Files
oldlinux-files/ftp-archives/tsx-11.mit.edu/1996-10-07/mail-archive/linux-devel/Volume2/digest248
2024-02-19 00:24:15 -05:00

553 lines
20 KiB
Plaintext

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: Sat, 1 Oct 94 04:13:12 EDT
Subject: Linux-Development Digest #248
Linux-Development Digest #248, Volume #2 Sat, 1 Oct 94 04:13:12 EDT
Contents:
Re: people using SCSI-IN2000 driver, please read this (John Shifflett)
boot disk for adaptec 7870 needed (Pong S Choa)
Re: SCSI PCI Adaptec (pp000547@interramp.com)
Re: Linux on CD (Jeff Kesselman)
Report on gcc with P5 optimizations (Erann Gat)
emufs (DOSEmu filesystem) - is it possible? (an128523@anon.penet.fi)
Re: [?] DIP with auto-redial? (Michael Will)
Re: Login USERID length bug? (Anat Faigon)
Obsolete route request made!? (Carlos Antunes)
Re: LOOK FIRST-- FORGED SPAM (Jeff Kesselman)
a bug in gcc 2.6.0/Linux (BARTHOLDI Laurent)
Re: OpenGL Extensions to X server? (Vince Taluskie)
Re: Shared Libs: working toward a permanent solution? (Robert Andrew Ryan)
Re: Special Sale On QNX! (Luke Howard)
Re: Porting from SunOS to Linux (Dynamic Linking) (Mitchum DSouza)
Re: Multiprocessing Pentium Systems (Aris Stathakis)
----------------------------------------------------------------------------
From: jshiffle@netcom.com (John Shifflett)
Subject: Re: people using SCSI-IN2000 driver, please read this
Date: Tue, 27 Sep 1994 23:49:29 GMT
doolitt@recycle.cebaf.gov (Larry Doolittle) writes:
>Stefan Markgraf (stefan@pippi.tu-bs.de) wrote:
>: Sometimes while swapping (to a swap-partition) the driver crashes.
>: No read/writes are possible and therefore every process begins to hang
>: when it needs access to the disk. The HD light burns constantly.
>I have seen this, always with the "new" fast driver on tsx-11. If
I also see this with the new fast driver & 1.51. BUT, I really
don't think it's related to *swapping*, per se. My experience
suggests that this hang occurs (somewhat randomly) whenever you
write _large_ amounts of data to disk. My system uses a swap file,
so it seems unlikely that it's a swap file versus partition issue.
I can recreate the hang by something as simple as untarring the
linux source tree, or doing 'dd if=zImage of=/boot/zImage'. Not
always, but often enough that I keep a boot floppy with 1.0.9 and
the old driver handy (never have been convinced that the newer
kernels don't play a part in the problem). Also, there was a
noticable increase in failures when I went from 8 megs to 16.
This is high quality RAM, by the way, and hasn't shown ANY sign
of giving trouble anywhere else. So, my theory is that with the
filesystem buffers a lot bigger than they used to be, large
amounts of stuff will be written to buffers before getting flushed
every 5 seconds or whatever by update. During flush, then,
something in the driver gets overloaded and ends up hanging.
>: (My memory "upgrade" 8->16 MB has shown, that the driver crashing occurs
>: more seldom --- this is a result of fewer swapping requests, I suppose).
Don't know what to say...
John S. jshiffle@netcom.com
------------------------------
From: pchoa@world.std.com (Pong S Choa)
Subject: boot disk for adaptec 7870 needed
Date: Sat, 1 Oct 1994 04:52:59 GMT
I have a 486 pc with a Adaptec 7870 SCSI controler (for PCI bus) and a
Viper video card. I tried to installed Linux on it with the SCSI boot
disk. It does not locate any SCSI devices. Do anyone know where I
can find a boot disk that work with a adaptec 7870 scsi controler.
