[FFmpeg-devel] [PATCH] avcodec: Extend fft to size 2^17

Paul B Mahol onemda at gmail.com
Fri Mar 4 13:12:45 CET 2016


On 3/4/16, Michael Niedermayer <michael at niedermayer.cc> wrote:
> On Fri, Mar 04, 2016 at 11:21:23AM +0100, Paul B Mahol wrote:
>> On 3/4/16, Michael Niedermayer <michael at niedermayer.cc> wrote:
>> > Hi
>> >
>> > patch to extend fft is attached (my git-send email atm doesnt work
>> > thanks to my ISP)
>> >
>> > --
>> > Michael     GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB
>> >
>> > I know you won't believe me, but the highest form of Human Excellence
>> > is
>> > to question oneself and others. -- Socrates
>> >
>>
>> This crashes whenever I tried to use it with sofalizer/afftfilt:
>
> fixed with af_afftfilt and applied
>
>
> sofalizer just crashes here even wthout any changes:

That's their awful library.

> #0  0x00007ffff50b447f in clearnccache () from /usr/lib/libnetcdf.so.6
> #1  0x00007ffff50af68d in freeNCDRNO3 () from /usr/lib/libnetcdf.so.6
> #2  0x00007ffff50c34f7 in nc4d_open_file () from /usr/lib/libnetcdf.so.6
> #3  0x00007ffff5076d1f in nc_open () from /usr/lib/libnetcdf.so.6
> #4  0x000000000042cf91 in load_sofa (ctx=0x7fffd8008040, filename=0x0,
> samplingrate=0x7fffd8008148) at libavfilter/af_sofalizer.c:137
> #5  0x000000000042d8b4 in init (ctx=0x7fffd8008040) at
> libavfilter/af_sofalizer.c:936
> #6  0x00000000004cb49b in avfilter_init_str (filter=0x7fffd8008040,
> args=<optimized out>) at libavfilter/avfilter.c:992
> #7  0x00000000004d9052 in create_filter (filt_ctx=0x7fffddf8c6f0,
> ctx=0x7fffd8129840, index=<optimized out>, filt_name=<optimized out>,
> args=0x0, log_ctx=0x0) at libavfilter/graphparser.c:133
> #8  0x00000000004da1d3 in parse_filter (log_ctx=0x0, index=0,
> graph=0x7fffd8129840, buf=0x7fffddf8c6c8, filt_ctx=0x7fffddf8c6f0) at
> libavfilter/graphparser.c:176
> #9  avfilter_graph_parse_ptr (graph=0x7fffd8129840, filters=0x1eda7c9 "",
> open_inputs_ptr=0x7fffddf8c758, open_outputs_ptr=0x7fffddf8c750,
> log_ctx=0x0) at libavfilter/graphparser.c:544
> #10 0x00000000004a8a31 in configure_filtergraph (graph=0x7fffd8129840,
> filtergraph=0x1eda7c0 "sofalizer", source_ctx=<optimized out>,
> sink_ctx=<optimized out>) at ffplay.c:1884
> #11 0x00000000004a8f2b in configure_audio_filters (is=0x7fffe1e8d040,
> afilters=0x1eda7c0 "sofalizer", force_output_format=0) at ffplay.c:2062
> #12 0x00000000004ae4e7 in stream_component_open (is=0x7fffe1e8d040,
> stream_index=2) at ffplay.c:2729
> #13 0x00000000004aef74 in read_thread (arg=0x7fffe1e8d040) at ffplay.c:2975
> #14 0x00007ffff5e80055 in ?? () from
> /usr/lib/x86_64-linux-gnu/libSDL-1.2.so.0
> #15 0x00007ffff5ec3a89 in ?? () from
> /usr/lib/x86_64-linux-gnu/libSDL-1.2.so.0
> #16 0x00007fffecc40e9a in start_thread () from
> /lib/x86_64-linux-gnu/libpthread.so.0
> #17 0x00007fffec96e38d in clone () from /lib/x86_64-linux-gnu/libc.so.6
> #18 0x0000000000000000 in ?? ()
>
>
>
>>
>> ==24326== Jump to the invalid address stated on the next line
>> ==24326==    at 0xEF83490005E6E2B8: ???
>> ==24326==    by 0x5F4BF5F: ff_fft_calc_avx (fft.asm:562)
