Linux-Hardware Digest #975, Volume #13 Sat, 2 Dec 00 19:13:07 EST
Contents:
Booting with 576KB conventional RAM ...or less. (Juan)
Re: Floppy drive won't spit out disk under Mandrake 7.2 (Noble Pepper)
Re: PCI modem under Mandrake 7.1 (Neuromancer)
Re: Booting with 576KB conventional RAM ...or less. ("Peter T. Breuer")
Sound modules: request_module always unresolved. (Gareth Randall)
Re: X windows (Andy Kinsey)
Problems with 3Com 3c905B ("Mark T. Dame")
Re: parallel (Gareth Randall)
Configuring Sound Cards in Red Hat
Re: Floppy drive won't spit out disk under Mandrake 7.2 (Jens)
Re: parallel ("robert smith")
Re: Problems with 3Com 3c905B (Dances With Crows)
Redhat hat professional 7.0 controller card (Drew cutter)
Corrupt Graphics, Voodoo 3000, DRI (John-Paul Stanford)
ftape_calibrate_data_rate failed; ditto 3200 (ECStahl)
----------------------------------------------------------------------------
From: Juan <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.setup,comp.os.linux.misc
Subject: Booting with 576KB conventional RAM ...or less.
Date: Sat, 02 Dec 2000 20:06:42 +0100
Hi.
I've got a problem. A real one.
About three years ago we bought a killer machine (by thoose days), a
Fujitsu-ICL teamserver M754i, a quad PPro with 512MB RAM . The machine
has been running OK until the moment we needed to upgrade the OS (that
was UnixWare 2.1.2.) in order to install Oracle8i. The option were to
spend lots of money and learning time with Unixware 7 or to install (
tachannn!) Linux. We took the second option.
I've been installing RedHat for several years, so I took the RedHat 7
CD-ROMs tha I had on my table, and I have succesfully installed several
times in dual Pentium machines, and went to install the above mentioned
machine. Installation went fine, but it was unable to boot the OS after
finishing. After sometime discarding possible causes of the problem, I
disabled the SCSI BIOS and got enought memory to boot from the boot disk
made by the installation procedure. It would have been OK (not fine, but
OK) except for the fact that the kernell in the floppy disk was a NON
SMP kernel.
The BIOS told that the conventional memory available was 576 KB, no
matter what options I enabled or disabled in the BIOS. There was no
means to disable the system BIOS shadow on RAM, the only option left to
increase the conventional memory. On the other hand, the kernell, when
it boots, recognizes all the installed memory. I think the problem is
with the boot loader that is unable to fit the whole kernel in memory
due to the low ammount of conventional memory. Do you agree?
>From this point on I've tried several booting options :
- Making a boot disk with the smp kernel installed in the machine.
- Using syslinux instead of lilo.
- Compiling lilo with LARGE_EBDA enabled and making a boot disk.
- Using GRUB instead of lilo
- Using loadlin from a msdos 6.22 disk instead of lilo.
In the best cases the machine frozes after printing "OK, booting the
kernel" in the screen.
In the meantime I've got BootMagic and I'll try it on monday (if I dont
get any better suggestion).
Oh! Iforgot to tell you, I called Fujitsu support (Fujitsu Customer
Services) and I was told that there was no way of disabling the system
BIOS shadowing, that was a feature included in the design of the machine
and, by the way, Linux was not a supported OS for this machine. Good
luck and goodbye.
I would appreciate any suggesions, hints or workarounds to solve or
overcome this problem.
(Crying)
I've tried to cut my veins with a Windows98 CDROM, but it did not work
(even for this). Please help me!
(Serious again)
Thanks in advance.
Juan.
------------------------------
From: Noble Pepper <[EMAIL PROTECTED]>
Subject: Re: Floppy drive won't spit out disk under Mandrake 7.2
Date: Sat, 2 Dec 2000 13:32:19 -0600
Peter Bloomfield wrote:
> Jens wrote:
>>
>> This is probably one of those real stupid questions but a search under
>> DEJA didn't help ......