Pong
------------------------------
From: pp000547@interramp.com
Subject: Re: SCSI PCI Adaptec
Date: 28 Sep 1994 03:16:33 GMT
Reply-To: pp000547@interramp.com
In article <1994Sep27.194907.3789@news.uit.no> frodef@stud.cs.uit.no (Frode Vatvedt Fjeld) writes:
Does anyone know if any of the adaptec SCSI-drivers will work with
"Adaptec AIC-7870 Fast and Wide SCSI interface connected to the
processor via the PCI bus" ?
If not, is this chip very different from the "ordinary" adaptec
SCSI controllers? Would anyone know where I can get the needed
information to write a driver?
Thanks,
--
FWIW, I happened to be reading a Microsoft document called "The
Chicago Beta Reviewer's Guide" the other day and in it I came across
something that may shed some light on the Adaptec picture.
In the Beta Reviewer's Guide, MS makes much of the idea of
splitting device support into hardware-dependent and hardware-independent
components, the former being the responsibility of MS and the latter being
the responsibility of the hardware mfg.
I wonder: if this strategy gains widespread acceptance among
hardware mfgs, isn't Linux going to have to find a way to come to
grips with it?
Bill
--
Bill Hogan <pp000547@interramp.com>
"Show me a wisdom that is greater than kindness." [J-J.Rousseau]
------------------------------
From: jeffpk@netcom.com (Jeff Kesselman)
Subject: Re: Linux on CD
Date: Wed, 28 Sep 1994 00:21:00 GMT
In article <butler.780613856@bert.cs.byu.edu>,
Kevin J. Butler <butler@bert.cs.byu.edu> wrote:
>pc@dale.dircon.co.uk (Pete Chown) writes:
>
>>If you have an IBM mainframe to spare, and run MVS on it, you can set
>>it up to move files to slower discs or to tape if they haven't been
>>used for a while. But they remain in the catalogue and are moved back
>>invisibly when you next use them.
>
>>This, IMO, is the sort of thing that is required here.
>Exactly.
>
>A multi-tiered (secondary) storage sol'n:
>disk
>compressed disk
>cd (if avail)
>tape (if avail)
>
>complications arise w/ cd/tape, because they are removable--you end up
>interrupting to tell the operator "Insert [media] volume X..."
>But hey--its cheaper than robot mounted tapes, etc. ;-)
>
>Anyone working on anything like this???
>
>kb
>--
>Kevin Butler butler@bert.cs.byu.edu 8-)
>This is a sample .sig file!
Lets go one step further... I'ld like such a system to keep a catalog of
my CDs, so if I want t access something that is out on CD, and not
currently mounted, itll ask me to mount the proper CD by name.
Now THATS convenient!
Jeff Kesselman
------------------------------
From: gat@robotics.jpl.nasa.gov (Erann Gat)
Subject: Report on gcc with P5 optimizations
Date: Wed, 28 Sep 1994 10:17:36 -0800
FWIW, FYI, ETC:
Several people asked me to report my experiences using gcc-2.5.8 with
pentium optimizations. I got the version from sunsite, which appears
to be slightly different from the version from intel. (For one thing,
the intel version claims to be revision 2.4.0.) It compiled without a
hitch, and the kernel recompiled OK using -O2 but not -O4. This is
consistent with earlier reports. Bootup dies during delay loop
calibration.
I tried recompiling just this part of the kernel (main.c) using -O2, but
it died again a little later during the boot sequence. I have not had
time to do a detailed investigation as to the cause of the problem, but
for now I think it is prudent to consider -O4 to be unsafe.
Unlike earlier reports, I perceive no performance change using the new
compiler. However, much of what I am doing is memory-intensive, and so
I would not expect to see much speedup.
E.
--
Erann Gat
gat@robotics.jpl.nasa.gov
------------------------------
From: an128523@anon.penet.fi
Reply-To: an128523@anon.penet.fi
Date: Fri, 30 Sep 1994 17:18:04 UTC
Subject: emufs (DOSEmu filesystem) - is it possible?
Hi,
First of all, excuse me for using the anon service, something seems to
be happening to my news host so I can't post.
There have been so many posts about mounting filesystems that are
supported by DOS but not by Linux (ie: Novell, Stacker, Doublespace,
etc), that it got me to thinking.
