The World of Computer Software/The World of Computer Software.iso/lnxd3-73.zip/DIGEST3.736

Internet Message Format | 1992-12-13 | 23KB

From [e-mail] Tue Dec  8 10:49:43 1992
Return-Path: <[e-mail]>
Received: from bloom-picayune (BLOOM-PICAYUNE.MIT.EDU) by montefiore.ulg.ac.be (4.1/SMI-4.1)
    id AA21344; Tue, 8 Dec 92 10:44:33 +0100
Received:  by bloom-picayune (5.57/25-eef)
    id AA26204; Tue, 8 Dec 92 04:15:45 -0500
Message-Id: <9212080915.AA26204@bloom-picayune>
From: Digestifier <[e-mail]>
To: [e-mail]
Reply-To: [e-mail]
Date:     Tue, 8 Dec 92 04:15:35 EST
Subject:  Linux-Activists Digest #736
Status: RO

Linux-Activists Digest #736, Volume #3            Tue, 8 Dec 92 04:15:35 EST

Contents:
  Will Linux/386BSD/XFree86 run on my Gateway 2000 4DX2-66V? (Mark A. Friedman)
  Re: .98.5 and .98.6: Infinite loop trying to init SMC Elite16 (Donald J. Becker)
  hardware hacks, was Re: MFM and IDE drives coexist under Linux (Cameron Spitzer)
  Troubles with X386-S3.2b: random lockups (Chris Newbold)
  Re: MFM and IDE drives coexist under Linux? (Donald E. Cope)
  Disk Defragger??? (Da Hubster!!!)
  TinyMUSE (Jon Anhold N8USK)
  Can't set line in Kermit ADDENDUM (Lenny Jacobs)
  Fonts: the crucial connection between IV and GS. (Peter MacDonald)
  SOLVED:Re: can't startx as non-root (Isaac Wong)
  Re: Troubles with X386-S3.2b: random lockups (Matthew A. Lewis)
  PROBLEM: Lilo and OS/2 bootmanager (Christoph Bernhardt)
  Re: Troubles with X386-S3.2b: random lockups (Matt Welsh)

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

Crossposted-To: comp.unix.pc-clone.32bit,comp.unix.bsd,comp.windows.x
From: [e-mail] (Mark A. Friedman)
Subject: Will Linux/386BSD/XFree86 run on my Gateway 2000 4DX2-66V?
Date: Tue, 8 Dec 1992 04:55:41 GMT


I am considering the purchase of a Gateway 2000 4DX2-66V system

(Micronics motherboard, 8MB RAM to be expanded to 16MB, VESA VL-Bus,
ATI Ultra Pro video card with Mac 32 68800 accelerator chip and 1 MB
VRAM, CrystalScan 1572FS monitor, Western Digital Caviar 2340 local
bus IDE hard disk)

with the intent of running Linux or 386BSD and XFree86 at home.

I've done my reading of the appropriate FAQ's and newsgroups over the
last few months.  Still, a few points confuse me about running the
software on this hardware.  (Ask me about SunOS and then I'll give
you an answer. :^) )

Thus, I would appreciate any advice or words of warning before I make
my purchase.  I'll summarize the responses I receive to the net.

1) Can someone verify that I will indeed be able to run Linux/386BSD
   with the above hardware configuration?  Or, am I making serious
   incompatibility errors with respect to the operating system?

2) Given that Gateway includes RIDE in its BIOS to interface with the
   32-bit local bus (which will not be used by Unix), will 386BSD/Linux
   take advantage of the 32-bit VL-Bus when interacting with the hard
   drive and monitor?  Will the VL-Bus offer noticable improvement?

3) From my reading of the XFree86 FAQ, the video card I intend to buy
   is not supported/recommended for use with XFree86.  However, I've been
   confused by several postings indicating success with various ATI
   accelerator cards.  Can someone verify that they indeed have had
   success running XFree86 with the ATI Ultra Pro video card with Mac 32
   68800 accelerator chip?  Can someone give a technical reason why it
   would be a difficult or next-to-impossible task to port XFree86 to
   this accelerator card?

4) What is the maximum resolution that I can expect to obtain with the
   ATI graphics card and the CrystalScan monitor?  Will I be able to
   obtain 1280x1024 NI resolution?

