Re: io_close proposal

[Roland McGrath]

> Second worry.  Programs calling exit shouldn't have to wait for
> possibly malicious servers to respond to io_close.  But the same
> synchronization issues happen for close-on-exit.  This is more
> serious, and I'd like to hear some thinking about it before we make
> this (rather significant) interface change.

This is arguably a concern for close as well as exit.
(e.g. a setuid program that closes its excess inherited file descriptors.)

The POSIX.1 synchronization requirements can only be in fact be fs actions
vs reap, not vs exit per se.  I don't know that this is useful, but just an
observation.  (i.e. some weird proc server interaction with the cleanup of
the dead task could be possible, though I haven't though of any such thing
that would be sane.)