Wouldn't it be possible to use userfs and hack DOSEMU so that it would
export its filesystems? What I'm thinking about is to have a user-defined
filesystem (the code would basically be a trimmed-down DOSEMU with hacks
to talk to userfs) that exports all drives available to a DOSEMU
session. Of course, the "emufs" driver would first do a standard DOS
boot (ie, loading CONFIG.SYS and AUTOEXEC.BAT) thus loading any drivers,
and then export them to Linux. The filesystem might look something like:
/dos
|
-----------------------------
| | | | |
A C D E LPT1
Of course, the LPT1 file is an idea for the future, so that emufs could
also make available, for example, Novell printers.
Now, is this possible? Am I crazy? I don't know enough about the
internals of DOSEMU of userfs to evaluate this.
Just an idea,
Mario Camou | MS-DOS is a bug
EDS Mexico Client-Server Technical Services | MS-Windows is a bug with a GUI
mario.camou@eng.eds.com | Linux: the best debugger
=========================================================================
To find out more about the anon service, send mail to help@anon.penet.fi.
Due to the double-blind, any mail replies to this message will be anonymized,
and an anonymous id will be allocated automatically. You have been warned.
Please report any problems, inappropriate use etc. to admin@anon.penet.fi.
------------------------------
From: zxmgv07@studserv.zdv.uni-tuebingen.de (Michael Will)
Subject: Re: [?] DIP with auto-redial?
Date: 27 Sep 94 13:18:26 GMT
I have moved from dip to
sunsite.unc.edu:/pub/Linux/system/network/serial/dslip203.tgz
which supports indial and outdial, lets specific users start/stop the connection
and automaticly redials when the line dropped or was busy.
Apart from the /dev/cua* itt even works with /dev/ttyS* which I need for
mgetty to work.
Cheers, Michael Will
------------------------------
From: anat@netcom.com (Anat Faigon)
Subject: Re: Login USERID length bug?
Date: Wed, 28 Sep 1994 06:36:35 GMT
Kenneth Tan (kenneth@bbs.sas.ntu.ac.sg) wrote:
: What bugs me more is that the _password_ is also limited to only 8 characters!
: (If you don't believe it, give it a try: Enter a very_long_password and try
: logging in. You can get in by just typing the first 8 characters of the
: password!)
:
Not a bug but an historic limitation. The Password encryption
algorithm of any traditional UNIX (Linux is not a special case)
is based on a 56-bit DES (Data Encryption Standard) key.
: Kenneth "Automan" Tan
-- Anat
--
Anat Faigon anat@netcom.com
------------------------------
From: cmsa@softsousa.pt (Carlos Antunes)
Subject: Obsolete route request made!?
Reply-To: Carlos.Antunes@softsousa.pt
Date: Wed, 28 Sep 1994 00:09:05 GMT
Hello, fellow linuxers!
After installing kernel 1.1.45, when starting, I receive a warning saying
'obsolete route request made'. Well, I removed the command
'route add 127.0.0.1' form rc.inet1 and the warning disapears, but cannot
telnet localhost, because there is no route info. What's up?
Thanx!
Regards,
Carlos Antunes.
--
Carlos Antunes @ SoftSousa Developing for 32bit MS Windows(tm)
Voice: 351-1-3975303 Windows NT(tm) and "Chicago"
Fax : 351-1-3975889 Console, GUI or Kernel Mode Drivers
------------------------------
Crossposted-To: comp.os.linux.help,comp.os.minix
From: jeffpk@netcom.com (Jeff Kesselman)
Subject: Re: LOOK FIRST-- FORGED SPAM
Date: Wed, 28 Sep 1994 07:15:26 GMT
In article <369f2n$12rt@tequesta.gate.net>,
Michael S. Scheidell <scheidel@gate.net> wrote:
>PLEASE READ THIS. I DID NOT MAKE THAT POST!
>
>This posting did not come from our office, but we are tracing it.
>If you look carefully at headers you will find sites that don't exist,
>and see that it does not trace back to scheidel@gate.net.
