Files
oldlinux-files/ftp-archives/tsx-11.mit.edu/1996-10-07/mail-archive/linux-devel/Volume2/digest234
2024-02-19 00:24:15 -05:00

692 lines
25 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: Tue, 27 Sep 94 13:13:15 EDT
Subject: Linux-Development Digest #234
Linux-Development Digest #234, Volume #2 Tue, 27 Sep 94 13:13:15 EDT
Contents:
Re: BSD/386 vs. Linux Performance (Grant Edwards)
Re: The ELF/PIC performance measurement. (Steven Buytaert)
Re: Scanning with Linux? (Mary Shenk)
Re: ELF-based Linux distribution? [Was: Shared Libs: working toward a permanent solution?] (Walter Wolfgang)
Re: Don't use Linux?! (Michael Dillon)
Re: 1+ Gig SCSI Drives // (Matthew S. Crocker)
Re: elm2.4 (Oliver Duesel)
Linux version of crypt(1) (David A. Vohwinkel)
Re: BSD/386 vs. Linux Performance (Martin Spott)
Re: Driver support for PS/2 (MCA) version of SMC/WD? (Donald Becker)
mmap and MAP_SHARED? (Brad Pepers)
Re: Does Quicken work under DOSEMU? (Dag Asheim)
Help: ftape 1.13b with kernel 1.1.49 (Andreas Weigand)
IBM Token Ring Support? (Clarence Chu)
Re: Is clock() broken in gcc 2.5.8? (Andreas Schwab)
Re: Digi Intelligent Boards? (Rusty L. Atkins)
Re: PROMISE DC4030VL-2 IDE Controller (David Boyd)
Re: The ELF/PIC performance measurement. (Steven Buytaert)
Re: [STATUS] Linus Floppy Driver Development (Mr D R Barlow)
multiple cdu31a drives (Robert Van.Zant)
Re: [STATUS] Linus Floppy Driver Development (Mr D R Barlow)
----------------------------------------------------------------------------
From: grante@reddwarf.rosemount.com (Grant Edwards)
Subject: Re: BSD/386 vs. Linux Performance
Date: Tue, 27 Sep 1994 07:27:17 GMT
Michael Nelson (mikenel@netcom.com) wrote:
: Anyone have any performance stats comparing BSD/386 (BSD/OS) with Linux?
: Network, disk, and overall...
DUCK!
--
Grant Edwards |Yow! Jesus is my POSTMASTER
Rosemount Inc. |GENERAL..
|
grante@rosemount.com |
------------------------------
From: buytaert@imec.be (Steven Buytaert)
Subject: Re: The ELF/PIC performance measurement.
Date: Tue, 27 Sep 1994 12:52:20 GMT
H.J. Lu (hjl@nynexst.com) wrote:
: I did an ELF/PIC performance measurement with gcc on a 486DX2/66 with
: 16MB RAM and no external cache.
[ measurement deleted ]
Sorry for asking, but I'm getting totaly clueless. What binary
format is used on Linux for binaries and objects ? Is it
ELF or a.out ? I know it isn't COFF since that is the format
I know best and I'm always out of luck :-/
I'm experimenting with the BFD library with format a.out-i386-linux
but with no success. Therefore my stupid question, where is the
ELF reference in the article comming from ?
Please give a short answer and then tell me to subscribe to the
GCC channel :-)
Stef
--
Steven Buytaert
WORK buytaert@imec.be
HOME buytaert@innet.be
'Imagination is more important than knowledge.'
(A. Einstein)
------------------------------
From: mkshenk@u.washington.edu (Mary Shenk)
Crossposted-To: comp.os.linux.misc
Subject: Re: Scanning with Linux?
Date: 27 Sep 1994 07:52:45 GMT
In article <366ord$kka@solaria.cc.gatech.edu>,
Byron A Jeff <byron@gemini.cc.gatech.edu> wrote:
>In article <NELSON.94Sep21212413@crynwr.crynwr.com>,
>Russell Nelson <nelson@crynwr.crynwr.com> wrote:
>>In article <35k0c6$qib@virgo.cc.gatech.edu> byron@cc.gatech.edu (Byron A Jeff) writes:
<stuff>
>In all honesty all we really need is the interface specifications for the
>interface card and the scanner. But getting the representatives I've talked
Or just the scanner, if we're talking SCSI. (proprietary interface cards
for things that should really be SCSI--yuk!)
