Undeliverable mail
PMDF Mail Server
Postmaster%TRINCC.BITNET at mitvma.mit.edu
Sat Jul 9 17:08:24 AEST 1988
The message could not be delivered to:
Addressee: TRIN4
Reason:
%MAIL-E-SYNTAX, error parsing 'DJA1::[TRIN4.BOX1024]'
----------------------------------------
Received: from JNET-DAEMON by TRINCC.BITNET; Mon, 4 Jul 88 03:17 EST
Received: From YALEVM(MAILER) by TRINCC with Jnet id 3884 for TRIN4 at TRINCC;
Mon, 4 Jul 88 03:14 EST
Received: by YALEVM (Mailer X1.24) id 3880; Mon, 04 Jul 88 03:12:36 EST
Date: Fri, 1 Jul 88 02:46:03 EST
From: Mike Muuss The Moderator <Info-Unix-Request at BRL.ARPA>
Subject: INFO-UNIX Digest V5#085
Sender: Info-Unix distribution list <I-UNIX at TCSVM.BITNET>
To: Robert Cummings <TRIN4 at TRINCC.BITNET>
Reply-to: INFO-UNIX at BRL.ARPA
Comments: To: INFO-UNIX at BRL.ARPA
INFO-UNIX Digest Fri, 01 Jul 1988 V5#085
Today's Topics:
wanted: L.Wall's MetaConfig
Re: RCS and SCCS
Re: has anybody put $VPATH into mk?
Re: UUCP Over TCP/IP
Re: Questions without answers
Re: Real-time UNIX - what is it & who h
Re: Need utils: head,yes,& nroff (-man) for SysV2 (CCI 2.21or2.22)
Re: vi vs emacs in a student environment
Re: Historical Fact
Re: Question: Driver access to memory mapped devices on 386 SysV.3
Re: plot, graph (summary)
OSF v. ATT/Sun :-)
Re: UUCP Over TCP/IP
Rn on a V.3 machine
NFS/YP alternative
MKS Toolkit/UNIX [r] System, SEA ARC/PKARC, ...
Re: how can I "." in csh?
Re: Question: Driver access to memory mapped devices on 386 SysV.3
-----------------------------------------------------------------
From: Dave Hammond <daveh at marob.masa.com>
Subject: wanted: L.Wall's MetaConfig
Date: 29 Jun 88 15:00:58 GMT
Posted: Wed Jun 29 11:00:58 1988
To: info-unix at brl-sem.arpa
I've been looking for a copy of Larry Wall's _MetaConfig_ for a few weeks
with no success. Searches of the mod.sources and comp.sources listings
have thus far proved fruitless.
Can anyone offer any clues where I might acquire a copy of the program ?
Please E-mail responses, don't post; Thanks in advance for any help.
Dave Hammond
UUCP: ...!marob!daveh
--------------------------------
-----------------------------
From: Joel Clark <joel at intelisc.uucp>
Subject: Re: RCS and SCCS
Date: 29 Jun 88 15:37:58 GMT
Posted: Wed Jun 29 11:37:58 1988
To: info-unix at brl-sem.arpa
In article <10406 at ulysses.homer.nj.att.com> ekrell at hector (Eduardo Krell)
writes:
>In article <290 at intelisc.UUCP> joel at intelisc.UUCP (Joel Clark) writes:
>
>>Can anyone explain to me how a program could store `the most recent version`
>>such that each line in the file does not need to be examined to determine
>>if it is in the most recent version?
>
>You store the most recent version at the beginning of the file in clear
>text followed by the reverse delta to get the previous version
>(followed by the reverse delta to get the version before that, etc.).
>
>Time to get the latest version is thus proportional only to the size of
>that version. Time to get version N is proportional to the size of
>the last version plus the size of all deltas necessary to get from there
>down to version N.
