Linux-Hardware Digest #168, Volume #14 Fri, 12 Jan 01 18:13:06 EST
Contents:
Re: Via on-board sound success! (Peter Christy)
Hardware detection sequence (Ragnar =?ISO-8859-1?Q?Wisl=F8ff?=)
Re: Hardware detection sequence ("D. Stimits")
Re: Linux IDE RAID Cards ("Aaron R. Kulkis")
Re: Linux on HP Pavilion 8670C (Jack Kelley)
Re: L2-Cache of Pentium2 with Linux ("Peter T. Breuer")
Re: Strange alpha clock issue (Marble Head)
Re: L2-Cache of Pentium2 with Linux ("Peter T. Breuer")
lost interrupt? (James Tappin)
intel express 100 known to the kernel but can't be identified? ("Alexander
Deztroyer")
Re: Linux IDE RAID Cards ("Steve Wolfe")
Re: Linux IDE RAID Cards (Chris Lopeman)
----------------------------------------------------------------------------
From: Peter Christy <[EMAIL PROTECTED]>
Subject: Re: Via on-board sound success!
Date: Fri, 12 Jan 2001 18:26:06 +0000
My only knowledge of sound has come from trying to get my systems working!
I know how to get ESS cards going, and CMI 8330s - I now also know how to
do the Via chips, and these have proved the most problematic.
My approach has always been to try the drivers built in to the kernel first
- you really need to know how to compile your own kernel to do this (not a
bad thing to know, anyway!). Do you know what drivers you are using? Do you
know what other drivers exist on your system? How are you calling them?
When building a kernel, don't put in things you don't need. Keep it as
small and light as practical. Try and do as much as possible as modules,
rather than set in stone in the kernel. If you can't get the modular
drivers to work, as a last resort try the Alsa drivers - those are what
finally worked for me with the Via chips. /etc/rc.local is one place to try
and load modules at boot time.
Sorry to sound a bit vague, but your question is really the equivalent to
the one about the length of a piece of string! Before you can start to
answer it, you need to know what you have in the first place.
One other thought, go into the bios on your system and make sure you
haven't got the onboard sound enabled - if you're using a seperate sound
card, it could be conflicting.
--
Pete
[EMAIL PROTECTED]
------------------------------
From: Ragnar =?ISO-8859-1?Q?Wisl=F8ff?= <[EMAIL PROTECTED]>
Subject: Hardware detection sequence
Date: Fri, 12 Jan 2001 21:11:27 +0100
I'm having trouble with a system with one onboard (for internal disks) and
one PCI SCSI controller (for external disks). It runs RHL 6.2. Before a
kernel switch from 2.2.14 to 2.2.16 all was fine, the internal disks came
up as sdax and the external ones as sdbx. But now, running 2.2.16, the
order is swapped, and the kernel panics as it cannot mount the root file
system. It looks for sdaN, but it really is on sdbN the way the h/w is
detected.
I've seen a thread some time ago regarding the order of detection of NICs,
and Donald Becker strongly advocating onboards last. That's OK with me, but
did Linus change his mind halfway through a kernel series? That cannot be
true.
I'm stuck with drivers a modules as they exist as binaries only (no naming
of the vendor here ...). Lilo loads an initrd image to get at the modules,
and I've tried to change the order of the module lines in
/etc/conf.modules. Also played with putting prepended numbers after the
scsi_hostadapater (like eth0, eth1 etc) before running mkinitrd. Now I'm
stuck. The thing insists on finding the external one before the internal
controller.
Is possible to force the kernel to detect the h/w (or load modules) in a
specific order?
If not, how does it decide in what order to do it?
Is there another way to solve my problem?
Thanks,
--
Ragnar Wisl�ff
==========
life is a reach. then you gybe.
------------------------------
Date: Fri, 12 Jan 2001 13:15:22 -0700
From: "D. Stimits" <[EMAIL PROTECTED]>
Reply-To: [EMAIL PROTECTED]
Subject: Re: Hardware detection sequence
Most bios setups for motherboards that have onboard devices have a
setting for "boot offboard devices first" or "reverse boot order" or
"reverse pci scan". Check the bios settings and reverse whatever it is
now.
