Linux-Hardware Digest #755, Volume #10 Wed, 14 Jul 99 00:13:40 EDT
Contents:
Re: Success in enabling FirstMouse+ S48-OEM scroll wheel ("Gene Heskett")
Re: Bogus hard disk sizes from manufacturers ("Daniel Tonks")
Re: Curious about root passwd ("Tim Izod")
cs4242.o modprobe fails (Dave Smith)
Canon BJC-5000 ("Alan Gibson")
Re: Celeron, what's the catch? (Keith R. Williams)
Re: Celeron, what's the catch? (Michael)
Re: DSL and Linux ("TURBO1010")
Driver for Asuscom ISDNLink 128k PCI? (Kari Brown)
Re: SCSI adaptor for ZIP 250? ([EMAIL PROTECTED])
mouse prob with 2.2.7 (Dennis Perisa)
Linux compatible Tablet? (Habin)
THANK YOU EVERYONE! ("Éric Paré")
Re: Avermedia TV Phone (bt848) TV card not working...help! ("Gene Heskett")
Re: Celeron, what's the catch? (Colin Andrew Percival)
Re: Compact Flash vs. SSFDC Smart Media (Phil Askey)
Re: 2.2 SMP brief lockups (Myca-Boo)
Compact Flash vs. SSFDC Smart Media (Ken)
----------------------------------------------------------------------------
Date: 13 Jul 99 20:33:39 -0500
From: "Gene Heskett" <[EMAIL PROTECTED]>
Subject: Re: Success in enabling FirstMouse+ S48-OEM scroll wheel
Unrot13 this;
Reply to: <[EMAIL PROTECTED]>
Gene Heskett sends Greetings to Andrew J.;
AJN> -----BEGIN PGP SIGNED MESSAGE-----
AJN> This is to let the community know that as of 16:25 on July 13,
AJN> 1999 the zilog controler version of logitech's FirstMouse+ PS/2
AJN> (known as S48-OEM) has been conquered. The scroll wheel can
AJN> enabled by a series of set sampling rate commands directed at
AJN> /dev/psaux.
AJN> A utility has been written, but is not ready for public
AJN> consuption, which will enable the wheel so that it registers
AJN> using the standard MouseManPlusPS2 protocol under Xwindows. In
AJN> addition there is a diagnositic utility that will allow for
AJN> trapping and interpretation of mouse events.
AJN> I will release both of these shortly along with a mini-howto.
AJN> When things are ready I will notify the community with offical
AJN> release info.
Thanks.
AJN> Untill then I have a Bob Dylan/Paul Simon concert to
AJN> attend......
Please, you're making me homesick, those *were and are* the days...
Cheers, Gene
--
Gene Heskett, CET, UHK |Amiga A2k Zeus040 50 megs fast/2 megs chip
Ch. Eng. @ WDTV-5 |A2091,GuruRom,1g Seagate,CDROM,Multiface III
|Buddha + 4 gig WDC drive, 525 meg tape
|Stylus Pro, EnPrint, Picasso-II, 17" vga
RC5-Moo! 690kkeys/sec isn't much, but it all helps
email gene underscore heskett at iolinc dot net
--
------------------------------
From: "Daniel Tonks" <[EMAIL PROTECTED]>
Crossposted-To: comp.sys.ibm.pc.hardware.storage
Subject: Re: Bogus hard disk sizes from manufacturers
Date: Wed, 14 Jul 1999 00:40:55 GMT
Robert Leong wrote in message <[EMAIL PROTECTED]>...
>Ian Tester <[EMAIL PROTECTED]> wrote:
>
>How about the speed ratings for CD-ROM drives, it started with simple
>2X, then 4X and so on, but some where along the way someone decided to
>change the de facto standards of X, and offers an incredible 24X (or
>higher) speed. BUT, this X is not the same X as before.
Actually, "X" is the same (150kb/sec) but it's now MAX "X". An "up-to"
statement, such as "you could win up to $25,000,000". The drive can actually
get 24X at one position on the disc - but at the other end it may be 11X.
Some companies such as Plextor calculate the "average" sustained transfer
you can expect - their "17/40" drive averages out to 27X.
---
Daniel Tonks
Remote Central - http://www.remotecentral.com/
~ Featuring universal remote & DVD movie reviews ~
------------------------------
From: "Tim Izod" <[EMAIL PROTECTED]>
Subject: Re: Curious about root passwd
Date: Tue, 13 Jul 1999 19:03:01 +0100
Markus Wandel wrote in message <7mfqng$deg$[EMAIL PROTECTED]>...
