Merge remote-tracking branch 'qatar/master'

* qatar/master: (26 commits)
  avconv: deprecate the -deinterlace option
  doc: Fix the name of the new function
  aacenc: make sure to encode enough frames to cover all input samples.
  aacenc: only use the number of input samples provided by the user.
  wmadec: Verify bitstream size makes sense before calling init_get_bits.
  kmvc: Log into a context at a log level constant.
  mpeg12: Pad framerate tab to 16 entries.
  kgv1dec: Increase offsets array size so it is large enough.
  kmvc: Check palsize.
  nsvdec: Propagate errors
  nsvdec: Be more careful with av_malloc().
  nsvdec: Fix use of uninitialized streams.
  movenc: cosmetics: Get rid of camelCase identifiers
  swscale: more generic check for planar destination formats with alpha
  doc: Document mov/mp4 fragmentation options
  build: Use order-only prerequisites for creating FATE reference file dirs.
  x86 dsputil: provide SSE2/SSSE3 versions of bswap_buf
  rtsp: Remove some unused variables from ff_rtsp_connect().
  avutil: make intfloat api public
  avformat_write_header(): detail error message
  ...

Conflicts:
	doc/APIchanges
	doc/ffmpeg.texi
	doc/muxers.texi
	ffmpeg.c
	libavcodec/kmvc.c
	libavcodec/x86/Makefile
	libavcodec/x86/dsputil_yasm.asm
	libavcodec/x86/pngdsp-init.c
	libavformat/movenc.c
	libavformat/movenc.h
	libavformat/mpegtsenc.c
	libavformat/nsvdec.c
	libavformat/utils.c
	libavutil/avutil.h
	libswscale/swscale.c
Merged-by: Michael Niedermayer <michaelni@gmx.at>

doc/APIchanges

@@ -16,6 +16,10 @@ API changes, most recent first:
 2012-01-24 - xxxxxxx - lavfi 2.60.100
   Add avfilter_graph_dump.
 
+2012-01-30 - xxxxxxx - lavu 51.22.0 - intfloat.h
+  Add a new installed header libavutil/intfloat.h with int/float punning
+  functions.
+
 2012-01-25 - lavf 53.22.0
   f1caf01 Allow doing av_write_frame(ctx, NULL) for flushing possible
           buffered data within a muxer. Added AVFMT_ALLOW_FLUSH for
@@ -109,7 +113,7 @@ API changes, most recent first:
   Add av_toupper()/av_tolower()
 
 2011-11-05 - b6d08f4 - lavf 53.13.0
-  Add avformat_network_init()/avformat_network_uninit()
+  Add avformat_network_init()/avformat_network_deinit()
 
 2011-10-27 - 512557b - lavc 53.15.0
   Remove avcodec_parse_frame.

doc/ffmpeg.texi

@@ -620,6 +620,8 @@ Enable Unlimited Motion Vector (h263+)
 
 @item -deinterlace
 Deinterlace pictures.
+This option is deprecated since the deinterlacing is very low quality.
+Use the yadif filter with @code{-filter:v yadif}.
 @item -ilme
 Force interlacing support in encoder (MPEG-2 and MPEG-4 only).
 Use this option if your input file is interlaced and you want

doc/muxers.texi

@@ -148,18 +148,70 @@ each of the YUV420P components. To read or write this image file format,
 specify the name of the '.Y' file. The muxer will automatically open the
 '.U' and '.V' files as required.
 
-@section mov
-
-MOV / MP4 muxer
-
-The muxer options are:
+@section MOV/MP4/ISMV
+
+The mov/mp4/ismv muxer supports fragmentation. Normally, a MOV/MP4
+file has all the metadata about all packets stored in one location
+(written at the end of the file, it can be moved to the start for
+better playback using the @command{qt-faststart} tool). A fragmented
+file consists of a number of fragments, where packets and metadata
+about these packets are stored together. Writing a fragmented
+file has the advantage that the file is decodable even if the
+writing is interrupted (while a normal MOV/MP4 is undecodable if
+it is not properly finished), and it requires less memory when writing
+very long files (since writing normal MOV/MP4 files stores info about
+every single packet in memory until the file is closed). The downside
+is that it is less compatible with other applications.
+
+Fragmentation is enabled by setting one of the AVOptions that define
+how to cut the file into fragments:
 
 @table @option
 @item -moov_size @var{bytes}
 Reserves space for the moov atom at the beginning of the file instead of placing the
 moov atom at the end. If the space reserved is insufficient, muxing will fail.
+@item -movflags frag_keyframe
+Start a new fragment at each video keyframe.
+@item -frag_duration @var{duration}
+Create fragments that are @var{duration} microseconds long.
+@item -frag_size @var{size}
+Create fragments that contain up to @var{size} bytes of payload data.
+@item -movflags frag_custom
+Allow the caller to manually choose when to cut fragments, by
+calling @code{av_write_frame(ctx, NULL)} to write a fragment with
+the packets written so far. (This is only useful with other
+applications integrating libavformat, not from @command{avconv}.)
+@end table
+
+Additionally, the way the output file is written can be adjusted
+through a few other options:
+
+@table @option
+@item -movflags empty_moov
+Write an initial moov atom directly at the start of the file, without
+describing any samples in it. Generally, an mdat/moov pair is written
+at the start of the file, as a normal MOV/MP4 file, containing only
+a short portion of the file. With this option set, there is no initial
+mdat atom, and the moov atom only describes the tracks but has
+a zero duration.
+
+Files written with this option set do not work in QuickTime.
+This option is implicitly set when writing ismv (Smooth Streaming) files.
+@item -movflags separate_moof
+Write a separate moof (movie fragment) atom for each track. Normally,
+packets for all tracks are written in a moof atom (which is slightly
+more efficient), but with this option set, the muxer writes one moof/mdat
+pair for each track, making it easier to separate tracks.
+
+This option is implicitly set when writing ismv (Smooth Streaming) files.
 @end table
 
