Thanks for the reply! I have to say because I'm so much interested in doing some kernel programming, that I didn't pay too much attention to your suggestions, although I did find them very good. [...good information stripped...] > versus kernel is the proper fit for your application. But in general, > you want to try to keep it in userspace (where things are more protected > in case of a stray pointer, etc) if at all possible. This last suggestion, however, was a truly convincing point on what should or should not be in kernel. I realize now that the segmentation faults I see once in a while, when I do something silly in my usual programs, will turn into complete denial of service requiring physical access to the machine for a reboot. I now see your other arguments as much more sensible, even for someone who's "really interested in doing some kernel programming"! I think my approach will be to make a first proof of concept completely in userspace, then move some parts of that program into kernel space. This also allows performance comparison and correct optimization of code (as opposed to the premature type). This also convinces me to use FUSE for my filesystem trials. My personal preference is to avoid FUSE, but I prefer much more to avoid rebooting often! ;) Thank you so much Valdis! Simon _______________________________________________ Kernelnewbies mailing list Kernelnewbies@xxxxxxxxxxxxxxxxx http://lists.kernelnewbies.org/mailman/listinfo/kernelnewbies