5) Finally, in general, is the above system consistent with the future
   of Linux/386BSD and XFree86?  That is, would you say for the
   configuration above that even though the current software does not
   take advantage of component X, component X is where the market is
   heading and the software will most certainly support that component in
   the near future.

Thanks for taking the time to read through/respond to my posting.

-- Mark

Mark A. Friedman                         [e-mail]
Engineering and Computer Science         (or [e-mail])
Trinity College                          Phone: [phone]
Hartford, Connecticut   06106            Fax: [phone]


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

From: [e-mail] (Donald J. Becker)
Subject: Re: .98.5 and .98.6: Infinite loop trying to init SMC Elite16
Date: Mon, 7 Dec 1992 21:21:01 GMT

In article <[e-mail]> [e-mail] (Adams) writes:
>In article <[e-mail]> [e-mail] (Jim Bray) writes:
>>     I just tried the 98.6 kernel. The SCSI driver seemed to work fine
>>   when I switched to enhanced mode.  The wd8013 driver went into the
>>   "wd8013_interrupt infinite loop detected: counter ..." loop when
>>   trying to autoconfig/initialize the SMC Elite16 card which is supposed
>>   to be wd8013-compatible (the sysv.4 wd8013 driver works OK with it). I
>>   tried changing parameters like IRQ and so forth but this appeared to
>>   have no effect.
>>     I have no manual for this card, unfortunately. I'm going to look at
>>   the code a bit anyway.
>       Strange at the glance, but occured and took about
>       3 hours to be detected.
>       Cable segment was terminated with a terminator supplied by the
>       vendor of the card at one end, and a terminator from Hirschmann
>       with proven 50 Ohms.  PC was the only member on its cable
>       segment, leading to a star coupler by Hirschmann.
>       Ping to other systems on *_same_ side
>       of the router* returned with "ICMP .... Host unreachable",
>       I knew for sure, our SUN was up and healthy ..
>       It turned out, that terminator from the card vendor had  75 Ohms,
>       yielding a total resistance of 30 Ohms against ground ( 75 || 50),
>       instead of 25 (50 || 50), as required.
>       May be CD (collision detection) would not work anymore reliably...

I'll second this: ALWAYS HAVE THE CARD HOOKED UP TO A KNOWN-GOOD
NETWORK!  While I was develoing the 3c503 driver I threw away lots of
good code, and wasted a lot of time finding this out.  The card will
sometimes work without a cable, but most of the time it won't.

>>   --Jim Bray ([e-mail])
>
>Software (we.c) seems to run, but:
>       I would like to see variables in shared memory declared
>       volatile..., as both CPU and Ethernet controller access them.
>
>best adams

Declaring them volatile wouldn't help: that shared memory won't
change while you are looking at it.  You _can_ worry about the cache...


--
Donald Becker                                  [e-mail]
Supercomputing Research Center
17100 Science Drive, Bowie MD 21114                301-805-7482

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

From: [e-mail] (Cameron Spitzer)
Subject: hardware hacks, was Re: MFM and IDE drives coexist under Linux
Date: 8 Dec 92 02:57:11 GMT
Reply-To: [e-mail]

In article [e-mail], [e-mail] (Delman Lee) writes:
>In article <[e-mail]> [e-mail] (Donald E. Cope) writes:
>   I have a system with two controlers running susessfully, both have two drives
>   each, for  a total of 210meg.  They both operate on IRQ 14 ..,
>   you need to isolate the IRQ lines on both controlers from the bus.  This was
>   accomplished by cutting solder paths on each board and adding a diode(in914)
>   and a resister to isolate the IRQ from the bus.
>
>Hi! I am MOST interested in the details...
>I have a couple of questions: Does BOTH IRQ14 lines on the 2
>controllers need to be isolated?


                                  Understand this.
                 "Seems to work for me, for now" is very far from
                           "guaranteed to work for you."

A passive diode OR is *not* reliable in TTL*, because of the 0.7 V drop across the
diode.  There are lots of ways to build a wired-OR, TTL-compatible gate, but
silicon signal diodes are not among them!  Use open-collector or 3-state buffers,
or low-drop diodes (germanium or Shottky) or build yourself an open-collector
inverter with a resistor and a transistor.  Details on request.

