Linux-Hardware Digest #701, Volume #12 Mon, 17 Apr 00 14:13:13 EDT
Contents:
Re: i810 Chipset - Error Installing (Scott Zielinski)
Printer problem ?? ("Daemon J")
Modem speed low: Olitec Self Memory 56K/V90 ([EMAIL PROTECTED])
printing with lj 1100 (James Phillpotts)
Linux and I820 chip set ("He who does the obeying (sometimes)")
Re: Joysticks and es1371 (Henrik Carlqvist)
Re: Phoneline NICs for Linux (Mark Darby)
Re: ati xpert128 video card problem (Joseph Szymczak)
Re: Guys,Guys BeOS For Linux Is Here !! (David C.)
Re: Linux Uses Less Power? (Jim Jerzycke)
Configure Kernal for using Promise 2300+ ("Roger E.")
Re: Where is my 128 MB RAM gone? (David C.)
Configure Kernal for using Promise 2300+ ("Roger E.")
Re: Where is my 128 MB RAM gone? (David C.)
Configure Kernal for using Promise 2300+ ("Roger E.")
Re: DAT backup (David C.)
Re: Dual Xeon Motherboard? (David C.)
Re: S3 Trio 3D graphic Chipset ("Jimmy")
----------------------------------------------------------------------------
Date: Mon, 17 Apr 2000 12:12:44 -0400
From: Scott Zielinski <[EMAIL PROTECTED]>
Subject: Re: i810 Chipset - Error Installing
Ok, I DO have the Macmillan Mandrake Linux 7.0 version...is there
anything I can do??
>vern wrote:
>
> I went through the exact same thing with i810 I and
> another guy on another newsgroup had bad experiences
> with the Macmillian "flavor" of the kernel. I have recently
> installed Mandrake 7.0 from the current issue of Maximum
> linux and it worked like a charm, so beware of the Macmillian
> version. So a nonstandard "patch" can make agpgart rebuild
> throw up on a 2.2 kernel.
> Vern
>
[snip]
> >> Two questions: Q1. What kernel (or distribution) are you using? The
> >> agpgart rpm was geared toward RedHat 6.0, which I think is kernel
> >> 2.2.5. I am using it on RedHat 6.1 with kernel 2.2.12 without
> >> problems. Q2. What does "rpm -q glibc-devel" tell you? Install it if
> >> it's not there. You might need a few more developer packages as well.
> >>
> >> --
> >> DG e-mail is: [EMAIL PROTECTED]
> >> (remove the Z's--they're what I do when I read SPAM!)
> > I got a better handle on it - and the installation seemed to complete
> > correctly - however, when i tried to run "startx" - I was given a new
> > error, that read simply "can't open /dev/agpgart"
> >
> > Thanks, Scott
------------------------------
From: "Daemon J" <[EMAIL PROTECTED]>
Subject: Printer problem ??
Date: Tue, 18 Apr 2000 00:24:00 +0800
I have a printer LEXMARK 3200 But I never can use in linux system
do there have any approach to slove this problem ??
------------------------------
From: [EMAIL PROTECTED]
Subject: Modem speed low: Olitec Self Memory 56K/V90
Date: Mon, 17 Apr 2000 16:43:02 GMT
My external Olitec Self Memory 2000 56K V.90 modem
only
connects at 26400, and occasionally at 28800 bps.
I
have tried everything I can think of, and cant
figure
out why it does not connect at 56K. It could be
the
ISP or the phone line quality that is at fault,
but I wonder if
there is something else that is wrong...
Here is my setup:
I have RH6.1 running on a 133 Pentium MMX machine,
and use KPPP to connect. Modem speed is set to
11520, and the init string is AT&F&K3B24, as
recommended by Olitec. I have used setserial
to change to "spd_hi" for /dev/ttyS0. I always
get
at least a 26400 bps connect, and sometimes a
28800 connect (as reported on the kppp icon).
Any suggestions as to what settings to change,
to get the 56K speed this modem is supposedly
capable of
Thanks
Ram
Btw. I have tried running the "cinfo" program that
Olitec supplies on its web site, but it reports
back (in French) that
I don't have an Olitec modem!
Sent via Deja.com http://www.deja.com/
Before you buy.
