Re: modutils, the next generation

Richard Henderson (richard@atheist.tamu.edu)
Tue, 14 Jan 1997 10:34:44 -0600 (CST)


> I think the crux of the matter is that one really does care that only
> proper modules get loaded into the kernel -- the result of loading a
> .o that *wasn't* a module would be ugly indeed.

Without conciously designing one, I don't think you'll just
run across a random object file that will pass initial sanity
checks to be loaded as a module.

> Better would be if
> modules were fully version-locked in some way, so that there was a
> high probability that if a module was incompatible with a particular
> kernel, it would not load at all.

There is a spot in both the old and new module formats to store
the version of the kernel that the module was built against. This
value _is_ checked before anything else happens.

> And, ideally, if the kernel were
> changed but still compatible with a particular module, the module
> could be loaded unchanged. (Unfortunately, I don't understand the
> module system well enough to know to what degree MODVERSIONS takes
> care of this.)

To an absurdly high degree of probability, MODVERSIONS does let
this happen safely.

r~