Linux-Hardware Digest #488, Volume #14 Fri, 16 Mar 01 13:13:11 EST
Contents:
Re: Promise FastTrak100 - OK, now what? (Larry Snyder)
Re: VIA82Cxxx Install Probs - MD 7.1 (Alberto BARSELLA)
Re: serial/parallel port card (Mark Bratcher)
10 gig disk in a 500 meg BIOS (Rick Griffiths)
Re: Reccomendations for system for high data throughput (Mark Bratcher)
onstream S-50 (Fleury =?iso-8859-1?Q?S=E9bastien?=)
Re: 10 gig disk in a 500 meg BIOS (John Joseph Trammell)
Re: Oki 320T & Decwriter printers on RH 6.1 ? ([EMAIL PROTECTED])
Promise Ultra66 + RAID1 problems (Dan Reagan)
Re: Processor ID
Motorola Cellphones serial protocol??? (bgeer)
Re: Processor ID
Re: logitech and others... (Rafael Rodriguez Pappalardo)
Re: ATA 100 / UDMA 4/5 on 815 board ([EMAIL PROTECTED])
Re: Processor ID (Roberto Nibali)
Re: 10 gig disk in a 500 meg BIOS (Vilmos Soti)
----------------------------------------------------------------------------
From: Larry Snyder <[EMAIL PROTECTED]>
Subject: Re: Promise FastTrak100 - OK, now what?
Date: Fri, 16 Mar 2001 15:35:23 GMT
After creating a boot partition using the supplied driver disk from
Promise I did modify lilo.conf and all is well - booting the stock RH
7 kernel. I would like to move forward and use 2.4.X - any progess
getting RAID0 to work with this platform ?
Larry
iQXth <[EMAIL PROTECTED]> wrote:
> Hey, I got it to work now.
> I WAS able to use the boot disk! I had to use the 'root=/dev/sda3'
> parameter. But shouldn't the boot disk know this already? It even
> makes reference to '/dev/sda3' in the paragraph that precedes the boot
> prompt.
> Anyway, after getting it up and running, it was easy to change
> 'lilo.conf', add the 'lba32' option, and apply the changes.
> Now I can boot to Red Hat Linux without a boot disk!
------------------------------
From: Alberto BARSELLA <[EMAIL PROTECTED]>
Subject: Re: VIA82Cxxx Install Probs - MD 7.1
Date: 16 Mar 2001 16:38:34 +0100
Patrick F Harris <[EMAIL PROTECTED]> writes:
> I got this chip set to work with sound with the redhat 6.95 without much
> trouble. I just tries insmod on a couple of the sound drivers, one of them
> worked, I am not using any of the ALSA stuff. When I get back to the system I
> will post the detailes.
Make sure you get 16-bit sound. When I tried the first time with the
kernel drivers xmms was running fine, but any app trying to set the
output to 16-bits would stop complaining that /dev/dsp didn't support
that option....
Bye,
Alberto
--
Alberto BARSELLA
PGP fingerprint = 13 3F 22 D2 0B 0A D3 25 F1 89 FE B5 82 AD 75 2A
** Beliefs are dangerous. Beliefs allow the mind to stop functioning.
A non-functioning mind is clinically dead. Believe in nothing... **
------------------------------
From: [EMAIL PROTECTED] (Mark Bratcher)
Subject: Re: serial/parallel port card
Date: Fri, 16 Mar 2001 15:47:46 GMT
On Thu, 15 Mar 2001 10:03:30 -0800, Rodney D. Myers <[EMAIL PROTECTED]> took up his/her
discourse and spake:
>Is there any particular serial/parallel port card that works with
>Linux?
>
>I need to try and get my printer working, and the onboard port isn't
>doing it's job.
>
>Thanks
>--
>Rodney D. Myers <[EMAIL PROTECTED]>
>Member of Digital Freedom Alliance
>Amateur: KG6ANX GMRS: WPOM592
>ICQ# : 18002350 Have A NORML Day
Rodney,
Any number of the SIIG PCI parallel port cards have worked for me.
Linux recognizes them right off and they work.
Why do you suspect your onboard port is faulty?
--
Mark Bratcher
To reply direct, remove both underscores (_) from my email name
===============================================================
Escape from Microsoft's proprietary tentacles: use Linux!
------------------------------
From: Rick Griffiths <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.misc
Reply-To: [EMAIL PROTECTED]
Subject: 10 gig disk in a 500 meg BIOS
Date: Fri, 16 Mar 2001 10:52:52 -0500
I've read the hard drive howto, and I am still unclear. Here is my
question:
Does it matter that the BIOS in my Nec Versa E (circa 1992) can't
handle a 10 gig drive if I am using Linux exclusively on the machine?
