[Ffmpeg-devel-irc] ffmpeg-devel.log.20191206

Sat Dec 7 03:05:03 EET 2019

[00:50:06 CET] <cone-831> ffmpeg 03James Almer 07master:964eb754b456: tools: add a fuzzer tool for bitstream filters
[00:50:29 CET] <jamrial> j-b: ^
[00:55:00 CET] <cone-831> ffmpeg 03Zhao Zhili 07master:71fd72070c86: avutil/buffer: use appropriate atomic operations
[05:50:44 CET] <cone-527> ffmpeg 03Gyan Doshi 07master:b66a800877d6: avfilter/crop: avoid premature eval error
[09:28:15 CET] <kurosu> listening to nttw4, Carl remarked "for most audio files and in broadcast, audio bitdepth as a property does not exist" (https://www.youtube.com/watch?v=cjJZ0cjtaz8&t=4910s)
[09:28:20 CET] <kurosu> err what?
[09:29:39 CET] <kurosu> (addressing archivists)
[09:30:09 CET] <nevcairiel> for most lossy codecs thats true, but those archivists probably backup lossless
[09:31:34 CET] <kurosu> even for lossy codecs I find this debatable
[09:32:26 CET] <kurosu> just because you do a transform and quantize, doesn't say you can have whatever bitdepth
[09:40:12 CET] <rcombs> you could do the math to figure out how many binary sigfigs the output has, I guess?
[09:40:49 CET] <rcombs> it's just not really a simple property
[09:41:11 CET] <rcombs> and the dynamic range encoded might differ in different frequency bands
[09:41:31 CET] <rcombs> or between frames
[09:42:20 CET] <nevcairiel> assuming thats not the first thing a lossy codec sacrifices to reduce size
[09:43:07 CET] <nevcairiel> just look at opus, it goes even further, and audio basically loses its sample rate as well
[09:43:32 CET] <rcombs> isn't conversion to 48kHz literally part of the encoder process
[09:43:40 CET] <kurosu> I mean, I think Carl was missing the point and potentially wrong, depending on his definition of "broadcast"
[09:43:41 CET] <rcombs> *encoding
[09:43:50 CET] <kurosu> does he make a difference with eg "distribution"
[09:43:56 CET] <nevcairiel> no clue how it works, I just know that all audio comes out as 48 afterwards
[09:44:08 CET] <rcombs> iirc it's literally resampled to 48 internally
[09:45:10 CET] <rcombs> but yeah all this applies just as much to "distribution" or "storage" or w/e lossy codecs as it does to "broadcast" ones, esp. since those are mostly the same codecs anyway
[09:46:38 CET] <kurosu> err, sorry, I've just backtracked, I meant "contribution"
[09:47:46 CET] <kurosu> (ie the thing the broadcasters receive from eg studios)
[09:49:12 CET] <kurosu> in any case, and from what I've seen, audio encoders have specific profile for sampling ranges
[09:49:47 CET] <kurosu> but in the end they apply transform (eg MDCT), and after some processing, ends up with frequency bands that they then handle
[09:50:00 CET] <kurosu> (masking, gain control, ...)
[09:50:46 CET] <kurosu> so yeah, in that sense, bitdepth can be "lost", but that's not what I meant
[09:51:49 CET] <kurosu> s/encoders/codecs
[09:51:59 CET] <kurosu> s/codecs/formats :D
[09:54:15 CET] <j-b> jamrial: you rock
[09:59:58 CET] <thardin> feels like hairsplitting tbh
[10:01:46 CET] <kurosu> thardin: what I'm arguing? likely, though I was mostly reacting to the "lossy" part, but my real point was that those people are indeed not using any of the things he's referring to
[10:02:37 CET] <kurosu> for video, there is bitdepth even for lossy, but that's also because things happen in the pixel domain and there is potential drift
[10:05:30 CET] <thardin> yes, as the bad old days taught us
[10:06:55 CET] <kurosu> it's a different drift (ancient days, because of iDCT mismatch between encoder and decoder)
[10:07:16 CET] <kurosu> here, if the encoder/format refers data that is N bits
[10:07:30 CET] <kurosu> the decoder must do the same, otherwise drift
[10:08:14 CET] <kurosu> (predictions would differ)
[10:08:55 CET] <thardin> of course
[10:09:02 CET] <thardin> some codecs don't have that property tho
[10:11:06 CET] <nevcairiel> very few audio codecs even use prediction at all, i believe some AAC profile has it, but noone uses it
[10:11:42 CET] <cone-712> ffmpeg 03Marton Balint 07master:648b8cca6c56: avdevice/xcbgrab: wrap non-shm image replies in a buffer ref
[10:12:29 CET] <Lynne> opus uses prediction for the band energies, but doesn't flag keyframes as such or even consider whether old ones exist
[10:12:43 CET] <kurosu> Lynne: AAC is LTP right?