>In article <7mbi40$aj1$[EMAIL PROTECTED]>,
>Davis Eric <[EMAIL PROTECTED]> wrote:
<snip>
>With a PC you can generally set up the BIOS and LILO that way. Your only
>way back in, should you forget the password, is to open the box and short
>the little jumper that makes it clear all its BIOS settings. And the cable
>locks that are used to discourage computers being taken away generally also
>inhibit access to the interior. Which brings you back to the guy with the
>bolt cutter.
>
>If I wanted to compromise a RedHat 5.2 system that has been secured in
>this fashion, what I would do is: Wait until it's accessing the disk a
>lot. Unplug it. Plug back in, watch it go through fsck on bootup and hope
>it gets messed up enough to dump me into a single-user shell to continue
>manually. My box did that once after a trip-over-the-power-cord event.
At University it wasn't unknown for savvy people to hit STOP-A as a
starting point for getting at the Ultra Sparcs running Solaris. Determined
people who know a bit will get in and all you can do is make it awkward.
</pessimist>
------------------------------
From: Dave Smith <[EMAIL PROTECTED]>
Subject: cs4242.o modprobe fails
Date: Wed, 14 Jul 1999 01:00:37 GMT
I need some help correcting this problem.
My soundcard had worked previously but seems to have given up the ghost.
I'm running RH6.0. The following is from /var/log/messages.
rc.sysinit: Finding module dependencies succeeded
modprobe: /lib/modules/2.2.5-15/misc/cs4232.o:
modprobe: init_module: Device or resource busy
rc.sysinit: Loading sound module failed
rc.sysinit: Loading midi module succeeded
This is a dual boot system running on a Dell Inspiron laptop.
The soundcard still works fine in Win95.
Could it be that the module is corrupt.
How would I replace a single module?
Thanks
------------------------------
From: "Alan Gibson" <[EMAIL PROTECTED]>
Subject: Canon BJC-5000
Date: Wed, 14 Jul 1999 11:04:03 +1000
Can anyone enlighten me as to how I can get Red Hat 5.2 to recognise my
Canon BJC-5000 colour printer as it is not listed in the installation
program, will I need a kernal update or a Red Hat upgrade?
Alan Gibson
[EMAIL PROTECTED]
------------------------------
From: [EMAIL PROTECTED] (Keith R. Williams)
Reply-To: [EMAIL PROTECTED]
Crossposted-To: comp.sys.ibm.pc.hardware.chips,comp.sys.intel
Subject: Re: Celeron, what's the catch?
Date: 14 Jul 1999 01:15:38 GMT
On Tue, 13 Jul 1999 16:51:30, Marc Mutz <[EMAIL PROTECTED]> wrote:
> Keith R. Williams wrote:
> >
> <snip>
> > > _Every_ read access is first checked against L1.
> >
> > Nope. Accesses to memory are broadcast and the first to "pick up"
> > the address as valid responds.
<spaces are nice>
> If that is right (and I'm not able to confirm nor tell you wrong) what
> will be the cache that 'wins'? L1! So practically, I'm right in my
> assertion that every read access is checked against L1. I withdraw the
> 'first' though. But that was not really my point...
That's not the point. There is no "search" or "check". If the
data of interest is in the L1 (or L2) it responds. There is no
loss of time if the Ln doesn't have the data.
> > > In 90% of the code
> > > accesses and - if you have a fine algorithm - also 80-90% of the data
> > > accesses the L1 hits. So the speed of the L1 is significant. Agreed?
> >
> > No. I don't agree. Not even close. Certainly the latency of the
> > L1 is important. The L1 is usually a line wide to the CPU and a
> > one cycle latency so the processor won't stall. Certainly nowhere
> > near 90% of the memory accesses are caught by the L1.
<Please leave spaces>
> Can you prove me wrong? I guess that modern CPU's do a prefetch of code
> cache lines so's to minimize the L1 fails. They calculate the target of
> a branch and load the code there speculatively, i.e. the cache line is
> filled even before the CPU needs to access that memory location, i.e.
> the L1 hits.
