Linux-Hardware Digest #486, Volume #9 Mon, 22 Feb 99 17:13:28 EST
Contents:
Re: Partitioning crashed HD? (Robert Schiele)
Re: HP Colorado 5 or 8 GB success story, anybody? ("David R. Christianson")
Re: AGP Graphics card? (Jason Clifford)
Re: Old ISA ethernet card - Recomendations? ("Jonas")
Help dial on KDE ! (Kevin Chu)
Re: PCI 128 (Jose Urena)
Re: Caught signal 11 - what does this mean ? (Klaus Kocheisen)
Re: Installing Linux on system with VIA chipset (Mike Nugent)
Re: PPP !!! (Michael Trausch)
Re: Help dial on KDE ! (tc)
Re: Problem with Linksys network card? (John Strange)
Re: Anyone out there with a RedHat installation success story? (John Thompson)
Re: SMC Tulip (John Strange)
Re: Hardware to recommend? For state of the art Linux box? (Eric Lee Green)
Re: sda : unsupported sector size 16712447. (aurthur)
Trantor T160 SCSI host adaptor - Trouble! (Tomas Bystrom)
Re: Same Disk RAID and Mirroring (Andy Glew)
Re: Please explain IRQs 2 and 9 (Craig Ruff)
----------------------------------------------------------------------------
From: Robert Schiele <[EMAIL PROTECTED]>
Subject: Re: Partitioning crashed HD?
Date: Sun, 21 Feb 1999 15:14:11 +0100
Hans Kingma wrote:
>
> Hi there,
>
> I'm using the SuSe 5.1 distribution CD set. After setting up a
> partitioning scheme my first harddisk (containing the part table) is
> offline, even the BIOS doesn't recognises it anymore. Is this a possible
> mistake of a MBR write action? Or is it just a mere coincedence that my
> WD31200 crashed while doing the boot initialisation? Any ideas of how to
> revive the harddisk?
>
> Any help is welcomed.
>
> hans
It shouldn't be possible to crash a harddisk with a simple write
command!
Does the drive spin up? If no: check power connection!
Else check data cable.
You could also test the drive in an other system. If it does not work
there, let's hope the data was not that important. :-(
Greetings,
Robert
------------------------------
From: "David R. Christianson" <[EMAIL PROTECTED]>
Subject: Re: HP Colorado 5 or 8 GB success story, anybody?
Date: Sun, 21 Feb 1999 15:52:32 GMT
Has IDE-Tape support changed since kernel 2.0.36? I couldn't get HP 5GB
to work with that. Something about not supporting non-DRQ drives. The
whole box locked up when I tried to access the tape. Only the 8GB and
Seagate clone were supported. If it is supported under 2.2.1, I'll have
to toss it back in!
John Thompson wrote:
> I have gotten the Colorado 5GB IDE unit to work with kernel
> 2.2.1 using tar. I've been struggling with taper but no joy
> so far. But tar works fine so I feel less desparate about
> not being able to use taper.
>
> --
>
> -John ([EMAIL PROTECTED])
------------------------------
From: Jason Clifford <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.misc,comp.os.linux.x,comp.os.linux.setup
Subject: Re: AGP Graphics card?
Date: Mon, 22 Feb 1999 17:57:14 +0000
On Mon, 22 Feb 1999, David Buckley wrote:
> I currently have an SiS 6326 AGP card. It runs X ok, but not as well as
> my old ViRGE dose. I want to upgrade, has anyone any suggestions on
> what card to get? Unfortunately I have to run Windoze, for reasons that
> have to many buts in for the moment. Any suggestions, I`d like to get a
> Savage 3D, but there aren't any drivers yet. I'm going to be using it
> for a lot of graphics so any suggestions should account for 8mb+ of
> memory (4mb just don't cut it any more)
Matrox Millenium G200 with 8MB of RAM is pretty much the best in its
class. The ATI cards are also quite good.
Jason Clifford
Definite Linux Systems
http://definite.ukpost.com/
------------------------------
From: "Jonas" <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.misc,comp.os.linux.networking
Subject: Re: Old ISA ethernet card - Recomendations?
Date: Sun, 21 Feb 1999 14:59:54 +0100
Intel also makes a 100 Mbit isa card but it's very expensive. About 50 %
more than a PCI-card.
