Linux-Hardware Digest #263, Volume #13 Wed, 19 Jul 00 05:13:05 EDT
Contents:
Re: Firewall/Gateway hardware requirements (Chris)
UPS suggestions? (James Kufrovich)
Re: MAXTOR 27G IDE, how to partition? ("Karl E. Jorgensen")
Raid Card .... Adaptec AAA131/U2 support under RedHat Linux 6.2 (2.2.16)? (Phil
Gordon)
Geforce GTS and Linux. ("Adan Morales")
Re: Solved? (was: Asus P2B-D, dual PIII-650MHz, SMP not working right) (Greg
Werstiuk)
Re: MAXTOR 27G IDE, how to partition? (hac)
Re: Problems with Promise PDC20267 RAID Controller (Tim Moore)
SCSI hangs with 2.2.1(4|6) on heavy I/O (Chris Adams)
partition gets corrupted on reboot (Ralph Gauges)
Re: This Adaptec SCSI bug ("James A Wilde")
Re: Question about cdrdao ([EMAIL PROTECTED])
Re: SCSI hangs with 2.2.1(4|6) on heavy I/O (Linus Torvalds)
Re: Geforce GTS and Linux. (smp root)
Re: Intel i810 chipset, AC97 sound driver ("Rob Sykes")
Re: This Adaptec SCSI bug (Wolfgang Zweimueller)
Re: Netgear FA-310TX with Tulip chip (sideband)
sound and Mandrake 7.1 ("Nick Bull")
Re: SCSI hangs with 2.2.1(4|6) on heavy I/O (Chris Adams)
----------------------------------------------------------------------------
From: [EMAIL PROTECTED] (Chris)
Subject: Re: Firewall/Gateway hardware requirements
Date: Wed, 19 Jul 2000 05:14:35 GMT
On Tue, 18 Jul 2000 15:08:11 -0500, Jason Loll <[EMAIL PROTECTED]>
wrote in comp.os.linux.hardware:
>Could you please give me some names of hardware routers so I could do more research?
>Also, a dumb quiestion, do you need a modem for the router, or how do you connect it
>to your computer?
The router sits between the cable/DSL modem and the local network
segments.
I am using a LinkSys BEFSR41, which is a cable/DSL firewall integrated
into a 4-port 10/100BaseT switch. It has NAT support (up to 253 local
addresses) and limited port forwarding capability.
See: http://www.linksys.com/products/product.asp?prid=20&grid=5
------------------------------
From: [EMAIL PROTECTED] (James Kufrovich)
Subject: UPS suggestions?
Reply-To: [EMAIL PROTECTED]
Date: Wed, 19 Jul 2000 05:28:06 GMT
Hi. I'm thinking of getting a UPS for my Linux box, since we've
been getting short (a few seconds) power outages around here recently, and
I'm just wondering what all to look for. Linux compatibility would
certainly be a plus. A serial port connection is a necessity for
automated shutdowns, if necessary (gotta get COM3 or 4 working, as 1 and 2
are taken already.) I'd only have my PC and monitor plugged into it,
nothing else. What load should I look for? 500VA like someone suggested
earlier? And I'd only really need a few minutes of juice, as, like I said
before, the power outages don't last long at all - I'd only need enough
time to do a safe shutdown myself, or for an automated shutdown.
A friend recommended the PK Electronics Blackout Buster (he run
Windows), but I don't know if that would work well with Linux? Something
from APC? Tripp Lite? Someone here suggested MGE? I'd appreciate any
tips. Thanks.
Jamie Kufrovich
--
Egg, [EMAIL PROTECTED]
FMSp3a/MS3a A- C D H+ M+ P+++ R+ T W Z+
Sp++/p# RLCT a+ cl++ d? e++ f h* i+ j p+ sm+
------------------------------
From: "Karl E. Jorgensen" <[EMAIL PROTECTED]>
Subject: Re: MAXTOR 27G IDE, how to partition?
