Linux-Hardware Digest #2, Volume #10 Mon, 12 Apr 99 07:13:22 EDT
Contents:
Re: DMA errors (Nick Lucent)
Re: Laser printer, which one will in Linux (Luca Filipozzi)
Re: How fast is your HD? (Jon Spreha)
Re: [Fwd: Source Code To Windows 98 (programmer humor) (fwd)] ("Martijn Lievaart")
Re: mounting hard drives (Lloyd Weehuizen)
Re: DRIVERS AGFA SCANNERS / CREATIVE WEBCAM / IOMEGA BUZZ (Frederic Dambreville)
pcmcia on desktop pc for linux (Pietro Marchionni)
ATTENTION UMAX and SANE (UNIX) developers: Please make a compatible network
scanning protocol! ("Scan")
Which driver for NCR 53C810 SCSI ? (**Nick Brown)
Re: Building a Better Distribution. (WAS: X munges the graphics card? (Re: Windows
2000 Rah! Rah! Session falls flat)) (Todd Ostermeier)
Re: Linux Networking Performance Question? (Marco BANO)
Re: [Fwd: Source Code To Windows 98 (programmer humor) (fwd)] (Michael Rubenstein)
----------------------------------------------------------------------------
From: [EMAIL PROTECTED] (Nick Lucent)
Crossposted-To:
alt.os.linux.caldera,alt.linux,alt.linux.sux,alt.os.linux,comp.os.linux.networking
Subject: Re: DMA errors
Date: 11 Apr 1999 23:19:04 GMT
Reply-To: [EMAIL PROTECTED]
On Fri, 09 Apr 1999 09:42:32 GMT, Wiley Hill <[EMAIL PROTECTED]> wrote:
<>On Tue, 30 Mar 1999 14:49:12 -0800, "TURBO1010"
<><[EMAIL PROTECTED]> wrote:
<>
<>>Why, after updating the bios on my Tyan 1564/D do I get DMA errors on all the hard
drives, and tells me DMA not available.
<>> It tells me bad CRC. What does it mean? It used to work fine with version 1.01,
now with 1.02 it gives me this error.
That board has an fx chipset, Im almost positive that fx doesnt support
udma, I got errors when I enabled some udma option in the bios, when I
disabled it the errors went away, Im using a wd 6.4 gb drive and I know its
udma, but disabling that option got rid of the errors which is the important
thing. Try it and see if it works.
Nick
<>
<>>>
<>>>Don't know. Did you save the old bios? I did have a simular problem though,
<>>>with a Tyan board. My hardisk was on their list (Tyan's) as being having
<>>>buggy firmware and so I could not use DMA with the CD-Rom that was
<>>>Slaving off of it. I believe that the real problem was that the CD-Rom
<>>>wasn't UDMA and the HD was. Later I added a CD, Moved the
<>>> old cd to the second interface as a master. I didn't have anymore
<>>> problems.
------------------------------
From: [EMAIL PROTECTED] (Luca Filipozzi)
Crossposted-To: comp.os.linux.misc,linux.debian.user
Subject: Re: Laser printer, which one will in Linux
Date: Mon, 12 Apr 1999 01:22:48 -0700
In article <7es4ke$v3o$[EMAIL PROTECTED]>, [EMAIL PROTECTED]
says...
> In comp.os.linux.misc henk van der knaap <[EMAIL PROTECTED]> wrote:
> : Dear Readers,
> : I intend to buy a laser printer. The following brands are available here:
> : Hewlett Packard, Canon, Oki, Brother (These are the main names)
> : Can anybody, please, advise me which one works particularly well under
> : Linux?
>
> If you have the money, get one that understands Postscript. That saves A
> LOT of trouble! NEC and Lexmark (not on your list, unfortunetly) offer
> small laser printers for about $300-$350US that have optional Postscript
> modules that cost an additional ~$100. Otherwise, most laser printers
> that understand Postscript cost at least $1200.
>
>
I strongly agree. Postscript!
--
Luca Filipozzi <[EMAIL PROTECTED]>
------------------------------
From: Jon Spreha <[EMAIL PROTECTED]>
Subject: Re: How fast is your HD?
Date: Sun, 11 Apr 1999 19:21:49 -0400
Interesting test... Here's my results.
