[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: Linux drivers, RT linux
- To: tass@wwa.com
- Subject: Re: Linux drivers, RT linux
- From: hjohnson@pluto.njcc.com (Herbert R Johnson)
- Date: Thu, 28 Aug 1997 15:31:36 -0400
- Old-Return-Path: <hjohnson@pluto.njcc.com>
- Organization: NJ Computer Connection for Astro Imaging Systems
- Reply-To: hjohnson@pluto.njcc.com
- Resent-Date: Thu, 28 Aug 1997 15:39:44 -0400 (EDT)
- Resent-From: tass@wwa.com
- Resent-Message-ID: <"xo2HU.A.LYF.fOdB0"@kani.wwa.com>
- Resent-Sender: tass-request@wwa.com
Aug 28 1997
I've been working on a Linux 486 system using the recommended Red Hat
distribution on the Tri-Linux CD's as previously described. I
have some "alpha" notes that Chris has commented on, mostly some
history and some how-to-installs. The pressures of other projects and my
chasing down various Linux/RT issues have delayed that document. But
I'm going to ask Michael Richmond to archive it on the TASS Web
page as an "alpha" release, even as I work on it. It's too
big to distribute by email and it will change anyway.
Chris has worked through the other Linux distributions on the recommended
Tri-Linux CD set (Red Hat, Slackware, Debian) and says the Red Hat
installation is the "least painful" of the three.
Thanks to Eric Thiele, who is building robots with Linux, it would appear
that the 2.0.30 kernal version of Linux "can" work
with the Real Time kernal patches developed for 2.0.29. This would
save us the trouble of replacing the 2.0.29 sources with 2.0.30
kernal sources from one of the other CD's in the Tri-Linux disks.
The 2.0.29 patches are "accepted" by the patching procedure opon the 2.0.30
kernel sources that Eric used and they appear to work.
Testing procedures seem to be "try it and see", so it's a matter
of time and sufficient trials, including by us TASS members. This may
be complicated by the fact that not all sets of kernal sources
are equal: each distributor may provide some slightly different
set of patches to those sources. Only a download from the archive
site is "pure". (sigh) The ancient computing metaphor for fighting this
is "nailing jelly to a tree". This is where I'm spending my time,
trying to resolve all this. The Tech Note for all this will be long
enough to cover all the possibilities.
This business of the "right" kernal is a real challenge, I've found.
Once a kernal's sources are patched successfully and the compiled
kernel runs OK, including the RT example programs, that seems to establish
a "successful" upgrade. Chris wants to make this work on "any clunker
PC system" so he's not going to make changes that will be hard to support
later on, whatever kernals come up.
Here's what else Chris says about all this:
"I am very unlikely
to make big changes to the RT portion of the driver. If I do and I
hear it causes problems on a 2.0.30 1/2 based RT system I'll fix it.
The RT portion of the system amounts to about 10% of the total and
has been tested out by Nick. It is basically finished now.
Most of the [TASS] driver is just normal C code that would compile and
run even on a Sun under Solaris. Let's say if the [RT] examples included
with the RT patch work then the driver will work. In other words
if someone has a system where the examples work and the driver fails
I'll fix the driver. The driver doesn't do anything that is not in
the example code so I can feel safe in saying that."
Meanwhile, of course, both Linux and the Real Time patches are
under further development. The Linux world is always in a state
of flux: you can stand still, but not for too long. But for my
Tech Note, I hope I can "nail down" both of these long enough for
our use. The virtue of having Linux on CD - with both of these
kernal versions - is that we will have convenient access to
these two Linux versions in any case. The RT patches are available
from the developer's Web site: it would be prudent for anyone
interested to download and archive the relevant files for both
Linux kernal 2.0.27 and 2.0.29. (The files for 2.0.27 are complete
for that kernal version; the files for 2.0.29 is a slightly different
kernal patch file for that Linux version, otherwise the RT files for 2.0.27
are to be used. The Tech note will have details.)
However, Chris tells me: " Today on the Redhat FTP site is a beta
version of RH4.8 (they skipped a few numbers) RH4.2 is still the shipping
version but they are letting people try out RH4.8 as of today.
I looked and it uses the same 2.0.30 kernel as RH4.2" I said "this
is like nailing jelly to a tree". Chris says it's more like "herding
cats".
The TASS driver code is not yet available on-line. Ask Chris if
you need it: eventually he'll find an archive place for it.
In the longer term, Chris reports that RT development will move
forward to more advanced versions of the Linux kernal, so some
new features may only be available to those who install later
kernals. (This is typical in the Linux world.) In addition, there
is a suggestion that these RT features may eventually be offered as another
"official variation" of Linux, like Linux for the Alpha or the 68K
processor, etc. perhaps in a year or so. So Chris made
an excellent decision in selecting the RT enhancements for his
TASS Mark III driver.
Meanwhile, in the MS-DOS world Caldera (www.caldera.com) has
offered a version of MS-DOS derived from Digital Research's
DR-DOS that supports networking, and may have other useful features.
This is available for downloading from their site, subject to
filling out an on-line form that permits educational, non-profit,
or personal use. Check their site for details. I don't have the
time to see if would support real-time stuff. It has been
suggested to me (maybe by Chris?) that Microsoft's version
of MS-DOS for Windows 95 (sometimes called MS-DOS version 7)
may also support some real-time and networking features. But of course
any TASS software would have to be written to make use of them, if
such features exist.
And finally, Chris is offering to TASS members some Linux-based tools like
IRAF on a custom CD-ROM. Many thanks, Chris, for this extra effort!
Herb Johnson
***** I do not want bulk email. Automated bulk mailings prohibited. ****
Herbert R. Johnson voice/FAX 609-771-1503 day/nite
hjohnson@pluto.njcc.com Ewing, NJ (near Princeton) USA
occasional amateur electronic astronomer
supporter of classic computers as "Dr. S-100"
and senior engineer at Astro Imaging Systems: old photons to new bits!