>
> Eduardo Krell AT&T Bell Laboratories, Murray Hill, NJ
>
> UUCP: {ihnp4,ucbvax}!ulysses!ekrell ARPA: ekrell at ulysses.att.com
How well does this work if (as we are) you are working from 3 branches at once.
We keep major branches for Development engineers, Sustaining engineers and
Evaluation all in the same source file. This allows us to proceed with project
that may not go into production for 12-18 months based on the same sources
that we ship, folding in the changes to the production code as they become
solid.
-----------------------------
From: Eric Krohn <krohn at u1100a.uucp>
Subject: Re: has anybody put $VPATH into mk?
Date: 29 Jun 88 21:12:36 GMT
To: info-unix at SEM.BRL.MIL
In article <732 at vsi.UUCP> friedl at vsi.UUCP writes:
] ... about wanting mk (from the AT&T Toolchest) to recognize $VPATH (like
] BSD make) for easier handling of files in multiple directories....
I was looking for this same capability last week and was disappointed
by its absence. Multiple metarules like
%.o: %.c
$CC -c -O $prereq
%.o: ../group/%.c
$CC -c -O $prereq
only produce errors about ambiguous metarules for making foo.o when foo.c
lives in both . and ../group (say I want to modify my own copy of one or two
files from the vi source and then build a new copy from scratch).
Regular expression metarules look promising but I have yet to get the desired
results:
'\(.*\)\.o':R: '\(\.|\.\./group\)/\1\.c'
cc -c -O $stem2/$stem1.c
Mk just says: don't know how to make 'foo.o'. Even if it would work given
sufficient backslashes and parentheses, it would be much too cryptic.
It would be nice if mk supported a $MKPATH or $VPATH directly and if the
"ambiguous metarules" checking were relaxed (maybe using the first applicable
metarule).
I'm almost desperate enough :-) to try GNU make which claims to support both
the %.o: %.c style metarules and $VPATH.
BTW, I have used nmake (also from the AT&T Toolchest) and it does support
multiple source directories through the .SOURCE target and it merely warns
about resulting ambiguities. However, I find that nmake knows too much about
how I "ought" to build my targets when I try to do something unconventional.
I am also a bit repulsed by the extra baggage nmake carries around
(its own cpp, a file of compiled default rules whose source is nearly
incomprehensible (look at the definition of the :: operator!), a compiled
makefile, and a state file). Don't get me wrong: nmake is nice about using
short makefiles when you do typical things and about automatically checking
dependencies on include files.
--
--
Eric J. Krohn
krohn at ctt.ctt.bellcore.com or {bcr,bellcore}!u1100a!krohn
Bell Communications Research, 444 Hoes Ln, Piscataway, NJ 08854
-----------------------------
From: Jeff Roth <ntm1569 at dsacg3.uucp>
Subject: Re: UUCP Over TCP/IP
Date: 30 Jun 88 12:56:05 GMT
To: info-unix at SEM.BRL.MIL
>From article <10036 at marduk.uucp>, by gww at marduk.uucp (Gary Winiger):
[responding to mparker at chip.UUCP's question on the TCP/IP UUCP service]
> Page 2 of L.sys(5) in the 4.3BSD documentation documents a ``caller''
> field value of TCP. That will cause the sender's 4.3BSD UUCP system to
> open a TCP connetion to the receiver's uucpd. The receiver, of course,
> needs to have configured access to uucpd. I don't happen to have a 4.3
> system on hand configured that way to tell you what the inetd.conf line
> has to look like, but I recall it all being in the source directories
> for uucp. (If it isn't already configured.)
