signed integer overflow at libswscale/output.c:2281

[Suhwan]

Summary of the bug:
There're two signed integer overflow at libswscale/output.c:2281

I compiled ffmpeg with "--toolchain=clang-usan" to check the undefined-behaviours and attached log file.
How to reproduce:

% ffmpeg_g -y -i $PoC -target vcd -c fits tmp.rawvideo

ffmpeg version N-95517-g0b8956b25c Copyright (c) 2000-2019 the FFmpeg developers
built with clang version 6.0.0-1ubuntu2 (tags/RELEASE_600/final)
configuration: --cc=clang --cxx=clang++ --ld=clang --enable-debug --toolchain=clang-usan

Here's UBSAN log

libswscale/output.c:2281:35: runtime error: signed integer overflow: 524280 * 4432 cannot be represented in type 'int'

Thread 1 "ffmpeg_g" hit Breakpoint 1, 0x00000000004288b0 in __ubsan::ScopedReport::~ScopedReport() ()
(gdb) bt
#0  0x00000000004288b0 in __ubsan::ScopedReport::~ScopedReport() ()
#1  0x000000000042b0eb in void handleIntegerOverflowImpl<__ubsan::Value>(__ubsan::OverflowData*, unsigned long, char const*, __ubsan::Value, __ubsan::ReportOptions) ()
#2  0x000000000042c9df in __ubsan_handle_mul_overflow ()
#3  0x0000000005706ada in yuv2gbrp16_full_X_c (c=0x93fee00, lumFilter=0x9416080, lumSrcx=0x9419000, 
    lumFilterSize=4, chrFilter=0x9413280, chrUSrcx=0x93fdec0, chrVSrcx=0x9419d00, chrFilterSize=4, 
    alpSrcx=0x9419e40, dest=0x7fffffffbba0, dstW=352, y=0) at libswscale/output.c:2281
#4  0x000000000556487c in any_vscale (c=<optimized out>, desc=<optimized out>, sliceY=<optimized out>, 
    sliceH=<optimized out>) at libswscale/vscale.c:198
#5  0x0000000005529255 in swscale (c=<optimized out>, src=<optimized out>, srcStride=<optimized out>, 
    srcSliceY=<optimized out>, srcSliceH=<optimized out>, dst=<optimized out>, dstStride=<optimized out>)
    at libswscale/swscale.c:503
#6  0x00000000055337c2 in sws_scale (c=0x93fee00, srcSlice=0x9408af4, srcStride=0x7fffffffc1d0, srcSliceY=0, 
    srcSliceH=60, dst=<optimized out>, dstStride=0x7fffffffc1c0) at libswscale/swscale.c:978
#7  0x0000000000d361ee in scale_slice (link=<optimized out>, out_buf=<optimized out>, cur_pic=<optimized out>, 
    sws=<optimized out>, y=<optimized out>, h=<optimized out>, mul=<optimized out>, field=<optimized out>)
    at libavfilter/vf_scale.c:411
#8  0x0000000000d31c9d in scale_frame (link=<optimized out>, in=0x9429080, frame_out=<optimized out>)
    at libavfilter/vf_scale.c:535
#9  0x0000000000d2f20d in filter_frame (link=0x93fb3c0, in=0x9429080) at libavfilter/vf_scale.c:549
#10 0x00000000005d002d in ff_filter_frame_framed (link=<optimized out>, frame=0x0)
    at libavfilter/avfilter.c:1084
#11 ff_filter_frame_to_filter (link=<optimized out>) at libavfilter/avfilter.c:1232
#12 ff_filter_activate_default (filter=<optimized out>) at libavfilter/avfilter.c:1281
#13 ff_filter_activate (filter=<optimized out>) at libavfilter/avfilter.c:1443
#14 0x00000000005f444f in push_frame (graph=0x93c7ec0) at libavfilter/buffersrc.c:187
#15 av_buffersrc_add_frame_internal (ctx=<optimized out>, frame=0x93f5700, flags=<optimized out>)
    at libavfilter/buffersrc.c:261
#16 0x00000000005f300e in av_buffersrc_add_frame_flags (ctx=0x93fac40, frame=0x93f5700, flags=4)
    at libavfilter/buffersrc.c:170
#17 0x00000000004cad80 in ifilter_send_frame (ifilter=<optimized out>, frame=<optimized out>)
    at fftools/ffmpeg.c:2186
#18 send_frame_to_filters (ist=0x93ca880, decoded_frame=0x93f5700) at fftools/ffmpeg.c:2260
#19 0x00000000004a07dc in decode_video (ist=<optimized out>, pkt=<optimized out>, got_output=<optimized out>, 
    duration_pts=<optimized out>, eof=<optimized out>, decode_failed=<optimized out>) at fftools/ffmpeg.c:2459
#20 process_input_packet (ist=<optimized out>, pkt=<optimized out>, no_eof=<optimized out>)
    at fftools/ffmpeg.c:2613
#21 0x00000000004ba058 in process_input (file_index=0) at fftools/ffmpeg.c:4303
#22 0x000000000048d60b in transcode_step () at fftools/ffmpeg.c:4628
#23 transcode () at fftools/ffmpeg.c:4682
#24 0x0000000000487dc4 in main (argc=11, argv=<optimized out>) at fftools/ffmpeg.c:4884
(gdb) c
Continuing.
libswscale/output.c:2281:19: runtime error: signed integer overflow: -1073741824 + -1971358336 cannot be represented in type 'int'

Thread 1 "ffmpeg_g" hit Breakpoint 1, 0x00000000004288b0 in __ubsan::ScopedReport::~ScopedReport() ()
(gdb) q

Please confirm.
Thanks

[Suhwan]

poc

[Balling]

​https://patchwork.ffmpeg.org/project/ffmpeg/patch/20200216194303.819-4-michael@niedermayer.cc/

[Carl Eugen Hoyos]

Fixed by Michael in 49ba1879add99d3f64d70d34fb0255c8a49d4b28