Messages in this thread | | | From | Richard Henderson <> | Subject | Re: Killing clones | Date | Wed, 13 Aug 1997 10:00:18 -0700 (PDT) |
| |
> Yeah, well since I didn't like the idea of a process waking up > periodically, I've implemented another scheme. The parent sets up a > pipe, I close the write end in the threads and do a select(2) for > input in the threads. When the parent dies, the select will > return. The child threads can then check if ppid has changed. It just > means I can no longer clone FDs, but that should not be a problem (I > don't expect to do I/O in the threads, just computations).
Hum. Well, that works for you, but not for LinuxThreads -- folks are going to expect to be able to share FDs.
I have another idea, that I'd like to bounce off folks, though:
A new flag, CLONE_SWAP, that reverses the sense of the return value and reverses who gets the new stack. This would make things so that the manager thread we were spawning becomes the parent, while the main thread becomes the child.
This could possibly break things in the main thread if someone was caching and relying on the thread's pid, but I can't think of what might want to do that.
r~
| |