from our /etc/inetd.conf:
uucpd stream tcp nowait uucp /etc/uucpd uucpd
from our /etc/services:
uucp 540/tcp uucpd # uucp daemon
from our /usr/lib/uucp/L.sys:
dsacg1 Any TCP uucp dsacg1 ogin:--ogin:--ogin: Udsacg3 ssword: PASSWORD
--
Jeff Roth (osu-cis!dsacg1!jroth) 614-238-9421 (Autovon 850-9421)
>From the Internet: jroth%dsacg1.uucp at daitc.arpa
US Defense Logistics Agency Systems Automation I speak for myself
Center, DSAC-TMP, Box 1605, Columbus, OH 43216
-----------------------------
From: Paul Houtz <gph at hpsemc.hp.com>
Subject: Re: Questions without answers
Date: 29 Jun 88 17:38:17 GMT
To: info-unix at SEM.BRL.MIL
Michael L. Siemon writes:
-->More to the point (I think Brad is being diplomatic), when somebody asks
-->a moderately hard question (that LOOKS easy) we get 20 wrong responses
-->posted to the net, sowing great confusion. Summaries are the way to go!
-->(Admittedly, the phrasing "email me, BECAUSE I don't have time to read
-->this group" is rather stupid.)
Amen, Michael! This has already happened to me, and I asked
a fairly simple question.
-----------------------------
From: Rick Schneider <rick at vsi1.uucp>
Subject: Re: Real-time UNIX - what is it & who h
Date: 30 Jun 88 15:11:54 GMT
To: info-unix at SEM.BRL.MIL
>From article <8188 at brl-smoke.ARPA>, by gwyn at brl-smoke.ARPA (Doug Gwyn ):
> In article <745 at vsi1.UUCP> rick at vsi1.UUCP (Rick Schneider) writes:
>>The only problem with GOULD's RT-UNIX is that it is only intended to
>>run on GOULD machines.
>
> Does this somehow surprise you? What IS your point?
The point is that if you want GOULD's RT-UNIX, you'll have to buy a GOULD
Powernode or NP series machine to run it on.
GOULD has in the past sold source licenses for all of its flavors of UNIX, so
I guess that the source for it RT-UNIX will also be available. If you have a
large budget and like hacking the kernal, buy a low-end Powernode with a
UTX/32-RT source license, and port it wherever you would like.
I worked for GOULD-CSD for several years and from my experience, their
implementations have been excellent, support is above average, and pricing
is in line with the rest of the industry.
--
Rick Schneider ...pyramid!vsi1!rick
VICOM SYSTEMS, INC.; 2520 Junction Avenue, San Jose, Ca, 95134, (408) 432-8660
The opinions expressed are not those of my employer or are they?
-----------------------------
From: Doug Gwyn <gwyn at brl-smoke.arpa>
Subject: Re: Need utils: head,yes,& nroff (-man) for SysV2 (CCI 2.21or2.22)
Date: 30 Jun 88 19:38:42 GMT
To: info-unix at SEM.BRL.MIL
In article <366 at theceg.UUCP> lkb at theceg.UUCP (Lawrence Keith Blische) writes:
> head [-count] [file ...]
But that's a violation of the command syntax standard (if count > 9).
Thus my version omits the -. It does make it hard to deal with files
whose names start with a digit, and is imperfect in other ways.
Actually I always type the direct "sed" command myself:
program | sed 23q
-----------------------------
From: Fred Fish <fnf at fishpond.uucp>
Subject: Re: vi vs emacs in a student environment
Date: 30 Jun 88 15:19:59 GMT
To: info-unix at brl-sem.arpa
In article <6056 at megaron.arizona.edu> lm at megaron.arizona.edu (Larry McVoy)
writes:
>As a consultant I'll volunteer the following advice: don't get people used to
>emacs. Please. Why? Because emacs is available on "some" unix machines.
>Vi is available on almost all unix machines. Old habits die hard, so I think
>it's better to start people out with something they can stay with...
Until recently you could use almost exactly the same argument for NOT teaching
vi, since vi was ONLY available on Unix systems, while some EMACS or workalike
was available on almost any OS. Now there are reasonable vi clones for many
of the more commonly used systems, so I don't think the "people portability"
factor is quite as onesided towards EMACS as it once was. There are also
enough EMACS's available in source form (GNU, microemacs, jove, scame, mg,
etc) that any EMACS addict that has to work for any length of time on a given
Unix system will find some way to get one installed and thus avoid having
to use vi.