>> I have an LS120 floppy drive that Linux reads just fine but pressing the
>> eject button doesn't get me my floppy disk back. It is a soft eject of
>> some kind (most floppy drives eject the disk mechanically but this drive
>> spits it out like a CD - press the button and a motor does the eject) so
>> it requires interaction (somehow) with the operating system. The drive
>> works fine under windows so it is working ok. Has anyone ever run across
>> this ?
>>
>> Jens
>
> man eject
>
or umount it (umount /mnt/superfloppy for me) then it will eject with the
button.
------------------------------
From: Neuromancer <[EMAIL PROTECTED]>
Crossposted-To: alt.os.linux.dial-up,alt.os.linux.mandrake,comp.os.linux.networking
Subject: Re: PCI modem under Mandrake 7.1
Date: 2 Dec 2000 20:05:58 GMT
Reply-To: [EMAIL PROTECTED]
Any idea when this limitation will be lifted? I have an Actiontec 56k
PCI Call Waiting modem (internal) which works fine, but I hate having to
use the setserial command. Here's why: I can't have ppp0 brought up on
boot because the modem is not yet initialized so things that require at
least ppp0 to be up (whether of not it's linked yet with /dev/modem)
like say a firewall script on bootup cannot be used. I *do* use a
firewall, but I have to have it run from my ip-up.local because it needs
ppp0 to be up and having "ifup ppp0" in my rc.local, ppp0 comes up a few
seconds AFTER I get my login prompt.
-Russ
Thus spake James Richard Tyrer about Life, the Universe, and Re: PCI modem under
Mandrake 7.1:
> PCI modems do not use standard COM port addresses. The I/O address and
> IRQ are assigned by the BIOS. All you need to do to get one working is
> to first find out the I/O address and interrupt number with (as root):
>
> lspci -v | less
>
> and then tell the serial driver where to find it (and the interrupt
> number) with the "setserial" command.
>
> JRT
------------------------------
From: "Peter T. Breuer" <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.setup,comp.os.linux.misc
Subject: Re: Booting with 576KB conventional RAM ...or less.
Date: Sat, 2 Dec 2000 21:10:52 +0100
In comp.os.linux.misc Juan <[EMAIL PROTECTED]> wrote:
> it boots, recognizes all the installed memory. I think the problem is
> with the boot loader that is unable to fit the whole kernel in memory
> due to the low ammount of conventional memory. Do you agree?
You are posting in the wrong forum, Take it to the kernel list.
But no, off the top of my head, 576KB should be fine for the loader,
particularly if it's a bzImage, which loads a bit, decompresses a bit,
loads a bit, decompresses a bit ... Yes, a zImage might have trouble.
> From this point on I've tried several booting options :
> - Making a boot disk with the smp kernel installed in the machine.
> - Using syslinux instead of lilo.
> - Compiling lilo with LARGE_EBDA enabled and making a boot disk.
> - Using GRUB instead of lilo
> - Using loadlin from a msdos 6.22 disk instead of lilo.
That should be the winner, provided it's a bzImage and you have a
loadlin that supports bzImage. Then again, maybe it's worth trying a
zImage with loadlin.
> In the best cases the machine frozes after printing "OK, booting the
> kernel" in the screen.
It looks like it's hitting the bios image in conventional memory, with
result death of bios, and dead machine because the bios is needed for
booting. You'll have to talk to the authors of the kernel boot code.
You might have to edit a bit of the kernels setup.c file that does
the reserviing of memory pages ... but no, you are getting hit before
that, when the machine is still in real mode.
I would guess that if you can boot NT, you might be able to use the NT
loader.
> In the meantime I've got BootMagic and I'll try it on monday (if I dont
> get any better suggestion).
Peter
------------------------------
From: Gareth Randall <[EMAIL PROTECTED]>
Subject: Sound modules: request_module always unresolved.