>
Hmm. Unfortunately most of us don't know all the sites that do our don't
exist. Due to the nature of the net, its also fairly common to find
things coming from places other then their originators home, so don't
think TOO harshly off us who bought it.
I'm glad you quickly responded though, that helps avoid confusion.
>We have a good idea of who it is and some admins are helping us trace it
>by comparing the posting on their machine to other postings.
>
>If you have the time and ability to help us trace this, we can give you
>info on who is doing it and maybe try to pressure HIS site to drop him.
>
I'm not in a position to help you, but I support your efforts. You
might point out to the site administrator that the poster violated both
your and QNX's trademarks, as well as enganing in what amounetd to
libel. If he were to be prosecuted, his site could find themselves in
the middle of some very messy leaglities (rmemebr what happend to Steve
Jackson...)
>Again, I did not do this post, and we have taken great pains to avoid
>offending anyone, and to date this is my first flame, even if undeserved.
>
>I thank those who looked at headers and saw it was a forgery, and look
>forward to a solution to this.
>
Let us know what happens.
>NOW TIME TO PAUSE AND PONDER:
>Seems like every time I get too full of myself, someone helps to humble
>me and remind me, that I too am just a sinner saved by the grace of God.
>
Oh, okay. Thats why you're happy in Florida, and I wasn't. To each
their own.
>NOW BACK TO BUSINESS:
>ps, reply with subject line "hacker" to avoid auto reply.
>
>--
>Michael S. Scheidell Florida Datamation, Inc.
>scheidel@gate.net (407) 241-2966
>Distributers of QNX Real Time OS Products (finger for more info)
>
>
------------------------------
From: lbartho@scsun.unige.ch (BARTHOLDI Laurent)
Subject: a bug in gcc 2.6.0/Linux
Date: Fri, 30 Sep 1994 17:28:41 GMT
Hi folks,
I got a weird problem with gcc 2.6.0:
main( int c ) { return c % 15; }
does not compile (gcc fills up all available memory and exits with a
'out of virtual memory' message). The simlilar program compiles fine:
const in xv = 15; main( int c ) { return c % xv; }
I suppose I should post this to some gcc group; I suppose it worked with
previous releases though I can't say.
--
==========
Larry `Hacker' Bartholdi, University of Geneva, Switzerland
E-Mail: lbartho@scsun.unige.ch or lbartho@cui.unige.ch
S-Mail: Larry Bartholdi / Barillette 11 / 1260 NYON / Switzerland
------------------------------
From: taluskie@utpapa.ph.utexas.edu (Vince Taluskie)
Subject: Re: OpenGL Extensions to X server?
Date: 30 Sep 1994 10:22:09 GMT
In article <CwuwxA.2EK@news.tudelft.nl>,
Robert Stockmann <stock@dutsh7.tudelft.nl> wrote:
>Yes it would certainly be nice! Then you could use your Linux -box
>as GL/X11 terminal for those SGI and IBM machines. Ive tried to
>get an SGI GL appliction to run with vogle but this package
>(although very nice) lacks about 90% of the true GL lib functions.
Have you tried Ygl yet ? I'm hardly expert about such things but I
remember some discussion on comp.unix.aix a while back about how it
was fast than IBM GL offering...
Vince
--
Vince Taluskie
UT Physics Computer Group taluskie@utpapa.ph.utexas.edu
UNIX Systems Administration http://www.ph.utexas.edu/~taluskie/home.html
"It's easier to be a result of the past than a cause of the future"
------------------------------
From: Robert Andrew Ryan <rr2b+@andrew.cmu.edu>
Subject: Re: Shared Libs: working toward a permanent solution?
Date: Wed, 28 Sep 1994 12:37:44 -0400
Excerpts from netnews.comp.os.linux.development: 28-Sep-94 Re: Shared
Libs: working to.. Richard Krehbiel@netcom1 (1432)
> So, share it. (I do understand that's what shared libraries are for.)
> The library image is loaded and relocated so that it exists at a
> single virtual address, and it's the same virtual address for all
> programs that link with it (as is currently the case). The difference
> is that the load address was chosen at load time, not at link time.