>
>My plan (which is stated in the Info-Sheet) is to check out EPSON because
>as far as I know they are the only flatbed scanner manufacturer that actually
>releases the hardware/interface specs for their interface boards.
Well, I'm going to have a go at writing a driver for the UMAX 840, a SCSI
flatbed. (Because I happen to have one, and would rather let it sit than
even boot winblows.) From what I hear, the UMAX folks should be good about
giving out the info...I will be talking to them sometime in the next few
days. This will be my first driver, but I fully intend to do it, as it
doesn't seem my scanner is going to be supported unless I do.
Anyway, being a newbie at this driver writing stuff, I'd like input on
any sort of hooks etc that should be in this thing for stuff folks might
want to add later. I know nothing about TWAIN, but if it's something I
should even be worrying about, I'd appreciate some pointers in that
direction as well... As long as I'm writing this thing, it might as well
be of use to as many others as possible, so be sure and speak up if there
anything I may easily do to make it more general, etc... I am wondering
what degree of similarity there is between SCSI flatbeds from different
mfrs.. Is it likely that once a driver for one SCSI scanner is written,
others will be a good deal easier?
Anyway, if this is the sort of thing you are interested in seeing,
please bombard me with any useful information you may have.
-Craig
(posting from wife's account.)
------------------------------
From: walterw@Informatik.TU-Muenchen.DE (Walter Wolfgang)
Subject: Re: ELF-based Linux distribution? [Was: Shared Libs: working toward a permanent solution?]
Date: 27 Sep 1994 12:54:36 GMT
If the shared lib type will be changed, is it possible to have a sort of
interface from the old one to the new one? I mean i.e a libc.so.... which seems to
be one of the old type, but will use the new libc.so..., mainly a "jump-table".
So the libs will not be in memory twice if programs linked for the old ones
are used.
Wolfgang
------------------------------
From: mpdillon@halcyon.com (Michael Dillon)
Subject: Re: Don't use Linux?!
Date: Sat, 24 Sep 1994 00:27:18 +0100
> WorkGroup Solutions and Multisoft have taken the plunge by porting FlagShip
> which is the Unix version of the CA/Clipper language from the DOS world.
>
> There are tons of commercial software packages written in Clipper (Possibly
> as many as all other languages put together), and all of these are now
> a simple compile away from Linux.
>
> There is a huge aftermarket for Clipper written products and add ons. Clipper
> is powerful enough to emulate Fox & Dbase ... the premise behind our FoxKit..
> FlagShip may be the key to bringin Linux to 200+ K Clipper programmers,
> Millions of users of Clipper written programs, and about 20 million Dbase &
> Foxers.....
>
> FlagShip sells for thousands of dollars on "Commercial" Unix systems, and we
> have thousands of installations.
>
> We are gambling that Linuxers want DOS pricing levels, and offer our
> Linux product inexpensively enough to qualify for this designation.
>
> We are also working at convincing other commercial vendors to take the
> plunge............
>
> What is missing? THE SUPPORT FROM THE LINUX COMMUNITY
>
> You want your OS to be to MicroSoft, as MicroSoft was to IBM (It's undoing)?
> The democratic solution to a potential Gates Dictatorship?
This guy is right. I have seen the mess that occurs when you try
to implement real-world business programs on a GUI platform. In the
real world, business runs on numbers and text and SHARED databases.
The ideal computer system to do real-world business applications
provides only what is needed and nothing more. This makes a Linux
box with a dozen terminals and Flagship's dBase software, the
IDEAL business computer system, especially when you look at
the needs of small and medium sized business.
There is real money to be made here by people who are willing to
design and build computer systems (hardware and software) to meet
the needs of your local business community.
While the rest of the PC world is trying to sell a Novell network
with 6 486's, you can walk in and sell 1 486, 5 terminals and
Linux. Your solution will be cheaper and will actually run FASTER
than the solution with 6 CPU's. That's because you don't have
any network overhead since you only transmit screens full of
information to the terminals rather than database chunks.
