, 0x7($11)                         \r\n"
+        "daddu $12, $10, $9                             \r\n"
+        "gsldrc1 $f10, 0x0($11)                         \r\n"
+        "gsldlc1 $f12, 0x7($12)                         \r\n"
+        "gsldrc1 $f12, 0x0($12)                         \r\n"
+        "daddu $14, $15, $15                            \r\n"
+        "punpckhbh $f14, $f0, $f2                       \r\n"
+        "punpcklbh $f0, $f0, $f2                        \r\n"
+        "punpckhbh $f2, $f4, $f6                        \r\n"
+        "punpcklbh $f4, $f4, $f6                        \r\n"
+        "punpckhbh $f6, $f8, $f10                       \r\n"
+        "punpcklbh $f8, $f8, $f10                       \r\n"
+        "daddu $12, $10, $14                            \r\n"
+        "sdc1 $f2, 0x10+%[stack]                        \r\n"
+        "gsldlc1 $f16, 0x7($12)                         \r\n"
+        "gsldrc1 $f16, 0x0($12)                         \r\n"
+        "daddu $13, $14, $14                            \r\n"
+        "punpckhbh $f10, $f12, $f16                     \r\n"
+        "punpcklbh $f12, $f12, $f16                     \r\n"
+        "punpckhhw $f2, $f0, $f4                        \r\n"
+        "punpcklhw $f0, $f0, $f4                        \r\n"
+        "punpckhhw $f4, $f8, $f12                       \r\n"
+        "punpcklhw $f8, $f8, $f12                       \r\n"
+        "ldc1 $f16, 0x10+%[stack]                       \r\n"
+        "punpckhwd $f0, $f0, $f8                        \r\n"
+        "sdc1 $f0, 0x0+%[stack]                         \r\n"
+        "punpckhhw $f12, $f14, $f16                     \r\n"
+        "punpcklhw $f14, $f14, $f16                     \r\n"
+        "punpckhhw $f0, $f6, $f10                       \r\n"
+        "punpcklhw $f6, $f6, $f10                       \r\n"
+        "punpcklwd $f12, $f12, $f0                      \r\n"
+        "punpckhwd $f10, $f14, $f6                      \r\n"
+        "punpcklwd $f14, $f14, $f6                      \r\n"
+        "punpckhwd $f6, $f2, $f4                        \r\n"
+        "punpcklwd $f2, $f2, $f4                        \r\n"
+        "sdc1 $f2, 0x10+%[stack]                        \r\n"
+        "sdc1 $f6, 0x20+%[stack]                        \r\n"
+        "sdc1 $f14, 0x30+%[stack]                       \r\n"
+        "sdc1 $f10, 0x40+%[stack]                       \r\n"
+        "sdc1 $f12, 0x50+%[stack]                       \r\n"
+        "daddu $8, $8, $13                              \r\n"
+        "daddu $10, $10, $13                            \r\n"
+        "gsldlc1 $f0, 0x7($8)                           \r\n"
+        "daddu $12, $8, %[stride]                       \r\n"
+        "gsldrc1 $f0, 0x0($8)                           \r\n"
+        "gsldlc1 $f2, 0x7($12)                          \r\n"
+        "daddu $11, $8, $15                             \r\n"
+        "gsldrc1 $f2, 0x0($12)                          \r\n"
+        "gsldlc1 $f4, 0x7($11)                          \r\n"
+        "gsldrc1 $f4, 0x0($11)                          \r\n"
+        "gsldlc1 $f6, 0x7($10)                          \r\n"
+        "daddu $12, $10, %[stride]                      \r\n"
+        "gsldrc1 $f6, 0x0($10)                          \r\n"
+        "gsldlc1 $f8, 0x7($12)                          \r\n"
+        "daddu $11, $10, $15                            \r\n"
+        "gsldrc1 $f8, 0x0($12)                          \r\n"
+        "gsldlc1 $f10, 0x7($11)                         \r\n"
+        "daddu $12, $10, $9                             \r\n"
+        "gsldrc1 $f10, 0x0($11)                         \r\n"
+        "gsldlc1 $f12, 0x7($12)                         \r\n"
+        "gsldrc1 $f12, 0x0($12)                         \r\n"
+        "punpckhbh $f14, $f0, $f2                       \r\n"
+        "punpcklbh $f0, $f0, $f2                        \r\n"
+        "punpckhbh $f2, $f4, $f6                        \r\n"
+        "punpcklbh $f4, $f4, $f6                        \r\n"
+        "punpckhbh $f6, $f8, $f10                       \r\n"
+        "punpcklbh $f8, $f8, $f10                       \r\n"
+        "daddu $12, $10, $14                            \r\n"
+        "sdc1 $f2, 0x18+%[stack]                        \r\n"
+        "gsldlc1 $f16, 0x7($12)                         \r\n"
+        "gsldrc1 $f16, 0x0($12)                         \r\n"
+        "punpckhhw $f2, $f0, $f4                        \r\n"
+        "punpckhbh $f10, $f12, $f16                     \r\n"
+        "punpcklbh $f12, $f12, $f16                     \r\n"
+        "punpcklhw $f0, $f0, $f4                        \r\n"
+        "punpckhhw $f4, $f8, $f12                       \r\n"
+        "punpcklhw $f8, $f8, $f12                       \r\n"
+        "punpckhwd $f0, $f0, $f8                        \r\n"
+        "ldc1 $f16, 0x18+%[stack]                       \r\n"
+        "sdc1 $f0, 0x8+%[stack]                         \r\n"
+        "punpckhhw $f12, $f14, $f16                     \r\n"
+        "punpcklhw $f14, $f14, $f16                     \r\n"
+        "punpckhhw $f0, $f6, $f10                       \r\n"
+        "punpcklhw $f6, $f6, $f10                       \r\n"
+        "punpckhwd $f10, $f14, $f6                      \r\n"
+        "punpcklwd $f14, $f14, $f6                      \r\n"
+        "punpckhwd $f6, $f2, $f4                        \r\n"
+        "punpcklwd $f2, $f2, $f4                        \r\n"
+        "punpcklwd $f12, $f12, $f0                      \r\n"
+        "sdc1 $f2, 0x18+%[stack]                        \r\n"
+        "sdc1 $f6, 0x28+%[stack]                        \r\n"
+        "sdc1 $f14, 0x38+%[stack]                       \r\n"
+        "sdc1 $f10, 0x48+%[stack]                       \r\n"
+        "sdc1 $f12, 0x58+%[stack]                       \r\n"
+        ::[pix]"r"(pix),[stride]"r"((int64_t)stride),[stack]"m"(stack[0])
+        : "$8","$9","$10","$11","$12","$13","$14","$15","$f0","$f2","$f4",
+          "$f6","$f8","$f10","$f12","$f14","$f16"
+    );
+
+    ff_deblock_v_luma_8_mmi((uint8_t *) &stack[6], 0x10, alpha, beta, tc0);
+
+    __asm__ volatile (
+        "daddu $15, %[stride], %[stride]                \r\n"
+        "daddiu $8, %[pix], -0x2                        \r\n"
+        "daddu $14, $15, $15                            \r\n"
+        "daddu $9, $15, %[stride]                       \r\n"
+        "daddu $13, $14, $14                            \r\n"
+        "daddu $10, $8, $9                              \r\n"
+        "ldc1 $f0, 0x10+%[stack]                        \r\n"
+        "ldc1 $f2, 0x20+%[stack]                        \r\n"
+        "ldc1 $f4, 0x30+%[stack]                        \r\n"
+        "ldc1 $f6, 0x40+%[stack]                        \r\n"
+        "punpckhwd $f8, $f0, $f0                        \r\n"
+        "punpckhwd $f10, $f2, $f2                       \r\n"
+        "punpckhwd $f12, $f4, $f4                       \r\n"
+        "punpcklbh $f0, $f0, $f2                        \r\n"
+        "punpcklbh $f4, $f4, $f6                        \r\n"
+        "punpcklhw $f2, $f0, $f4                        \r\n"
+        "punpckhhw $f0, $f0, $f4                        \r\n"
+        "gsswlc1 $f2, 0x3($8)                           \r\n"
+        "gsswrc1 $f2, 0x0($8)                           \r\n"
+        "daddu $12, $8, %[stride]                       \r\n"
+        "punpckhwd $f2, $f2, $f2                        \r\n"
+        "daddu $11, $8, $15                             \r\n"
+        "gsswlc1 $f2, 0x3($12)                          \r\n"
+        "gsswrc1 $f2, 0x0($12)                          \r\n"
+        "gsswlc1 $f0, 0x3($11)                          \r\n"
+        "gsswrc1 $f0, 0x0($11)                          \r\n"
+        "punpckhwd $f0, $f0, $f0                        \r\n"
+        "punpckhwd $f6, $f6, $f6                        \r\n"
+        "gsswlc1 $f0, 0x3($10)                          \r\n"
+        "gsswrc1 $f0, 0x0($10)                          \r\n"
+        "punpcklbh $f8, $f8, $f10                       \r\n"
+        "punpcklbh $f12, $f12, $f6                      \r\n"
+        "punpcklhw $f10, $f8, $f12                      \r\n"
+        "daddu $12, $10, %[stride]                      \r\n"
+        "punpckhhw $f8, $f8, $f12                       \r\n"
+        "gsswlc1 $f10, 0x3($12)                         \r\n"
+        "gsswrc1 $f10, 0x0($12)                         \r\n"
+        "daddu $12, $10, $15                            \r\n"
+        "punpckhwd $f10, $f10, $f10                     \r\n"
+        "daddu $11, $10, $9                             \r\n"
+        "gsswlc1 $f10, 0x3($12)                         \r\n"
+        "gsswrc1 $f10, 0x0($12)                         \r\n"
+        "gsswlc1 $f8, 0x3($11)                          \r\n"
+        "gsswrc1 $f8, 0x0($11)                          \r\n"
+        "daddu $12, $10, $14                            \r\n"
+        "punpckhwd $f8, $f8, $f8                        \r\n"
+        "daddu $8, $8, $13                              \r\n"
+        "gsswlc1 $f8, 0x3($12)                          \r\n"
+        "gsswrc1 $f8, 0x0($12)                          \r\n"
+        "daddu $10, $10, $13                            \r\n"
+        "ldc1 $f0, 0x18+%[stack]                        \r\n"
+        "ldc1 $f2, 0x28+%[stack]                        \r\n"
+        "ldc1 $f4, 0x38+%[stack]                        \r\n"
+        "ldc1 $f6, 0x48+%[stack]                        \r\n"
+        "daddu $15, %[stride], %[stride]                \r\n"
+        "punpckhwd $f8, $f0, $f0                        \r\n"
+        "daddu $14, $15, $15                            \r\n"
+        "punpckhwd $f10, $f2, $f2                       \r\n"
+        "punpckhwd $f12, $f4, $f4                       \r\n"
+        "punpcklbh $f0, $f0, $f2                        \r\n"
+        "punpcklbh $f4, $f4, $f6                        \r\n"
+        "daddu $12, $8, %[stride]                       \r\n"
+        "punpcklhw $f2, $f0, $f4                        \r\n"
+        "punpckhhw $f0, $f0, $f4                        \r\n"
+        "gsswlc1 $f2, 0x3($8)                           \r\n"
+        "gsswrc1 $f2, 0x0($8)                           \r\n"
+        "punpckhwd $f2, $f2, $f2                        \r\n"
+        "daddu $11, $8, $15                             \r\n"
+        "gsswlc1 $f2, 0x3($12)                          \r\n"
+        "gsswrc1 $f2, 0x0($12)                          \r\n"
+        "gsswlc1 $f0, 0x3($11)                          \r\n"
+        "gsswrc1 $f0, 0x0($11)                          \r\n"
+        "punpckhwd $f0, $f0, $f0                        \r\n"
+        "punpckhwd $f6, $f6, $f6                        \r\n"
+        "gsswlc1 $f0, 0x3($10)                          \r\n"
+        "gsswrc1 $f0, 0x0($10)                          \r\n"
+        "punpcklbh $f8, $f8, $f10                       \r\n"
+        "punpcklbh $f12, $f12, $f6                      \r\n"
+        "daddu $12, $10, %[stride]                      \r\n"
+        "punpcklhw $f10, $f8, $f12                      \r\n"
+        "punpckhhw $f8, $f8, $f12                       \r\n"
+        "gsswlc1 $f10, 0x3($12)                         \r\n"
+        "gsswrc1 $f10, 0x0($12)                         \r\n"
+        "daddu $12, $10, $15                            \r\n"
+        "punpckhwd $f10, $f10, $f10                     \r\n"
+        "daddu $11, $10, $9                             \r\n"
+        "gsswlc1 $f10, 0x3($12)                         \r\n"
+        "gsswrc1 $f10, 0x0($12)                         \r\n"
+        "gsswlc1 $f8, 0x3($11)                          \r\n"
+        "gsswrc1 $f8, 0x0($11)                          \r\n"
+        "daddu $12, $10, $14                            \r\n"
+        "punpckhwd $f8, $f8, $f8                        \r\n"
+        "gsswlc1 $f8, 0x3($12)                          \r\n"
+        "gsswrc1 $f8, 0x0($12)                          \r\n"
+        ::[pix]"r"(pix),[stride]"r"((int64_t)stride),[stack]"m"(stack[0])
+        : "$8","$9","$10","$11","$12","$13","$14","$15","$f0","$f2","$f4",
+          "$f6","$f8","$f10","$f12","$f14","$f16"
+    );
+}
+
+void ff_deblock_h_luma_intra_8_mmi(uint8_t *pix, int stride, int alpha,
+        int beta)
+{
+    uint64_t ptmp[0x11];
+    uint64_t pdat[4];
+
+    __asm__ volatile (
+        "daddu $12, %[stride], %[stride]                \r\n"
+        "daddiu $10, %[pix], -0x4                       \r\n"
+        "daddu $11, $12, %[stride]                      \r\n"
+        "daddu $13, $12, $12                            \r\n"
+        "daddu $9, $10, $11                             \r\n"
+        "daddu $8, $10, %[stride]                       \r\n"
+        "gsldlc1 $f0, 0x7($10)                          \r\n"
+        "gsldrc1 $f0, 0x0($10)                          \r\n"
+        "daddu $14, $10, $12                            \r\n"
+        "gsldlc1 $f2, 0x7($8)                           \r\n"
+        "gsldrc1 $f2, 0x0($8)                           \r\n"
+        "gsldlc1 $f4, 0x7($14)                          \r\n"
+        "gsldrc1 $f4, 0x0($14)                          \r\n"
+        "daddu $8, $9, %[stride]                        \r\n"
+        "gsldlc1 $f6, 0x7($9)                           \r\n"
+        "gsldrc1 $f6, 0x0($9)                           \r\n"
+        "daddu $14, $9, $12                             \r\n"
+        "gsldlc1 $f8, 0x7($8)                           \r\n"
+        "gsldrc1 $f8, 0x0($8)                           \r\n"
+        "daddu $8, $9, $11                              \r\n"
+        "gsldlc1 $f10, 0x7($14)                         \r\n"
+        "gsldrc1 $f10, 0x0($14)                         \r\n"
+        "gsldlc1 $f12, 0x7($8)                          \r\n"
+        "gsldrc1 $f12, 0x0($8)                          \r\n"
+        "daddu $8, $9, $13                              \r\n"
+        "punpckhbh $f14, $f0, $f2                       \r\n"
+        "punpcklbh $f0, $f0, $f2                        \r\n"
+        "punpckhbh $f2, $f4, $f6                        \r\n"
+        "punpcklbh $f4, $f4, $f6                        \r\n"
+        "punpckhbh $f6, $f8, $f10                       \r\n"
+        "punpcklbh $f8, $f8, $f10                       \r\n"
+        "gsldlc1 $f16, 0x7($8)                          \r\n"
+        "gsldrc1 $f16, 0x0($8)                          \r\n"
+        "punpckhbh $f10, $f12, $f16                     \r\n"
+        "punpcklbh $f12, $f12, $f16                     \r\n"
+        "sdc1 $f6, 0x0+%[ptmp]                          \r\n"
+        "punpckhhw $f6, $f0, $f4                        \r\n"
