Subject: Linux-Development Digest #967
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:     Sun, 31 Jul 94 00:13:04 EDT

Linux-Development Digest #967, Volume #1         Sun, 31 Jul 94 00:13:04 EDT

Contents:
  Re: Driver for Adaptec 2940 SCSI ? (Harald Milz)
  Windows swap/bdflush/1.1.3x (Paul Sitz)
  Re: lint on linux? (David Fox)
  Re: Truecolor support for XFree86? (David Fox)
  Re: Ooops! SCSI CD-Rom broken in 1.1.33 (Rob Janssen)
  Re: Park Routine for old RLL/MFM drives (David Lyle Robinson)
  Re: Linux 1.1.35 and ld -qmagic (Van Dao Mai)
  Re: Linux backup of MSDOS? (Dennis Heltzel)
  Re: Realtime? (Arthur C. Smith)
  Re: Fatal Signal 11 - reproduceable ! (David Flood)
  Re: NE2100: anyone got it working? (Donald Becker)
  NE2100: anyone got it working? (Matt Warnock)
  Re: Xfree86: increase pallate? (Bernd U Meyer)
  Re: APC UPS owners or potential buyers, trying to show user base (Michael K. Johnson)
  Re: APC UPS owners or potential buyers, trying to show user base (Michael K. Johnson)
  Re: APC UPS owners or potential buyers, trying to show user base (Terry Lambert)
  Re: IPX (Ryan A Whelan)

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

From: hm@seneca.ix.de (Harald Milz)
Subject: Re: Driver for Adaptec 2940 SCSI ?
Date: Tue, 26 Jul 1994 21:16:35 GMT
Reply-To: hm@seneca.ix.de

Hauke Jans (jans@plumbum) wrote:
            ^^^^^^^^^^^^
No one can answer by personal e-mail, so ...

: >     Hi. Is there anybody developing a driver for 
: > the Adaptec 2940 PCI SCI host adapter ? 

Yes and no, this is in the Projects-FAQ. BTW: Why would you want that?


-- 
Harald Milz                             office: hm@ix.de
iX Multiuser Multitasking Magazine      home:   hm@seneca.ix.de
Opinions are mine, not my employer's -- the answer is Forty-two
PS: I apologize for my bad return address - I hope this will be fixed soon.

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

From: psitz@empros.com (Paul Sitz)
Subject: Windows swap/bdflush/1.1.3x
Date: Sat, 30 Jul 1994 17:42:54 GMT

I am sharing a swap partition between linux and Windows. Something
in a recent upgrade has broken the restoration of the Windows
information on a shutdown.  Specifically, I am running Slackware
1.2, and kernel 1.1.36 (the problem was evident at earlier levels).
My /etc/rc.d/rc.0 is

########

  PATH=/sbin:/bin:/usr/sbin:/usr/bin
  /sbin/swapoff /dev/hda5
  if [ -f /local/win.swap ]; then
    echo Restoring Windows swap drive.....
    /bin/dd if=/local/win.swap of=/dev/hda5
    sync
    echo Windows swap drive restored.
  fi
  echo Unmounting file systems.....
  umount -a
  echo Done.

########

When  I shutdown, I get all the appropriate messages from dd 
indicating that the restoration has occurred.   But it doesn't
work (i.e. When the other user of this machine tries to bring
up Windows, it hangs with a blank screen.)  I have found two
workarounds.  
   
     1) Boot up an old (1.0.9) version of linux and
        shut it down.

     2) Before calling shutdown, execute
         /sbin/swapoff /dev/hda5
         /bin/dd if=/local/win.swap of=/dev/hda5
        manually.

This appears to me to be some kind of caching problem.  I don't
like either of my workarounds.  Any suggestions for a reliable
method to force the write to get flushed to disk in /etc/rc.d/rc.0?

