Linux-Hardware Digest #301, Volume #10 Sat, 22 May 99 18:13:35 EDT
Contents:
Sound card IDE (=?iso-8859-1?Q?S=E9bastien?=)
Re: UDMA under Linux 2.2.5 on Asus P5A-b (Ali M15xx chipset) ("marco viola")
Re: what is a voice/fax modem exactly? ("Larry Brasfield")
ATI video card info ("Matt")
Re: accurate timer - HELP! (Armin Steinhoff)
Re: With dual-processor system, is SCSI a must or is Ultra-DMA enough? (David C)
Re: All the current OSes are idiotic (was Re: Is Windows for idiots?) (Anthony Ord)
Re: how's this for a cheap webserver? ([EMAIL PROTECTED])
voodoo3 support (Chrysophylax Dives)
ethernetcard NE2000 ("eelco of viola")
Re: Adaptec 2940UW PCI problem ("E.W. Veltman")
The World Wide Expo 7633 ([EMAIL PROTECTED])
Re: Linux on Dual Pentium-II machines ([EMAIL PROTECTED])
Re: voodoo3 support ("Lee Sharp")
IBM 36XP Harddrive and Symbios 53x875 and linux Kernel 2.0.36/2.2.5 (The Man)
Toshiba cd-rom (D E G)
Configuration for high-spec NFS server (Nick Williams)
----------------------------------------------------------------------------
From: =?iso-8859-1?Q?S=E9bastien?= <[EMAIL PROTECTED]>
Subject: Sound card IDE
Date: Mon, 17 May 1999 11:38:46 +0200
Hi
I have a PNP sound card with an IDE connector.
The card is working and configuration of PNP also.
The problem is that linux detect IDE hardware before the launch
of PNP config, so , IDE interface n�3 is not seen by the kernel.
When I boot under DOS and run the proprietary activating tool, and
then run Linux (with loadlin) , IDE2 is here (with IDE0 and IDE1).
How to tell to the Kernel to detect IDE2 after the launch of
PNP configuration (by isapnp tools), without running the proprietary
tool under dos.
Thanks for any answer.
Seb.
------------------------------
From: "marco viola" <[EMAIL PROTECTED]>
Subject: Re: UDMA under Linux 2.2.5 on Asus P5A-b (Ali M15xx chipset)
Date: Sat, 22 May 1999 14:41:01 +0200
Simone Piccardi ha scritto nel messaggio
<[EMAIL PROTECTED]>...
>Hi,
>I have the same problem with this board (and I'm using 2.2.9). At boot
http://www.dyer.vanderbilt.edu/
ciao
marco
------------------------------
From: "Larry Brasfield" <[EMAIL PROTECTED]>
Subject: Re: what is a voice/fax modem exactly?
Date: Sat, 22 May 1999 18:38:12 GMT
Warning: Potentially inflammatory material within.
Robert Young <[EMAIL PROTECTED]> wrote in message
news:tEA13.7$[EMAIL PROTECTED]...
> Please kindly keep me posted when you received any response.
OK.
...
> >===== Original Message From [EMAIL PROTECTED] (Walter Tautz) =====
> >My apologies for asking this here, but I couldn't find
> >an explanation in the current howto on modems (the author
> >says it will appear in the next version). Is it correct
> >to assume that a voice modem allows one to use the phone
> >for ordinary phone calls while remaining connected?
Assumption can only be correct coincidentally.
A modern modem actively "connected" to a
common voice line is using far too much of
the channel capacity for it to also carry a full
bandwidth voice signal. It might be possible
for some modem to share the digital channel
between the "bits" port and a voice port, but
that would require similar provisions on the
other end which have not been made, yet.
A voice modem is one that provides a pathway
for voice signals (digital sample sequences)
between the line and other subsystems within
the computer such as the sound card. This is
actually one side-benefit of the much-dreaded
"WinModems". But the "voice" pathway can
only carry the voice signal when the QPAM
(V.##) signals are not passing thru the line.
