2003-03-14 22:52:45

by Philip Dodd

[permalink] [raw]
Subject: SCSI errors in logs

Hello,

Running debian testing with 2.4.20 + preempt + bttv kernel patches with
1GB of RAM - high memory enabled, I get big bunches of the following in
/var/log/messages.


Mar 14 20:41:08 gandalf kernel: scsi0: Transceiver State Has Changed to
SE mode
Mar 14 20:41:08 gandalf kernel: scsi0: Transceiver State Has Changed to
LVD mode
Mar 14 20:41:08 gandalf kernel: scsi0: Transceiver State Has Changed to
SE mode
Mar 14 20:41:27 gandalf kernel: scsi0: Transceiver State Has Changed to
LVD mode

On several occasions my logs have been filled with theses messages.
Adaptec driver is compiled in the kernel, and scsi0 is unused (scsi1 is
the other channel on my adaptec 39160, scsi2 is ide-scsi). Nothing is
physically plugged in to this channel (yet!).

I'd be grateful for any ideas - is this hardware?

Thanks,

Philip


2003-03-15 14:37:24

by John Bradford

[permalink] [raw]
Subject: Re: SCSI errors in logs

> Running debian testing with 2.4.20 + preempt + bttv kernel patches with
> 1GB of RAM - high memory enabled, I get big bunches of the following in
> /var/log/messages.
>
>
> Mar 14 20:41:08 gandalf kernel: scsi0: Transceiver State Has Changed to
> SE mode
> Mar 14 20:41:08 gandalf kernel: scsi0: Transceiver State Has Changed to
> LVD mode
> Mar 14 20:41:08 gandalf kernel: scsi0: Transceiver State Has Changed to
> SE mode
> Mar 14 20:41:27 gandalf kernel: scsi0: Transceiver State Has Changed to
> LVD mode
>
> On several occasions my logs have been filled with theses messages.
> Adaptec driver is compiled in the kernel, and scsi0 is unused (scsi1 is
> the other channel on my adaptec 39160, scsi2 is ide-scsi). Nothing is
> physically plugged in to this channel (yet!).
>
> I'd be grateful for any ideas - is this hardware?

Plug a device in to it, make sure it's properly terminated, and it
should work :-).

I suspect that the SE/LVD negotiation is confused, because there are
no devices on the bus.

John.