More importantly, can you prove me wrong? I state for a fact
that a Pentium class processor will *only* fetch (or store) in a
full cache line to cacheable memory. Cacheable memory is
subordinate to the L2, which is subordinate to the L1. P5/6
class processors don't do less than full cache line reads/writes
to cacheable memory.
Here's the gauntlet, prove me wrong. It only takes one example
for you to prove me wrong. I want to see how a byte can be
written to main memory without a full cache line written.
If I'm wrong, burst memory doesn't work. Show me one case! I'd
*really* love to see how to a byte write to main memory works
without a full burst.
>
> >
> > > But also the _size_ of the L1 is significant, because it is so much
> > > smaller than the L2. Going from 16K to 32K will bring you some kind of
> > > performance 'boost' (like 10%), going from 256K to 512K L2 will give you
> > > <2%.
> >
> > Nope. But I have to laugh. You say that 90% of the accesses are
> > caught by the L1 and you get a 10% boost by going from 16K to
> > 32K. That leaves 1% of the performance for the L2 and main
> > memory. Neat!
<spaces please>
> Well, if 16K of L1 catch 90% of accesses then 32K will just lift that
> to, say, 95%.
Proof please.
> The 10% were more meant as an upper bound. Don't take
> those numbers to be precise. If it was that 80% of all accesses hit L1,
> how would that change my argument (line-of-tought? what's the word?)?
What was your line of thought again? I see no substantiated
numbers, only wild-ass guesses.
>
> <snip>
> > > To sum up: L1 speed and size is 90% significant.
> > > L2 speed and size is ~10% significant.
> > > L3 speed and size is <~2% significant.
> >
> > We cannot agree on the premis of this argument, so we are surely
> > not going to agree on the conclusion. If you'd run your own
> > numbers you'd see that you're over 100%.
> You missed the tilde, indicating that the numbers be approximate.
>
> >
> > > This does not hold, if you have many CPU's in a UMA. In this case you
> > > can see significant speed-ups from larger L2/L3 caches, given every CPU
> > > has its own.
> >
> > UMA? Me thinks you have to learn the jargon. SMP is what you're
> > discussing (UMA is a graphics architecture). I wasn't touching
> > SMP. That's a different argument, but not unrelated.
> >
> I thought of UMA as a MP system where all (or large groups) of the CPUs
> share a common main memory (e.g. std. intel SMP), as opposed to systems
> where each CPU has it's own piece of RAM (e.g. transputers or DSP
> farms).
Yes, I was reminded of NUMA for such architectures (which implies
UMA - duh). On the "Intel" side UMA refers to the asinine
graphics architecture foisted by Intel. I do understand what
you're saying, but SMP is the popular term.
I do apologize for that slam. Sorry.
> I was fully aware of UMA being (also?) used for framebuffer
> integration into main memory. If you want to deny the question mark
> above, please tell me the 'jargon' acronym for what I have just
> described (maybe SMA?).
SMP does is the usual term in the "Intel" lexicon.
----
Keith
------------------------------
From: [EMAIL PROTECTED] (Michael)
Crossposted-To: comp.sys.ibm.pc.hardware.chips,comp.sys.intel
Subject: Re: Celeron, what's the catch?
Date: Wed, 14 Jul 1999 01:16:31 GMT
"FM" <[EMAIL PROTECTED]> wrote:
>How much L3 cache does your system have? I thought I read somewhere that
>K6-III allows up to 8MB of onboard L3 cache although it results in minimal
>performance increase. I can see your point that such a large cache would
>definitely smoothen the context-switching operation.
>
That's what's funny, its only 512K, about double the L2 cache.
>Hmm, btw, how much is this lag? I'm rather used to computers scratching the
>harddisk that I can't see how much this lag could be, since even RAM access
>should be rather instant from a user's point of view, unless heavy
>iteration's involved.
>
>Dan.
Its not much, at all, but you know the way it is. Once you get used
to a certain speed improvement, you go back and the other way seems
slow. The hard disk isn't touched at all (because when I noticed
this, I did fresh reboots to test). I have 128 Megs main memory. I
can of course force things to swap by opening a large photograph, and
then I can see the lag as the start menu goes to the swap file each
time. Its not that kind of lag. The real question I guess I have is
just how much code can fit into 128K, 256K, 512K, and 1 cache meg
increments after this, in a Windows 98 or NT environment. Obviously
certain data sets aren't getting loaded into cache like an entire
large.doc or .jpg image. But how much of the critical operating
system code fits into it as we increment cache? We all know that no
cache means meager performance, and that larger caches don't buy us
much via benchmarks. But exactly what is loading into this extra
space of cache? Under dos, the principle of locality was rather
simple. I am not certain how much "locality" we need anymore with
this bloated operating system. How much less waiting is there and on
what code is what I am after. I don't have the expertise to argue or
discuss this point, and some of the people here have raised a number
of issues, such as Marc (did I get it right this time) and Keith. I
can only do the Aristotle "let us observe" technique and throw it out
for discussion.