> >Similarly for fax??
The same situation holds for FAX, (although
FAX uses much less of the channel capacity).
> >Anyone have experience using this type of modem under
Not yet that I've seen.
--
Larry Brasfield
Above opinions may be mine alone.
(Humans may reply at unundered [EMAIL PROTECTED] )
------------------------------
From: "Matt" <[EMAIL PROTECTED]>
Subject: ATI video card info
Date: Sat, 22 May 1999 13:52:31 -0500
I just put in a different video card. It is a ATI mach64 GX, found out
about the GX from ATI's web page. I got it used. It works fine, but it
seems that I can't get higher resolutions. I would like to have the 24 bit
color, but it seems to be stuck on th 8 bit. If I try to set it up for
higher res. it fails to load Xwindows. I don't know what kind of clock chip
is on it and also what kind of DAC it has. Any help with this would be
greatful.
Thanks
------------------------------
From: Armin Steinhoff <Armin@Steinhoff_de>
Crossposted-To: comp.realtime,comp.os.linux.development.system,comp.hardware
Subject: Re: accurate timer - HELP!
Date: 22 May 1999 11:21:25 -0700
In article <7i4ntv$83p$[EMAIL PROTECTED]>, [EMAIL PROTECTED]
says...
>
>In article <[EMAIL PROTECTED]>,
>Vladimir G. Stanishev <[EMAIL PROTECTED]> wrote:
>)killbill wrote:
>)
>)> In article <[EMAIL PROTECTED]>,
>)> Dorin-Ioan Marinca <[EMAIL PROTECTED]> wrote:
>)>
>)> > How can I count the time (*less than 1us* - even x*10ns) very accurate
>)> > on Linux? I search something not depended on hardware or, if not,
>)> > something Pentium specific.
>)
>)honnestly I doubt that you'll be able to get to less than 1 microsecond
>)resolution. You may have already found the UTIME
>
>[snip]
>
>On PC hardware, the highest frequency clock which is directly visible to
>software (discounting those which a specific processor might provide)
>runs at 14.31818MHz/12 = 1.19318 MHz, giving 838 ns resolution.
That's true for the system timers ... but you have no chance to use that
resolution :-(
The real-time clock (RTC,IRQ8)is more practical. It goes down to 122.070 us ...
Armin
>
>Mike
>--
>----
>char *p="char *p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);}
>This message made from 100% recycled bits.
>I don't speak for Alcatel <- They make me say that.
------------------------------
From: David C <[EMAIL PROTECTED]>
Subject: Re: With dual-processor system, is SCSI a must or is Ultra-DMA enough?
Date: 22 May 1999 15:29:39 -0400
[EMAIL PROTECTED] (David Morgan) writes:
>
> Are the numbers at storage review all wet? I know they don't compair
> linux, but my guess is that threadmarks are not that far off.
>
> (for example, at storagereview.com, database, compaire an ibm Deskstar
> 14GXP and Maxtor DiamondMax Plus 2500 against a seagate barracuda and
> cheta scsi drives.)
Most of the time, they test the drive in isolation. That is, they run a
program to send a pile of read and write commands at the drive, and they
time the results. This does a good job of testing the top speed of the
drive and its interface, but it doesn't reflect real-world situations.
If instead, they would run 30 copies of their program - so the disk
subsystem gets hit with a dozen or so simultaneous commands, then you'd
see a difference.
The IDE drive would have no choice but to handle them one at a time,
making the other requests wait. (The operating system could re-order
the requests, but once a request goes to the drive, no other request can
be sent until the first one finishes.)
A proper SCSI drive, however, can accept all the commands at once. The
drive's firmware should be capable of re-ordering the requests for
optimum performance, completing them in whatever order is most efficient
for the mechanism.
Furthermore, I don't think any of these publications test the
performance of two drives on a single EIDE channel together. If you put
two drives on one IDE channel and access them both at once, you'll find
that only one drive can work at a time - when a request is sent to one,
the other can't send or receive any data until the first drive releases
the channel. On the other hand, proper SCSI drives (that implement
disconnect) don't have this problem - the drive can receive a request,
disconnect from the bus (so some other device can use it), and then
reconnect later to send back the requested data.