Sheesh!  Next we'll have the Liquid Solder fans telling us we don't need
soldering irons any more.  And let me tell you how I patch my bicycle tires
with duct tape...  :-)

Cameron, continuously amazed


*TTL == Transistor Transistor Logic, used on the Industry Standard Architecture
(aka AT-compatible) bus.  A valid TTL zero *output* is 0.5 V or lower.

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

From: [e-mail] (Chris Newbold)
Subject: Troubles with X386-S3.2b: random lockups
Date: 8 Dec 92 04:50:03 GMT

I think there may be something wrong with X386-S3.2b. It locks
up randomly on my machine. It seems to happen more when memory
is low and swap space is close being exhausted. Sometimes it
seems to happen during heavy disk access, but it has happened
while all is quiet.

Typically, I'll switch to a window and type a command like
tar or man and then the thing locks. No keyboard no mouse.

I first thought it may be trouble with the SCSI disk drivers;
people had been reporting random lockups with the aha1542.
Having an EISA-bus system with an ESIA SCSI controller compatible
with the aha1542, I don't think I could be experiencing the
troubles others are.

I did give the kernel a thourough thrashing outside X; I got
the load average up around 3.00 by simultainiously:
        compiling the kernel & ghostscript
        running iozone
        running a csh loop using groff to format a man
                page through /dev/null forever
        running several finds on the whole filesystem
        loading about 6-7 emacs
        and telnetting to myself several times...

This kept both my disks rather busy :-) Solid as a
rock, though. iozone even turned in rates of 475kb/sec
under this load!

Anyway, I don't think it's the disk drivers as this was
the hardest I've ever pushed the system. It could be
a swapping bug in the kernel... I couldn't really test
this without X as the shared libs made it very hard to
eat up all of the RAM and swapspace.

I'm running 0.98-pl6. I did not have these troubles under
0.98-pl5, so maybe there is something in the kernel.

I can't run X as non-root anymore either. It compians ot
can't open /dev/console, but if I su and chown /dev/console
and then X it just locks up.

Help!

-Chris

--
>>>> Chris Newbold <<<< * "If you fool around with a thing for very long you *
University of Rochester *                 will screw it up."                 *
Disclaimer: "All warranties expire upon payment of invoice."
[e-mail] * uhura.cc.rochester.edu!ctne_ltd@uunet

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

Subject: Re: MFM and IDE drives coexist under Linux?
From: [e-mail] (Donald E. Cope)
Date: Mon,  7 Dec 92 14:17:01 EST

Delman Lee ([e-mail]) wrote:
: In article <[e-mail]> [e-mail] (Donald E. Cope) writes:
:    I have a system with two controlers running susessfully
:
: Hi! I am MOST interested in the details. I have 1IDE and 2MFM's that I
: am trying to work with both MS-DOS and Linux. Please do email me the
: details and patches to Linux.
:
: I have a couple of questions: Does BOTH IRQ14 lines on the 2
: controllers need to be isolated? Won't isolating the IRQ14 on the
: controller card operating at the secondary I/O port do?
:
: Many thanks --- Delman.
: --
: ______________________________________________________________________________
:
:   Delman Lee                                    Tel.: [phone]
:   Medical Image Processing Group,               Fax.: [phone]
:   University of Pennsylvania,
:   4th floor Blockley Hall,
:   418 Service Drive,
:   Philadelphia, PA 19104-6021,
:   U.S.A..                                   Internet: [e-mail]
: ______________________________________________________________________________
:
Both controlers need the isolation, because the one that does not will tend
to clamp the bus to a low sig(close to ground) and the other one under most
cases will not be able to override this conditon.  This condition also exists
for the COM boards that are trying to use the same IRQ line (ie. 4).
 I'll get a diff for the two files and post them. They will be against .98p1

 Don.
 email <vtrm01!vtrx33![e-mail]>


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

From: [e-mail] (Da Hubster!!!)
Subject: Disk Defragger???
Date: 7 Dec 1992 23:34:36 -0600



Is there a disk defragger avaible for Linux???

