Feedback stress-testing for comedi

Hello,

I decided to write a "functional" test routine for comedi. I was inspired by a 
few recent bug reports (e.g., from Jack Culpepper and Markus Dostal) and a 
general, long-standing, and occasionally useful paranoia about making sure my 
data acquisition system is doing what I think it is. Now that I've run it on 
my systems, I can sleep easily...

To make this useful, you need to have both inputs and outputs available 
(either analog or digital), not necessarily on the same board (in theory, at 
least; I only have one board per machine so haven't tested this). The basic 
idea is to generate outputs and measure them with the inputs; since you 
generated them yourself, you know what the answer is supposed to be, and you 
can compare the two. While we all do that "by eye," an automated system can 
check several hours of data with a high degree of accuracy.

I've implemented a somewhat flexible function for doing this. It handles both 
digital and analog channels, single-sample and streaming types (sorry, 
digital streaming isn't yet implemented, since I don't use it and don't even 
have a board which can do it). You can do multiple channels simultaneously; 
the waveforms it generates are unique for each channel. Currently you need 
one output channel for each input; it might be worth generalizing this code 
to allow multiple inputs to "listen" to the same output, but because of other 
duties I am unlikely to implement this in the near future. Since this code is 
already useful in its current form, I thought I'd share it with the list. 
While it works well on my systems when I (the author) am running it, I expect 
other users will find bugs; please let me know when you find them.

It's a single file, which you can compile with
# gcc -o feedback feedback.c -lm -lcomedi

Execute
# feedback -h
to find out how to use it and to get a sense for the options available.

If it seems generally useful, feel free to include it in comedlib/testing. 
Since it makes fairly extensive use of comedilib, it's a handy complement to 
the routines in that directory.

Best,

-- 
Tim Holy
Asst. Prof. of Neurobiology
Washington University School of Medicine
Department of Anatomy and Neurobiology
4401 North Building
Campus Box 8108, 660 S. Euclid Avenue
St. Louis, MO 63110-1093
tel: 314-362-0086
fax: 314-362-3446
email: holy_at_pcg.wustl.edu