In other words, a UDMA-IDE drive and a SCSI drive may perform
identically in a system where there are no other drives and all disk
access is coming from one application, but if you put those same drives
in a configuration with several other devices and multiple simultaneous
accesses, you'll find the SCSI drive performing better. Not because the
bus is any faster (although it might be) but because things block on
each other less often.
-- David
------------------------------
From: [EMAIL PROTECTED] (Anthony Ord)
Crossposted-To:
comp.lang.java.advocacy,comp.os.linux.advocacy,comp.os.ms-windows.advocacy
Subject: Re: All the current OSes are idiotic (was Re: Is Windows for idiots?)
Date: Sat, 22 May 1999 19:42:15 GMT
On Fri, 21 May 1999 12:17:56 +0200, [EMAIL PROTECTED]
(Guest) wrote:
>In article
><[EMAIL PROTECTED]>
>[EMAIL PROTECTED] (Mike Bartman) writes:
>
>> >Other examples include plugins that seamlessly allow your web browser to work with
>new data
>> >types.
>>
>> Compilers and plugins are one thing...OSs are another.
>>
>> If the OS dies, *everything* dies.
>
>I think you guys are mixing up concepts here. Where one of you sees 'OS' in
>the Windows NT way, i.e. including GUI, standard applications, etc., the other
>sees 'OS' as the kernel.
>
>I'm not going to argue for either one, but I would like say that I think the
>Linux 'OS' (in the broad meaning) is much more modular than NT. In Linux you
>can choose the GUI you want: do you want to create something with a simple
>character interface, something with SVGALib or do you want full-blown X
>windows ? The user has the choice. On NT4, you get Microsoft's idea of a GUI
>(ok, it's not too bad) and I have yet to see a way for someone to tweak it the
>way you can select your own Windows manager (and therefor decorations) under
>X.
Litestep and EFX, though EFX may only be for Win9x. Ok,
neither of them make any fundamental changes, but they are
good for a bit of harmless fun.
>The same thing goes for the API issue. If you really want to create such a
>module for NT you have to use Microsoft APIs that are almost completely
>undocumented.
Even with the DDK?
>Writing a filesystem for NT is a real pain which is probably
>why Linux offers a lot more 'foreign' filesystems than NT.
>
>The fact that Linux is more modular also has a drawback: because even large
>things like the GUI aren't part of the kernel (so are a kind of module),
>several of them may exist, which in turn have different APIs, meaning that
>modules created to interact with one of them, will not run on the
>other. That's my biggest objection agains the Gnome vs. KDE issue. Both seem
>promising, but what will happen if I need to use 2 applications and one
>uses specific Gnome features, while the other uses specific KDE stuff ?
You simply don't get the Gnome/KDE specific feature. They
are often referred to as "window manager *hints*". Of course
this presupposes that no one has specifically gone out of
their way to be just compatible with one.
>Having the same API would be the solution,
You use ORBit or whatever.
>but then what would the difference
>be between the two ?
The difference is in the look I guess.
>Anyhow: Linux definately beats Micro$hit on the point of modularity.
Regards
Anthony
--
=========================================
| And when our worlds |
| They fall apart |
| When the walls come tumbling in |
| Though we may deserve it |
| It will be worth it - Depeche Mode |
=========================================
------------------------------
Subject: Re: how's this for a cheap webserver?
From: [EMAIL PROTECTED]
Date: Sat, 22 May 1999 20:25:31 GMT
According to Lyn A Headley <[EMAIL PROTECTED]>:
> I'm putting together an experimental webserver on a shoestring
> budget.
What do you mean by experimental? It isn't to difficult to find
someone who will *give* you an old 486 which will run run linux and a
web server just fine with 32+megs of RAM.