+Smooth Streaming content can be pushed in real time to a publishing
+point on IIS with this muxer. Example:
+@example
+avconv -re @var{<normal input/transcoding options>} -movflags isml+frag_keyframe -f ismv http://server/publishingpoint.isml/Streams(Encoder1)
+@end example
+
 @section mpegts
 
 MPEG transport stream muxer.

ffmpeg.c

@@ -471,7 +471,7 @@ static void reset_options(OptionsContext *o, int is_input)
 static int alloc_buffer(AVCodecContext *s, InputStream *ist, FrameBuffer **pbuf)
 {
     FrameBuffer  *buf = av_mallocz(sizeof(*buf));
-    int ret, i;
+    int i, ret;
     const int pixel_size = av_pix_fmt_descriptors[s->pix_fmt].comp[0].step_minus1+1;
     int h_chroma_shift, v_chroma_shift;
     int edge = 32; // XXX should be avcodec_get_edge_width(), but that fails on svq1
@@ -4939,6 +4939,13 @@ static int opt_vsync(const char *opt, const char *arg)
     return 0;
 }
 
+static int opt_deinterlace(const char *opt, const char *arg)
+{
+    av_log(NULL, AV_LOG_WARNING, "-%s is deprecated, use -filter:v yadif instead\n", opt);
+    do_deinterlace = 1;
+    return 0;
+}
+
 #define OFFSET(x) offsetof(OptionsContext, x)
 static const OptionDef options[] = {
     /* main options */
@@ -5018,8 +5025,8 @@ static const OptionDef options[] = {
       "use same quantizer as source (implies VBR)" },
     { "pass", HAS_ARG | OPT_VIDEO, {(void*)opt_pass}, "select the pass number (1 or 2)", "n" },
     { "passlogfile", HAS_ARG | OPT_VIDEO, {(void*)&opt_passlogfile}, "select two pass log file name prefix", "prefix" },
-    { "deinterlace", OPT_BOOL | OPT_EXPERT | OPT_VIDEO, {(void*)&do_deinterlace},
-      "deinterlace pictures" },
+    { "deinterlace", OPT_EXPERT | OPT_VIDEO, {(void*)opt_deinterlace},
+      "this option is deprecated, use the yadif filter instead" },
     { "psnr", OPT_BOOL | OPT_EXPERT | OPT_VIDEO, {(void*)&do_psnr}, "calculate PSNR of compressed frames" },
     { "vstats", OPT_EXPERT | OPT_VIDEO, {(void*)&opt_vstats}, "dump video coding statistics to file" },
     { "vstats_file", HAS_ARG | OPT_EXPERT | OPT_VIDEO, {(void*)opt_vstats_file}, "dump video coding statistics to file", "file" },

libavcodec/aacenc.c

@@ -223,8 +223,9 @@ WINDOW_FUNC(eight_short)
     const float *pwindow = sce->ics.use_kb_window[1] ? ff_aac_kbd_short_128 : ff_sine_128;
     const float *in = audio + 448;
     float *out = sce->ret;
+    int w;
 
-    for (int w = 0; w < 8; w++) {
+    for (w = 0; w < 8; w++) {
         dsp->vector_fmul        (out, in, w ? pwindow : swindow, 128);
         out += 128;
         in  += 128;
@@ -476,7 +477,7 @@ static void put_bitstream_info(AVCodecContext *avctx, AACEncContext *s,
  * Channels are reordered from Libav's default order to AAC order.
  */
 static void deinterleave_input_samples(AACEncContext *s,
-                                       const float *samples)
+                                       const float *samples, int nb_samples)
 {
     int ch, i;
     const int sinc = s->channels;
@@ -490,10 +491,12 @@ static void deinterleave_input_samples(AACEncContext *s,
         memcpy(&s->planar_samples[ch][1024], &s->planar_samples[ch][2048], 1024 * sizeof(s->planar_samples[0][0]));
 
         /* deinterleave */
-        for (i = 2048; i < 3072; i++) {
+        for (i = 2048; i < 2048 + nb_samples; i++) {
             s->planar_samples[ch][i] = *sptr;
             sptr += sinc;
         }
+        memset(&s->planar_samples[ch][i], 0,
+               (3072 - i) * sizeof(s->planar_samples[0][0]));
     }
 }
 
@@ -507,14 +510,12 @@ static int aac_encode_frame(AVCodecContext *avctx,
     int chan_el_counter[4];
     FFPsyWindowInfo windows[AAC_MAX_CHANNELS];
 
-    if (s->last_frame)
+    if (s->last_frame == 2)
         return 0;
 