>
Mike
Mike
------------------------------
From: "TURBO1010" <[EMAIL PROTECTED]>
Subject: Re: DSL and Linux
Date: Tue, 13 Jul 1999 19:03:12 -0700
I have my linux box setup with a 3Com ISA etherlink III, and a $15 NE2000
card. The 3com card is going to the dsl modem, with a static ip and the
NE2000 to the hub, and out to the internal lan, 3 computers. Linux runs IP
masquerading, and is also my firewall.
Albert Goins <[EMAIL PROTECTED]> wrote in message
news:[EMAIL PROTECTED]...
> I am having DSL put in my place and have three computers to hook up. I
> was wondering if anyone can tell me what kind of hub and ethernet cards
> to buy to hook them all up. I am looking for affordability, and
> compatability with both Win98 and Linux. If you could tell me where to
> purchase them that would be great too. Thanks!
>
> -Al
>
------------------------------
From: Kari Brown <[EMAIL PROTECTED]>
Subject: Driver for Asuscom ISDNLink 128k PCI?
Date: Wed, 14 Jul 1999 03:24:15 +0300
Hi,
I was wondering if someone is working on a driver for the card on the
topic? The card uses Winbond 6692 chip whose specifications are
available from www.winbond.com. I really hope that someone could do
something about this because I can't :/
------------------------------
From: [EMAIL PROTECTED]
Subject: Re: SCSI adaptor for ZIP 250?
Date: Wed, 14 Jul 1999 01:43:51 GMT
In article <7mfp5v$[EMAIL PROTECTED]>,
[EMAIL PROTECTED] (Pat Palmer) wrote:
>
> I looked at Adaptec's web page, and it seems that the 2906 would be
> adequate. But, will that work with my version of Linux?
I got the 2906 to work with Red Hat 5.1. Use the aic7xxx driver.
Another SCSI interface that would probably work is a Tekram DC-310
if you can find one. It uses a popular NCR SCSI chip.
Sent via Deja.com http://www.deja.com/
Share what you know. Learn what you don't.
------------------------------
From: Dennis Perisa <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.x
Subject: mouse prob with 2.2.7
Date: Wed, 14 Jul 1999 11:47:06 +1000
I recently upgraded my Linux box to SuSE 6.1 and found that when I
recompiled my kernel (v2.2.7), my mouse failed to work in X and in
terminal mode. It works fine, however, with the original SuSE kernel
and with kernel version 2.2.6. I've compiled the kernels exactly the
same way so I can't figure out what the problem might be. If anyone can
tell me something, I would love to hear it!
Thanks.
Dennis.
------------------------------
From: Habin <[EMAIL PROTECTED]>
Subject: Linux compatible Tablet?
Date: Wed, 14 Jul 1999 01:45:00 GMT
Hi there,
Could you advice me a Linux compatible tablet?
The price is near $100 is good for me.
Blue Skies~
--
Software Engineer (VC++, Linux)
Skydiver (FAI certificate, total freefall time is
50m 45s)
Company URL is "http://www.mizi.co.kr"
Sent via Deja.com http://www.deja.com/
Share what you know. Learn what you don't.
------------------------------
From: "Éric Paré" <[EMAIL PROTECTED]>
Subject: THANK YOU EVERYONE!
Date: Wed, 14 Jul 1999 01:58:34 GMT
Thank you to all of you!
The solution to be able to boot is to use the "resc1440tecra.bin" disk.
Effectively, it was a cache memory problem because, I tryed to disable
the CPU cache and I was able to boot (but the system was fast like a
386!).
Now, the system is installed and I have to configure X Window, the mouse
and the network. I'll try by myself.
Thank you, your help is appreciated.
