Opened 8 years ago
Last modified 8 years ago
#5842 open enhancement
"unsupported" mtaf file
| Reported by: | Anon-san | Owned by: | |
|---|---|---|---|
| Priority: | wish | Component: | avcodec |
| Version: | git-master | Keywords: | adpcm |
| Cc: | Blocked By: | ||
| Blocking: | Reproduced by developer: | no | |
| Analyzed by developer: | no |
Description
Well, it does use the mtaf codec but the output doesn't even look stable. It's just noises at this point.
>ffprobe E:\MGS_HDC\PS3\BLUS30847\PS3_GAME\USRDIR\MGS3\mgs3\us\mg1_bgm03_main.mtaf
ffprobe version N-81516-gbe07c25 Copyright (c) 2007-2016 the FFmpeg developers
built with gcc 5.4.0 (GCC)
configuration: --enable-gpl --enable-version3 --disable-w32threads --enable-dxva2 --enable-libmfx --enable-nvenc --enable-avisynth --enable-bzlib --enable-libebur128 --enable-fontconfig --enable-frei0r --enable-gnutls --enable-iconv --enable-libass --enable-libbluray --enable-libbs2b --enable-libcaca --enable-libfreetype --enable-libgme --enable-libgsm --enable-libilbc --enable-libmodplug --enable-libmp3lame --enable-libopencore-amrnb --enable-libopencore-amrwb --enable-libopenh264 --enable-libopenjpeg --enable-libopus --enable-librtmp --enable-libschroedinger --enable-libsnappy --enable-libsoxr --enable-libspeex --enable-libtheora --enable-libtwolame --enable-libvidstab --enable-libvo-amrwbenc --enable-libvorbis --enable-libvpx --enable-libwavpack --enable-libwebp --enable-libx264 --enable-libx265 --enable-libxavs --enable-libxvid --enable-libzimg --enable-lzma --enable-decklink --enable-zlib
libavutil 55. 29.100 / 55. 29.100
libavcodec 57. 54.101 / 57. 54.101
libavformat 57. 48.101 / 57. 48.101
libavdevice 57. 0.102 / 57. 0.102
libavfilter 6. 58.100 / 6. 58.100
libswscale 4. 1.100 / 4. 1.100
libswresample 2. 1.100 / 2. 1.100
libpostproc 54. 0.100 / 54. 0.100
Input #0, mtaf, from 'E:\MGS_HDC\PS3\BLUS30847\PS3_GAME\USRDIR\MGS3\mgs3\us\mg1_bgm03_main.mtaf':
Duration: 00:01:40.23, bitrate: 408 kb/s
Stream #0:0: Audio: adpcm_mtaf, 48000 Hz, 2 channels, s16p
Attachments (1)
Change History (10)
by , 8 years ago
| Attachment: | mg1_bgm03_main.rar added |
|---|
comment:1 by , 8 years ago
| Keywords: | crash added |
|---|---|
| Priority: | normal → important |
| Version: | unspecified → git-master |
comment:4 by , 8 years ago
| Keywords: | SIGSEGV adpcm added |
|---|
I wonder what kind of proof you mean, if it does not crash for you, there most likely is no issue...
$ valgrind ffmpeg_g -i mg1_bgm03_main.mtaf -f null -
==22290== Memcheck, a memory error detector
==22290== Copyright (C) 2002-2011, and GNU GPL'd, by Julian Seward et al.
==22290== Using Valgrind-3.7.0 and LibVEX; rerun with -h for copyright info
==22290== Command: ffmpeg_g -i mg1_bgm03_main.mtaf -f null -
==22290==
ffmpeg version N-81667-g0222602 Copyright (c) 2000-2016 the FFmpeg developers
built with gcc 4.7 (SUSE Linux)
configuration:
libavutil 55. 29.100 / 55. 29.100
libavcodec 57. 55.101 / 57. 55.101
libavformat 57. 49.100 / 57. 49.100
libavdevice 57. 0.102 / 57. 0.102
libavfilter 6. 62.100 / 6. 62.100
libswscale 4. 1.100 / 4. 1.100
libswresample 2. 1.100 / 2. 1.100
Guessed Channel Layout for Input Stream #0.0 : stereo
Input #0, mtaf, from 'mg1_bgm03_main.mtaf':
Duration: 00:01:40.23, bitrate: 408 kb/s
Stream #0:0: Audio: adpcm_mtaf, 48000 Hz, 2 channels, s16p
[null @ 0xb52cba0] Using AVStream.codec to pass codec parameters to muxers is deprecated, use AVStream.codecpar instead.