[10:13:49 CET] <Lynne> ltp technically doesn't qualify as it predicts within the mdct frame, but the main profile which delta encodes coeffs from the previous frame does
[10:15:11 CET] <Lynne> but every mdct codec relies on some part of the previous frame, so if you think that's prediction then all codecs which use it predict
[10:15:19 CET] <kurosu> both happen in the transform domain, but I did mention my comment applies to video, I was just commenting on the "prediction" (at all)
[10:15:31 CET] <kurosu> yeah, the overlap part of the MDCT
[10:29:07 CET] <thardin> codec2 does prediction, but it will converge to the right output quickly
[10:29:24 CET] <thardin> necessary when operating over a lossy channel
[11:27:16 CET] <durandal_1707> oh noes, here goes the spam
[11:33:08 CET] <cone-712> ffmpeg 03Limin Wang 07master:f9d6addd60b3: avformat/libsrt: fix for the memory leak if passphrase has been configured by option
[11:33:09 CET] <cone-712> ffmpeg 03Ting Fu 07master:039a0ebe6fd8: libswscale/swscale_unscaled.c: remove redundant code
[11:33:10 CET] <cone-712> ffmpeg 03leozhang 07master:0c7f9f714d87: avfilter/vf_yaepblur: add yaepblur filter
[11:33:11 CET] <cone-712> ffmpeg 03Limin Wang 07master:1d757b111aed: avfilter/vsrc_cellauto: Fix for random_seed type
[11:33:12 CET] <cone-712> ffmpeg 03Limin Wang 07master:5ee4c12ec277: avfilter/vsrc_life: Fix for random_seed type
[11:33:13 CET] <cone-712> ffmpeg 03Michael Niedermayer 07master:e7011a0ca654: avcodec/mvha: Check remaining space when reading VLC table probabilities
[11:33:14 CET] <cone-712> ffmpeg 03Limin Wang 07master:79d907774d59: avformat/libsrt: change tlpktdrop, nakreport, messageapi options to boolean type
[11:58:40 CET] <BtbN> Patch 001/224. Oh God.
[12:00:23 CET] <thardin> big boi
[12:00:48 CET] <BtbN> The series was posted an hour ago, and still only the first 10 mails reached me
[12:01:04 CET] <nevcairiel> he likely didnt send all of them but only a few
[12:01:46 CET] <BtbN> I hope so
[12:01:54 CET] <BtbN> otherwise this can get the list banned from various providers
[12:02:20 CET] <nevcairiel> if high volume gets you banned, you should get a better provider
[12:11:39 CET] <Lynne> so could the new channel api properly describe ambisonics or is it changing things for the sake of changing things?
[12:13:51 CET] <kurosu> very good question I was about to ask
[12:14:24 CET] <kurosu> you suggesting ambisonics makes it sound like there's likely one big user
[12:15:43 CET] <kurosu> at the risk of being bureaucratic (and not having checked enough the patchset), I'd say 1) deprecation warning long in advance 2) make sure there's not a non-trivial amount of users that would complain
[12:16:32 CET] <BtbN> What do you mean, long in advance? Warning about deprecation before an alternative exists is only confusing people.
[12:16:35 CET] <kurosu> ok deprecation is there
[12:17:39 CET] <kurosu> the alternative is implemented there, didn't pay attention (until now) whether it was bundled with deprecation warnings
[12:17:54 CET] <nevcairiel> the important part is compat with old api, as long as thats there, they can complain all they want, and then adjust later anyway
[12:19:22 CET] <nevcairiel> advanced channel layouts are coming (or are already here, just not in ffmpeg), so eventually we have to define a new api to represent them
[12:19:56 CET] <BtbN> Only thing I'm worried about is running out of bits in a 64bit integer
[12:20:19 CET] <nevcairiel> well you cant expect more advanced layouts to fit into the compat api
[12:20:30 CET] <nevcairiel> thats why the new one exists, since 64-bit integers are just limited
[12:20:44 CET] <BtbN> Oh, those new flags are purely for compat with the old API?
[12:21:40 CET] <nevcairiel> well sort of, not everything needs full special magic
[12:22:22 CET] <kurosu> or we could use gmplib :o
[12:22:23 CET] <nevcairiel> look at the new struct, it has two options to signal channel, either its a 64-bit mask, similar to before, or it can have an allocated array of complex structures to describe channel, in which case only the sky (or your memory) is the limit
[12:23:14 CET] <nevcairiel> so all simple cases will continue to use the 64-bit mask, it'll just be wrapped into a structure
[12:23:23 CET] <nevcairiel> anything that doesnt fit into that, has a new method to signal
[12:29:54 CET] <Lynne> kurosu: well, 2, I think vlc were working on something ambisonics related, but I just want google to not hack up the libopus wrapper
[12:31:25 CET] <kurosu> I have no opinion except I don't like changes without reasons
[12:31:49 CET] <kurosu> but then I see a few people here already thinking it's useful (besides the people working on it)
[19:13:12 CET] <cone-637> ffmpeg 03Andriy Gelman 07master:02a83e26de6a: lavc/cbs_h2645: Fix incorrect max size of nalu unit
[20:58:35 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:c78d5c9ce93e: avcodec/truemotion2: fix integer overflows in tm2_low_chroma()
[20:58:36 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:b84cda4b6399: avcodec/msmpeg4dec: Skip frame if its smaller than 1/8 of the minimal size
[20:58:37 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:d6b647e22138: avcodec/wmv2dec: Skip I frame if its smaller than 1/8 of the minimal size
[20:58:38 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:b3dab0f894d3: avcodec/msvideo1: Check for too small dimensions
[20:58:39 CET] <cone-637> ffmpeg 03chcunningham 07release/2.8:acd49222eb03: lavf/mov: ensure only one tkhd per trak
[20:58:40 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:8381cbf692f4: avformat/nutenc: Document trailer index assert better
[20:58:41 CET] <cone-637> ffmpeg 03chcunningham 07release/2.8:68be54dfbecb: lavf/id3v2: fail read_apic on EOF reading mimetype
