> > I wonder if library routines (like dstate_*) are thread-safe? > > I can't speak for the other NUT developers, but I would prefer that we > not open that can of worms. > > Most hardware can't handle multithreaded access without some sort of > mutex, so I am not sure I see why the driver should be multithreaded. > > Can you explain a little more about this driver?
I can but I'm afraid you won't be happy. :) I plan to write a version of "clone" that polls more than one physical UPS drivers. We have 4 UPS with parallel output and some 200 servers with several system administrators. (Some of them cannot be called "guru".) I want to offer a very simple, easy to use, single virtual ups that collect data from real drivers then calculates simulated values and offers a single status variable (OL/OB/LB) that my clients' upsmon. This is the only I can imagine that all of the is able to configure w/o any errors. OK, I can build a non threaded program too but it will be more complex a bit. Gabor -- Mz mn fszm mhrmnc cntm, mk nm mhsz bvhmt. _______________________________________________ Nut-upsdev mailing list [email protected] http://lists.alioth.debian.org/mailman/listinfo/nut-upsdev
