Linux-Hardware Digest #842, Volume #10 Sun, 25 Jul 99 03:14:04 EDT
Contents:
Re: Mobo for linux server (Bryan)
Occasional timeouts DMA'ing to IDE drive (Jeremy Buhler)
Re: Promise Ultra/66 EIDE Controller (Tim Moore)
Help Kouwell Ultra scsi under linux (John Esterson)
Re: How do I change SCSI cards in Red Hat 6.0? (David Koretzky)
Re: "The LinuxStore" - are they good ?? (Jeremy Fincher)
Can't get Microphone to work. ([EMAIL PROTECTED])
Re: Writing to Windows Partition From Linux Partition (Nelson Ricciardi)
Hardware Problems causing Sig 11.... (Colin Wong)
Re: speed tests (Glenn Valenta)
Re: Matrox Millenium G 200 SD 16 MD ("Bobby D. Bryant")
Re: my hdparm specs question (Tim Moore)
----------------------------------------------------------------------------
From: Bryan <Bryan@[EMAIL PROTECTED]>
Subject: Re: Mobo for linux server
Date: Sun, 25 Jul 1999 04:04:38 GMT
you couldn't be more wrong about dual celerons. the speedup is about
70% on kernel builds (as an example). for folks who need to compile a
lot of files (or similar activity), dual cpu system are a SERIOUS
contender in price/performance.
I just built an abit bp6 with 2 cel 433's for a system for work and it
builds the 2.2 kernel (with my set of options: scsi, lan, etc) in
about 2:30. my system at home (dual cel 300a o/c to 450 ea.) builds
that in 1:50. as a comparison, my older work dual p-pro 200/512k (the big
expensive one in its day) builds that kernel in about 4:00.
I can understand going from a dual cel to a dual xeon (or quad), but
dual cel to dual p2 or p3 just isn't a significant enough jump to
justify what intel charges.
and when the dual k7 comes out, -that- should be a killer system, far
overshaddowing what intel will have around at that time.
but for right now, a dual cel system is a -very- nice system for linux.
Axis <[EMAIL PROTECTED]> wrote:
: DON'T waste your money on a dual Celeron configuration--Celeron's
: onboard cache is not large enough to support dual processor utilization
: effectively. A single Celeron 333A will give you more than double the
: performance of a P166, and currently offers the best price/performance
: ratio on the market. If you need more horsepower, go with dual PII
: 400. ASUS also makes dual CPU boards.
--
Bryan, http://www.Grateful.Net - Linux/Web-based Network Management
->->-> to email me, you must hunt the WUMPUS and kill it.
------------------------------
From: Jeremy Buhler <[EMAIL PROTECTED]>
Subject: Occasional timeouts DMA'ing to IDE drive
Date: Sun, 25 Jul 1999 04:30:18 GMT
Every few days, I get one copy of the following lines in my syslog:
Jul 24 15:12:54 jbuhler kernel: hda: timeout waiting for DMA
Jul 24 15:12:54 jbuhler kernel: hda: irq timeout: status=0xd0 { Busy }
Jul 24 15:12:55 jbuhler kernel: ide0: reset: success
Occasionally, programs hitting the disk intensively (eg a big
untarring or rm -rf) will appear to pause for a second or two, but
otherwise, I haven't noticed any unusual behavior. I certainly
haven't seen any signs of data loss or flaky system components, and a
recent forced fsck showed the drive to be clean. Are the syslog
messages harmless, or are they signs of impending doom?
I've observed this behavior with Linux 2.2.10, running on a new (< 2
months) Dell P3/550 with 384 MB RAM, an Intel 82371AB PIIX4 chipset
(rev01), and a 22 GB IBM IDE drive. The drive is the master (hda) on
ide0 (irq 14, not shared with anything). hdb is empty, hdc is a
CDROM, and hdd is an ATAPI tape drive. The messages don't correlate
to any activity on hdc or hdd. The offending DMA's are almost certainly
either to my swap partition (hda5, 384 MB) or to my Linux root partition
(hda6, 15 GB), both of which are on the high-numbered cylinders of the
drive.