--
ADIOS!!!!!!                                            -----------------------
                                                      | Da  Hubster!!!        |
                                                      | [e-mail]  |
                                                       -----------------------

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

From: [e-mail] (Jon Anhold N8USK)
Subject: TinyMUSE
Date: Tue, 08 Dec 92 00:21:39 EST

If anyone can tell me how to get TinyMUSE running, please help! I grabbed
version 1.5 from michael.ai.mit.edu, and tried to coax it to compile, but
since I'm a novice at all of this, I was unsuccessful.

-j


--
|   |  |  |   -Jon Anhold N8USK- @ Dreamland Network Systems
+-*-+--+--+   (...{uunet|backbone}!dreaml!jga) ([e-mail])
| # |  |  |   Packet: N8USK @ NO8M    TCP/IP: [e-mail]
  #           "It's a fax from your dog, sir. It looks like your cat."

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

From: [e-mail] (Lenny Jacobs)
Subject: Can't set line in Kermit ADDENDUM
Date: 8 Dec 92 05:57:27 GMT


12/06/92

I'm using Linux 0.97 p2 and I can't use Kermit from any user other
than root. When I try to set line to ttys0 or ttys1 I get "Can't
open connection: operation not permitted. It seems that although
/dev/ttys2 and /dev/ttys3 are owned by bin, group tty; /dev/ttts0
and /dev/ttys1 are owned by root, group root. Also ttys2 and 3 are
marked read,write by user, group and other but ttys0 and 1 are
marked read, write by user and write only by group and other. The
system won't let me change the permissions to read, write for all
on these devices even though I'm logged in as root. How can
I change these permissions?

ADDENDUM 12/07/92

I experimented with this problem today and realised that the problem
was caused because I had "getty" lines in my inittab which addressed
ttys0 and ttys1. It seems that getty removes the group and other read
permissions from ttys lines that it is using so that root and either
getty or group & other can have control of serial lines. I wonder if
this is standard "UNIX" or just a quirk of Linux?
--
+----------------------------------------------------------------------+
|    <-> lenny <-> [e-mail] <-> [e-mail] <->     |
+----------------------------------------------------------------------+

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

Crossposted-To: gnu.ghostscript.bugs,comp.windows.interviews
From: pmacdona@sanjuan (Peter MacDonald)
Subject: Fonts: the crucial connection between IV and GS.
Date: Tue, 8 Dec 92 03:12:50 GMT

For some time SLS (a free 386 Linux/Unix distribution) has bundled
both Doc, idraw and ghostscript.  The idea was that IV would
provide (among other things) Desktop Publishing and GS would
provide diverse device output support for PS.

Last week I got a coprocessor, and started to test this out,
only to find that IV uses the X fonts, while GS uses it's
own.  I have heard that GS will also work with the Type1
fonts, but have been unable to get them to work.

So what I would like to do is 1) get GS to support some or most
X11 fonts, or failing that, 2) find a way to map to fonts GS
includes and supports.   Obviously, 1) is preferred, even if
I have to hack GS myself.

Has anyone done anything on this?  Or am I missing some easy
solution (is my fly open)?

Peter
[e-mail]


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

From: [e-mail] (Isaac Wong)
Subject: SOLVED:Re: can't startx as non-root
Date: Tue, 8 Dec 1992 06:15:49 GMT


a few people have asked me to summarize, so here it comes.
the only response i've received is from James Chacon and it solved the problem.
thanks james. the only thing i may add is to make sure xinit is also setuid
root (ie chmod u+s /usr/bin/X11/xinit)


======================================================================
From: [e-mail] (James Michael Chacon)

Check to make sure xterm and the X386 binary are setuid root.
If you ls -l them you should see:
-rwsr-xr-x  1 root      wheel   123908 Dec 5 23:33 xterm
-rwsr-xr-x  1 root      wheel   893956 Dec 5 23:33 X386

The group can be different depending on how you installed everything.
The main thing, is that the s is there on the 4th character.

If it isn't, do the next 2 commands:
chmod 4755 /usr/X386/bin/xterm
chmod 4755 /usr/X386/bin/X386

That should fix any problems. Oh yea, make sure /tmp has the right permissions
also. It should be read-writable by anyone with the sticky bit set also.

Do a chmod 1777 /tmp to make sure.



