Re: Microtek E6 - mode_sense_1 problem

Rogier Wolff (R.E.Wolff@BitWizard.nl)
Wed, 6 May 1998 13:07:49 +0200 (MET DST)

Messages sorted by: [ date ][ thread ][ subject ][ author ]
Next message: Donald Burr: "[umax] using custom gamma table dumps core"
Previous message: Paul D. Robertson: "Re: [umax] Can't use 10-bit scan: "preview cannot handle depth 16""
In reply to: Randall Hopper: "Re: Microtek E6 - mode_sense_1 problem"

Randall Hopper wrote:
>
>
> --mYCpIKhGyMATD0i+
> Content-Type: text/plain; charset=us-ascii
>
> Matto Marjanovic:
> |Short answer: The latest version of the backend (0.7, available from
> | the backend webpage "http://www.mir.com/mtek/", and sent upstream for
> | inclusion on SANE-0.73) has MODE_SENSE_1 disabled, so that problem
> | should disappear for the time being.
>
> Ok.
>
> One other piece of useful info. Here's what the NCR driver says about that
> "extraneous data discarded" message that's being printed when the
> MODE_SENSE_1 cmd is issued:
>
> ** The target wants to tranfer too much data
> ** or in the wrong direction.
> ** Remember that in extended error.
>
> (The internal error is XE_EXTRA_DATA.)
>
> |Long answer: MODE_SENSE_1 used to cause problems for folks using Solaris,
> | because apparently the Solaris scsi driver insisted that this was some
> | command reserved for tape drives. Why and how it is causing problems for
> | FreeBSD I don't know, this is a new complaint. I am curious, however,
> | to know why (if you feel like poking around even more), and I would very
> | much appreciate it if you sent me an INQUIRY dump from the Microtek backend
> | so that I can compare firmware revisions/etc between your E6 and mine.
> | Is it some scsi driver quirk, or is the scanner complaining? The big Q.
>
> Right, I'm also interested in what's really going on. It's a good excuse
> to get in there and dig around anyway. :-)

What I know about this issue is that the linux ncr8xx driver (ported
from BSD) is sensitive to the "device wants to send more data than I
expected". The other ncr7xx/8xx driver doesn't notice this.

Thus using the ncr8xx driver proved helpful in figuring out how many
bytes the scanner actually wanted to send. By carefully examining the
logs I was able to determine where things went wrong, and determining
where I had mis-read the specs.

Roger.

--
Source code, list archive, and docs: http://www.mostang.com/sane/
To unsubscribe: echo unsubscribe sane-devel | mail majordomo@mostang.com

Next message: Donald Burr: "[umax] using custom gamma table dumps core"
Previous message: Paul D. Robertson: "Re: [umax] Can't use 10-bit scan: "preview cannot handle depth 16""
In reply to: Randall Hopper: "Re: Microtek E6 - mode_sense_1 problem"