-Fred
--
# Fred Fish, 1346 West 10th Place, Tempe, AZ 85281, USA
# noao!nud!fishpond!fnf (602) 921-1113
-----------------------------
From: Doug Gwyn <gwyn at brl-smoke.arpa>
Subject: Re: vi vs emacs in a student environment
Date: 30 Jun 88 19:46:20 GMT
To: info-unix at SEM.BRL.MIL
In article <6056 at megaron.arizona.edu> lm at megaron.arizona.edu (Larry McVoy)
writes:
>As a consultant I'll volunteer the following advice: don't get people used to
>emacs. Please. Why? Because emacs is available on "some" unix machines.
>Vi is available on almost all unix machines. Old habits die hard, so I think
>it's better to start people out with something they can stay with...
If "vi" weren't such a crappy editor this might be good advice.
However, many users spend much of their time text-editing, so they
should use the best editor available rather than suffer with inferior
tools just because they are more universal. (If they really have to
deal with a wide variety of UNIX systems, then it makes more sense to
emphasize universality. It would also make sense in that case to
provide better tools on all those systems.)
I actually do use "vi" on my Sun, until I get "sam" running. (The
SunTools text editor is a joke.) Given a choice between "vi" or an
EMACS variant I'll choose EMACS, but those aren't the only choices.
-----------------------------
From: Ron Natalie <ron at topaz.rutgers.edu>
Subject: Re: Historical Fact
Date: 30 Jun 88 17:27:50 GMT
To: info-unix at SEM.BRL.MIL
I would suspect the later. There was a program to print
planet and sattelite orbits, but I believe it was called azel.
I left my V6 manual at home.
-Ron
-----------------------------
From: Ron Natalie <ron at topaz.rutgers.edu>
Subject: Re: Question: Driver access to memory mapped devices on 386 SysV.3
Date: 30 Jun 88 17:31:39 GMT
Keywords: drivers
To: info-unix at brl-sem.arpa
Then I suppose you shouldn't teach them vi either. I don't know vi,
I have never used it. I've been doing UNIX since 1977. I went straight
from /bin/ed to EMACS. If I can't find EMACS on a machine, I use ed.
It means that I am a true master of regular expressions. By the way,
EMACS is available on machines other than UNIX ones, which makes it
more advantageous than vi. Remember EMACS started on DEC-10's.
-Ron
-----------------------------
From: Mitchell Wyle <wyle at solaris.uucp>
Subject: Re: plot, graph (summary)
Date: 29 Jun 88 10:37:55 GMT
To: info-unix at SEM.BRL.MIL
Last week, I asked for plot(1) enhancements:
...
>and graph(1) always puts a key at the bottom left of how he scaled the
>graph. The home-grown packages here are too applications specific.
>I'm looking for a robust scatter-plot package.
>
>I don't have and therefore haven't looked at: gnuplot, graph+,
>crc_plot, vplot, quickplot, or lplot. If anyone (preferably in Europe)
>uses any of these PD packages, and knows it's what I want, I would
>appreciate his sending it my way.
I received the obligatory me-too's and am therefore posting a summary.
---Bill Jones <mcvax!calgary!vaxb!jonesb at cernvax.uucp> said:
You might want to check out "S", a language for data analysis available
from Bell Labs. It's an interactive system which allows you to define
datasets, perform analysis on them, then produce graphical output using
many different styles. My memory says (and you should verify this
before posting) that the price for source code is around $10000 for
commercial users, $100 for educational institutions.
---Marc Majka <mcvax!ubc-cs!majka at cernvax.uucp> said:
Hello. I have recently written a scatterplot program which runs on
top of a superset of the UNIX plot(5) format, known as Superplot. I
think that you may have Superplot there already. Contact Gladys Wong
<wong at ethz.uucp> or phone her: 256 52 33 and see if she has Superplot.