[20:58:42 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:01b179f39bc4: avcodec/mjpegdec: Fix indention of ljpeg_decode_yuv_scan()
[20:58:43 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:6cc642cafa41: postproc/postprocess_template: Avoid using %4 for the threshold compare
[20:58:44 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:75c53cf09e77: avcodec/fic: Fail on invalid slice size/off
[20:58:45 CET] <cone-637> ffmpeg 03gxw 07release/2.8:a7f94ae58381: avcodec/mips: Fix failed case: hevc-conformance-AMP_A_Samsung_* when enable msa
[20:58:46 CET] <cone-637> ffmpeg 03David Bryant 07release/2.8:12dcf732ff73: avformat/wvdec: detect and error out on WavPack DSD files
[20:58:47 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:415dd357af2b: avcodec/mjpegbdec: Fix some misplaced {} and spaces
[20:58:47 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:1e41754792e4: avcodec/4xm: Fix returned error codes
[20:58:49 CET] <cone-637> ffmpeg 03James Almer 07release/2.8:d9590afafcc3: configure: bump year
[20:58:50 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:af857cb62f93: avutil/mem: Optimize fill32() by unrolling and using 64bit
[20:58:51 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:54c01b6f9404: avcodec/fic: Check that there is input left in fic_decode_block()
[20:58:52 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:90fbb18a7c2f: avformat/rtsp: Clear reply in every iteration in ff_rtsp_connect()
[20:58:53 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:61894a612b23: avformat/rtsp: Check number of streams in sdp_parse_line()
[20:58:54 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:5f0699d39abf: avcodec/pgssubdec: Check for duplicate display segments
[20:58:55 CET] <cone-637> ffmpeg 03chcunningham 07release/2.8:4123064b5f25: avformat/mov.c: require tfhd to begin parsing trun
[20:58:56 CET] <cone-637> ffmpeg 03Wenxiang Qian 07release/2.8:7426a7c958f8: avformat/ftp: Fix Out-of-Bounds Access and Information Leak in ftp.c:393
[20:58:57 CET] <cone-637> ffmpeg 03Wenxiang Qian 07release/2.8:423319d1575a: avformat/http: Fix Out-of-Bounds access in process_line()
[20:58:58 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:8dde01b57b3c: avformat/webmdashenc: Check id in adaption_sets
[20:58:59 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:da696f996641: avcodec/jvdec: Check available input space before decode8x8()
[20:58:59 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:1911fdb7b894: avcodec/mlpdec: Insuffient typo
[20:59:01 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:d555be4ba5b8: avcodec/error_resilience: Use a symmetric check for skipping MV estimation
[20:59:02 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:4797b6092df9: avcodec/jpeg2000dwt: Fix integer overflow in dwt_decode97_int()
[20:59:03 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:1ca27e5215f9: avcodec/bethsoftvideo: Check block_type
[20:59:04 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:9a5150b40ca8: avcodec/aic: Check remaining bits in aic_decode_coeffs()
[20:59:05 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:6e926090d31c: avcodec/qpeg: Limit copy in qpeg_decode_intra() to the available bytes
[20:59:06 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:d3423fb7d40c: avcodec/cdgraphics: Use ff_set_dimensions()
[20:59:07 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:f8cb091b6e51: avcodec/dvbsubdec: Check object position
[20:59:08 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:ce5fb015acee: avcodec/dfa: Check the chunk header is not truncated
[20:59:09 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:0b93f593184c: avcodec/truemotion2: Fix integer overflow in tm2_null_res_block()
[20:59:10 CET] <cone-637> ffmpeg 03Carl Eugen Hoyos 07release/2.8:8d3764a54dcd: lavc/bmp: Avoid a heap buffer overwrite for 1bpp input.
[20:59:11 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:4bc338278054: avcodec/hevcdec: Avoid only partly skiping duplicate first slices
[20:59:12 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:12e4877ae251: avcodec/truemotion2: Fix integer overflow in tm2_decode_blocks()
[20:59:13 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:059f291aafbf: avcodec/diracdec: Use 64bit in intermediate of global motion vector field generation
[20:59:14 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:89b9c728623f: avcodec/ivi: Move buffer/block end check to caller of ivi_dc_transform()
[20:59:15 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:aa152bcaeccb: avcodec/ccaption_dec: Add a blank like at the end to avoid rollup reading from outside
[20:59:16 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:ef7bef2ba4c2: avformat/aadec: Check for scanf() failure
[20:59:17 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:37b2a19afa25: avutil/avstring: Fix bug and undefined behavior in av_strncasecmp()
[20:59:18 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:7a5f875e9898: avcodec/aacdec_fixed: Fix undefined shift in noise_scale()
[20:59:19 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:a4bd19c53afe: avcodec/truemotion2: Fix 2 integer overflows in tm2_update_block()
[20:59:20 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:9bb9aa9085e5: avcodec/jvdec: Use ff_get_buffer() when the content is not reused
[20:59:21 CET] <cone-637> ffmpeg 03Andreas Rheinhardt 07release/2.8:2df8814ae7ce: lavf/webm_chunk: Respect buffer size
[20:59:22 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:332de4c1e42e: avcodec/hq_hqa: Check available space before reading slice offsets
[20:59:23 CET] <cone-637> ffmpeg 03Adam Richter 07release/2.8:b55cf5f53d3e: libswcale: Fix possible string overflow in test.
