Segfault when ogg123 exits

Hi,

It seems ogg123 segfaults when it finishes playing. Note that I'm not pressing ^C -- I'm not even sure whether the "^C bug" is still around.

reported with SDL_Quit() and timer_create():

#0  0x00000000 in ?? ()
#1  0x400a7aac in _dl_close_worker () from /lib/ld.so.1
#2  0x400a7c08 in _dl_close () from /lib/ld.so.1
#3  0x400190dc in dlclose_doit () from /lib/libdl.so.2
#4  0x400a1834 in _dl_catch_error () from /lib/ld.so.1
#5  0x40019590 in _dlerror_run () from /lib/libdl.so.2
#6  0x40019118 in dlclose () from /lib/libdl.so.2
#7  0x402f9ffc in ao_shutdown () from /usr/lib/libao.so.2
#8  0x40c57bd4 in exit () from /lib/libc.so.6
#9  0x000182fc in main ()

Would any glibc hackers feel like taking a closer look at this?

Cheers,

Stuart Brady

Can you distill this into a smaller testcase? Maybe a call to dlopen() to load a library that has a function called by timer_create?

Cheers,
Carlos.

I just ran ogg123 in gdb with a breakpoint set on timer_create(), and it didn't get triggered, so timer_create() itself seems unlikely to be the problem, but I can certainly try chopping libao down a bit.

BTW, this happens with both the OSS and ALSA plugins.

Side note: I've found that while running 'while (true); do true; done', mmapped ALSA output seems to work! If I drop the brackets, I just get noise -- presumably because it doesn't fork without them.

Stuart Brady

Excellent, this is exactly the way you want to track this down. You may wish to run strace to see what types of system calls the program is making.

Cheers,
Carlos.