------------------------------
From: Kevin Chu <[EMAIL PROTECTED]>
Crossposted-To:
hk.comp.os.linux,comp.os.linux,comp.os.linux.help,comp.os.linux.x,uk.comp.os.linux,tw.bbs.comp.linux
Subject: Help dial on KDE !
Date: Mon, 22 Feb 1999 21:39:20 +0800
i am using SuSE 6.0
and i want to dial to ISP by KDE
after i dial to isp , it show a error message
"The pppd daemon died unexpextedly"
I am using IMS 56K (USR x2)
i also want to know how to dial to isp on command mode (tty1)
Thank you very very much ^_^
------------------------------
From: Jose Urena <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.setup,comp.os.linux.misc
Subject: Re: PCI 128
Date: Mon, 22 Feb 1999 13:11:18 -0500
Assuming you mean AudioPCI128 and Compatibles
the drivers are part of the linux 2.2.x files, you might have to toggle
on/off the prompting for new drivers
use AudioPCI or Creative Labs AudioPCI
there is no distinction between the 64 and 128 intrument card, Why? because
the intruments are software based and that is not part of the driver
David Buckley wrote:
> Has any one got any idea where to get drivers, and how to get a PCI 128
> working? I've got it working in windoze 95 & NT, but would like to get
> it working in linux.
------------------------------
From: Klaus Kocheisen <[EMAIL PROTECTED]>
Subject: Re: Caught signal 11 - what does this mean ?
Date: Mon, 22 Feb 1999 19:46:32 +0100
Jim McIntyre schrieb:
>
> I'm tyring to run X with Rh 5.1, using an S3 Virge DX 4Mb, and a 17"
> CyberVision monitor.
> When I run startx, I get a monochrome screen appears briefly, and then I
> get the message "caught signal 11". I also get the message xinit:connect
> refused errno111, and xinit no such process (errno3). Can anybody tell me
> what these mean?
> I'm pretty the problem is with my monitor, but I don"t know what to do>
> Thanx in advance.
> Jim McIntyre.
The errors errno* are system error numbers defined in
/usr/include/errno.h or other errno.h files included by
/usr/include/errno.h. I think these errors are caused by the first error
occuring in your X startup (caught signal 11)
"caught signal 11" means, that your X server was terminated by the
operating system by signal 11 (SIGSEGV - segmentation violation), which
simply means that your X server crashed.
------------------------------
From: Mike Nugent <[EMAIL PROTECTED]>
Subject: Re: Installing Linux on system with VIA chipset
Date: Sun, 21 Feb 1999 11:35:48 -0500
Tim wrote:
> Lisa Charles <[EMAIL PROTECTED]> writes:
>
> > Hello All
> > I just want to clarify. I am trying to install Linux on a system with an
> > EPOX MVP3C motherboard, with a VIA chip set and AWARD BIOS, using an AMD
> > K6-3D processor and IDE drives.
> > I have created the boot disks, but when I boot from it the devices are
> > not recognized. The primary drive on the primary channel is found but
> > the primary drive on the secondary channel ( cdrom) is not found. Come
>
I don't think this is an issue with the chipset/bios. I run a amd k6-3d-300
on an FIC-2012A with a VIA and Award and also a k6-2-400 on a FIC-VA503+ with
a VIA and award. I've had absolutely no problems with the ide drives at all.
> Try at the boot prompt;
> LILO>linux hdc=cdrom
> or whatevere the appropriate image name is. AFAIK this is
> right for RedHat.
>
> > to that, the ethernet card ( NE2000 clone ISA ) is also not found. I
> > also get some messages about unknown PCI devices but they scroll of
> > the screen so I am not sure what it says. Since my cdrom is not
> > detected I can not continue with the installation. I have also
> > considered the possibility of trying to mount the cd on a windoze
> > machine and sharing it so that I can do an install with an SMB
> > source, but my ethernet card is not found. So for now I'm screwed it
> > seems. Any help will be appreciated TIA Thanks for the suggestion
> > Trever but that won't help William
>
> I've never had any joy with the NE2000 clone cards. Sorry.
>
The problems I've had with ne2* clones are mostly due to the kernel not
looking at what IO port it's on. Look in /usr/src/linux/drivers/net/ne.c and
check line 65-70 or so. If your io port is not there, add it.
Hope this helps.
