Subject: Linux-Development Digest #18
From: Digestifier <Linux-Development-Request@senator-bedfellow.MIT.EDU>
To: Linux-Development@senator-bedfellow.MIT.EDU
Reply-To: Linux-Development@senator-bedfellow.MIT.EDU
Date:     Fri, 12 Aug 94 19:13:11 EDT

Linux-Development Digest #18, Volume #2          Fri, 12 Aug 94 19:13:11 EDT

Contents:
  Re: Lilo Support for NRC-SCSI (Michael Will)
  Re: Interesting idea for lilo developers (Darin Johnson)
  Re: Typo in kernel 1.1.43 (Gaybeul)
  Re: compiling new version of Mosaic (turnbull@turnbull.sk.tsukuba.ac.jp)
  Re: -= good programmer's editor for X? (Pete Fenelon)
  Re: Anyone want T1 access from Linux? (root)
  Re: (Some) NFS-mounted Executables Won't Run on 1.1.41 (Piercarlo Grandi)
  Re: gcc and Linux (compiling)... (Lars Wirzenius)
  Re: PowerGlove and linux? (Stuart N. John)
  Re: XFree86 and backing store? (James P. Trainor)
  Floppy: No mdir on DD-Disks (1.1.41/42) (Dirk Hillbrecht)
  BOCA ioAT66 serial card?? (Steven Charlton)
  sbpcd ejects on umount in 1.1.44 (Ross Boswell)
  Sony CDU31A/Floppy driver fix (Corey Minyard)
  Re: Suggestion: Lets have a standard for numeric uid/gids (Donald Becker)
  Re: More IO ports :) (Donald Becker)
  Recent kernel versions (S. Joel Katz)

----------------------------------------------------------------------------

Crossposted-To: comp.os.linux.admin
From: michaelw@desaster.student.uni-tuebingen.de (Michael Will)
Subject: Re: Lilo Support for NRC-SCSI
Date: Tue, 9 Aug 1994 12:45:20 GMT

martin@wilbur.zdv.Uni-Mainz.DE (Christoph Martin) writes:
>I use a 540MB Quantum disk and tried disktab entries like

>> 0x800        0x80    32      64      ..

>I'm just guessing these values, but there is no difference in the
>error message.

The /etc/disktab is the key to your problem. 

I had to create one with an entry for every of my partitions. it looks like 
this: 
0x800   0x80    32      64      1006    0
0x801   0x80    32      64      1006    
0x802   0x80    32      64      1006    
0x803   0x80    32      64      1006

You can find out the correct parameters for 32/64/1006 when running the
program dparam.com on DOS. It comes with LILO and is included in the
SCSI-HOWTO. But I wonder why lilo does not work correctly with the
bios-diskgeometry-syscall - fdisk does. 

Cheers, Michael Will
-- 
. .         Michael Will <michaelw@desaster.student.uni-tuebingen.de> 
 .      cs-student in Tuebingen, Germany, Europe, Earth, Solar-System, [...]

        <Lestat> set novice off

------------------------------

From: djohnson@arnold.ucsd.edu (Darin Johnson)
Subject: Re: Interesting idea for lilo developers
Date: 12 Aug 1994 05:57:21 GMT

>  - Not everyone runs DOS.  (My home machine has been DOS-free for over a
>    year and I plan to keep it that way.)

Except you can easily get older versions of DOS from people who've
upgraded with little problem.

Then you can use it to boot to linux, and play all that neato
games.  Then tell your friend who you got the old version of
DOS from that it makes a great game machine...

>  - LILO is a better boot manager for Linux.  (Multiple kernels, better
>    error reporting, better documentation, etc.)

Loadlin can do multiple kernela and has decent error reporting and
docs as well.  Plus it doesn't have things saying "warning, if you
don't know what you're doing you can screw up your computer."

But who really cares - if you don't like it, don't use it.  Other
people like it and use it.  (some people have to use it)
--
Darin Johnson
djohnson@ucsd.edu
    Ensign, activate the Wesley Crusher!

------------------------------

From: liphy02@frbdx12.cribx1.u-bordeaux.fr (Gaybeul)
Subject: Re: Typo in kernel 1.1.43
Date: 12 Aug 1994 14:16:40 GMT

Linus Torvalds (torvalds@cc.Helsinki.FI) wrote:
: In article <1994Aug11.123718.375@spectre.apana.org.au>,
: Richard Lindner <rjl@spectre.apana.org.au> wrote:
: >the 1.1.43 patches include a typo - in tpqic02.c, SA_INTERRUP should be
: >SA_INTERRUPT.