Ragnar Wisl�ff wrote:
>
> I'm having trouble with a system with one onboard (for internal disks) and
> one PCI SCSI controller (for external disks). It runs RHL 6.2. Before a
> kernel switch from 2.2.14 to 2.2.16 all was fine, the internal disks came
> up as sdax and the external ones as sdbx. But now, running 2.2.16, the
> order is swapped, and the kernel panics as it cannot mount the root file
> system. It looks for sdaN, but it really is on sdbN the way the h/w is
> detected.
>
> I've seen a thread some time ago regarding the order of detection of NICs,
> and Donald Becker strongly advocating onboards last. That's OK with me, but
> did Linus change his mind halfway through a kernel series? That cannot be
> true.
>
> I'm stuck with drivers a modules as they exist as binaries only (no naming
> of the vendor here ...). Lilo loads an initrd image to get at the modules,
> and I've tried to change the order of the module lines in
> /etc/conf.modules. Also played with putting prepended numbers after the
> scsi_hostadapater (like eth0, eth1 etc) before running mkinitrd. Now I'm
> stuck. The thing insists on finding the external one before the internal
> controller.
>
> Is possible to force the kernel to detect the h/w (or load modules) in a
> specific order?
> If not, how does it decide in what order to do it?
>
> Is there another way to solve my problem?
>
> Thanks,
>
> --
> Ragnar Wisl�ff
> ----------
> life is a reach. then you gybe.
------------------------------
From: "Aaron R. Kulkis" <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.misc,comp.os.linux.advocacy
Subject: Re: Linux IDE RAID Cards
Date: Fri, 12 Jan 2001 15:16:36 -0500
Chris Lopeman wrote:
>
> Can anyone recommend a good raid IDE controller for Linux. Preferably
> ATA 100. We have tried using the Promise controller with limited
> success. We probably want to run 2 controllers in the server (for more
> speed) with a total of 6 drives. 4 of the drives running raid 1+0 and
> the other 2 forming a separate mirror. The 2 in the mirror we also want
>
> to boot off of.
>
> If you can't recommend a good one maybe you can let us know your
> experience with the AMI or Escalade cards. We are considering going to
> one of these.
http://www.research.att.com/~gjm/linux/ide-raid.html
>
> Thanks,
>
> Chris Lopeman
> Object Link Inc.
--
Aaron R. Kulkis
Unix Systems Engineer
DNRC Minister of all I survey
ICQ # 3056642
H: "Having found not one single carbon monoxide leak on the entire
premises, it is my belief, and Willard concurs, that the reason
you folks feel listless and disoriented is simply because
you are lazy, stupid people"
I: Loren Petrich's 2-week stubborn refusal to respond to the
challenge to describe even one philosophical difference
between himself and the communists demonstrates that, in fact,
Loren Petrich is a COMMUNIST ***hole
J: Other knee_jerk reactionaries: billh, david casey, redc1c4,
The retarded sisters: Raunchy (rauni) and Anencephielle (Enielle),
also known as old hags who've hit the wall....
A: The wise man is mocked by fools.
B: Jet Silverman plays the fool and spews out nonsense as a
method of sidetracking discussions which are headed in a
direction that she doesn't like.
C: Jet Silverman claims to have killfiled me.
D: Jet Silverman now follows me from newgroup to newsgroup
...despite (C) above.
E: Jet is not worthy of the time to compose a response until
her behavior improves.
F: Unit_4's "Kook hunt" reminds me of "Jimmy Baker's" harangues against
adultery while concurrently committing adultery with Tammy Hahn.
G: Knackos...you're a retard.
------------------------------
From: Jack Kelley <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.setup
Subject: Re: Linux on HP Pavilion 8670C
Date: Fri, 12 Jan 2001 20:57:46 GMT
"Kent R. Frazier" wrote:
>
> Has anyone installed Linux on an HP Pavilion 8670C Desktop machine?
>
> My Linux box has died, and I would like to make my Windows machine a
> dual-boot until I can afford to replace / rebuild my Linux box.