Date: Tue, 18 Jul 2000 22:45:51 +0100
Jim White wrote:
>
> I want use half of this MAXTOR 27G IDE for window, half for linux.
> Is it possible? the "MAX BLAST PLUS" from MAXTOR only work for window.
At that size harddisk you're bound to run against the 1024 cylinder (or
is it 1023 cylinder) limit. That is: Whatever you're booting from has to
be within the first 1024 cylinders. Nothing to do with Linux - it's
BIOS, it's legacy, it's ancient, but true.
There _is_ a way around it though; if you create a small partition
('bout 1 meg should do it) close to the beginning of the disk (i.e.
below cylinder 1024), and put the kernel there, you should be able to
get started. Typically, you would mount that partition as /boot, and
tell the kernel to use some other partition (/dev/hda??) as the root
file system. Once the kernel is up and running, you shouldn't suffer
from the 1024 cylinder limit.
PS: Have a look at the BigDisk-Howto as well.
Hope this helps
--
Karl E. Jorgensen
------------------------------
From: [EMAIL PROTECTED] (Phil Gordon)
Subject: Raid Card .... Adaptec AAA131/U2 support under RedHat Linux 6.2 (2.2.16)?
Date: Wed, 19 Jul 2000 06:02:32 GMT
Hello...
Does anyone know if there is support for the Adaptec AAA131/U2 Raid
Card under RedHat Linux 6.2 (2.2.16)??
I have looked on Adaptec's site for information and could not seem to
find anything that said it would work.
This URL is the closest I could get for driver support:
http://www.adaptec.com/support/overview/aaa130u2.html#drivers
I checked on RedHat's site, and it is listed in their Hardware
Compability Listing as the following:
The following SCSI adapters are compatible with Red Hat Linux, but are
not supported: AAA-13x Raid Adapters and AAA-113x Raid Port Card
http://www.redhat.com/support/hardware/intel/62/rh6.2-hcl-i.ld-5.html
What I have tried:
- Installed the card, and Linux recognizes the Chipset, but not the
card. ... basically I can use the DOS Array Config utility to setup
the Array, but the OS still sees the individual disks.
What I am looking for:
- any information on how to get it working, beta drivers, etc.
------------------------------
From: "Adan Morales" <[EMAIL PROTECTED]>
Subject: Geforce GTS and Linux.
Date: Wed, 19 Jul 2000 01:21:55 -0500
Can I assume that the nVidia Linux drivers for the Geforce 256 will work
with a Geforce 2 GTS and XFree86 4.0? Has anybody got one to work yet?
Thanks.
------------------------------
From: Greg Werstiuk <[EMAIL PROTECTED]>
Subject: Re: Solved? (was: Asus P2B-D, dual PIII-650MHz, SMP not working right)
Date: Tue, 18 Jul 2000 23:24:41 -0700
Crossposted-To: alt.comp.periphs.mainboard.asus,alt.comp.hardware
Dan -
I see many people discussing "older motherboards cannot supply the
correct CPU voltage of 1.65V (can't supply below 1.8V)" as the reason
one cannot run Coppermine processors on P2B series MoBos.
ASUS states many times in their damned impossibly slow to use, NetQ
replacement for their newsgroup, that the P2B family does not support
PIII processors faster than 600MHz. They do *not* state it is due to
the CPU voltage issue. In fact, they do not state the cause, ever.
I recently tried running a late production P2B-D rev 1.05 at 650MHz.
The late version can and does provide the correct CPU voltage.
With either one or two 450MHz PII processors, the system is and has
been always rock solid.
Testing with a 500MHz 512K cache PIII, system is rock solid.
Using either one or two 650MHZ PIII 256K cache processors, the system
boots, one can access the BIOS settings, etc. However, one cannot run
NT. Hit a BSOD every boot. In addition, the MoBo reset switch
circuitry no longer functions correctly.
Some other issue exists with the P2B family motherboards and PIII
processors and ASUS is not talking about it.
- Greg
On Sat, 15 Jul 2000 16:29:02 -0700, Dan Kegel
<[EMAIL PROTECTED]> wrote:
>Some progress to report:
>
>It seems that Red Hat 6.2 has some sort of problem with the Asus P2B-D.