--
Paul Sitz                                                          
Empros Power Systems Control
A Division of Siemens Energy & Automation, Inc.
2300 Berkshire Lane North          email:          psitz@empros.com
Plymouth, MN  55441-3694           screammail:     (612) 553-4516


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

From: fox@graphics.cs.nyu.edu (David Fox)
Subject: Re: lint on linux?
Date: 30 Jul 1994 15:08:26 GMT

] >     <sigh> This question comes up every two weeks. No there is no 
] >lint for Linux. 'gcc -Wall' should do everything you need though.
] 
] Or if you really want fun: gcc -Wall -pedantic

Turning on the optimizer causes additional flow analysis, and
thus even more messages are generated.
--
David Fox                                               xoF divaD
NYU Media Research Lab                     baL hcraeseR aideM UYN

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

From: fox@graphics.cs.nyu.edu (David Fox)
Subject: Re: Truecolor support for XFree86?
Date: 30 Jul 1994 15:14:20 GMT

In article <Ctr9n8.Fzn@aib.com> hohndel@aib.com (Dirk Hohndel) writes:

] as I said, you will have a very hard time with packed 24 bit modes, and
] with 32 bit modes 1280k videoram are needed for 640x480. On the other
] hand with 16bit color even 800x600 is possible in 1MB.

I for one will go out and buy whatever I need to run a 32 bit
server at a decent resolution, so if you are worrying that no
one would use it, you can set your mind at ease...
--
David Fox                                               xoF divaD
NYU Media Research Lab                     baL hcraeseR aideM UYN

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

From: rob@pe1chl.ampr.org (Rob Janssen)
Subject: Re: Ooops! SCSI CD-Rom broken in 1.1.33
Reply-To: pe1chl@rabo.nl
Date: Tue, 26 Jul 1994 21:25:30 GMT

In <1994Jul26.021022.964@encomix.exnet.com> rampa@encomix.exnet.com (Ramon Martinez) writes:


>:>Error:
>:>root:/home/cjs# mount /dev/cdrom /system_cd/ -t iso9660
>:>mount: block device /dev/cdrom is not permitted on its filesystem
>:>root:/home/cjs#


>  Not only on SCSI. i have the same here with a Double Speed Panasonic. 
>(SBPCD).

>  It works fine in 1.1.31 but fails in 1.1.32 to 1.1.35

> -Ramon-

How about first reading some of the other postings on this subject??

Rob
-- 
=========================================================================
| Rob Janssen                | AMPRnet:   rob@pe1chl.ampr.org           |
| e-mail: pe1chl@rabo.nl     | AX.25 BBS: PE1CHL@PI8UTR.#UTR.NLD.EU     |
=========================================================================

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

From: robinson@ichips.intel.com (David Lyle Robinson)
Crossposted-To: comp.lang.c
Subject: Re: Park Routine for old RLL/MFM drives
Date: 26 Jul 1994 19:19:07 GMT

volker@cs.sun.ac.za (Volker Otto) writes:
        Hi there
        all you need to do is check on max cylinder and seek the
        drive there. Under DOG you would use BIOS calls (int 13h)
        ie.     mov ah,08
                mov dl,80
                int 13
        this returns drive specs for 1st. hard disk
                mov ah,0c
                int 13
        this seeks disk (to last cylinder) since the drive specs
        should be correct (I hope)
        hope this helps ever so slightly
        regards
        volker



Does anyone know how to do this in C under Linux?

Thanks,
David Robinson 


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

From: mai@vietgate.apana.org.au (Van Dao Mai)
Subject: Re: Linux 1.1.35 and ld -qmagic
Date: Fri, 29 Jul 1994 23:01:34 GMT

Look for the later version of gcc. Loader should come with it. Install the
whole lot to make sure everything is compatible.

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

From: dheltzel@crl.com (Dennis Heltzel)
Subject: Re: Linux backup of MSDOS?
Date: 30 Jul 1994 15:24:44 -0700

Chris J. Lansley (ssrlansl@reading.ac.uk) wrote:
:     I am considering using "dd" to get a binary image of the disc, and then
:  compressing it, then putting that on tape.  But this may have problems?

