Subject: Linux-Development Digest #948
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:     Tue, 26 Jul 94 07:13:04 EDT

Linux-Development Digest #948, Volume #1         Tue, 26 Jul 94 07:13:04 EDT

Contents:
  Re: ERROR: Not mounting SCSI cd-rom with 1.1.35 (Jim Balter)
  Kernel 34 and Virtual Memory (Yeo Ann Kian)
  threads in kernel (G Dinesh Dutt)
  Anyone working on IDE CDROM support? (Dave Snowdon)
  Re: [BUG?] VM problem -- possible security hole (Rob Janssen)
  Re: Dosemu won't work with >1.1.29 <1.1.35 and recompile won't help (Rob Janssen)
  Re: When will 1.1.?? become 1.2.0 (Rob Janssen)
  Re: Floppy error on TI Travelmate Laptop (Rob Janssen)
  Re: gcc-2.6.0: help w/asm (Caleb Epstein)
  SWAP broken in v1.1.35 ?! (Klaus Schneider)
  SLIP hangs in newer kernels (Klaus Schneider)
  DOS EMU under 1.1.34 (Steve Brown)
  Linux 1.1.35 and ld -qmagic (Robert KNOTEK)
  Re: Ooops! SCSI CD-Rom broken in 1.1.33 (Keith Pritchard)
  Re: what's this stuff? (pI[ITE PO-RU ...) (Jim Balter)
  Re: Ooops! SCSI CD-Rom broken in 1.1.33 (Ramon Martinez)
  QLogic status and anyone doing PCMCIA support? (Thomas E Zerucha)

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

From: jqb@netcom.com (Jim Balter)
Subject: Re: ERROR: Not mounting SCSI cd-rom with 1.1.35
Date: Tue, 26 Jul 1994 03:45:55 GMT

In article <310sq1$gs8@bee.uspnet.usp.br>,
Andre U Manoel <andre@futuro.usp.br> wrote:
>
>       Please do forget about the above message. I just read a post on
>comp.os.misc.development and found that all I have to do is to mount the
>CD-ROM as readonly. Otherwise it won't. 

I know it's a little bit backwards, but the error message would at least
point to the problem if sys_mount returned EROFS instead of EACCES.
-- 
<J Q B>

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

From: yeoak@cyberspace.org (Yeo Ann Kian)
Subject: Kernel 34 and Virtual Memory
Date: 25 Jul 1994 06:01:10 GMT

Kernel 1.1.31 to 1.1.34 seems to be eating up virtual memory.
I have only 4 meg of phisical memory and so I make use of my 10meg
swap partition extensively.  It turns out that my swap uses up quickly 
after running a few processes.  I ran top to trace the swap memory and I
discovered that a program does not or releases very little of its virtual
memory used once it ends.  I was told to get kernel 35 but I would like to know
if it does address the problem problem as it takes me hours to get a kernel compiled.

BTW is there anyway which I could release (or compact?) the virtual memory
manually?

Thanks for any help offered.
A.K.

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

From: G Dinesh Dutt <brat@htilbom.ernet.in>
Subject: threads in kernel
Date: 26 Jul 1994 02:00:53 -0400
Reply-To: brat@htilbom.ernet.in

Hi everybody,

I have been seeing a lot of messages with people interested in seeing threads
in Linux kernel. Well, development for one has already started. We'll let you
know as and when results come in.

Dinesh
----
There's more to life than just increasing its speed.
                                                - Mahatma Gandhi.
G. Dinesh Dutt,                                 email : brat@htilbom.ernet.in
Project Leader,                                 voice : 8321144/8349393/8391478
Hinditron Tektronix Instruments Ltd.,
SDF-2, Unit 63-A, SEEPZ, Andheri (east), Bombay - 400096.


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

From: dns@cs.nott.ac.uk (Dave Snowdon)
Subject: Anyone working on IDE CDROM support?
Date: Mon, 25 Jul 94 21:04:19 GMT

Hi
I've got a Gateway 2000 PCI bus 486DX2-66 (P4D-66) which comes with an
NEC CDR-260 IDE CDROM drive.

The current CDROM-HOWTO indicates that IDE CDROM drives are not
supported - is anyone working on adding support for them?

