Re: SCSI changes Interrupts? I Confirm Problems!

[Jameson Burt]

I confirm possible problems with scsi.
Here's my story

For four months, I noticed no problems with my tape drive or my disk drive.
In the last two weeks, I have updated a few packages in hamm, but not many.
I now have two indicators that just one partition (out of 8+ partitions) on my 
disk drive gets its signals diverted to the tape drive.

1. mke2fs -c /dev/sdc12
   rather than working on my disk drive,  results in a churning of the tape 
/dev/st0 for 30 minutes until I kill the process.
   Then I have difficulty typing a character as the load meter shows 17 
horizontal lines.
2. When making backups, only the partition /dev/sdc12 fails.
   If I duplicate its contents to another partition, dump succeeds.  Again, 
dumping to this /dev/sdc12 partition churns the tape drive for at least 30 
minutes until I kill the dump command.

The above disk drive is a wide 9GB Fujitsu.  The above tape drive is an old 
Exabyte 8200 8mm.  I have two other narrow 2GB Seagate drives externally 
attached.  All these scsi devices are attached to an Adaptec 2940UW adapter.

Because of this mixing of scsi signals between one disk drive partition and 
the tape drive, I no longer use the partition /dev/sdc12. The setup of my 
Adaptec 2940UW refuses any ultra scsi protocol, preventing an ultra problem. 
  This cross-signal problem seems inexplicable.  Has anyone any 
explanations/solutions?



--
TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to
[EMAIL PROTECTED] . 
Trouble?  e-mail to [EMAIL PROTECTED] .

SCSI changes Interrupts?

[Timothy M. Hospedales]

I just got my Adaptec 152x SCSI adapter to work However, two weird
things happened in the process that I wonder if anyone could shed some
light on

Firstly, before I got it to work, cat /proc/interrupts reported my PS2
mouse as being on IRQ 8,  SoundScape as being on IRQ 10, and had no
listing for sound. After I got it to work by setting the jumpers on
the  SCSI adapter to IRQ 9, and giving the kernel the appropriate
parameters at boot time, cat /proc/interrupts reports the PS2 mouse on
IRQ 12, sound on IRQ 5, and does not list  SoundScape any more.
(The SoundScape support in the kernel is configured to use IRQ 10 for
MIDI, and IRQ 5 for Audio... (DSP?)).
Any ideas on the discrepancy?

Also, before it started working,  I tried setting the IRQ jumpers on 11.
(Not listed as in use by /proc/interrupts at the time), however, at boot
time, it failed to initalize that IRQ, and suggested that it might be
wrong..  So, after that I tried IRQ 9, (also not listed as in use),
which worked.
Any ideas on why it wouldn't work on an interrupt not listed as in use?
I suspect my /proc/interrupts is abit crazy, because even now that it
doesn't show the SoundScape on IRQ 10 any more, it still initalizes fine
on that IRQ at boot time, and I can still play mp3s fine. :).

Thanks in advance for any thoughts on the matter,
Timothy.

PS: I use aha152x=0x140,11,7,1 at boot time now. Can anyone fill me in
on what the '7' and '1' are about?  The SCSI howto used them, but I
couldn't find anything on their significance.


--
TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to
[EMAIL PROTECTED] . 
Trouble?  e-mail to [EMAIL PROTECTED] .