[294] in linux-scsi channel archive
Re: scsi problems with debian
daemon@ATHENA.MIT.EDU (Stuart Marshall)
Fri Jun 23 18:25:41 1995
Date: Fri, 23 Jun 95 14:41:30 PDT
From: Stuart Marshall <stuart@lensing.physics.ucsb.edu>
To: debian-user-request@Pixar.com, linux-scsi@vger.rutgers.edu,
ncr53c810@cs.colorado.edu
Hi,
Thanks very much for the responses I have received so far. But I am still
not fixed! Let me apologize for the length of this message. If you are
interested in the problem the read on.
I include below the responses that I got and an attempt to
answer any questions that were asked. Also at the end of the message
is the "dmesg" output from booting that may include more info than I have
already provided. Particularly worisome is the beginning:
bios32_init : BIOS32 Service Directory structure at 0x000e8130
bios32_init : BIOS32 Service Directory entry at 0xf382c
bios32_init : BIOS32 Service Directory structure at 0x000f0130
bios32_init : multiple entries, mail drew@colorado.edu
pcibios_init : PCI BIOS revision 2.00 entry at 0xf3800
Probing PCI hardware.
Calibrating delay loop.. ok - 33.55 BogoMips
But my old setup said exactly the same thing about multiple entries and as
I have said it worked fine for 9 months..
I have been able to make the system die on reads and writes. The latest
was:
#cat /usr/bin/* > /dev/null
cat: /usr/bin/X11: Is a directory
Kernel panic: EXT2fs panic (device 8/2): ext2_write_inode: unable to read i-node block - inode=63273, block=253962
Kernel panic: kernel stack overflow
I have also had it die on a long write doing cp -R /usr /usr/local/tmp.
And another sequence just as e2fsck was finishing up was:
scsi0: illegal instruction 00209870: 0x00000000 0xc0000004
scsi0: DSP=0x209870, DCMD|DBC=0xff, DSA=0x212e40
DSPS=0x0, TEMP=0x2099b0, DMODE=0x80,
o DNAD=0x0
Kernel panic: mail drew@Colorado.EDU
In swapper task - not syncing
I would like to hear the appropriate termination scheme for 2 (or n) internal
scsi devices. They come with resistor packs installed. In my case I called
quantum and a not terribly knowledgeable person suggested that only the last
drive should have the terminating resistor packs. That sounded correct to
me and it is the current configuration.
It is difficult for me to believe that this is a hardware problem because
the same hardware worked just fine under the old 1.1.19 kernel.
Please let me know what you think is going on if you are interested. If
I find the cause (likely something I did!) I will post the answers.
thanks very much,
Stuart
#------ responses that people sent are indented with >>
#------ and my answers are marked with a *
from: grogg@holly.ColoState.EDU
>>I get this exact same error whenever I upgrade/recompile the kernel.
>>ASUS PCI/I-486SP3 (no G)
>>Slack 2.x, kernel 1.1.x - 1.2.x
>>I'd be interested to know what's going on, too.
from: werner@itap.physik.uni-stuttgart.de
>>On wich kernel version the system dead occures? Whats
>>about 1.2.10?
*kernel version is 1.2.1 from debian installation.
*kernel 1.2.10 is available but I haven't tried to upgrade to it.
>>What version of the second extended file system tools
>>do you use for making ext2fs partition?
*I did not repartition the drive but used the existing partitions made
last september. I did reformat the linux partition however. Also I
reinitialized the swap partition.
*the ext2fs package is labeled e2fsprogs-0.5b-2.deb. So the version appears
to be 0.5b.
>>And, whats about other interesting things like
>> Setting of the motherboard/BIOS??
>> what scsi-disk??
>> kernel configuration??
*the two scsi disks are quantum lps540 and lt730. the bios is ami and
the ncr card has a bios also? It is 14 months old and labeled 3.01.04.
After the computer boots should the ncr bios matter?
The kernel is a large kernel that is the installation kernel for the
debian distribtution. It has the EATA whatever that is.
The output of "dmesg" appears at the end of this message and that
may shed more light on the kernel configuration.
from: berg@POOL.Informatik.RWTH-Aachen.DE
>> scsi: DANGER: command in running list, can not abort
>>Two things are happening here:
>>1. The current NCR driver can't deal with problems on the SCSI bus, if a
>> problem occurs, it cannot recover.
>>2. The fact that problems occur at all, usually indicates that your
>> cabling is bad, over- or underterminated.
*I have two internal drives. First has resistor packs removed, second
has resistor packs installed. I had same problem with resistor packs
installed on first drive but have reformated since then.
*I put my old cable back in just in case I made a bad cable. I still have
the same problem. ( I made the new cable so I could install and extra device.)
from: dwoodard@Quasar.Tach.Net
>>Also, certain drives may be incompatible. On approach is to remove all
>>the scsi drives and add them back one at a time. I had problems because
>>one of my external devices was internally terminated, and I also had an
>>external terminator on the bus.