If no-one's currently working on this, perhaps someone could give me a few
pointers on where to start in order to code a driver myself - I haven't
written a UNIX device driver before, but theres always a first time :-)

                        cheers
                                Dave

-- 
Dave Snowdon                 
Communications Research Group                 Tel:    +44 (0)602 514226
Department of Computer Science                Fax:    +44 (0)602 514254
The University of Nottingham                  E-mail: dns@cs.nott.ac.uk
Nottingham NG7 2RD, UK    <http://www.crg.cs.nott.ac.uk/~dns/dave.html>

                "This mind intentionally left blank"

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

From: rob@pe1chl.ampr.org (Rob Janssen)
Subject: Re: [BUG?] VM problem -- possible security hole
Reply-To: pe1chl@rabo.nl
Date: Mon, 25 Jul 1994 21:53:05 GMT

In <30vrfr$8nh@nippur.irb.hr> marino@ds5000.irb.hr (Marino Ladavac) writes:

>a program froze the VM system.  This is the offending code:

>#include <stdlib.h>

>main()
>{
>       char  buffer[200];

>       memmove( buffer, buffer + 30, -22 );
>}

>on memmove call the VM started paging new memory, exhausting all the 
>available memory (RAM and swap) and then everything froze.

>-22 for the count was a bug, but shouldn't I have run into an unmapped
>page or non-writable page? (note: buffer is allocated on the stack!!)
>During execution the program occupied more than 20 MB before it froze,
>but the top was not showing this program size (top managed to work due
>to the slow swap)

>security hole is, of course, the ability of everyone to freeze the 
>system.

You can use "ulimit" to limit the amount of memory and/or stack available
to a program.  When this is done from /etc/profile and some reasonable
values are being used, it should prevent a user from allocating all
memory.

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

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

From: rob@pe1chl.ampr.org (Rob Janssen)
Subject: Re: Dosemu won't work with >1.1.29 <1.1.35 and recompile won't help
Reply-To: pe1chl@rabo.nl
Date: Mon, 25 Jul 1994 21:55:35 GMT

In <1994Jul25.110259.4029@tudedv.et.tudelft.nl> lcvanveen@et.tudelft.nl writes:

>Last week I rebuild Linux to version 1.33 and noticed dosemu0.52 
>wouldn't work anymore. The screen went black and the only thing 
>I could do was shutdown with crtl-alt-del. This morning I went 
>over to Linux 1.1.35 and recompiled dosemu again. I still won't 
>work. I does in an xterm, but goes black in a VC. I then set 
>commented aout the S3-support and the direct access to the 
>videoport and got a segmentation fault that brought Linux to a 
>complete standstill. The only thing I could do was reset my poor 
>machine.
>Does anybody know what's changed within Linux VC support that 
>Thanks a lot in advance.

It is not VC support but VM support.  Dosemu people are working on it.

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

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

From: rob@pe1chl.ampr.org (Rob Janssen)
Subject: Re: When will 1.1.?? become 1.2.0
Reply-To: pe1chl@rabo.nl
Date: Mon, 25 Jul 1994 21:57:52 GMT

In <310f2n$k6r@inet.up.ac.za> phil@lostris.up.ac.za (Phillip Stofberg) writes:

>The stable version of the linux kernel is currently at 1.0.9. A lot of us want 
>the features and enhancements that is in 1.1.??, but don't want to experiment 
>with a unstable kernal. 

>What is involved in going to 1.2.0 and when can we start to expect this? 
>Can't Linus call a code freeze on 1.1.35+ and let it stabelize into 1.2.0?

He actually has announced he would do so, but then there were so many
last-minute additions that the whole thing became immensely unstable :-(

I think some patience is required.  Linus is still releasing at a much
higher rate than some commercial suppliers :-)

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

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

From: rob@pe1chl.ampr.org (Rob Janssen)
Subject: Re: Floppy error on TI Travelmate Laptop
Reply-To: pe1chl@rabo.nl
Date: Mon, 25 Jul 1994 22:00:19 GMT

In <1994Jul25.133746.4060@newsserver.rrzn.uni-hannover.de> hillbrec@informatik.uni-hannover.de (Dirk Hillbrecht) writes:

>Hello everyone,

>I have here a laptop TI Travelmate 4000 WinDX2/50 running Linux with the
>following problem: The disc drive works only if I move the mouse under X.
>This is only
>with the newer kernel versions: 1.1.29 and 1.1.34 are prooved to have
>this bug, 1.1.8 is prooved to have it not. Exact error description:
>Insert disc - make something like mdir or mcopy - disc LED goes on but
>nothing happens - some minutes later: disc access and program stops without 
>any error but also without any productive results... - - - do the same 
>under X and move the mouse whilst Linux reading from/writing to the floppy: 
>everything goes well and with the appropriate speed, but only as long as
>the mouse is moved. If the mouse stops, the disk drive stops too. :-(

>This happens not (and in fact, never happened) to the disk drives in my
>big computer. Is anyone of the kernel hackers able to remove this annoying
>bug?

I had floppy drive problems as well, for the first time in the history
of Linux running on my machine...  After some close examination of the
startup messages I found another driver auto-detected its IRQ to be 6,
the IRQ of the floppy.  For me it was an ethernet driver, for you it seems
to be the mouse.  I sent a patch to fix this to Linus.