-    if (data) {
-        deinterleave_input_samples(s, data);
+    deinterleave_input_samples(s, data, data ? avctx->frame_size : 0);
     if (s->psypp)
         ff_psy_preprocess(s->psypp, s->planar_samples, s->channels);
-    }
 
     if (!avctx->frame_number)
         return 0;
@@ -645,7 +646,7 @@ static int aac_encode_frame(AVCodecContext *avctx,
     }
 
     if (!data)
-        s->last_frame = 1;
+        s->last_frame++;
 
     return put_bits_count(&s->pb)>>3;
 }
@@ -686,11 +687,12 @@ static av_cold int dsp_init(AVCodecContext *avctx, AACEncContext *s)
 
 static av_cold int alloc_buffers(AVCodecContext *avctx, AACEncContext *s)
 {
+    int ch;
     FF_ALLOCZ_OR_GOTO(avctx, s->buffer.samples, 3 * 1024 * s->channels * sizeof(s->buffer.samples[0]), alloc_fail);
     FF_ALLOCZ_OR_GOTO(avctx, s->cpe, sizeof(ChannelElement) * s->chan_map[0], alloc_fail);
     FF_ALLOCZ_OR_GOTO(avctx, avctx->extradata, 5 + FF_INPUT_BUFFER_PADDING_SIZE, alloc_fail);
 
-    for(int ch = 0; ch < s->channels; ch++)
+    for(ch = 0; ch < s->channels; ch++)
         s->planar_samples[ch] = s->buffer.samples + 3 * 1024 * ch;
 
     return 0;

libavcodec/kmvc.c

@@ -33,6 +33,7 @@
 #define KMVC_KEYFRAME 0x80
 #define KMVC_PALETTE  0x40
 #define KMVC_METHOD   0x0F
+#define MAX_PALSIZE   256
 
 /*
  * Decoder context
@@ -43,7 +44,7 @@ typedef struct KmvcContext {
 
     int setpal;
     int palsize;
-    uint32_t pal[256];
+    uint32_t pal[MAX_PALSIZE];
     uint8_t *cur, *prev;
     uint8_t *frm0, *frm1;
     GetByteContext g;
@@ -376,14 +377,15 @@ static av_cold int decode_init(AVCodecContext * avctx)
     }
 
     if (avctx->extradata_size < 12) {
-        av_log(NULL, 0, "Extradata missing, decoding may not work properly...\n");
+        av_log(avctx, AV_LOG_WARNING,
+               "Extradata missing, decoding may not work properly...\n");
         c->palsize = 127;
     } else {
         c->palsize = AV_RL16(avctx->extradata + 10);
-        if (c->palsize > 255U) {
+        if (c->palsize >= (unsigned)MAX_PALSIZE) {
             c->palsize = 127;
-            av_log(NULL, AV_LOG_ERROR, "palsize too big\n");
-            return -1;
+            av_log(avctx, AV_LOG_ERROR, "KMVC palette too large\n");
+            return AVERROR_INVALIDDATA;
         }
     }
 

libavcodec/mpc7.c

@@ -53,7 +53,7 @@ static av_cold int mpc7_decode_init(AVCodecContext * avctx)
     int i, j;
     MPCContext *c = avctx->priv_data;
     GetBitContext gb;
-    uint8_t buf[16];
+    LOCAL_ALIGNED_16(uint8_t, buf, [16]);
     static int vlc_initialized = 0;
 
     static VLC_TYPE scfi_table[1 << MPC7_SCFI_BITS][2];

libavcodec/mpeg12data.c

@@ -305,7 +305,7 @@ const uint8_t ff_mpeg12_mbMotionVectorTable[17][2] = {
 { 0xc, 10 },
 };
 
-const AVRational avpriv_frame_rate_tab[] = {
+const AVRational avpriv_frame_rate_tab[16] = {
     {    0,    0},
     {24000, 1001},
     {   24,    1},

libavcodec/wmadec.c

@@ -891,6 +891,8 @@ static int wma_decode_superframe(AVCodecContext *avctx, void *data,
 
         /* read each frame starting from bit_offset */
         pos = bit_offset + 4 + 4 + s->byte_offset_bits + 3;
+        if (pos >= MAX_CODED_SUPERFRAME_SIZE * 8)
+            return AVERROR_INVALIDDATA;
         init_get_bits(&s->gb, buf + (pos >> 3), (MAX_CODED_SUPERFRAME_SIZE - (pos >> 3))*8);
         len = pos & 7;
         if (len > 0)

libavcodec/x86/Makefile

@@ -48,6 +48,7 @@ MMX-OBJS-$(CONFIG_GPL)                 += x86/idct_mmx.o
 MMX-OBJS-$(CONFIG_LPC)                 += x86/lpc_mmx.o
 YASM-OBJS-$(CONFIG_PRORES_LGPL_DECODER)     += x86/proresdsp.o
 MMX-OBJS-$(CONFIG_PRORES_LGPL_DECODER)      += x86/proresdsp-init.o
+YASM-OBJS-$(CONFIG_PNG_DECODER)        += x86/pngdsp.o
 MMX-OBJS-$(CONFIG_PNG_DECODER)         += x86/pngdsp-init.o
 YASM-OBJS-$(CONFIG_PRORES_DECODER)     += x86/proresdsp.o
 MMX-OBJS-$(CONFIG_PRORES_DECODER)      += x86/proresdsp-init.o

libavcodec/x86/dsputil_mmx.c

@@ -2392,6 +2392,9 @@ void ff_apply_window_int16_ssse3     (int16_t *output, const int16_t *input,
 void ff_apply_window_int16_ssse3_atom(int16_t *output, const int16_t *input,
                                       const int16_t *window, unsigned int len);
 
