Subject: Re: NCR Driver Problems
To: None <current-users@NetBSD.ORG>
From: Christoph Badura <bad@flatlin.ka.sub.org>
List: current-users
Date: 01/26/1996 02:47:00
Stefan Esser writes:
>On Jan 24,  2:10, Christoph Badura wrote:
>} Stefan Esser wrote:

>} >It is well possible, that it fully supports tags. 
>} >It is just picky about non-tagged commands (even
    ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
>} >NOP commands) sent when any tagged command is 
    ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
>} >active.
    ^^^^^^

>} I figure the problem isn't with the generic SCSI code but with your
>} driver violating the standard.

>The GENERIC SCSI code issues the start unit command,
>and the driver sends it to the drive. It includes a 
>simple tag message, if tags are in use on this drive.

Now, why then would you suggest that the reason that someone's drive
doesn't work with your driver is that your driver sends an untagged
start unit command?  I don't see how anyone could interpret your
previous message differently.  In fact, your messages says that your
driver violates the SCSI standard in sending untagged commands when
tagged commands are in the targets queue.  I, naively, assumed that
you would know what your driver is doing.

>Now there seem to be drives that don't like that 
>START STOP UNIT command being accompanied by a tag.

So retry it without a tag.

>And there seem to be drives, that do some kind of 
>soft reset, making the drive unresponsive for as 
>much as a second or two.

Gee, my drives take much longer then two seconds to spin up.  Why
don't you increase the timeout?

>Now, where is the problem in my driver ?

How about a different set of problems?  Like assuming that all drives
will spin up in less then two seconds?

Actually, from reading the code I think the driver will send tagged
commands when a contingent allegiance or extended contingent
allegiance conditions exists, which it must not do.

>The GENERIC code should remember, whether a drive has
>been started (i.e. opened), and should not send any
>more START UNIT requests. 

Oh, it does.

-- 
Christoph Badura	bad@flatlin.ka.sub.org

You don't need to quote my .signature.  Everyone has seen it by now.
Besides, it doesn't add anything to the current thread.