Date: Sat, 02 Dec 2000 20:33:42 +0000
Intro
=====
I have been trying to compile sound support into my 2.2.17 kernel as a module. The
card itself is a PAS2 (Pro Audio Studio) and is known good.
It works unreliably when sound support is compiled into the kernel i.e. not using
modules. In this case DMA IRQs are lost (DMA timeout) and the sound stops at random
points. It will sustain over a minute and interrupts are registered, so the IRQ
mechanism works most of the time, but not well enough.
I have tried:
1. hdparm -u 1 (unmask IRQs - seemed to have some effect but problem remained).
2. pre-allocate 64K DMA buffer in lower 16MB (no discernable effect)
Consequently I am inclined to assume that this is a bug in the kernel-only
implementation and go for modules as is recommended anyway.
Main Question
=============
However, and this is my main question, I simply cannot build these modules without
getting the unresolved symbols error:
fermat:/lib/modules/2.2.17# depmod -a -v -e
/lib/modules/2.2.17/net/ppp_deflate.o
/lib/modules/2.2.17/net/bsd_comp.o
/lib/modules/2.2.17/net/dummy.o
/lib/modules/2.2.17/misc/pas2.o
/lib/modules/2.2.17/misc/sb.o
depmod: *** Unresolved symbols in /lib/modules/2.2.17/misc/sound.o
depmod: request_module
/lib/modules/2.2.17/misc/sound.o
depmod: *** Unresolved symbols in /lib/modules/2.2.17/misc/soundcore.o
depmod: request_module
/lib/modules/2.2.17/misc/soundcore.o
/lib/modules/2.2.17/misc/uart401.o
This is the same problem, and the same missing symbol every time no matter what I try.
As you can see, the other modules are okay.
I am using Debian 2.2. (kernel 2.2.17-pre6). No updated packages appear to be
available. I have tried replacing the standard /usr/include/asm and /usr/include/linux
with links to /usr/src/linux/include/asm and /usr/src/linux/include/linux. The problem
remains exactly the same.
I have removed sound support other than in modules. I have versions set on everything.
I am doing:
make xconfig
make dep
make clean
make bzImage
cp arch/i386/boot/bzImage /boot/vmlinuz-2.2.17-date
cp System.map /boot/System.map-2.2.17-date
cd /
rm vmlinuz System.map
ln -s /boot/vmlinuz-2.2.17-date vmlinuz
ln -s /boot/System.map-2.2.17-date System.map
/sbin/lilo
#Reboot
make modules
make modules_install
depmod -a # Always fails
Some minor variations have also been tried. e.g. build modules straight after kernel.
If anyone has seen this error and knows a solution then I'd be delighted to hear from
them. If more details are needed, such .config entries, then just say.
Yours,
======= Gareth Randall =======
------------------------------
From: Andy Kinsey <[EMAIL PROTECTED]>
Crossposted-To:
alt.linux,alt.os.linux,comp.os.linux.questions,comp.os.linux.redhat,comp.os.linux.setup
Subject: Re: X windows
Date: Sat, 02 Dec 2000 15:41:32 -0500
What if you did a re-install, and instead of actually formatting and
re-installing, selected "Upgrade" and installed the packages you wanted?
Just a thought...
Andy
Darren and Marla Welson wrote:
> I installed RH62 on a router box without X, but now I want to be able to
> start X remotely. When I go to install an RPM, I keep getting NUMEROUS
> dependencies and do not know where to begin. Can someone tell me what X
> related packages I need and in which order to add the RPMs?
>
>
------------------------------
Date: Sat, 02 Dec 2000 15:44:39 -0500
From: "Mark T. Dame" <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.networking
Subject: Problems with 3Com 3c905B
I have a couple of RedHat boxes (6.0 and 6.2 running 2.4 kernels) with
multiple 3c905B NICs. I keep getting the following error:
eth0: command 0x3002 did not complete! Status=0xf000
on both boxes. It seems to result in periodic network hangs. I think I
was getting the same error with the 2.3 kernel series too.