Output #0, null, to 'pipe:':
Metadata:
encoder : Lavf57.49.100
Stream #0:0: Audio: pcm_s16le, 48000 Hz, stereo, s16, 1536 kb/s
Metadata:
encoder : Lavc57.55.101 pcm_s16le
Stream mapping:
Stream #0:0 -> #0:0 (adpcm_mtaf (native) -> pcm_s16le (native))
Press [q] to stop, [?] for help
==22290== Invalid read of size 2
==22290== at 0xD9E872: adpcm_decode_frame (adpcm.c:352)
==22290== by 0xAC8C0D: avcodec_decode_audio4 (utils.c:2326)
==22290== by 0x499343: decode_audio (ffmpeg.c:1960)
==22290== by 0x49D118: transcode (ffmpeg.c:2336)
==22290== by 0x47EADF: main (ffmpeg.c:4313)
==22290== Address 0x1569ea0 is not stack'd, malloc'd or (recently) free'd
==22290==
==22290==
==22290== Process terminating with default action of signal 11 (SIGSEGV)
==22290== Access not within mapped region at address 0x1569EA0
==22290== at 0xD9E872: adpcm_decode_frame (adpcm.c:352)
==22290== by 0xAC8C0D: avcodec_decode_audio4 (utils.c:2326)
==22290== by 0x499343: decode_audio (ffmpeg.c:1960)
==22290== by 0x49D118: transcode (ffmpeg.c:2336)
==22290== by 0x47EADF: main (ffmpeg.c:4313)
==22290== If you believe this happened as a result of a stack
==22290== overflow in your program's main thread (unlikely but
==22290== possible), you can try to increase the size of the
==22290== main thread stack using the --main-stacksize= flag.
==22290== The main thread stack size used in this run was 8388608.
==22290==
==22290== HEAP SUMMARY:
==22290== in use at exit: 132,185 bytes in 318 blocks
==22290== total heap usage: 910 allocs, 592 frees, 239,934 bytes allocated
==22290==
==22290== LEAK SUMMARY:
==22290== definitely lost: 0 bytes in 0 blocks
==22290== indirectly lost: 0 bytes in 0 blocks
==22290== possibly lost: 2,736 bytes in 9 blocks
==22290== still reachable: 129,449 bytes in 309 blocks
==22290== suppressed: 0 bytes in 0 blocks
==22290== Rerun with --leak-check=full to see details of leaked memory
==22290==
==22290== For counts of detected and suppressed errors, rerun with: -v
==22290== ERROR SUMMARY: 1 errors from 1 contexts (suppressed: 2 from 2)
Killed
(gdb) r -i mg1_bgm03_main.mtaf -f null -
Starting program: ffmpeg_g -i mg1_bgm03_main.mtaf -f null -
[Thread debugging using libthread_db enabled]
Using host libthread_db library "/lib64/libthread_db.so.1".
ffmpeg version N-81667-g0222602 Copyright (c) 2000-2016 the FFmpeg developers
built with gcc 4.7 (SUSE Linux)
configuration:
libavutil 55. 29.100 / 55. 29.100
libavcodec 57. 55.101 / 57. 55.101
libavformat 57. 49.100 / 57. 49.100
libavdevice 57. 0.102 / 57. 0.102
libavfilter 6. 62.100 / 6. 62.100
libswscale 4. 1.100 / 4. 1.100
libswresample 2. 1.100 / 2. 1.100
Guessed Channel Layout for Input Stream #0.0 : stereo
Input #0, mtaf, from 'mg1_bgm03_main.mtaf':
Duration: 00:01:40.23, bitrate: 408 kb/s
Stream #0:0: Audio: adpcm_mtaf, 48000 Hz, 2 channels, s16p
[New Thread 0x7ffff1709700 (LWP 22331)]
[New Thread 0x7ffff0f08700 (LWP 22332)]
[New Thread 0x7ffff0707700 (LWP 22333)]
[New Thread 0x7fffeff06700 (LWP 22334)]
[New Thread 0x7fffef705700 (LWP 22335)]
[New Thread 0x7fffeef04700 (LWP 22336)]
[New Thread 0x7fffee703700 (LWP 22337)]
[New Thread 0x7fffedf02700 (LWP 22338)]
[New Thread 0x7fffed701700 (LWP 22339)]
[null @ 0x1e75200] Using AVStream.codec to pass codec parameters to muxers is deprecated, use AVStream.codecpar instead.
Output #0, null, to 'pipe:':
Metadata:
encoder : Lavf57.49.100
Stream #0:0: Audio: pcm_s16le, 48000 Hz, stereo, s16, 1536 kb/s
Metadata:
encoder : Lavc57.55.101 pcm_s16le
Stream mapping:
Stream #0:0 -> #0:0 (adpcm_mtaf (native) -> pcm_s16le (native))
Press [q] to stop, [?] for help
Program received signal SIGSEGV, Segmentation fault.