+void ff_bswap32_buf_ssse3(uint32_t *dst, const uint32_t *src, int w);
+void ff_bswap32_buf_sse2(uint32_t *dst, const uint32_t *src, int w);
+
 void ff_add_hfyu_median_prediction_mmx2(uint8_t *dst, const uint8_t *top, const uint8_t *diff, int w, int *left, int *left_top);
 int  ff_add_hfyu_left_prediction_ssse3(uint8_t *dst, const uint8_t *src, int w, int left);
 int  ff_add_hfyu_left_prediction_sse4(uint8_t *dst, const uint8_t *src, int w, int left);
@@ -2880,6 +2883,7 @@ void dsputil_init_mmx(DSPContext* c, AVCodecContext *avctx)
                     c->apply_window_int16 = ff_apply_window_int16_sse2;
                 }
             }
+            c->bswap_buf = ff_bswap32_buf_sse2;
 #endif
         }
         if (mm_flags & AV_CPU_FLAG_SSSE3) {
@@ -2892,6 +2896,7 @@ void dsputil_init_mmx(DSPContext* c, AVCodecContext *avctx)
             if (!(mm_flags & (AV_CPU_FLAG_SSE42|AV_CPU_FLAG_3DNOW))) { // cachesplit
                 c->scalarproduct_and_madd_int16 = ff_scalarproduct_and_madd_int16_ssse3;
             }
+            c->bswap_buf = ff_bswap32_buf_ssse3;
 #endif
         }
 

libavcodec/x86/dsputil_yasm.asm

@@ -30,6 +30,7 @@ pb_zzzz3333zzzzbbbb: db -1,-1,-1,-1,3,3,3,3,-1,-1,-1,-1,11,11,11,11
 pb_zz11zz55zz99zzdd: db -1,-1,1,1,-1,-1,5,5,-1,-1,9,9,-1,-1,13,13
 pb_revwords: db 14, 15, 12, 13, 10, 11, 8, 9, 6, 7, 4, 5, 2, 3, 0, 1
 pd_16384: times 4 dd 16384
+pb_bswap32: db 3, 2, 1, 0, 7, 6, 5, 4, 11, 10, 9, 8, 15, 14, 13, 12
 
 SECTION_TEXT
 
@@ -1180,3 +1181,126 @@ BUTTERFLIES_FLOAT_INTERLEAVE
 INIT_YMM avx
 BUTTERFLIES_FLOAT_INTERLEAVE
 %endif
+
+INIT_XMM sse2
+; %1 = aligned/unaligned
+%macro BSWAP_LOOPS_SSE2  1
+    mov      r3, r2
+    sar      r2, 3
+    jz       .left4_%1
+.loop8_%1:
+    mov%1    m0, [r1 +  0]
+    mov%1    m1, [r1 + 16]
+    pshuflw  m0, m0, 10110001b
+    pshuflw  m1, m1, 10110001b
+    pshufhw  m0, m0, 10110001b
+    pshufhw  m1, m1, 10110001b
+    mova     m2, m0
+    mova     m3, m1
+    psllw    m0, 8
+    psllw    m1, 8
+    psrlw    m2, 8
+    psrlw    m3, 8
+    por      m2, m0
+    por      m3, m1
+    mova     [r0 +  0], m2
+    mova     [r0 + 16], m3
+    add      r1, 32
+    add      r0, 32
+    dec      r2
+    jnz      .loop8_%1
+.left4_%1:
+    mov      r2, r3
+    and      r3, 4
+    jz       .left
+    mov%1    m0, [r1]
+    pshuflw  m0, m0, 10110001b
+    pshufhw  m0, m0, 10110001b
+    mova     m2, m0
+    psllw    m0, 8
+    psrlw    m2, 8
+    por      m2, m0
+    mova     [r0], m2
+    add      r1, 16
+    add      r0, 16
+%endmacro
+
+; void bswap_buf(uint32_t *dst, const uint32_t *src, int w);
+cglobal bswap32_buf, 3,4,5
+    mov      r3, r1
+    and      r3, 15
+    jz       .start_align
+    BSWAP_LOOPS_SSE2  u
+    jmp      .left
+.start_align:
+    BSWAP_LOOPS_SSE2  a
+.left:
+    and      r2, 3
+    jz       .end
+.loop2:
+    mov      r3d, [r1]
+    bswap    r3d
+    mov      [r0], r3d
+    add      r1, 4
+    add      r0, 4
+    dec      r2
+    jnz      .loop2
+.end
+    RET
+
+; %1 = aligned/unaligned
+%macro BSWAP_LOOPS_SSSE3  1
+    mov      r3, r2
+    sar      r2, 3
+    jz       .left4_%1
+.loop8_%1:
+    mov%1    m0, [r1 +  0]
+    mov%1    m1, [r1 + 16]
+    pshufb   m0, m2
+    pshufb   m1, m2
+    mova     [r0 +  0], m0
+    mova     [r0 + 16], m1
+    add      r0, 32
+    add      r1, 32
+    dec      r2
+    jnz      .loop8_%1
+.left4_%1:
+    mov      r2, r3
+    and      r3, 4
+    jz       .left2
+    mov%1    m0, [r1]
+    pshufb   m0, m2
+    mova     [r0], m0
+    add      r1, 16
+    add      r0, 16
+%endmacro
+
+INIT_XMM ssse3
+; void bswap_buf(uint32_t *dst, const uint32_t *src, int w);
+cglobal bswap32_buf, 3,4,3
+    mov      r3, r1
+    mova     m2, [pb_bswap32]
+    and      r3, 15
+    jz       .start_align
+    BSWAP_LOOPS_SSSE3  u
+    jmp      .left2
+.start_align:
+    BSWAP_LOOPS_SSSE3  a
+.left2:
+    mov      r3, r2
+    and      r2, 2
+    jz       .left1
+    movq     m0, [r1]
+    pshufb   m0, m2
+    movq     [r0], m0
+    add      r1, 8
+    add      r0, 8
+.left1:
+    and      r3, 1
+    jz       .end
+    mov      r2d, [r1]
+    bswap    r2d
+    mov      [r0], r2d
+.end:
+    RET
+

libavcodec/x86/pngdsp-init.c

@@ -19,118 +19,33 @@
  * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
  */
 
