Messages in this thread | | | From | David Wragg <> | Date | Thu, 11 Feb 1999 01:13:14 GMT | Subject | Re: clone, signal and pids |
| |
Luis Irun <lirun@iti.upv.es> writes: Mime-Version: 1.0 (generated by tm-edit 7.108) Content-Type: text/plain; charset=US-ASCII From: David Wragg <dpw@sytry.doc.ic.ac.uk> Date: 11 Feb 1999 01:13:14 +0000 Message-ID: <y7r4sotspxx.fsf@sytry.doc.ic.ac.uk> Lines: 21 X-Mailer: Gnus v5.5/XEmacs 20.3 - "Vatican City"
> The matter is: I don't agree the > signal behaviour for clones sharing its pid's.
The impression I've got from the code is that CLONE_PID isn't supposed to have any particular semantics, as long as it doesn't expose security holes (which it diesn't seem to). I assume sensible semantics were supposed to be implemented at a later date, but no-one has got round to it. The obvious user-space semantics are those for POSIX thread signals, but it isn't quite so obvious which kernel additions would be most useful for a low-overhead user-space implementation.
> [snip] > I think the wrong step is allowing the clones to share its pid's.
I suspect no-one has actually made real-world use of CLONE_PID from user-space. It is however used inside the kernel (creating the idle task for each processor on SMP, which should all have the same pid of 0).
Dave Wragg
- 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/
| |