The way I read the howto, all I have to do is make a 5meg first
partition as the root. Is this correct? If I make the first partition 5
meg, will the installation system find another partition and use it if
I mount the second partition with /root as the mount point?
Need I even do all this if I am only using Linux? Can I, in other
words, make the first partition 5 gig and still expect the BIOS to find
the master boot record in the first 1024 cylinders?
--
Use reply-to address.
------------------------------
From: [EMAIL PROTECTED] (Mark Bratcher)
Subject: Re: Reccomendations for system for high data throughput
Date: Fri, 16 Mar 2001 16:03:15 GMT
I thought I had read that kernel 2.4.x supports file sizes up
to 16TB. What I did not investigate is whether this is a kernel
issue or a file system issue (since 2.4.x supports other
file system extensions). I suspect it's a kernel issue, in the
word size used to represent file sizes.
--
Mark Bratcher
To reply direct, remove both underscores (_) from my email name
===============================================================
Escape from Microsoft's proprietary tentacles: use Linux!
On Thu, 15 Mar 2001 21:07:52 GMT, xuare <[EMAIL PROTECTED]> took up his/her discourse and
spake:
>Correct me if I am wrong, but I think ext2fs has a 2GB limit on file sizes.
>This may be an issue for consideration...
>
>x
>-
>The best defense is a good offense
>
>perl -e 'print "begin 664
>/dev/stdout\n\;\>\&UA\<S\`P0\&AE\;\&QG871E\+FAO\;65I\<\"YN970\*\n\`\nend\n"' |
>uudecode
>
>
>
>In article <[EMAIL PROTECTED]>,
> Pranab Kumar Nag <[EMAIL PROTECTED]> writes:
>>I am in the process of designing a linux system for large scale data
>>analysis. The overall parameters are:
>>
>>a. large storage requirements (1-4 Terabytes)
>> The data files themeselves are 1-5 Gb each.
>> . The files will be read only.
>>b. The computing platform will be executing fairly simple repeated
>>analysis.
>>
>>Are there any such projects around which discusses pros-cons of various
>>configurations of SANs, NAS and multi-processor environments which can
>>keep the cpu(s) fed with data all the time.
>>
>>Thanks
>>
>>-Pranab
------------------------------
From: Fleury =?iso-8859-1?Q?S=E9bastien?= <[EMAIL PROTECTED]>
Subject: onstream S-50
Date: Fri, 16 Mar 2001 17:02:09 +0100
Reply-To: [EMAIL PROTECTED]
bonjour,
nous avons un lecteur de bande onstream S-50 avec un systeme linux
Mandrake 7.2
Le lecteur est reconnue au d�marrage, dmesg nous donne :
(scsi0:0:4:0) Synchronous at 10.0 Mbyte/sec, offset 7.
Vendor: OnStream Model: SC-50 Rev: 1.05
Type: Sequential-Access ANSI SCSI revision: 02
le probl�me, c'est que nous ne r�cup�rons pas l'interface dans /dev
nous avons :
[root@saturne cups]# file /dev/*|grep "(9/0)"
/dev/md0: block special (9/0)
/dev/st0: character special (9/0)
or, si nous tentons un affichage de l'archivage qui a �t� a fait sur la
meme distribution a une p�riode ou le syst�me fonctionnait (depuis, on a
eu un probl�me, on a r�install� la distribution):
[root@saturne cups]# tar tvf /dev/st0
tar: /dev/st0: Cannot read: Erreur d'entr�e/sortie
tar: Au d�but du ruban, fin pr�matur�e.
tar: Error is not recoverable: exiting now
je vous remercie de m'�clairer � ce sujet.
--
FLEURY S�bastien
[EMAIL PROTECTED]
Apprenti Ing�nieur Informatique et R�seaux
SESIN
7 avenue de l'europe 92310 Sevres
------------------------------
From: [EMAIL PROTECTED] (John Joseph Trammell)
Crossposted-To: comp.os.linux.misc
Subject: Re: 10 gig disk in a 500 meg BIOS
Date: Fri, 16 Mar 2001 16:13:31 GMT
On Fri, 16 Mar 2001 10:52:52 -0500, Rick Griffiths wrote:
> Does it matter that the BIOS in my Nec Versa E (circa 1992) can't
> handle a 10 gig drive if I am using Linux exclusively on the machine?
> The way I read the howto, all I have to do is make a 5meg first
> partition as the root. Is this correct? If I make the first partition 5
> meg, will the installation system find another partition and use it if
> I mount the second partition with /root as the mount point?