+        "punpcklhw $f0, $f0, $f4                        \r\n"
+        "punpckhhw $f4, $f8, $f12                       \r\n"
+        "punpcklhw $f8, $f8, $f12                       \r\n"
+        "punpckhhw $f12, $f14, $f2                      \r\n"
+        "punpcklhw $f14, $f14, $f2                      \r\n"
+        "sdc1 $f4, 0x20+%[ptmp]                         \r\n"
+        "ldc1 $f4, 0x0+%[ptmp]                          \r\n"
+        "punpckhhw $f2, $f4, $f10                       \r\n"
+        "punpcklhw $f4, $f4, $f10                       \r\n"
+        "punpckhwd $f10, $f0, $f8                       \r\n"
+        "punpcklwd $f0, $f0, $f8                        \r\n"
+        "punpckhwd $f8, $f14, $f4                       \r\n"
+        "punpcklwd $f14, $f14, $f4                      \r\n"
+        "sdc1 $f0, 0x0+%[ptmp]                          \r\n"
+        "sdc1 $f10, 0x10+%[ptmp]                        \r\n"
+        "sdc1 $f14, 0x40+%[ptmp]                        \r\n"
+        "sdc1 $f8, 0x50+%[ptmp]                         \r\n"
+        "ldc1 $f16, 0x20+%[ptmp]                        \r\n"
+        "punpckhwd $f0, $f6, $f16                       \r\n"
+        "punpcklwd $f6, $f6, $f16                       \r\n"
+        "punpckhwd $f10, $f12, $f2                      \r\n"
+        "punpcklwd $f12, $f12, $f2                      \r\n"
+        "daddu $8, $13, $13                             \r\n"
+        "sdc1 $f6, 0x20+%[ptmp]                         \r\n"
+        "sdc1 $f0, 0x30+%[ptmp]                         \r\n"
+        "sdc1 $f12, 0x60+%[ptmp]                        \r\n"
+        "sdc1 $f10, 0x70+%[ptmp]                        \r\n"
+        "daddu $10, $10, $8                             \r\n"
+        "daddu $9, $9, $8                               \r\n"
+        "daddu $8, $10, %[stride]                       \r\n"
+        "gsldlc1 $f0, 0x7($10)                          \r\n"
+        "gsldrc1 $f0, 0x0($10)                          \r\n"
+        "daddu $14, $10, $12                            \r\n"
+        "gsldlc1 $f2, 0x7($8)                           \r\n"
+        "gsldrc1 $f2, 0x0($8)                           \r\n"
+        "gsldlc1 $f4, 0x7($14)                          \r\n"
+        "gsldrc1 $f4, 0x0($14)                          \r\n"
+        "daddu $8, $9, %[stride]                        \r\n"
+        "gsldlc1 $f6, 0x7($9)                           \r\n"
+        "gsldrc1 $f6, 0x0($9)                           \r\n"
+        "daddu $14, $9, $12                             \r\n"
+        "gsldlc1 $f8, 0x7($8)                           \r\n"
+        "gsldrc1 $f8, 0x0($8)                           \r\n"
+        "daddu $8, $9, $11                              \r\n"
+        "gsldlc1 $f10, 0x7($14)                         \r\n"
+        "gsldrc1 $f10, 0x0($14)                         \r\n"
+        "gsldlc1 $f12, 0x7($8)                          \r\n"
+        "gsldrc1 $f12, 0x0($8)                          \r\n"
+        "daddu $8, $9, $13                              \r\n"
+        "punpckhbh $f14, $f0, $f2                       \r\n"
+        "punpcklbh $f0, $f0, $f2                        \r\n"
+        "punpckhbh $f2, $f4, $f6                        \r\n"
+        "punpcklbh $f4, $f4, $f6                        \r\n"
+        "punpckhbh $f6, $f8, $f10                       \r\n"
+        "punpcklbh $f8, $f8, $f10                       \r\n"
+        "gsldlc1 $f16, 0x7($8)                          \r\n"
+        "gsldrc1 $f16, 0x0($8)                          \r\n"
+        "punpckhbh $f10, $f12, $f16                     \r\n"
+        "punpcklbh $f12, $f12, $f16                     \r\n"
+        "sdc1 $f6, 0x8+%[ptmp]                          \r\n"
+        "punpckhhw $f6, $f0, $f4                        \r\n"
+        "punpcklhw $f0, $f0, $f4                        \r\n"
+        "punpckhhw $f4, $f8, $f12                       \r\n"
+        "punpcklhw $f8, $f8, $f12                       \r\n"
+        "punpckhhw $f12, $f14, $f2                      \r\n"
+        "punpcklhw $f14, $f14, $f2                      \r\n"
+        "sdc1 $f4, 0x28+%[ptmp]                         \r\n"
+        "ldc1 $f4, 0x8+%[ptmp]                          \r\n"
+        "punpckhhw $f2, $f4, $f10                       \r\n"
+        "punpcklhw $f4, $f4, $f10                       \r\n"
+        "punpckhwd $f10, $f0, $f8                       \r\n"
+        "punpcklwd $f0, $f0, $f8                        \r\n"
+        "punpckhwd $f8, $f14, $f4                       \r\n"
+        "punpcklwd $f14, $f14, $f4                      \r\n"
+        "sdc1 $f0, 0x8+%[ptmp]                          \r\n"
+        "sdc1 $f10, 0x18+%[ptmp]                        \r\n"
+        "sdc1 $f14, 0x48+%[ptmp]                        \r\n"
+        "sdc1 $f8, 0x58+%[ptmp]                         \r\n"
+        "ldc1 $f16, 0x28+%[ptmp]                        \r\n"
+        "punpckhwd $f0, $f6, $f16                       \r\n"
+        "punpcklwd $f6, $f6, $f16                       \r\n"
+        "punpckhwd $f10, $f12, $f2                      \r\n"
+        "punpcklwd $f12, $f12, $f2                      \r\n"
+        "sdc1 $f6, 0x28+%[ptmp]                         \r\n"
+        "sdc1 $f0, 0x38+%[ptmp]                         \r\n"
+        "sdc1 $f12, 0x68+%[ptmp]                        \r\n"
+        "sdc1 $f10, 0x78+%[ptmp]                        \r\n"
+        "sd $10, 0x00+%[pdat]                           \r\n"
+        "sd $11, 0x08+%[pdat]                           \r\n"
+        "sd $12, 0x10+%[pdat]                           \r\n"
+        "sd $13, 0x18+%[pdat]                           \r\n"
+        ::[pix]"r"(pix),[stride]"r"((uint64_t)stride),[ptmp]"m"(ptmp[0]),
+          [pdat]"m"(pdat[0])
+        : "$8","$9","$10","$11","$12","$13","$14","$f0","$f2","$f4","$f6",
+          "$f8","$f10","$f12","$f14","$f16"
+    );
+
+    ff_deblock_v_luma_intra_8_mmi((uint8_t *) &ptmp[8], 0x10, alpha, beta);
+
+    __asm__ volatile (
+        "ld $10, 0x00+%[pdat]                           \r\n"
+        "ld $11, 0x08+%[pdat]                           \r\n"
+        "ld $12, 0x10+%[pdat]                           \r\n"
+        "ld $13, 0x18+%[pdat]                           \r\n"
+        "daddu $9, $10, $11                             \r\n"
+        "ldc1 $f0, 0x8+%[ptmp]                          \r\n"
+        "ldc1 $f2, 0x18+%[ptmp]                         \r\n"
+        "ldc1 $f4, 0x28+%[ptmp]                         \r\n"
+        "ldc1 $f6, 0x38+%[ptmp]                         \r\n"
+        "ldc1 $f8, 0x48+%[ptmp]                         \r\n"
+        "ldc1 $f10, 0x58+%[ptmp]                        \r\n"
+        "ldc1 $f12, 0x68+%[ptmp]                        \r\n"
+        "punpckhbh $f14, $f0, $f2                       \r\n"
+        "punpcklbh $f0, $f0, $f2                        \r\n"
+        "punpckhbh $f2, $f4, $f6                        \r\n"
+        "punpcklbh $f4, $f4, $f6                        \r\n"
+        "punpckhbh $f6, $f8, $f10                       \r\n"
+        "punpcklbh $f8, $f8, $f10                       \r\n"
+        "ldc1 $f16, 0x78+%[ptmp]                        \r\n"
+        "punpckhbh $f10, $f12, $f16                     \r\n"
+        "punpcklbh $f12, $f12, $f16                     \r\n"
+        "gssdlc1 $f6, 0x7($10)                          \r\n"
+        "gssdrc1 $f6, 0x0($10)                          \r\n"
+        "daddu $8, $10, $12                             \r\n"
+        "punpckhhw $f6, $f0, $f4                        \r\n"
+        "punpcklhw $f0, $f0, $f4                        \r\n"
+        "punpckhhw $f4, $f8, $f12                       \r\n"
+        "punpcklhw $f8, $f8, $f12                       \r\n"
+        "punpckhhw $f12, $f14, $f2                      \r\n"
+        "punpcklhw $f14, $f14, $f2                      \r\n"
+        "gssdlc1 $f4, 0x7($8)                           \r\n"
+        "gssdrc1 $f4, 0x0($8)                           \r\n"
+        "gsldlc1 $f4, 0x7($10)                          \r\n"
+        "gsldrc1 $f4, 0x0($10)                          \r\n"
+        "punpckhhw $f2, $f4, $f10                       \r\n"
+        "punpcklhw $f4, $f4, $f10                       \r\n"
+        "punpckhwd $f10, $f0, $f8                       \r\n"
+        "punpcklwd $f0, $f0, $f8                        \r\n"
+        "punpckhwd $f8, $f14, $f4                       \r\n"
+        "punpcklwd $f14, $f14, $f4                      \r\n"
+        "daddu $8, $10, %[stride]                       \r\n"
+        "gssdlc1 $f0, 0x7($10)                          \r\n"
+        "gssdrc1 $f0, 0x0($10)                          \r\n"
+        "daddu $14, $9, %[stride]                       \r\n"
+        "gssdlc1 $f10, 0x7($8)                          \r\n"
+        "gssdrc1 $f10, 0x0($8)                          \r\n"
+        "daddu $8, $9, $12                              \r\n"
+        "gssdlc1 $f14, 0x7($14)                         \r\n"
+        "gssdrc1 $f14, 0x0($14)                         \r\n"
+        "daddu $14, $10, $12                            \r\n"
+        "gssdlc1 $f8, 0x7($8)                           \r\n"
+        "gssdrc1 $f8, 0x0($8)                           \r\n"
+        "gsldlc1 $f16, 0x7($14)                         \r\n"
+        "gsldrc1 $f16, 0x0($14)                         \r\n"
+        "daddu $8, $10, $12                             \r\n"
+        "punpckhwd $f0, $f6, $f16                       \r\n"
+        "punpcklwd $f6, $f6, $f16                       \r\n"
+        "punpckhwd $f10, $f12, $f2                      \r\n"
+        "punpcklwd $f12, $f12, $f2                      \r\n"
+        "gssdlc1 $f6, 0x7($8)                           \r\n"
+        "gssdrc1 $f6, 0x0($8)                           \r\n"
+        "daddu $8, $9, $11                              \r\n"
+        "gssdlc1 $f0, 0x7($9)                           \r\n"
+        "gssdrc1 $f0, 0x0($9)                           \r\n"
+        "daddu $14, $9, $13                             \r\n"
+        "gssdlc1 $f12, 0x7($8)                          \r\n"
+        "gssdrc1 $f12, 0x0($8)                          \r\n"
+        "daddu $8, $13, $13                             \r\n"
+        "gssdlc1 $f10, 0x7($14)                         \r\n"
+        "gssdrc1 $f10, 0x0($14)                         \r\n"
+        "dsubu $10, $10, $8                             \r\n"
+        "dsubu $9, $9, $8                               \r\n"
+        "ldc1 $f0, 0x0+%[ptmp]                          \r\n"
+        "ldc1 $f2, 0x10+%[ptmp]                         \r\n"
+        "ldc1 $f4, 0x20+%[ptmp]                         \r\n"
+        "ldc1 $f6, 0x30+%[ptmp]                         \r\n"
+        "ldc1 $f8, 0x40+%[ptmp]                         \r\n"
+        "ldc1 $f10, 0x50+%[ptmp]                        \r\n"
+        "ldc1 $f12, 0x60+%[ptmp]                        \r\n"
+        "punpckhbh $f14, $f0, $f2                       \r\n"
+        "punpcklbh $f0, $f0, $f2                        \r\n"
+        "punpckhbh $f2, $f4, $f6                        \r\n"
+        "punpcklbh $f4, $f4, $f6                        \r\n"
+        "punpckhbh $f6, $f8, $f10                       \r\n"
+        "punpcklbh $f8, $f8, $f10                       \r\n"
+        "ldc1 $f16, 0x70+%[ptmp]                        \r\n"
+        "punpckhbh $f10, $f12, $f16                     \r\n"
+        "punpcklbh $f12, $f12, $f16                     \r\n"
+        "gssdlc1 $f6, 0x7($10)                          \r\n"
+        "gssdrc1 $f6, 0x0($10)                          \r\n"
+        "daddu $8, $10, $12                             \r\n"
+        "punpckhhw $f6, $f0, $f4                        \r\n"
+        "punpcklhw $f0, $f0, $f4                        \r\n"
+        "punpckhhw $f4, $f8, $f12                       \r\n"
+        "punpcklhw $f8, $f8, $f12                       \r\n"
+        "punpckhhw $f12, $f14, $f2                      \r\n"
+        "punpcklhw $f14, $f14, $f2                      \r\n"
+        "gssdlc1 $f4, 0x7($8)                           \r\n"
+        "gssdrc1 $f4, 0x0($8)                           \r\n"
+        "gsldlc1 $f4, 0x7($10)                          \r\n"
+        "gsldrc1 $f4, 0x0($10)                          \r\n"
+        "punpckhhw $f2, $f4, $f10                       \r\n"
+        "punpcklhw $f4, $f4, $f10                       \r\n"
+        "punpckhwd $f10, $f0, $f8                       \r\n"
+        "punpcklwd $f0, $f0, $f8                        \r\n"
+        "punpckhwd $f8, $f14, $f4                       \r\n"
+        "punpcklwd $f14, $f14, $f4                      \r\n"
+        "daddu $8, $10, %[stride]                       \r\n"
+        "gssdlc1 $f0, 0x7($10)                          \r\n"
+        "gssdrc1 $f0, 0x0($10)                          \r\n"
+        "daddu $14, $9, %[stride]                       \r\n"
+        "gssdlc1 $f10, 0x7($8)                          \r\n"
+        "gssdrc1 $f10, 0x0($8)                          \r\n"
+        "daddu $8, $9, $12                              \r\n"
+        "gssdlc1 $f14, 0x7($14)                         \r\n"
+        "gssdrc1 $f14, 0x0($14)                         \r\n"
+        "daddu $14, $10, $12                            \r\n"
+        "gssdlc1 $f8, 0x7($8)                           \r\n"
+        "gssdrc1 $f8, 0x0($8)                           \r\n"
+        "gsldlc1 $f16, 0x7($14)                         \r\n"
+        "gsldrc1 $f16, 0x0($14)                         \r\n"
+        "daddu $8, $10, $12                             \r\n"
+        "punpckhwd $f0, $f6, $f16                       \r\n"
+        "punpcklwd $f6, $f6, $f16                       \r\n"
+        "punpckhwd $f10, $f12, $f2                      \r\n"
+        "punpcklwd $f12, $f12, $f2                      \r\n"
+        "gssdlc1 $f6, 0x7($8)                           \r\n"
+        "gssdrc1 $f6, 0x0($8)                           \r\n"
+        "daddu $8, $9, $11                              \r\n"
+        "gssdlc1 $f0, 0x7($9)                           \r\n"
+        "gssdrc1 $f0, 0x0($9)                           \r\n"
+        "daddu $14, $9, $13                             \r\n"
+        "gssdlc1 $f12, 0x7($8)                          \r\n"
+        "gssdrc1 $f12, 0x0($8)                          \r\n"
+        "gssdlc1 $f10, 0x7($14)                         \r\n"
+        "gssdrc1 $f10, 0x0($14)                         \r\n"
+        ::[pix]"r"(pix),[stride]"r"((uint64_t)stride),[ptmp]"m"(ptmp[0]),
+          [pdat]"m"(pdat[0])
+        : "$8","$9","$10","$11","$12","$13","$14","$f0","$f2","$f4","$f6",
+          "$f8","$f10","$f12","$f14","$f16"
+    );
+}