-Mike Nugent
--
"Who took the cork off my lunch?"
-W.C. Fields
------------------------------
From: Michael Trausch <[EMAIL PROTECTED]>
Crossposted-To:
hk.comp.os.linux,comp.os.linux,comp.os.linux.help,tw.bbs.comp.linux,uk.comp.os.linux
Subject: Re: PPP !!!
Date: Mon, 22 Feb 1999 09:02:27 -0500
=====BEGIN PGP SIGNED MESSAGE=====
use YaST to configure your PPP, it's under "Networking"
- Mike
=====================================================================
Michael B. Trausch [EMAIL PROTECTED]
V: (419) 838-8104 F: (815) 846-9374
"Curiosity is the very basis of education and if you tell me that
curiosity killed the cat, I say only the cat died nobly."
- Arnold Edinborough
** Apparently, there are problems with my PGP key. This message **
** will go away after I have fixed these problems **
If you do not have my public PGP key, you are encouraged to obtain it
from my website at http://www.wcnet.org/~mtrausch/mykey.zip. You need
to have PGP 5.0i or newer to use the key.
=====================================================================
On Mon, 22 Feb 1999, Kevin Chu wrote:
> can you send the PPP script file to me ?
>
> and use what software to dial to the ISP ?
>
> ??? can you help me
> i am using SuSE 6.0 Help Help !!!
> Thank you very very much ^_^
>
>
>
=====BEGIN PGP SIGNATURE=====
Version: PGPfreeware 5.0i for non-commercial use
Charset: noconv
iQCVAwUBNtFjcBjf2mx8Omu9AQGbSgP/eGB19+QJhIlVSnOsywQrozVQvrGflMKZ
k3c0hI0lkSGxlac+MRDMKAEIfZMATb/bL6pHfI0pmca/EYxuKGWDcnAmyzvLhgu2
AYhg+q9bK94XYSmuxWWSpYsoLriwkkxcD88AlsclVj8UyHzYXk+cR/Y27+IRBkG5
B9jRMwfkeH0=
=SGgI
=====END PGP SIGNATURE=====
------------------------------
From: tc <[EMAIL PROTECTED]>
Crossposted-To:
hk.comp.os.linux,comp.os.linux,comp.os.linux.help,comp.os.linux.x,uk.comp.os.linux,tw.bbs.comp.linux
Subject: Re: Help dial on KDE !
Date: Mon, 22 Feb 1999 17:02:55 +0100
Hi!
Remove "lock" in the file "options" in /etc/ppp/ (or thereabouts).
Worked for me.
bye
Thorsten
Kevin Chu wrote:
>
> i am using SuSE 6.0
> and i want to dial to ISP by KDE
>
> after i dial to isp , it show a error message
> "The pppd daemon died unexpextedly"
> I am using IMS 56K (USR x2)
> i also want to know how to dial to isp on command mode (tty1)
> Thank you very very much ^_^
------------------------------
From: [EMAIL PROTECTED] (John Strange)
Subject: Re: Problem with Linksys network card?
Date: 21 Feb 1999 16:36:50 GMT
You need a new tulip.c driver I use .090f
http://www.linksys.com/support/solution/nos/linux.htm
http://cesdis.gsfc.nasa.gov/linux/drivers
snowman ([EMAIL PROTECTED]) wrote:
--
While Alcatel may claim ownership of all my ideas (on or off the job),
Alcatel does not claim any responsibility for them. Warranty expired when u
opened this article and I will not be responsible for its contents or use.
------------------------------
From: John Thompson <[EMAIL PROTECTED]>
Subject: Re: Anyone out there with a RedHat installation success story?
Date: Mon, 22 Feb 1999 12:25:26 -0600
JT wrote:
>
> I'm planning to build a low cost (under $600) RedHat system and need advise
> as to all components (the entire enchilade - as we say here in Texas).
> I am hopping to hear from someone who has simply followed the RH
> installation instructions and finally shutdown with a smile. Is there
> anyone out there with this experience or is this system as cantankerous as
> most sites seem to indicate. I'm hopping to setup as many as 1000 machines
> per year and that why I'm looking for simplicity. I have an opportunity to
> disrupt the Microsoft tax system on a corporate level but I need help to do
> it.
>
> If you can help please document every component in detail. This system
> requires the following options: sound card, CD ROM and 56k modem.
