Re: /dev/null doesn't get created

On Sunday 01 August 2004 05:09 pm, Derek Tattersall wrote:
> * Mark Murray (mark_at_grondar.org) [040801 17:03]:
> > From: Mark Murray <mark_at_grondar.org>
> > Message-Id: <200408011825.i71IPZgG018406_at_grimreaper.grondar.org>
> > To: dlt_at_mebtel.net
> > Cc: current_at_FreeBSD.ORG
> > Subject: Re: /dev/null doesn't get created
> > Date: Sun, 01 Aug 2004 19:25:34 +0100
> >
> > Derek Tattersall writes:
> > > CURRENT as of this morning dint'create /dev/null.  My previous
> > > CURRENT, from July 25 created it fine so all was well.Can anybody
> > > offer me a clue as to what this is all about?  I tried to figure out
> > > what causes /dev/null to get created, but was unsuccessful.
> >
> > BOOBOO of mine. Please always check src/UPDATING for clues :-)
> >
> > M
> > --
> > Mark Murray
> > iumop ap!sdn w,I idlaH
>
> Well I appear to have cvsup'ed just before you added the text to
> UPDATING.  Oh well.  That brings up an interesting question: Under
> what circumstances would it be advantageous to not compile the drivers
> into the kernel, but leave them as modules and load from
> /boot/loader.conf?  I haven't had any luck imagining why you would
> want to leave them as modules.

It's a preference thing.  I pretty much never use modules myself (things 
compiled into the kernel are more optimized, mutex and atomic ops are inlined 
rather than function calls for example) unless doing development on a device 
driver in which case I leave the device driver as a single module.

-- 
John Baldwin <jhb_at_FreeBSD.org>  <><  http://www.FreeBSD.org/~jhb/
"Power Users Use the Power to Serve"  =  http://www.FreeBSD.org