You must have a *humongous* Linux partition! Why not skip the compression 
and go directly to tape, making a "disk image" backup? This will mandate 
that you restore to a partition of the same size, however, and that is 
why "image" backups are not popular under DOS (plus, it would probably 
break most copy protection schemes :))

Dennis


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

From: asmith@nwdkit2.nw.att.com (Arthur C. Smith)
Subject: Re: Realtime?
Date: Sat, 30 Jul 1994 18:58:18 GMT

In article <31bo29$1j8@smurf.noris.de>,
Matthias Urlichs <urlichs@smurf.noris.de> wrote:
>In comp.os.linux.development, article <JOHNSONM.94Jul28232913@calypso-2.oit.unc.edu>,
[ Munch ]
>Don't forget that one big contributor to high response times is the time it
>takes to page a program in from disk. Surprise, Linux has a process-specific
>flag to prevent swapping. It's currently used for /sbin/init, but should be
>usable for other processes as soon as somebody adds a system call to
>actually set the flag. ;-)

I wrote a small system call to try this out and it works. I tested it by
grabbing most of my memory, locking the process in memory, then starting up
X. It took forever to start, and when I looked at free the locked process
was still all in memory. This did slow down X a bit, though 8-). I wonder what
would happen if I grabbed the memory, forced some of it to be swapped out,
and then locked it? Hmmm.
  Art

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

From: dcflood@u.washington.edu (David Flood)
Subject: Re: Fatal Signal 11 - reproduceable !
Date: 30 Jul 1994 19:16:08 GMT

wosch@rbg.informatik.th-darmstadt.de (olav woelfelschneider) writes:

>So you can not be absolutely sure wether this is a compiler bug or a linux bug.

Or a hardware bug.  I had a bad swap partition that neither re-lowleveling
(MFM) or mkswap -c would find the bug but by spliting the partition in 1/2
eliminated the consistant signal 11's when compiling the kernel.  It also
eliminated some serious disk thrashing.

-- 
=============================================================================
dcflood@u.washington.edu

The above opinions are mine alone and do not reflect anyone elses.
Besides, who wants my opinion anyway?
=============================================================================

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

From: becker@cesdis.gsfc.nasa.gov (Donald Becker)
Subject: Re: NE2100: anyone got it working?
Date: 30 Jul 1994 19:11:11 -0400

In article <31cpcj$1ij@garlic.com>, Matt Warnock <mwarnock@garlic.com> wrote:
>I have 2 NE2100 cards set up as a small net.  Trying to get Linux running
>on them.  They are jumpered for IRQ 9, DMA 3, base 360.  These are real
>Novell/Eagle, not clones.  
>
>Lance.c 0.14g detects the right base address, but auto-irq fails.

This seems to happen a lot with real NE2100 cards.  I'll have a new version
of lance.c in a few days with some extra code that might work better with
the 7990 chip.  (Sorry, I don't have a real NE2100 to test with.)

>Setting LANCE_DMA to 3 doesn't help the auto-IRQ.  So I added 3 lines

No, it wouldn't.  In fact the current auto-IRQ code expects to get a
DMA_FAILURE interrupt, since the DMA controller isn't yet initialized.

>to define LANCE_IRQ, and I use the printk line "unable to get IRQ
>line" (or something like that) to set dev->irq to LANCE_IRQ (9), and I
>commented out the return just below, so that it does go on to
>initialize the card with the compiled-in defaults, even though
>auto-irq failed.  Then the card seemed to init okay, I can ping back
>and forth-- for a while.

OK, great.  Once an old LANCE is initialized it should work just fine.

>After a few minutes, no pattern that I can yet discern, I get "Network
>unreachable".  ifconfig shows nothing changed.  I have not tried other

I'll be this is unrelated.  I'll bet that you are running routed, and it's
removing your routes.  Stop running routed.

>I don't know enough about Linux networking to identify what layer is
>shutting down, I only suspect the card/driver because of the auto-IRQ
>failure and hack set forth above.   It happens on both 1.0.8 and 1.1.33.

Something seem to have changed in the recent 1.1.* kernel to break the
auto-IRQ for the lance.  I'm trying to track the problem down.

-- 
Donald Becker                                     becker@cesdis1.gsfc.nasa.gov
USRA 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: mwarnock@garlic.com (Matt Warnock)
Subject: NE2100: anyone got it working?
Date: 29 Jul 1994 22:44:51 -0700

I have 2 NE2100 cards set up as a small net.  Trying to get Linux running
on them.  They are jumpered for IRQ 9, DMA 3, base 360.  These are real
Novell/Eagle, not clones.  

