Re: Deadlock resolution

["Nick Roberts" <nick.roberts@acm.org>]

On Tue, 27 Jul 2004 18:52:11 +0200, Jano <notelacreas@porfavor.no> wrote:

> ...
> If I were extra paranoid about certain tasks and were to detect  
> deadlocks, as a first though I suppose I would use the ATC
> construct, ...

I think there is generally great value in using an ATC to provide
a timeout to a call to a service which is external (to the program
or module), since this helps to ensure that your program or module
is not stymied by something outside its control, as well as
protecting against 'big ring' deadlock.

I find that the gory details can be neatly hidden inside a
procedure, to which you pass -- in addition to other relevant
parameters -- the timeout value, which can have a convenient
default.

> Are queued entry calls abortable? I seem to remember something
> about that.

There are various rules and caveats. An entry call can be
cancelled if it is still waiting in the queue, but the body of an
accept statement is an 'abort deferred region', and cannot be
aborted as such. An ATC is not permitted within any abort deferred
region, which includes Initialize, Finalize, and Adjust. And so
on.

> I find the client (tasks)/server (monitors) architecture safe
> and convenient, if you can stick to it...

Indeed.

-- 
Thanks for your reply,
Nick Roberts