ffmpeg-2.8.1.tar.bz2/libavcodec/mjpegdec.c -> ffmpeg-2.8.3.tar.bz2/libavcodec/mjpegdec.c Changed

@@ -98,6 +98,15 @@
         av_log(s->avctx, AV_LOG_INFO, "AVID: len:%d %d\n", len, len > 14 ? buf[12] : -1);
 }
 
+static void init_idct(AVCodecContext *avctx)
+{
+    MJpegDecodeContext *s = avctx->priv_data;
+
+    ff_idctdsp_init(&s->idsp, avctx);
+    ff_init_scantable(s->idsp.idct_permutation, &s->scantable,
+                      ff_zigzag_direct);
+}
+
 av_cold int ff_mjpeg_decode_init(AVCodecContext *avctx)
 {
     MJpegDecodeContext *s = avctx->priv_data;
@@ -112,9 +121,7 @@
     s->avctx = avctx;
     ff_blockdsp_init(&s->bdsp, avctx);
     ff_hpeldsp_init(&s->hdsp, avctx->flags);
-    ff_idctdsp_init(&s->idsp, avctx);
-    ff_init_scantable(s->idsp.idct_permutation, &s->scantable,
-                      ff_zigzag_direct);
+    init_idct(avctx);
     s->buffer_size   = 0;
     s->buffer        = NULL;
     s->start_code    = -1;
@@ -267,7 +274,6 @@
 
     /* XXX: verify len field validity */
     len     = get_bits(&s->gb, 16);
-    s->avctx->bits_per_raw_sample =
     bits    = get_bits(&s->gb, 8);
 
     if (bits > 16 || bits < 1) {
@@ -275,6 +281,11 @@
         return AVERROR_INVALIDDATA;
     }
 
+    if (s->avctx->bits_per_raw_sample != bits) {
+        av_log(s->avctx, AV_LOG_INFO, "Changeing bps to %d\n", bits);
+        s->avctx->bits_per_raw_sample = bits;
+        init_idct(s->avctx);
+    }
     if (s->pegasus_rct)
         bits = 9;
     if (bits == 9 && !s->pegasus_rct)
@@ -1093,7 +1104,10 @@
                         dc = mjpeg_decode_dc(s, s->dc_index[i]);
                         if(dc == 0xFFFFF)
                             return -1;
-                        if(bits<=8){
+                        if (   h * mb_x + x >= s->width
+                            || v * mb_y + y >= s->height) {
+                            // Nothing to do
+                        } else if (bits<=8) {
                         ptr = s->picture_ptr->data[c] + (linesize * (v * mb_y + y)) + (h * mb_x + x); //FIXME optimize this crap
                         if(y==0 && toprow){
                             if(x==0 && leftcol){
@@ -1161,7 +1175,10 @@
                         dc = mjpeg_decode_dc(s, s->dc_index[i]);
                         if(dc == 0xFFFFF)
                             return -1;
-                        if(bits<=8){
+                        if (   h * mb_x + x >= s->width
+                            || v * mb_y + y >= s->height) {
+                            // Nothing to do
+                        } else if (bits<=8) {
                             ptr = s->picture_ptr->data[c] +
                               (linesize * (v * mb_y + y)) +
                               (h * mb_x + x); //FIXME optimize this crap

ffmpeg-2.8.1.tar.bz2/libavcodec/mpeg12dec.c -> ffmpeg-2.8.3.tar.bz2/libavcodec/mpeg12dec.c Changed

ffmpeg-2.8.1.tar.bz2/libavcodec/mpegvideo.c -> ffmpeg-2.8.3.tar.bz2/libavcodec/mpegvideo.c Changed

ffmpeg-2.8.1.tar.bz2/libavcodec/opusdec.c -> ffmpeg-2.8.3.tar.bz2/libavcodec/opusdec.c Changed

ffmpeg-2.8.1.tar.bz2/libavcodec/pngdec.c -> ffmpeg-2.8.3.tar.bz2/libavcodec/pngdec.c Changed

@@ -663,7 +663,10 @@
                 break;
 
             default:
-                av_assert0(0);
+                avpriv_request_sample(avctx, "bit depth %d "
+                        "and color type %d with TRNS",
+                        s->bit_depth, s->color_type);
+                return AVERROR_INVALIDDATA;
             }
 
             s->bpp += byte_depth;
@@ -1012,7 +1015,7 @@
             for (x = s->x_offset; x < s->x_offset + s->cur_w; ++x, foreground += s->bpp, background += s->bpp) {
                 size_t b;
                 uint8_t foreground_alpha, background_alpha, output_alpha;
-                uint8_t output[4];
+                uint8_t output[10];
 
                 // Since we might be blending alpha onto alpha, we use the following equations:
                 // output_alpha = foreground_alpha + (1 - foreground_alpha) * background_alpha
@@ -1052,6 +1055,8 @@
 
                 output_alpha = foreground_alpha + FAST_DIV255((255 - foreground_alpha) * background_alpha);
 
+                av_assert0(s->bpp <= 10);
+
                 for (b = 0; b < s->bpp - 1; ++b) {
                     if (output_alpha == 0) {
                         output[b] = 0;

ffmpeg-2.8.1.tar.bz2/libavcodec/sbr.h -> ffmpeg-2.8.3.tar.bz2/libavcodec/sbr.h Changed

ffmpeg-2.8.1.tar.bz2/libavcodec/smacker.c -> ffmpeg-2.8.3.tar.bz2/libavcodec/smacker.c Changed

ffmpeg-2.8.1.tar.bz2/libavcodec/takdec.c -> ffmpeg-2.8.3.tar.bz2/libavcodec/takdec.c Changed

ffmpeg-2.8.1.tar.bz2/libavcodec/utils.c -> ffmpeg-2.8.3.tar.bz2/libavcodec/utils.c Changed

ffmpeg-2.8.1.tar.bz2/libavcodec/vp8.c -> ffmpeg-2.8.3.tar.bz2/libavcodec/vp8.c Changed

ffmpeg-2.8.1.tar.bz2/libavcodec/wmaprodec.c -> ffmpeg-2.8.3.tar.bz2/libavcodec/wmaprodec.c Changed

ffmpeg-2.8.1.tar.bz2/libavcodec/x86/videodsp.asm -> ffmpeg-2.8.3.tar.bz2/libavcodec/x86/videodsp.asm Changed

ffmpeg-2.8.1.tar.bz2/libavformat/brstm.c -> ffmpeg-2.8.3.tar.bz2/libavformat/brstm.c Changed

ffmpeg-2.8.1.tar.bz2/libavformat/concatdec.c -> ffmpeg-2.8.3.tar.bz2/libavformat/concatdec.c Changed

ffmpeg-2.8.1.tar.bz2/libavformat/hls.c -> ffmpeg-2.8.3.tar.bz2/libavformat/hls.c Changed