: There is also a "&" in drivers/sound/soundcard.c before the second
: argument to the register_irq() call.  You'd better remove it if you
: include the sound-drivers, or you'll get a warning at compile-time and a
: system crash when trying to use the sound driver at run-time.  Oops. 
: Silly me for not even trying to compile it,

should be request_irq() instead of register_irq() no?

Lokh.

------------------------------

From: turnbull@turnbull.sk.tsukuba.ac.jp
Subject: Re: compiling new version of Mosaic
Date: 12 Aug 94 04:23:04 GMT
Reply-To: turnbull@shako.sk.tsukuba.ac.jp

In article <1994Aug4.025553.21603@freenet.victoria.bc.ca>,
Jason Fiset <uc452@freenet.Victoria.BC.CA> wrote:
>
>I would like to compile Mosaic (I already know there is a binary)
>but it doesn't seem to be working.  It's complaining about not
>being able to find /Xm/<filename>.h

That's Motif.

>Does anyone know how to configure the Makefile so that I don't
>have to have Mosaic in order to compile it?  Better yet, could

You mean Motif, not Mosaic.  Answer is, no.  Buy Motif, I understand
SWiM is only about $150.

>someone post the changes they made to the Makefile for it to
>work?  Thanks

Change the Makefile.  Raaiiiiight.  Nope, you'd have to rewrite the
whole thing to use a different set of widgets.  Motif may look like a
cosmetic difference, but it's not.  It involves changes all the back
into the X Toolkit.  Buy it, or use the binaries contributed by
someone who shelled out the $$$ for Motif.
    You think you've got problems?  I'm looking for a copy that does
*Japanese*.  Believe me, *that* is more than just cosmetic.

>Jason Fiset
>-- 
>Sea to Sky Freenet Member

--
Steve Turnbull
Sysop *says* I'm in the DNS, DNS can't find me.  For now, replies to:
turnbull@shako.sk.tsukuba.ac.jp 

------------------------------

From: pete@minster.york.ac.uk (Pete Fenelon)
Subject: Re: -= good programmer's editor for X?
Date: 12 Aug 1994 15:19:07 BST

In article <31qf4v$hoh@hippo.shef.ac.uk> Stuart Herbert wrote:
; : : Are there any good programmer's editors for X - not term-like ports?

; : : kristof-
; : : nybakken@world.std.com

Well, I've ported the X version of Rob Pike's ``sam'' to Linux, but at
present it's a little too flaky to distribute the port... that's the
best editor _full stop_.

However, if you want a slightly more conventional X-based editor, the Linux
binaries for Crisp are available on some of the FTP sites, and it's quite a
tolerable piece of software. Not _quite_ to my taste (give me sam, or vi,
and I'm a happy man!) but infinitely better than emacs.

On the other hand, if you want a damn fine terminal-based editor, the
Thompson-Davis Editor (TDE) is very impressive -- good regexps, nice
keyboard handling, reasonable level of features... but it's far nicer in
console mode on Linux than in an xterm.

pete
--
Peter Fenelon - Research Associate - High Integrity Systems Engineering Group,
Dept. of Computer Science, University of York, York, YO1 5DD +44 (0)904 433388
pete@minster.york.ac.uk WWWpage http://dcpu1.cs.york.ac.uk:6666/pete/pete.html
``Having the time of my life... or something quite like it'' - Elvis Costello.

------------------------------

From: root@coetech7-pc.uncc.edu (root)
Subject: Re: Anyone want T1 access from Linux?
Date: 11 Aug 1994 06:42:59 GMT

>
>We would be interested in an ISDN connection.
>-- 
>Irving Greisman
>

Bill,

I enthusiastically second this suggestion!  From what I've seen from an
administrative standpoint (I'm new to this game...) it's a relatively
short step from providing a T1 product to providing an ISDN or B-ISDN
product. 

Here at the UNCC College of Engineering, we have been monitoring the
ISDN market as it develops as an alternative to dialup modems.  The
key link that is missing at this point is a good ISDN product for
the remote user.  I'm sure I could encourage more than a few purchases
here if you were to produce such a product.

Sorry to get so far off the topic, but Linux running ISDN is a goal of
mine.  Email me if you'd like input on what I think the Linux community
would like in terms of ISDN CPE.

