master/slave translators (was IPv6)

[Kalle Niemitalo <tosi@ees2.oulu.fi>]

On Wed, Sep 13, 2000 at 10:44:18AM +0200, Mark Kettenis wrote:
> There is a possible race when
> both passive translators are activated simultaniously and I'm not 100%
> sure the mechanism used by /hurd/term protects against that.

Would it be better to set up the nodes as master and slave too?

- An application opens the slave node.
- The file system starts the translator of the slave node.
- The slave translator opens the master node.
- The file system starts the translator of the master node.
- The master translator sets itself as the active translator of the slave
  node, but does not kill the slave process.
- The slave translator now has a send right to the master translator.
- The slave translator asks the master translator to take over listening
  to the port which the application got when it opened the slave node.
- The master translator says ok, and starts serving the application.
- The slave translator exits.

Is this doable and free of races?


-- 
To UNSUBSCRIBE, email to debian-hurd-request@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org