+#include "libavutil/common.h"
 #include "libavutil/cpu.h"
-#include "libavutil/x86_cpu.h"
 #include "libavcodec/pngdsp.h"
-#include "dsputil_mmx.h"
 
-#define PAETH(cpu, abs3)\
-static void add_png_paeth_prediction_##cpu(uint8_t *dst, uint8_t *src, uint8_t *top, int w, int bpp)\
-{\
-    x86_reg i, end;\
-    if(bpp>4) add_png_paeth_prediction_##cpu(dst+bpp/2, src+bpp/2, top+bpp/2, w-bpp/2, -bpp);\
-    if(bpp<0) bpp=-bpp;\
-    i= -bpp;\
-    end = w-3;\
-    __asm__ volatile(\
-        "pxor      %%mm7, %%mm7 \n"\
-        "movd    (%1,%0), %%mm0 \n"\
-        "movd    (%2,%0), %%mm1 \n"\
-        "punpcklbw %%mm7, %%mm0 \n"\
-        "punpcklbw %%mm7, %%mm1 \n"\
-        "add       %4, %0 \n"\
-        "1: \n"\
-        "movq      %%mm1, %%mm2 \n"\
-        "movd    (%2,%0), %%mm1 \n"\
-        "movq      %%mm2, %%mm3 \n"\
-        "punpcklbw %%mm7, %%mm1 \n"\
-        "movq      %%mm2, %%mm4 \n"\
-        "psubw     %%mm1, %%mm3 \n"\
-        "psubw     %%mm0, %%mm4 \n"\
-        "movq      %%mm3, %%mm5 \n"\
-        "paddw     %%mm4, %%mm5 \n"\
-        abs3\
-        "movq      %%mm4, %%mm6 \n"\
-        "pminsw    %%mm5, %%mm6 \n"\
-        "pcmpgtw   %%mm6, %%mm3 \n"\
-        "pcmpgtw   %%mm5, %%mm4 \n"\
-        "movq      %%mm4, %%mm6 \n"\
-        "pand      %%mm3, %%mm4 \n"\
-        "pandn     %%mm3, %%mm6 \n"\
-        "pandn     %%mm0, %%mm3 \n"\
-        "movd    (%3,%0), %%mm0 \n"\
-        "pand      %%mm1, %%mm6 \n"\
-        "pand      %%mm4, %%mm2 \n"\
-        "punpcklbw %%mm7, %%mm0 \n"\
-        "paddw     %%mm6, %%mm0 \n"\
-        "paddw     %%mm2, %%mm3 \n"\
-        "paddw     %%mm3, %%mm0 \n"\
-        "pand      %6   , %%mm0 \n"\
-        "movq      %%mm0, %%mm3 \n"\
-        "packuswb  %%mm3, %%mm3 \n"\
-        "movd      %%mm3, (%1,%0) \n"\
-        "add       %4, %0 \n"\
-        "cmp       %5, %0 \n"\
-        "jle 1b \n"\
-        :"+r"(i)\
-        :"r"(dst), "r"(top), "r"(src), "r"((x86_reg)bpp), "g"(end),\
-         "m"(ff_pw_255)\
-        :"memory"\
-    );\
-}
-
-#define ABS3_MMX2\
-        "psubw     %%mm5, %%mm7 \n"\
-        "pmaxsw    %%mm7, %%mm5 \n"\
-        "pxor      %%mm6, %%mm6 \n"\
-        "pxor      %%mm7, %%mm7 \n"\
-        "psubw     %%mm3, %%mm6 \n"\
-        "psubw     %%mm4, %%mm7 \n"\
-        "pmaxsw    %%mm6, %%mm3 \n"\
-        "pmaxsw    %%mm7, %%mm4 \n"\
-        "pxor      %%mm7, %%mm7 \n"
-
-#define ABS3_SSSE3\
-        "pabsw     %%mm3, %%mm3 \n"\
-        "pabsw     %%mm4, %%mm4 \n"\
-        "pabsw     %%mm5, %%mm5 \n"
-
-PAETH(mmx2, ABS3_MMX2)
-#if HAVE_SSSE3
-PAETH(ssse3, ABS3_SSSE3)
-#endif
-
-static void add_bytes_l2_mmx(uint8_t *dst, uint8_t *src1, uint8_t *src2, int w)
-{
-    x86_reg i=0;
-    __asm__ volatile(
-        "jmp 2f                         \n\t"
-        "1:                             \n\t"
-        "movq   (%2, %0), %%mm0         \n\t"
-        "movq  8(%2, %0), %%mm1         \n\t"
-        "paddb  (%3, %0), %%mm0         \n\t"
-        "paddb 8(%3, %0), %%mm1         \n\t"
-        "movq %%mm0,  (%1, %0)          \n\t"
-        "movq %%mm1, 8(%1, %0)          \n\t"
-        "add $16, %0                    \n\t"
-        "2:                             \n\t"
-        "cmp %4, %0                     \n\t"
-        " js 1b                         \n\t"
-        : "+r" (i)
-        : "r"(dst), "r"(src1), "r"(src2), "r"((x86_reg) w - 15)
-    );
-    for (; i < w; i++)
-        dst[i] = src1[i] + src2[i];
-}
+void ff_add_png_paeth_prediction_mmx2 (uint8_t *dst, uint8_t *src,
+                                       uint8_t *top, int w, int bpp);
+void ff_add_png_paeth_prediction_ssse3(uint8_t *dst, uint8_t *src,
+                                       uint8_t *top, int w, int bpp);
+void ff_add_bytes_l2_mmx (uint8_t *dst, uint8_t *src1,
+                          uint8_t *src2, int w);
+void ff_add_bytes_l2_sse2(uint8_t *dst, uint8_t *src1,
+                          uint8_t *src2, int w);
 
 void ff_pngdsp_init_x86(PNGDSPContext *dsp)
 {
+#if HAVE_YASM
     int flags = av_get_cpu_flags();
 