[20:59:24 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:2f8d30fb9b53: avcodec/cpia: Check input size also against linesizes and EOL
[20:59:25 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:3fa90098a9fb: avformat/webm_chunk: Check header filename length
[20:59:26 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:fa0caf43c46c: avformat/webm_chunk: Specify expected argument length of get_chunk_filename()
[20:59:27 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:805e6923035e: avcodec/truemotion2: Fix several integer overflows in tm2_update_block()
[20:59:28 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:d418ca3786b1: avformat/mp3enc: Avoid SEEK_END as it is unsupported
[20:59:29 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:3d7b7e533138: avcodec/aacdec_fixed: ssign seems always -1 in noise_scale(), simplify
[20:59:30 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:7f4ac04f3096: avcodec/aacdec_template: Merge 3 #ifs related to noise handling
[20:59:31 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:53b314938ff4: avcodec/aacdec_fixed: Handle more extreem cases in noise_scale()
[20:59:32 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:380289beab8d: avcodec/aacdec_template: skip apply_tns() if max_sfb is 0 (from previous header decode failure)
[20:59:33 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:b4aeb407a581: avcodec/vmnc: Check available space against chunks before reget_buffer()
[20:59:34 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:7e7f52deda5f: avcodec/diracdec: Fix integer overflow in global_mv()
[20:59:35 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:386368fd918a: avcodec/mss4: Check input size against skip bits
[20:59:36 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:ea3df2f7becf: avcodec/mjpegdec: Check for non ls PAL8
[20:59:37 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:28e2f0c51d3e: avformat/sbgdec: Fixes integer overflow in str_to_time() with hours
[20:59:38 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:5cca0755c721: avcodec/bitstream: Check for integer code truncation in build_table()
[20:59:39 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:1bf0dd08c526: avcodec/bitstream: Check for more conflicting codes in build_table()
[20:59:40 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:990753e81f2d: avformat/wtvdec: Avoid (32bit signed) sectors
[20:59:41 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:b9be7f3adf0b: avcodec/bink: Reorder operations in init to avoid memleak on error
[20:59:42 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:d3a1a59c22a1: avcodec/iff: finetune the palette size check in the mask case
[20:59:43 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:5a7a2180f8d5: avcodec/truemotion2: Fix integer overflow in last loop in tm2_update_block()
[20:59:44 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:1df86144e086: avcodec/aacpsdsp_template: Fix integer overflow in ps_hybrid_analysis_c()
[20:59:45 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:677f461508e0: avformat/mov: Set fragment.found_tfhd only after TFHD has been parsed
[20:59:46 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:1d288d16d871: avcodec/loco: Limit lossy parameter so it is sane and does not overflow
[20:59:47 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:7ddb5fcbace0: avcodec/motionpixels: Check for vlc error in mp_get_vlc()
[20:59:48 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:274423dd468a: avcodec/binkdsp: Fix integer overflows in idct
[20:59:49 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:58c95520ee4e: avcodec/bintext: Check font height
[20:59:50 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:7b9ca45a4317: avcodec/tta: Fix undefined shift
[20:59:51 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:006d1a2f8098: avcodec/vc1dsp: Avoid undefined shifts in vc1_v_s_overlap_c / vc1_h_s_overlap_c
[20:59:52 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:40a614a5bfd1: avcodec/ffwavesynth: Check ts_end - ts_start for overflow
[20:59:53 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:3af9317a113a: avcodec/videodsp_template: Fix overflow of addition
[20:59:54 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:1894a36d6cbf: avcodec/libvorbisdec: Check extradata size
[20:59:55 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:1b491fbf2505: avcodec/qdm2: Move fft_order check up
[20:59:56 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:43ced88379a5: avcodec/apedec: Add k < 24 check to the only k++ case which lacks such a check
[20:59:57 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:7a4e69fed433: avcodec/alsdec: Fixes invalid shifts in read_var_block_data() and INTERLEAVE_OUTPUT()
[20:59:58 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:91c70cc895c8: avcodec/alsdec: Fix undefined behavior in decode_rice()
[20:59:59 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:cabdd900b7bf: avcodec/alsdec: Fix integer overflow with shifting samples
[21:00:00 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:a6d53b3028b4: avcodec/alsdec: Check opt_order / sb_length in ra_block handling
[21:00:01 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:4e2bf59270b0: avcodec/alsdec: Fix integer overflow with buffer number
[21:00:02 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:fa89d0ec88e2: avcodec/alsdec: Add FF_CODEC_CAP_INIT_CLEANUP
[21:00:03 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:c583154829d2: avcodec/alac: Check lpc_quant
[21:00:04 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:05a164a4dcd3: avcodec/vc1_block: Check for vlc error in vc1_decode_ac_coeff()
[21:00:05 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:3bb800525e5e: avcodec/ffwavesynth: Fix backward lcg_seek()
[21:00:06 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:d8c85a346484: avcodec/ffwavesynth: Simplify lcg_seek(), avoid negative case
[21:00:07 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:ace8a759e67e: avcodec/ffwavesynth: use uint32_t to compute difference, it is enough
[21:00:08 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:a58dfb447eac: avcodec/iff: Check ham vs bpp
[21:00:09 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:7b4cd6ac8e88: avcodec/svq3: Use ff_set_dimension()
[21:00:10 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:8f5aff9e2d1b: avcodec/qdm2: Do not read out of array in fix_coding_method_array()
