Re: caching and userapi

[Matthew Dillon <dillon@xxxxxxxxxxxxxxxxxxxx>]

:Is it possible to start work on the caching and userapi without the
:threads and messaging work being done?

    It depends what you mean by 'the caching'.  I would say, in general,
    that work on per-cpu caches for certain critical subsystems (anything
    except the VFS name cache which is going to be shredded by the VFS
    work) could be started without the messaging, but remember the MP lock
    is going to remain intact for several months to come so the caches
    will have only a very limited effect on performance.   Still, the 
    caches *will* be almost fully testable even with the MP lock so there
    is no reason to hold off putting together a plan for what to cache.

    The messaging syscall interface (is that what you mean by userapi?)
    should probably wait until I get DEV done along with the initial
    kernel messaging infrastructure.  I will then do a pass on the syscall
    interface (to get the really nasty stuff that requires touching dozens
    of files out of the way), after which it would be possible to start
    working on userland threads support.

					-Matt
					Matthew Dillon 
					<dillon@xxxxxxxxxxxxx>