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

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

From: cae@cae.ny.jpmorgan.com (Caleb Epstein)
Crossposted-To: gnu.gcc.help
Subject: Re: gcc-2.6.0: help w/asm
Date: 25 Jul 1994 13:51:04 GMT

I wrote:

Caleb> [ I'm having trouble compiling the Linux kernel w/gcc-2.6.0 ]

Caleb> static inline unsigned char xchgb(unsigned char reg,
Caleb>                            volatile unsigned char *mem)
Caleb> {
Caleb>   asm("xchgb %0,%1" :
Caleb>       "=r" (reg), "=m" (*(unsigned char *)mem) :
Caleb>       "0" (reg), "1" (*(unsigned char *)mem));        /* line 309 */
Caleb>   return reg;
Caleb> }

Caleb> [ gcc complains "inconsistent operand constraints in an 'asm'" 
Caleb>   on line 309; this is an error not a warning. ]

Richard Earnshaw wrote:

Richard> I'm not a 386 expert, but try:
Richard> static inline unsigned char xchgb(unsigned char reg,
Richard>                                  volatile unsigned char *mem)
Richard> {
Richard>    asm ("xchgb %0, %1" :
Richard>        "+r" (reg), "+m" (*(unsigned char *)mem));
Richard>    return reg;
Richard> }

Richard> '+' in a constraint says that the operand is read as well as written.

        Well, I tried Richard's suggestion and unfortunately it didn't
work (thanks for trying, though, Richard).  I get the error 

        ultrastor.c:308: output operand constraint contains '+'

        Line 308 is the line with the +r/+m stuff.  There was a recent
patch (1.1.35) to the Linux kernel which changed the "=r" in the
original xchgb() function to an "=q", but this also fails to correct
the problem - I get the same error as with the original "=r".
Combining the two patches so that Richard's suggestion becomes
"+q"... also fails.

David Kastrup wrote:

David> The "r" constraint is definitely wrong, as it can indicate an SI or
David> DI register as well.  Look into your docs.

        Which docs?  I've looked briefly at the gcc info pages, but as
I know zero assembler this really isn't that helpful.  It is written
from the point of view of someone who already knows what they are
doing, and I don't.  Your note about the "r" constraint is correct, as
the official kernel has been patched to change the "r" to a "q".
Unfortunately it still doesn't fix the error when compiling the code.

        Is it possible that this is simply a bug in gcc-2.6.0?  I've
tried every combination of "=" and "+" constraints on the original
code and on Richard's suggested patch.  Each one fails with either the
original error or the "constraint contains '+'" error.  I'd really
appreciate any help on this.  Thanks,

--
[ cae@jpmorgan.com ][ Caleb Epstein ][ JP Morgan & Co. Incorporated ]

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

From: uk0q@rzstud1.rz.uni-karlsruhe.de (Klaus Schneider)
Subject: SWAP broken in v1.1.35 ?!
Date: 25 Jul 1994 23:04:02 GMT

It appears that for the first time I use Linux (since 1st quarter of
1993) swapping is broken in kernel v1.1.35.

While I let my 'puter run idle, I discovered it gave me countless
errors:

        swap_duplicate: trying to duplicate unused page

I tried to switch off the swapping and got (besides a successful
"unmount" of the two swap-partitions):

        swap_free: swap-space map bad (entry 0003b000)
        swap_free: swap-space map bad (entry 0003c000)

Seems something is broken here.  :-(

Klaus
===============================================================================
Klaus Schneider                         Email: root@ks1i486.dialup.xlink.net
Student of Informatics                         uk0q@rzstud1.rz.uni-karlsruhe.de
University of Karlsruhe, Germany               klsc@delphi.com
===============================================================================




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

From: uk0q@rzstud1.rz.uni-karlsruhe.de (Klaus Schneider)
Subject: SLIP hangs in newer kernels
Date: 25 Jul 1994 23:06:12 GMT

Since about kernel 1.1.25 my SLIP connections hangs for a few seconds
or even minutes before resuming its work.  This still happens in
v1.1.35.  Has anybody else experienced a similar problem?

Klaus
===============================================================================
Klaus Schneider                         Email: root@ks1i486.dialup.xlink.net
Student of Informatics                         uk0q@rzstud1.rz.uni-karlsruhe.de
University of Karlsruhe, Germany               klsc@delphi.com
===============================================================================


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

From: stevebr@cherokee (Steve Brown)
Subject: DOS EMU under 1.1.34
Date: 26 Jul 1994 09:23:57 GMT

Hi Folks,
        I want to use Dosemu0.52, and the instructions tell me I need to
upgrade my kernel ( I was on 1.0.8 ) so I did. Only snag is that Dosemu
wont compile, it moans about too many things to mention ( and they
fly off my screen at a rare rate of knots ), in sigsegv.c and its
associated header files ( one of them is net.h! ).