I don't think I've seen the error on eth1, but I can't swear to it.
eth1 doesn't have as much traffic as eth0, so that may be why I haven't
seen the error on it.
Has anyone seen this error or know how to fix the problem?
-m
--
## Mark T. Dame: mailto:[EMAIL PROTECTED]
## WWW: http://www.mfm.com/~mdame/
## MFM Communication Software: http://www.mfm.com/
"The opposed thumb is one of the defining characteristics of our
species, and mine are strongly opposed to working pointing devices."
-- Bill Machrone, PC Magazine, Sep. 28, 1993, discussing
miniature trackballs
------------------------------
From: Gareth Randall <[EMAIL PROTECTED]>
Subject: Re: parallel
Date: Sat, 02 Dec 2000 21:04:55 +0000
robert smith wrote:
>
> has anyone tried parallel processing using a LAN on linux.i think i saw
> something about it in suse, i just wondered if anyone had tried it...
>
> rob.
Many people have! The method that you use depends on the task you want to perform. A
popular application is in "Beowulf" clusters to solve problems that can be
parallelised but require significant communication between threads of execution across
the network. Such clusters may use PVM (parallel virtual machine) or MPI (message
passing interface) as an API to this interaction. There are some great pages at, and
especially linked to by: www.beowulf.org
Just take a look at Avalon: http://cnls.lanl.gov/avalon/
(actually, look at Avalon last as it is a very big one, and you can build up to it...)
Yours,
======= Gareth Randall =======
------------------------------
From: <[EMAIL PROTECTED]>
Subject: Configuring Sound Cards in Red Hat
Date: Sat, 02 Dec 2000 21:30:06 -0000
I can't make work my soundcard and I just don't know where to start...
Please HELP ME!
--
Posted via CNET Help.com
http://www.help.com/
------------------------------
From: Jens <[EMAIL PROTECTED]>
Subject: Re: Floppy drive won't spit out disk under Mandrake 7.2
Date: Sat, 02 Dec 2000 22:01:46 GMT
Peter Bloomfield wrote:
> Jens wrote:
> >
> > This is probably one of those real stupid questions but a search under
> > DEJA didn't help ......
> > I have an LS120 floppy drive that Linux reads just fine but pressing the
> > eject button doesn't get me my floppy disk back. It is a soft eject of
> > some kind (most floppy drives eject the disk mechanically but this drive
> > spits it out like a CD - press the button and a motor does the eject) so
> > it requires interaction (somehow) with the operating system. The drive
> > works fine under windows so it is working ok. Has anyone ever run across
> > this ?
> >
> > Jens
>
> man eject
Duuuuhhhhh .... I knew it was a stupid question but I didn't know it was
THAT stupid .....
Thanks !
Jens
------------------------------
From: "robert smith" <[EMAIL PROTECTED]>
Subject: Re: parallel
Date: Sat, 2 Dec 2000 22:57:01 -0000
if i had two pentiums, on a peer to peer, would that give me in effect a
pentium II?
rob.
"Gareth Randall" <[EMAIL PROTECTED]> wrote in message
news:[EMAIL PROTECTED]...
> robert smith wrote:
> >
> > has anyone tried parallel processing using a LAN on linux.i think i saw
> > something about it in suse, i just wondered if anyone had tried it...
> >
> > rob.
>
> Many people have! The method that you use depends on the task you want to
perform. A popular application is in "Beowulf" clusters to solve problems
that can be parallelised but require significant communication between
threads of execution across the network. Such clusters may use PVM (parallel
virtual machine) or MPI (message passing interface) as an API to this
interaction. There are some great pages at, and especially linked to by:
www.beowulf.org
>
> Just take a look at Avalon: http://cnls.lanl.gov/avalon/
>
> (actually, look at Avalon last as it is a very big one, and you can build
up to it...)