------------------------------
From: James Phillpotts <[EMAIL PROTECTED]>
Subject: printing with lj 1100
Date: Mon, 17 Apr 2000 16:48:48 GMT
ok, so i can't get my Laserjet 1100 so be detected by linux (RH6.1) at
all. I've tried all the ports and a number of other HP drivers - does
anyone know of a driver i can install to get this working?
------------------------------
Date: Mon, 17 Apr 2000 17:14:39 +0000
From: "He who does the obeying (sometimes)" <[EMAIL PROTECTED]>
Subject: Linux and I820 chip set
I am using SuSE 6.4 with a SuperMicro I820 Chipset and find that the
CMOS Clock setup part of the boot routine does not work. This is witht
he 2.2.14 kernel. I am getting the impression that this and earlier
versions of Linux are not fully I820 compatible yet. I have also tried
Mandrake 7.0 and Corel (forget which version but fairly recent) and they
show problems though differrent ones. Free BSD 2.4 also does not
install cleanly on the M'board, I have regeisterd the problem with them.
Does any one else have experiences of this chipset and LINUX working or
otherwise???
--
John
------------------------------
From: Henrik Carlqvist <[EMAIL PROTECTED]>
Subject: Re: Joysticks and es1371
Date: Sun, 16 Apr 2000 21:24:36 +0200
"Arthur J. Yarwood" wrote:
> I'm trying to get my ickle joystick working, it's plugged into my
> SB128 PCI, which happens to be of the es1371 nature.
>
> After consulting the kernal docs, it would seem this could be because
> the port on the soundcard needs to be initialised first. Anyone got
> any idea how to do that?
I have an SB64 PCI, but have no experience from any joystick. However,
I'm using ALSA and found the following lines in my /etc/asound.conf:
soundcard("CARD_1") {
control {
; The type is 'bool'.
switch("Joystick", false)
; The type is 'word'.
; The accepted switch range is from 512 to 536.
; Available addresses - 0x200, 0x208, 0x210, 0x218
switch("Joystick Address", 0x200)
}
I suppose that I should change "false" to "true" if I had any joystick
and wanted to use it. It isn't very easy to install the ALSA drivers,
but there is a mini howto which explains how to do it.
regards Henrik
--
spammer strikeback:
root@localhost
[EMAIL PROTECTED]
------------------------------
From: Mark Darby <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.networking
Subject: Re: Phoneline NICs for Linux
Date: Mon, 17 Apr 2000 13:18:39 -0400
Haven't seen any direct support for any of the 10Mbps HPNA cards
available (3Com, LinkSys, Netgear, Intel, ...). Broadcom makes
the controller chip that these NICs are based on.
1Mbps HPNA NICs based on the AMD controller (PCNet-Home) are supported
with the pcnet32 driver, starting with kernel 2.2.13. I've tried
the Diamond HomeFree 1Mbps PhoneLine NIC, and a similar one from
ActionTec. These were using a pcnet32 driver with LRP's
Materhorn release (Linux Router
Project, www.linuxrouter.org). However, the driver doesn't support
switching between HPNA and 10BaseT (which the controller chip is
capable of doing). The driver will come up with HPNA active, which
is OK for the two cards I mentioned above.
AMD has a Linux driver for NICs based on it's 1Mbps HPNA controller chip
that they advertise works only for RedHat 6.0. I've tried it as well
on my RH6.0 system with the same two NICs above without problems.
Their driver also has the flexibility to configure for 10BASET or HPNA.
The driver is called pcnethme. However, I didn't see the source code
for it. It appears to be a pcnet32 variant, though.
Also, I haven't seen direct support for NICs based on Intel's 1Mbps HPNA
controller (the 21145, on Intel's AnyPoint 1Mbps NIC), although it seems that
there
may be hooks in the tulip driver source code, in the 2.2.14 release.
Regards,
Mark Darby
Stuart Anderson wrote:
>
> What Phoneline NICs (3Com 3c410 For E.G.) are supported under Linux?
> Actually, I'm wandering if _any_ of these cards are supported?
