On 2024/11/23 17:06, Kirill A. Korinsky wrote:
> >
> > Where is it frozen? On which sleep channel is ffplay sleeping? Can you
> > enter ddb and get a trace of the ffplay process?
> >
>
> When ffplay is complaining ioctl(VIDIOC_DQBUF): Invalid argument, it creates
> a few threads and everything works. When I press C-c on terminal or use
> pkill -KILL ffplay nothing happens, but ffplay changes its state via ps from
>
> ~ $ ps aux | grep fplay
> catap 20332 0.8 0.3 35204 53348 p2 S+ 2:05PM 0:00.51
> ffplay -f v4l2 -i /dev/video1
> catap 77165 0.0 0.0 796 1376 p4 S+p 2:05PM 0:00.01
> grep fplay
> ~ $
>
> into
>
> $ ps aux | grep ffplay
> catap 20332 0.0 0.3 29280 47628 p2 RE+/2 2:05PM 0:00.63
> (ffplay)
> catap 97615 0.0 0.0 100 360 C0 R+/2 2:07PM 0:00.00
> grep ffplay
> $
>
> and there is no ffplay processes by ps in ddb anymore.
that's odd! The 'wait' column in top(1) would be of interest, that's the
wait aka sleep channel.
> and there is no ffplay processes by ps in ddb anymore. I see only uvideo1
> which has stacktrace:
>
> ddb{1}> tr /t 0t393427
> sleep_finish(0,1) at sleep_finish+0x17f
> usbd_transfer(fffffdb461bb4c08) at usbd_transfer+0x275
> uvideo_vs_start_bulk_thread(ffff800001b25000) at
> uvideo_vs_start_bulk_thread+0x7e
> end trace frame: 0x0, count: -3
> ddb{1}>
>
> and when I deattach the webcam from the machine, it crashes with stacktrace:
>
> uvm_fault(0xfffffdb32e87f840, 0x3f, 0, 2) -> e
> kernel: page fault trap, code=2
> Stopped at usbd_close_pipe+0x91: movq %rcx,0x40(%rax)
> TID PID UID PRFLAGS PFLAGS CPU COMMAND
> 420740 59811 1000 0x100002 0 0 xconsole
> 421818 10645 0 0x100000 0x80 3 hotplugd
> 205493 61124 0 0x14000 0x200 1 reaper
> usbd_close_pipe(ffff800001c99000) at usbd_close_pipe+0x91
> uvideo_close(ffff8000019bf000) at uvideo_close+0x57
> videoclose(2c81,3,2000,ffff80004d2c16c8) at videoclose+0x63
> spec_close(ffff80004bf56488) at spec_close+0x107
> VOP_CLOSE(fffffdd4518e6e78,3,fffffdd3e18fe480,ffff80004d2c16c8) at
> VOP_CLOSE+0x59
> vn_closefile(fffffdd825a0edb58,ffff80004d2c16c8) at vn_closefile+0xaf
> fdrop(fffffdd825a0edb58,ffff80004d2c16c8) at fdrop+0xa5
> closef(fffffdd825a0edb58,ffff80004d2c16c8) at closef+0xaf
> fdfree(ffff80004d2c16c8) at fdfree+0xa6
> exit1(ffff80004d2c16c8,7b,0,1) at exit1+0x29b
> sys_exit(ffff80004d2c16c8,ffff80004bf56780,ffff80004bf56f80) at
> sys_exit+0x1a
> syscall(ffff80004bf56780) at syscall+0x5e7
> Xsyscall() at Xsyscall+0x128
> end of kernel
> end trace frame: 0x761dc4ff9130, count: 2
> https://www.openbsd.org/ddb.html describes the minimum info required
> in bug
> reports. Insufficient info makes it difficult to find and fix bugs.
> ddb(2)>
>
> Here, I double-checked the traces and it shouldn't contain any typos/errors.
>
> Also, if I wait a few minutes after C-c or kill ffpplay, I can deattach the
> webcam without crashing, and the ffplay process just disappears.
>
> --
> wbr, Kirill
>