>
> Additionally I must incorporate a touch screen which I think few of you have
> experience with.
I have installed RedHat on two machines so far and was
impressed at how easily it installed each time. The first
was RH 5.0 on a 5x86/133 with 32MB, SCSI HD, CDROM and tape,
and S3 VLB video and a Newcomm 33.6 modem. I didn't bother
with sound on that machine.
The second (only a few short months later) was RH5.1 on a
P2/350 with 128MB and IDE HD, CDROM and tape and a Cardinal
56k modem. I did have trouble with the tape drive and kernel
2.0.34 but going to kernel 2.2.1 solved that. Sound only
required a minor tweak: the sndconfig program wouldn't let
me choose the proper IRQ for the sound board, so I just
manually edited /etc/conf.modules to put it right.
If you're looking to set up multiple identical machines you
might want to burn your own installation CD once you get it
working on one machine.
--
-John ([EMAIL PROTECTED])
------------------------------
From: [EMAIL PROTECTED] (John Strange)
Subject: Re: SMC Tulip
Date: 21 Feb 1999 16:26:57 GMT
tulip.c 0.90f is at
http://cesdis.gsfc.nasa.gov/linux/drivers/
matthew.r.pavlovich.1 ([EMAIL PROTECTED]) wrote:
: I am able to get the driver to load and recognize the card, no link and no
: data connection. Any ideas? Kernel 2.2.1 w/ driver patched from 0.89 to
: 0.90 (latest)
: -Matt
--
While Alcatel may claim ownership of all my ideas (on or off the job),
Alcatel does not claim any responsibility for them. Warranty expired when u
opened this article and I will not be responsible for its contents or use.
------------------------------
From: [EMAIL PROTECTED] (Eric Lee Green)
Subject: Re: Hardware to recommend? For state of the art Linux box?
Date: 21 Feb 1999 17:40:00 GMT
Reply-To: [EMAIL PROTECTED]
On Sat, 20 Feb 1999 16:22:47 -0700, Plinio Barbeito <[EMAIL PROTECTED]>
wrote:
>My question is slightly different -- I'm trying to figure out what is
>the most *maxed-out* system for Linux I can build that will work
>perfectly with available/supported drivers (I don't mind playing around
>with kernels as long as I know it's going to work). Therefore, I
>address this question to anyone that has tried to add fairly late model
>parts to their system and succeeded.
>
>I'd also like to know if the manufacturer has a good record of linux
>support and which version worked with the part in question. Even *with*
>slightly older hardware I know a linux system is still going to blow
>away some of the OSes bloating out of control out there...
>
>Motherboard: BX; Have you been able to get a 133Mhz bus clock board with
>heat alarms, sleep mode clock speed reduction, AGP (2x AGP?) to work?
Not at 133Mhz. We don't overclock or etc. because our systems are designed
for 24/7 operation (i.e. they're never turned off, ever).
>Hard disk: What is the largest overall size or partition that can be
>used?
You can't afford it :-). I think it's in the terabyte range. The biggest
setup I have personally done had ten 18gb Barracudas in it for a total
capacity of 180gb. This was attached to a 5-channel RAID controller.
>SCSI host adaptor: Looks like Mylex and Advansys might have the best
>linux support up to now, but does anyone have a multichannel 80
>Mbyte/sec card working? Or one that will allow adding a slower device
Cards based upon the Symbios chipsets are also a best bet, because
Symbios (now owned by LSI Logic) is a chip vendor, not a board
vendor. That means that they release all the technical details so that
people can build boards with it. Meaning better drivers. Intraserver
and DigitalScape are two companies that explicitly support Linux and
release cards based on these chipsets.
We have successfully used an Intraserver 6201 board with Linux. This
is a 2-channel board based on the Symbios 53c896 chipset. However, it
did require Gerard's latest 53c8xx driver. Under some BIOS's it may
require a slight hack to the 2.0 kernel's PCI bus scan to get both
functional devices reported. On the 2.2 kernel, the PCI bus scan is a
"make xconfig" option, and if your BIOS does not report both
functional devices, you can configure the kernel to do its own BIOS
scan. This works better with most modern motherboards.
For more channels, the 5-channel RAID controller that I used is an
ICP-Vortex GDT controller. It works great, has excellent performance,
etc.
