Linux-Hardware Digest #589, Volume #12 Sat, 1 Apr 00 06:13:05 EST
Contents:
Re: winmodem confusion !! ("Jason Byrne")
Re: Multi Processor Systems ("D. Stimits")
Re: Dual processor ("D. Stimits")
Re: SCSI and IDE disk problems ("D. Stimits")
Parallel Port Programming ([EMAIL PROTECTED])
Re: BE6-II mobo no > 64MB (Peter T. Breuer)
Re: Parallel Port Programming (Prasanth Kumar)
what harware is best? (Ray Tayek)
Any Promise Fasttrak66 running? ("Jack")
Re: Optra 40 + Samba + Win98 != happy (Martin Vonwald)
Soundblaster Live! Value card not supported ("John T. H. Wong")
Re: Multi Processor Systems (Keith)
Re: Soundblaster Live! Value card not supported (Stephen)
Re: I Need Mouse Help ! (Peter T. Breuer)
Re: RH6.1 install to RAID (Peter T. Breuer)
----------------------------------------------------------------------------
From: "Jason Byrne" <[EMAIL PROTECTED]>
Subject: Re: winmodem confusion !!
Date: Fri, 31 Mar 2000 21:29:13 -0800
I would never *suggest* a winmodem... but I have three computers with PCTel
HSP56 Micromodem (winmodem) - working under Linux.
Good source of information... try http://www.o2.net/~gromitkc/winmodem.html
Ray <[EMAIL PROTECTED]> wrote in message
news:[EMAIL PROTECTED]...
> anyone out there that can tell me if there is such a thing as a
> winmodem that will work with linux(redhat 6.1 or mandrake 7.0)
> I don't have any spare ISA slots only PCI. Nd which do you feel is
> better to use for internet publishing. Mandrake or RedHat ?
> would appreciate input on these matters
> Thanks
------------------------------
Date: Fri, 31 Mar 2000 22:41:36 -0700
From: "D. Stimits" <[EMAIL PROTECTED]>
Reply-To: [EMAIL PROTECTED]
Subject: Re: Multi Processor Systems
Roaster wrote:
>
> Hi folks,
>
> I'm interested in building a multi-processor based system.
>
> But- I know that current Windows (non NT/2000) systems dont support it, and
> am relatively knowlde-less with Linux.
>
> Anyone know the improved performance that you get from a MP based system?
>
> i.e. figures from single and multi-cpu systems of the same type.
>
> Also, I've only seen multi processor based Celeron systems available - can
> you get multi cpu based Athalon boards? (and I havent seen multi proc p3
> boards either.)
>
> Cheers
>
> Chris.
Another reply already mentioned SuperMicro, which does well with linux
SMP. I heard that an Athlon dual-cpu chipset has just been announced,
but I wouldn't expect any motherboards to hit the streets for a long
time. The i840 chipset for pII and III's is out, but it is hard to
find dealers that actually have them in stock.
*Before you buy one, ask the dealer if they are in stock first.*
The BX and GX chipsets are good, but the new generation of AGP 4x
video cards are not supported here. While it is true that AGP 4x isn't
really useful with the current generation of video cards, due to
bandwidth limitations, it's possible that the SMP setups might be able
to use this a tad bit better (define a tad, anyone). A couple of real
advantages of going *beyond* the BX/GX are:
1) The AGP Pro electrical specification will supply 50 watts to the
AGP slot, which can come in handy on a number of existing AGP video
cards...BX/GX boards don't have this power delivery;
2) Some are coming out with a 66 MHz 64 bit PCI slot (Intel and
SuperMicro I know). This slot is able to handle quite a bit more
bandwidth than the 33 MHz or 32 bit slots. New drive controllers, such
as SCSI 3 and fiber channel, can finally exceed the capacity of the
old PCI.
As for performance, it depends greatly. To use both processors from a
single program, the program has to be threaded. Two separate
single-threaded programs could benefit as well, in terms of overall
performance of the machine; individually the programs would not
benefit, but the system as a whole won't slow down. Some applications
are highly threaded, and do quite well...my favorite example is
compiling C/C++. Kernel compiles, for example, do come close to being
twice as fast. Even when your individual programs are not going to
take any advantage of the 2nd cpu, some operating systems can use the
2nd cpu for its own purposes (linux is exceptionally good at this from
2.1.x kernels on up). For example, your program might need to access a
hard drive...if you have only one cpu, then this access requires
stopping the program for a short time while doing so; if you have a
2nd cpu, the drives can be running and getting the next data while the
program continues. Two cpus, even when overall speed doesn't improve
noticably, tend to give a qualitative smoothness improvement that is
highly noticable. If you have enough ram to run your system with a
large amount of caching, and no disk swap, it becomes difficult to
make the machine "burp"; if you are low on ram, you'll lose
performance of the 2nd cpu quite fast.