Once you get it running, sick a Perl script on it to see how many hits
it can take per second. The results may surprise you. If you need
more performance, spend your money where it counts. In most cases,
this means (1) getting more RAM and (2) going SCSI.
-p.
------------------------------
From: Chrysophylax Dives <[EMAIL PROTECTED]>
Subject: voodoo3 support
Date: 22 May 1999 19:54:22 GMT
I'm having trouble finding out what the situation is with 3dfx voodoo3 support
is for linux. /Is/ there any? voodoo2/banshee seems to be supported, but
i've seen nothing on voodoo3.
-giles
------------------------------
From: "eelco of viola" <[EMAIL PROTECTED]>
Subject: ethernetcard NE2000
Date: Sat, 22 May 1999 22:31:36 +0200
hello there,
As I am new to linux and ethernetcards i`ve been looking at different cards.
I think its going to be an ne 2000 since it isn`t that expensive. However
the ethernet-howto says to be careful with clones. Can anyone tel me the
brand name of an originel NE 2000 (I cant find it annywhere) or other cheap
cards names that work under linux aswel as windows.
Anny comment would be greatly appriciated
greets Eelco
------------------------------
From: "E.W. Veltman" <[EMAIL PROTECTED]>
Subject: Re: Adaptec 2940UW PCI problem
Date: Sat, 22 May 1999 21:57:49 +0200
Hello Nigel,
Nigel Hawkes heeft geschreven in bericht <01bea30a$003498c0$0100007f@dan>...
>Fred Horan <[EMAIL PROTECTED]> wrote in article
><[EMAIL PROTECTED]>...
>> I installed RedHat v5.2 and ran into a problem with an adaptec 2940UW pci
>> board. At bootup I get the following message:
>>
>> scsi: aborting command due to timeout: pid 0, scsi0, channel 0, id 0, lun
>0,
>> test unit ready 00 00 00 00 00
>> SCSI host 0 abort (pid 0) timed out - resetting
>
>Sorry, I can't offer help, only sympathy. I've had very similar problems
>trying to install SuSE 6.1 on a system with AIC-7895 SCSI on the
>motherboard. Typing "linux aic7xxx=no_reset" at the Boot: prompt didn't
>help - I just got lots of messages about data overrun, followed by the same
>endless series of resets.
Add me to the list of people having the same problem ...
I didn't even have to go back to SuSE5.3, I couldn't _start_ installing
SuSE6.1 :-(
I also disabled "automatic" termination of my AHA2940UW, verified
termination
on all SCSI devices, tried the "aic7xxx=no_reset" parameter, but it doesn't
help.
Kernel 2.0.35 + 2.0.36 both work properly.
I've got the feeling it's got something to do with non harddisk SCSI
devices.
Just about all messages I saw on this topic had non harddisk devices
connected to the SCSI bus. ( Things like CD-RWs, SCSI tape drives and so
on )
Best regards,
Eric
------------------------------
From: [EMAIL PROTECTED]
Crossposted-To:
comp.os.linux.development.system,comp.os.linux.m68k,comp.os.linux.misc,comp.os.linux.networking,comp.os.linux.portable,comp.os.linux.powerpc,comp.os.linux.setup,comp.os.linux.x,comp.os.lynx
Subject: The World Wide Expo 7633
Date: 22 May 1999 19:36:15 GMT
voskyxupefprrncchlgzkoxkkmpxeiuncxjxvwfryih
------------------------------
Crossposted-To: comp.os.linux.misc
Subject: Re: Linux on Dual Pentium-II machines
From: [EMAIL PROTECTED]
Date: Sat, 22 May 1999 20:11:39 GMT
According to mumford <[EMAIL PROTECTED]>:
> You're requesting info about building a number crunching system... I'm
> almost positive that you could expect a significant performance hit be-
> cause of the decreased cache size if you chose celerons instead of true
> P-II's (celerons have 128K cache, true P-II's have 512K).
The Celeron cache size is smaller, but it runs at twice the speed. In
almost all real-world cases, this leads to better performance from the
128K cache Celerons.