I think 5 Mb is small, but I tend to tinker a lot. Here's what's
worked for me in the past:
/dev/hda1 /boot small (20 Mb?)
/dev/hda2 / big (500 Mb?)
/dev/hda3 /usr big (1.5 Gb?)
:
:
Put your kernels in /boot, which means they're well below
the 504 Mb limit. More docs at:
http://judi.greens.org/c/h/get/lilodocs.html
------------------------------
From: [EMAIL PROTECTED]
Subject: Re: Oki 320T & Decwriter printers on RH 6.1 ?
Date: Fri, 16 Mar 2001 16:37:35 GMT
Brian Richardson <[EMAIL PROTECTED]> wrote:
> Thanks Kris....I figured an Epson emulation might be the way to
> go - but considering how old the Decwriter is, I'll have to scrounge
> for some info. on checking/setting the emulation mode....if you
> happen to have a source for that info., it would be most appreciated.
> I haven't messed with a Decwriter for more years than I would care
> to admit.
If I recall correctly (good luck, eh?) the emulation on that is set
very much like on the Oki - ie. software controlled (from the main
panel.) There should be a way to get to the menu on the printer
from the buttons on the front...
I might be able to get some more info - who knows. Have you checked
the Compaq web site?
Kris
------------------------------
From: Dan Reagan <[EMAIL PROTECTED]>
Subject: Promise Ultra66 + RAID1 problems
Date: 16 Mar 2001 08:03:49 -0800
I am trying to get a machine running with RAID1 on two identical drives hanging
off of a Promise Ultra66 IDE card. I initially built the machine with a primary
drive on the onboard controller so that I could configure the arrays and copy
the system over to them. The first time that I tried to initialize the array I
got errors saying timeout waiting for DMA (typical error messages as well as
more detailed hardware and software info at the bottom).
I was confused (not an uncommon occurrence for me lately <g>) but I sorted
through a lot of stuff in usenet and made some changes (moved the cards so that
the Promise was no longer sharing an irq with the video card) and it allowed me
to build the array. I copied the system over there, worked through some lilo
issues and thought that I had it sorted out. It booted to the raid devices, ran
X happily, I beat on it by compiling a few kernels and generally working it as
hard as I felt that I could and all was right in my computer world. This lasted
for about 36 hours. After that I started getting similar errors and couldn't get
the thing to remain running stably for more than a few minutes.
After that I proceeded to flail about and do just about anything that occured to
me to try. I have run with and without and in every combination the generic IDE
driver, the pdc202xx driver and the piixx driver in the kernel and the error
seems to happen with any combination. I tried turning dma off with hdparm before
initializing the arrays and again, no real change. I figured compiling a kernel
with no smp support might be worth a try but also no change. I've swapped the
cards around so much that I'm thinking about starting a three card monty table,
always with no change. I've also tried forcing the Promise card to use various
interrupts as I disabled things to try and free up as many as I could and try
them, still no changes. Through all of this the kernel booted fine, recognized
all hardware and didn't complain until I actually started to work with the two
devices on the promise card.
If I simply remove hde the machine will come up and run fine with the raid
device in degraded mode. I have plugged the drive that was connected to hde into
the onboard controller making it hda and added it to the arrays and they rebuilt
fine. I've also swapped the drives around so that the original hdg is on the
onboard controller and the original hde is back on the promise card and so far
so good. I've also swapped my udma cables around quite a few different ways and
haven't seen a difference with either one.
It appears that I have problems when both devices are on the promise card at the
same time.
I'm basically at a complete loss and I'd greatly appreciate any advice anyone
had to offer before I do something drastic like test the aerodynamic stability
of these drives ;')
Thanks very much for any help,
Dan Reagan,
watching the snow melt in Walpole, NH.
[EMAIL PROTECTED]
Hardware: Tyan S1692 MB with 440LX chipset, Intell PIIX4 onboard IDE controller,
Two PII-333 processors, Matrox MilleniumII PCI, Promise Ultra66 (pdc20262), 3com
3c905c, two Seagate ST310210A 10.2 GB drives each with a UDMA cables, one is a
promise cable and the other a Belkin cable.