Trevor

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Trevor A. Fiatal------------------------Computer Engineering, UNC-Charlotte
tafiatal@uncc.edu-------------------Project Mosaic PC Network Administrator
PGP 2.6 Public Key available by: Finger (local) Email request (Off-campus)

------------------------------

From: pcg@aber.ac.uk (Piercarlo Grandi)
Subject: Re: (Some) NFS-mounted Executables Won't Run on 1.1.41
Reply-To: pcg@aber.ac.uk (Piercarlo Grandi)
Date: Fri, 12 Aug 1994 15:29:50 GMT

>>> On Tue, 9 Aug 94 22:34:27 CDT, brianc@saintjoe.edu (Brian Capouch) said:

Brian> I think there are some VFS problems in the more recent kernels.
Brian> I only had around 1.1.35, 1.1.40, and 1.1.41--the problem occurs
Brian> in the 1.1.4x but not the older kernel.

Brian> Some, but not all, executables that I have NFS-mounted on another
Brian> Linux box give a segmentation violation when run under the newer
Brian> kernels but are just fine under the older ones.

Brian> Running ldd on the executable enables prediction of failure: if
Brian> it returns silently then the app isn't going to execute, if it
Brian> shows anything at all it will run just fine.

I have noticed this happening; it also happens with CDROM mounted
executables. A VFS problem indeed?

------------------------------

From: wirzeniu@cc.Helsinki.FI (Lars Wirzenius)
Subject: Re: gcc and Linux (compiling)...
Date: 12 Aug 1994 18:50:16 +0300

> >>   void main(int argc, char *argv[])
>
> >The strictly correct version is:
> >int main(int argc, char *argv[])
>
> Actually, the strictly correct version is:
> int main(int argc, char *argv[], char *envp[])

Actually, the strictly correct version depends on the standard you wish
to follow.  The C standard specifies two strictly correct versions
(ISO 9899-1990, 5.1.2.2.1):

        int main(void)
        int main(int, char **)

(and compatible versions, e.g., leaving off the void, or using
``char *argv[]'' for the second argument).

The version you mentioned is a common extension.  Returning void is also
a common extension (although usually just because it happens to work, not
by design).  Neither is strictly conforming to the C standard.  I don't
know what POSIX says.

(Note that it is preferable to use getenv if one knows the name of the
environment variable.)

-- 
Lars.Wirzenius@helsinki.fi  (finger wirzeniu@klaava.helsinki.fi)
ftp.cs.helsinki.fi:pub/Software/Local/Publib -- general C function library

------------------------------

From: stuart@zen (Stuart N. John)
Crossposted-To: alt.cyberpunk.tech
Subject: Re: PowerGlove and linux?
Date: 12 Aug 1994 09:16:49 GMT

Daniel Garcia (kender@esu.edu) wrote:

: I'm pondering the idea of porting the powerglove control code to linux,
: but, I wanted to check and see if this had already been done somewhere,
: and if so - if someone could point me to it.  If not, let me know if
: any of you out there would be interested in it, and i'll look more into
: porting the driver.

: D

: ps. I started a little earlier tonight - but the timings under linux
:     seem to be a little bit different than under DOS ;)  Anybody out
:     there know the linux equivalent of biostime?  I don't have a dos
:     compiler - or i'd look it up myself.



i don't think anyone has bothered to do this...

It's a hell of a lot easier to build/buy a Menelli box and
access it through the serial driver...

the box can take all the work out of accessing the glove,
and you just work with packets of 7 bytes...

worked for me...

Stu

--

Stuart N. John              stuart@jba.co.uk  (Email) 
JBA Software Products Int   +44 (0)527 491259 (Voice)     //) /\
Studley, Warwickshire, UK   +44 (0)527 857146 (Fax)    (_//_)/  \

------------------------------

From: trainor@Avalon.M3iSystems.QC.CA (James P. Trainor)
Subject: Re: XFree86 and backing store?
Date: 11 Aug 1994 00:24:23 GMT