> --
> Stuart Anderson -- echelon fodder follows
> nuclear iraq bioweapon encryption cocaine korea terrorist NRA assassin
> sarin
------------------------------
From: Joseph Szymczak <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.misc,comp.os.linux.setup,comp.os.linux.x
Subject: Re: ati xpert128 video card problem
Date: Mon, 17 Apr 2000 13:22:27 -0400
I also had probloems with this chip, however Installing v.4.0 of xfree was
easy and it works wonderfully!!
-joe
Adam K Kirchhoff wrote:
> Xfree86 3.3.6 supports the Rage128 line of cards, including the Xpert128.
> However, the 2D support in XFree86 4.0 is much better, and if your card is
> agp there is now a beta of the DRI available for 3D support for the card
> (with PCI support coming soon).
>
> Adam
>
> On Mon, 17 Apr 2000, Parminder Lehal wrote:
>
> > Hello,
> >
> > I bought an ATI expert128 video card for my computer. I am
> > unable to run X on this card. Are there any xservers
> > avalable for this card?
> >
> > Please,.... help...............
> >
> >
> > Thanks,
> >
> > Parminder Lehal
> >
> >
------------------------------
From: [EMAIL PROTECTED] (David C.)
Subject: Re: Guys,Guys BeOS For Linux Is Here !!
Date: 17 Apr 2000 13:25:55 -0400
Rod Smith writes:
>
> BeOS was never available for 68k systems. NeXTStep did originally run
> on 68k hardware (68040, I believe, although there may have been a
> 68030 version). AFAIK, though, it only ran on proprietary NeXT
> hardware, not on Apple hardware.
NeXT later ported NeXTStep to Intel hardware. This port is named
OpenStep, and may still be available from some vendors. It's a nice
system to use and write code for, but your selection of apps is very
slim.
I don't know if they ported this system to other hardware platforms.
> If you want to use a 68k Macintosh with something other than MacOS,
> AFAIK your only choices are various UNIX offshoots. Apple offered one
> for a while, although I think it's no longer available.
A/UX is dead. Don't even bother with it.
It is quite different from the BSD/SysV-derived UNIX systems we're all
used to. It actually looks and feels a lot like MacOS - except that
there is a layer of UNIX APIs underneath, where UNIX apps can run.
It was written purely to get a foot in the door of customers (like the
US government) which were demanding UNIX at the time. (Must like the
reason why Microsoft put a stripped-down POSIX layer in Windows NT.)
> Linux is certainly available for 68k, although distributions are
> slim.
See http://mac.linux-m68k.org.
According to this page, the Debian distribution includes code for
several 68K Mac architectures, but it is out of date. They say you
should get that distribution and then download the newer kernels from
the web site.
> Some of the BSDs may be available for that hardware, too.
NetBSD (http://www.netbsd.org) has a port in progress. See their web
page for info.
Ditto for OpenBSD (http://www.openbsd.org/)
I didn't see any mac compatibility mentioned on the FreeBSD web pages.
-- David
------------------------------
From: Jim Jerzycke <[EMAIL PROTECTED]>
Subject: Re: Linux Uses Less Power?
Date: Tue, 18 Apr 2000 00:29:37 +0000
Thought I posted a reply, but guess something happened.
"rdg" means "Reading", as in the meter reads 2.001 Volts.
"lsd" means "Least Significant Digit" as in the ".001".
And you're right, measuring TRUE power in an AC system gets very
complicated when you take Power Factor into account. If you just measure
the Voltage, and the Current, and multiply them, you'll wind up with the
APPARENT Power, NOT the True Power. The Power Factor (pf) is calculated
by taking the cosine of the phase angle between the voltage and current,
and in a switching type supply under a variable load, it gets messy REAL
fast, as the power factor is constantly changing. BTW....Switching
supplies generally have really crummy power-factors, even under a
constant load, as they basically hang a bunch of capacitors on the
output of the rectifier. The caps serve as an "Energy Storage Bank" for
the switching regulator, and allow it to respond to rapid changes in
load impedance. Most of the better supplies have a "Pre-Regulator"
(Triacs or SCR's) that allows a soft-start, and can be turned on harder
when the supply needs to output some extra oomph.
HTH, and good luck if you try and make any accurate measurements.
Regards, Jim
------------------------------
From: "Roger E." <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.questions,comp.os.linux.setup,linux.redhat
Subject: Configure Kernal for using Promise 2300+
Date: Mon, 17 Apr 2000 17:31:06 GMT
Hello
I've worked with Linux before, but am somewhat new to its installation.
