Subject: Re: README: VM stats programs need recompile
To: None <Thilo.Manske@HEH.Uni-Oldenburg.DE>
From: Jason Thorpe <thorpej@nas.nasa.gov>
List: current-users
Date: 07/03/1999 14:13:02
On Sat, 3 Jul 1999 02:37:57 +0200
Thilo Manske <Thilo.Manske@HEH.Uni-Oldenburg.DE> wrote:
> UBC branch?!? What's that?
>
> I suppose that UBC means something like "Unified Buffer Cache",
> but I didn't know that there's already code for/in NetBSD. (Cool!)
> Have I missed something?
{Unified,UVM} Buffer Cache, sure :-)
The name is slightly borrowed from Digital UNIX's subsystem of the
same functionality (well, similar; UBC will actually be able to handle
some cases that the Digital UNIX UBC can't :-)
There's been some discussion of it on tech-kern, and there have
been quite a few commits to the chs-ubc and chs-ubc2 branches of the
source tree.
Chuck Silvers has been doing all the work. What I've done on it so
far is sync the branch w/ -current (so that I can hack on it :-) and
do a few minor things that make non-converted file systems continue
to work (with the old buffer cache).
> I really would like to know do when we will see UBC in -current
> (maybe optional) or if patches for -current are available anywhere.
I doubt patches against -current will ever be distributed; it's
fairly invasive in some places. This is also why there isn't simply
#ifdef UBC all over.
However, once anoncvs is finished being deployed (there *was* an
experimental server, but it's being moved to a production server),
you'll be able to check out the chs-ubc2 branch of the syssrc module.
> BTW: I don't find any hints to the "UBC project" on the NetBSD
> Project Server (http://mail-index.netbsd.org/cgi-bin/projects.cgi).
Nope, you're right. The author evidently didn't choose to list
it there.
-- Jason R. Thorpe <thorpej@nas.nasa.gov>