@@ -73,6 +73,8 @@
     char *key;
     enum KeyType key_type;
     uint8_t iv[16];
+    /* associated Media Initialization Section, treated as a segment */
+    struct segment *init_section;
 };
 
 struct rendition;
@@ -110,6 +112,13 @@
     int64_t cur_seg_offset;
     int64_t last_load_time;
 
+    /* Currently active Media Initialization Section */
+    struct segment *cur_init_section;
+    uint8_t *init_sec_buf;
+    unsigned int init_sec_buf_size;
+    unsigned int init_sec_data_len;
+    unsigned int init_sec_buf_read_offset;
+
     char key_url[MAX_URL_SIZE];
     uint8_t key[16];
 
@@ -135,6 +144,11 @@
      * multiple (playlist-less) renditions associated with them. */
     int n_renditions;
     struct rendition **renditions;
+
+    /* Media Initialization Sections (EXT-X-MAP) associated with this
+     * playlist, if any. */
+    int n_init_sections;
+    struct segment **init_sections;
 };
 
 /*
@@ -205,16 +219,29 @@
     pls->n_segments = 0;
 }
 
+static void free_init_section_list(struct playlist *pls)
+{
+    int i;
+    for (i = 0; i < pls->n_init_sections; i++) {
+        av_freep(&pls->init_sections[i]->url);
+        av_freep(&pls->init_sections[i]);
+    }
+    av_freep(&pls->init_sections);
+    pls->n_init_sections = 0;
+}
+
 static void free_playlist_list(HLSContext *c)
 {
     int i;
     for (i = 0; i < c->n_playlists; i++) {
         struct playlist *pls = c->playlists[i];
         free_segment_list(pls);
+        free_init_section_list(pls);
         av_freep(&pls->renditions);
         av_freep(&pls->id3_buf);
         av_dict_free(&pls->id3_initial);
         ff_id3v2_free_extra_meta(&pls->id3_deferred_extra);
+        av_freep(&pls->init_sec_buf);
         av_free_packet(&pls->pkt);
         av_freep(&pls->pb.buffer);
         if (pls->input)
@@ -352,6 +379,60 @@
     }
 }
 
+struct init_section_info {
+    char uri[MAX_URL_SIZE];
+    char byterange[32];
+};
+
+static struct segment *new_init_section(struct playlist *pls,
+                                        struct init_section_info *info,
+                                        const char *url_base)
+{
+    struct segment *sec;
+    char *ptr;
+    char tmp_str[MAX_URL_SIZE];
+
+    if (!info->uri[0])
+        return NULL;
+
+    sec = av_mallocz(sizeof(*sec));
+    if (!sec)
+        return NULL;
+
+    ff_make_absolute_url(tmp_str, sizeof(tmp_str), url_base, info->uri);
+    sec->url = av_strdup(tmp_str);
+    if (!sec->url) {
+        av_free(sec);
+        return NULL;
+    }
+
+    if (info->byterange[0]) {
+        sec->size = atoi(info->byterange);
+        ptr = strchr(info->byterange, '@');
+        if (ptr)
+            sec->url_offset = atoi(ptr+1);
+    } else {
+        /* the entire file is the init section */
+        sec->size = -1;
+    }
+
+    dynarray_add(&pls->init_sections, &pls->n_init_sections, sec);
+
+    return sec;
+}
+
+static void handle_init_section_args(struct init_section_info *info, const char *key,
+                                           int key_len, char **dest, int *dest_len)
+{
+    if (!strncmp(key, "URI=", key_len)) {
+        *dest     =        info->uri;
+        *dest_len = sizeof(info->uri);
+    } else if (!strncmp(key, "BYTERANGE=", key_len)) {
+        *dest     =        info->byterange;
+        *dest_len = sizeof(info->byterange);
+    }
+}
+
 struct rendition_info {
     char type[16];
     char uri[MAX_URL_SIZE];
@@ -525,6 +606,14 @@
     return ret;
 }
 
+static void update_options(char **dest, const char *name, void *src)
+{
+    av_freep(dest);
+    av_opt_get(src, name, 0, (uint8_t**)dest);
+    if (*dest && !strlen(*dest))
+        av_freep(dest);
+}
+
 static int open_url(HLSContext *c, URLContext **uc, const char *url, AVDictionary *opts)
 {
     AVDictionary *tmp = NULL;
@@ -534,6 +623,12 @@
     av_dict_copy(&tmp, opts, 0);
 
     ret = ffurl_open(uc, url, AVIO_FLAG_READ, c->interrupt_callback, &tmp);
+    if( ret >= 0) {
+        // update cookies on http response with setcookies.
+        URLContext *u = *uc;
+        update_options(&c->cookies, "cookies", u->priv_data);
+        av_dict_set(&opts, "cookies", c->cookies, 0);
+    }
 
     av_dict_free(&tmp);
 
@@ -557,6 +652,7 @@
     uint8_t *new_url = NULL;
     struct variant_info variant_info;
     char tmp_str[MAX_URL_SIZE];
+    struct segment *cur_init_section = NULL;
 
     if (!in) {
 #if 1
@@ -642,6 +738,14 @@
                 pls->type = PLS_TYPE_EVENT;
             else if (!strcmp(ptr, "VOD"))
                 pls->type = PLS_TYPE_VOD;
+        } else if (av_strstart(line, "#EXT-X-MAP:", &ptr)) {
+            struct init_section_info info = {{0}};
+            ret = ensure_playlist(c, &pls, url);
+            if (ret < 0)
+                goto fail;
+            ff_parse_key_value(ptr, (ff_parse_key_val_cb) handle_init_section_args,
+                               &info);
+            cur_init_section = new_init_section(pls, &info, url);
         } else if (av_strstart(line, "#EXT-X-ENDLIST", &ptr)) {
             if (pls)
                 pls->finished = 1;
@@ -720,6 +824,8 @@
                     seg->url_offset = 0;
                     seg_offset = 0;
                 }
+
+                seg->init_section = cur_init_section;
             }
         }
     }
@@ -733,17 +839,22 @@
     return ret;
 }
 
+static struct segment *current_segment(struct playlist *pls)
+{
+    return pls->segments[pls->cur_seq_no - pls->start_seq_no];
+}
+
 enum ReadFromURLMode {
     READ_NORMAL,
     READ_COMPLETE,
 };
 
 /* read from URLContext, limiting read to current segment */
-static int read_from_url(struct playlist *pls, uint8_t *buf, int buf_size,
+static int read_from_url(struct playlist *pls, struct segment *seg,
+                         uint8_t *buf, int buf_size,
                          enum ReadFromURLMode mode)
 {
     int ret;
-    struct segment *seg = pls->segments[pls->cur_seq_no - pls->start_seq_no];
 
      /* limit read if the segment was only a part of a file */
     if (seg->size >= 0)
@@ -866,12 +977,13 @@
     int bytes;
     int id3_buf_pos = 0;
     int fill_buf = 0;
+    struct segment *seg = current_segment(pls);
 
     /* gather all the id3 tags */
     while (1) {
         /* see if we can retrieve enough data for ID3 header */
         if (*len < ID3v2_HEADER_SIZE && buf_size >= ID3v2_HEADER_SIZE) {
-            bytes = read_from_url(pls, buf + *len, ID3v2_HEADER_SIZE - *len, READ_COMPLETE);
+            bytes = read_from_url(pls, seg, buf + *len, ID3v2_HEADER_SIZE - *len, READ_COMPLETE);
             if (bytes > 0) {
 
                 if (bytes == ID3v2_HEADER_SIZE - *len)
@@ -892,7 +1004,6 @@
             break;
 
         if (ff_id3v2_match(buf, ID3v2_DEFAULT_MAGIC)) {
-            struct segment *seg = pls->segments[pls->cur_seq_no - pls->start_seq_no];
             int64_t maxsize = seg->size >= 0 ? seg->size : 1024*1024;
             int taglen = ff_id3v2_tag_len(buf);
             int tag_got_bytes = FFMIN(taglen, *len);
@@ -924,7 +1035,7 @@
 
             if (remaining > 0) {
                 /* read the rest of the tag in */
-                if (read_from_url(pls, pls->id3_buf + id3_buf_pos, remaining, READ_COMPLETE) != remaining)
+                if (read_from_url(pls, seg, pls->id3_buf + id3_buf_pos, remaining, READ_COMPLETE) != remaining)
                     break;
                 id3_buf_pos += remaining;
                 av_log(pls->ctx, AV_LOG_DEBUG, "Stripped additional %d HLS ID3 bytes\n", remaining);
@@ -938,7 +1049,7 @@
 
     /* re-fill buffer for the caller unless EOF */
     if (*len >= 0 && (fill_buf || *len == 0)) {
-        bytes = read_from_url(pls, buf + *len, buf_size - *len, READ_NORMAL);
+        bytes = read_from_url(pls, seg, buf + *len, buf_size - *len, READ_NORMAL);
 
         /* ignore error if we already had some data */
         if (bytes >= 0)
@@ -958,20 +1069,10 @@
         pls->is_id3_timestamped = (pls->id3_mpegts_timestamp != AV_NOPTS_VALUE);
 }
 
-static void update_options(char **dest, const char *name, void *src)
-{
-    av_freep(dest);
-    av_opt_get(src, name, 0, (uint8_t**)dest);
-    if (*dest && !strlen(*dest))
-        av_freep(dest);
-}
-
-static int open_input(HLSContext *c, struct playlist *pls)
+static int open_input(HLSContext *c, struct playlist *pls, struct segment *seg)
 {
     AVDictionary *opts = NULL;
-    AVDictionary *opts2 = NULL;
     int ret;
-    struct segment *seg = pls->segments[pls->cur_seq_no - pls->start_seq_no];
 
     // broker prior HTTP options that should be consistent across requests
     av_dict_set(&opts, "user-agent", c->user_agent, 0);
@@ -979,9 +1080,6 @@
     av_dict_set(&opts, "headers", c->headers, 0);
     av_dict_set(&opts, "seekable", "0", 0);
 
-    // Same opts for key request (ffurl_open mutilates the opts so it cannot be used twice)
-    av_dict_copy(&opts2, opts, 0);
-
     if (seg->size >= 0) {
         /* try to restrict the HTTP request to the part we want
          * (if this is in fact a HTTP request) */
@@ -999,14 +1097,12 @@
         char iv[33], key[33], url[MAX_URL_SIZE];
         if (strcmp(seg->key, pls->key_url)) {
             URLContext *uc;
-            if (open_url(pls->parent->priv_data, &uc, seg->key, opts2) == 0) {
+            if (open_url(pls->parent->priv_data, &uc, seg->key, opts) == 0) {
                 if (ffurl_read_complete(uc, pls->key, sizeof(pls->key))
                     != sizeof(pls->key)) {
                     av_log(NULL, AV_LOG_ERROR, "Unable to read key file %s\n",
                            seg->key);
                 }
-                update_options(&c->cookies, "cookies", uc->priv_data);
-                av_dict_set(&opts, "cookies", c->cookies, 0);
                 ffurl_close(uc);
             } else {
                 av_log(NULL, AV_LOG_ERROR, "Unable to open key file %s\n",
@@ -1055,11 +1151,70 @@
 
 cleanup:
     av_dict_free(&opts);
-    av_dict_free(&opts2);
     pls->cur_seg_offset = 0;
     return ret;
 }
 
+static int update_init_section(struct playlist *pls, struct segment *seg)
+{
+    static const int max_init_section_size = 1024*1024;
+    HLSContext *c = pls->parent->priv_data;
+    int64_t sec_size;
+    int64_t urlsize;
+    int ret;
+
+    if (seg->init_section == pls->cur_init_section)
+        return 0;
+
+    pls->cur_init_section = NULL;
+
+    if (!seg->init_section)
+        return 0;
+
+    /* this will clobber playlist URLContext stuff, so this should be
+     * called between segments only */
+    ret = open_input(c, pls, seg->init_section);
+    if (ret < 0) {
+        av_log(pls->parent, AV_LOG_WARNING,
+               "Failed to open an initialization section in playlist %d\n",
+               pls->index);
+        return ret;
+    }
+
+    if (seg->init_section->size >= 0)
+        sec_size = seg->init_section->size;
+    else if ((urlsize = ffurl_size(pls->input)) >= 0)
+        sec_size = urlsize;
+    else
+        sec_size = max_init_section_size;
+
+    av_log(pls->parent, AV_LOG_DEBUG,
+           "Downloading an initialization section of size %"PRId64"\n",
+           sec_size);
+
+    sec_size = FFMIN(sec_size, max_init_section_size);
+
+    av_fast_malloc(&pls->init_sec_buf, &pls->init_sec_buf_size, sec_size);
+
+    ret = read_from_url(pls, seg->init_section, pls->init_sec_buf,
+                        pls->init_sec_buf_size, READ_COMPLETE);
+    ffurl_close(pls->input);
+    pls->input = NULL;
+
+    if (ret < 0)
+        return ret;
+
+    pls->cur_init_section = seg->init_section;
+    pls->init_sec_data_len = ret;
+    pls->init_sec_buf_read_offset = 0;
+
+    /* spec says audio elementary streams do not have media initialization
+     * sections, so there should be no ID3 timestamps */
+    pls->is_id3_timestamped = 0;
+
+    return 0;
+}
+
 static int64_t default_reload_interval(struct playlist *pls)
 {
     return pls->n_segments > 0 ?
@@ -1080,6 +1235,7 @@
 
     if (!v->input) {
         int64_t reload_interval;
+        struct segment *seg;
 