>(CD) to it without throwing a wrench into the transfer rates? I've been
>told that it's better not to mix and match with IDE, is this true?
>Anyone have RAID working?
ICP-Vortex directly supports Linux. Their GDT driver was written by
them, themselves, and they have a Linux version of all the support
tools for their GDT RAID controllers. As far as I can tell they were the
second RAID controller to be supported under Linux (DPT was the first).
Most Ultra2 cards based on the Symbios chipset have a "SCSI Buddy"
chip that creates a separate bus for the CD-ROM and tape drives. It is
still on the same channel, but a separate bus. Thus the SCSI chip
slows down only when it's talking to the CD-ROM, and runs full speed
when talking to the rest of the hard drives. The LVD ICP-Vortex boards
use this "SCSI Buddy" chip on one of the channels so that a SCSI CD-ROM
can be connected without affecting the rest of the bus.
There's no problem with running IDE CD-ROMs with SCSI hard drives. But
do not try to use an IDE ZIP and SCSI hard drives. This confuses many
BIOS's (which will try to boot off the IDE ZIP rather than off the
SCSI hard drive), as well as confusing the Red Hat installer. Also, if
you have both SCSI and IDE hard drives, the installer and many BIOS'es
will automatically default to the IDE.
>Memory: PC100 SRAM DIMMS ought to be transparent, but let me know if
>they're not. Max size is not the biggest concern, but if you've
>populated your board to 1G and can use all of it I wouldn't mind hearing
>about it.
Boards based upon the BX chipset need *BUFFERED* 256mb SDRAM modules
to get to 1G. Many of the older ones will also require a BIOS update
to properly configure that much memory. I have successfully populated
an ASUS P2B-D board to 1G before (it was an older board that did
require a BIOS update). The default Linux kernel will only use around
900mb of that memory, but can be fairly easily patched to access up to
2gb (at the expense of some virtual memory space). The patches are
floating around on the Internet.
>Sound: Have the PCI sound cards reached "supported" status yet (they're
>supposed to be less interrupt intensive than the ISA cards and therefore
>more conducive to performance)?
Sort of. There is actually no difference in performance unless you use
bigger buffers than the default OSS drivers will do.
>Network: 100Mbit cards are mainstream by now, right?
Yes, but be careful. With the purchase of Digital Semiconductor by Intel,
cards based upon the Digital Tulip chipsets are becoming rare and far
between. Most former Tulip customers have switched to something called the
"Lite-On PNIC", which is a sort of cut-rate Tulip clone that does not work
well at all.
So far, the two most reliable 10/100 cards are also the two most
expensive -- the Intel EtherExpress Pro/100, and the Digital/Cabletron
DE-500. Both will set you back somewhere around $100-$120 retail (less
wholesale, but not much less). Some people swear by 3Com, and some
people swear at 3Com, so I avoid them. Basically, for most uses the
EEPRO100 is slightly cheaper and has slightly better performance, but
the DE-500 is preferable if you're going to be using gated (apparently
the eepro100 driver has some problems with that). We sell the DE-500
because we don't know what use they're going to make of the machine
and we don't want to have to stock two different cards (one for gated,
and one for everything else), but for most people the EEPRO100 is also
an excellent choice.
--
Eric Lee Green [EMAIL PROTECTED] http://www.linux-hw.com/~eric
"Microsoft will compete ... by adding features" -- Ed Muth, Microsoft
------------------------------
From: aurthur <[EMAIL PROTECTED]>
Subject: Re: sda : unsupported sector size 16712447.
Date: Sun, 21 Feb 1999 17:42:37 +0000
[EMAIL PROTECTED] wrote:
> In article <[EMAIL PROTECTED]>, Shane Bearham
><[EMAIL PROTECTED]> wrote:
> >[EMAIL PROTECTED] wrote:
> >>
> >> Linux kernel 2.0.34 or 2.0.36, same problem.
> >>
> >> with "modprobe aha152x.o aha152x="0x140,10,7,1" I get the following:
> >>
> >> aha152x: BIOS test: passed, detected 1 controller(s)
> >> aha152x0: vital data: PORTBASE=0x140, IRQ=10, SCSI ID=7, reconnect=enabled,
> >> parity=enabled, synchronous=disabled, delay=100, extended
> > translation=disabled
> >> aha152x: trying software interrupt, ok.
