hi,

> Here's a snapshot of my buffer cache implementation changes. Included
> are MD patches for archs named a* + i386 + sparc. The resulting kernel has
> run on i386 & sparc.
> 
> I would appreciate it if port-masters could look over the MD changes
> (which mostly boil down to deleting calls to allocsys() and the special
> buffer memory mapping code) and verify that it does the right thing.
> 
> -pk

currently, the MD part deals with some allocation policies,
eg. preferablly allocate buffers from DMAable memory.
how do you think about it?
i guess "file" page caches can use the same or similar mechanism.

YAMAMOTO Takashi