------------------------------
Date: Fri, 31 Mar 2000 22:50:54 -0700
From: "D. Stimits" <[EMAIL PROTECTED]>
Reply-To: [EMAIL PROTECTED]
Subject: Re: Dual processor
Tom Haeck wrote:
>
> Hello everybody,
>
> I have a CPU-board with 2 x PIII-500. I installed SUSE 6.2 on it. It was
> installed, and the CPU monitor in KDE only showed me one CPU. So I concluded
> that the basic installation of SUSE doesn't use the 2nd processor (Is this
> so?).
> So, I recompiled my Kernel, but still I can see only 1 CPU in the
> CPU-monitor. How can I be sure that Linux uses the 2 processors? And what
> must I do in order to make them work?
>
> Thanks in advance,
>
> TH
I don't know of any basic cpu monitors that show dual cpu when there
(xosview has a patch, but I don't think any dist's use it by default).
/proc/cpuinfo, however, is straight from the kernel. If it shows cpu 0
and 1, you are using 2 cpu's. /proc/cpuinfo should be your test for
SMP/non-SMP support.
------------------------------
Date: Fri, 31 Mar 2000 22:58:06 -0700
From: "D. Stimits" <[EMAIL PROTECTED]>
Reply-To: [EMAIL PROTECTED]
Crossposted-To: comp.os.linux.misc
Subject: Re: SCSI and IDE disk problems
"Knut A. Nilsen" wrote:
>
> About 6 months ago I installed a new Seagate Barracuda ST39175LW 9,1 Gb SCSI
> disk on a small fileserver. The disk has one partition taking up the entire
> disk, most of which is 'shared' to Windows clients using Samba (2.0.5a). The
> system has been running happily until yesterday, when I discovered that
> files and directories were missing over the samba share. Trying to list the
> files in linux gave me IO errors. Unmounting the drive and running e2fsck
> gave the following output:
>
> Error reading block XXX (Attempt to read block from filesystem resulted in
> short read) while doing inode scan. Ignore error<y>?
>
> Ignoring the error would give me more messages like this. Running e2fsck -c
> caused the scsi bus to attempt to reset several times. Finally, after
> several hours of errors, SCSI bus reset attempts etc etc, rebooted the
> machine, unmounted the drive and ran e2fsck -c, which now seamed to work.
> The disk is up and running again with no new error messages...
>
> /var/log/messages:
> Mar 29 12:29:01 arthur kernel: SCSI disk error : host 0 channel 0 id 0 lun 0
> return code = 28000002
> Mar 29 12:29:01 arthur kernel: [valid=0] Info fld=0x0, Current sd08:01:
> sense key Not Ready
> Mar 29 12:29:01 arthur kernel: Additional sense indicates Logical unit is in
> process of becoming ready
> Mar 29 12:29:01 arthur kernel: scsidisk I/O error: dev 08:01, sector 8536140
> Mar 29 12:29:01 arthur kernel: EXT2-fs error (device sd(8,1)):
> ext2_write_inode: unable to read inode block - inode=1067
> 010, block=4268070
> Mar 29 12:29:06 arthur kernel: SCSI disk error : host 0 channel 0 id 0 lun 0
> return code = 28000002
> Mar 29 12:29:06 arthur kernel: [valid=0] Info fld=0x0, Current sd08:01:
> sense key Not Ready
> Mar 29 12:29:06 arthur kernel: Additional sense indicates Logical unit is in
> process of becoming ready
> Mar 29 12:29:06 arthur kernel: scsidisk I/O error: dev 08:01, sector 2
> Mar 29 12:30:03 arthur kernel: (scsi0:0:0:-1) Unexpected busfree, LASTPHASE
> = 0x40, SEQADDR = 0x5f
> Mar 29 12:30:07 arthur kernel: SCSI disk error : host 0 channel 0 id 0 lun 0
> return code = 28000002
> Mar 29 12:30:07 arthur kernel: [valid=0] Info fld=0x0, Current sd08:01:
> sense key Not Ready
> Mar 29 12:30:07 arthur kernel: Additional sense indicates Logical unit is in
> process of becoming ready
> Mar 29 12:30:07 arthur kernel: scsidisk I/O error: dev 08:01, sector 983962
> Mar 29 12:30:07 arthur kernel: SCSI disk error : host 0 channel 0 id 0 lun 0
> return code = 28000002
>
> (same message repeated a _lot: of times)
>
> Mar 29 12:36:11 arthur kernel: scsi0 channel 0 : resetting for second half
> of retries.