I'm
wanting to install RedHat Linux v4.1 on a i486DX2-66 system. The
motherboard BIOS does not support EIDE only IDE (1 channel). I have a
Promise EIDE 2300+ Vesa LB controller card with onboard BIOS
that I'd like to use with my 1.6GB & 1GB hard drives. I could not find
info on
this card in the Linux Hardware Compatibility HOWTO file.
Does anyone have any experience in using this card with Linux?
I understand Linux will not work with this card in using its EIDE BIOS
to supercede the IDE BIOS on the motherboard? It includes drivers for
other OS's like DOS, Win95, NT 3.x, OS/2, SCO UNIX 3.2.x and AT&T UNIX
SVR 4.0. All are system files, no sources. Would the SCO or AT&T drivers
work w/ Linux???
If I do source a driver, the next problem is then how about installing
Linux using this card
with its EIDE BIOS active right from the start in order to complete the
installation on a 1 GB drive?
Neither Promise nor Red Hat have made available a Linux driver for the
Promise EIDE 2300+.
I did get some information from someone on the net. He has the same
concerns I do.
His note is below........
===============================================================
I suspect the answer is "no". The Promise EIDE 2300+ BIOS programs
seem to play a lot of DOS/WINDOWS specific tricks, from my brief
investigations. E.g. the BIOS "grabs" a piece of memory just below
the 640K boundary, then hooks the required BIOS calls to lie to DOS
when it asks about memory so that it looks like only about 638K is
physically present. I am not a Linux Kernel expert, but since a
multitasking kernel like Linux can't rely on the BIOS routines, I
suspect that it would require some major hacking to accommodate tricks
like that. That would be compounded if Promise considered the tricks
"proprietary" and were therefore unwilling to release any documentation
on exactly what that BIOS was doing. (That would take reverse-
engineering the BIOS code on the board, time consuming, and possibly
a violation of the Promise license).
The board will operate as an EIDE controller using just the BIOS in the
machine, IF the BIOS
supports the size drive you have; I originally got the Promise board
simplly to give me the second IDE chain on a machine that had only one
chain. This allowed me to support a CDRom on the second chain; you
could do that without support in the BIOS for the second chain, since
you were only supporting a CD, not a hard drive on the second chain.
However, my BIOS does not support a 1.6 G drive either, so that is not
an option.
There is an interesting possibility, however. Depending on the drive's
CYL, HEAD, Sector, arrangement, it is possible that the BIOS on the
mother board would exactly see the "first 504 MB" of the drive. This
depends critically on the drive not reporting more heads than the BIOS
can handle, and the BIOS correctly truncating just at Cylinder 1023 (not
wrapping back to 0, if you understand what I mean).
Under Linux, only the Lilo boot uses the BIOS to read the disk, so
setting up either the root ("/") int he first 504 MB, or at least
separating the
"/boot" directory into a separate partition and putting it in the first
504 MB would allow booting to the Linux Kernel. The kernel accesses the
hardware directly, without using the BIOS and therefore will see the
full 1.6 GB.
It isn't quite automatic, though. You have to make sure that the BIOS
and the Kernel see the same "map" of CYLS, HEADS, and Sectors,
otherwise,
what the boot sees in the partition table and what the kernel sees might
make the file systems totally corrupt. Linux is quite configurable,
with all sorts of extra boot parameters, so as long as there is
something
compatibile with the BIOS seeing the first 504 MB, you might be in
business.
==============================================================================
>>>
I would have to do some work on the kernel such that the BIOS and the
Kernel see the same "map" of CYLS, HEADS, and Sectors. However I'm new
to this level of knowledge on Linux at this time. Can anyone provide me
some
advice on getting the card to work with Linux, or would this be a futile
attempt due to proprietary issues?
Thanks for your help,
Roger
------------------------------
From: [EMAIL PROTECTED] (David C.)
Subject: Re: Where is my 128 MB RAM gone?
Date: 17 Apr 2000 13:30:55 -0400
[EMAIL PROTECTED] writes:
> David C. wrote:
>>
>> (Oh, one more thing - don't forget the "M". If you accidentally say
>> "mem=128", then the kernel tries to run with only 128 _bytes_ of RAM,
>> and panics as soon as it tries to boot. Recovering can be a royal
>> pain in the neck.
