> I brought up this with Linus once, the conversation was in reference
> to how we thought we could do on database benchmarks etc. which is
> essentially all over raw block devices these days. (which we both
> agreed was entirely stupid, the kernel should be doing buffer caching,
> not some silly Oracle disk I/O layer)
Not necessarily. A well-designed database probably has a better idea of
its future disk usage pattern than the kernel does. Remember, LRU is just an
approximation (and sometimes a very poor one) to the optimal FFU.
--- Bob Forsman thoth@gainesville.fl.us http://www.gainesville.fl.us/~thoth/