underflow field

underflow field

[Justin T. Gibbs]

I received a recent report of the aic7xxx driver reporting:

(scsi0:A:0:0): CDB: 0x12 0x1 0x80 0x0 0x60 0x0
(scsi0:A:0:0): Saw underflow (80 of 96 bytes). Treated as error

This is because the aic7xxx driver (and the aic79xx driver) honors
the underflow field.  Has this field been deprecated and thus should
be ignored by HBA drivers?  If not, why is it that scsi_scan.c doesn't
set underflow to 0 (or maybe 1 in this case) for commands that can have
variable length responses?  The above is a fetch of the serial number
from VPD page 0x80.

--
Justin

Re: underflow field

[Doug Ledford]

On Thu, Jan 30, 2003 at 04:45:48PM -0700, Justin T. Gibbs wrote:
> I received a recent report of the aic7xxx driver reporting:
> 
> (scsi0:A:0:0): CDB: 0x12 0x1 0x80 0x0 0x60 0x0
> (scsi0:A:0:0): Saw underflow (80 of 96 bytes). Treated as error
> 
> This is because the aic7xxx driver (and the aic79xx driver) honors
> the underflow field.  Has this field been deprecated and thus should
> be ignored by HBA drivers?  If not, why is it that scsi_scan.c doesn't
> set underflow to 0 (or maybe 1 in this case) for commands that can have
> variable length responses?  The above is a fetch of the serial number
> from VPD page 0x80.

Because it was an oversight that needs fixing.  When the VPD Inquiry code 
was added it should have had a correct underflow initializer and it 
doesn't look like it does.

-- 
  Doug Ledford <dledford@redhat.com>     919-754-3700 x44233
         Red Hat, Inc. 
         1801 Varsity Dr.
         Raleigh, NC 27606