*BSD News Article 87302


Return to BSD News archive

Path: euryale.cc.adfa.oz.au!newshost.carno.net.au!harbinger.cc.monash.edu.au!lucy.swin.edu.au!news.rmit.EDU.AU!news.unimelb.EDU.AU!munnari.OZ.AU!news.ecn.uoknor.edu!feed1.news.erols.com!howland.erols.net!news.mathworks.com!fu-berlin.de!irz401!orion.sax.de!uriah.heep!news
From: j@uriah.heep.sax.de (J Wunsch)
Newsgroups: comp.unix.bsd.freebsd.misc
Subject: Re: Embedded FreeBSD
Date: 21 Jan 1997 22:48:33 GMT
Organization: Private BSD site, Dresden
Lines: 62
Message-ID: <5c3h41$mtn@uriah.heep.sax.de>
References: <32B744C0.2DA9@wdc.net> <5bdbd0$uih@news.aus.world.net>
  <32DA62A7.A91@onramp.net> <5be62a$14b@qnx.com> <5bqs4v$oar@uriah.heep.sax.de>
  <5c08m3$dvk@qnx.com> <5c14ih$7nn@uriah.heep.sax.de> <5c2kub$6uu@qnx.com>
Reply-To: joerg_wunsch@uriah.heep.sax.de (Joerg Wunsch)
NNTP-Posting-Host: localhost.heep.sax.de
Mime-Version: 1.0
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 8bit
X-Newsreader: knews 0.9.6
X-Phone: +49-351-2012 669
X-PGP-Fingerprint: DC 47 E6 E4 FF A6 E9 8F  93 21 E0 7D F9 12 D6 4E
Xref: euryale.cc.adfa.oz.au comp.unix.bsd.freebsd.misc:34405

doug@qnx.com (Doug Santry) wrote:

> >Sure, 2 µs is certainly out of the question with a Unix kernel.  20 µs
> >probably not.  Hey!, my old CP/M BIOS had to pick up one byte after
> >each 16 µs from the FDC -- without DMA. :-)
> 
> yikes!

I forgot: this was a 2.5 MHz Z80.  I really had to count clock cycles,
do manual loop enrolling (macros were fun :), and had to play dirty
tricks like jumping halfwards into the enrolled loop to support
different sector sizes.  IIRC, the worst-case loop time maxed out at
13 or 14 µs...

Btw., just this experience with the scarey NEC765 has finally brought
me into the FreeBSD project. :-)

> BSD is bigger than a meg.  With a 16 meg machine it swaps(running
> X).  The same setup on QNX has megs of free memory.

That surprises me a little.  For a machine running X11, it's not the
kernel but the applications that usually waste the memory, and they
shouldn't be that much different in size for the same architecture.

Of course, if you're using a window system that is not X11, that's a
different thing.

> >Never trust an operating system you don't have sources for. ;-)
> 
> ya ya :-)

Sure, but this .sig is not tailored to QNX. ;) I used to have it for
more than a year now, so it's really a `signature', not just an
opinion only...

> Never trust an operating system with a kernel bigger than 32k! ;-)

That's unfair when comparing a microkernel OS, however.  You have to
at least account for those processes that perform kernel functionality
in a monolithic kernel world.  Though i have no doubts that it won't
be much more than 200...300 KB in QNX.

OTOH, memory is cheap, we aren't running a PDP-11 with hand-wired core
memory.  I still feel much more comfortable to get a ``blahblah sd0
blah UNIT ATTENTION: Bus-device reset, device reset, or power-on
reset'' instead of just being faced with ``ASC(Q): 29,0'', even though
i realize that this collection of strings accounts for some KB of the
mentioned general bloat.

What we sorely need, looking from the monolith-kernel point of view,
is pageable kernel memory.  While this doesn't solve the real-time
problems (so we are heavily diverging from the subject now), it might
help against the waste of physical memory on small machines.  Disks
are even cheaper than memory these days (though the ratio is not that
drastic as it used to be in the PDP-11 era).

-- 
cheers, J"org

joerg_wunsch@uriah.heep.sax.de -- http://www.sax.de/~joerg/ -- NIC: JW11-RIPE
Never trust an operating system you don't have sources for. ;-)