>After fiddling around, I noticed that 'make -j4 bzImage' would choke
>even if I was booted single processor. Then I noticed that if
>I installed the vanilla 2.2.16 + alan cox's combo errata patch
>and configured a very simple kernel, the problem went away.
>
>I've build successfully three times now under various stressful
>conditions. I can't be sure the problem is gone for a week or so
>of stress testing, I suppose, but I'm quite hopeful.
>- Dan
>
>Dan Kegel wrote:
>> Just bought a system with an Asus P2B-D and two Pentium III-650's.
>> Installed Red Hat 6.2. When I compile the Linux kernel (a big job
>> that stresses the computer pretty well for two or three minutes) with
>> make -j2 bzImage
>> it fails on a different source file each time, complaining of
>> an assembler error. Rerunning 'make -j2 bzImage' immediately
>> causes it to compile the same file without errors and move on.
>> This is the classic sign of a hardware problem -- but it appears
>> to be specific to having both processors actively working and
>> using the disk at the same time.
>>
>> I can reproduce the problem three ways:
>> 1. boot SMP with all 32 MB RAM, run X, and 'make -j2 bzImage'
>> 2. boot SMP with 'linux mem=24M', run in console mode, and 'make -j2 bzImage'
>> 3. boot SMP with all 32 MB RAM, run in console mode, and 'make -j4 bzImage'
>> It's very reproducible with the first two methods, and fairly so
>> with the third.
>> Booting uniprocessor and doing 'make -j2 bzImage' doesn't do it,
>> booting SMP and running 'make bzImage' doesn't do it,
>> booting SMP in console mode with all 32 MB RAM and running 'make -j2 bzImage'
>doesn't do it.
>>
>> ...
>> gcc -D__KERNEL__ -I/usr/src/linux-2.2.14/include -Wall -Wstrict-prototypes -O2
>> -fomit-frame-pointer -fno-strict-aliasing -D__SMP__ -pipe -fno-strength-reduce
>-m386 -DCPU=386 -c -o buffer.o buffer.c
>> gcc -D__KERNEL__ -I/usr/src/linux-2.2.14/include -Wall -Wstrict-prototypes -O2
>> -fomit-frame-pointer -fno-strict-aliasing -D__SMP__ -pipe -fno-strength-reduce
>-m386 -DCPU=386 -c -o scsi.o scsi.c
>> {standard input}: Assembler messages:
>> {standard input}:1699: Error: invalid character '_' in mnemonic
>> {standard input}:1699: Error: Rest of line ignored. First ignored character is ` '.
>> make[2]: *** [buffer.o] Error 1
------------------------------
From: hac <[EMAIL PROTECTED]>
Subject: Re: MAXTOR 27G IDE, how to partition?
Date: Wed, 19 Jul 2000 06:31:40 GMT
"Karl E. Jorgensen" wrote:
>
> Jim White wrote:
> >
> > I want use half of this MAXTOR 27G IDE for window, half for linux.
> > Is it possible? the "MAX BLAST PLUS" from MAXTOR only work for window.
>
> At that size harddisk you're bound to run against the 1024 cylinder (or
> is it 1023 cylinder) limit. That is: Whatever you're booting from has to
> be within the first 1024 cylinders. Nothing to do with Linux - it's
> BIOS, it's legacy, it's ancient, but true.
>
> There _is_ a way around it though; if you create a small partition
> ('bout 1 meg should do it) close to the beginning of the disk (i.e.
> below cylinder 1024), and put the kernel there, you should be able to
> get started. Typically, you would mount that partition as /boot, and
> tell the kernel to use some other partition (/dev/hda??) as the root
> file system. Once the kernel is up and running, you shouldn't suffer
> from the 1024 cylinder limit.
>
Believe it or not, you may need a bit more than 1MB. Red Hat, for
some reason unknown to me, likes to install an uncompressed kernel, as
well as the compressed one that you actually use. And the system map,
etc. 4MB seems to be the minimum for Red Hat; others may vary.