         /* Check that the playlist is still needed before opening a new
          * segment. */
@@ -1133,7 +1289,14 @@
             goto reload;
         }
 
-        ret = open_input(c, v);
+        seg = current_segment(v);
+
+        /* load/update Media Initialization Section, if any */
+        ret = update_init_section(v, seg);
+        if (ret)
+            return ret;
+
+        ret = open_input(c, v, seg);
         if (ret < 0) {
             if (ff_check_interrupt(c->interrupt_callback))
                 return AVERROR_EXIT;
@@ -1145,7 +1308,15 @@
         just_opened = 1;
     }
 
-    ret = read_from_url(v, buf, buf_size, READ_NORMAL);
+    if (v->init_sec_buf_read_offset < v->init_sec_data_len) {
+        /* Push init section out first before first actual segment */
+        int copy_size = FFMIN(v->init_sec_data_len - v->init_sec_buf_read_offset, buf_size);
+        memcpy(buf, v->init_sec_buf, copy_size);
+        v->init_sec_buf_read_offset += copy_size;
+        return copy_size;
+    }
+
+    ret = read_from_url(v, current_segment(v), buf, buf_size, READ_NORMAL);
     if (ret > 0) {
         if (just_opened && v->is_id3_timestamped != 0) {
             /* Intercept ID3 tags here, elementary audio streams are required
@@ -1613,6 +1784,7 @@
     int ret, i, minplaylist = -1;
 
     recheck_discard_flags(s, c->first_packet);
+    c->first_packet = 0;
 
     for (i = 0; i < c->n_playlists; i++) {
         struct playlist *pls = c->playlists[i];

ffmpeg-2.8.1.tar.bz2/libavformat/isom.h -> ffmpeg-2.8.3.tar.bz2/libavformat/isom.h Changed

ffmpeg-2.8.1.tar.bz2/libavformat/jvdec.c -> ffmpeg-2.8.3.tar.bz2/libavformat/jvdec.c Changed

ffmpeg-2.8.1.tar.bz2/libavformat/matroskadec.c -> ffmpeg-2.8.3.tar.bz2/libavformat/matroskadec.c Changed

ffmpeg-2.8.1.tar.bz2/libavformat/mov.c -> ffmpeg-2.8.3.tar.bz2/libavformat/mov.c Changed

@@ -3294,7 +3294,7 @@
     MOVFragment *frag = &c->fragment;
     MOVTrackExt *trex = NULL;
     MOVFragmentIndex* index = NULL;
-    int flags, track_id, i;
+    int flags, track_id, i, found = 0;
 
     avio_r8(pb); /* version */
     flags = avio_rb24(pb);
@@ -3312,15 +3312,6 @@
         av_log(c->fc, AV_LOG_ERROR, "could not find corresponding trex\n");
         return AVERROR_INVALIDDATA;
     }
-    for (i = 0; i < c->fragment_index_count; i++) {
-        MOVFragmentIndex* candidate = c->fragment_index_data[i];
-        if (candidate->track_id == frag->track_id) {
-            av_log(c->fc, AV_LOG_DEBUG,
-                   "found fragment index for track %u\n", frag->track_id);
-            index = candidate;
-            break;
-        }
-    }
 
     frag->base_data_offset = flags & MOV_TFHD_BASE_DATA_OFFSET ?
                              avio_rb64(pb) : flags & MOV_TFHD_DEFAULT_BASE_IS_MOOF ?
@@ -3334,24 +3325,33 @@
     frag->flags    = flags & MOV_TFHD_DEFAULT_FLAGS ?
                      avio_rb32(pb) : trex->flags;
     frag->time     = AV_NOPTS_VALUE;
-    if (index) {
-        int i, found = 0;
-        for (i = index->current_item; i < index->item_count; i++) {
-            if (frag->implicit_offset == index->items[i].moof_offset) {
-                av_log(c->fc, AV_LOG_DEBUG, "found fragment index entry "
-                        "for track %u and moof_offset %"PRId64"\n",
-                        frag->track_id, index->items[i].moof_offset);
-                frag->time = index->items[i].time;
-                index->current_item = i + 1;
-                found = 1;
+    for (i = 0; i < c->fragment_index_count; i++) {
+        int j;
+        MOVFragmentIndex* candidate = c->fragment_index_data[i];
+        if (candidate->track_id == frag->track_id) {
+            av_log(c->fc, AV_LOG_DEBUG,
+                   "found fragment index for track %u\n", frag->track_id);
+            index = candidate;
+            for (j = index->current_item; j < index->item_count; j++) {
+                if (frag->implicit_offset == index->items[j].moof_offset) {
+                    av_log(c->fc, AV_LOG_DEBUG, "found fragment index entry "
+                            "for track %u and moof_offset %"PRId64"\n",
+                            frag->track_id, index->items[j].moof_offset);
+                    frag->time = index->items[j].time;
+                    index->current_item = j + 1;
+                    found = 1;
+                    break;
+                }
             }
-        }
-        if (!found) {
-            av_log(c->fc, AV_LOG_WARNING, "track %u has a fragment index "
-                   "but it doesn't have an (in-order) entry for moof_offset "
-                   "%"PRId64"\n", frag->track_id, frag->implicit_offset);
+            if (found)
+                break;
         }
     }
+    if (index && !found) {
+        av_log(c->fc, AV_LOG_DEBUG, "track %u has a fragment index but "
+               "it doesn't have an (in-order) entry for moof_offset "
+               "%"PRId64"\n", frag->track_id, frag->implicit_offset);
+    }
     av_log(c->fc, AV_LOG_TRACE, "frag flags 0x%x\n", frag->flags);
     return 0;
 }
@@ -3541,7 +3541,106 @@
         return AVERROR_EOF;
 
     frag->implicit_offset = offset;
-    st->duration = sc->track_end = dts + sc->time_offset;
+
+    sc->track_end = dts + sc->time_offset;
+    if (st->duration < sc->track_end)
+        st->duration = sc->track_end;
+
+    return 0;
+}
+
+static int mov_read_sidx(MOVContext *c, AVIOContext *pb, MOVAtom atom)
+{
+    int64_t offset = avio_tell(pb) + atom.size, pts;
+    uint8_t version;
+    unsigned i, track_id;
+    AVStream *st = NULL;
+    MOVStreamContext *sc;
+    MOVFragmentIndex *index = NULL;
+    MOVFragmentIndex **tmp;
+    AVRational timescale;
+
+    version = avio_r8(pb);
+    if (version > 1) {
+        avpriv_request_sample(c->fc, "sidx version %u", version);
+        return AVERROR_PATCHWELCOME;
+    }
+
+    avio_rb24(pb); // flags
+
+    track_id = avio_rb32(pb); // Reference ID
+    for (i = 0; i < c->fc->nb_streams; i++) {
+        if (c->fc->streams[i]->id == track_id) {
+            st = c->fc->streams[i];
+            break;
+        }
+    }
+    if (!st) {
+        av_log(c->fc, AV_LOG_ERROR, "could not find corresponding track id %d\n", track_id);
+        return AVERROR_INVALIDDATA;
+    }
+
+    sc = st->priv_data;
+
+    timescale = av_make_q(1, avio_rb32(pb));
+
+    if (version == 0) {
+        pts = avio_rb32(pb);
+        offset += avio_rb32(pb);
+    } else {
+        pts = avio_rb64(pb);
+        offset += avio_rb64(pb);
+    }
+
+    avio_rb16(pb); // reserved
+
+    index = av_mallocz(sizeof(MOVFragmentIndex));
+    if (!index)
+        return AVERROR(ENOMEM);
+
+    index->track_id = track_id;
+
+    index->item_count = avio_rb16(pb);
+    index->items = av_mallocz_array(index->item_count, sizeof(MOVFragmentIndexItem));
+
+    if (!index->items) {
+        av_freep(&index);
+        return AVERROR(ENOMEM);
+    }
+
+    for (i = 0; i < index->item_count; i++) {
+        uint32_t size = avio_rb32(pb);
+        uint32_t duration = avio_rb32(pb);
+        if (size & 0x80000000) {
+            avpriv_request_sample(c->fc, "sidx reference_type 1");
+            av_freep(&index->items);
+            av_freep(&index);
+            return AVERROR_PATCHWELCOME;
+        }
+        avio_rb32(pb); // sap_flags
+        index->items[i].moof_offset = offset;
+        index->items[i].time = av_rescale_q(pts, st->time_base, timescale);
+        offset += size;
+        pts += duration;
+    }
+
+    st->duration = sc->track_end = pts;
+
+    tmp = av_realloc_array(c->fragment_index_data,
+                           c->fragment_index_count + 1,
+                           sizeof(MOVFragmentIndex*));
+    if (!tmp) {
+        av_freep(&index->items);
+        av_freep(&index);
+        return AVERROR(ENOMEM);
+    }
+
+    c->fragment_index_data = tmp;
+    c->fragment_index_data[c->fragment_index_count++] = index;
+
+    if (offset == avio_size(pb))
+        c->fragment_index_complete = 1;
+
     return 0;
 }
 
@@ -3799,6 +3898,7 @@
 { MKTAG('a','l','a','c'), mov_read_alac }, /* alac specific atom */
 { MKTAG('a','v','c','C'), mov_read_glbl },
 { MKTAG('p','a','s','p'), mov_read_pasp },
+{ MKTAG('s','i','d','x'), mov_read_sidx },
 { MKTAG('s','t','b','l'), mov_read_default },
 { MKTAG('s','t','c','o'), mov_read_stco },
 { MKTAG('s','t','p','s'), mov_read_stps },
@@ -3922,9 +4022,9 @@
                 return err;
             }
             if (c->found_moov && c->found_mdat &&
-                ((!pb->seekable || c->fc->flags & AVFMT_FLAG_IGNIDX) ||
+                ((!pb->seekable || c->fc->flags & AVFMT_FLAG_IGNIDX || c->fragment_index_complete) ||
                  start_pos + a.size == avio_size(pb))) {
-                if (!pb->seekable || c->fc->flags & AVFMT_FLAG_IGNIDX)
+                if (!pb->seekable || c->fc->flags & AVFMT_FLAG_IGNIDX || c->fragment_index_complete)
                     c->next_root_atom = start_pos + a.size;
                 c->atom_depth --;
                 return 0;
@@ -4529,6 +4629,52 @@
     return 1;
 }
 
+static int mov_switch_root(AVFormatContext *s, int64_t target)
+{
+    MOVContext *mov = s->priv_data;
+    int i, j;
+    int already_read = 0;
+
+    if (avio_seek(s->pb, target, SEEK_SET) != target) {
+        av_log(mov->fc, AV_LOG_ERROR, "root atom offset 0x%"PRIx64": partial file\n", target);
+        return AVERROR_INVALIDDATA;
+    }
+
+    mov->next_root_atom = 0;
+
+    for (i = 0; i < mov->fragment_index_count; i++) {
+        MOVFragmentIndex *index = mov->fragment_index_data[i];
+        int found = 0;
+        for (j = 0; j < index->item_count; j++) {
+            MOVFragmentIndexItem *item = &index->items[j];
+            if (found) {
+                mov->next_root_atom = item->moof_offset;
+                break; // Advance to next index in outer loop
+            } else if (item->moof_offset == target) {
+                index->current_item = FFMIN(j, index->current_item);
+                if (item->headers_read)
+                    already_read = 1;
+                item->headers_read = 1;
+                found = 1;
+            }
+        }
+        if (!found)
+            index->current_item = 0;
+    }
+
+    if (already_read)
+        return 0;
+
+    mov->found_mdat = 0;
+
+    if (mov_read_default(mov, s->pb, (MOVAtom){ AV_RL32("root"), INT64_MAX }) < 0 ||
+        avio_feof(s->pb))
+        return AVERROR_EOF;
+    av_log(s, AV_LOG_TRACE, "read fragments, offset 0x%"PRIx64"\n", avio_tell(s->pb));
+
+    return 1;
+}
+
 static int mov_read_packet(AVFormatContext *s, AVPacket *pkt)
 {
     MOVContext *mov = s->priv_data;
@@ -4539,19 +4685,11 @@
     mov->fc = s;
  retry:
     sample = mov_find_next_sample(s, &st);
-    if (!sample) {
-        mov->found_mdat = 0;
+    if (!sample || (mov->next_root_atom && sample->pos > mov->next_root_atom)) {
         if (!mov->next_root_atom)
             return AVERROR_EOF;
-        if (avio_seek(s->pb, mov->next_root_atom, SEEK_SET) != mov->next_root_atom) {
-            av_log(mov->fc, AV_LOG_ERROR, "next root atom offset 0x%"PRIx64": partial file\n", mov->next_root_atom);
-            return AVERROR_INVALIDDATA;
-        }
-        mov->next_root_atom = 0;
-        if (mov_read_default(mov, s->pb, (MOVAtom){ AV_RL32("root"), INT64_MAX }) < 0 ||
-            avio_feof(s->pb))
-            return AVERROR_EOF;
-        av_log(s, AV_LOG_TRACE, "read fragments, offset 0x%"PRIx64"\n", avio_tell(s->pb));
+        if ((ret = mov_switch_root(s, mov->next_root_atom)) < 0)
+            return ret;
         goto retry;
     }
     sc = st->priv_data;
@@ -4629,12 +4767,41 @@
     return 0;
 }
 