If you would like the scatterplot program, I can mail it to you. It
needs the Superplot layer to draw labels. If you want the whole package
(It comes with Gnuplot and other goodies), send me a tape, self-addressed
mailer, and CAN$10 for postage. I can put it on a small reel of tape
(600 foot) at 800, 1600, or 6250 BPI, or on a SUN cartridge tape.
I queried the csnet sources server and got crc_plot which satisfies the
need here. The Makefiles and .h files are weak, but the package
compiles with no .c code changes on Sun OS EXPORT. The language
interfaces look strong, and I have tested "qplot" with success. The
plot3d program looks exciting as well. Thank you Rich Kulawiec! For
those unfamiliar with Rich's crc plot package, here are excerpts from
the README:
---Rich Kulawiec, rsk at j.cc.purdue.edu writes:
The CRC plotting package is a device independent graphics system.
It includes subroutines for generating graphics which may be called
from Fortran or C, a two-dimensional plotting utility, and a
three-dimensional plotting utility.
The CRC package was originally developed at the Purdue University School
of Electrical Engineering by Carl Crawford; additional work has been
contributed by Mani Azimi and Malcolm Slaney, notably "plot3d". This
software has been in use locally for several years, and so most of the
obvious bugs have hopefully been caught and fixed. Although nobody's
willing to promise to fix future bugs immediately, it is not unreasonable
to assume that this package will continue to be supported, so please
do report bugs. (If you like, send them to me, and I'll forward them
to the folks at EE.) HOWEVER, no guarantees, folks.
This software has been developed on Vaxen running 4.XBSD; it's likely that
it will work on most machines running some variant of 4.XBSD. The two
user programs contained herein (qplot and plot3d) are probably of some use
to folks who need quick plots with reasonable resolution and labels
and annotation and so on without a lot of bother. Nice features of qplot
and plot3d include the ability to overlay multiple plots, tolerance of
a lot of different data formats, automatic or explicit scaling, logarithmic
plotting, ability to plot "bar graphs", and adjustable surface tilt (plot3d).
Thank you all for the replies, advice, etc. Long live usenet!
-Mitch (wyle%ifi.ethz.ch at relay.cs.net)
--
-Mitchell F. Wyle wyle at ethz.uucp
Institut fuer Informatik wyle%ifi.ethz.ch at relay.cs.net
ETH Zentrum
8092 Zuerich, Switzerland +41 1 256-5237
-----------------------------
From: Steve Schlesinger <steves at ncr-sd.sandiego.ncr.com>
Subject: OSF v. ATT/Sun :-)
Date: 30 Jun 88 19:32:19 GMT
To: info-unix at brl-sem.arpa
Going from ATT/Sun control of Unix to OSF is like going
from the Shah to the Ayatolluh.
(attributed to an anonymous Unisys employee).
-----------------------------
From: Greg Hackney <hack at bellboy.uucp>
Subject: Re: UUCP Over TCP/IP
Date: 30 Jun 88 17:57:39 GMT
Keywords: uucp, uucpd, bsd4.3
To: info-unix at SEM.BRL.MIL
In article <175 at chip.UUCP> mparker at chip.UUCP (M. D. Parker) writes:
>In bits of the BSD documentation, there is a mention of the TCP/IP UUCP server
>deamon (i.e. /etc/uucpd). My question is if you are the sending party, how
>do you tell the system to use the UUCP protocol? I find nothing in this
>in connection with the L.sys file, in fact, I have really found nothing at
>all.
>Can anybody enlighten me on /etc/uucpd and its operation, invokation, etc.?