There is good money to be made in designing and writing business
programs, advising what hardware/software to buy and babysitting
the whole system to make it work. I often find that I have to
redo hardware installations that the local computer retailler
has screwed up because they all seem clueless. Things like
incorrect IRQ settings on add-in cards like SCSI host adapters
or Digiboard PC/8e's.
cruisin' down the information highway, lookin' for a blast
breakin' all the speed limits as I come zoomin' past!
--
Michael Dillon Internet: mpdillon@halcyon.halcyon.com
C-4 Powerhouse Fidonet: 1:353/350
RR #2 Armstrong, BC V0E 1B0 Voice: +1-604-546-8022
Canada BBS: +1-604-546-2705
------------------------------
From: matthew@crocker.com (Matthew S. Crocker)
Subject: Re: 1+ Gig SCSI Drives //
Date: 27 Sep 1994 12:57:36 GMT
I have a Seagate 2.1GB baraccuda drive installed off an AHA1542CF
formated as one big partition under ext2. This drive is my
/usr/spool/news drive and is about 1.5GB with news. working just fine
and dandy... Linux = 1.1.51, After formatting I get @2.0GB usable
space.
-Matt
--
-Matthew S Crocker "The mask, given time, comes
mcrocker@crocker.com to be the face itself." -anonymous
*OS/2*OS/2*OS/2*OS/2*OS/2*OS/2*OS/2*OS/2*OS/2*OS/2*OS/2*OS/2*OS/2*OS/2*OS/2*
*linux*linux*linux*linux*linux*linux*linux*linux*linux*linux*linux*linux*
------------------------------
From: oli@wnbbs.nbg.sub.org (Oliver Duesel)
Subject: Re: elm2.4
Date: Sun, 25 Sep 1994 19:39:49 GMT
hi there,
wolle@anguish.ancient.trillium.se (Wolfgang Feldmann) writes:
>I get this nice error when I try to compile elm2.4:
>opt_utils.c:90: argument `size' doesn't match prototype
>Any ideas please?
:
#ifndef GETHOSTNAME
gethostname(cur_hostname,size) /* get name of current host */
char *cur_hostname;
int size; <-- change this to size_t size;
:
should do it ...
By(t)e, Oli.
--
Oliver Duesel email: oli@wnbbs.nbg.sub.org
Fido : Oliver Duesel, 2:2490/2081
snail: Oliver Duesel, P.O.Box 210 512, 90123 Nuernberg
------------------------------
Crossposted-To: comp.os.linux.misc
From: vohwi-d@acsu.buffalo.edu (David A. Vohwinkel)
Subject: Linux version of crypt(1)
Date: Tue, 27 Sep 1994 12:56:24 GMT
I have a file that has been encrypted with crypt(1), is there anyway under
linux that I can decrypt it?
Thanks
-Dave-
--
David A Vohwinkel
Unix Consulting ^ ^ vohwi-d@acsu.buffalo.edu
& Operations 0 0 @ The State University of New York at Buffalo
==============oOO=(_)=OOo====================================================
------------------------------
From: sk001sp@unidui.uni-duisburg.de (Martin Spott)
Subject: Re: BSD/386 vs. Linux Performance
Date: 25 Sep 1994 11:31:59 GMT
Michael Nelson (mikenel@netcom.com) wrote:
: Anyone have any performance stats comparing BSD/386 (BSD/OS) with Linux?
: Network, disk, and overall...
--> comp.os.amusement !!!
--
EMail: bevorzugt privat: Martin@smigel.mitropa.com
notfalls auch "dienstlich": Martin.Spott@uni-duisburg.de
------------------------------
From: becker@cesdis.gsfc.nasa.gov (Donald Becker)
Crossposted-To: comp.os.linux.misc
Subject: Re: Driver support for PS/2 (MCA) version of SMC/WD?
Date: 27 Sep 1994 10:33:48 -0400
In article <jfmorrisCwruMH.4rq@netcom.com>,
James F. Morris <jfmorris@netcom.com> wrote:
>I have been working on getting Linux up and running on an IBM PS/2 Model
>70, which has not been an enviable task, to say the least! But, beggars
>can't be choosers, either.