>> ==24326==    by 0x508AC34: filter_frame (af_afftfilt.c:242)
>> ==24326==    by 0x50C422D: ff_filter_frame_framed (avfilter.c:1123)
>> ==24326==    by 0x50C576D: ff_filter_frame (avfilter.c:1221)
>> ==24326==    by 0x50C9D4A: request_frame (buffersrc.c:450)
>> ==24326==    by 0x50C98A9: av_buffersrc_add_frame_internal
>> (buffersrc.c:239)
>> ==24326==    by 0x50C9635: av_buffersrc_add_frame_flags (buffersrc.c:164)
>> ==24326==    by 0x421F8B: decode_audio (ffmpeg.c:2046)
>> ==24326==    by 0x421F8B: process_input_packet (ffmpeg.c:2325)
>> ==24326==    by 0x41E047: process_input (ffmpeg.c:3998)
>> ==24326==    by 0x41E047: transcode_step (ffmpeg.c:4086)
>> ==24326==    by 0x41E047: transcode (ffmpeg.c:4140)
>> ==24326==    by 0x419331: main (ffmpeg.c:4331)
>> ==24326==  Address 0xef83490005e6e2b8 is not stack'd, malloc'd or
>> (recently) free'd
>> ==24326==
>> ==24326==
>> ==24326== Process terminating with default action of signal 11 (SIGSEGV)
>> ==24326==  Bad permissions for mapped region at address
>> 0xEF83490005E6E2B8
>> ==24326==    at 0xEF83490005E6E2B8: ???
>> ==24326==    by 0x5F4BF5F: ff_fft_calc_avx (fft.asm:562)
>> ==24326==    by 0x508AC34: filter_frame (af_afftfilt.c:242)
>> ==24326==    by 0x50C422D: ff_filter_frame_framed (avfilter.c:1123)
>> ==24326==    by 0x50C576D: ff_filter_frame (avfilter.c:1221)
>> ==24326==    by 0x50C9D4A: request_frame (buffersrc.c:450)
>> ==24326==    by 0x50C98A9: av_buffersrc_add_frame_internal
>> (buffersrc.c:239)
>> ==24326==    by 0x50C9635: av_buffersrc_add_frame_flags (buffersrc.c:164)
>> ==24326==    by 0x421F8B: decode_audio (ffmpeg.c:2046)
>> ==24326==    by 0x421F8B: process_input_packet (ffmpeg.c:2325)
>> ==24326==    by 0x41E047: process_input (ffmpeg.c:3998)
>> ==24326==    by 0x41E047: transcode_step (ffmpeg.c:4086)
>> ==24326==    by 0x41E047: transcode (ffmpeg.c:4140)
>> ==24326==    by 0x419331: main (ffmpeg.c:4331)
>> ==24326==
>> ==24326== HEAP SUMMARY:
>> ==24326==     in use at exit: 10,552,359 bytes in 2,983 blocks
>> ==24326==   total heap usage: 4,458 allocs, 1,475 frees, 11,530,798
>> bytes allocated
>> ==24326==
>> ==24326== LEAK SUMMARY:
>> ==24326==    definitely lost: 0 bytes in 0 blocks
>> ==24326==    indirectly lost: 0 bytes in 0 blocks
>> ==24326==      possibly lost: 1,680 bytes in 5 blocks
>> ==24326==    still reachable: 10,550,679 bytes in 2,978 blocks
>> ==24326==         suppressed: 0 bytes in 0 blocks
>> ==24326== Rerun with --leak-check=full to see details of leaked memory
>> ==24326==
>> ==24326== For counts of detected and suppressed errors, rerun with: -v
>> ==24326== ERROR SUMMARY: 1 errors from 1 contexts (suppressed: 0 from 0)
>> Killed
>>
>> Setting cpuflags to 0 gives output with heavy artifacts.
>> _______________________________________________
>> ffmpeg-devel mailing list
>> ffmpeg-devel at ffmpeg.org
>> http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
>
> --
> Michael     GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB
>
> If you think the mosad wants you dead since a long time then you are either
> wrong or dead since a long time.
>


More information about the ffmpeg-devel mailing list