Lance.c 0.14g detects the right base address, but auto-irq fails.
Setting LANCE_DMA to 3 doesn't help the auto-IRQ.  So I added 3 lines
to define LANCE_IRQ, and I use the printk line "unable to get IRQ
line" (or something like that) to set dev->irq to LANCE_IRQ (9), and I
commented out the return just below, so that it does go on to
initialize the card with the compiled-in defaults, even though
auto-irq failed.  Then the card seemed to init okay, I can ping back
and forth-- for a while.

After a few minutes, no pattern that I can yet discern, I get "Network
unreachable".  ifconfig shows nothing changed.  I have not tried other
DMA/IRQ settings, but the card works fine under Netware Lite 1.1 and
Personal Netware.  This IRQ is the ONLY one I can use without conflict
(unless I can stack 2 16550As on IRQ 4, I dunno).

I don't know enough about Linux networking to identify what layer is
shutting down, I only suspect the card/driver because of the auto-IRQ
failure and hack set forth above.   It happens on both 1.0.8 and 1.1.33.
Any ideas, anyone?

Has anyone gotten a true NE2100 up?  if so mail me your experience, I'll 
summarize to the net.

-- 
W. Matthew Warnock, Attorney (mwarnock@garlic.com)          Tel:408.778.7273 
60 West Main Avenue, Suite 12A, Morgan Hill CA  95037-4553  Fax:408.778.7989

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

From: berndm@cs.monash.edu.au (Bernd U Meyer)
Subject: Re: Xfree86: increase pallate?
Date: Fri, 29 Jul 1994 15:06:58 GMT

ericy@cais.cais.com (Eric Youngdale) writes:

>Phil Johnson (johnsonp@sscnet.ucla.edu) wrote:
>: In article <30ne0u$6gm@nic.umass.edu>,
>: >I heard that millions of colors would arrive with XFree 3.1
>:
>: Perhaps you are joking, but someone out there will believe you.  
>: XFree86[tm] 3.1 will not increase the colors available.

>       I heard that you will only get the additional colors if you are 
>using FreeLSD or NetLSD :-).

After reading this thread, I thought I would share with you one of the
startup messages this Silicon Graphics just gave me:

  starting netlsd.....

And in case you wonder, it has 24 bit coulour :-) 

Bernie

P.S.: I'll have to reset one of those 8 bit Silicons some time, wonder
if they also run netlsd :-))
--
"And the band played 'Waltzing Mathilda' /  as we stopped to bury our slain;
And we buried ours / and the Turks buried theirs  | ..... living in Oz ....
And it started all over again"                    | 
(The Pogues, "Waltzing Mathilda", orig by Eric Bogle, "And the band played WM")

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

From: johnsonm@calypso-2.oit.unc.edu (Michael K. Johnson)
Crossposted-To: comp.unix.bsd
Subject: Re: APC UPS owners or potential buyers, trying to show user base
Date: 31 Jul 1994 01:10:21 GMT


In article <CtpoIA.JJH@newsserver.aggregate.com> rhealey@sirius.aggregate.com (Rob Healey) writes:

   In article <JOHNSONM.94Jul28233853@calypso-2.oit.unc.edu>,
   Michael K. Johnson <johnsonm@calypso-2.oit.unc.edu> wrote:
   >   Really now! Those boxes use an RS-232 interface, right? What do they tell 
   >   the computer? If they only communicate one thing (power fail) then it
   >   is probably something as simple as shorting the RD and SD lines. Get a
   >   technician to check it out for you while you pull the plug.

I did not write that.  Please keep your attributions straight.  I
wrote a response which contained that text, but that was written by
someone else.

           The Smart-UPS's use some sort of dongle RS232 cable if I remember
           correctly, i.e. it's not a normal RS232 cable since it has extra
           components in it. Once you have one of those cables its some sort
           of 1 or 2 letter commands sort of thing to get info out of the
           UPS.

           It's the cables that hurt since they cost more than a simple 232
           cable.

Hm.  I thought that they told me it was normal serial...  If this is
true, I'll have to manufacture a cable myself.  Won't be the first time...

           The UNIXware software is SCO version too, I don't think they have
           a specific version for any other OS.

They appear to have at least two different versions, since I've seen
them for two different prices...  I didn't buy it; I don't want
software to which I don't have the source controlling when I shut my
machine down, even if I *can* run it.