>
> Yours,
>
> ======= Gareth Randall =======
------------------------------
From: [EMAIL PROTECTED] (Dances With Crows)
Crossposted-To: comp.os.linux.networking
Subject: Re: Problems with 3Com 3c905B
Date: 2 Dec 2000 23:06:49 GMT
Reply-To: [EMAIL PROTECTED]
On Sat, 02 Dec 2000 15:44:39 -0500, Mark T. Dame wrote:
>I have a couple of RedHat boxes (6.0 and 6.2 running 2.4 kernels) with
>multiple 3c905B NICs. I keep getting the following error:
>
>eth0: command 0x3002 did not complete! Status=0xf000
>
>on both boxes. It seems to result in periodic network hangs. I think I
>was getting the same error with the 2.3 kernel series too.
Development kernels = periodic driver problems? Sounds fairly normal to
me. Does 2.2.x show the same behavior? And which module are you
loading for your 905s? 3c90x in my experience is flakier than the old
standby 3c59x, so you might try switching. And report this problem to
the driver's maintainer if you haven't already done so....
--
Matt G|There is no Darkness in Eternity/But only Light too dim for us to see
Brainbench MVP for Linux Admin / Workin' in a code mine, hittin' Ctrl-Alt
http://www.brainbench.com / Workin' in a code mine, whoops!
=============================/ I hit a seg fault....
------------------------------
Date: Sun, 03 Dec 2000 18:16:09 -0500
From: Drew cutter <[EMAIL PROTECTED]>
Reply-To: [EMAIL PROTECTED]
Subject: Redhat hat professional 7.0 controller card
looking for a controller card for redhat 7.0 professional server . Raid
? eide ? scsi ?
------------------------------
From: John-Paul Stanford <[EMAIL PROTECTED]>
Subject: Corrupt Graphics, Voodoo 3000, DRI
Date: Sat, 02 Dec 2000 23:13:53 GMT
Hi,
I'm hoping someone can help. I've been having problems setting up the
DRI drives for the Voodoo 3000. I'm running XFree 4.0.1 and the DRI
drives from the 3dfx homepage.
Most games seem to work. Doom legacy works in OpenGl mode with out
problems, so does Heavy Gear solid demo and GLChess. But Unreal
Tournament does not and gltron does not work either.
On unreal the textures of walls seem to be corrupted but everything
else works. It's fast and graphics of the bots are fine. GLtron has a
corrupt walls and floor graphics.
Thanks in advance
JP.
Sent via Deja.com http://www.deja.com/
Before you buy.
------------------------------
From: [EMAIL PROTECTED] (ECStahl)
Date: 02 Dec 2000 23:31:06 GMT
Subject: ftape_calibrate_data_rate failed; ditto 3200
Greetings,
I am hoping that somebody out there in newsland will help me to resolve the
following tape drive problem.
I'm attempting to configure an internal Iomega ditto easy 3200, connected to
floppy cable. Running Linux kernel 2.2.14.SuSE, with ftape v3.04 and tools
v1.07.
I did the following:
# modprobe zftape
# modprobe ftape
# ls -l /dev/qft0
# grep char-major-27 /etc/conf.modules (it was "ftape")
# mt -f /dev/qft0 status
mt: /dev/qft0: Device not configured
# tail -40 /var/log/messages
Dec 2 14:22:41 bombur kernel: (c) 1996-1997 Claus-Justus Heine ([EMAIL PROTECTED])
Dec 2 14:22:41 bombur kernel: QIC-117 driver for QIC-40/80/3010/3020 floppy tas
Dec 2 14:22:41 bombur kernel: Compiled for Linux version 2.2.14
Dec 2 14:22:41 bombur kernel: [000] ftape-init.c (ftape_init) - installing QIC.
Dec 2 14:22:41 bombur kernel: [001] ftape-init.c (ftape_init) - ftape_init @ 0.