[21:00:11 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:2a1e1bc0b59b: avcodec/qdm2: error out of qdm2_fft_decode_tones() before entering endless loop
[21:00:12 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:76a60fc760bb: avcodec/qdm2: Check checksum_size for 0
[21:00:13 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:a70dd95c43c0: avcodec/4xm: Fix signed integer overflows in idct()
[21:00:14 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:4696370756e5: avcodec/rv10: Fix integer overflow in aspect ratio compare
[21:00:15 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:54b3727a494c: avcodec/hq_hqa: Use ff_set_dimensions()
[21:00:16 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:65c25ad3688a: avcodec/mpc8: Fixes invalid shift in mpc8_decode_frame()
[21:00:17 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:40c3eedd40da: avcodec/huffyuv: remove gray8a (the format is listed but not supported by the implementation)
[21:00:18 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:f2d81316e8ea: avcodec/pngdec: Check that previous_picture has same w/h/format
[21:00:19 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:332a03042dc6: avformat/xmv: Make bitrate 64bit
[21:00:20 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:42734c45785e: avcodec/ivi: Ask for samples with odd tiles
[21:00:21 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:9ad8710d1cdd: avcodec/parser: Check next index validity in ff_combine_frame()
[21:00:22 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:73556665ec90: avformat/utils: Check rfps_duration_sum for overflow
[21:00:23 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:561aefbf8add: avcodec/ffwavesynth: Check sample rate before use
[21:00:24 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:af0aea414fac: avcodec/ffwavesynth: More correct cast in wavesynth_seek()
[21:00:25 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:996accb0d7d1: avcodec/ffwavesynth: Check if there is enough extradata before allocation
[21:00:26 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:45a306e4c42f: avcodec/dvbsubdec: Use ff_set_dimensions()
[21:00:27 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:d36bc33bcca5: avcodec/flicvideo: Make line_packets int
[21:00:28 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:b4185cb15426: avcodec/alsdec: Fix 2 integer overflows
[21:00:29 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:7700ef673784: avcodec/apedec: Fix multiple integer overflows in predictor_update_filter()
[21:00:30 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:0d922de167ba: avcodec/apedec: Fix various integer overflows
[21:00:31 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:b217b8ed59cd: avcodec/vorbisdec: amplitude bits can be more than 25 bits
[21:00:32 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:b47758964a8e: avcodec/vorbisdec: Check vlc for floor0 dec vector offset
[21:00:33 CET] <cone-637> ffmpeg 03Matt Wolenetz 07release/2.8:82094a703ed8: lafv/wavdec: Fail bext parsing on incomplete reads
[21:00:34 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:e46c1c68c65c: avcodec/brenderpix: Check input size before allocating image
[21:00:35 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:8c66aed22bd5: avcodec/assdec: undefined use of memcpy()
[21:00:36 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:f2eb30e65469: avcodec/eatgv: Check remaining size after the keyframe header
[21:00:37 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:f9749f3335f8: avcodec/apedec: Fix multiple integer overflows and undefined behaviorin filter_3800()
[21:00:38 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:f7f3c07282d3: avcodec/apedec: make left/right unsigned to avoid undefined behavior
[21:00:39 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:e6701fa94b61: avcodec/truemotion2: Fix several integer overflows in tm2_motion_block()
[21:00:40 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:5c1a6ecb9832: avcodec/vc1_block: Fix integer overflow in ff_vc1_pred_dc()
[21:00:41 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:8d6608c04d4b: avcodec/vc1_pred: Fix invalid shift in scaleforsame()
[21:00:42 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:d1c2e5a36254: avcodec/vp3: Check that theora is theora
[21:00:43 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:18013f760b04: avcodec/hnm4video: Forward errors of decode_interframe_v4()
[21:00:44 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:cc600fe4b58e: avcodec/apedec: Do not partially clear data array
[21:00:45 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:4748da854a3e: avcodec/vc1_block: Check get_vlc2() return before use
[21:00:46 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:b4530a6ac8bb: avcodec/mss3: Check for the rac stream being invalid in rac_normalize()
[21:00:47 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:787501ae7f78: avcodec/apedec: Fix 2 signed overflows
[21:00:48 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:979d982bc1b8: avcodec/vorbisdec: Check parameters in vorbis_floor0_decode() before divide
[21:00:49 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:ef938d8efdd7: avcodec/vorbisdec: Implement vr->classifications = 1
[21:00:50 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:0e69785d5368: avcodec/ralf: Fix integer overflow in apply_lpc()
[21:00:51 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:21d93e0254da: avcodec/ralf: Fix undefined pointer in decode_channel()
[21:00:52 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:af58de8e490c: avcodec/dirac_parser: Fix overflow in dts
[21:00:53 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:5a7b7fc0258f: avcodec/mss1: check for overread and forward errors
[21:00:54 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:1d752c2a057d: avcodec/sanm: Check extradata_size before allocations
[21:00:55 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:5818ae3cff32: avcodec/vqavideo: Set video size
[21:00:56 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:92c75a313ec6: avcodec/alsdec: Check for block_length <= 0 in read_var_block_data()
[21:00:57 CET] <cone-637> ffmpeg 03Andreas Rheinhardt 07release/2.8:a1233d39c62b: libavcodec/iff: Use unsigned to avoid undefined behaviour
[21:00:58 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:3ee68a677d67: avcodec/aacdec_template: fix integer overflow in imdct_and_windowing()