Here's what our local Linux box reports:



name of display:    torrezmo:0.0
version number:    11.0
vendor string:    XFree86
vendor release number:    2110
maximum request size:  262140 bytes
motion buffer size:  0
bitmap unit, bit order, padding:    32, LSBFirst, 32
image byte order:    LSBFirst
number of supported pixmap formats:    2
supported pixmap formats:
    depth 1, bits_per_pixel 1, scanline_pad 32
    depth 8, bits_per_pixel 8, scanline_pad 32
keycode range:    minimum 8, maximum 134
focus:  window 0x100000d, revert to Parent
number of extensions:    6
    XTestExtension1
    SHAPE
    MIT-SHM
    Multi-Buffering
    XTEST
    MIT-SUNDRY-NONSTANDARD
default screen number:    0
number of screens:    1
 
screen #0:
  dimensions:    1152x900 pixels (390x304 millimeters)
  resolution:    75x75 dots per inch
  depths (2):    1, 8
  root window id:    0x29
  depth of root window:    8 planes
  number of colormaps:    minimum 1, maximum 1
  default colormap:    0x27
  default number of colormap cells:    256
  preallocated pixels:    black 1, white 0
  options:    backing-store YES, save-unders YES  <<<<<<<<<<<<<<<<<<<<
  current input event mask:    0x50003d
    KeyPressMask             ButtonPressMask          ButtonReleaseMask        
    EnterWindowMask          LeaveWindowMask          SubstructureRedirectMask 
    PropertyChangeMask       
  number of visuals:    6


--

======================================================================
James P. Trainor    trainor@m3isystems.qc.ca      (514) 928-4600 x2427


------------------------------

From: hillbrec@informatik.uni-hannover.de (Dirk Hillbrecht)
Subject: Floppy: No mdir on DD-Disks (1.1.41/42)
Reply-To: hillbrec@informatik.uni-hannover.de
Date: Fri, 12 Aug 1994 16:59:06 GMT

Hi everyone,

I have never checked this out with kernels earlier than 1.1.41, but
beginning with this kernel and ongoing up to 1.1.42 (I've not installed
later versions) I have the following problem:

boot - insert 3,5"-HD-disc - "mdir" - very FAST the dir is shown -
now insert 3,5"-DD-disc in the very same drive - "mdir" - error:
"failed to initialize A:"

Dirk

--

Dirk Hillbrecht
snail mail: Bangemannweg 8 A, 30459 Hannover, Germany
Internet:   hillbrec@informatik.uni-hannover.de
FidoNet:    coming soon           === End of File ===

------------------------------

From: steven@gsb019.cs.ualberta.ca (Steven Charlton)
Subject: BOCA ioAT66 serial card??
Date: 12 Aug 1994 14:17:37 GMT

Has anyone hacked the kernel to properly support this serial card?  I
have one, and must get it working in fairly short order.  This is a 6
port, 3 UART (16550) serial card.  I noticed support for some BOCA
serial card, but the IRQs and base addresses seemed wrong for this
one.  Any and all help would be greatly appreciated.  E-mail response
would be preferable.

Steve

steven@cs.ualberta.ca   or   steve@mainland.ab.ca

------------------------------

From: drb@chem.canterbury.ac.nz (Ross Boswell)
Subject: sbpcd ejects on umount in 1.1.44
Date: 12 Aug 1994 20:44:40 GMT

I don't want sbpcd to eject the tray when the device is unmounted --
it's likely to crunch my fingers!
At 1.1.41 I edited drivers/block/sbpcd.c to define EJECT (and JUKEBOX)
to zero, and this vicious behaviour was cured.
At 1.1.44 it has returned, although the defines remain zero and there
is no new call to yy_SpinDown introduced into sbpcd.c.  

Any suggestions?

--
| Ross Boswell                    | Email : drb@chmeds.ac.nz          |
| Department of Pathology         | FAX   : +64 3 364 0525            |
| Christchurch School of Medicine | Phone : +64 3 364 0590            |
| NEW ZEALAND                     | Post  : PO Box 4345, Christchurch |

------------------------------

From: minyard@crchh7b9 (Corey Minyard)
Subject: Sony CDU31A/Floppy driver fix
Date: 12 Aug 1994 14:49:52 GMT

I have found the problem with the CDU31A.  Actually the floppy driver
was clobbering my base I/O address value.  I think someone owes me a beer
for this one.  At least a virtual beer :-).

This also might fix a lot of problems people were having with the new
floppy driver.