> Mar 29 12:36:11 arthur kernel: SCSI bus is being reset for host 0 channel 0.
> Mar 29 12:36:14 arthur kernel: SCSI disk error : host 0 channel 0 id 0 lun 0
> return code = 28000002
> Mar 29 12:36:14 arthur kernel: [valid=0] Info fld=0x0, Current sd08:01:
> sense key Not Ready
> Mar 29 12:36:14 arthur kernel: Additional sense indicates Logical unit is in
> process of becoming ready
> Mar 29 12:36:14 arthur kernel: scsidisk I/O error: dev 08:01, sector 2883722
>
> Can anyone explain this behaviour?
>
> Now, for chapter two:
>
> In 'panic', while I was fsck'ing the SCSI drive, I went an bought a second
> 10GB IDE drive, to backup my disk in case I needed to try some extensive
> data recovery. I makde this dosk inot one big partition also (/dev/hdc1).
> Running mke2fs, /var/log/messages is filled with this kind of errors:
>
> Mar 30 11:27:44 arthur kernel: hdc: dma_intr: status=0x51 { DriveReady
> SeekComplete Error }
> Mar 30 11:27:44 arthur kernel: hdc: dma_intr: error=0x10
> { SectorIdNotFound }, CHS=770/0/137, sector=197074
> Mar 30 11:28:02 arthur kernel: hdc: dma_intr: status=0x51 { DriveReady
> SeekComplete Error }
> Mar 30 11:28:02 arthur kernel: hdc: dma_intr: error=0x10
> { SectorIdNotFound }, CHS=1666/0/79, sector=426450
> Mar 30 11:28:21 arthur kernel: hdc: dma_intr: status=0x51 { DriveReady
> SeekComplete Error }
> Mar 30 11:28:21 arthur kernel: hdc: dma_intr: error=0x10
> { SectorIdNotFound }, CHS=3650/0/23, sector=934354
> Mar 30 11:28:22 arthur kernel: hdc: dma_intr: status=0x51 { DriveReady
> SeekComplete Error }
> Mar 30 11:28:22 arthur kernel: hdc: dma_intr: error=0x10
> { SectorIdNotFound }, CHS=3650/0/22, sector=934354
>
> ..and runing e2fsck on this disk gives:
>
> Error reading block 3358818 (Attempt to read block from filesystem resulted
> in short read) while doing inode scan. Ignore error<y>?
>
> ...which is the same message I got on the SCSI disk in the first place!
>
> Which means that I can't get my brand new 10 GB drive to work at all. The
> drive is an IBM DTTA-371010 CHS=19590/16/63
>
> On top of this, as I am writing this post, I see that I get the same error
> messages on /dev/hda and IO errors as I am parsing through /var/log/messages
> to copy the error messages. This disk is an IBM DTTA-350640 CHS=790/255/63
>
> This is beginning to become frustrating! Any ideas, anyone?
>
> Knut Nilsen
The faster the data setup, such as newer 80 MB/s versus old 20 MB/s,
the stricter the cable requirements. And cable connectors, constantly
warming up and cooling down, can become unseated...not necessarily
enough to be obvious, but enough to cause errors. If you were writing
to a drive when it was marginal, you might have lost data, but not
have lost the drive. Do the simple test: keep yourself grounded to the
computer case, and unseat/reseat all cable connections related to the
drives. Try again to run e2fsck.
------------------------------
From: [EMAIL PROTECTED]
Subject: Parallel Port Programming
Date: Sat, 01 Apr 2000 06:30:05 GMT
Hi all,
I am in the process of purchasing electronic parts for a project that I
am building. Basically, this project will send a series of infrared
signals to a Sony Pocketstation so that I can "trick" a mini-game. I
want to use a PC to send specific signals to my device using the
parallel port.
Is there a way to write to the parallel port directly, either in C or
assembly (preferrably C)? I tried sending a stream of data to /dev/lp0,
but the application waited until I powered on an imaginary printer
connected to the port. :*) I just want to send signals to the parallel
port and not expect anything back in return.