>
> Sounds like someone is speaking from experience?!? :-)
Yep. I made that mistake once. I won't ever be making that mistak
again. Took me severl hours of searching newsgroups to figure out how
to get my system back without a reinstall. (The trick is to provide the
"mem=128M" line at the LILO prompt when launching the kernel.)
-- David
------------------------------
From: "Roger E." <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.questions,comp.os.linux.setup,linux.redhat
Subject: Configure Kernal for using Promise 2300+
Date: Mon, 17 Apr 2000 17:32:49 GMT
Hello
I've worked with Linux before, but am somewhat new to its installation.
I'm
wanting to install RedHat Linux v4.1 on a i486DX2-66 system. The
motherboard BIOS does not support EIDE only IDE (1 channel). I have a
Promise EIDE 2300+ Vesa LB controller card with onboard BIOS
that I'd like to use with my 1.6GB & 1GB hard drives. I could not find
info on
this card in the Linux Hardware Compatibility HOWTO file.
Does anyone have any experience in using this card with Linux?
I understand Linux will not work with this card in using its EIDE BIOS
to supercede the IDE BIOS on the motherboard? It includes drivers for
other OS's like DOS, Win95, NT 3.x, OS/2, SCO UNIX 3.2.x and AT&T UNIX
SVR 4.0. All are system files, no sources. Would the SCO or AT&T drivers
work w/ Linux???
If I do source a driver, the next problem is then how about installing
Linux using this card
with its EIDE BIOS active right from the start in order to complete the
installation on a 1 GB drive?
Neither Promise nor Red Hat have made available a Linux driver for the
Promise EIDE 2300+.
I did get some information from someone on the net. He has the same
concerns I do.
His note is below........
===============================================================
I suspect the answer is "no". The Promise EIDE 2300+ BIOS programs
seem to play a lot of DOS/WINDOWS specific tricks, from my brief
investigations. E.g. the BIOS "grabs" a piece of memory just below
the 640K boundary, then hooks the required BIOS calls to lie to DOS
when it asks about memory so that it looks like only about 638K is
physically present. I am not a Linux Kernel expert, but since a
multitasking kernel like Linux can't rely on the BIOS routines, I
suspect that it would require some major hacking to accommodate tricks
like that. That would be compounded if Promise considered the tricks
"proprietary" and were therefore unwilling to release any documentation
on exactly what that BIOS was doing. (That would take reverse-
engineering the BIOS code on the board, time consuming, and possibly
a violation of the Promise license).
The board will operate as an EIDE controller using just the BIOS in the
machine, IF the BIOS
supports the size drive you have; I originally got the Promise board
simplly to give me the second IDE chain on a machine that had only one
chain. This allowed me to support a CDRom on the second chain; you
could do that without support in the BIOS for the second chain, since
you were only supporting a CD, not a hard drive on the second chain.
However, my BIOS does not support a 1.6 G drive either, so that is not
an option.
There is an interesting possibility, however. Depending on the drive's
CYL, HEAD, Sector, arrangement, it is possible that the BIOS on the
mother board would exactly see the "first 504 MB" of the drive. This
depends critically on the drive not reporting more heads than the BIOS
can handle, and the BIOS correctly truncating just at Cylinder 1023 (not
wrapping back to 0, if you understand what I mean).
Under Linux, only the Lilo boot uses the BIOS to read the disk, so
setting up either the root ("/") int he first 504 MB, or at least
separating the
"/boot" directory into a separate partition and putting it in the first
504 MB would allow booting to the Linux Kernel. The kernel accesses the
hardware directly, without using the BIOS and therefore will see the
full 1.6 GB.
It isn't quite automatic, though. You have to make sure that the BIOS
and the Kernel see the same "map" of CYLS, HEADS, and Sectors,
otherwise,
what the boot sees in the partition table and what the kernel sees might
make the file systems totally corrupt. Linux is quite configurable,
with all sorts of extra boot parameters, so as long as there is
something
compatibile with the BIOS seeing the first 504 MB, you might be in
business.