Here are my typical drive settings from hdparm -v and -i:
/dev/hda:
multcount = 16 (on)
I/O support = 1 (32-bit)
unmaskirq = 0 (off)
using_dma = 1 (on)
keepsettings = 1 (on)
nowerr = 0 (off)
readonly = 0 (off)
readahead = 8 (on)
geometry = 2748/255/63, sectors = 44150400, start = 0
Model=IBM-DJNA-372200, FwRev=J71OA30K, SerialNo=GX0GXFA1508
Config={ HardSect NotMFM HdSw>15uSec Fixed DTR>10Mbs }
RawCHS=43800/16/63, TrkSize=0, SectSize=0, ECCbytes=34
BuffType=3(DualPortCache), BuffSize=1966kB, MaxMultSect=16, MultSect=16
DblWordIO=no, maxPIO=2(fast), DMA=yes, maxDMA=2(fast)
CurCHS=2748/255/63, CurSects=44150400, LBA=yes
LBA CHS=687/1020/63 Remapping, LBA=yes, LBAsects=44150400
tDMA={min:120,rec:120}, DMA modes: mword0 mword1 mword2
IORDY=on/off, tPIO={min:240,w/IORDY:120}, PIO modes: mode3 mode4
UDMA modes: mode0 mode1 *mode2
I previously ran with unmaskirq set on, but changing that setting
apparently didn't fix the problem.
For reference, here is a typical output of hdparm -t -T. I think the
observed throughput is about what's expected for sustained transfers
in UDMA/33 mode.
/dev/hda:
Timing buffer-cache reads: 128 MB in 1.05 seconds =121.90 MB/sec
Timing buffered disk reads: 64 MB in 3.78 seconds =16.93 MB/sec
--
## Jeremy Buhler * peace through superior algorithms * U. Washington ##
------------------------------
Date: Sat, 24 Jul 1999 21:25:22 -0700
From: Tim Moore <[EMAIL PROTECTED]>
Subject: Re: Promise Ultra/66 EIDE Controller
> I use Redhat 6.0 and the experimental kernel 2.3.10, the card is recognized
> correctly.
Does it reflect 33MHz or 66MHz timing?
[tim@asus /raid]# uname -a
Linux asus 2.0.37 #8 Tue Jul 20 01:46:42 PDT 1999 i686 unknown
[tim@asus /raid]# cat /proc/pci
PCI devices found:
...
Bus 0, device 4, function 1:
IDE interface: Intel 82371AB PIIX4 IDE (rev 1).
Medium devsel. Fast back-to-back capable. Master Capable.
Latency=32.
I/O at 0xd800.
...
Bus 0, device 10, function 0:
RAID storage controller: Promise Technology IDE UltraDMA/33 (rev 1).
Medium devsel. IRQ 3. Master Capable. Latency=32.
I/O at 0xb800.
I/O at 0xb404.
I/O at 0xb000.
I/O at 0xa804.
I/O at 0xa400.
[dmesg]
...
ide: Intel 82371 (single FIFO) DMA Bus Mastering IDE # PIIX4
Controller on PCI bus 0 function 33
ide: timings == a307e377
ide0: BM-DMA at 0xd800-0xd807
ide1: BM-DMA at 0xd808-0xd80f
ide: PDC20246 UDMA Bus Mastering # U/33
Controller on PCI bus 0 function 80
ide: timings == 000003ee
ide2: BM-DMA at 0xa400-0xa407
ide3: BM-DMA at 0xa408-0xa40f
--
timothymoore "Everything is permitted. Nothing is forbidden."
bigfoot WS Burroughs.
com
------------------------------
From: [EMAIL PROTECTED] (John Esterson)
Subject: Help Kouwell Ultra scsi under linux
Date: Mon, 19 Jul 1999 22:10:45 GMT
I bought an Kouwell ultra scsi card WK-910U but can't run it under redhat 6.0
is there a driver for this card?
thanks is advance
------------------------------
From: David Koretzky <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.misc
Subject: Re: How do I change SCSI cards in Red Hat 6.0?
Date: Sun, 25 Jul 1999 01:09:01 -0400
Anthony Valentine wrote:
>
> I currently use and Advansys SCSI card (which works fine) but I woulf
> like to upgrade it to an Adaptec 2940UW. The aic7xxx.o module supports
> the Adaptec card, but I cannot get Red Hat to load this module on boot.
You have to recompile the kernel with support for the Adaptec as a
module.
Then you have to make a ram drive to load the module at boot time.
Otherwise you have to compile support in the kernel for it.
Look in the RedHat manual for how to build a ram drive.
The short version is
/sbin/mkinitrd /boot/initrd-image 2.x.x <- whatever kernel you are
running
Then in lilo.conf, reference the new ram drive in the
initrd=initrd-image
and the make sure to put in the new kernel image as well.
