>Why not just provide a way for the kernel to unload some baggage,
   >maybe make the GENERIC kernel do mostly lkms?

uh.  "just".  huh?
   
   However, LKMs are an orthogonal issue because they won't solve the problem
   I'm asking after: architecture specific kernels that eliminate the need for
   indirection through function pointers in critical (i.e. very often used)
   code paths in the kernel, and thereby perform better than GENERIC. We have
   a model of use which presumes that GENERIC is to get you bootstrapped and
   the system installed, and then you compile up a kernel that more closely
   matches your needs.

we used to have SUN4, SUN4C and SUN4M kernels as i recall.. perhaps
if they were properly generated from GENERIC it would be great.
   
   What this really comes down to is: do typical users
   configure/compile/install an arch-specific kernel after install, or do they
   just run GENERIC in production? If the former, then my suggestion is
   useless. If the latter, then I think it would be helpful if we could give
   the users a step-up in the right direction.

i bet that there are a large number of users who do both.