On my Pyramid, it is set up like this in the L.sys file:
sitename Any TCP uucp sitename in:--in: nuucp password: hispassword
The site I send to is a Unisys system that doesn't have the /etc/uucpd
daemon, but does have telnet, so I use:
sitename Any TCP telnet sitename in:--in: nuucp password: hispassword
--
Greg
-----------------------------
From: "J. Steven Harrison" <jsh0 at isg300.uucp>
Subject: Rn on a V.3 machine
Date: 30 Jun 88 15:30:53 GMT
To: info-unix at brl-sem.arpa
I thought I would ask the net before diving into the code of rn about this
first.
Standard distribution tapes on AT&T V.3 machines with more than 1 drive
sets up the user file system on the second drive with root, swap, etc on
the first drive.
It seems that in the file util.c getwd only works for root on the root
partition (drive 1) without getting the error "getwd: read error in ..".
If run as root from the second drive as root you get the same error and
if run as any user on the system (that doesn't have uid=0) the same error
also results.
Has anyone already solved this problem?? We are running on AT&T 3B2 600's,
700's and a 4000.
Thanks in advance for any help.
Steven Harrison
VP Programming/Support
Information Systems Group Inc., San Diego, CA
(619) 234-3405
{nitro!isg100, ucsdhub}!jack!steve
-----------------------------
From: "Larry W. Virden" <lwv at n8emr.uucp>
Subject: NFS/YP alternative
Date: 30 Jun 88 17:28:08 GMT
To: info-unix at brl-sem.arpa
I have two questions. First, can anyone explain why Yellow Pages requires
the files served to be in a DBM format? Secondly, does anyone know of a similar
service that could be used for non-dbm files? For instance, if for
administrative
reasons I did NOT want to NFS mount the /usr/man and /usr/local/man pages but
instead distribute them in a manner similar to Yellow Pages, what could I do to
achieve this?
Thanks!
--
Larry W. Virden 75046,606 (CIS)
674 Falls Place, Reynoldsburg, OH 43068 (614) 864-8817
osu-cis!n8emr!lwv (UUCP) osu-cis!n8emr!lwv at TUT.CIS.OHIO-STATE.EDU (INTERNET)
We haven't inherited the world from our parents, but borrowed it from our
children.
-----------------------------
From: "The Resource, Poet of Quality" <bicker at hoqax.uucp>
Subject: MKS Toolkit/UNIX [r] System, SEA ARC/PKARC, ...
Date: 30 Jun 88 12:34:29 GMT
Keywords: AT&T, lawsuit, CSS, PC/Tools, PC/VI
To: info-unix at brl-sem.arpa
> In article <166 at skep2.ATT.COM>, wcs at skep2.ATT.COM (Bill.Stewart.<ho95c>)
writes:
> > b) MKS has ported some UNIX tools and reimplemented others and is presumably
> > following the rules for the products they use,
This raises a question in my mind in light of all the talk about
SEA suing PK Ware over the "look" of the archive tools. If SEA has
a case doesn't AT&T?
--
/kohn/brian.c AT&T Bell Laboratories Semantic Engineering Center
The Resource, Poet of Quality ...ihnp4!hoqam!bicker (201) 949-5850
"It is useless for sheep to pass resolutions in favor of vegetarianism
while wolves remain of a different opinion." - Wm. Ralph Inge, D.D.
-----------------------------
From: These terminals don't do umlauts!? <greywolf at unisoft.uucp>
Subject: Re: how can I "." in csh?
Date: 30 Jun 88 19:17:24 GMT
To: info-unix at brl-sem.arpa
In article <3513 at ncrcae.Columbia.NCR.COM> rogerc at ncrcae.Columbia.NCR.COM (Roger
Collins) writes:
: In csh, is there an easy way to duplicate the function of Bourne's "."
: command? I want to input commands from a file and have them
: change the current layer's environment.