The manual gives a pretty good description of what you need. An
important note is that you MUST compile in support for a ramdrive (in
block devices), otherwise the module won't get loaded and you can't get
access to your drive. Also loopback device support is needed.
--
David Koretzky
[EMAIL PROTECTED]
------------------------------
From: [EMAIL PROTECTED] (Jeremy Fincher)
Subject: Re: "The LinuxStore" - are they good ??
Date: 25 Jul 1999 05:15:56 GMT
>e-biz will be coming out, allegedly, with a 200
>usa dollar computer...only spec was a 2 meg hd.
A 2 meg hd? for $200? What a deal!! :-)
Jeremy
==================================
If i ever forget to capitalize a proper noun, forgive me. i'm a big fan of ee
cummings
My ICQ # is 28153190. My AIM/AOL name is either jemfinch02 or Cassius80.
Have a good day, and good luck in your endeavors!
------------------------------
From: [EMAIL PROTECTED]
Subject: Can't get Microphone to work.
Date: Sun, 25 Jul 1999 05:43:27 GMT
Hi, I've been using Linux for about a year now; I consider myself fairly
competent, but one thing I've never managed to get working is my
microphone.
I have a Creative AudioPCI (ES1371, I believe) with support compiled in
(not a module). Audio playback works fine, but I have never gotten audio
recording to work at all. I have the microphone plugged into the correct
outlet, and I have my TV card looped into the audio in. If I speak into
the microphone, I can hear myself coming out of the speakers (in Linux),
so I know it's connected properly. However, nothing seems to work for
recording. For instance, I just tried the command:
dd bs=8k count=4 </dev/audio >/tmp/sample.au
while running as root (having killed esd). Nothing. It gives a file, but
it has no sound in it (au files work otherwise). Same thing for
/dev/dsp. One thing I did notice, though, is that if I have an
application using the TV card, that same command will record _that_
sound. "Great," I think. I'll just switch the plugs. Unfortunately, that
doesn't work. I don't even get the sound through the speakers when I
talk into the microphone after I do that. I've reread the howto, and it
doesn't seem to offer any advice on fixing this problem.
In case any of this helps:
[root@fizgig /dev]# ls -l audio* dsp*
crw-rw-rw- 1 root sys 14, 4 Jul 25 01:19 audio
crw-rw-rw- 1 root sys 14, 20 Jul 25 01:19 audio1
crw-rw-rw- 1 root sys 14, 3 Jul 25 01:19 dsp
crw-rw-rw- 1 root sys 14, 19 Jul 25 01:19 dsp1
Trying to use dsp1 or audio1 with dd gives "bash: /dev/audio1 [or dsp]:
Invalid argument".
So, can anyone help me here? Is there any more information I can
provide? /dev/sndstat doesn't exist. Or rather, it does, but I get a "No
such device" error when I try to access it.
Other than this, sound is perfect. It's just that I'd like to be able to
use my microphone.
Thanks,
Andrew
Sent via Deja.com http://www.deja.com/
Share what you know. Learn what you don't.
------------------------------
From: [EMAIL PROTECTED] (Nelson Ricciardi)
Subject: Re: Writing to Windows Partition From Linux Partition
Date: Sun, 25 Jul 1999 04:55:16 GMT
Hi there. I'm a newbie but I belive you can mount the dos/windows
partition and then print the file to it.
When you reboot your machine, the file will be there.
Easy,
On Thu, 22 Jul 1999 22:59:07 -0400, Wretch <[EMAIL PROTECTED]>
wrote:
>Hi -- I received a Lexmark 3200 color inkjet printer as a present,
>and long after I finally set the thing up on my home computer
>I had the horrific realization that Lexmark is yet another Microsoft
>minion, and printing directly from Linux (Redhat 5.1) seems hopeless.
>It's too late to return the printer, so I'm kind of stuck.
>
>So, I'm wondering if I can generate files in my linux partition
>and write them from there directly to my Windows 98 partition,
>and then print the files from there once I've rebooted to the
>Windows OS. If anyone can point me in the right direction (a link, any
>literature, etc.) I'd be greatly appreciative.
>
>Thanks,
>
>AC
------------------------------
From: Colin Wong <[EMAIL PROTECTED]>
Subject: Hardware Problems causing Sig 11....
Date: Sun, 25 Jul 1999 05:50:28 GMT
Hi everyone,
I have been experiencing hardware problems (I'm guessing) that have
caused my efforts to compile the kernel and module to 'crash' with a
signal 11 error most of the time. By most of the time I mean about 60+%
chance of a failed compilation. Here is a list of my specs, and what I
have done.....