==============================================================================
>>>
I would have to do some work on the kernel such that the BIOS and the
Kernel see the same "map" of CYLS, HEADS, and Sectors. However I'm new
to this level of knowledge on Linux at this time. Can anyone provide me
some
advice on getting the card to work with Linux, or would this be a futile
attempt due to proprietary issues?
Thanks for your help,
Roger
------------------------------
From: [EMAIL PROTECTED] (David C.)
Subject: Re: Where is my 128 MB RAM gone?
Date: 17 Apr 2000 13:32:31 -0400
Mathias Mildenberger <[EMAIL PROTECTED]> writes:
>
> As we are talking 'bout xperience: Even in some systems seeming to be
> up to date, you have to tell lilo/kernel the amount of RAM. A good
> example for this is my GA7ix/Athlon 500, which had to be told it had
> 128M, not 64M. This was with SuSE 6.3 and a kernel 2.2.13. Now it's
> working almost perfect.
It's all in the BIOS. Even newer ones don't always provide what LILO is
looking for. Whenever the BIOS doesn't provide the right value, you use
the "mem=###M" line to override it.
-- David
------------------------------
From: "Roger E." <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.questions,comp.os.linux.setup,linux.redhat
Subject: Configure Kernal for using Promise 2300+
Date: Mon, 17 Apr 2000 17:36:41 GMT
Hello
I've worked with Linux before, but am somewhat new to its installation.
I'm
wanting to install RedHat Linux v4.1 on a i486DX2-66 system. The
motherboard BIOS does not support EIDE only IDE (1 channel). I have a
Promise EIDE 2300+ Vesa LB controller card with onboard BIOS
that I'd like to use with my 1.6GB & 1GB hard drives. I could not find
info on
this card in the Linux Hardware Compatibility HOWTO file.
Does anyone have any experience in using this card with Linux?
I understand Linux will not work with this card in using its EIDE BIOS
to supercede the IDE BIOS on the motherboard? It includes drivers for
other OS's like DOS, Win95, NT 3.x, OS/2, SCO UNIX 3.2.x and AT&T UNIX
SVR 4.0. All are system files, no sources. Would the SCO or AT&T drivers
work w/ Linux???
If I do source a driver, the next problem is then how about installing
Linux using this card
with its EIDE BIOS active right from the start in order to complete the
installation on a 1 GB drive?
Neither Promise nor Red Hat have made available a Linux driver for the
Promise EIDE 2300+.
I did get some information from someone on the net. He has the same
concerns I do.
His note is below........
===============================================================
I suspect the answer is "no". The Promise EIDE 2300+ BIOS programs
seem to play a lot of DOS/WINDOWS specific tricks, from my brief
investigations. E.g. the BIOS "grabs" a piece of memory just below
the 640K boundary, then hooks the required BIOS calls to lie to DOS
when it asks about memory so that it looks like only about 638K is
physically present. I am not a Linux Kernel expert, but since a
multitasking kernel like Linux can't rely on the BIOS routines, I
suspect that it would require some major hacking to accommodate tricks
like that. That would be compounded if Promise considered the tricks
"proprietary" and were therefore unwilling to release any documentation
on exactly what that BIOS was doing. (That would take reverse-
engineering the BIOS code on the board, time consuming, and possibly
a violation of the Promise license).
The board will operate as an EIDE controller using just the BIOS in the
machine, IF the BIOS
supports the size drive you have; I originally got the Promise board
simplly to give me the second IDE chain on a machine that had only one
chain. This allowed me to support a CDRom on the second chain; you
could do that without support in the BIOS for the second chain, since
you were only supporting a CD, not a hard drive on the second chain.
However, my BIOS does not support a 1.6 G drive either, so that is not
an option.
There is an interesting possibility, however. Depending on the drive's
CYL, HEAD, Sector, arrangement, it is possible that the BIOS on the
mother board would exactly see the "first 504 MB" of the drive. This
depends critically on the drive not reporting more heads than the BIOS
can handle, and the BIOS correctly truncating just at Cylinder 1023 (not
wrapping back to 0, if you understand what I mean).
Under Linux, only the Lilo boot uses the BIOS to read the disk, so
setting up either the root ("/") int he first 504 MB, or at least
separating the
"/boot" directory into a separate partition and putting it in the first
504 MB would allow booting to the Linux Kernel. The kernel accesses the
hardware directly, without using the BIOS and therefore will see the
full 1.6 GB.