+#if ARCH_X86_32
     if (flags & AV_CPU_FLAG_MMX)
-        dsp->add_bytes_l2         = add_bytes_l2_mmx;
+        dsp->add_bytes_l2         = ff_add_bytes_l2_mmx;
+#endif
     if (flags & AV_CPU_FLAG_MMX2)
-        dsp->add_paeth_prediction = add_png_paeth_prediction_mmx2;
-    if (HAVE_SSSE3 && flags & AV_CPU_FLAG_SSSE3)
-        dsp->add_paeth_prediction = add_png_paeth_prediction_ssse3;
+        dsp->add_paeth_prediction = ff_add_png_paeth_prediction_mmx2;
+    if (flags & AV_CPU_FLAG_SSE2)
+        dsp->add_bytes_l2         = ff_add_bytes_l2_sse2;
+    if (flags & AV_CPU_FLAG_SSSE3)
+        dsp->add_paeth_prediction = ff_add_png_paeth_prediction_ssse3;
+#endif
 }

libavcodec/x86/pngdsp.asm

+;******************************************************************************
+;* x86 optimizations for PNG decoding
+;*
+;* Copyright (c) 2008 Loren Merritt <lorenm@u.washington.edu>
+;* Copyright (c) 2012 Ronald S. Bultje <rsbultje@gmail.com>
+;*
+;* This file is part of Libav.
+;*
+;* Libav is free software; you can redistribute it and/or
+;* modify it under the terms of the GNU Lesser General Public
+;* License as published by the Free Software Foundation; either
+;* version 2.1 of the License, or (at your option) any later version.
+;*
+;* Libav is distributed in the hope that it will be useful,
+;* but WITHOUT ANY WARRANTY; without even the implied warranty of
+;* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+;* Lesser General Public License for more details.
+;*
+;* You should have received a copy of the GNU Lesser General Public
+;* License along with Libav; if not, write to the Free Software
+;* 51, Inc., Foundation Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+;******************************************************************************
+
+%include "x86inc.asm"
+%include "x86util.asm"
+
+SECTION_RODATA
+
+cextern pw_255
+
+section .text align=16
+
+; %1 = nr. of xmm registers used
+%macro ADD_BYTES_FN 1
+cglobal add_bytes_l2, 4, 6, %1, dst, src1, src2, wa, w, i
+%if ARCH_X86_64
+    movsxd             waq, wad
+%endif
+    xor                 iq, iq
+
+    ; vector loop
+    mov                 wq, waq
+    and                waq, ~(mmsize*2-1)
+    jmp .end_v
+.loop_v:
+    mova                m0, [src1q+iq]
+    mova                m1, [src1q+iq+mmsize]
+    paddb               m0, [src2q+iq]
+    paddb               m1, [src2q+iq+mmsize]
+    mova  [dstq+iq       ], m0
+    mova  [dstq+iq+mmsize], m1
+    add                 iq, mmsize*2
+.end_v:
+    cmp                 iq, waq
+    jl .loop_v
+
+%if mmsize == 16
+    ; vector loop
+    mov                 wq, waq
+    and                waq, ~7
+    jmp .end_l
+.loop_l:
+    movq               mm0, [src1q+iq]
+    paddb              mm0, [src2q+iq]
+    movq  [dstq+iq       ], mm0
+    add                 iq, 8
+.end_l:
+    cmp                 iq, waq
+    jl .loop_l
+%endif
+
+    ; scalar loop for leftover
+    jmp .end_s
+.loop_s:
+    mov                wab, [src1q+iq]
+    add                wab, [src2q+iq]
+    mov          [dstq+iq], wab
+    inc                 iq
+.end_s:
+    cmp                 iq, wq
+    jl .loop_s
+    REP_RET
+%endmacro
+
+%if ARCH_X86_32
+INIT_MMX mmx
+ADD_BYTES_FN 0
+%endif
+
+INIT_XMM sse2
+ADD_BYTES_FN 2
+
+%macro ADD_PAETH_PRED_FN 1
+cglobal add_png_paeth_prediction, 5, 7, %1, dst, src, top, w, bpp, end, cntr
+%if ARCH_X86_64
+    movsxd            bppq, bppd
+    movsxd              wq, wd
+%endif
+    lea               endq, [dstq+wq-(mmsize/2-1)]
+    sub               topq, dstq
+    sub               srcq, dstq
+    sub               dstq, bppq
+    pxor                m7, m7
+
+    PUSH              dstq
+    lea              cntrq, [bppq-1]
+    shr              cntrq, 2 + mmsize/16
+.bpp_loop:
+    lea               dstq, [dstq+cntrq*(mmsize/2)]
+    movh                m0, [dstq]
+    movh                m1, [topq+dstq]
+    punpcklbw           m0, m7
+    punpcklbw           m1, m7
+    add               dstq, bppq
+.loop:
+    mova                m2, m1
+    movh                m1, [topq+dstq]
+    mova                m3, m2
+    punpcklbw           m1, m7
+    mova                m4, m2
+    psubw               m3, m1
+    psubw               m4, m0
+    mova                m5, m3
+    paddw               m5, m4
+%if cpuflag(ssse3)
+    pabsw               m3, m3
+    pabsw               m4, m4
+    pabsw               m5, m5
+%else ; !cpuflag(ssse3)
+    psubw               m7, m5
+    pmaxsw              m5, m7
+    pxor                m6, m6
+    pxor                m7, m7
+    psubw               m6, m3
+    psubw               m7, m4
+    pmaxsw              m3, m6
+    pmaxsw              m4, m7
+    pxor                m7, m7
+%endif ; cpuflag(ssse3)
+    mova                m6, m4
+    pminsw              m6, m5
+    pcmpgtw             m3, m6
+    pcmpgtw             m4, m5
+    mova                m6, m4
+    pand                m4, m3
+    pandn               m6, m3
+    pandn               m3, m0
+    movh                m0, [srcq+dstq]
+    pand                m6, m1
+    pand                m2, m4
+    punpcklbw           m0, m7
+    paddw               m0, m6
+    paddw               m3, m2
+    paddw               m0, m3
+    pand                m0, [pw_255]
+    mova                m3, m0
+    packuswb            m3, m3
+    movh            [dstq], m3
+    add               dstq, bppq
+    cmp               dstq, endq
+    jle .loop
+
+    mov               dstq, [rsp]
+    dec              cntrq
+    jge .bpp_loop
+    POP               dstq
+    RET
+%endmacro
+
+INIT_MMX mmx2
+ADD_PAETH_PRED_FN 0
+
+INIT_MMX ssse3
+ADD_PAETH_PRED_FN 0