>
> I'm looking for a fairly easy install, as I currently don't have the time to
> trouble shoot hardware right now.
>
> Thanks,
>
> Kent
Kent,
I have a HP Pavilion 8660C
It works great with redhat 7.0 since the intel 810 video board is
supported however I had to replace that Conexant Chamelion peice of crap
because the drivers cost 30 bucks. At comp usa they sell a CMI (Muse
by Guillemot) based board that is fully supported. besides messing with
the scsi emulation stuff for the CDR is was pretty straight foward.
Good Luck,
Jack
------------------------------
From: "Peter T. Breuer" <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.misc
Subject: Re: L2-Cache of Pentium2 with Linux
Date: Fri, 12 Jan 2001 21:47:38 +0100
In comp.os.linux.hardware fred smith <[EMAIL PROTECTED]> wrote:
> Jean-David Beyer <[EMAIL PROTECTED]> wrote:
> : In the little test, the dropoff as I exceeded the size of the L1 and L2
> : caches was extremely well-pronounced, even though the machine was 0%
> Just for kicks I ran it on my K6-2/350 (RH 6.2) and got what seem like
> interesting results. I've got a FIC VA503+ motherboard which purports
> to have a 1meg L2 cache. here's the test results:
> *** MEMORY WRITE PERFORMANCE (256 MB LOOP) ***
> size = 64 bytes: 853.333 MB/s
> size = 128 bytes: 948.148 MB/s
> size = 256 bytes: 984.615 MB/s
> size = 512 bytes: 1024.000 MB/s
> size = 1024 bytes: 984.615 MB/s
> size = 2048 bytes: 1024.000 MB/s
> size = 4096 bytes: 1066.667 MB/s
> size = 8192 bytes: 1024.000 MB/s
> size = 16384 bytes: 1024.000 MB/s
> size = 32768 bytes: 984.615 MB/s
> size = 65536 bytes: 193.939 MB/s <======== This makes sense, it's a
> size = 131072 bytes: 193.939 MB/s 64kb L1 cache
> size = 262144 bytes: 148.837 MB/s
> size = 524288 bytes: 126.108 MB/s
> size = 1048576 bytes: 100.787 MB/s <===== The drop-off around 1 meg
> size = 2097152 bytes: 80.503 MB/s doesn't seem to be any greater than
> size = 4194304 bytes: 75.294 MB/s the drop-offs that occur in a gradation
> size = 8388608 bytes: 74.419 MB/s above it.
Indeed, I see no evidence of an L2 cache. The results are roughly consistent with
there
not being one. I'm not prepared to go into a second order analysis!
By the way, I have an old P100 on which the results are flat at 20MB/s as far as the
eye
can see. I suspect it has a broken mobo cache.
> *** MEMORY READ PERFORMANCE (256 MB LOOP) ***
> size = 16384 bytes: 1163.636 MB/s
> size = 32768 bytes: 1113.043 MB/s
> size = 65536 bytes: 400.000 MB/s <===== Again, this looks right.
> size = 131072 bytes: 400.000 MB/s
> size = 262144 bytes: 290.909 MB/s
> size = 524288 bytes: 241.509 MB/s
> size = 1048576 bytes: 186.861 MB/s <==== more obvious dropoff here than
> size = 2097152 bytes: 147.126 MB/s the one above, but not major.
You'd have to graph it to see if the slope changes here. You might have a read cache.
> size = 4194304 bytes: 134.031 MB/s
> size = 8388608 bytes: 132.642 MB/s
> note that there's a sharp drop-off at the 64kb point on both tests,
> but much less sharp at 1 meg, especially on the first test. What I find
> curious is that between 64kb and 1 meg there's a continuous and gradual
> drop-off which seems to continue at about the same rate right on down
> past the 1 meg point, at least in the first test. Does this tell me
> something about the cahce-handling on my motherboard, or about how well
> the K6-2 uses the cache, or maybe something else? Maybe it simply means
> that the L2 cache is of low value for memory writes, less than for reads.
It tells you at least what percentage of your write falls in the L1 cache!
You need to look at the first and second derivative to learn more.