It was also noticed several years ago that increasing cache size in
some processors, assuming constant speed, actually reduced performance
due to the increased overhead of cache management. I do not believe
this factors into the difference between the PII/Celeron, though.
-p.
------------------------------
From: "Lee Sharp" <[EMAIL PROTECTED]>
Subject: Re: voodoo3 support
Date: Sat, 22 May 1999 16:04:21 -0500
Chrysophylax Dives wrote in message <7i721e$ioi$[EMAIL PROTECTED]>...
|I'm having trouble finding out what the situation is with 3dfx voodoo3
support
|is for linux. /Is/ there any? voodoo2/banshee seems to be supported, but
|i've seen nothing on voodoo3.
The Banshee driver has V3 support. It is still somewhat raw, but it
works. It is on the edge 3d website who's url escapes me at the moment.
Lee
--
SCSI is *NOT* magic. There are *fundamental technical reasons* why it is
necessary to sacrifice a young goat to your SCSI chain now and then. * Black
holes are where God divided by zero. - I am speaking as an individual, not
as a representative of any company, organization or other entity. I am
solely responsible for my words.
------------------------------
From: The Man <[EMAIL PROTECTED]>
Subject: IBM 36XP Harddrive and Symbios 53x875 and linux Kernel 2.0.36/2.2.5
Date: Sat, 22 May 1999 22:10:43 +0200
Does anybody out there have any experiance with this combination
descriped above..
My system has turned really unstabel after changing my harddrive from an
old Quantum Atlas II 4,3 G
and to this IBM 36,4 G Wide LVD harddisk
No, the Symbios controller is not a LVD controller but the harddrive is
backware compatible though.
Everytime I run fsck after boot it finds new errors on the disk, and if
I run it twice it finds differant errors
co there is no continuerity in the failure report.
The drive and Controller has ben testet on a Win98 (sorry) platform and
I got no failures yet
Cheers
Please reply with E-Mail
------------------------------
From: D E G <[EMAIL PROTECTED]>
Subject: Toshiba cd-rom
Date: 22 May 1999 13:23:42 PDT
i have a toshiba 12/24x cd rom drive (sorry don't know the model number,
but it's the kind that ships with gateway gp5 machines).
anyone know how to get it to work under RH6?
TIA
-eli
------------------------------
From: [EMAIL PROTECTED] (Nick Williams)
Crossposted-To: comp.os.linux.misc,comp.os.linux.networking
Subject: Configuration for high-spec NFS server
Date: 22 May 1999 22:07:51 +0100
Hi there,
I'm involved in specifying a new server for a research group,
which requires a high performance NFS server. Under
consideration are a Sun Enterprise 450 with a StorEdge A1000
external RAID box running Solaris, or a PC-based Linux solution,
as detailed below.
I'm interested in the likely NFS throughputs that could be
expected from this configuration, as well as any helpful
comments as to the general suitability of the design.
Server configuration:
* Dual/Quad Intel Pentium III Xeon (512K LII) at 500MHz
* 1GB ECC EDO RAM
* Mylex eXtremeRAID 1100 64bit PCI RAID Controller (64MB cache)
* 6 x 36.4GB Seagate Barracuda LVD Ultra-2 SCSI disks (3 per
channel on the RAID controller - in RAID 5 array)
* 3Com 3C985 Gigabit Ethernet NIC
Network configuration:
* 3Com Switch 3300 10/100 (48 ports total) with Gigabit modules
Client configuration:
* Mixed Sun hardware, from SparcStation 4 to Ultra 60
* Miscellaneous PC Linux boxes (Pentium II)
System load:
* 15-20 simultaneous users
* Typically multi-gigabyte datasets stored as UNIX files
* 50/50 read/write load
--
[ Nick Williams Office - 01865-(2)79252 ]
[ Computing Officer Mobile - 07775-637124 ]
[ New College, Oxford http://www.new.ox.ac.uk/~nick/ ]
------------------------------
** 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
******************************