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

From: [e-mail] (Matthew A. Lewis)
Subject: Re: Troubles with X386-S3.2b: random lockups
Date: 8 Dec 1992 07:00:13 GMT
Reply-To: [e-mail] (Matthew A. Lewis)


In a previous article, [e-mail] (Chris Newbold) says:

>Typically, I'll switch to a window and type a command like
>tar or man and then the thing locks. No keyboard no mouse.
>
>I did give the kernel a thourough thrashing outside X; I got
>the load average up around 3.00 by simultainiously:
>       compiling the kernel & ghostscript
>       running iozone
>       running a csh loop using groff to format a man
>               page through /dev/null forever
>       running several finds on the whole filesystem
>       loading about 6-7 emacs
>       and telnetting to myself several times...

I suspect this is due to the socket overload that Matt Welsh was
telling me about. I hear there is new code coming out
soon. Watch for it as i am :)

Later

--
maThiEu a. LeWis                     THE NUCLEAR CLUB
cyber-mercenary   USA,Russia,Ukraine,Belarus,Kazakhstan,England,France,China
Thieves Like Us   Secret Members:S.Africa,India,Israel,Pakistan
[e-mail] Wannabees:Iraq,Iran,N.Korea,Libya,Brazil,Argentina,Taiwan,ME

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

From: [e-mail] (Christoph Bernhardt)
Subject: PROBLEM: Lilo and OS/2 bootmanager
Date: 8 Dec 92 09:40:14


After successfully installing Hlu's rootdisk (0.98pl5) I encountered a problem
when I tried to boot Linux from the OS/2 bootmanager.

Making the Linux partition bootable -- while disabling the bootmanager -- works
just fine. Lilo comes up and bootstraps correctly. But when I reactivate the
bootmanager and chose the Linux partition from the bootmanagers menu the screen
blanks and the first "L" of the Lilo bootmessage appears, and that's it. The
harddrive light remains on and nothing happens. At least Ctrl-Alt-Del solves
this problem, but instead of Linux I see the bootmanager again.

Can anyone tell what is happening?
Is there another way to chose a boot partition during the boot process?

My configuration:
        386/33&387 12Mb 420Mb Fuj & 170MB Quant IDE


--
===============================================================================
Christoph Bernhardt
University of Karlsruhe
Institute of Telematics
email: [e-mail]

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

From: [e-mail] (Matt Welsh)
Subject: Re: Troubles with X386-S3.2b: random lockups
Date: Tue, 8 Dec 1992 07:29:24 GMT

In article <[e-mail]> [e-mail] (Matthew A. Lewis) writes:
>
>In a previous article, [e-mail] (Chris Newbold) says:
>
>>Typically, I'll switch to a window and type a command like
>>tar or man and then the thing locks. No keyboard no mouse.
>>
>>I did give the kernel a thourough thrashing outside X; I got
>>the load average up around 3.00 by simultainiously:
>>      compiling the kernel & ghostscript
>>      running iozone
>>      running a csh loop using groff to format a man
>>              page through /dev/null forever
>>      running several finds on the whole filesystem
>>      loading about 6-7 emacs
>>      and telnetting to myself several times...
>
>I suspect this is due to the socket overload that Matt Welsh was
>telling me about. I hear there is new code coming out
>soon. Watch for it as i am :)

Not exactly the same problem, unfortunately. What I was describing
is a bug (that should be fixed with today's tcpip patches) which causes
a null pointer dereference in the kernel, in dev_tint() when transmitting
a lot of packets. It's more prominent than a simple hang. :)

The patches for this are on Sunsite. I don't know about this X problem.
Doesn't sound like a load problem, but rather a hang somewhere.
If you have anything compiled with -g, send a kill to one or more
of the processes and look at the core dump. May be the best way to
figure out where it's stuck.

mdw

--
Matt Welsh     [e-mail]     Cornell Theory Center
  "We're going away now. I fed the cat."

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


** 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: [e-mail]

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

    Internet: [e-mail]

Linux may be obtained via one of these FTP sites:
    nic.funet.fi                pub/OS/Linux
    tsx-11.mit.edu                pub/linux
    tupac-amaru.informatik.rwth-aachen.de    pub/msdos/replace

The current version of Linux is 0.97 released on August 1, 1992

End of Linux-Activists Digest
******************************