Does anyone else have this problem, or could someone tar the binaries
up for me? Is my source tree corrupted ( the kernel compiled just fine! )
Or is Dosemu0.52 not good with 1.1.34, in which case which source tree should
I use?

Thanks in Advance,
Steve.

--
=============================================================================
Steve Brown           |  stevebr@summer.bt.co.uk  | Act upon what matters,
Room 209, B81. PP35   | brown_s_m@bt-web.bt.co.uk | Not how you're perceived.
BT Labs               |                           | 
Ipswich, England.     |    Tel: +44 473 640523    | D A Gibson.
=============================================================================

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

From: idefix@venus (Robert KNOTEK)
Subject: Linux 1.1.35 and ld -qmagic
Date: 26 Jul 94 11:08:25 +0100

Trying to compile Linux kernel 1.1.35 with gcc 2.5.8 

compiling works fine but at the end there ist an ld switch (-qmagic)
which my version of ld dont know.

Is there a newer version of ld (how to find, i searched on nic.funet.fi and 
tsx-11.mit.edu but can't find).

                                                        regards
-- 
Robert KNOTEK                                           R.Knotek@elin.co.at
ELIN-Energieanwendung GesmbH            
A-1141 Vienna, Austria, Penzingerstr.76
Phone: +43 1 89100-2176                                 Fax: +43 1 8946046

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

From: keith@datawell.demon.co.uk (Keith Pritchard)
Subject: Re: Ooops! SCSI CD-Rom broken in 1.1.33
Date: Sat, 23 Jul 1994 13:33:56 GMT

Christopher Shaulis (cjs@netcom.com) wrote:
: Kernel Version: 1.1.33
: 'puter: 486XDX/33, 8M ram
: SCSI Host: Adaptec 1542C
: CD-ROM:
:    Vendor: TEAC      Model: CD-ROM CD-50      Rev: 1.06
:    Type:   CD-ROM                             ANSI SCSI revision: 02
:    Detected scsi CD-ROM sr0 at scsi0, id 2, lun 0
:    scsi : detected 1 SCSI disk 1 CD-ROM drive total.

: 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#

Try mounting it Read only as follows

mount -t iso9660 /dev/cdrom -o ro /system_cd/


cheers
hope it works :)
Keith
--
 =============================================================================
 Name:-            |   Email:-                      |  Quote of the week:-
 Keith Pritchard   |   keith@datawell.demon.co.uk   |  Its Hard to work in a
 Aughton           |   keith@sc.lancsp.ac.uk        |  team when youre
 Lancashire,UK     |   100101.1450@compuserve.com   |  omnipotent ! 'Q' ST:TNG
 ============= PGP Public Key Available from the usual Servers ===============

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

From: jqb@netcom.com (Jim Balter)
Subject: Re: what's this stuff? (pI[ITE PO-RU ...)
Date: Tue, 26 Jul 1994 09:56:11 GMT

In article <RSANDERS.94Jul25153654@hrothgar.mindspring.com>,
Robert Sanders <rsanders@mindspring.com> wrote:
>On 25 Jul 1994 14:59:52 GMT, vox@asterix.uni-muenster.de (Andreas Vox) said:
>
>> A friend of mine suggested this solution, in case you set it off by
>> mistake:
>
>> type (blindly): cat /dev/hda ============
>
>> then interrupt with CTRL-C
>
>Come on, I keep seeing "cat /vmlinuz", "cat /dev/hda" and other
>totally bogus tricks.  It's very simple: ^N messes up your screen, ^O
>fixes it.  So, all you need is a way to send ^O to your screen, so
>let's make that file.
>
>Type exactly what you see below, with one exception: where it says ^O,
>hold down the CTRL key and press the 'O' key, and where it says ^D,
>hold down the CTRL key and press the 'D' key.
>
>cat > fix
>#!/bin/sh
>echo ^O
>^D
>chmod 755 fix
>
>Now, just put that somewhere in your PATH and you'll be able to make
>your screen happy by typing "fix" at the command line.

Much more effective, for a wider array of problems, is <escape>c
-- 
<J Q B>

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

From: rampa@encomix.exnet.com (Ramon Martinez)
Subject: Re: Ooops! SCSI CD-Rom broken in 1.1.33
Date: Tue, 26 Jul 1994 02:10:22 GMT


:>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-


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

From: zerucha@shell.portal.com (Thomas E Zerucha)
Subject: QLogic status and anyone doing PCMCIA support?
Date: 24 Jul 1994 02:09:57 GMT

I was wondering how the Q-Logic driver is going.  Since I have two such cards,
it would be nice to have.

On the other front (and since my second card is in my laptop), is anyone
doing anything about PCMCIA support (including hot-swaps and enabling)
for Linux?  Does anyone have a working Linux laptop using PCMCIA cards?
---
zerucha@shell.portal.com - main email address


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


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