+static int mov_seek_fragment(AVFormatContext *s, AVStream *st, int64_t timestamp)
+{
+    MOVContext *mov = s->priv_data;
+    int i, j;
+
+    if (!mov->fragment_index_complete)
+        return 0;
+
+    for (i = 0; i < mov->fragment_index_count; i++) {
+        if (mov->fragment_index_data[i]->track_id == st->id) {
+            MOVFragmentIndex *index = index = mov->fragment_index_data[i];
+            for (j = index->item_count - 1; j >= 0; j--) {
+                if (index->items[j].time <= timestamp) {
+                    if (index->items[j].headers_read)
+                        return 0;
+
+                    return mov_switch_root(s, index->items[j].moof_offset);
+                }
+            }
+        }
+    }
+
+    return 0;
+}
+
 static int mov_seek_stream(AVFormatContext *s, AVStream *st, int64_t timestamp, int flags)
 {
     MOVStreamContext *sc = st->priv_data;
     int sample, time_sample;
     int i;
 
+    int ret = mov_seek_fragment(s, st, timestamp);
+    if (ret < 0)
+        return ret;
+
     sample = av_index_search_timestamp(st, timestamp, flags);
     av_log(s, AV_LOG_TRACE, "stream %d, timestamp %"PRId64", sample %d\n", st->index, timestamp, sample);
     if (sample < 0 && st->nb_index_entries && timestamp < st->index_entries[0].timestamp)

ffmpeg-2.8.1.tar.bz2/libavformat/mp3dec.c -> ffmpeg-2.8.3.tar.bz2/libavformat/mp3dec.c Changed

@@ -42,6 +42,9 @@
 
 #define XING_TOC_COUNT 100
 
+#define SAME_HEADER_MASK \
+   (0xffe00000 | (3 << 17) | (3 << 10) | (3 << 19))
+
 typedef struct {
     AVClass *class;
     int64_t filesize;
@@ -54,7 +57,7 @@
     int is_cbr;
 } MP3DecContext;
 
-static int check(AVIOContext *pb, int64_t pos);
+static int check(AVIOContext *pb, int64_t pos, uint32_t *header);
 
 /* mp3 read */
 
@@ -374,12 +377,21 @@
 
     off = avio_tell(s->pb);
     for (i = 0; i < 64 * 1024; i++) {
+        uint32_t header, header2;
+        int frame_size;
         if (!(i&1023))
             ffio_ensure_seekback(s->pb, i + 1024 + 4);
-        if (check(s->pb, off + i) >= 0) {
-            av_log(s, AV_LOG_INFO, "Skipping %d bytes of junk at %"PRId64".\n", i, off);
-            avio_seek(s->pb, off + i, SEEK_SET);
-            break;
+        frame_size = check(s->pb, off + i, &header);
+        if (frame_size > 0) {
+            avio_seek(s->pb, off, SEEK_SET);
+            ffio_ensure_seekback(s->pb, i + 1024 + frame_size + 4);
+            if (check(s->pb, off + i + frame_size, &header2) >= 0 &&
+                (header & SAME_HEADER_MASK) == (header2 & SAME_HEADER_MASK))
+            {
+                av_log(s, AV_LOG_INFO, "Skipping %d bytes of junk at %"PRId64".\n", i, off);
+                avio_seek(s->pb, off + i, SEEK_SET);
+                break;
+            }
         }
         avio_seek(s->pb, off, SEEK_SET);
     }
@@ -420,7 +432,7 @@
 
 #define SEEK_WINDOW 4096
 
-static int check(AVIOContext *pb, int64_t pos)
+static int check(AVIOContext *pb, int64_t pos, uint32_t *ret_header)
 {
     int64_t ret = avio_seek(pb, pos, SEEK_SET);
     unsigned header;
@@ -434,6 +446,8 @@
     if (avpriv_mpegaudio_decode_header(&sd, header) == 1)
         return -1;
 
+    if (ret_header)
+        *ret_header = header;
     return sd.frame_size;
 }
 
@@ -461,7 +475,7 @@
             continue;
 
         for(j=0; j<MIN_VALID; j++) {
-            ret = check(s->pb, pos);
+            ret = check(s->pb, pos, NULL);
             if(ret < 0)
                 break;
             if ((target_pos - pos)*dir <= 0 && abs(MIN_VALID/2-j) < score) {

ffmpeg-2.8.1.tar.bz2/libavformat/mpegts.c -> ffmpeg-2.8.3.tar.bz2/libavformat/mpegts.c Changed

ffmpeg-2.8.1.tar.bz2/libavformat/mxfdec.c -> ffmpeg-2.8.3.tar.bz2/libavformat/mxfdec.c Changed

ffmpeg-2.8.1.tar.bz2/libavformat/mxfenc.c -> ffmpeg-2.8.3.tar.bz2/libavformat/mxfenc.c Changed

@@ -19,7 +19,8 @@
  * License along with FFmpeg; if not, write to the Free Software
  * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
  *
- * signal_standard, color_siting and klv_fill_key version fixes sponsored by NOA GmbH
+ * signal_standard, color_siting, store_user_comments and klv_fill_key version
+ * fixes sponsored by NOA GmbH
  */
 
 /*
@@ -321,6 +322,7 @@
     int signal_standard;
     uint32_t tagged_value_count;
     AVRational audio_edit_rate;
+    int store_user_comments;
 } MXFContext;
 
 static const uint8_t uuid_base[]            = { 0xAD,0xAB,0x44,0x24,0x2f,0x25,0x4d,0xc7,0x92,0xff,0x29,0xbd };
@@ -378,7 +380,6 @@
     { 0x4404, {0x06,0x0E,0x2B,0x34,0x01,0x01,0x01,0x02,0x07,0x02,0x01,0x10,0x02,0x05,0x00,0x00}}, /* Package Modified Date */
     { 0x4402, {0x06,0x0E,0x2B,0x34,0x01,0x01,0x01,0x01,0x01,0x03,0x03,0x02,0x01,0x00,0x00,0x00}}, /* Package Name */
     { 0x4403, {0x06,0x0E,0x2B,0x34,0x01,0x01,0x01,0x02,0x06,0x01,0x01,0x04,0x06,0x05,0x00,0x00}}, /* Tracks Strong reference array */
-    { 0x4406, {0x06,0x0E,0x2B,0x34,0x01,0x01,0x01,0x02,0x03,0x02,0x01,0x02,0x0C,0x00,0x00,0x00}}, /* User Comments */
     { 0x4701, {0x06,0x0E,0x2B,0x34,0x01,0x01,0x01,0x02,0x06,0x01,0x01,0x04,0x02,0x03,0x00,0x00}}, /* Descriptor */
     // Track
     { 0x4801, {0x06,0x0E,0x2B,0x34,0x01,0x01,0x01,0x02,0x01,0x07,0x01,0x01,0x00,0x00,0x00,0x00}}, /* Track ID */
@@ -398,9 +399,6 @@
     { 0x1501, {0x06,0x0E,0x2B,0x34,0x01,0x01,0x01,0x02,0x07,0x02,0x01,0x03,0x01,0x05,0x00,0x00}}, /* Start Time Code */
     { 0x1502, {0x06,0x0E,0x2B,0x34,0x01,0x01,0x01,0x02,0x04,0x04,0x01,0x01,0x02,0x06,0x00,0x00}}, /* Rounded Time Code Base */
     { 0x1503, {0x06,0x0E,0x2B,0x34,0x01,0x01,0x01,0x01,0x04,0x04,0x01,0x01,0x05,0x00,0x00,0x00}}, /* Drop Frame */
-    // Tagged Value
-    { 0x5001, {0x06,0x0E,0x2B,0x34,0x01,0x01,0x01,0x02,0x03,0x02,0x01,0x02,0x09,0x01,0x00,0x00}}, /* Name */
-    { 0x5003, {0x06,0x0E,0x2B,0x34,0x01,0x01,0x01,0x02,0x03,0x02,0x01,0x02,0x0A,0x01,0x00,0x00}}, /* Value */
     // File Descriptor
     { 0x3F01, {0x06,0x0E,0x2B,0x34,0x01,0x01,0x01,0x04,0x06,0x01,0x01,0x04,0x06,0x0B,0x00,0x00}}, /* Sub Descriptors reference array */
     { 0x3006, {0x06,0x0E,0x2B,0x34,0x01,0x01,0x01,0x05,0x06,0x01,0x01,0x03,0x05,0x00,0x00,0x00}}, /* Linked Track ID */
@@ -446,6 +444,12 @@
     { 0x3D0A, {0x06,0x0E,0x2B,0x34,0x01,0x01,0x01,0x05,0x04,0x02,0x03,0x02,0x01,0x00,0x00,0x00}}, /* Block Align */
 };
 
+static const MXFLocalTagPair mxf_user_comments_local_tag[] = {
+    { 0x4406, {0x06,0x0E,0x2B,0x34,0x01,0x01,0x01,0x02,0x03,0x02,0x01,0x02,0x0C,0x00,0x00,0x00}}, /* User Comments */
+    { 0x5001, {0x06,0x0E,0x2B,0x34,0x01,0x01,0x01,0x02,0x03,0x02,0x01,0x02,0x09,0x01,0x00,0x00}}, /* Name */
+    { 0x5003, {0x06,0x0E,0x2B,0x34,0x01,0x01,0x01,0x02,0x03,0x02,0x01,0x02,0x0A,0x01,0x00,0x00}}, /* Value */
+};
+
 static void mxf_write_uuid(AVIOContext *pb, enum MXFMetadataSetType type, int value)
 {
     avio_write(pb, uuid_base, 12);
@@ -523,10 +527,12 @@
 
 static void mxf_write_primer_pack(AVFormatContext *s)
 {
+    MXFContext *mxf = s->priv_data;
     AVIOContext *pb = s->pb;
     int local_tag_number, i = 0;
 
     local_tag_number = FF_ARRAY_ELEMS(mxf_local_tag_batch);
+    local_tag_number += mxf->store_user_comments * FF_ARRAY_ELEMS(mxf_user_comments_local_tag);
 
     avio_write(pb, primer_pack_key, 16);
     klv_encode_ber_length(pb, local_tag_number * 18 + 8);
@@ -534,10 +540,15 @@
     avio_wb32(pb, local_tag_number); // local_tag num
     avio_wb32(pb, 18); // item size, always 18 according to the specs
 
-    for (i = 0; i < local_tag_number; i++) {
+    for (i = 0; i < FF_ARRAY_ELEMS(mxf_local_tag_batch); i++) {
         avio_wb16(pb, mxf_local_tag_batch[i].local_tag);
         avio_write(pb, mxf_local_tag_batch[i].uid, 16);
     }
+    if (mxf->store_user_comments)
+        for (i = 0; i < FF_ARRAY_ELEMS(mxf_user_comments_local_tag); i++) {
+            avio_wb16(pb, mxf_user_comments_local_tag[i].local_tag);
+            avio_write(pb, mxf_user_comments_local_tag[i].uid, 16);
+        }
 }
 
 static void mxf_write_local_tag(AVIOContext *pb, int size, int tag)
@@ -1251,14 +1262,15 @@
     int user_comment_count = 0;
 
     if (type == MaterialPackage) {
-        user_comment_count = mxf_write_user_comments(s, s->metadata);
+        if (mxf->store_user_comments)
+            user_comment_count = mxf_write_user_comments(s, s->metadata);
         mxf_write_metadata_key(pb, 0x013600);
         PRINT_KEY(s, "Material Package key", pb->buf_ptr - 16);
-        klv_encode_ber_length(pb, 104 + name_size + (16*track_count) + (16*user_comment_count));
+        klv_encode_ber_length(pb, 92 + name_size + (16*track_count) + (16*user_comment_count) + 12*mxf->store_user_comments);
     } else {
         mxf_write_metadata_key(pb, 0x013700);
         PRINT_KEY(s, "Source Package key", pb->buf_ptr - 16);
-        klv_encode_ber_length(pb, 124 + name_size + (16*track_count)); // 20 bytes length for descriptor reference
+        klv_encode_ber_length(pb, 112 + name_size + (16*track_count) + 12*mxf->store_user_comments); // 20 bytes length for descriptor reference
     }
 
     // write uid
@@ -1293,10 +1305,12 @@
         mxf_write_uuid(pb, type == MaterialPackage ? Track : Track + TypeBottom, i);
 