*Given that these two drives worked fine together under my old setup
(slackware with 1.1.19 kernel) drive incompatibility seems unlikely.
I am confused about the termination scheme however since it worked with
both drives having their resistor packs installed before.
I have not actually removed that second drive and terminated the first.
from: hjl@nynexst.com
>>I posted a note last week about my problem with NCR 53C810 SCSI card
>>and Quantum 2 GB SCSI-2 hard drive. Thanks to all who sent me
>>suggestions. What I did was I low level formatted my Quantum
>>drive under DOS, which was from a Mac mail-order vendor. It seems
>>to work ok now.
>>>From what I have seen, I think
>>1. Always do a low level format on a new SCSI hard drive.
>>2. The error handling code in the generic/NCR 53C810 SCSI driver
may not work at all.
*This is confusing. You mean I should reformat the disk under dos and
the make an ext2 filesystem on that? I was using the installation program
under debian and don't know for sure what commands are used but I could
certainly do that part in a shell if it helps.
What is the distinction low level formating and making a file system?
#--------------------------- dmesg output ----------------------------
Console: colour EGA+ 80x25, 1 virtual console (max 63)
bios32_init : BIOS32 Service Directory structure at 0x000e8130
bios32_init : BIOS32 Service Directory entry at 0xf382c
bios32_init : BIOS32 Service Directory structure at 0x000f0130
bios32_init : multiple entries, mail drew@colorado.edu
pcibios_init : PCI BIOS revision 2.00 entry at 0xf3800
Probing PCI hardware.
Calibrating delay loop.. ok - 33.55 BogoMips
Serial driver version 4.11 with no serial options enabled
tty00 at 0x03f8 (irq = 4) is a 16450
tty03 at 0x02e8 (irq = 3) is a 16550A
Floppy drive(s): fd0 is 1.44M
FDC 0 is a 8272A
scsi-ncr53c7,8xx : at PCI bus 0, device 1, function 0
scsi-ncr53c7,8xx : NCR53c810 at memory 0xffbfff00, io 0xfc00, irq 9
scsi0 : using io mapped access
scsi0 : using initiator ID 7
scsi0 : using level active interrupts.
scsi0 ; burst length 8
scsi0 : using 40MHz SCSI clock
scsi0 : m_to_n = 0x90, n_to_m = 0xa0, n_to_n = 0xb0
scsi0 : NCR code relocated to 0x20973c
scsi0 : testing
scsi0 : test 1 started
scsi0 : tests complete.
EATA (Extended Attachment) driver version: 2.31a
developed in co-operation with DPT
(c) 1993-95 Michael Neuffer neuffer@goofy.zdv.uni-mainz.de
Registered HBAs:
HBA no. Boardtype: Revis: EATA: Bus: BaseIO: IRQ: DMA: Ch: ID: Pr: QS: SG: CPL:
scsi0 : NCR53c{7,8}xx (rel 3)
scsi : 1 host.
Vendor: QUANTUM Model: LPS540S Rev: 5900
Type: Direct-Access ANSI SCSI revision: 02
Detected scsi disk sda at scsi0, id 0, lun 0
Vendor: QUANTUM Model: LIGHTNING 730S Rev: 241E
Type: Direct-Access ANSI SCSI revision: 02
Detected scsi disk sdb at scsi0, id 1, lun 0
scsi : detected 2 SCSI disks total.
SCSI Hardware sector size is 512 bytes on device sda
SCSI Hardware sector size is 512 bytes on device sdb
lance.c: PCI bios is present, checking for devices...
Memory: 14876k/16384k available (848k kernel code, 384k reserved, 276k data)
This processor honours the WP bit even when in supervisor mode. Good.
Swansea University Computer Society NET3.019
Swansea University Computer Society TCP/IP for NET3.019
IP Protocols: ICMP, UDP, TCP
eth0: SMC Ultra at 0x300, 00 00 C0 90 9C 8E, IRQ 11 memory 0xc8000-0xcbfff.
smc-ultra.c:v1.12 1/18/95 Donald Becker (becker@cesdis.gsfc.nasa.gov)
Checking 386/387 coupling... Ok, fpu using exception 16 error reporting.
Checking 'hlt' instruction... Ok.
Linux version 1.2.1 (imurdock@debra.debian.org) (gcc version 2.6.3) #1 Tue Mar 21 12:31:42 EST 1995
Partition check:
sda: sda1 sda2 sda3 sda4
sdb: sdb1 sdb2
VFS: Mounted root (ext2 filesystem) readonly.
lp1 at 0x0378, using polling driver
CSLIP: code copyright 1989 Regents of the University of California
PPP: version 0.2.7 (4 channels) NEW_TTY_DRIVERS OPTIMIZE_FLAGS
TCP compression code copyright 1989 Regents of the University of California
PPP line discipline registered.
Adding Swap: 15548k swap-space