This is my Multia machine:
alpha:~# hdparm -tT /dev/sda1
/dev/sda1:
Timing buffer-cache reads: 64 MB in 7.22 seconds = 8.87 MB/sec
Timing buffered disk reads: 32 MB in 18.73 seconds = 1.71 MB/sec
This is my 486 router machine:
slow:~# hdparm -tT /dev/hd{a1,b1}
/dev/hda1:
Timing buffer-cache reads: 64 MB in 7.35 seconds = 8.71 MB/sec
Timing buffered disk reads: 32 MB in 14.98 seconds = 2.14 MB/sec
/dev/hdb1:
Timing buffer-cache reads: 64 MB in 8.94 seconds = 7.16 MB/sec
Timing buffered disk reads: 32 MB in 14.19 seconds = 2.26 MB/sec
Are these normal?
Jon
---
[EMAIL PROTECTED]
PGP Public Key available on the servers.
------------------------------
From: "Martijn Lievaart" <[EMAIL PROTECTED]>
Crossposted-To:
comp.lang.c,comp.os.linux.setup,comp.os.linux,comp.os.linux.x,alt.os.linux,comp.lang.c++
Subject: Re: [Fwd: Source Code To Windows 98 (programmer humor) (fwd)]
Date: 12 Apr 1999 08:34:58 GMT
James Goldman wrote in message
<[EMAIL PROTECTED]>...
>Olaf Walkowiak wrote:
>> Useful for things like
>>
>> test || echo "Something went wrong";
>> test && echo "Cool, it worked!";
>
>Ah.. I see. The UNIX command line is pretty good, eh?
>
As this was also about MSDos originally, let me point out that there are
better shells than command.com, that do allow this syntax [1]. As there are
for NT and other OSses. Under Unix I also never work in sh ;^>. I find this
a more a case of braindead users than braindead OS. You wouldn't want to cut
your meat with a spoon now do you? But to me that is exactly what millions
of lusers are doing, because M$ says that is the way to do it.
HTH,
Martijn
[1] And filename completion, real history, a real script language, etc....
--
My reply-to address is intentionally set to /dev/null
reply to mlievaart at orion in nl
------------------------------
From: Lloyd Weehuizen <[EMAIL PROTECTED]>
Crossposted-To:
alt.linux,alt.os.linux,alt.uu.comp.os.linux,alt.uu.comp.os.linux.questions,comp.os.linux,comp.os.linux.misc,comp.os.linux.setup,comp.os.linux.x
Subject: Re: mounting hard drives
Date: Mon, 12 Apr 1999 20:56:31 +1200
On Sun, 11 Apr 1999, Upali Bandara wrote:
>If you want to mount the first partition of "C:" try hda0 or hda1 (I'm
>not sure).
hda1, there is no such thing as hdX0, for some strange reason that on of the
exceptions to the rule, Always start on 0. :)
Maybe somebody else has and idea to why this is???
cya
Lloyd
------------------------------
From: Frederic Dambreville <[EMAIL PROTECTED]>
Subject: Re: DRIVERS AGFA SCANNERS / CREATIVE WEBCAM / IOMEGA BUZZ
Date: Mon, 12 Apr 1999 10:59:19 +0200
Greg Bruhin wrote:
>
> Hi,
>
> If a driver for one of these devices exists or if you know someone who is
> creating these drivers please mail me his/her e-mail address.
>
> Thanks
>
> Greg
>
> [EMAIL PROTECTED]
For Buz :
http://www.lysator.liu.se/~gz/buz/
http://www.munich-vision.de/buz/index.html
------------------------------
From: Pietro Marchionni <[EMAIL PROTECTED]>
Subject: pcmcia on desktop pc for linux
Date: Mon, 12 Apr 1999 11:28:23 +0200
Hi all
I need to connect some PCMCIA cards into a desktop pc with linux inside.
I found some hw stuff but there are no linux drivers for it.
Do you know some hw with linux drivers that allows to plug pcmcia cards
into a desktop pc?
Thanks in advance
Pietro
------------------------------
From: "Scan" <[EMAIL PROTECTED]>
Subject: ATTENTION UMAX and SANE (UNIX) developers: Please make a compatible network
scanning protocol!
Crossposted-To:
comp.periphs.scanners,comp.graphics.apps.gimp,comp.os.linux.networking,comp.os.ms-windows.nt.misc,alt.comp.periphs.scanner
Date: Mon, 12 Apr 1999 10:05:35 GMT
topic says it all, im sick of only being able to network my UMAX 610 scsi
scanner with another windows machine with the same operatin system...
------------------------------
From: **Nick Brown <[EMAIL PROTECTED]>
Subject: Which driver for NCR 53C810 SCSI ?
Date: Mon, 12 Apr 1999 12:16:10 +0200
Reply-To: [EMAIL PROTECTED]
In the kernel setup, there are two drivers for the NCR 53C810 SCSI
card. One claims to support the 7xx and 8xx drivers, the other just the
8xx family.
