>>>>> "WF" == William O Ferry <WOFerry+@CMU.EDU> writes:

WF>     BTW, why are things done this way?  Are kernel modules always
WF> hard-wired to a specific version like this?  Given how rarely the
WF> modules seem to break, I guess I don't see the point of having it
WF> disable itself if the version doesn't match.  Could this be changed to
WF> simply issuing a *warning* that the version doesn't match, but still
WF> letting you run it?  Seems to me that so long as the module will link
WF> it'll probably work.

I do it this way in AFS because it depends on many system interfaces,
the sorts of things which induce version number changes when they're
altered.

It is a bit more strict than it needs to be, but that's more acceptable
to me than trying to debug a crash from someone with undetected
interface mismatches.

In this case, I gave up on keeping up with 1.2<foo> when <foo> changed
on a less than weekly basis.  When 1.3 comes out, I'll put together a
new module.

==John