Posts

Showing posts from July, 2006

More adventures in multithreading

Just wanted to capture a few more thoughts; over the weekend, I experienced what I think was my first case of deadlock. My chaosnet-on-OpenMCL implementation had worked just fine when the two sides of a connection were careful to exchange only one packet at a time, taking turns. When I tried to listen to a SEND (the Lisp machine analogue to an instant message), the CADR emulator sent two packets in quick succession, and my OpenMCL process would freeze trying to receive it. As an aside, I get the feeling that I don't really know how to debug this kind of thing in OpenMCL: looking at backtraces does not always make it completely obvious which semaphore or lock is being waited for in the various threads, and I don't think it is possible to look at semaphores to see which thread(s) are waiting. In any case, I concluded that my somewhat haphazard use of a "count lock" to protect access to packet number information in a connection was either being used recursively, or v

Apple II Disk Transfer

It seems that others have also had the desire to upgrade the ADT (Apple Disk Transfer) tool to work with disks other than the Disk II 5-1/4 inch floppy format under DOS 3.3. ADTPro is a ProDOS-based version of ADT. It should presumably work with all block devices recognized by ProDOS. I e-mailed some code to David Schmidt which allowed ADT to work with the ordinary IIgs serial ports; direct access to the Z8530 chip might allow for faster transfers; that's a potential project.

Clickr: Lisp API for Flickr

Mark Probst has developed a Common Lisp API for Flickr and mentioned it on his blog .

More thinking about multithreaded queueing

Looking back on my previous post about multithreaded queues, and based on a little hacking I was able to do last night on my multithreaded-Lisp-user-mode-Chaos-client, and having read a bit more in the Little Book of Semaphores, (must stop typing 'sempahores'!), in particular the turnstile pattern, and the barbershop problem, there are a few other ways to slice the onion. A "waiting room", i.e., the privilege of blocking on the queue is exclusive. That actually makes a certain amount of sense when, for instance, Chaos packets are already sorted by the destination connection. Does it really make sense for multiple consumers to be concurrently transacting on a single connection? What would make sense, in a C10k kind of way, is rapidly generating server connections in response to RFCs, and handing these connections to waiting server threads. Or, if a connection is known to have independent incoming packets, or is deliberately recycled, connections can queue through a tur