...
>So, to make a long story short, I was able to scrounge an IBM PS/2 Ethernet
>Adapter /A. This is a microchannel version of an SMC Ultra, from what I
>have been able to tell. Under DOS, the same SMC_WD.COM packet driver is used
>for this card, as for an ISA based SMC Ultra or WD8013. Biggest difference
>is that the I/O port address range is up at 0x800, rather than 0x300 - no
>big deal, and the RAM address is fixed in the PS/2 setup program.
The WD80*3 and SMC Ultra use different methods of enabling memory.
Do you know which one this card is similar to?
>Problem is, even with the 0x800 I/O port range added to the scan list in the
>WD.C and SMC-ULTRA.C kernel driver, the card just doesn't work right (under
>Linux - works fine under DOS). I get TX status 3 (timeout on TX) when
>attempting to access the network.
Please include the exact message -- is that the Tx status register or the
general status register?
>Looking through the Crynwr packet drivers shows that there are differences
>between operation of the NS (8390?) chip used on these boards on the ISA
>vs. MCA bus. Specifically, the MCA bus has twice as fast a clock cycle as
>the ISA bus. The Crynwr driver inserts extra delays if it detects operation
>on a Microchannel machine. Other differences are some extra setup in
>certain registers on the NIC's ASIC.
Those delays shouldn't be necessary with the UltraChip. They were for
earlier chips that needed a recovery period between accesses. That was a
bug, not a feature.
BTW, which driver are you trying to use, the SMC Ultra or WD80*3?
Is the ethernet address correctly detected?
Are you receiving any packets at all? Even errors? Check /proc/net/dev.
What addresses can the board end up at?
--
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: pepersb@cuug.ab.ca (Brad Pepers)
Subject: mmap and MAP_SHARED?
Date: Tue, 27 Sep 1994 04:38:26 GMT
I am working on a database and was intending to mmap the file as
PROT_READ | PROT_WRITE and MAP_FILE | MAP_SHARED. When experimenting
with this I found out that the Linux mmap doesn't allow PROT_WRITE
with MAP_SHARED. Can anyone tell me if this is planned to be added
soon? Any pointers on how to make it work? I just checked 1.1.51 and
it still seemed to not allow it.
Any help on this would be appreciated!
+----------------------------Ren & Stimpy--------------------------------+
| "Psst. Hey Guido. It's all so clear to me now. I'm the keeper of the |
| cheese. And you're the lemon merchant. Get it? And he knows it. That's |
| why he's gonna kill us. So we gotta beat it. Yeah. Before he lets |
| loose the marmosets on us! Don't worry, little missy! I'll save you!" |
+------------------ Brad Pepers -- pepersb@cuug.ab.ca -------------------+
------------------------------
From: Dag Asheim <dash@ifi.uio.no>
Subject: Re: Does Quicken work under DOSEMU?
Date: 27 Sep 1994 13:10:56 +0100
Reply-To: dash
>Related Q: Does Quicken for Windows run under WINE?
Not yet.
By the way, this thread does not fit in the comp.os.linux.development
group. I put in a Followup-To: comp.os.linux.misc.
Dag
------------------------------
From: usera@wema10.elektro.uni-wuppertal.de (Andreas Weigand)
Subject: Help: ftape 1.13b with kernel 1.1.49
Date: 27 Sep 1994 14:20:58 GMT
Hello all !
Two weeks ago i send the following message, but i get no responce.
So I post it again and hope someone can answer me:
Can someone help me ? I want to use my Colorado streamer DJ-10 with
controller card FC-10 on my Linux system ! I get the patched ftape-driver
from sunsite and compiled it successfully. The modules were compiled same
and I'am using kernel version 1.1.49.
Now when I give the command "mt -f /dev/ftape reten " the tape go's from the
beginning to the end and return. That is it what the tape should do, but
when I want to read from the tape with:
mt -f /dev/ftape erase
or
mt -f /dev/ftape status
the tape moves and I get the message: I/O error !
Can someone help me about this problem ?