Thanks for the info.  I ordered my Smart-UPS 900 yesterday, and it will
arrive next week, baring disaster.  Then I get to work on it...

michaelkjohnson

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

From: johnsonm@calypso-2.oit.unc.edu (Michael K. Johnson)
Crossposted-To: comp.unix.bsd
Subject: Re: APC UPS owners or potential buyers, trying to show user base
Date: 31 Jul 1994 01:13:37 GMT


In article <1994Jul30.102905.23539@news.dkrz.de> hll@lidarh.dkrz.de (Holger Linne) writes:

   Strange -- I own a Smart UPS 600 and it does not need 'some sort of dongle
   RS232 cable'. Okay, what you need is a 'special' cable to connect 'pin x' at the
   UPS-port to some 'pin y' at the computers port. But the pins are very well
   described in the Owner's manual that comes with the UPS (or did they change
   their policy ?-- i bought my UPS  in 1993).

Are you sure that this is a Smart UPS 600 and not a Back UPS 600?
Your description sounds like a Back UPS 600, from what I have read.

   According to this manual, there is a pin to turn off the UPS, one to signal
   a transfer to battery backup, one to signal if less than 2 minutes of 
   backup time left (can be changed to 5 min) and one to turn on the UPS.

This sounds very much like the Back-UPS product.

I'll have more to say later this week when my unit shows up, assuming
I have time to play with it and disaster doesn't strike.

michaelkjohnson

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

From: terry@cs.weber.edu (Terry Lambert)
Crossposted-To: comp.unix.bsd
Subject: Re: APC UPS owners or potential buyers, trying to show user base
Date: 31 Jul 1994 02:17:10 GMT

In article <JOHNSONM.94Jul30211022@calypso-2.oit.unc.edu> johnsonm@calypso-2.oit.unc.edu (Michael K. Johnson) writes:
] 
] In article <CtpoIA.JJH@newsserver.aggregate.com> rhealey@sirius.aggregate.com (Rob Healey) writes:

[ ... ]

] I did not write that.  Please keep your attributions straight.  I
] wrote a response which contained that text, but that was written by
] someone else.
] 
]          The Smart-UPS's use some sort of dongle RS232 cable if I remember
]          correctly, i.e. it's not a normal RS232 cable since it has extra
]          components in it. Once you have one of those cables its some sort
]          of 1 or 2 letter commands sort of thing to get info out of the
]          UPS.

It is difficult to keep this particular "indentation is quoting" style
of reply straight, since there isn't a physical marker and you don't
also indent the article attribution ("In article" ...).  Generally,
indentation is used for paragraph starts or for emphasis.

There is also a fundamental assumption in doing this that the news reader
that a potential responder is using is not using variable width fonts
anyway; the columnar form using spaces doesn't guarantee a clean left
margin to allow the viewer to visually line it up... for instance, many
X based news readers don't have fixed-width (terminal) fonts.

Do you use this method to get around the "more new than quoted material"
restriction some news reader/posters have?

If so, a better method would be to turn the default quote character (on
my system, ']') to another printable character (like '[' -- notice that
sometimes my "quoted material" markers are reversed?) instead of a series
of spaces.


                                        Regards,
                                        Terry Lambert
                                        terry@cs.weber.edu
---
Any opinions in this posting are my own and not those of my present
or previous employers.

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

From: rwhelan@site.gmu.edu (Ryan A Whelan)
Subject: Re: IPX
Date: 31 Jul 1994 02:31:17 GMT

David G. Boney (dboney@cs.ttu.edu) wrote:
: Hi, 
:       Where can I find info on how to use the ipx stuff in the kernel?

Is there anything documented on the Linux IPX API?  Also, is there a
way in IPX to so a sorta 'PING' on a workstation with the LAN drivers loaded
(Its attached to a File Server but I want to and think I have to stick to 
the IPX layer)
Can a workstation reply with a IPX equivalent (if there is one) of IAH, and 
AYT.

thanks
ryan


--
Ryan A. Whelan            Unix and the World Unixes with you. Vax and you
                                                               Vax ALONE!
ryan@absolut.labs.gmu.edu  rwhelan@gmu.edu                      
Gopher: absolut.labs.gmu.edu:70         URL: http://absolut.labs.gmu.edu:80/

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


** 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
******************************