Software is Debian unstable, current as of a few days ago, kernel version
2.4.2-ac19.
typical error messages when drives are being cranky:
ide_dmaproc: chipset supported ide_dma_timeout func only: 14
hde: irq timeout: status=0x50 { DriveReady Seek Complete }
hdg: timeout waiting for DMA
ide_dmaproc: chipset supported ide_dma_timeout func only: 14
hdg: irq timeout: status=0x50 { DriveReady Seek Complete }
hde: timeout waiting for DMA
ide_dmaproc: chipset supported ide_dma_timeout func only: 14
hde: irq timeout: status=0x50 { DriveReady Seek Complete }
hdg: timeout waiting for DMA
ide_dmaproc: chipset supported ide_dma_timeout func only: 14
irq timeout: status=0x50 { DriveReady Seek Complete }
hde: timeout waiting for DMA
ide_dmaproc: chipset supported ide_dma_timeout func only: 14
hde: irq timeout: status=0x50 { DriveReady Seek Complete }
hdg: timeout waiting for DMA
ide_dmaproc: chipset supported ide_dma_timeout func only: 14
hdg: irq timeout: status=0x50 { DriveReady Seek Complete }
------------------------------
From: [EMAIL PROTECTED] ()
Crossposted-To: comp.os.linux.development.apps,comp.os.linux.development.system
Subject: Re: Processor ID
Date: Fri, 16 Mar 2001 16:57:25 -0000
In article <[EMAIL PROTECTED]>,
Matt Woodyard <[EMAIL PROTECTED]> wrote:
>You can't track the MAC addr anywhere but the local lan segment .
Even if you could, the MAC address is easily changed.
--
http://www.spinics.net/linux
------------------------------
From: [EMAIL PROTECTED] (bgeer)
Subject: Motorola Cellphones serial protocol???
Date: 16 Mar 2001 09:58:22 -0700
I use a Motorola TimePort cellphone which has memory for 99 names & 4
phone numbers for each. I purchased their serial port hardware &
software package [designed for Win95/98] for downloading that info.
It works ok but I hate to boot Win98 just to download my phone.
Especially since the phone numbers I want to download are in my
Handspring Prism which I manage with Linux & Coldsync.
With 30 years of software experience with all kinds of hardware,
operating systems, & programming languages exotic & plain, programming
a download utility should be, oh, maybe a *full* weekend's spare time
hacking for me if I make it fancy...:-)
So, I nicely & reasonably described my credentials, explained my
reasoning, & requested documentation on the TimePort's serial
communications protocol from Motorola customer support & quickly
received:
"The information you are requesting is considered proprietary and
will not be sent out the individual consumers. We reserve the right
to save these for our technicians that are affiliated with Motorola."
Ok, so what's their problem? [Rhetorical question!]
My real request is for pointers to any useful information so I can
program a download utility. Anyone know where I might find some
useful info?
Naturally, anything useful I come up with will be GPL'd or less...:-)
Much obliged, Bob
--
<> Robert Geer & Donna Tomky | |||| |||| <>
<> [EMAIL PROTECTED] | == == Suddenly, == == <>
<> [EMAIL PROTECTED] | == == We feel enchanted! == == <>
<> Albuquerque, NM USA | |||| |||| <>
------------------------------
From: [EMAIL PROTECTED] ()
Crossposted-To: comp.os.linux.development.apps,comp.os.linux.development.system
Subject: Re: Processor ID
Date: Fri, 16 Mar 2001 16:59:38 -0000
In article <98sfdf$de6$[EMAIL PROTECTED]>,
Rolf Magnus <[EMAIL PROTECTED]> wrote:
>> You can't track the MAC addr anywhere but the local lan segment .
>But it is unique, so it it can be used as your ID.
No, the MAC address is not unique. Maybe it is supposed to be but reality
doesn't work that way. Besides, it can be changed.
--
http://www.spinics.net/linux
------------------------------
From: Rafael Rodriguez Pappalardo <[EMAIL PROTECTED]>
Subject: Re: logitech and others...
Date: 16 Mar 2001 16:27:52 GMT
ERI <[EMAIL PROTECTED]> wrote:
> My Logitech mouses work fine under Linux... Is yours not supported ? What is
> it ?
Mine is a Marble FX trackball. I am not able to use the red button (the 4th one)
to do scrolling.
Regards,
--
Dr. Rafael R. Pappalardo
Dept. Quimica Fisica, Fac. de Quimica, Univ. de Sevilla. 41012-Sevilla (Spain)
e-mail: [EMAIL PROTECTED]
------------------------------
From: [EMAIL PROTECTED]
Subject: Re: ATA 100 / UDMA 4/5 on 815 board
Date: Fri, 16 Mar 2001 17:10:46 +0000
On Thu, 15 Mar 2001 22:32:36 GMT
"Me" <[EMAIL PROTECTED]> wrote:
> In article <[EMAIL PROTECTED]>,
> [EMAIL PROTECTED] wrote:
>
> Try this:
> hdparm -c3 -d1 -X68 or X-69
> If it speeds up your HD, then put it into rc.sysinit or rc.local
Hi,
Thanks for the response. Unfortunately I had already tried that, but it does not
produce any improvement. The only thing that seems to make any difference is setting
DMA on or off. When it is off I get about 4 MB/sec and when it is on I get about 15.8
MB/sec. All other hdparm settings just seem to be ignored. If you actually have a
setup similar to mine I would be very interested to know whether the suggestion you
make has actually improved your transfer rates.