> >> scsi2 : Adaptec 152x SCSI driver; $Revision: 1.18 $
> >> scsi : 1 host.
> >> Vendor: IMG Model: ZP10PU Rev: J6
> >> Type: Direct-Access ANSI SCSI revision: 02
> >> Detected scsi removable disk sda at scsi0, channel 0, id 5, lun 0
> >> sda : unsupported sector size 16712447.
> >> SCSI device sda: hdwr sector= 16712447 bytes. Sectors= 0 [0 MB] [0.0 GB]
> >> sda: Write Protect is off
> >> sda:scsidisk I/O error: dev 08:00, sector 0, absolute sector 0
> >> unable to read partition table
> >>
> >> Any idea what the problem is?
> >
> >I believe if you enable extend translation... that should fix your
> >problem
> >Unfortunately I cannot advise on this bit
>
> Nope, doesn't help. I have tried all combinations of the flags and still get
> the same error.
unfortunately, I had the same problem with an IDE hard drive just last week. It seems
to be
nothing more than an incompatiblity between the hard drive's bios and linux. the only
way I fixed
it was to switch brand of hard drive - once I did it came right up!f
------------------------------
From: [EMAIL PROTECTED] (Tomas Bystrom)
Subject: Trantor T160 SCSI host adaptor - Trouble!
Date: 21 Feb 1999 17:38:30 GMT
Reply-To: [EMAIL PROTECTED]
I recently stumbled ovar an old SCSI controller and a CD-ROM, the SCSI
card was a Trantor T160 (NCR 53C400-based) cadr that most probably shipped
with the CD-ROM reader, a NEC CDR-501.
The kernel (2.2.1) seems to detect the card and the CD-ROM(s) fine, snip
from dmesg output (right now I have two CD-ROM readers connected to check
termination and such):
scsi0 : at port 0x368 irq 10 options CAN_QUEUE=16 CMD_PER_LUN=2 release=1
generic options AUTOPROBE_IRQ AUTOSENSE PSEUDO DMA USLEEP, USLEEP_POLL=1
USLEEP_SLEEP=20 generic release=7 ncr53c400 release=2
scsi0 : Generic NCR5380/53C400 Driver
scsi : 1 host.
Vendor: NEC Model: CD-ROM DRIVE:501 Rev: 2.2
Type: CD-ROM ANSI SCSI revision: 02
Detected scsi CD-ROM sr0 at scsi0, channel 0, id 1, lun 0
Vendor: TOSHIBA Model: CD-ROM XM-3701TA Rev: 3415
Type: CD-ROM ANSI SCSI revision: 02
Detected scsi CD-ROM sr1 at scsi0, channel 0, id 6, lun 0
scsi : detected 2 SCSI cdroms total.
Uniform CDROM driver Revision: 2.52
Also, the entries in /proc/scsi seems reasonable
> cat /proc/scsi/scsi
Attached devices:
Host: scsi0 Channel: 00 Id: 01 Lun: 00
Vendor: NEC Model: CD-ROM DRIVE:501 Rev: 2.2
Type: CD-ROM ANSI SCSI revision: 02
Host: scsi0 Channel: 00 Id: 06 Lun: 00
Vendor: TOSHIBA Model: CD-ROM XM-3701TA Rev: 3415
Type: CD-ROM ANSI SCSI revision: 02
>cat /proc/scsi/g_NCR5380/0
SCSI host number 0 : Generic NCR5380/NCR53C400 Scsi Driver
Generic NCR5380 driver version 1
NCR5380 core version 7
NCR53C400 extension version 2
NCR53C400 card detected
NCR53C400 pseudo DMA used
Using port mapping at port 0x368, on interrupt 10
T:1 CD-ROM NEC CD-ROM DRIVE:501 2.2
0 kb read in 0 secs
0 kb written in 0 secs
T:6 CD-ROM TOSHIBA CD-ROM XM-3701TA 3415
0 kb read in 0 secs
0 kb written in 0 secs
REQ not asserted, phase unknown.
No currently connected command
issue_queue
disconnected_queue
However, when i try to mount the CD-ROM (or play audio CDs with workman)
the machine locks up, hard! I have to reset it to boot, also the CD-ROM
readers don't rev down until their power is completely cut. Right before
the system locks up, something like this is displayed:
Command aborted due to timeout [PID info]
NCR5380: Coroutine isn't running!