It isn't quite automatic, though. You have to make sure that the BIOS
and the Kernel see the same "map" of CYLS, HEADS, and Sectors,
otherwise,
what the boot sees in the partition table and what the kernel sees might
make the file systems totally corrupt. Linux is quite configurable,
with all sorts of extra boot parameters, so as long as there is
something
compatibile with the BIOS seeing the first 504 MB, you might be in
business.
==============================================================================
>>>
I would have to do some work on the kernel such that the BIOS and the
Kernel see the same "map" of CYLS, HEADS, and Sectors. However I'm new
to this level of knowledge on Linux at this time. Can anyone provide me
some
advice on getting the card to work with Linux, or would this be a futile
attempt due to proprietary issues?
Thanks for your help,
Roger
------------------------------
From: [EMAIL PROTECTED] (David C.)
Subject: Re: DAT backup
Date: 17 Apr 2000 13:39:14 -0400
"Philippe PIERARD" <[EMAIL PROTECTED]> writes:
>
> I try to use a tape to make a backup of my linux system.
>
> It is a scsi DAT tape (model 'python').
>
> The tape recorder is reconize in SCSI hosts as sequentiel block
> device, but i can't mount a tape in /dev/rst*
You can't mount tapes. They're sequential devices. You can only mount
random-access devices - like disks.
Furthermore, your device name is not /dev/rst. Use /dev/st0 or
/dev/nst0. (st0 will rewind the tape after every operation. nst0 will
not.)
I make my backups with tar:
mt -f /dev/nst0 rewind
tar cpPMf /dev/nst0 /
mt -f /dev/nst0 rewoffl
If you need to activate any hardware features (like setting block sizes
or turning on data compression), do it by adding more mt commands before
the call to tar.
-- David
------------------------------
From: [EMAIL PROTECTED] (David C.)
Subject: Re: Dual Xeon Motherboard?
Date: 17 Apr 2000 13:54:09 -0400
"Jack L. Owens" <[EMAIL PROTECTED]> writes:
>
> I think I screwed up. I just purchased a couple of Xeon PII 400-1M
> pulls with the intention of using them in a dual processor
> motherboard.
Oops. Good processors, but expensive and requires Slot-2.
> My current mb is a two year old ASUS P2B that has had few faults. So,
> I was looking at the ASUS P2BD. Unfortunately, the two Xeon processors
> require a Slot 2 motherboard.
Intel makes Slot-2 boards. So does Asus. Some other people here
mentioned some other brands as well.
The Xeon architecture supports up to 8 processors, although some of the
earlier chips don't have this support enabled due to some problems early
on in the Xeon's development. (The first-generation models only
supported 2 CPUs, later they supported 4. The ones made today should
support 8.)
Motherboards with enough slots for 4 or 8 processors can be bought, but
they will be very expensive. Actually, I would expect all Slot-2 boards
to be expensive, because they're normally only sold to vendors selling
expensive server systems.
> I have two IBM ATA66 7200RPM 20.6 GIG drives which could be used in a
> RAID 5 setup.
not with only two drives. With two drives, you can get RAID-0
(striping), RAID-1 (mirroring) or RAID-0/1 (mirroring and striping).
For RAID-5, you need at least three drives. And IMO, you don't see much
efficiency from RAID-5 until you get to 5 drives.
> Any suggestions for motherboard selection would be appreciated.
Can't help here. I've never gone shopping for one.
-- David
------------------------------
From: "Jimmy" <[EMAIL PROTECTED]>
Subject: Re: S3 Trio 3D graphic Chipset
Date: Mon, 17 Apr 2000 19:54:05 +0200
Robert Hampf <[EMAIL PROTECTED]> schrieb in im Newsbeitrag:
[EMAIL PROTECTED]
> Jimmy <[EMAIL PROTECTED]> hélt ŝessu fram:
> : Help!
> :
> : Is there any possibility to get the Trio 3D AGP chipset working properly
> : under Linux with an X-Server?
>
> Have you checked the X-FAQ (F19)?
>
> rh
Yes, I did an tryed it out....no way to get this f* thing working! Is there
any special X-Server available???
------------------------------
** 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
******************************