Regards,
Geoff
------------------------------
From: Roberto Nibali <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.development.apps,comp.os.linux.development.system
Subject: Re: Processor ID
Date: Fri, 16 Mar 2001 18:12:19 +0100
[EMAIL PROTECTED] wrote:
>
> In article <[EMAIL PROTECTED]>,
> Matt Woodyard <[EMAIL PROTECTED]> wrote:
>
> >You can't track the MAC addr anywhere but the local lan segment .
>
> Even if you could, the MAC address is easily changed.
>
> --
> http://www.spinics.net/linux
First off, I haven't followed this thread. I just want to drop
in and share some thoughts about your statement, although I
reckon I'm pretty offtopic herewith.
You correctly state that the MAC address basically is
easily to change, however this is only true for the so
called Station MAC address or the OEM Station MAC address.
The Node Address is not easily changeable since this is
burnt into the PROM. For example the 3com PCI cards under
Linux:
MAC changeable
====================+==========================
station address | ifconfig ... hw ether MAC ....
OEM station address | vortex-diag -# 1 -D -f -H MAC -w
3com Node address | not changeable (although there was
| a paper on how to do it)
But for the problem you answered you're absolutely right,
even if someone could track down the MAC address, it's
changed very easily and also done very often for HA systems
to be able to do a faster failover/failback with the same
IP.
... just my 2 cents
Roberto Nibali, ratz
--
mailto: `echo [EMAIL PROTECTED] | sed 's/[NOSPAM]//g'`
------------------------------
Crossposted-To: comp.os.linux.misc
Subject: Re: 10 gig disk in a 500 meg BIOS
From: Vilmos Soti <[EMAIL PROTECTED]>
Date: Fri, 16 Mar 2001 17:32:15 GMT
Rick Griffiths <[EMAIL PROTECTED]> writes:
> Does it matter that the BIOS in my Nec Versa E (circa 1992) can't
> handle a 10 gig drive if I am using Linux exclusively on the machine?
I have no idea about your machine, but my primary machine is a P133
and I have a 13GB disk in it. The BIOS recognizes it correctly, but
it hangs at boot. I can either make BIOS ignore the disk (this is hdb)
or recognize only as 8GB. In these cases it boots. However, when I
partitioned the disk under Linux, in the fdisk I went to the expert
mode and reconfigured the geometry to reflect the true 13GB size.
Linux doesn't use the BIOS for the disk stuff. It handles it entirely
on her own. This is why I could use it even it was not recognized by
the BIOS or the BIOS had the wrong idea about it.
> The way I read the howto, all I have to do is make a 5meg first
> partition as the root. Is this correct? If I make the first partition 5
> meg, will the installation system find another partition and use it if
> I mount the second partition with /root as the mount point?
You don't need the first 5MB as root. What it recommends (I think)
is that you have a 5MB partition at the beginning which would be
/boot and this would be where your kernel image is. The kernel has
to be under 1023 cylinders entirely (newer Lilos don't need this anymore),
and this is the reason that /boot partition is mentioned. I never had it,
I do these things on my own, but for beginners, this is clearly a
good way to go.
> Need I even do all this if I am only using Linux? Can I, in other
> words, make the first partition 5 gig and still expect the BIOS to find
> the master boot record in the first 1024 cylinders?
The master boot record (mbr) is not only "in the first 1024 cylinders".
It is the very first sector on your disk. It is in your first one
cylinder. When the system boots, it reads the contents of the first
sector on the disk (head0, cylinder0, sector1) and executes it.
In that sector you have lilo which in turn will load the kernel.
By definition, at this time, you still rely on the BIOS' idea of the
disk geometry, and it has this 1024 cylinder limitation. This is
why the compressed kernel image (vmlinuz) should be under it.
Once the kernel is up and running it will probe your machine for
everything, including the hard disks, and will recognize them correctly.
As I know, never Lilos don't have this 1024 cylinder limitation,
but I have no experience with it.
Vilmos
------------------------------
** 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 by posting to comp.os.linux.hardware.
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
******************************
