Weird SCSI problem with 2.4.13-ac8

Weird SCSI problem with 2.4.13-ac8

Post by Jeffrey S. Mullike » Thu, 14 Feb 2002 18:55:24



Hey folks,

        I don't know if these are the correct newsgroups in which to post this,
but here goes.  Maybe someone can at least point me in the right
direction.

        I just upgraded on of the QA machines to RH7.2, and built a 'custom'
kernel of 2.4.13, with Mr. Cox's -ac8 patch, and the bigphysarea patch.
  The 'bigphys' is needed for the development group, and the 2.4.13-ac8
just seemed to be the most recent/stable release that we could all
standardize on.  Back to my story...

        This particular QA machine (PIII 550 on a SuperMicro motherboard, I
believe) has an Adaptec AHA-2940U PCI SCSI card, and an Exabyte 8200 8mm
tape drive.

        I will attach the simple test script that we are running, but the deal
is, when I run this script on the stock 2.4.7-10 kernel that ships with
RH7.2, everything works fine.  If I run the same script with 2.4.13-ac8,
then I get the errors that I've pasted below. I've repeated this several
times.  I spent some time in the support section of the Exabyte website,
and saw some note about timing issues with certain levels of firmware,
with respect to error recovery.  That gave me an idea.
        I edited the script below, inserting a 'sleep 3' line in all the places
that I've marked with a -->> in the script below, and the script runs
without errors on the 2.4.13-ac8 kernel, with the sleeps in between
consecutive tape I/O operations.

        Anyone got any ideas?  Need more input?

        Thanks, in advance.  ( I'll post the errors first, then the script)

                Jeff.

ERRORS:
------------------------------------------------------------

kernel: st0: Block limits 1 - 245760 bytes.
kernel: st0: Error 8 (sugg. bt 0x0, driver bt 0x0, host bt 0x0).
kernel: st0: Error with sense data: Info fld=0x1f (nonstd), Current
st09:00: sense key Illegal Request

SCRIPT:
-------------------------------------------------------------

#!/bin/sh
clear
echo

if [ $# -lt 1 ]; then
  echo "Tape Device Name Required - Example: ./ddtest /dev/nst0"
  exit 1
fi

tape=$1

echo Save Current Directory...
curdir=`pwd`

echo Get Tape Status...
mt -f $tape status

-->>

echo Rewind Tape...
mt -f $tape rewind

echo Write Test Data to Tape

-->>

dd bs=1024 if=/private/qa/qadddata.tgz of=$tape

-->>

echo Get Tape Status...
mt -f $tape status

-->>

echo Rewind Tape...
mt -f $tape rewind

-->>

echo Remove Temporary Test Data - /private/qa/tmp/dddata.tgz
rm -rf /private/qa/tmp/dddata.tgz

echo Read Test Data from Tape into Temporary Directory

dd bs=1024 if=$tape of=/private/qa/tmp/dddata.tgz

echo Compare Orginal Test Data with Test Data Read from Tape...
diff /private/qa/qadddata.tgz /private/qa/tmp/dddata.tgz

echo Change to Initial Directory...
cd $curdir

echo End DD Test
exit 0

 
 
 

1. 2.4.13-ac8 weird hangs

Hi!

I recently upgraded from 2.4.12-acsomething to 2.4.13-ac8 and now I keep
getting weird hangs with that kernel. It just randomly completely hangs and
then the CAPS LOCK and SCROLL LOCK LEDs light up.

Does anyone know if this is a known problem? Or maybe some ideas as to what
might cause it and how to fix it?

Thanks a lot!

Marten Klencke

2. Deleting a partition On Windows Me...

3. Tulip driver problem with 2.4.13-ac8

4. rp-pppoed.so DOD

5. PROBLEM: system hangs on dual 1GHz PIII system with 2.4.13-ac8

6. Heathkit Terminal

7. 2.4.13-ac8: oops in update_one_process

8. Diskimage NTFS partition

9. oops with 2.4.13-ac8

10. 2.4.13-ac8: crash on IBM Thinkpad 600x

11. Weird /proc/meminfo output on 2.4.13-ac5

12. Problems with 2.4.13, kt133a, and IBM DLTA drives

13. Google's mm problem - not reproduced on 2.4.13