     // write user comment refs
-    mxf_write_local_tag(pb, user_comment_count*16 + 8, 0x4406);
-    mxf_write_refs_count(pb, user_comment_count);
-    for (i = 0; i < user_comment_count; i++)
-         mxf_write_uuid(pb, TaggedValue, mxf->tagged_value_count - user_comment_count + i);
+    if (mxf->store_user_comments) {
+        mxf_write_local_tag(pb, user_comment_count*16 + 8, 0x4406);
+        mxf_write_refs_count(pb, user_comment_count);
+        for (i = 0; i < user_comment_count; i++)
+            mxf_write_uuid(pb, TaggedValue, mxf->tagged_value_count - user_comment_count + i);
+    }
 
     // write multiple descriptor reference
     if (type == SourcePackage) {
@@ -2036,6 +2050,9 @@
         return -1;
     }
 
+    if (!av_dict_get(s->metadata, "comment_", NULL, AV_DICT_IGNORE_SUFFIX))
+        mxf->store_user_comments = 0;
+
     for (i = 0; i < s->nb_streams; i++) {
         AVStream *st = s->streams[i];
         MXFStreamContext *sc = av_mallocz(sizeof(*sc));
@@ -2654,6 +2671,8 @@
 
 static const AVOption mxf_options[] = {
     MXF_COMMON_OPTIONS
+    { "store_user_comments", "",
+      offsetof(MXFContext, store_user_comments), AV_OPT_TYPE_INT, {.i64 = 1}, 0, 1, AV_OPT_FLAG_ENCODING_PARAM},
     { NULL },
 };
 
@@ -2668,6 +2687,8 @@
     { "d10_channelcount", "Force/set channelcount in generic sound essence descriptor",
       offsetof(MXFContext, channel_count), AV_OPT_TYPE_INT, {.i64 = -1}, -1, 8, AV_OPT_FLAG_ENCODING_PARAM},
     MXF_COMMON_OPTIONS
+    { "store_user_comments", "",
+      offsetof(MXFContext, store_user_comments), AV_OPT_TYPE_INT, {.i64 = 0}, 0, 1, AV_OPT_FLAG_ENCODING_PARAM},
     { NULL },
 };

ffmpeg-2.8.1.tar.bz2/libavformat/rtmpcrypt.c -> ffmpeg-2.8.3.tar.bz2/libavformat/rtmpcrypt.c Changed

ffmpeg-2.8.1.tar.bz2/libavformat/rtsp.c -> ffmpeg-2.8.3.tar.bz2/libavformat/rtsp.c Changed

ffmpeg-2.8.1.tar.bz2/libavformat/utils.c -> ffmpeg-2.8.3.tar.bz2/libavformat/utils.c Changed

ffmpeg-2.8.1.tar.bz2/libavformat/xmv.c -> ffmpeg-2.8.3.tar.bz2/libavformat/xmv.c Changed

ffmpeg-2.8.1.tar.bz2/libavutil/channel_layout.c -> ffmpeg-2.8.3.tar.bz2/libavutil/channel_layout.c Changed

ffmpeg-2.8.1.tar.bz2/libavutil/common.h -> ffmpeg-2.8.3.tar.bz2/libavutil/common.h Changed

ffmpeg-2.8.1.tar.bz2/libavutil/file_open.c -> ffmpeg-2.8.3.tar.bz2/libavutil/file_open.c Changed

ffmpeg-2.8.1.tar.bz2/libavutil/softfloat.c -> ffmpeg-2.8.3.tar.bz2/libavutil/softfloat.c Changed

@@ -97,6 +97,48 @@
     sf1 = av_int2sf(0xE0000001, 0);
     printf("test4 softfloat: %.10lf (0x%08x %d)\n", (double)av_sf2double(sf1), sf1.mant, sf1.exp);
 
+
+    sf1 = (SoftFloat){ 0x20000000,   MIN_EXP };
+    sf1 = av_mul_sf(sf1, sf1);
+    printf("test5 softfloat: %.10lf (0x%08x %d)\n", (double)av_sf2double(sf1), sf1.mant, sf1.exp);
+
+    sf1 = (SoftFloat){ 0x20000000,   MIN_EXP };
+    sf2 = (SoftFloat){ 0x20000000,   MAX_EXP };
+    i = av_cmp_sf(sf1, sf2);
+    j = av_cmp_sf(sf2, sf1);
+    sf1 = av_div_sf(sf1, sf2);
+    printf("test6 softfloat: %.10lf (0x%08x %d) %d %d\n", (double)av_sf2double(sf1), sf1.mant, sf1.exp, i, j);
+
+    for(i= -50; i<50; i++) {
+        sf1= av_int2sf(i, 0);
+        for(j= -50; j<50; j++) {
+            int c;
+            sf2= av_int2sf(j, 0);
+            c = av_cmp_sf(sf1, sf2);
+            if (FFDIFFSIGN(i,j) != c && (FFDIFFSIGN(i,j)^c)<0) {
+                printf("av_cmp_sf failed at %d %d as %X\n", i, j, c);
+            }
+            c = av_gt_sf(sf1, sf2);
+            if ((i>j) != c) {
+                printf("av_gt_sf failed at %d %d as %X\n", i, j, c);
+            }
+        }
+        sf1 = av_int2sf(1, i);
+        for(j = -50; j < 50; j++) {
+            int c;
+            sf2 = av_int2sf(1, j);
+            c = av_cmp_sf(sf2, sf1);
+            if (FFDIFFSIGN(i,j) != c && (FFDIFFSIGN(i,j)^c) < 0) {
+                printf("av_cmp_sf failed2 at %d %d as %X\n", i, j, c);
+            }
+            c = av_gt_sf(sf1, sf2);
+            if ((i<j) != c) {
+                printf("av_gt_sf failed2 at %d %d as %X\n", i, j, c);
+            }
+        }
+    }
+
+
     for(i= 0; i<4*36; i++){
         int s, c;
         double errs, errc;

ffmpeg-2.8.1.tar.bz2/libavutil/softfloat.h -> ffmpeg-2.8.3.tar.bz2/libavutil/softfloat.h Changed

@@ -36,7 +36,7 @@
     int32_t  exp;
 }SoftFloat;
 
-static const SoftFloat FLOAT_0          = {          0,   0};
+static const SoftFloat FLOAT_0          = {          0,   MIN_EXP};
 static const SoftFloat FLOAT_05         = { 0x20000000,   0};
 static const SoftFloat FLOAT_1          = { 0x20000000,   1};
 static const SoftFloat FLOAT_EPSILON    = { 0x29F16B12, -16};
@@ -90,7 +90,7 @@
 }
 
 /**
- * @return Will not be more denormalized than a+b. So if either input is
+ * @return Will not be more denormalized than a*b. So if either input is
  *         normalized, then the output will not be worse then the other input.
  *         If both are normalized, then the output will be normalized.
  */
@@ -98,7 +98,10 @@
     a.exp += b.exp;
     av_assert2((int32_t)((a.mant * (int64_t)b.mant) >> ONE_BITS) == (a.mant * (int64_t)b.mant) >> ONE_BITS);
     a.mant = (a.mant * (int64_t)b.mant) >> ONE_BITS;
-    return av_normalize1_sf((SoftFloat){a.mant, a.exp - 1});
+    a = av_normalize1_sf((SoftFloat){a.mant, a.exp - 1});
+    if (!a.mant || a.exp < MIN_EXP)
+        return FLOAT_0;
+    return a;
 }
 
 /**
@@ -108,20 +111,27 @@
 static inline av_const SoftFloat av_div_sf(SoftFloat a, SoftFloat b){
     a.exp -= b.exp;
     a.mant = ((int64_t)a.mant<<(ONE_BITS+1)) / b.mant;
-    return av_normalize1_sf(a);
+    a = av_normalize1_sf(a);
+    if (!a.mant || a.exp < MIN_EXP)
+        return FLOAT_0;
+    return a;
 }
 
 static inline av_const int av_cmp_sf(SoftFloat a, SoftFloat b){
     int t= a.exp - b.exp;
-    if(t<0) return (a.mant >> (-t)) -  b.mant      ;
-    else    return  a.mant          - (b.mant >> t);
+    if      (t <-31) return                  -  b.mant      ;
+    else if (t <  0) return (a.mant >> (-t)) -  b.mant      ;
+    else if (t < 32) return  a.mant          - (b.mant >> t);
+    else             return  a.mant                         ;
 }
 
 static inline av_const int av_gt_sf(SoftFloat a, SoftFloat b)
 {
     int t= a.exp - b.exp;
-    if(t<0) return (a.mant >> (-t)) >  b.mant      ;
-    else    return  a.mant          > (b.mant >> t);
+    if      (t <-31) return 0                >  b.mant      ;
+    else if (t <  0) return (a.mant >> (-t)) >  b.mant      ;
+    else if (t < 32) return  a.mant          > (b.mant >> t);
+    else             return  a.mant          >  0           ;
 }
 
 static inline av_const SoftFloat av_add_sf(SoftFloat a, SoftFloat b){
@@ -143,7 +153,12 @@
  * @returns a SoftFloat with value v * 2^frac_bits
  */
 static inline av_const SoftFloat av_int2sf(int v, int frac_bits){
-    return av_normalize_sf((SoftFloat){v, ONE_BITS + 1 - frac_bits});
+    int exp_offset = 0;
+    if(v == INT_MIN){
+        exp_offset = 1;
+        v>>=1;
+    }
+    return av_normalize_sf(av_normalize1_sf((SoftFloat){v, ONE_BITS + 1 - frac_bits + exp_offset}));
 }
 
 /**
@@ -163,7 +178,9 @@
     int tabIndex, rem;
 
     if (val.mant == 0)
-        val.exp = 0;
+        val.exp = MIN_EXP;
+    else if (val.mant < 0)
+        abort();
     else
     {
         tabIndex = (val.mant - 0x20000000) >> 20;

ffmpeg-2.8.1.tar.bz2/libswresample/resample.c -> ffmpeg-2.8.3.tar.bz2/libswresample/resample.c Changed

ffmpeg-2.8.1.tar.bz2/tests/fate/avformat.mak -> ffmpeg-2.8.3.tar.bz2/tests/fate/avformat.mak Changed

ffmpeg-2.8.1.tar.bz2/tests/fate/libswresample.mak -> ffmpeg-2.8.3.tar.bz2/tests/fate/libswresample.mak Changed

ffmpeg-2.8.1.tar.bz2/tests/ref/lavf/mxf -> ffmpeg-2.8.3.tar.bz2/tests/ref/lavf/mxf Changed

ffmpeg-2.8.1.tar.bz2/tests/ref/lavf/mxf_d10 -> ffmpeg-2.8.3.tar.bz2/tests/ref/lavf/mxf_d10 Changed

ffmpeg-2.8.1.tar.bz2/tests/ref/lavf/mxf_opatom -> ffmpeg-2.8.3.tar.bz2/tests/ref/lavf/mxf_opatom Changed

ffmpeg-2.8.1.tar.bz2/tests/ref/lavf/mxf_opatom_audio -> ffmpeg-2.8.3.tar.bz2/tests/ref/lavf/mxf_opatom_audio Changed

ffmpeg-2.8.1.tar.bz2/tests/ref/seek/lavf-mxf -> ffmpeg-2.8.3.tar.bz2/tests/ref/seek/lavf-mxf Changed