libavformat/movenc.h

@@ -44,7 +44,7 @@ typedef struct MOVIentry {
     uint64_t     pos;
     int64_t      dts;
     unsigned int size;
-    unsigned int samplesInChunk;
+    unsigned int samples_in_chunk;
     unsigned int chunkNum;              ///< Chunk number if the current entry is a chunk start otherwise 0
     unsigned int entries;
     int          cts;
@@ -79,21 +79,21 @@ typedef struct MOVIndex {
     int         entry;
     unsigned    timescale;
     uint64_t    time;
-    int64_t     trackDuration;
-    long        sampleCount;
-    long        sampleSize;
+    int64_t     track_duration;
+    long        sample_count;
+    long        sample_size;
     long        chunkCount;
-    int         hasKeyframes;
+    int         has_keyframes;
 #define MOV_TRACK_CTTS         0x0001
 #define MOV_TRACK_STPS         0x0002
     uint32_t    flags;
     int         language;
-    int         trackID;
+    int         track_id;
     int         tag; ///< stsd fourcc
     AVCodecContext *enc;
 
-    int         vosLen;
-    uint8_t     *vosData;
+    int         vos_len;
+    uint8_t     *vos_data;
     MOVIentry   *cluster;
     int         audio_vbr;
     int         height; ///< active picture (w/o VBI) height for D-10/IMX

libavformat/mpegtsenc.c

@@ -991,7 +991,7 @@ static int mpegts_write_packet(AVFormatContext *s, AVPacket *pkt)
 
         if (pkt->size < 5 || AV_RB32(pkt->data) != 0x0000001) {
             av_log(s, AV_LOG_ERROR, "H.264 bitstream malformed, "
-                   "no startcode found, use the h264_mp4toannexb bitstream filter\n");
+                   "no startcode found, use the h264_mp4toannexb bitstream filter (-bsf h264_mp4toannexb)\n");
             return AVERROR_INVALIDDATA;
         }
 

libavformat/nsvdec.c

@@ -313,7 +313,7 @@ static int nsv_parse_NSVf_header(AVFormatContext *s)
         char *token, *value;
         char quote;
 
-        p = strings = av_mallocz(strings_size + 1);
+        p = strings = av_mallocz((size_t)strings_size + 1);
         if (!p)
             return AVERROR(ENOMEM);
         endp = strings + strings_size;
@@ -350,6 +350,8 @@ static int nsv_parse_NSVf_header(AVFormatContext *s)
         if((unsigned)table_entries_used >= UINT_MAX / sizeof(uint32_t))
             return -1;
         nsv->nsvs_file_offset = av_malloc((unsigned)table_entries_used * sizeof(uint32_t));
+        if (!nsv->nsvs_file_offset)
+            return AVERROR(ENOMEM);
 
         for(i=0;i<table_entries_used;i++)
             nsv->nsvs_file_offset[i] = avio_rl32(pb) + size;
@@ -357,6 +359,8 @@ static int nsv_parse_NSVf_header(AVFormatContext *s)
         if(table_entries > table_entries_used &&
            avio_rl32(pb) == MKTAG('T','O','C','2')) {
             nsv->nsvs_timestamps = av_malloc((unsigned)table_entries_used*sizeof(uint32_t));
+            if (!nsv->nsvs_timestamps)
+                return AVERROR(ENOMEM);
             for(i=0;i<table_entries_used;i++) {
                 nsv->nsvs_timestamps[i] = avio_rl32(pb);
             }
@@ -527,11 +531,16 @@ static int nsv_read_header(AVFormatContext *s)
     for (i = 0; i < NSV_MAX_RESYNC_TRIES; i++) {
         if (nsv_resync(s) < 0)
             return -1;
-        if (nsv->state == NSV_FOUND_NSVF)
+        if (nsv->state == NSV_FOUND_NSVF) {
             err = nsv_parse_NSVf_header(s);
+            if (err < 0)
+                return err;
+        }
             /* we need the first NSVs also... */
         if (nsv->state == NSV_FOUND_NSVS) {
             err = nsv_parse_NSVs_header(s);
+            if (err < 0)
+                return err;
             break; /* we just want the first one */
         }
     }