You can reach me via Internet at the adress:
usera@wema10.elektro.uni-wuppertal.de
or you can post an articel at this place !
Thank you !
Andreas Weigand
----
usera@wema10.elektro.uni-wuppertal.de
------------------------------
From: Clarence.Chu@f132.n700.z6.ftn.air.org (Clarence Chu)
Subject: IBM Token Ring Support?
Date: Tue, 27 Sep 1994 21:28:00 +0800
Hi netter,
I would like to know whether IBM Token Ring Adaptor is
being supported by Linux. I don't find it in FreeBSD, and
in hope that the device driver is in Linux.
Thanks for any information.
Clarence Chu
* Origin: (6:700/132)
------------------------------
From: schwab@ls5.informatik.uni-dortmund.de (Andreas Schwab)
Subject: Re: Is clock() broken in gcc 2.5.8?
Date: 27 Sep 1994 12:46:14 GMT
In article <368b9m$dh@tusol.cs.trinity.edu>, jmccaski@Leda.CS.Trinity.Edu (Scott McCaskill) writes:
|> Does anyone know of any problems with clock() in gcc 2.5.8?
|> I recently downloaded and installed the slackware release of linux
|> (about 3 or 4 weeks ago) and this function does not appear to work as
|> advertised. The code is extremely simple:
|> #include <time.h>
|> #include <iostream.h>
|> int main()
|> {
|> cout << "\n\nclock() = " << clock() << endl;
|> cin.get(); // wait a few seconds, then hit return
|> cout << "clock() = " << clock() << endl;
|> return 0;
|> }
|> CLK_TCK is reported as being 100, so 100 ticks per second. The output
|> from the above code always indicates that clock() is returning the same
|> value the second time it is called as the first time (usually 2 or 3),
|> even when I wait several seconds before pressing return.
|> Any ideas?
I'd expect this behaviour with every compiler. clock(3) reports the
*cpu* time, not real (wall clock) time. Of course, while waiting for
input your program does not need any cpu time. To get real time use
time(3), or gettimeofday(2) for better resolution.
--
Andreas Schwab "And now for something
schwab@issan.informatik.uni-dortmund.de completely different"
------------------------------
From: atkins@cc5037.pms.ford.com (Rusty L. Atkins)
Subject: Re: Digi Intelligent Boards?
Date: 26 Sep 1994 19:33:55 GMT
In article <troyd.779983677@digibd> troyd@digibd.digibd.com (Troy DeJongh) writes:
: >Well, I'm responsible for the maintenance/development of our SCO serial
: >drivers and the development of the Linux driver for our PC/Xe product,
: >and I haven't heard anything like that yet. Plans are to release a
: >Linux driver for our PC/Xe line of products this fall.
Yes! Yes! Cool! Kicka*s! Thank-you! I've had one of these things sitting on
my desk collecting dust ever since I switched to linux about 6 months ago.
--
================================================================================
Rusty Atkins atkins@mail.msen.com
A not-so-proud citizen of the USSA... Where only crimimals can have guns.
________________________________________________________________________________
Don't Tread On Me.
------------------------------
From: dwb@ITD.Sterling.COM (David Boyd)
Crossposted-To: comp.os.linux,comp.os.linux.help,comp.os.linux.misc
Subject: Re: PROMISE DC4030VL-2 IDE Controller
Date: 26 Sep 1994 19:43:25 GMT
I have already replied directly to the originator of this thread concerning
this but will follow up this response to the net for completeness.
In article <366vjj$k9a@csugrad.cs.vt.edu>,
Christian Nelson <cnelson@csugrad.cs.vt.edu> wrote:
>I also have one of these controllers, though I only use two drives
>with it... so I'm not going to be much help. I'm almost positive
>you'll need to acquire the patch that allows you to use two IDE
>controllers simutaneously. The promise card, as far as using a 3d and
>4th drive is concerned, acts like two controllers.
>
NO the Promise does not, as reported to me by Promise. According to
the promise people all 3&4 drive support is done through the primary (i.e
standard) controller IO ports and addresses. I have the source to
Promise drivers for ATT & SCO under non-disclosure. Unfortunately I
have not had the time to do much more than experiment and those
experiments have not been successful. I know that they read the drive
configurations from the controller nvram but have not successfully be
able to get that to work under Linux.