Is there any particular reason to use one rather than the other for an
8xx card ? And does anyone know why there are two drivers ?
--
===============================================================
Nick Brown, Strasbourg, France (Nick(dot)Brown(at)coe(dot)fr)
Protect yourself against Word 95/97 viruses, free - check out
http://www.geocities.com/NapaValley/Vineyard/1446/atlas-t.html
===============================================================
------------------------------
From: Todd Ostermeier <[EMAIL PROTECTED]>
Crossposted-To:
comp.lang.java.advocacy,comp.os.linux.advocacy,comp.os.ms-windows.advocacy,comp.os.linux.setup
Subject: Re: Building a Better Distribution. (WAS: X munges the graphics card? (Re:
Windows 2000 Rah! Rah! Session falls flat))
Date: Sun, 11 Apr 1999 18:56:54 -0500
On 11 Apr 1999, Bloody Viking wrote:
: In comp.os.linux.advocacy Jeff Szarka <[EMAIL PROTECTED]> wrote:
:
: : I'd probably use linux myself if it weren't for a few things that
: : still make it impossible to switch. Maybe once corel gets their distro
: : out I'll think about it. Here's a perfect example of why I can't/won't
: : use linux:
:
: : Problem: I want to change my resolution.
:
: : Win9x/NT:
: : Right click on desktop - properties - settlings - done
:
: : Linux:
: : Edit config files and god know's what else
Actually, assuming everything is set up properly, just use
ctrl-alt-keypadplus or ctrl-alt-keypadminus to switch resolutions. IF
it's not set up properly, you will have to edit /etc/XF86Config (or
/etc/X11/XF86Config, depending on which you have).
:
: : Why does everything have to be so inconvenient? I don't even care if
: : you can get a program to let you do this, it HAS to come built in,
: : ready to go, I don't want someone telling me I can download
: : whateverdalskjflsa;djfla;sjdfl;.rpm or download the source and compile
: : it. I just want it as simple as win9x/nt, end of story.
:
: It's not a problem of Linux itself. Rather, the problem is in the
: distributions. On a pre-configured computer, it would have a "Packard
: Bell" distribution for that computer. It sounds like you are not satisfied
: with the current state of distributions. You have 3 options:
:
: 1: Put up with it as it stands.
:
: 2: Come up with a "better mousetrap" distribution.
:
: 3: Wait until that perfect distribution comes out.
:
: It sounds like you selected Option 3. That's fine. Of the 3, it's the most
: logical choice unless the cost of software for Windows is too much for you
: to afford. That factor caused me to take Option 1. I have toyed with the
: idea of making a "better mousetrap" distribution designed to be easy to
: set up like an OEM CD. THAT is the standard to shoot for with distribution
: designing. All you should need is one boot floppy and the CD and one
: blank floppy.
:
: Here's how the "perfect" install would go:
Perfect for an OEM, perhaps. Perfect for a normal user? Not even close
(and remember, your average Joe who builds his own machine can get an OEM
copy of Windows for it, rather than buying off-the-shelf). Manufacturers
who use the Recovery Disk style of system setup are considered bad, in my
book (what happens if an app that came pre-installed needs to be
reinstalled, but you only have it in the tarball on the Recovery Disk?).
:
: You slip in the install floppy and the CD. You then use Fdisk to partition
: your drive, the format. You could have a small script to format the drive
: where after Fdisk-ing, the user just types "format" and it formats. You
: can even automate partitioning!
:
: Next, after getting the hard drive ready, you type "setup", and the root
: partition is mounted and the CD is mounted. Then, one giant tar file is
: untarred into place!
So far, sounds pretty similar to current installations. You boot into a
text UI, answer a few questions (do you have a color display? etc.), and
then setup the drive. Of course, the single tarball solution is not
neccessarily a good one. Agreed, an OEM might like the idea, but what
about other users? No, customizability is a good thing, and that's what
we have now (in fact, most distributions allow you to customize everything
or pick from a few preset installs -- this is good, IMHO, and is also a
better way for an OEM to install. They can simply make their own preset.
SuSE allows you do this, I'm not sure what other distro's do).
:
: Then, it's tome to set up the boot loader. During the partitioning, a
: small DOS boot partition is made. After the tarball is untarred, then a
: FreeDOS-based boot floppy is made that has the FreeDOS, the Linux kernel,
: and Loadlin. After the boot floppy is made, the computer shuts down with
: shutdown -r now to reboot into the Loadlin floppy. Upon booting, the
: AUTOEXEC.BAT then formats the DOS partition as bootable, then the files
: are copied except for AUTOEXEC.BAT which is created anew on the hard drive
: to boot with Loadlin to Linux.