Eric Pare B.Eng (electrical)
Montreal (Quebec) CANADA
[EMAIL PROTECTED] wrote:
> Hello Debian 2.1 users,
>
> I'm trying to install GNU/Debian 2.1 on a laptop (Toshiba Satellite
> 2545XCDT) and the system reset (sometimes freeze) during the load of
> linux.
>
> Desciption of the system :
> The detailed specifications are in attachement (html format).
>
> I'm booting from the install CDROM disk #1 or from the "rescue floppy"
> and I always have the same problem. I already installed Debian, from
> this disk, on a 486 desktop and it was a success. Also, I tryed to
> install RedHat 5.0 and I did not have any problem except for
> configuring
>
> X.
>
> Following, is what I can see on the screen during boot of Debian 2.1 :
>
> Loading root.bin..............
> Loadind linux..................
>
> After "loading linux...", the system reboot or just freeze.
>
> Could you help me to find the way to install Debian 2.1 on my laptop?
>
> Thank you.
>
> Note : You can Reply in french or in english.
>
>
> Eric Pare B.Eng (electrical)
> Montreal (Quebec) CANADA
>
>
------------------------------
Date: 13 Jul 99 20:38:55 -0500
From: "Gene Heskett" <[EMAIL PROTECTED]>
Subject: Re: Avermedia TV Phone (bt848) TV card not working...help!
Unrot13 this;
Reply to: <[EMAIL PROTECTED]>
Gene Heskett sends Greetings to Ken Overgard;
Ken; The newsgroups are text based, and most of the worlds readers show
html encoded stuff *with* all the html. I can't find your message in
all the html trash. Please use text ONLY for newsgroup postings and
replies, with a line length of about 70 characters, and autowrap turned
on.
KO> <!doctype html public "-//w3c//dtd html 4.0 transitional//en">
KO> <html>
Cheers, Gene
--
Gene Heskett, CET, UHK |Amiga A2k Zeus040 50 megs fast/2 megs chip
Ch. Eng. @ WDTV-5 |A2091,GuruRom,1g Seagate,CDROM,Multiface III
|Buddha + 4 gig WDC drive, 525 meg tape
|Stylus Pro, EnPrint, Picasso-II, 17" vga
RC5-Moo! 690kkeys/sec isn't much, but it all helps
email gene underscore heskett at iolinc dot net
--
------------------------------
From: [EMAIL PROTECTED] (Colin Andrew Percival)
Crossposted-To: comp.sys.ibm.pc.hardware.chips,comp.sys.intel
Subject: Re: Celeron, what's the catch?
Date: 14 Jul 1999 02:57:09 GMT
[EMAIL PROTECTED] wrote:
: If the data is not available on the cache of one processor, does the
: first processor look for it on the cache of the second processor?
: You surely know a lot about caches do you?
I wasn't going to wade into this flamewar, but actually, yes, the first
processor _does_ look for data in the 2nd processor's cache.
The first processor (after possibly asserting BR0 to ask for the bus)
puts a memory read request onto the processor-chipset (frontside) bus
indicating that it wants to read a certain location in memory.
The chipset sees this and starts working on it.
Simultaneously, the 2nd processor's cache looks at the request, and
checks to see if that data is inside it.
If the data is not in the 2nd processor's cache, all proceeds as in a
single processor machine, and data is read from memory via the chipset.
However, if it is in the 2nd processor's cache, the 2nd processor
asserts #BOFF (backoff) to tell the chipset to release the bus, and
provides the data itself.
Colin Percival
PS. I've simplified somewhat in my description (the P6 bus is
transaction-oriented) but the above is essentially correct.
------------------------------
From: [EMAIL PROTECTED] (Phil Askey)
Crossposted-To: rec.photo.digital,comp.arch.embedded
Subject: Re: Compact Flash vs. SSFDC Smart Media
Date: Wed, 14 Jul 1999 03:19:55 GMT
Ken <[EMAIL PROTECTED]> wrote:
>formats? I'm designing a new gadget that will need Flash memory and I
>might go with a card slot instead of soldered-down chips and serial
>downloading. I figure anything that can read a PCMCIA hard disk will
>read either format, but just want to confirm that.
>From what I understand CF has a controller chip inside it which makes
it completely PCMCIA compatible, the PCMCIA->CF adapters simply acting
as a "shell" for the CF, the pins connectivity being the same. This
means that no matter what the capacity the device is simply accessing
an ATAPI "flash device", in theory any size card (assuming it fits -
CF Type I and Type II are slightly different package sizes) can be
used in any device which takes CF.