>
>Have you have any trouble with your controllers, ie: system locking-up
>when doing intensive (disk intensive) activities? Mine does, but ONLY
>when I have it set on defered write. Let me know if you also have
>this problem. I think it might me related to my drive
>configuration... That Promise controller doesn't like WD drives in a
>two drive configuration.
I have two WDs on the primary port of my promise and get occasional
lockups also. Mine do not appear to be related to heavy disk activity
however. I would almost suspect a timing problem with defered writes
(which is more likely to happen under heavy load).
Another thing about the Promise controller is that it does not
support the multisector (hdparm) transfers. I am ready to sell mine
and just get a generic VL bus controller.
--
David W. Boyd UUCP: uunet!sparky!dwb
Sterling Software ITD INTERNET: Dave_Boyd@Sterling.COM
1404 Ft. Crook Rd. South Phone: (402) 291-8300
Bellevue, NE. 68005-2969 FAX: (402) 291-4362
I survived - Seoul Sea of Fire Tour 94
------------------------------
From: buytaert@imec.be (Steven Buytaert)
Subject: Re: The ELF/PIC performance measurement.
Date: Tue, 27 Sep 1994 13:07:59 GMT
Steven Buytaert (buytaert@imec.be) wrote:
: H.J. Lu (hjl@nynexst.com) wrote:
: : I did an ELF/PIC performance measurement with gcc on a 486DX2/66 with
: : 16MB RAM and no external cache.
: [ measurement deleted ]
: Sorry for asking, but I'm getting totaly clueless. What binary
: format is used on Linux for binaries and objects ? Is it
: ELF or a.out ? I know it isn't COFF since that is the format
: I know best and I'm always out of luck :-/
Yup, I put it in there again, all the way :-)
Sorry, I was away for 10 days and am totally out of sync with news.
Yes, currently it's a.out, trend is to move over to ELF.
You can still mail me to subscribe to the GCC channel though...
--
Steven Buytaert
WORK buytaert@imec.be
HOME buytaert@innet.be
'Imagination is more important than knowledge.'
(A. Einstein)
------------------------------
From: xuuah@csv.warwick.ac.uk (Mr D R Barlow)
Subject: Re: [STATUS] Linus Floppy Driver Development
Date: 27 Sep 1994 16:10:53 +0100
In article <35tg47$28i@ccnet.ccnet.com>,
kender@esu.edu writes:
>What if they did something like on the sgi's, where you can specify
>a mount point for dos disks, and run a daemon that will mount the floppy
>automagically (at least, for a dos disk). Pretty neat though.
I've used something like this system on Solaris 2, and ended up in the
ludicrous state of being unable to format disks, as every time I put
them in they were mounted, and then fdformat would give 'device busy'...
I'm sure there's a way round this but do bear it in mind :-)
>Perhaps
>if we had something like that for msdos disks only, all others needing
>to be mounted manually (therefore, by root)?
There's no reason to do it for dos disks only -- see other threads on
the 'user' option to mount.
Daniel.
------------------------------
From: rvanzant@solaria.mil.wi.us (Robert Van.Zant)
Subject: multiple cdu31a drives
Date: 27 Sep 1994 15:41:42 GMT
Thought I'd ask here before I bite off more than I care to chew.
Has anyone tried to dress up the cdu31a driver to allow more than 1
drive per linux box?
thanks,
rvz
------------------------------
From: xuuah@csv.warwick.ac.uk (Mr D R Barlow)
Subject: Re: [STATUS] Linus Floppy Driver Development
Date: 27 Sep 1994 16:12:30 +0100
In article <368ssc$glc@deneb.tm.informatik.uni-frankfurt.de>,
Anselm Lingnau <lingnau@tm.informatik.uni-frankfurt.de> writes:
>How are you going to umount the disk if it's already been ejected? It seems
>to me that this would rule out anything but read-only mounts.
Or synchronous mounts. I know sync means a big performance hit, but
floppies are so slow anyway I doubt anyone'd notice too much.
Daniel.
------------------------------
** 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
******************************