Thanks.
Sent via Deja.com http://www.deja.com/
Before you buy.
------------------------------
From: [EMAIL PROTECTED] (Peter T. Breuer)
Subject: Re: BE6-II mobo no > 64MB
Date: 1 Apr 2000 07:11:30 GMT
Reply-To: [EMAIL PROTECTED]
Felonius Monk ([EMAIL PROTECTED]) wrote:
: Okay, so I have this Abit BE6-II that won't read any more than 64 megs of
: Ram in Linux. It turns out that it's because they banished a call in the
: BIOS to make the board ACPI compliant or something, but Linux (at least
: 2.2.14) relies on that call to accurately detect the right amount of RAM.
: There isn't an update for the board, but I was told that I had a couple of
: choices:
: 1) Manually tell the kernel how much RAM I have in the system
: 2) 2.3 kernels seem to be detecting the right amount of Ram on these boards
: My questions:
: 1) How would I go about telling the kernel how much RAM it should be reading
: manually?
This question has been asked �very day for years. Read the newsgroups
and find out! Search deja for 64M+linux. Or just scan 10 posts upwards
in this group. Or read your BootPrompt-HOWTO, or grep for the obvious in
the kernel Documentation/ ....
: 2)How stable are the prerelease 2.3 kernels?
How would anyone know? I don't think anyone can say they've had them running
for long! They seem OKish to me.
: Obviously, since I asked the first question, I'm far from a kernel hacker,
: but I'd like to get my system reading the right amount of memory....
Then tell it how much you have.
: TIA!
Peter
------------------------------
From: Prasanth Kumar <[EMAIL PROTECTED]>
Subject: Re: Parallel Port Programming
Date: Sat, 01 Apr 2000 08:12:58 GMT
[EMAIL PROTECTED] wrote:
<snip>
> Is there a way to write to the parallel port directly, either in C or
> assembly (preferrably C)? I tried sending a stream of data to /dev/lp0,
> but the application waited until I powered on an imaginary printer
> connected to the port. :*) I just want to send signals to the parallel
> port and not expect anything back in return.
<snip>
This link may help:
http://www.linuxdoc.org/HOWTO/mini/IO-Port-Programming.html
--
Prasanth Kumar
[EMAIL PROTECTED]
------------------------------
From: [EMAIL PROTECTED] (Ray Tayek)
Subject: what harware is best?
Date: Sat, 1 Apr 2000 00:28:13 -0800
hi, looking at getting dsl from telco with a static ip and running linux.
i want to run a web server and a java servet engine and route to a few
other machine (with static ip's) also do some nat stuff for a lan with
private ip's.
my hardware choices are:
1. 486/50 32mb
2. 120mhz cyrix 686 pr-150 96mb
3. some other relatively inexpensive machine (pentium 2 or so).
which of these would work well?
what brands of nics, modems, graphics and sound cards are preferred or to
be avoided?
any pointers would be appreciated.
thanks
--
Ray (will hack java for food) http://home.pacbell.net/rtayek/
vice chair orange county java users group http://www.ocjug.org/
hate spam? http://samspade.org/ssw/
------------------------------
From: "Jack" <[EMAIL PROTECTED]>
Subject: Any Promise Fasttrak66 running?
Date: Sat, 01 Apr 2000 08:33:30 GMT
as title.
--
Rgds,
Jack
Homepage: http://members.xoom.com/jackl (Xoom mirror)
Email: [EMAIL PROTECTED]
------------------------------
From: Martin Vonwald <[EMAIL PROTECTED]>
Subject: Re: Optra 40 + Samba + Win98 != happy
Date: Sat, 1 Apr 2000 10:58:47 +0200
Vincent Fox wrote:
> However the Win9x drivers for the Optra 40
> seem to produce some weird stuff. Week, 27/9/95
Hi!
Try the generic postscript driver from Adobe. Go to www.adobe.com and
search for "postscript printer driver generic".
Martin
------------------------------
From: "John T. H. Wong" <[EMAIL PROTECTED]>
Subject: Soundblaster Live! Value card not supported
Date: Sat, 01 Apr 2000 17:39:12 +0800
Dear Linux users,
My computer has the soundblaster live! value soundcard and I can
just find a soundmodule for kernel 2.2.5 (now kernel 2.2.12) So it
doesn't work
Can somebody tell me where can i find the driver or work around for
this soundcard?