STATUS_REG: 07,PARITY,IO,SEL
BASR: 10
ICR: 00
MODE: 00
scsi0: REQ not asserted, phase unknown.
scsi0: No currently connected command
scsi0: issue_queue
scsi0: disconnected_queue
scsi0: command=Read(10) 00 00 00 00 10 00 00 01 00
or
scsi0: command=40(0x28) 00 00 00 00 10 00 00 01 00
It's probably a lame question, I am however dumbfounded. It does not seem
to be IRQ or I/O-adress conflicts (after looking in
/proc/{interrupts,ioports}. The chain is terminated in both ends (no
difference in the behavior if the SCSI host adaptor card is at the end with
CD-ROMS connected to the internal or external connectors or if the card is
in the middle of the chain with the CD-ROMs connected both to the internal
and external connectors, of course with termination resistors moved
accordingly).
I have seen news posts with the same behavior with Trantor T130, but I
did not find replies. Also, there was (in 1996) a post about patches to
the SCSI-drivers in kernel 1.3.66 for the T160, but the ftp-site reported
did not exist anymore.
The rest of the system is more or less vanilla Debian hamm (2.0).
Any hints would be greatly appreciated! Also reports from people that got
their T160 to work would be nice.
/Tomas Bystr�m
--
Caps and foobar are normally not parts of my address.
------------------------------
From: Andy Glew <[EMAIL PROTECTED]>
Crossposted-To: comp.arch,comp.arch.storage,alt.os.linux,comp.periphs
Subject: Re: Same Disk RAID and Mirroring
Date: Sun, 21 Feb 1999 11:51:38 -0600
> In general, my experience is that drives tend to stick the ECC at the end of
> the sector, since it is mainly used for detection, rather than correction.
This is pretty much the same as all of the ECC codes I am familiar
with from memories and busses: the data bits are in one place, and
the ECC bits are at the end.
However, you said something different: you said that an error in
the data bits could be recovered from, but an error in the ECC bits could
not be detected or recovered from. *That* is quite surprising - it means
that the ECC bits are not themselves checked or corrected. My suspicion
is that you are wrong here. but I am willing to be corrected.
In memory and interconnect ECC codes the data bits can be looked
at straight, and the ECC bits used only to detect an error, and to correct
an error once detected. This is "optimistic" ECC - if an error is detected,
you somehow have to stop the incorrect data bits from being used,
and either retry or correct using the ECC. The alternative is "flow-through"
ECC - you never send the data bits on right away, but you always check
the ECC and correct any error. "Flow-through" ECC has simpler pipeline
control, and if the correction can be done at high speed may have equivalent
bandwidth. "Optimistic" ECC has lower latency,; if the ECC correction is
hard to calculate, detection may be done at full bandwidth, but correction may
not be; and it has more complicated pipeline control. In at least one of the
chips I am familiar with, ECC had to be disabled in first silicon because
the pipeline control for optimistic ECC was broken, not the ECC itself.
Therefore, the data bits and the ECC bits may be different in how errors
affect TIMING of results. However, in all of the codes I am familiar with,
the data bits and the ECC bits are identical wrt how detectable and correctable
errors in them are.
> oing this simplifies the data path in the drive electronics somewhat:
> stream the data bit by bit from the serializer directly into the drive's
> buffer via the ECC logic, and then compare the calculated ECC against the
> observed codes read off the media.
>
> In most cases, the error handling logic uses retries to get the data, only
> resorting to the ECC code when that avenue is exhausted.
This sounds pretty consistent with "optimistic" ECC.
------------------------------
From: [EMAIL PROTECTED] (Craig Ruff)
Subject: Re: Please explain IRQs 2 and 9
Date: 22 Feb 1999 14:06:03 -0700
In article <[EMAIL PROTECTED]>,
Rodney M. Bates <[EMAIL PROTECTED]> wrote:
>What is the story on IRQ 2 and 9? Can I assign a device to
>neither, both, only one? only a specific one?
IRQ 9 is usable. IRQ 2 is the cascade for the second interrupt
controller, and effectively proxies IRQ 8-15.
--
Craig Ruff NCAR [EMAIL PROTECTED]
(303) 497-1211 P.O. Box 3000
Boulder, CO 80307 Amateur Call KI0NO
------------------------------
** 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
******************************