I am running a P100 chip produced in Intel's SU099 group, with 32MB of
70ns 5V FPM RAM (uses SIEMENS chips with part number HYB514400BJ-70) on
an ASUS P/I-P55TP4XE (rev 2.1) motherboard.
The CPU itself comes with an "Intel certified" heatsink and fan glued on
it, and I have mounted a 3.7" brushless DC Fan directly on the metal
casing and provided sufficient air flow through the case, so overheating
of the CPU is not likely.
I have vacuumed the memory slots (in fact I initially took apart the
case and removed all dust) and I do not 'believe' the motherboard is at
fault.
I have attempted to run the CPU at 75MHz, but this is even worse... when
I attempt to compile and compilation fails, it usually crashes the
entire system, and requires a cold boot.
I do not have any extra 72 pin memory lying around which I could
test...... does anyone think I should go out and buy some?
Lastly, this Linux installation was previously working flawlessly in a
486 DX 33. The Pentium box is second hand, and the only common
components reused in the 'upgrade' were the hard drives, cd-rom, floppy,
sound card and 2 RTL8019 network cards. Memory was not reused b/c the
486 used 70-pin RAM common in older AST machines.
I hope that someone can help me, but I will port everything back to the
486 if necessary. (a flawless 486 is better than any flaky computer).
Thanks,
Colin Wong
(post responsed to comp.os.linux.hardware or email me at
[EMAIL PROTECTED])
------------------------------
From: Glenn Valenta <[EMAIL PROTECTED]>
Reply-To: [EMAIL PROTECTED]
Subject: Re: speed tests
Date: Sun, 25 Jul 1999 05:43:12 GMT
Justin Miller wrote:
>
> I'm going to be doing some tweaking to my system soon. I'll be adding
> another harddrive and reconfiguring the partitions and will be playing
> with overclocking. What I'm looking for is some sort of benchmark test
> so I can see how much performance I gain. I would like to test for
> inproved cpu power and inproved disk i/o independantly if possible.
>
> I was thinking that compiling a kernel might be a good test of cpu
> speed, and perhaps large file writes and reads, but I'm not sure of the
> best way to test that.
>
> Can anyone suggest anything?
>
> Thanks,
> Justin
I use lmbench. I can't find where I got it but it works for me.
Glenn Valenta Engineering @ http://www.coloradostudios.com
[EMAIL PROTECTED] http://ouray.cudenver.edu/~gavalent/
[EMAIL PROTECTED] Personal mail
[EMAIL PROTECTED] Work mail
------------------------------
From: "Bobby D. Bryant" <[EMAIL PROTECTED]>
Subject: Re: Matrox Millenium G 200 SD 16 MD
Date: Sun, 25 Jul 1999 01:42:40 -0500
Lars Amsel wrote:
> I just downloaded the newest XFree-Version for SuSE 6.0 with no effect. I
> think there should be no problems with my XF86Config. The monitor (Eizo F56)
> should display 1024x768 with no problems. Maybe you know a logfile which
> tells me what went wrong?
Do you want me to mail you my config file? I have a different monitor and use
a different resolution, but if you think it would be helpful to compare it
against yours, drop me a line and I'll send it to you.
Bobby Bryant
Austin, Texas
------------------------------
Date: Sat, 24 Jul 1999 21:18:15 -0700
From: Tim Moore <[EMAIL PROTECTED]>
Subject: Re: my hdparm specs question
> The rpm -qa line didn't work, but "hdparm -qa /dev/hda" did:
hdparm -qa is meaningless.
I was looking for the version you're running. My UDMA drive has similar
hdparm numbers as yours. Try a dd test similar to what I've done and
compare data rates to what you're seeing with hdparm. Change the count=
value to be about equal to your memory (300k for 300MB memory in my
case).
The values at the end indicate write and read speed for large,
sequential I/O about the same size as physical memory to a particular
partition. I got 9.25MB/s write, 10.55MB/s read. Note the read speed
is close enough to the hdparm value. You'll want to run the test
sequence several times and look for average repeatable values. This
will test the whole I/O subsystem (disk, disk cache, buffer cache, bus).
