Re: problem with MFS and init

[David Young <dyoung%pobox.com@localhost>]

On Tue, Nov 01, 2005 at 05:18:10PM -0600, David Young wrote:
> On Tue, Nov 01, 2005 at 08:54:06PM +0100, Jachym Holecek wrote:
> > Hello,
> > 
> > > If you don't have /dev populated, init is supposed to "make one" in an 
> > > MFS filesystem.
> > 
> > Yup, there's a compile-time option for this.
> > 
> > > I probably should file this as a PR, but I wanted to inquire before I 
> > > did that.  FWIW my unit has ~192MB, so its not RAM that is exhausted -- 
> > > just inodes in the MFS.
> > 
> > Bump NINODE in sbin/init/init.c (1024 currently). IMO, the right solution
> > here is to teach init(8) to create /dev on tmpfs instead of mfs
> > (tmpfs can grow and shrink as needed, unlike mfs). This is somewhere
> > deep down on my TODO list...
> 
> I did that last night.  It's pretty simple.  Here is a patch.

I found some bugs in tmpfs /dev/.  Update tmpfs to HEAD to get the fixes
by yamt and I.

Dave

-- 
David Young             OJC Technologies
dyoung%ojctech.com@localhost      Urbana, IL * (217) 278-3933