How about:
alias . source
Of course, source doesn't look along a path, so we have to hack this a bit.
alias . 'source `ffo \!*`'
----
#! /bin/csh -f
# (well, why not csh if you're going to be using '.' like bourne shell, but
# using csh instead?)
# ffo: find first occurrence of a file along the path
foreach i ( $argv )
foreach i ( $path )
if ( -e $path/$argv ) then
echo $path/$argv
exit
endif
end
end
: --
: Roger Collins
: NCR - E&M Columbia
: rogerc at ncrcae.Columbia.NCR.COM
--
** All views expressed herein are my own and not those of UniSoft Corp. **
# Re: removing -c from diff and adopting NIH attitude: See Figure 1. #
# Roan (RO-an) Anderson, UniSoft Corporation, (415) 420-6400 x 165. #
# ...sun!unisoft!greywolf # /harley/davidson: Mount device busy.###########
-----------------------------
From: Eric Varsanyi <ewv at zippy.berkeley.edu>
Subject: Re: Question: Driver access to memory mapped devices on 386 SysV.3
Date: 1 Jul 88 02:29:06 GMT
Sender: usenet at agate.berkeley.edu
Keywords: drivers
To: info-unix at brl-sem.arpa
In article <1196 at award.UUCP> brian at apt.UUCP (Brian Litzinger) writes:
>I am running Microport Unix System V.3 on a discrete 386. I have
>successfully written drivers to control various I/O space mapped
>hardware devices. However, I cannot access the memory space of
>a memory mapped device that I have.
>
>My documentation doesn't cover sptalloc, spbyte & fpbyte which were used
>in the Hercules driver. An explanation of these might solve my problem.
There is actually 1 lone man page in the software development set that
sort of explains these. In writing a driver for a board that uses
exclusively memory mapped IO, I used sptalloc for a couple of things:
#include <sys/types.h>
#include <sys/immu.h>
#include <sys/param.h>
extern char *sptalloc();
[ during init routine... ]
/* Allocate and build the free list */
npages = (dxnbuf*sizeof(struct dqe))/4096+1; /* Pages to get */
DBGMISC(printf("dxinit: Allocating %d pages via sptalloc\n",npages));
if ((cdqe = (struct dqe *) sptalloc(npages,PG_P,NULL,NOSLEEP))==NULL) {
printf("dxinit: Buffer allocate failed, will not initialize\n");
return;
}
[ This first part grabs memory from the system memory map for IO buffer,
I didn't want to use buffers from the buffer pool since these buffers
sometimes have to hang out for a long time.
The PG_P gives the type of memory we want (always use PG_P), the NULL means
we want memory from the system memory map (not a particular address), and
the NOSLEEP means don't wait if its not available (its always available)
The driver goes on to build a free list with its freshly allocated memory,
then... ]
/* Map boards in and initialize them */
for (cnt=0; cnt<dxnbrd; cnt++) {
if ( (
dxbrd[cnt].base=(struct srcp *)sptalloc(2,PG_P,dxbase[cnt], 0)
) == NULL) {
printf("Error mapping physical/virtual memory\n");
} else { /* Initialize board */
dxbinit(&dxbrd[cnt]);
}
}
[ This loops through the all the boards (there can be more than 1 in a
system) and maps their physical (dxbase[cnt]) memory into kernel
virtual memory. The base is the physical PAGE number we want to map:
0xD0 would point to D000:0000 (or D0000 for you linear folks)
The 2 is the number of pages I want to map, and the NOSLEEP parameter
has no meaning here, so I set it to 0
When this loop finishes, I can get to the board memory via:
dxbrd[cnt].base->xyzzy (Where xyzzy is some register I want
to talk to)
]
>I particularly like the idea of obtaining a pointer to the memory mapped
>space of my device and using val=*(base+offset) and *(base+offset)=val
>for reading and writing. However, I have had only PANICs so far.
No problem, instead of using a structure, get the kv address to a character
pointer, then have fun with *xx+yy=c all you like.
-----------------------------
End of INFO-UNIX Digest
***********************
More information about the Comp.unix.questions
mailing list