Peter
------------------------------
From: [EMAIL PROTECTED] (Marble Head)
Crossposted-To: comp.os.linux.alpha
Subject: Re: Strange alpha clock issue
Date: 12 Jan 2001 21:37:34 GMT
[EMAIL PROTECTED] (Joakim Roubert) wrote in
<[EMAIL PROTECTED]>:
>On 12 Jan 2001 16:23:04 GMT Marble Head wrote:
>>Sorry 'bout that. No sarcasm intended. There are a lot of users out
>>there who don't know the things that the rest of consider to be basic.
>
>Assuming that a person who asks a question is a complete jerk isn't
>really nice. Then there's enough with the excuse, and no rotten
>explanation which actually say that you're acutally not sorry at all.
>That attitude stinks.
>
>Your original posting was sarcastic, your "excuse" an insult.
>
>/Jokke tmw.
>���
> http://www.efd.lth.se/~d97jro/
>
I didn't assume anybody was a jerk.
I chose NOT to assume that you have experience with linux, unix, or alpha.
You can't tell me when I'm sarcastic or not.
And I'm sorry for insulting you.
I choose to have no further conversation with you on this subject.
------------------------------
From: "Peter T. Breuer" <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.misc
Subject: Re: L2-Cache of Pentium2 with Linux
Date: Fri, 12 Jan 2001 22:28:29 +0100
In comp.os.linux.hardware Peter T. Breuer <[EMAIL PROTECTED]> wrote:
> In comp.os.linux.hardware fred smith <[EMAIL PROTECTED]> wrote:
>> Jean-David Beyer <[EMAIL PROTECTED]> wrote:
>> : In the little test, the dropoff as I exceeded the size of the L1 and L2
>> : caches was extremely well-pronounced, even though the machine was 0%
>> Just for kicks I ran it on my K6-2/350 (RH 6.2) and got what seem like
>> interesting results. I've got a FIC VA503+ motherboard which purports
>> to have a 1meg L2 cache. here's the test results:
Actually, now I look closer, there is evidence for a 32KB cache, and then
a 512KB or a 1MB one. It may be that your cache is not all of it set up to
cache writes (writeback? writethru?)
>> *** MEMORY WRITE PERFORMANCE (256 MB LOOP) ***
>> size = 64 bytes: 853.333 MB/s
>> size = 128 bytes: 948.148 MB/s
>> size = 256 bytes: 984.615 MB/s
>> size = 512 bytes: 1024.000 MB/s
>> size = 1024 bytes: 984.615 MB/s
>> size = 2048 bytes: 1024.000 MB/s
>> size = 4096 bytes: 1066.667 MB/s
>> size = 8192 bytes: 1024.000 MB/s
>> size = 16384 bytes: 1024.000 MB/s
>> size = 32768 bytes: 984.615 MB/s
>> size = 65536 bytes: 193.939 MB/s <======== This makes sense, it's a
>> size = 131072 bytes: 193.939 MB/s 64kb L1 cache
>> size = 262144 bytes: 148.837 MB/s
>> size = 524288 bytes: 126.108 MB/s
>> size = 1048576 bytes: 100.787 MB/s <===== The drop-off around 1 meg
>> size = 2097152 bytes: 80.503 MB/s doesn't seem to be any greater than
>> size = 4194304 bytes: 75.294 MB/s the drop-offs that occur in a gradation
>> size = 8388608 bytes: 74.419 MB/s above it.
> Indeed, I see no evidence of an L2 cache. The results are roughly consistent with
>there
> not being one. I'm not prepared to go into a second order analysis!
> By the way, I have an old P100 on which the results are flat at 20MB/s as far as the
>eye
> can see. I suspect it has a broken mobo cache.
>> *** MEMORY READ PERFORMANCE (256 MB LOOP) ***
>> size = 16384 bytes: 1163.636 MB/s
>> size = 32768 bytes: 1113.043 MB/s
>> size = 65536 bytes: 400.000 MB/s <===== Again, this looks right.
Here it's clear that on read there is a 64KB cache.