[21:00:59 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:09a6d81db0e9: avcodec/alsdec: Fix integer overflow of raw_samples in decode_blocks()
[21:01:00 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:781aa8f07f6e: avcodec/alsdec: Fix integer overflows of raw_samples in decode_var_block_data()
[21:01:01 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:3538c3476f16: avcodec/mpc8: Fix 32bit mask/enum
[21:01:02 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:514164d4e202: avcodec/dds: Use ff_set_dimensions()
[21:01:03 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:296a167b3ec0: avcodec/hevc_refs: Optimize 16bit generate_missing_ref()
[21:01:04 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:ba3f7c14af6d: avcodec/hnm4video: Optimize postprocess_current_frame()
[21:01:05 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:6eef056cf56e: avcodec/4xm: Check for end of input in decode_p_block()
[21:01:06 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:94242dd739a0: avcodec/vp56: Consider the alpha start as end of the prior header
[21:01:07 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:6201a6063155: avcodec/indeo2: Check remaining input more often
[21:01:08 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:479361eb92b6: avcodec/ffwavesynth: Fix integer overflow for some corner case values
[21:01:09 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:a982163627a3: avcodec/ffwavesynth: Fixes invalid shift with pink noise seeking
[21:01:10 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:e32ef596a910: avcodec/loco: Check left column value
[21:01:11 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:1b0aac8f4282: avcodec/anm: Check input size for a frame with just a stop code
[21:01:12 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:59f2ec706267: avcodec/alsdec: Limit maximum channels to 512
[21:01:13 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:f4c287ed23a1: avcodec/alsdec: Fix integer overflow in decode_var_block_data()
[21:01:14 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:d5ab2f5023d1: avformat/realtextdec: free queue on error
[21:01:15 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:72a0e74db986: avcodec/rl2: set dimensions
[21:01:16 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:db0b4479050e: avcodec/alac: Fix multiple integer overflows in lpc_prediction()
[21:01:17 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:9a46d7223580: avcodec/cavsdec: Limit the number of access units per packet to 2
[21:01:18 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:c4981bcf07b2: avcodec/vb: Check input packet size to be large enough to contain flags
[21:01:19 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:6d6965b7c41d: avcodec/tta: Fix integer overflow in prediction
[21:01:20 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:7954bc8f63c0: avcodec/vorbisdec: Check get_vlc2() failure
[21:01:21 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:720c2deb0221: avcodec/vc1_block: Check for double escapes
[21:01:22 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:4b2c2d787c41: avcodec/vc1dec: Require res_sprite for wmv3images
[21:01:23 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:d35b7197ddf2: avcodec/truemotion2: Fix multiple integer overflows in tm2_null_res_block()
[21:01:24 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:b74d119ee67d: avcodec/bgmc: Check input space in ff_bgmc_decode_init()
[21:01:25 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:1da3f1084eb6: avcodec/ralf: fix undefined shift
[21:01:26 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:ac0ffb67f55a: avcodec/ralf: fix undefined shift in extend_code()
[21:01:27 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:6bfec5efc4e2: avcodec/apedec: Fix 32bit int overflow in do_apply_filter()
[21:01:28 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:08e46a341a40: avcodec/iff: Check for overlap in cmap_read_palette()
[21:01:29 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:feaf42eb0d73: avcodec/vorbisdec: fix FASTDIV usage for vr_type == 2
[21:01:30 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:c01055a1ffe1: avcodec/vc1_pred: Fix refdist in scaleforopp()
[21:01:31 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:0e704512b574: avcodec/qdm2: Check frame size
[21:01:32 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:04acafdc7c36: avcodec/vp56: Add vpX_rac_is_end() to check for the end of input
[21:01:33 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:8fecba0aa7e3: avcodec/vp5/6/8: use vpX_rac_is_end()
[21:01:34 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:80d27aaf7078: avcodec/vp56rac: delay signaling an error on truncated input
[21:01:35 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:668d797a5723: avcodec/alsdec: Check k from being outside what our implementation can handle
[21:01:36 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:2a306730a4a3: avcodec/hevcdec: repeat character in skiped
[21:01:37 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:a6f50a53cfd1: avformat/cdxl: Fix integer overflow in intermediate
[21:01:38 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:fc6664cb2abb: avcodec/sunrast: Fix return type for "unsupported (compression) type"
[21:01:39 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:0e406d200816: avcodec/sunrast: Fix indention
[21:01:40 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:40d3103d36e4: avcodec/ffwavesynth: Fix integer overflow in timestamps
[21:01:41 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:2f095670c88d: avcodec/alac: fix undefined behavior with INT_MIN in lpc_prediction()
[21:01:42 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:e15fd9bc2562: avcodec/smacker: Fix integer overflow in signed int multiply in SMK_BLK_FILL
[21:01:43 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:880579a76387: avcodec/aacdec: Check if we run out of input in read_stream_mux_config()
[21:01:44 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:bf5e3a4962cb: avcodec/g729_parser: Check block_size
[21:01:45 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:e60504a71dcb: avcodec/alsdec: Avoid dereferencing context pointer in inner interleave loop
[21:01:46 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:e5ea4e727c55: avcodec/atrac3: Check block_align
[21:01:47 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:bdcff3ef0dcb: avcodec/4xm: Check index in decode_i_block() also in the path where its not used.
