Mixing tagged and untagged tasks

gop at us.ibm.com gop at us.ibm.com
Mon Sep 27 13:25:41 PDT 1999


* From the T10 Reflector (t10 at t10.org), posted by:
* gop at us.ibm.com
*
--0__=EjZWXhGDI3IgLTToPnw8yU4EqxpZsLycb0oMVEGabSnfeYjJyza4pRqJ
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Content-Disposition: inline

Gerard,
The main reason we make the change was to make what everyone was already doing
legal. That activity is the practice of sending an untagged request sense
command to recover information after a check condition with outstanding tagged
commands still in the queue. (Although whether this is actually allowed or not
in SCSI-2 has been the topic of some heated debates). In any case it is now
clearly allowed.

As for how initiators should handle the case were is they send an untagged
command to a device that they also have outstanding tagged commands is up to
them to figure out. I would not recommend initiators do this.

Bye for now,
George Penokie

Dept Z9V  114-2 N212
E-Mail:    gop at us.ibm.com
Internal:  553-5208
External: 507-253-5208   FAX: 507-253-2880




Gerard Roudier <groudier at club-internet.fr> on 09/24/99 06:19:05 PM

To:   George Penokie/Rochester/IBM at IBMUS
cc:   Mark_Heath at notes.seagate.com, t10 at t10.org
Subject:  Re: Mixing tagged and untagged tasks




--0__=EjZWXhGDI3IgLTToPnw8yU4EqxpZsLycb0oMVEGabSnfeYjJyza4pRqJ
Content-Type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: Quoted-Printable
Content-Disposition: inline




On Fri, 24 Sep 1999 gop at us.ibm.com wrote:

> * From the T10 Reflector (t10 at t10.org), posted by:
> * gop at us.ibm.com
> *
> Mark,
> The change from SCSI-2 to SCSI-3 on allowing a single untagged task to be=

mixed
> in with tagged tasks was intentional.

It seem to me that this has some impact on the way initiators have to deal
with reselections when both an untagged and tagged tasks are used (if
untagged one has DISC granted), especially in the handling of reselection
errors. May be, you may want to elaborate on this topic.
This point does not seem to be covered by the standard.
It would also be interesting to know about the actual advantage of
allowing such a situation.

Thanks,
  G=E9rard.

> Bye for now,
> George Penokie
>
> Dept Z9V  114-2 N212
> E-Mail:    gop at us.ibm.com
> Internal:  553-5208
> External: 507-253-5208   FAX: 507-253-2880
>
>
>
> Mark_Heath at notes.seagate.com on 09/24/99 04:04:28 PM
>
> To:   t10 at t10.org
> cc:
> Subject:  Mixing tagged and untagged tasks
>
>
>
>
> * From the T10 Reflector (t10 at t10.org), posted by:
> * Mark_Heath at notes.seagate.com
> *
> Hello all,
>
> I recently noticed a discrepancy between the SCSI-2 standards documents a=
nd
> the SCSI-3 standards documents related to the legality of mixing tagged a=
nd
> untagged tasks.  SCSI-2 disallows the simultaneous use of tagged and
> untagged tasks by a single initiator (except when a contingent allegiance=

> condition exists), as stated in section 7.5.2 of X3T9.2/375R revision 10L=
:
>
> > An incorrect initiator connection also occurs on an initial connection
> when
> > an initiator:
> >   a) attempts to establish an I_T_L_Q nexus when an I_T_L nexus already=

> exists
> >      from a previous connection, or
> >   b) attempts to establish an I_T_L nexus when an I_T_L_Q nexus already=

> exists,
> >      unless there is a contingent allegiance or extended contingent
> allegiance
> >      condition present for the logical unit.
> >
> > A target that detects an incorrect initiator connection shall abort all=

> I/O
> > processes for the initiator on the logical unit or target routine and
> shall
> > return CHECK CONDITION status. The sense key shall be set to ABORTED
> COMMAND
> > and the additional sense code shall be set to OVERLAPPED COMMANDS
> ATTEMPTED.
>
> However, it appears that SCSI-3 allows the simultaneous use of a single
> untagged task with any number of tagged tasks, as stated in section 5.6.2=

> of SAM-2 T10/1157-D revision 12:
>
> > An overlapped command occurs when an application client reuses a Task
> Address
> > (see 4.9.3) in a new command before a previous task to which that addre=
ss
> was
> > assigned completes its task lifetime as described in 5.4. Each SCSI
> protocol
> > standard shall specify whether or not a logical unit is required to
> detect
> > overlapped commands.  A logical unit that detects an overlapped command=

> shall
> > abort all tasks for the initiator in the task set and shall return CHEC=
K
> CONDITION
> > status for that command. If the overlapped command condition was caused=

> by an
> > untagged task or a tagged task with a tag value exceeding FFh, then the=

> sense
> > key shall be set to ABORTED COMMAND and the additional sense code shall=

> be set
> > to OVERLAPPED COMMANDS ATTEMPTED. Otherwise, an additional sense code o=
f
> TAGGED
> > OVERLAPPED TASKS shall be returned with the additional sense code
> qualifier byte
> > set to the value of the duplicate tag.
>
> Since a tagged Task Address and an untagged Task Address are by definitio=
n
> different Task Addresses, mixing a single untagged task with any number o=
f
> tagged tasks doesn't result in "reuse" of a Task Address, and thus appear=
s
> to be legal.
>
> Is this an intentional deviation from SCSI-2's rules, or is this an
> oversight?
>
> Cheers,
>
> Mark A. Heath
> Seagate Technology, Inc.
>
>
> *
> * For T10 Reflector information, send a message with
> * 'info t10' (no quotes) in the message body to majordomo at t10.org
>
>
>
> *
> * For T10 Reflector information, send a message with
> * 'info t10' (no quotes) in the message body to majordomo at t10.org
>
>



--0__=EjZWXhGDI3IgLTToPnw8yU4EqxpZsLycb0oMVEGabSnfeYjJyza4pRqJ--

*
* For T10 Reflector information, send a message with
* 'info t10' (no quotes) in the message body to majordomo at t10.org






More information about the T10 mailing list