>> size = 131072 bytes: 400.000 MB/s
>> size = 262144 bytes: 290.909 MB/s
>> size = 524288 bytes: 241.509 MB/s
>> size = 1048576 bytes: 186.861 MB/s <==== more obvious dropoff here than
and not so obvious. Looks more like 512KB cache to me.
>> size = 2097152 bytes: 147.126 MB/s the one above, but not major.
> You'd have to graph it to see if the slope changes here. You might have a read cache.
I meant to say: a cache that works for reading but doesn't do a lot on writing .. i.e.
it
goes through synchronously on write. Writethru?
>> size = 4194304 bytes: 134.031 MB/s
>> size = 8388608 bytes: 132.642 MB/s
Peter
------------------------------
From: James Tappin <[EMAIL PROTECTED]>
Subject: lost interrupt?
Date: Fri, 12 Jan 2001 22:30:06 +0000
Hi,
I've just been resurrecting an old machine to run as a sort of
try-things-out box. I was installing Debian potato on it, and things were
going fine until it got into one of the configuration stages quite late on
(sorry I didn't remember to not down exactly what was configuring) at which
point I got the error message:
hda: lost interrupt
which repeated every few seconds with no other signs of activity until I
pressed the reset button.
Does anyone have any clue what might be the cause of the error (I've been
using Linux for about 3 1/2 years and it's a new one on me). My worry is
that possible the IDE controller is flakey -- the machine had been unused
for a couple of years since it's power supply failed [slowly and painfully].
Any ideas?
TIA
James
--
James Tappin, O__ "I forget the punishment for using
[EMAIL PROTECTED] -- \/` Microsoft --- Something lingering
http://www.xena.uklinux.net/ with data loss in it I fancy"
------------------------------
From: "Alexander Deztroyer" <[EMAIL PROTECTED]>
Subject: intel express 100 known to the kernel but can't be identified?
Date: Sat, 13 Jan 2001 11:44:03 +1300
Greeting.
I've attempted to get my on board LAN card to work with Linux but failed.
The problem is my kernel (2.1.125) has configured to identify my express PRO
100, but when I type ifconfig eth0 ..... the system said unknown device!!
The same problem not just only happened to my Debian 2.1, but also RH 5.2.
Red Hat can't even "auto detect" my LAN card. The Express Pro worked
perfectly under Windows environment, so I assume it is working fine. Now,
what should I do in order to solve the problem?
Your help will be appreciated.
Alex
------------------------------
From: "Steve Wolfe" <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.misc,comp.os.linux.advocacy
Subject: Re: Linux IDE RAID Cards
Date: Fri, 12 Jan 2001 15:56:35 -0700
> Can anyone recommend a good raid IDE controller for Linux. Preferably
> ATA 100. We have tried using the Promise controller with limited
> success. We probably want to run 2 controllers in the server (for more
> speed) with a total of 6 drives. 4 of the drives running raid 1+0 and
> the other 2 forming a separate mirror. The 2 in the mirror we also want
>
> to boot off of.
If you really need that much speed that you're going to go with two IDE
RAID controllers, you really might want to look into a SCSI RAID setup.
Having used SCSI RAID, I can say that it just makes you go "ooooooh" with
delight. It's definitely worth the extra money.
steve
------------------------------
From: Chris Lopeman <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.misc,comp.os.linux.advocacy
Subject: Re: Linux IDE RAID Cards
Date: Fri, 12 Jan 2001 23:06:15 GMT
As I said. We already have started. We have the IDE drive. But thanks for
posting.
Steve Wolfe wrote:
> > Can anyone recommend a good raid IDE controller for Linux. Preferably
> > ATA 100. We have tried using the Promise controller with limited
> > success. We probably want to run 2 controllers in the server (for more
> > speed) with a total of 6 drives. 4 of the drives running raid 1+0 and
> > the other 2 forming a separate mirror. The 2 in the mirror we also want
> >
> > to boot off of.
>
> If you really need that much speed that you're going to go with two IDE
> RAID controllers, you really might want to look into a SCSI RAID setup.
> Having used SCSI RAID, I can say that it just makes you go "ooooooh" with
> delight. It's definitely worth the extra money.
>
> steve
------------------------------
** 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
******************************