======================================
[tim@asus /raid]# uname -a
Linux asus 2.0.37 #8 Tue Jul 20 01:46:42 PDT 1999 i686 unknown
[tim@asus /raid]# df
Filesystem 1024-blocks Used Available Capacity Mounted on
/dev/hda5 46632 25329 18895 57% /
/dev/hda11 1095993 601007 438359 58% /usr
/dev/hda12 303251 64220 223370 22% /var
/dev/hda13 396500 353831 22188 94% /opt
/dev/hda14 754160 228286 486918 32% /home
/dev/hde1 482061 1754 480307 0% /tmp
/dev/hde6 497667 452453 45214 91% /kits
/dev/md0 5043117 448947 4594170 9% /raid
[tim@asus /raid]# dmesg | grep Mem
Memory: sized by int13 0e801h
Memory: 321648k/327616k available (744k kernel code, 384k reserved,
4840k data)
======================================
[tim@asus /raid]# time dd if=/dev/zero of=/tmp/write.test bs=1k
count=300k
307200+0 records in
307200+0 records out
0.220u 9.840s 0:34.00 29.5% 0+0k 0+0io 83pf+0w
[tim@asus /raid]# ls -l /tmp/write.test
-rw-r--r-- 1 root root 314572800 Jul 24 20:53 /tmp/write.test
[tim@asus /raid]# time dd if=/tmp/write.test of=/dev/null bs=1k
307200+0 records in
307200+0 records out
0.230u 6.890s 0:29.80 23.8% 0+0k 0+0io 76883pf+0w
[tim@asus /raid]# rm -vf /tmp/write.test
/tmp/write.test
[tim@asus /raid]# sync
[tim@asus /raid]# echo 314572800/34.00 | bc -q
9252141
[tim@asus /raid]# echo 314572800/29.80 | bc -q
10556134
======================================
[tim@asus /raid]# hdparm -tT /dev/hde1
/dev/hde1:
Timing buffer-cache reads: 64 MB in 0.61 seconds =104.92 MB/sec
Timing buffered disk reads: 32 MB in 3.35 seconds = 9.55 MB/sec
[tim@asus /raid]# hdparm -tT /dev/hde1
[tim@asus /raid]# hdparm -vi /dev/hde
/dev/hde:
multcount = 0 (off)
I/O support = 0 (default 16-bit)
unmaskirq = 0 (off)
using_dma = 1 (on)
keepsettings = 0 (off)
nowerr = 0 (off)
readonly = 0 (off)
readahead = 8 (on)
geometry = 790/255/63, sectors = 12692736, start = 0
Model=IBM-DHEA-36481, FwRev=HP6OA20C, SerialNo=SG0SG0K4
Config={ HardSect NotMFM HdSw>15uSec Fixed DTR>10Mbs }
RawCHS=12592/16/63, TrkSize=0, SectSize=0, ECCbytes=28
BuffType=3(DualPortCache), BuffSize=472kB, MaxMultSect=16, MultSect=off
DblWordIO=no, maxPIO=2(fast), DMA=yes, maxDMA=2(fast)
CurCHS=790/255/63, CurSects=12692736, LBA=yes, LBAsects=12692736
tDMA={min:120,rec:120}, DMA modes: sword0 sword1 sword2 mword0 mword1
mword2
IORDY=on/off, tPIO={min:240,w/IORDY:120}, PIO modes: mode3 mode4
>
> Linux johnjac 2.2.6 #95 Tue Apr 27 19:10:37 CDT 1999 i586 unknown
>
> /dev/hda:
> Timing buffer-cache reads: 128 MB in 4.24 seconds =30.19 MB/sec
> Timing buffered disk reads: 64 MB in 6.15 seconds =10.41 MB/sec
>
> /dev/hda:
> Timing buffer-cache reads: 128 MB in 5.91 seconds =21.66 MB/sec
> Timing buffered disk reads: 64 MB in 5.34 seconds =11.99 MB/sec
> Hmm.. suspicious results: probably not enough free memory for a proper
> test.
>
> /dev/hda:
> Timing buffer-cache reads: 128 MB in 3.58 seconds =35.75 MB/sec
> Timing buffered disk reads: 64 MB in 5.77 seconds =11.09 MB/sec
>
> /dev/hda:
> multcount = 0 (off)
> I/O support = 0 (default 16-bit)
> unmaskirq = 0 (off)
> using_dma = 1 (on)
> keepsettings = 0 (off)
> nowerr = 0 (off)
> readonly = 0 (off)
> readahead = 8 (on)
> geometry = 1313/255/63, sectors = 21095424, start = 0
> ---------------------------------
--
timothymoore "Everything is permitted. Nothing is forbidden."
bigfoot WS Burroughs.
com
------------------------------
** 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
******************************