You think maybe a choice should be given here? Yes, we all know you hate
LILO. However, in an OEM situation, either the hardware will work with
Linux and LILO, or the OEM shouldn't be offering Linux. Not to mention
the confusion on the user's end when they buy the machine, plug it in,
turn it on, and see it boot into DOS. I'm sure a number of clueless users
would have some irrate words for whatever OEM went this route (even if
they used FreeDOS, and went straight to Linux after booting to DOS, it
still boots to DOS first). No, LILO makes a much better boot loader in
this area, especially if you're talking about OEM's, as that is a
very controlled environment (hey, they built the machine, after all).
:
: Then the user removes the floppy and reboots into a new Linux box! Then,
: the user logs in as root and X auto-starts and then the user can then use
: the tools to config the server stuff.
While you're at it, why not have the user supply a username/password combo
during the setup, so that they need not log in as root (or are you one of
those people who does everything as root?).
:
: Some problems with this approach:
:
: Detecting hardware. This is where Windows wins hands down. The above
: approach would be great for an OEM CD like a Packrat Bell CD. For a
: general purpose distro, it will be harder. For the latter, it would have
: to be more like Red Hat as opposed to a glorified Slackware.
Most hardware that would be in a new machine can actually be auto-detected
by the kernel (assuming PCI), if the proper modules are available.
Otherwise, a bit of work would need to be done (perhaps automated by a
sufficiently intelligent script that I can't write).
:
: Adding hardware. Again, Windows wins out. That's becuse of the problem of
: kernel compiles. The modules are a potentially big help. Of course,
: there's the age-old problem of unsupported hardware.
If the kernel compiles are an issue, why not just compile everything in as
modules, have the modules available, and when a new piece of hardware is
installed, include an installation script that loads the module? Out of
curiosity, though, is it really too much to ask that an admin be able to
configure/compile a kernel? I don't think so. Of course, someone could
probably script around the procedure (ie, the user only takes care of the
config part, while the script runs the make dep clean bzImage ..., places
the kernel in the proper place, and reruns lilo ... er, I mean edits the
autoexec.bat file for FreeDOS to boot the new kernel).
:
: X Window. This is the biggest problem with this idea. You could have as
: the default the generic X server and client. For an OEM, the
: "distribution" would be configured for the computer being manufactured.
Simple. Just run one of the many excellent X setup tools during the
install (hmm... that's exactly how it's done now. Imagine that!).
:
: If you read the above install sequence of a "better mousetrap" distro,
: you'll realise it sounds like a Slackware of sorts. It's indeed inspired
: by Slackware and my convoluted booting methods. (Since I don't use LILO.)
No, not really. Even slackware allowed selection of what to install. The
major difference between what you've listed and what any sane installation
does is that you allow the user no choice in what is installed (hey, even
the Windows setup allows you to choose whether or not you want to install
solitaire -- contrary to popular opinion, it is not an integral part of
the operating system, and not installing it does not hurt Windows).
: All that's different is the approach of having one giant tarball. However,
See a couple lines up for my view of the tarball.
: you could make the OEM distro out of Red Hat as you'd install Red Hat on
: one of the computers at the factory, tar up the filesystem, and make this
: wierd Slackware out of it. Indeed, this approach would be a crossbreed
: between Red Hat and Slackware. The Red Hat is the installed filesystem and
: the Slackware is how it's installed, with my added thing with Loadlin.
Yes, this process almost makes sense for an OEM. Of course, that's until
you realize that with a small amount of work (smaller than installing a
whole system and tarring it all up), you can create a preset installation
profile (using SuSE) while installing the initial system, and then simply
use that preset to install the rest. Then all that's left is general
setup, which could easily be propogated over a network from the initial
install. This way, there's no need to work with a huge tarball.
________________________________
Todd Ostermeier
[EMAIL PROTECTED]
http://www.ews.uiuc.edu/~ostermer/index.html
ICQ UIN: 2253928
A-723
________________________________
------------------------------
From: Marco BANO <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.networking,comp.periphs.scsi
Subject: Re: Linux Networking Performance Question?