0x0000000000d9e872 in adpcm_mtaf_expand_nibble (nibble=0 '\000', c=<optimized out>)
at libavcodec/adpcm.c:352
352 c->predictor += ff_adpcm_mtaf_stepsize[c->step][nibble];
(gdb) bt
#0 0x0000000000d9e872 in adpcm_mtaf_expand_nibble (nibble=0 '\000', c=<optimized out>)
at libavcodec/adpcm.c:352
#1 adpcm_decode_frame (avctx=0x1e60e20, data=0x1e85980, got_frame_ptr=0x7fffffffd7bc,
avpkt=0x7fffffffd460) at libavcodec/adpcm.c:920
#2 0x0000000000ac8c0e in avcodec_decode_audio4 (avctx=avctx@entry=0x1e60e20,
frame=frame@entry=0x1e85980, got_frame_ptr=got_frame_ptr@entry=0x7fffffffd7bc,
avpkt=avpkt@entry=0x7fffffffdac0) at libavcodec/utils.c:2326
#3 0x0000000000499344 in decode_audio (ist=ist@entry=0x1e61540,
pkt=pkt@entry=0x7fffffffdac0, got_output=got_output@entry=0x7fffffffd7bc)
at ffmpeg.c:1960
#4 0x000000000049d119 in process_input_packet (no_eof=0, pkt=0x7fffffffda60, ist=0x1e61540)
at ffmpeg.c:2336
#5 process_input (file_index=2048) at ffmpeg.c:3976
#6 transcode_step () at ffmpeg.c:4064
#7 transcode () at ffmpeg.c:4118
#8 0x000000000047eae0 in main (argc=<optimized out>, argv=0x7fffffffdd08) at ffmpeg.c:4313
(gdb) disass $pc-32,$pc+32
Dump of assembler code from 0xd9e852 to 0xd9e892:
0x0000000000d9e852 <adpcm_decode_frame+6722>: rex.WX clc
0x0000000000d9e854 <adpcm_decode_frame+6724>: lea (%r15,%rdx,4),%rcx
0x0000000000d9e858 <adpcm_decode_frame+6728>: movzbl (%r12,%rax,1),%r11d
0x0000000000d9e85d <adpcm_decode_frame+6733>: mov 0x8(%rcx),%edi
0x0000000000d9e860 <adpcm_decode_frame+6736>: movslq %edi,%rdx
0x0000000000d9e863 <adpcm_decode_frame+6739>: shl $0x4,%rdx
0x0000000000d9e867 <adpcm_decode_frame+6743>: movzbl %r11b,%r10d
0x0000000000d9e86b <adpcm_decode_frame+6747>: and $0xf,%r11d
0x0000000000d9e86f <adpcm_decode_frame+6751>: add %r11,%rdx
=> 0x0000000000d9e872 <adpcm_decode_frame+6754>: movswl 0x1371ba0(%rdx,%rdx,1),%edx
0x0000000000d9e87a <adpcm_decode_frame+6762>: add (%rcx),%edx
0x0000000000d9e87c <adpcm_decode_frame+6764>: lea 0x8000(%rdx),%r8d
0x0000000000d9e883 <adpcm_decode_frame+6771>: mov %edx,%esi
0x0000000000d9e885 <adpcm_decode_frame+6773>: and $0xffff0000,%r8d
0x0000000000d9e88c <adpcm_decode_frame+6780>: je 0xd9e898 <adpcm_decode_frame+6792>
0x0000000000d9e88e <adpcm_decode_frame+6782>: sar $0x1f,%edx
0x0000000000d9e891 <adpcm_decode_frame+6785>: mov %edx,%esi
End of assembler dump.
(gdb) info register
rax 0x0 0
rbx 0x1 1
rcx 0x1e612c0 31855296
rdx 0xfc180 1032576
rsi 0x2 2
rdi 0xfc18 64536
rbp 0x1e612c0 0x1e612c0
rsp 0x7fffffffcf60 0x7fffffffcf60
r8 0x1e85988 32004488
r9 0x1e85e00 32005632
r10 0x0 0
r11 0x0 0
r12 0x1e63630 31864368
r13 0x100 256
r14 0x0 0
r15 0x1e612c0 31855296
rip 0xd9e872 0xd9e872 <adpcm_decode_frame+6754>
eflags 0x10202 [ IF RF ]
cs 0x33 51
ss 0x2b 43
ds 0x0 0
es 0x0 0
fs 0x0 0
gs 0x0 0
comment:6 by , 8 years ago
Replying to richardpl:
Crash should be fixed.
Thank you, but how would an unknowing observer know why you have committed that patch?
comment:7 by , 8 years ago
Because lookup table works only for steps from 0 to 15, anything other is not yet supported.
comment:8 by , 8 years ago
What I meant is: If an innocent outsider looks at b82c1a37 how would he know why this commit was needed, or who told you about the issue?
And how would somebody looking at this ticket know which commit fixed the crash (before this post)?
I often forget to add the necessary information but I wonder if you really forgot it twice today...
comment:9 by , 8 years ago
| Component: | undetermined → avcodec |
|---|---|
| Keywords: | crash SIGSEGV removed |
| Priority: | important → wish |
| Status: | new → open |
| Type: | defect → enhancement |
Note:
See TracTickets
for help on using tickets.
Crashes on 64bit here.