Any drive with more than 1024 cylinders can spare the room for a /boot
partition of several MB. I keep my current production kernel, the
previous revision, and the occasional development kernel there, so
more space is needed. A monolithic 2.4.0-test kernel runs close to a
meg by itself.
--
Howard Christeller Irvine, CA [EMAIL PROTECTED]
------------------------------
From: Tim Moore <[EMAIL PROTECTED]>
Subject: Re: Problems with Promise PDC20267 RAID Controller
Date: Wed, 19 Jul 2000 06:41:31 GMT
The Promise controller family (U/33, U/66, U/100, FastTrak) all share nearly
identical hardware and BIOS, which is why the FastTrak booted exactly like a
U/100 controller.
FastTrak RAID functions are primarily implemented in the software driver
which is why this card is significantly cheaper than real hardware RAID
controllers by Mylex, AMI, etc. Promise has threatened to port their
drivers to linux any moment now for about a year. Currently non-hardware
RAID 0,1,4,5 is available through kernel raid options and raidtools.
Reading assignments:
http://www.linuxdoc.org/HOWTO/Software-RAID-HOWTO.html
or
http://www.linuxdoc.org/cgi-bin/ldpsrch.cgi (type in 'RAID')
http://people.redhat.com/mingo/
ftp.us.kernel.org/pub/linux/daemons/raid/
Jon Morby wrote:
>
> I'm trying to get one of the new Fastrak/100 ATA RAID controllers working
> under 2.4.0-test4 ...
>
> After much fiddling and searching of faq's I've managed to get the kernel to
> see the controller by adding append="ide4=0x9800,0x9c02 ide5=0xa000,0xa402"
> to my lilo.conf but it now sees the drives as seperate drives, even though
> they're configured as a stripe on the controller ...
> ...
--
timothymoore
bigfoot
com
------------------------------
From: [EMAIL PROTECTED] (Chris Adams)
Crossposted-To: comp.os.linux.redhat
Subject: SCSI hangs with 2.2.1(4|6) on heavy I/O
Reply-To: [EMAIL PROTECTED]
Date: Wed, 19 Jul 2000 06:54:19 GMT
I ran into a major problem with a development system here (K6-III400/128MB). It
has an Initio-9100UW which has a Quantum Atlus IV on the internal interface and
a Seagate ST423451WEXT on the external interface. Everything is terminated
properly, the card recognizes it just fine, Linux recognizes it just, etc.
The problem was that disk I/O would frequently cause the system to hang. First
there'd be some timeouts and then the familar reset/trying harder cycle which
would lock the system up & require a power-cycle. There was more than just pure
disk I/O as I had transfered over several gigabytes of webserver log files
without any problems. I had a reliable kill, however, with a Perl script which
ran wusage (a webserver log analysis program) for a number of sites:
1) Get the next site to run the log reports
2) Use wget to retrieve logfiles for that site
3) Fork: the child used exec() to run wusage; the parent returned to
step #1
This would rapidly hang the system as soon after the first fork. I think this
had something to do with the combination of simultaneous heavy read and write
activity.
Based on a fair amount of time in the linux-kernel & linux-scsi lists and
dejanews.com, I'm starting to reach the conclusion that SCSI in 2.2.14+ isn't
reliable as there appear to be a number of similar posts for a variety of SCSI
adapters.
I went from 2.2.14 to 2.2.16 (both Redhat's RPM and a manual compile with the
Initio support compiled into the kernel instead of as a module) with no change.
Since I had little left to lose, I installed 2.4.0-test4. Build & install was
seamless and the system has yet to so much as hiccup.