libavformat/rtsp.c

@@ -1352,7 +1352,6 @@ int ff_rtsp_connect(AVFormatContext *s)
 {
     RTSPState *rt = s->priv_data;
     char host[1024], path[1024], tcpname[1024], cmd[2048], auth[128];
-    char *option_list, *option, *filename;
     int port, err, tcp_fd;
     RTSPMessageHeader reply1 = {0}, *reply = &reply1;
     int lower_transport_mask = 0;

libavformat/utils.c

@@ -2988,7 +2988,11 @@ int avformat_write_header(AVFormatContext *s, AVDictionary **options)
             if(av_cmp_q(st->sample_aspect_ratio, st->codec->sample_aspect_ratio)
                && FFABS(av_q2d(st->sample_aspect_ratio) - av_q2d(st->codec->sample_aspect_ratio)) > 0.004*av_q2d(st->sample_aspect_ratio)
             ){
-                av_log(s, AV_LOG_ERROR, "Aspect ratio mismatch between encoder and muxer layer\n");
+                av_log(s, AV_LOG_ERROR, "Aspect ratio mismatch between encoder "
+                       "(%d/%d) and muxer layer (%d/%d)\n",
+                       st->sample_aspect_ratio.num, st->sample_aspect_ratio.den,
+                       st->codec->sample_aspect_ratio.num,
+                       st->codec->sample_aspect_ratio.den);
                 ret = AVERROR(EINVAL);
                 goto fail;
             }

libavformat/yuv4mpeg.c

@@ -249,6 +249,9 @@ static int yuv4_read_header(AVFormatContext *s)
             } else if (strncmp("420paldv", tokstart, 8) == 0) {
                 pix_fmt = PIX_FMT_YUV420P;
                 chroma_sample_location = AVCHROMA_LOC_TOPLEFT;
+            } else if (strncmp("420", tokstart, 3) == 0) {
+                pix_fmt = PIX_FMT_YUV420P;
+                chroma_sample_location = AVCHROMA_LOC_CENTER;
             } else if (strncmp("411", tokstart, 3) == 0)
                 pix_fmt = PIX_FMT_YUV411P;
             else if (strncmp("422", tokstart, 3) == 0)

libavutil/Makefile

@@ -19,6 +19,7 @@ HEADERS = adler32.h                                                     \
           fifo.h                                                        \
           file.h                                                        \
           imgutils.h                                                    \
+          intfloat.h                                                    \
           intfloat_readwrite.h                                          \
           intreadwrite.h                                                \
           lfg.h                                                         \

libavutil/avutil.h

@@ -154,8 +154,8 @@
  */
 
 #define LIBAVUTIL_VERSION_MAJOR 51
-#define LIBAVUTIL_VERSION_MINOR 35
-#define LIBAVUTIL_VERSION_MICRO 101
+#define LIBAVUTIL_VERSION_MINOR 36
+#define LIBAVUTIL_VERSION_MICRO 100
 
 #define LIBAVUTIL_VERSION_INT   AV_VERSION_INT(LIBAVUTIL_VERSION_MAJOR, \
                                                LIBAVUTIL_VERSION_MINOR, \

libswscale/swscale.c

@@ -2804,7 +2804,7 @@ static int swScale(SwsContext *c, const uint8_t* src[],
         }
     }
 
-    if (isPlanarYUV(dstFormat) && isALPHA(dstFormat) && !alpPixBuf)
+    if (isPlanar(dstFormat) && isALPHA(dstFormat) && !alpPixBuf)
         fillPlane(dst[3], dstStride[3], dstW, dstY-lastDstY, lastDstY, 255);
 
 #if HAVE_MMX2

tests/Makefile

@@ -14,20 +14,18 @@ ffservertest: ffserver$(EXESUF) tests/vsynth1/00.pgm tests/data/asynth1.sw
 	@echo
 	$(SRC_PATH)/tests/ffserver-regression.sh $(FFSERVER_REFFILE) $(SRC_PATH)/tests/ffserver.conf
 
-tests/vsynth1/00.pgm: tests/videogen$(HOSTEXESUF)
-	@mkdir -p tests/vsynth1
+OBJDIRS += tests/data tests/vsynth1 tests/vsynth2
+
+tests/vsynth1/00.pgm: tests/videogen$(HOSTEXESUF) | tests/vsynth1
 	$(M)./$< 'tests/vsynth1/'
 
-tests/vsynth2/00.pgm: tests/rotozoom$(HOSTEXESUF)
-	@mkdir -p tests/vsynth2
+tests/vsynth2/00.pgm: tests/rotozoom$(HOSTEXESUF) | tests/vsynth2
 	$(M)./$< 'tests/vsynth2/' $(SRC_PATH)/tests/lena.pnm
 
-tests/data/asynth1.sw: tests/audiogen$(HOSTEXESUF)
-	@mkdir -p tests/data
+tests/data/asynth1.sw: tests/audiogen$(HOSTEXESUF) | tests/data
 	$(M)./$< $@
 
-tests/data/asynth-16000-1.sw: tests/audiogen$(HOSTEXESUF)
-	@mkdir -p tests/data
+tests/data/asynth-16000-1.sw: tests/audiogen$(HOSTEXESUF) | tests/data
 	$(M)./$< $@ 16000 1
 
 tests/data/mapchan-6ch.sw: tests/audiogen$(HOSTEXESUF)