![]() | |||||||||||||
Messages in this thread |
> > This is not really a Linux related question, as I experienced it first > > under Solaris. But it seems, that the linux behavior WRT this problem > > changed in recent dev. kernels, so I thought, maybe someone knows the > > answer. > > > > The symptom is that a process is Kill-ed, when the file descriptors 0, > > 1 or 2 are close()-ed (or dup2()-ed). These file descriptors are > > actually connected to /dev/null before closing. > > > > The circumstances are rather foggy, while this all happens in a shared > > library, which overrides some libc functions. This actually happens > > inside an opendir() syscall made from tcsh, when TAB is pressed for > > filename completion. > > > > The following code and test results show that what you observe is not > what is happening. Try it yourself. No, of course not. I didn't say it was the normal behavior of programs to be killed when a file descriptor is closed. I'm actually interested in HOW can it happen at all (as it _does_ happen). What I don't see is, how can a process get a KILL signal from kernel at all (except of course by the kill() syscall). Miklos Szeredi - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/faq.html | ||||||||||||
| Last update: 2005-03-22 13:44 [from the cache] ©2003-2008 | |||||||||||||