@@ -1,48 +1,48 @@
-ret: 0         st: 0 flags:1 dts:-0.040000 pts: 0.000000 pos:   6656 size: 24801
+ret: 0         st: 0 flags:1 dts:-0.040000 pts: 0.000000 pos:   6144 size: 24801
 ret: 0         st:-1 flags:0  ts:-1.000000
-ret: 0         st: 0 flags:1 dts:-0.040000 pts: 0.000000 pos:   6656 size: 24801
+ret: 0         st: 0 flags:1 dts:-0.040000 pts: 0.000000 pos:   6144 size: 24801
 ret: 0         st:-1 flags:1  ts: 1.894167
-ret: 0         st: 0 flags:1 dts: 0.840000 pts: 0.960000 pos: 460800 size: 24711
+ret: 0         st: 0 flags:1 dts: 0.840000 pts: 0.960000 pos: 460288 size: 24711
 ret: 0         st: 0 flags:0  ts: 0.800000
-ret: 0         st: 0 flags:1 dts: 0.840000 pts: 0.960000 pos: 460800 size: 24711
+ret: 0         st: 0 flags:1 dts: 0.840000 pts: 0.960000 pos: 460288 size: 24711
 ret: 0         st: 0 flags:1  ts:-0.320000
-ret: 0         st: 0 flags:1 dts:-0.040000 pts: 0.000000 pos:   6656 size: 24801
+ret: 0         st: 0 flags:1 dts:-0.040000 pts: 0.000000 pos:   6144 size: 24801
 ret:-1         st: 1 flags:0  ts: 2.576667
 ret: 0         st: 1 flags:1  ts: 1.470833
-ret: 0         st: 0 flags:1 dts: 0.840000 pts: 0.960000 pos: 460800 size: 24711
+ret: 0         st: 0 flags:1 dts: 0.840000 pts: 0.960000 pos: 460288 size: 24711
 ret: 0         st:-1 flags:0  ts: 0.365002
-ret: 0         st: 0 flags:1 dts: 0.360000 pts: 0.480000 pos: 211968 size: 24786
+ret: 0         st: 0 flags:1 dts: 0.360000 pts: 0.480000 pos: 211456 size: 24786
 ret: 0         st:-1 flags:1  ts:-0.740831
-ret: 0         st: 0 flags:1 dts:-0.040000 pts: 0.000000 pos:   6656 size: 24801
+ret: 0         st: 0 flags:1 dts:-0.040000 pts: 0.000000 pos:   6144 size: 24801
 ret:-1         st: 0 flags:0  ts: 2.160000
 ret: 0         st: 0 flags:1  ts: 1.040000
-ret: 0         st: 0 flags:1 dts: 0.840000 pts: 0.960000 pos: 460800 size: 24711
+ret: 0         st: 0 flags:1 dts: 0.840000 pts: 0.960000 pos: 460288 size: 24711
 ret: 0         st: 1 flags:0  ts:-0.058333
-ret: 0         st: 0 flags:1 dts:-0.040000 pts: 0.000000 pos:   6656 size: 24801
+ret: 0         st: 0 flags:1 dts:-0.040000 pts: 0.000000 pos:   6144 size: 24801
 ret: 0         st: 1 flags:1  ts: 2.835833
-ret: 0         st: 0 flags:1 dts: 0.840000 pts: 0.960000 pos: 460800 size: 24711
+ret: 0         st: 0 flags:1 dts: 0.840000 pts: 0.960000 pos: 460288 size: 24711
 ret:-1         st:-1 flags:0  ts: 1.730004
 ret: 0         st:-1 flags:1  ts: 0.624171
-ret: 0         st: 0 flags:1 dts: 0.360000 pts: 0.480000 pos: 211968 size: 24786
+ret: 0         st: 0 flags:1 dts: 0.360000 pts: 0.480000 pos: 211456 size: 24786
 ret: 0         st: 0 flags:0  ts:-0.480000
-ret: 0         st: 0 flags:1 dts:-0.040000 pts: 0.000000 pos:   6656 size: 24801
+ret: 0         st: 0 flags:1 dts:-0.040000 pts: 0.000000 pos:   6144 size: 24801
 ret: 0         st: 0 flags:1  ts: 2.400000
-ret: 0         st: 0 flags:1 dts: 0.840000 pts: 0.960000 pos: 460800 size: 24711
+ret: 0         st: 0 flags:1 dts: 0.840000 pts: 0.960000 pos: 460288 size: 24711
 ret:-1         st: 1 flags:0  ts: 1.306667
 ret: 0         st: 1 flags:1  ts: 0.200833
-ret: 0         st: 0 flags:1 dts:-0.040000 pts: 0.000000 pos:   6656 size: 24801
+ret: 0         st: 0 flags:1 dts:-0.040000 pts: 0.000000 pos:   6144 size: 24801
 ret: 0         st:-1 flags:0  ts:-0.904994
-ret: 0         st: 0 flags:1 dts:-0.040000 pts: 0.000000 pos:   6656 size: 24801
+ret: 0         st: 0 flags:1 dts:-0.040000 pts: 0.000000 pos:   6144 size: 24801
 ret: 0         st:-1 flags:1  ts: 1.989173
-ret: 0         st: 0 flags:1 dts: 0.840000 pts: 0.960000 pos: 460800 size: 24711
+ret: 0         st: 0 flags:1 dts: 0.840000 pts: 0.960000 pos: 460288 size: 24711
 ret: 0         st: 0 flags:0  ts: 0.880000
-ret: 0         st: 0 flags:1 dts: 0.840000 pts: 0.960000 pos: 460800 size: 24711
+ret: 0         st: 0 flags:1 dts: 0.840000 pts: 0.960000 pos: 460288 size: 24711
 ret: 0         st: 0 flags:1  ts:-0.240000
-ret: 0         st: 0 flags:1 dts:-0.040000 pts: 0.000000 pos:   6656 size: 24801
+ret: 0         st: 0 flags:1 dts:-0.040000 pts: 0.000000 pos:   6144 size: 24801
 ret:-1         st: 1 flags:0  ts: 2.671667
 ret: 0         st: 1 flags:1  ts: 1.565833
-ret: 0         st: 0 flags:1 dts: 0.840000 pts: 0.960000 pos: 460800 size: 24711
+ret: 0         st: 0 flags:1 dts: 0.840000 pts: 0.960000 pos: 460288 size: 24711
 ret: 0         st:-1 flags:0  ts: 0.460008
-ret: 0         st: 0 flags:1 dts: 0.360000 pts: 0.480000 pos: 211968 size: 24786
+ret: 0         st: 0 flags:1 dts: 0.360000 pts: 0.480000 pos: 211456 size: 24786
 ret: 0         st:-1 flags:1  ts:-0.645825
-ret: 0         st: 0 flags:1 dts:-0.040000 pts: 0.000000 pos:   6656 size: 24801
+ret: 0         st: 0 flags:1 dts:-0.040000 pts: 0.000000 pos:   6144 size: 24801

ffmpeg-2.8.1.tar.bz2/tests/ref/seek/lavf-mxf_opatom_audio -> ffmpeg-2.8.3.tar.bz2/tests/ref/seek/lavf-mxf_opatom_audio Changed

@@ -1,53 +1,53 @@
-ret: 0         st: 0 flags:1 dts: 0.000000 pts: 0.000000 pos:   5145 size:  3840
+ret: 0         st: 0 flags:1 dts: 0.000000 pts: 0.000000 pos:   4633 size:  3840
 ret: 0         st:-1 flags:0  ts:-1.000000
-ret: 0         st: 0 flags:1 dts: 0.000000 pts: 0.000000 pos:   5145 size:  3840
+ret: 0         st: 0 flags:1 dts: 0.000000 pts: 0.000000 pos:   4633 size:  3840
 ret: 0         st:-1 flags:1  ts: 1.894167
-ret: 0         st: 0 flags:1 dts: 0.999979 pts: 0.999979 pos: 101143 size:     2
+ret: 0         st: 0 flags:1 dts: 0.999979 pts: 0.999979 pos: 100631 size:     2
 ret: 0         st: 0 flags:0  ts: 0.788333
-ret: 0         st: 0 flags:1 dts: 0.788333 pts: 0.788333 pos:  80825 size:  3840
+ret: 0         st: 0 flags:1 dts: 0.788333 pts: 0.788333 pos:  80313 size:  3840
 ret: 0         st: 0 flags:1  ts:-0.317500
-ret: 0         st: 0 flags:1 dts: 0.000000 pts: 0.000000 pos:   5145 size:  3840
+ret: 0         st: 0 flags:1 dts: 0.000000 pts: 0.000000 pos:   4633 size:  3840
 ret: 0         st:-1 flags:0  ts: 2.576668
-ret: 0         st: 0 flags:1 dts: 0.999979 pts: 0.999979 pos: 101143 size:     2
+ret: 0         st: 0 flags:1 dts: 0.999979 pts: 0.999979 pos: 100631 size:     2
 ret: 0         st:-1 flags:1  ts: 1.470835
-ret: 0         st: 0 flags:1 dts: 0.999979 pts: 0.999979 pos: 101143 size:     2
+ret: 0         st: 0 flags:1 dts: 0.999979 pts: 0.999979 pos: 100631 size:     2
 ret: 0         st: 0 flags:0  ts: 0.365000
-ret: 0         st: 0 flags:1 dts: 0.365000 pts: 0.365000 pos:  40185 size:  3840
+ret: 0         st: 0 flags:1 dts: 0.365000 pts: 0.365000 pos:  39673 size:  3840
 ret: 0         st: 0 flags:1  ts:-0.740833
-ret: 0         st: 0 flags:1 dts: 0.000000 pts: 0.000000 pos:   5145 size:  3840
+ret: 0         st: 0 flags:1 dts: 0.000000 pts: 0.000000 pos:   4633 size:  3840
 ret: 0         st:-1 flags:0  ts: 2.153336
-ret: 0         st: 0 flags:1 dts: 0.999979 pts: 0.999979 pos: 101143 size:     2
+ret: 0         st: 0 flags:1 dts: 0.999979 pts: 0.999979 pos: 100631 size:     2
 ret: 0         st:-1 flags:1  ts: 1.047503
-ret: 0         st: 0 flags:1 dts: 0.999979 pts: 0.999979 pos: 101143 size:     2
+ret: 0         st: 0 flags:1 dts: 0.999979 pts: 0.999979 pos: 100631 size:     2
 ret: 0         st: 0 flags:0  ts:-0.058333
-ret: 0         st: 0 flags:1 dts: 0.000000 pts: 0.000000 pos:   5145 size:  3840
+ret: 0         st: 0 flags:1 dts: 0.000000 pts: 0.000000 pos:   4633 size:  3840
 ret: 0         st: 0 flags:1  ts: 2.835833
-ret: 0         st: 0 flags:1 dts: 0.999979 pts: 0.999979 pos: 101143 size:     2
+ret: 0         st: 0 flags:1 dts: 0.999979 pts: 0.999979 pos: 100631 size:     2
 ret: 0         st:-1 flags:0  ts: 1.730004
-ret: 0         st: 0 flags:1 dts: 0.999979 pts: 0.999979 pos: 101143 size:     2
+ret: 0         st: 0 flags:1 dts: 0.999979 pts: 0.999979 pos: 100631 size:     2
 ret: 0         st:-1 flags:1  ts: 0.624171
-ret: 0         st: 0 flags:1 dts: 0.624167 pts: 0.624167 pos:  65065 size:  3840
+ret: 0         st: 0 flags:1 dts: 0.624167 pts: 0.624167 pos:  64553 size:  3840
 ret: 0         st: 0 flags:0  ts:-0.481667
-ret: 0         st: 0 flags:1 dts: 0.000000 pts: 0.000000 pos:   5145 size:  3840
+ret: 0         st: 0 flags:1 dts: 0.000000 pts: 0.000000 pos:   4633 size:  3840
 ret: 0         st: 0 flags:1  ts: 2.412500
-ret: 0         st: 0 flags:1 dts: 0.999979 pts: 0.999979 pos: 101143 size:     2
+ret: 0         st: 0 flags:1 dts: 0.999979 pts: 0.999979 pos: 100631 size:     2
 ret: 0         st:-1 flags:0  ts: 1.306672
-ret: 0         st: 0 flags:1 dts: 0.999979 pts: 0.999979 pos: 101143 size:     2
+ret: 0         st: 0 flags:1 dts: 0.999979 pts: 0.999979 pos: 100631 size:     2
 ret: 0         st:-1 flags:1  ts: 0.200839
-ret: 0         st: 0 flags:1 dts: 0.200833 pts: 0.200833 pos:  24425 size:  3840
+ret: 0         st: 0 flags:1 dts: 0.200833 pts: 0.200833 pos:  23913 size:  3840
 ret: 0         st: 0 flags:0  ts:-0.905000
-ret: 0         st: 0 flags:1 dts: 0.000000 pts: 0.000000 pos:   5145 size:  3840
+ret: 0         st: 0 flags:1 dts: 0.000000 pts: 0.000000 pos:   4633 size:  3840
 ret: 0         st: 0 flags:1  ts: 1.989167
-ret: 0         st: 0 flags:1 dts: 0.999979 pts: 0.999979 pos: 101143 size:     2
+ret: 0         st: 0 flags:1 dts: 0.999979 pts: 0.999979 pos: 100631 size:     2
 ret: 0         st:-1 flags:0  ts: 0.883340
-ret: 0         st: 0 flags:1 dts: 0.883333 pts: 0.883333 pos:  89945 size:  3840
+ret: 0         st: 0 flags:1 dts: 0.883333 pts: 0.883333 pos:  89433 size:  3840
 ret: 0         st:-1 flags:1  ts:-0.222493
-ret: 0         st: 0 flags:1 dts: 0.000000 pts: 0.000000 pos:   5145 size:  3840
+ret: 0         st: 0 flags:1 dts: 0.000000 pts: 0.000000 pos:   4633 size:  3840
 ret: 0         st: 0 flags:0  ts: 2.671667
-ret: 0         st: 0 flags:1 dts: 0.999979 pts: 0.999979 pos: 101143 size:     2
+ret: 0         st: 0 flags:1 dts: 0.999979 pts: 0.999979 pos: 100631 size:     2
 ret: 0         st: 0 flags:1  ts: 1.565833
-ret: 0         st: 0 flags:1 dts: 0.999979 pts: 0.999979 pos: 101143 size:     2
+ret: 0         st: 0 flags:1 dts: 0.999979 pts: 0.999979 pos: 100631 size:     2
 ret: 0         st:-1 flags:0  ts: 0.460008
-ret: 0         st: 0 flags:1 dts: 0.460000 pts: 0.460000 pos:  49305 size:  3840
+ret: 0         st: 0 flags:1 dts: 0.460000 pts: 0.460000 pos:  48793 size:  3840
 ret: 0         st:-1 flags:1  ts:-0.645825
-ret: 0         st: 0 flags:1 dts: 0.000000 pts: 0.000000 pos:   5145 size:  3840
+ret: 0         st: 0 flags:1 dts: 0.000000 pts: 0.000000 pos:   4633 size:  3840

Overview

Request 2278 (accepted)

Request History