Dec 2 14:22:41 bombur kernel: [002] ftape-buffer.c (add_one_buffer) - buffer.
Dec 2 14:22:41 bombur kernel: [003] ftape-buffer.c (add_one_buffer) - buffer.
Dec 2 14:22:41 bombur kernel: [004] ftape-buffer.c (add_one_buffer) - buffer.
Dec 2 14:22:41 bombur kernel: [005] ftape-calibr.c (time_inb) - inb() durati.
Dec 2 14:22:41 bombur kernel: [006] ftape-calibr.c (ftape_calibrate) - TC for.
Dec 2 14:22:41 bombur kernel: [007] ftape-calibr.c (ftape_calibrate) - TC for.
Dec 2 14:22:41 bombur kernel: zftape for ftape v3.04d 25/11/97
Dec 2 14:22:41 bombur kernel: (c) 1996, 1997 Claus-Justus Heine ([EMAIL PROTECTED])
Dec 2 14:22:41 bombur kernel: vfs interface for ftape floppy tape driver.
Dec 2 14:22:41 bombur kernel: Support for QIC-113 compatible volume table, dynn
Dec 2 14:22:41 bombur kernel: and builtin compression (lzrw3 algorithm).
Dec 2 14:22:41 bombur kernel: Compiled for Linux version 2.2.14
Dec 2 14:22:41 bombur kernel: [008] zftape-init.c (zft_init) - zft_init @ 0xc4.
Dec 2 14:22:41 bombur kernel: [009] zftape-init.c (zft_init) - installing zfta.
Dec 2 14:28:47 bombur kernel: [010] fdc-io.c (fdc_config) - fdc base: 0x3f.
Dec 2 14:28:47 bombur kernel: [011] fdc-io.c (fdc_probe) - Type 8272A/765A.
Dec 2 14:28:49 bombur kernel: [012] ftape-ctl.c (ftape_get_drive_status) - .
Dec 2 14:28:49 bombur kernel: [013] ftape-io.c (ftape_report_error) - No e.
Dec 2 14:28:49 bombur kernel: [014] ftape-ctl.c (ftape_get_drive_status) - .
Dec 2 14:28:50 bombur kernel: [015] ftape-ctl.c (ftape_log_vendor_id) - tap.
Dec 2 14:28:50 bombur kernel: [016] ftape-io.c (ftape_report_configuration.
Dec 2 14:28:50 bombur kernel: [017] ftape-ctl.c (ftape_init_drive) - post Q.
Dec 2 14:28:50 bombur kernel: [018] ftape-ctl.c (ftape_calibrate_data_rate.
Dec 2 14:28:51 bombur kernel: [019] ftape-io.c (ftape_report_configurati.
Dec 2 14:28:51 bombur kernel: [020] ftape-io.c (ftape_set_data_rate) - Ra.
Dec 2 14:28:51 bombur kernel: [021] ftape-io.c (ftape_set_data_rate) - fd.
Dec 2 14:28:51 bombur kernel: [022] ftape-ctl.c (ftape_calibrate_data_rate.
Dec 2 14:28:51 bombur kernel: [023] ftape-ctl.c (ftape_init_drive) - ftape_.
Dec 2 14:28:51 bombur kernel: [024] ftape-ctl.c (ftape_enable) - ftape_init_.
Dec 2 14:28:51 bombur kernel: [025] zftape-ctl.c (_zft_open) - ftape_enable f.
Dec 2 14:28:51 bombur kernel: [026] zftape-init.c (zft_open) - _ftape_open fai.
My next Linux system will of course have SCSI, but I'd like to be able
to back up my current inexpensive training system to the ditto 3200. My ditto
H/W
worked fine under the Windows OS that used to be on the PC - though I had the
accelerator card in use then and not the floppy connection. Had heard somewhere
that the floppy connection was to be preferred.
Thank you in advance for any help you might offer.
Eric Stahl
------------------------------
** 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
******************************