The biggest drawback to this solution is that libraries relocated at
runtime would differ from the version of the libraries on disk, thus
swap space would have to be allocated to swap them out. (with ELF and
pic or the current shared libraries nearly all the pages can be paged
from the library files on the filesystem.)
-Rob
------------------------------
From: lukeh@zola.apana.org.au (Luke Howard)
Subject: Re: Special Sale On QNX!
Date: Thu, 29 Sep 94 09:57:41 +1000
In <jeffpkCwrMHE.D64@netcom.com> jeffpk@netcom.com (Jeff Kesselman) writes:
>In article <CwpIGD.GyH@pe1chl.ampr.org>, Rob Janssen <pe1chl@rabo.nl> wrote:
>>In <CwoFHF.Hxs@undergrad.math.uwaterloo.ca> scheidel@gate.net () writes:
>>
>>>Why settle for slow and obselete Unix such as UnixWare, Sun Solaris,
>>>SCO, Linux or BSD when you can have POWER & RELIABILITY with QNX 4.21!
>>>Stop playing games with these inferior o/s's and switch to QNX today.
>>
>>Hey come on, a bit more humble announcements would be nice!
>>Especially from someone who posts the same article twice in the same
>>group. Done under QNX, I suppose?
Must be QNX's fail-safe redundancy? :^)
L.
------------------------------
From: Mitchum.DSouza@mrc-apu.cam.ac.uk (Mitchum DSouza)
Subject: Re: Porting from SunOS to Linux (Dynamic Linking)
Date: 30 Sep 1994 10:28:46 GMT
In article <36f4ph$o8n@io.rdc.puc-rio.br>, draco@xvnews.unconfigured.domain
(DRACO) writes:
|>
|> Dear friends,
|>
|> I am trying to port a system from SunOS to Linux. The porting
|> has been a breeze except for the fact we dont have libdl available.
|> The system depends on dlopen and dlsym, since C++ code is generated
|> at run-time and then linked to the application.
|> Does anybody is working in the port of libdl???
|> Does anybody had already faced the same problem? If so, how it had been
|> treated?
|> Does anybody has an idea on how to solve the problem???
|>
Use the GNU dld library instead
tsx-11.mit.edu:/pub/linux/sources/libs/dld-3.2.4.src.tar.gz
Mitch
------------------------------
From: aris@mpd.co.za (Aris Stathakis)
Subject: Re: Multiprocessing Pentium Systems
Date: Fri, 30 Sep 1994 09:01:28 GMT
In <huw.157.2E832221@isgtec.com> huw@isgtec.com (Huw Leonard) writes:
>In article <CwJyxM.KxC@gcs.com> mark@gcs.com (Mark Bolzern) writes:
>>>I've just seen some new dual processor pentium systems in Computer
>>>Shopper. They look swell for the money, but there isn't a single OS
>>>that can take advantage of them. Anybody have any thoughts about how
>>>hard it might be to make Linux one of the first OS's to take advantage
>>>of these systems?
>>Not quite true... SCO MPX will work with many of them. But it would still
>>be good to have Linux available......
>WinNT works well on many multi-processor systems. IBM seems to be on the verge
>of releasing an SMP OS/2, as well. The problem, as I understand it, is that
>there is no firm multi-processor standard yet. The problem with working up an
>SMP or MPX version of Linux would be having to support all of the wierd
>variations of the architecture. Perhaps when manufacturers have a consistent
>standard, multi-processor Linux would be more feasable.
>Is this opinion incorrect? Comments?
There is the APIC stanard from Intel (now at version 1.1) which most of
these dual-processor boards are supporting (probably because they use
the Intel chip-set :-)
Aris
--
Aris Stathakis Tel: +27 11 887 1040 Snail Mail:
SCO ACE / Novell CNE Tel: +27 11 887 5150 P.O. Box 781228
M&PD (Pty) Ltd. Fax: +27 11 887-5158 Sandton, 2146
E-Mail: aris@mpd.co.za Cell:+27 83 601 0206 R.S.A.
------------------------------
** 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
******************************