[21:01:48 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:4b619bad46ff: avcodec/hevc_cabac: Tighten the limit on k in ff_hevc_cu_qp_delta_abs()
[21:01:49 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:76c86b0ad963: avcodec/apedec: Fix several integer overflows in predictor_update_filter() and do_apply_filter()
[21:01:50 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:f7218d155e5a: avcodec/vc1: check REFDIST
[21:01:51 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:bdacbbba69c8: avcodec/vc1: Check for excessive resolution
[21:01:52 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:e5796aa1b97a: avcodec/vc1_block: Fix invalid shift with rangeredfrm
[21:01:53 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:446238d170ca: avcodec/vc1_pred: Fix invalid shifts in scaleforopp()
[21:01:54 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:4d9776589861: vcodec/vc1: compute rangex/y only for P/B frames
[21:01:55 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:923bcf0c3f92: avcodec/ralf: Fix integer overflow in decode_channel()
[21:01:56 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:5812f834d4fe: avcodec/motionpixels: Mark 2 functions as always_inline
[21:01:57 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:214556b9d9db: avcodec/g2meet: Check if adjusted pixel was on the stack
[21:01:58 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:3690b8684f5a: avcodec/g2meet: Check for end of input in jpg_decode_block()
[21:01:59 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:9b424ca5d208: avformat/subtitles: Check nb_subs in ff_subtitles_queue_finalize()
[21:02:00 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:18f91bc819d0: avcodec/sunrast: Check for availability of maplength before allocating image
[21:02:01 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:e4667a699196: avcodec/sunrast: Check that the input is large enough for the maximally compressed image
[21:02:02 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:5591681af13b: avformat/mpsubdec: Clear queue on error
[21:02:03 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:9c6e76eee70f: avcodec/truemotion1: Check that the input has enough space for a minimal index_stream
[21:02:04 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:bd5a214c3c1c: avcodec/ituh263dec: Check input for minimal frame size
[21:02:05 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:560770d32be8: avfilter/vf_geq: Use av_clipd() instead of av_clipf()
[21:02:06 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:68b62e36eb2e: avcodec/adpcm: Fix left shifts in AV_CODEC_ID_ADPCM_EA
[21:02:07 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:23d226fe42b1: avcodec/lsp: Fix undefined shifts in lsp2poly()
[21:02:08 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:79532fa945a3: avcodec/g729postfilter: Fix undefined shifts
[21:02:09 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:7187b14d9bac: avcodec/g729postfilter: Fix undefined intermediate pointers
[21:02:10 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:dad5ed01a82a: avcodec/apedec: Fix integer overflow in predictor_update_3930()
[21:02:11 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:68a8bdf5f198: avcodec/adpcm: Check initial predictor for ADPCM_IMA_EA_EACS
[21:02:12 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:8b462c958a91: avcodec/binkaudio: Check sample rate
[21:02:13 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:366cfe1615e2: avcodec/g729postfilter: Fix left shift of negative value
[21:02:14 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:a3692f404175: avcodec/wmalosslessdec: Check block_align
[21:02:15 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:d9123585d152: avcodec/vc1_block: Fixes integer overflow in vc1_decode_i_block_adv()
[21:02:16 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:77847fd19e71: avcodec/ffwavesynth: Fix integer overflows in pink noise addition
[21:02:17 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:cdb193839ac5: avcodec/aliaspixdec: Check input size against minimal picture size
[21:02:18 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:684fbde54347: avcodec/smacker: Fix integer overflows in pred[] in smka_decode_frame()
[21:02:19 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:50947f65e0ad: avcodec/alac: Fix integer overflow in LPC
[21:02:20 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:d85af33e05cc: avcodec/ptx: Check that the input contains at least one line
[21:02:21 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:cf5945613fc0: avformat/pjsdec: Check duration for overflow
[21:02:22 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:08a0c3cd179a: avcodec/ralf: Skip initializing unused filter variables
[21:02:23 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:f1a20096f76a: avcodec/vp5: Check render_x/y
[21:02:24 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:62eced0903ae: avcodec/libvorbisdec: Fix insufficient input checks leading to out of array reads
[21:02:25 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:1f183e8cc80f: avcodec/apedec: Only clear the needed buffer space, instead of all
[21:02:26 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:5ebb902caf04: avcodec/adpcm: Fix invalid shifts in ADPCM DTK
[21:02:27 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:d7ebf3a074f9: avcodec/truemotion2: Fix several integer overflows in tm2_low_res_block()
[21:02:28 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:5729bbfdc769: avformat/nutenc: Do not pass NULL to memcmp() in get_needed_flags()
[21:02:29 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:ef9340a9dc6e: avcodec/snowenc: Fix 2 undefined shifts
[21:02:30 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:7a5c2173fec5: avcodec/apedec: Fix undefined integer overflow in long_filter_ehigh_3830()
[21:02:31 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:31ae31c21e7f: avcodec/ffv1dec: Use a different error message for the slice level CRC
[21:02:32 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:f5fdab6b0e41: avcodec/apedec: Fix integer overflow in filter_3800()
[21:02:33 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:f1b32668dc33: avcodec/wmavoice: Check sample_rate
[21:02:34 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:6f4c963d6168: avcodec/atrac3plus: Check split point in fill mode 3
[21:02:35 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:6ef666f4a699: avcodec/cook: Enlarge gain table
[21:02:36 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:20facd705edd: avcodec/twinvq: Check block_align
[21:02:37 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:e53182c467ef: avcodec/cook: Move up and extend block_align check
[21:02:38 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:c8f306fc0b9a: avcodec/adpcm: Fix undefined behavior with negative predictions in IMA OKI
[21:02:39 CET] <cone-637> ffmpeg 03Carl Eugen Hoyos 07release/2.8:fbc6717495d9: lavc/tableprint_vlc: Remove avpriv_request_sample() from included files.