Chris
------------------------------
From: Ralph Gauges <[EMAIL PROTECTED]>
Subject: partition gets corrupted on reboot
Date: Wed, 19 Jul 2000 09:09:06 +0200
Hi,
I am having a slight problem with my hardisk settings. I
have a system with three IDE harddisks
1. hda: Maxtor 40GB HD (48??/255/63)
2. hdc: WD136BA 13,6 GB HD (26500/16/63)
3. hdd: IBM DTA 10,1GB HD (19???/16/63)
I just added the 40GB disk, before that the syste was
running fine with the other two disks a masters.
hda and hdd are still well behaved, but hdc is causing me
headaches now. I just created one partition on the disk from
cyl 1 to 9000 (~4.5 GB) the rest of the disk is
unpartitioned, but whenever I reboot the partition table
seems to get corrupted somehow. After reboot fdisk report
wrong settings and that the partition does not end on the
cylinder boundry. When I copy the saved bootsektor back to
hdc, it is O.K. until the next reboot.
This disk worked before, so there has to be a way to get it
working again, but I am lost.
I already tried different bios setting, but to no avail so
far.
Does anybody know where the partition table gets corrupted?
Or could somebody even tell me how to eliminate this
problem?
Thanks a lot
Ralph
------------------------------
Reply-To: "James A Wilde" <[EMAIL PROTECTED]>
From: "James A Wilde" <[EMAIL PROTECTED]>
Subject: Re: This Adaptec SCSI bug
Date: Wed, 19 Jul 2000 09:09:46 +0200
What mobo, Wolfgang?
Wolfgang Zweimueller <[EMAIL PROTECTED]> wrote in message
news:[EMAIL PROTECTED]...
> Hmm. The DC-390U2W runs fine here! SCSI-Tape, -Disk and -CDROM. Never
> had a problem. The only annoying thing is that it is kinda slow on
> scanning the bus at the boot stage.
>
> BTW, I run the 390 with the NCDR53C8XX-driver under Linux 2.2.16 .
------------------------------
From: [EMAIL PROTECTED]
Crossposted-To: comp.os.linux
Subject: Re: Question about cdrdao
Date: Wed, 19 Jul 2000 06:57:01 GMT
In article <[EMAIL PROTECTED]>,
Andras <[EMAIL PROTECTED]> wrote:
> Does anyone know whether Andreas Mueller's cdrdao can handle
> .WAV files (with header) correctly or do I have to convert them to
raw?
>
You can use .wav files.
Andreas
Sent via Deja.com http://www.deja.com/
Before you buy.
------------------------------
From: [EMAIL PROTECTED] (Linus Torvalds)
Crossposted-To: comp.os.linux.redhat
Subject: Re: SCSI hangs with 2.2.1(4|6) on heavy I/O
Date: 19 Jul 2000 00:11:32 -0700
In article <[EMAIL PROTECTED]>,
Chris Adams <[EMAIL PROTECTED]> wrote:
>
>I went from 2.2.14 to 2.2.16 (both Redhat's RPM and a manual compile with the
>Initio support compiled into the kernel instead of as a module) with no change.
>Since I had little left to lose, I installed 2.4.0-test4. Build & install was
>seamless and the system has yet to so much as hiccup.
Btw, thanks. It's sometimes quite depressing to only get bug-reports
for the new kernels. Good to get a virtual pat on the shoulder every
once in a while,
Linus
------------------------------
Subject: Re: Geforce GTS and Linux.
From: smp root <[EMAIL PROTECTED]>
Date: Wed, 19 Jul 2000 00:21:23 -0700
yes, i've been running one for two months, even before the
drivers were supposed to recognise it (the 0.9-2 drivers make no
mention of it, even in /var/log/XFree.0.log)
however the new 0.9-4 drivers available on their site are very
good. they offer FSAA, even though some games appear not to
like it. it works well and is spectacularly fast.
===========================================================
Got questions? Get answers over the phone at Keen.com.
Up to 100 minutes free!
http://www.keen.com
------------------------------
From: "Rob Sykes" <[EMAIL PROTECTED]>
Subject: Re: Intel i810 chipset, AC97 sound driver
Date: Wed, 19 Jul 2000 09:22:43 +0100
"seumas" <[EMAIL PROTECTED]> wrote in message
news:963979566.618509@news...