Thanx
John
------------------------------
From: Keith <[EMAIL PROTECTED]>
Subject: Re: Multi Processor Systems
Date: Sat, 01 Apr 2000 09:53:28 GMT
In article <01bf9aea$0fac2d30$bad4280a@ntbcare19>,
"Roaster" <[EMAIL PROTECTED]> wrote:
> Hi folks,
>
> I'm interested in building a multi-processor based system.
>
> But- I know that current Windows (non NT/2000) systems dont support
it, and
> am relatively knowlde-less with Linux.
Linux supports SMP (symmetric multi-processor), and so does Windows NT
(and I'd assume Terminal Server does too).
> Anyone know the improved performance that you get from a MP based
system?
Most software only uses a single processor, even if two are available.
So, for games, office work, etc there is no big improvement. For
software development the improvement comes because you can compile on
one processor and continue to work on the other one. Also, some graphics
programs (Photoshop for example) will use the second processor to speed
up filtering etc.
The big wins come from multi-threaded software and server applications.
A web server for example can spawn several copies of the connection
process and share the load across both processors. Similarly a
multi-threaded database server can have lots of processes running
simultaneously (data reader, log writer, user connections etc). These
can be shared across both processors.
>
> i.e. figures from single and multi-cpu systems of the same type.
Sorry - I don't have any numbers, but I'm sure someone will post some.
They vary from 0% improvement for single programs to 80% improvement for
some small re-entrant processes.
> Also, I've only seen multi processor based Celeron systems available -
can
> you get multi cpu based Athalon boards? (and I havent seen multi proc
p3
> boards either.)
AFAIK the Athlon doesn't support multi-processor. But you can use
Pentiums, Pentium-Pros, some Celerons, P-II and P-III and Slot-1
Coppermine processors in SMP systems with the appropriate motherboards.
Keith.
--
My employer bears no responsibility for my newsgroup postings.
Sent via Deja.com http://www.deja.com/
Before you buy.
------------------------------
Subject: Re: Soundblaster Live! Value card not supported
From: Stephen <[EMAIL PROTECTED]>
Date: Sat, 01 Apr 2000 10:19:37 GMT
"John T. H. Wong" <[EMAIL PROTECTED]> wrote:
>Dear Linux users,
>
> My computer has the soundblaster live! value soundcard and I can
>just find a soundmodule for kernel 2.2.5 (now kernel 2.2.12) So it
>doesn't work
> Can somebody tell me where can i find the driver or work around for
>this soundcard?
>
> Thanx
>
>John
I got Redhat 6.2 and the soundblaster live! is in the list of soundcfg
so it should work on redhat 6.2
------------------------------
From: [EMAIL PROTECTED] (Peter T. Breuer)
Subject: Re: I Need Mouse Help !
Date: 1 Apr 2000 10:50:08 GMT
Reply-To: [EMAIL PROTECTED]
Quinn ([EMAIL PROTECTED]) wrote:
: Lunix Was Installed & Seemed To Run Fine Except No Mouse...Is Ther A Fix
: For This ? Please E-Mail Me @ [EMAIL PROTECTED]
You Have A Problem With Capital Letters. Your Mouse Will Not Run
While Your CapsLock Key Behaves Like This.
Peter
------------------------------
From: [EMAIL PROTECTED] (Peter T. Breuer)
Subject: Re: RH6.1 install to RAID
Date: 1 Apr 2000 10:48:50 GMT
Reply-To: [EMAIL PROTECTED]
David Trimmell ("dwt*NoSpam*"@pond.net) wrote:
: The more I look into installing RH6.1 on a HP Netserver LM with a EISA
: RAID array, the less I like it. RH claims that they do not support a
If you're talking about hardware raid, there's nothing to do. ALl you need
is support for the controller. If the controller is supported, then you
don't know you're running from raid. Is your controller supported?
: root install onto RAID :-(. I have found a great deal of info regarding
: RAID, but is there a easy step by step guide for installing to a RAID
: system for a relative newbee? I hope I am not forced to go the
There's nothing special to do.
Even if redhat don't have a driver for your controller on their bootdisk
(which is all they mean by "support"), you can always boot without it,
say from floppy or from cd or from the net, and load the driver
afterwards. Then you can make yourself a kernel which does have the
driver support built in, and toss BugHat out the window.
Peter
------------------------------
** 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
******************************