Here's the fix:


--- linux/drivers/block/floppy.c.old    Fri Aug 12 09:31:58 1994
+++ linux/drivers/block/floppy.c        Fri Aug 12 09:31:36 1994
@@ -2920,7 +2920,9 @@
        config_types();
 
        fdc_state[0].address = 0x3f0;
+#if N_FDC == 2
        fdc_state[1].address = 0x370;
+#endif
        for(fdc = 0 ; fdc < N_FDC; fdc++){
                FDCS->dtr = -1;
                FDCS->dor = 0;

------------------------------

From: becker@cesdis.gsfc.nasa.gov (Donald Becker)
Subject: Re: Suggestion: Lets have a standard for numeric uid/gids
Date: 10 Aug 1994 14:44:31 -0400

In article <root.776460959@olivia.ping.dk>,
Torben Fjerdingstad <tfj@olivia.ping.dk> wrote:
>Today I compiled at-2.7a on my fastest machine.
>Then I nfs'et to it from another machine and did an 'make install'.
>
>It did not work because the numeric user- and group id's are compiled
>into the programs, and they do not agree on the two machines.
>(at: Operation not permitted, cannot cd to /var/spool/atspool, or
>something like that).

You are in luck -- unlike most other NFS servers, the Linux NFS server has
the ability to maintain a non-one-to-one mapping of user and group ids.
Look at the 'ugidd' manual for details.

This is one of the advantages of the user-level NFS server that Linux uses.
For historical reasons Sun-derived NFS servers (most major NFS servers are
licensed from Sun) are implemented as in-kernel servers where this mapping
would be difficult to maintain.

>I wish there were standards for root,bin,sys,wheel,mail,daemon,
>sync,news,uucp,adm, (or admin?)... and so on.

I wish some Posix group had published a *suggested* list that we could all
migrate towards.  Just converting Linux boxes is easy -- converting the rest
of the world will be hard.  (And any standard coming from the Linux
community would be especially resisted -- there are still "old-timers" that
naively think of Linux as just another toy PC program.)


-- 
Donald Becker                                     becker@cesdis.gsfc.nasa.gov
USRA-CESDIS, Center of Excellence in Space Data and Information Sciences.
Code 930.5, Goddard Space Flight Center,  Greenbelt, MD.  20771
301-286-0882         http://cesdis.gsfc.nasa.gov/pub/people/becker/whoiam.html

------------------------------

From: becker@cesdis.gsfc.nasa.gov (Donald Becker)
Subject: Re: More IO ports :)
Date: 10 Aug 1994 14:55:42 -0400

In article <1f.6179.1566.0NC8F237@compart.fi>,
Riku Saikkonen <riku.saikkonen@compart.fi> wrote:
>But now it complains of a segmentation fault. It seems that the ioperm()
>routine returns -1 (marking an error, as per the man page) but does not
>set errno (perror reports 'Unknown error') as it should.

I suspect your code.
The only time ioperm() should fail is with invalid arguments, EINVAL, or
when you aren't root, EPERM.  I use the following code in my ethercard
diagnostics, and it works fine:

#include <unistd.h>
#define EC_IO_EXTENT 16
int port_base = 0x300;
...
    if (ioperm(port_base, EC_IO_EXTENT, 1)
        || ioperm(0x80, 1, 1)) {                /* Needed for SLOW_DOWN_IO. */
        perror("io-perm");
        fprintf(stderr, " You must be 'root' to run hardware diagnotics.\n");
        return 1;
    }

This functions correctly.

>Then inb()/outb() fails with a seg fault, probably because we're not
>supposed to access the ports without calling ioperm() first.

Yup.

>E-mail or post, I will post a summary with some sample code if/when I
>get this to work...

Check my ethercard diagnostic programs.  Some of them also demonstrate how
to mmap() in the shared memory regions used by some cards.

-- 
Donald Becker                                     becker@cesdis.gsfc.nasa.gov
USRA-CESDIS, Center of Excellence in Space Data and Information Sciences.
Code 930.5, Goddard Space Flight Center,  Greenbelt, MD.  20771
301-286-0882         http://cesdis.gsfc.nasa.gov/pub/people/becker/whoiam.html

------------------------------

From: stimpson@panix.com (S. Joel Katz)
Subject: Recent kernel versions
Date: 12 Aug 1994 12:44:07 -0400

        I've been looking at the latest patches (40-44). Who ran a spell 
check on the kernel?

        SJK


------------------------------


** FOR YOUR REFERENCE **

The service address, to which questions about the list itself and requests
to be added to or deleted from it should be directed, is:

    Internet: Linux-Development-Request@NEWS-DIGESTS.MIT.EDU

You can send mail to the entire list (and comp.os.linux.development) via:

    Internet: Linux-Development@NEWS-DIGESTS.MIT.EDU

Linux may be obtained via one of these FTP sites:
    nic.funet.fi				pub/OS/Linux
    tsx-11.mit.edu				pub/linux
    sunsite.unc.edu				pub/Linux

End of Linux-Development Digest
******************************