> I'm running oss from http://www.opensound.com/ and it works perfectly on
the
> i810. It does ask for registration and about $30 but seems to run while
> just giving me warnings. Very easy install.
>
Many thanks - I've just downloaded the evaluation and will be hacking away
this evening.
All the best to you
--
Rob Sykes
Remove NOSPAM to e-mail
------------------------------
Subject: Re: This Adaptec SCSI bug
From: Wolfgang Zweimueller <[EMAIL PROTECTED]>
Date: 19 Jul 2000 10:07:19 +0200
"James A Wilde" <[EMAIL PROTECTED]> writes:
> What mobo, Wolfgang?
Abit KA7.
cu, Wolfgang.
------------------------------
From: sideband <[EMAIL PROTECTED]>
Subject: Re: Netgear FA-310TX with Tulip chip
Date: Wed, 19 Jul 2000 04:59:13 -0400
Are you sure it's a switch in the DSL router? That sounds a bit off.
Most of the switches/routers, such as the NETGEAR ISDN routers, etc,
come with a 4 port HUB built in, AFAIK. Make sure the Linksys has a
switch in it before forcing full duplex mode. If the FA310TX didn't
detect full duplex, then the Linksys probably isn't capable of it, or
it's a manual setting at the DSL router.
At any rate, luck to you. Let me know how it turns out.
-SSB
heyday wrote:
> Thanks for the info. I have my Netgear Cards hooked up to the
> new Linksys Cable/DSL Router. It is a 10/100 Switch and works at
> full Duplex. I will check my cabling also. I thout it was
> Cat-5.
>
> Thanks,
>
> Heyday
>
> http://www.phone4less.nu
>
> -----------------------------------------------------------
>
> Got questions? Get answers over the phone at Keen.com.
> Up to 100 minutes free!
> http://www.keen.com
------------------------------
From: "Nick Bull" <[EMAIL PROTECTED]>
Subject: sound and Mandrake 7.1
Date: Wed, 19 Jul 2000 10:00:38 +0100
I've just put Mandrake 7.1 onto my system at home and I'm having a problem
getting the sound to work. The sound card (VIA 686 on board) is detected
during the boot up, and when I do dmesg it says (and I'm going from memory
here) something about There are other ways to configure this card, but if it
really is a SB Pro, then leave it as it is.
The problem is that I can't get any sound to work. I've made sure the
volume is turned up in the mixers and nothing is muted. Sndconfig detects
the VIA on board sound, then it says it will play the sound, but I never
hear anything.
Anyone know what I can do to solve this ? AFAIK, the sound that comes with
Mandrake 7.1 is OSS/Free.
Cheers,
Nick.
------------------------------
From: [EMAIL PROTECTED] (Chris Adams)
Crossposted-To: comp.os.linux.redhat
Subject: Re: SCSI hangs with 2.2.1(4|6) on heavy I/O
Reply-To: [EMAIL PROTECTED]
Date: Wed, 19 Jul 2000 09:08:21 GMT
On 19 Jul 2000 00:11:32 -0700, Linus Torvalds <[EMAIL PROTECTED]>
wrote:
>In article <[EMAIL PROTECTED]>,
>Chris Adams <[EMAIL PROTECTED]> wrote:
>>
>>I went from 2.2.14 to 2.2.16 (both Redhat's RPM and a manual compile with the
>>Initio support compiled into the kernel instead of as a module) with no change.
>>Since I had little left to lose, I installed 2.4.0-test4. Build & install was
>>seamless and the system has yet to so much as hiccup.
>
>Btw, thanks. It's sometimes quite depressing to only get bug-reports
>for the new kernels. Good to get a virtual pat on the shoulder every
>once in a while,
Uhoh. Now I feel pretty bad mentioning that I just managed to get 2.4.0-test4
to hang, even though it took ~30 simultaneous heavy I/O processes to do it (vs.
one before). I'll have to see if I can track down just where it's going wrong.
Fortunately, this system won't run into that sort of load under normal usage.
------------------------------
** 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
******************************