Date: Mon, 12 Apr 1999 12:00:20 +0200
Reply-To: [EMAIL PROTECTED]
Chris Pitzel wrote:
> > It would cost us over �10,000 pounds to upgrade the HP9000 hard disks and
> > controller to Ultra2 SCSI (80Mbs) and have considered storing the data on
>
> One thing to consider is the fact that your existing 20mb/sec
> controllers on the HP9000 probably aren't even being utilized to their
> fullest capacity. Modern hard drives only put out around 4-5mb/sec each
> at the most, and it's unlikely that all of the drives in your HP 9000
> are active at any given time.
>
> There are other considerations, of course, with the HP9000, such as the
> fact that you likely will have to buy 'HP-brand' drives (which usually
> are just Seagate Barracudas anyways..), and of course, you will need to
> consider your backup needs as well.
>
> > linux boxes over the 100BaseT network.
> > Our primary data is UniData database files. A linux box with Ultra2 drives
> > and controller can be purchased for under �2000 with increased storage
> > space.
>
> Yes, or you could just add some Ultra2 drives to the HP 9000 machine.
> Since SCSI is forwards and backwards compatible, you should be able to
> do this without any problems whatsoever. The fact that the HP
> controller and existing drives only support 20mb/sec isn't really an
> issue, considering that you're dealing with database files which are not
> terribly taxing on the bandwidth of your SCSI controller.
>
> > 1) What would the performance increase be like if I were to go for this
> > method, considering our current local HP drives are rated at 20Mbs.
>
> You wouldn't see a performance increase. Your HP machine has plenty of
> capacity for additional drives.
>
> > 2) Where would the primary bottleneck be.
>
> Unless the HP9000 is multihomed (ie: fitted with multiple network
> cards), the network card would be busy doing the NFS traffic from the
> Linux box, and would be impaired in it's capability to serve the network
> it's intended to serve.
>
> With 100mb/sec ethernet, you only get around 10mb/sec of effective
> capacity, if even. If you're doing some heavy work on Linux mounted
> SCSI drives that you propose, that could nearly eat up all the
> bandwidth, leaving none for your real clients to use.
>
> > 3) As the network is switched, would there be a degredation in network
> > performance.
>
> Definitely, due to saturation of the network interface on the HP machine
> itself, not to mention the increased load on the switch, and
> corresponding decrease in reliability as well (because when you add an
> extra machine, you've just established another point of failure..).
>
> > Any tips or information would be most appreciated.
>
> I would look into fitting Ultra2 drives onto your existing 20mb/sec
> controller. You won't see any performance degradation.
Sorry for the intrusion...
But Are you sure that you could add Ultra2 drives on a 20mb/sec controller ? it
depends on which type of controller you have . I have in a D350 a FWD, so I need
differential (high voltage) disks . isn't it ?
I wish I could add "normal Ultra2" drives (for investiment.. save).
Please drop me a line ... to help me.
--
Marco BANO
Network administrator Consultant
EUMETSAT
Am Kavalleriesand 31
64295 DARMSTADT
Germany
Office : ++49 6151 807536
------------------------------
From: [EMAIL PROTECTED] (Michael Rubenstein)
Crossposted-To:
comp.lang.c,comp.os.linux.setup,comp.os.linux,comp.os.linux.x,alt.os.linux,comp.lang.c++
Subject: Re: [Fwd: Source Code To Windows 98 (programmer humor) (fwd)]
Date: Mon, 12 Apr 1999 10:44:54 GMT
On Mon, 12 Apr 1999 00:31:38 -0700, Matt Brunk <[EMAIL PROTECTED]>
wrote:
>Michael Rubenstein wrote:
>>
>> On Sat, 10 Apr 99 23:01:43 GMT, [EMAIL PROTECTED]
>> (Lawrence Kirby) wrote:
>>
>> >A slighly different question is whether void main() is a true MS-C extension
>> >i.e. is explicitly documented to work. Many things work by accident on
>> >many compiler versions, but they can't be considered true extensions
>> >unless the compiler provides some guarantee, and more than just usage in
>> >code examples.
>>
>> Microsoft's documentation for main doesn't mention void main, but
>> many (most?) of the examples do use void as the return type.
>> --
>> Michael M Rubenstein
>
>What version of Microsoft's compiler are you using?
>
>If you're using Microsoft VC++ 5.0 or 6.0, the documentation *does*
>mention void main. Just search for "main" and then click the "Program
>Startup: the main Function" topic.
>
>As to whether Microsoft implements void main as an extension, that is
>another topic for discussion.
Sorry. I just noticed that this is cross posted to
comp.lang.c++. I was reading this in comp.lang.c. You are
correct, the C++ documentation does document void main, so it is
an extenstion to C++. The C documentation does not.
--
Michael M Rubenstein
------------------------------
** 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
******************************
