Jos> I total agree with you on the profits of modules, but what I mean
Jos> is not that it should not be allowed to keep them in the kernel
Jos> tree, but that the interface with the kernel is more structured
Jos> and more flexible, in such way that if someone (for example a
Jos> commercial company) wants to deliver a driver, he has a
Jos> welldefined interface towards the kernel. And indeed, some
Jos> kernel-structures change sometimes, but I think that at least in
Jos> the stable kernel branch, these changes are not frequent enough
Jos> to implicit that a different driver must be compiled for every
Jos> kernel.
This has been discussed and the conclusion was basically that if
something needs to be changed to solve a problem, it has to be
changed, whether it is a stable kernel or not is irrelevant. This
happend some time during 2.0.x for instance.
The point is that we are not willing to pay the price of guaranteeing
an API just for the sake of making life easier for people who wants to
do binary only drivers. The solution is simple, the people who really
wants binary only drivers can implement a thin wrapper layer and Open
Source it and then allow people to change it if a kernel interface
changes.
Jes
-
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/