Thats also a bad example. A V6 fs has no impact on the rest of the code.
Thats actually an important criteria in two senses
For a driver "How much non specific code do I touch"
- If its just to add my_setup, entries in /proc and call
to the initialiser function this is good.
- If it redesigns pieces of the memory manager expect to have
to do a lot of justification
An example here is the IRDA layer which went in just before 2.2. It
went in because it added only 3 or 4 lines of code to the kernel other than
its own code. That is it doesnt impact other layers - it being good clean
code helped too of course
For a new subsystem the same applies
- If a user needs to add little code to add a driver to the subsystem
its good
- If the interfaces are clear and modular to help this it is good
Free Software slightly re-invents the Software Engineering model in the sense
that because the module is modifiable its better to be clean and cover the
cases you understand not an all encompassing "reusable object". Since you
aren't shipping another bunch of people a dead binary brick, you can let
them handle stuff you completely missed. At the same time they modularity,
clean expandability stuff is very important.
I've actually got a V7fs code set lying here as pieces for SYS51K. Unfortunatelyits author recycled it by accident so there is an extant paper copy only. But
it did its job. It rescued something of great historic interest from an old
disk. More on that when people are ready to announce it ..
Alan
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/