SmartMedia on the otherhand isn't so "smart" as it's purely a flash
chip, the controller has to be in the device, so for example digital
cameras which take smartmedia have a special smartmedia controller
built into them. Problem with that is that that controller typically
only supports smartmedia upto a certain capacity and the camera itself
has to be upgraded to take larger sizes. From what I understand this
is also a problem with the SmartMedia Clik! drive which won't take
32MB SmartMedia (I think).
--
Phil Askey <[EMAIL PROTECTED]>
http://photo.askey.net/
Digital cameras, news, reviews, photography tips, forums,
60 camera comparison database, 850+ photos in the gallery.
(remove NoSPAM to email)
------------------------------
From: [EMAIL PROTECTED] (Myca-Boo)
Subject: Re: 2.2 SMP brief lockups
Date: 14 Jul 1999 03:51:29 GMT
I am using a Netgear DEC Tulip based PCI NIC and I've not swapped it nor have I
adjusted irq's. Why do you suggest messing with the NIC? I remember hearing about
particular NIC's causing weird locks on one of the early 2.2's but didn't think it was
a tulip and thought that bug was squashed back in the early 2.2's. Perhaps I should
try swapping the NIC and changing irq's.
Thanks for the ideas!
Dave
In article <W0Bh3.355$[EMAIL PROTECTED]>,
Bryan <Bryan@[EMAIL PROTECTED]> writes:
> what ether card are you using? have you tried swapping it? maybe
> changing irq's?
>
> Myca-Boo <[EMAIL PROTECTED]> wrote:
>
>: Howdy folks,
>
>: I have a dual PII tyan thunder II that I run kernel 2.2.10 on. With SMP enabled,
>and both CPU's loaded (two setiathome clients) I can watch xosview freeze up briefly
>until a packet comes in from the network and interrupts it. I've seen it freeze up
>in the middle of a kernel compile and then unfreeze as soon as I ping it. Has anyone
>had this type of problem? I've experienced it from 2.2.2 on up and I'm not sure if
>it's a hardware problem or not. this same machine did not exhibit this problem with
>NT
> (previous operator).
>
>: thanks for the help,
>
>: Dave
>
>: Email: [EMAIL PROTECTED]
>: --
>: Linux, the choice of a GNU generation
>
--
Linux, the choice of a GNU generation
------------------------------
Date: Tue, 13 Jul 1999 19:57:57 -0700
From: Ken <[EMAIL PROTECTED]>
Subject: Compact Flash vs. SSFDC Smart Media
Crossposted-To: rec.photo.digital,comp.arch.embedded
Ok, getting a card-based camera sounds like the best option. Now I'm
wondering whether I should target CF or SSFDC formats? I looked at the
SSFDC website (http://www.ssfdc.or.jp) and it implies that SSFDC is only
available up to 8mb. Is the site out of date? I'm seeing reference to 48
mb and 64 mb CF cards in rec.photo.digital, so CF looks like it has the
lead in capacity. How about market penetration? Is CF more popular?
(CF's website is http://www.compactflash.org.)
BTW, anybody have experience designing equipment that uses these
formats? I'm designing a new gadget that will need Flash memory and I
might go with a card slot instead of soldered-down chips and serial
downloading. I figure anything that can read a PCMCIA hard disk will
read either format, but just want to confirm that.
Albert Hall wrote:
>
> If you have a camera that uses compact flash and your system supports PCMIA
> as a drive, all you have to do is plug in the memory card and viola! a drive
> containing your pictures. Smartmedia has an adapter to fit in a floppy drive.
> I don't know if you had a Mac with a floppy (not made any more) that it would
> read it as it would be a Wintel format. Compact flash card readers are
> available in a USB flavor so a Mac wouldn't be left out if you went that way.
>
> Al
--
Ken
mailto:[EMAIL PROTECTED]
http://www.sewingwitch.com/ken/
http://www.215Now.com/
------------------------------
** 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: [EMAIL PROTECTED]
You can send mail to the entire list (and comp.os.linux.hardware) via:
Internet: [EMAIL PROTECTED]
Linux may be obtained via one of these FTP sites:
ftp.funet.fi pub/Linux
tsx-11.mit.edu pub/linux
sunsite.unc.edu pub/Linux
End of Linux-Hardware Digest
******************************