[21:02:40 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:e710fd6a838c: avcodec/sonic: Fix integer overflow in predictor_calc_error()
[21:02:41 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:0e40a839bb78: avcodec/apedec: Fixes integer overflow of res+*data in do_apply_filter()
[21:02:42 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:40872346be2e: avcodec/vc1_block: Fix undefined behavior in ac prediction rescaling
[21:02:43 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:b48dcc5613ca: avcodec/wmadec: Require previous exponents for reuse
[21:02:44 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:6248ed92fd63: avcodec/ralf: use multiply instead of shift to avoid undefined behavior in decode_block()
[21:02:45 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:67103c565c34: avcodec/atrac3: Check for huge block aligns
[21:02:46 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:aa8498c41783: avcodec/iff: Move index use after check in decodeplane8()
[21:02:47 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:1e57ce1dd5a1: avcodec/cngdec: Remove AV_CODEC_CAP_DELAY
[21:02:48 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:98510f3c8482: avutil/lfg: Correct index increment type to avoid undefined behavior
[21:02:49 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:22b24fe8ae76: avcodec/wmavoice: Fix integer overflow in synth_frame()
[21:02:50 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:c005d08103f5: avcodec/nuv: Move comptype check up
[21:02:51 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:8470f8fd8180: avcodec/mxpegdec: Check for multiple SOF
[21:02:52 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:33fff42d8d98: avcodec/g729dec: Use 64bit and clip in scalar product
[21:02:53 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:971a6493da7b: avcodec/ralf: Fix integer overflows with the filter coefficient in decode_channel()
[21:02:54 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:032650458cce: avcodec/ffwavesynth: Fix integer overflow with pink_ts_cur/next
[21:02:55 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:ce61019eb197: avcodec/nuv: Use ff_set_dimensions()
[21:02:56 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:9640ed3da01d: avformat/mxfdec: Clear metadata_sets_count in mxf_read_close()
[21:02:57 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:4bfb9fa3fcf1: avcodec/vmdaudio: Check chunk counts to avoid integer overflow
[21:02:58 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:f6c63ac83ddc: avcodec/vc1_block: Fix integer overflow in AC rescaling in vc1_decode_i_block_adv()
[21:02:59 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:50dfc2dbf95a: avcodec/wmaprodec: Fix buflen computation in save_bits()
[21:03:00 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:fcc8c187fbde: avcodec/alac: Fix integer overflow in lpc_prediction() with sign
[21:03:01 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:4708c410e21e: avcodec/g729dec: require buf_size to be non 0
[21:03:02 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:7d76759b5999: avcodec/utils: Check sample_rate before opening the decoder
[21:03:03 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:29ea453b05a0: avcodec/utils: Check block_align
[21:04:48 CET] <durandal_1707> why is 2.8 still maintained?
[21:09:33 CET] <williamto> Hi guys, today I tried to profile `ffmpeg` with scaling option on an aarch64 system and found a high overhead in a load instruction.
[21:09:35 CET] <williamto> https://imgur.com/a/PuWLniP
[21:09:55 CET] <williamto> I was wondering if there would be optimization opportunities for me in this part?
[21:10:22 CET] <JEEB> that's probably swscale if it's vf_scale
[21:10:44 CET] <williamto> the function is called `ff_hscale_8_to_15_neon`
[21:11:21 CET] <JEEB> ah, so then it already has NEON... then it depends on how many chances there are to optimize it further
[21:13:00 CET] <michaelni> durandal_1707, because some distros use it, see https://trac.ffmpeg.org/wiki/Downstreams
[21:13:31 CET] <williamto> also I was wondering if it is normal for the `add` instruction to have high overhead
[21:20:57 CET] <J_Darnley> I can't comment on arm instructions but if it loads from memory then it could be
[21:21:46 CET] <J_Darnley> or maybe there's a bug like on x86 that causes an off by 1 error in theww profile
[21:22:15 CET] <Lynne> its arm, they can't touch memory without a load/store instruction
[21:23:58 CET] <Lynne> perf reports for which instruction took the most amount of time seemed random to me under aarch64, including on in-order CPUs
[21:24:13 CET] <williamto> oh really?
[21:26:34 CET] <williamto> i realized most overheads are mostly in load instructions
[21:26:45 CET] <williamto> they are probably hardware-related constraints
[21:27:43 CET] <williamto> on aarch64 systems that is
[21:28:10 CET] <williamto> is there any other parts of the ffmpeg code where I can find potential opportunities for optimization?
[21:37:07 CET] <cone-637> ffmpeg 03Michael Niedermayer 07release/2.8:5df3efbdd6bf: Update for 2.8.16
[00:00:00 CET] --- Sat Dec  7 2019