Minutes of X3T10 ATA -3 Working Group #1 October 19, 1994
Gene_Milligan at notes.seagate.com
Wed Oct 26 21:58:40 PDT 1994
The following minutes were sent to the ATA reflector several days ago and have
not emerged from it as of this moment. I don't know if there was a problem with
our gateway or an unwanted blockage in the ATA reflector. Since the problem is
not known, this repeat transmission is being sent both to the ATA and SCSI
To: ata @ dt.wdc.com @ internet
From: Gene Milligan
Date: 10/24/94 05:16:03 PM
Subject: Minutes of X3T10 ATA -3 Working Group #1 October 19, 1994
Accredited Standards Committee**Operating under the procedures of The American
National Standards Institute.
X3 Secretariat, Computer and Business Equipment Manufacturers Association
1250 Eye Street NW, Suite 200, Washington, DC 20005-3922
Telephone: 202-737-8888 (Press 1 twice) FAX: 202-638-4922 or 202-628-2829
X3, Information Processing Systems
Doc. No. X3T10/94-ataR0
Date: October 19, 1994
Reply to: G. Milligan
To: Membership of X3T10
Subject: Minutes of X3T10 ATA -3 Working Group #1 October 19, 1994
1. Openning Remarks
2. Attendance and Membership, Introductions
3. Approval of Agenda
4. Document Distribution
5. Old Business
5.1 ATA-3 Project Plan
5.2 ATA Suspended Issues List
5.3 ATA Command Queuing Proposals
5.4 ATA Multi-threading
5.5 Data prefetching across the ATA interface
5.6 Device 1 only on ATA
5.7 20 MB/s and greater ATA transfer rates
6. New Business
7. Call for Patents
8. Liaison Reports
9. Action Items
10. Meeting Schedule
Gene Milligan convened the meeting at 9:30 am. He thanked Jim McGrath of
Quantum for hosting the meeting. He also requested that Larry Lamers take the
minutes and thanked him for accepting the request.
As is customary, the people attending introduced themselves. A copy of the
attendance list was circulated for attendance and corrections.
It was stated that the meeting had been authorized by X3T10 and would be
conducted under the X3 rules. Ad hoc meetings take no final actions, but
prepare recommendations for approval by the X3T10 task group. The voting rules
for the meeting are those of the parent committee, X3T10. For the ad hoc,
other than straw votes, the voing rules are: one vote per participating
The minutes of this meeting will be posted to the X3T10 BBS and the ATA
Reflector and will be included in the next X3T10 committee mailing.
Attendance and Membership, Introductions
Attendance at working group meetings does not count toward minimum attendance
requirements for X3T10 membership. Working group meetings are open to any
person or company to attend and to express their opinion on the subjects being
The following people attended the meeting.
Name _Company Email Address
Hale Landis Consultant landis at sugs.tware.com
John Packer Adaptec packer at eng.adaptec.com
Lawrence J. Lamers Adaptec ljlamers at aol.com
Richard Kalish Adaptec rkalish at corp.adaptec.com
Kevin James AMD kevin.james at amd.com
Nicos Syrimis Cirrus Logic
Mark Gurkowski Conner mark.gurkowski at conner.com
Bill Willette Dell Computer bill_willette at us1.dell1.com
Dal Allan ENDL
Jeff Epstein Future Domain jeffe at fdc.mhs.compuserve.com
Dan Colegrove IBM colegrove at vnet.ibm.com
Pete McLean Maxtor pete_mclean at maxtor.com
Ron Roberts Maxtor rkroberts at aol.com
Tom Newman Mission Peak Designs 71246.1573 at compuserve.com
Robbie Shergill National Semiconductor rss at berlioz.nsc.com
Peter Brown Oak Technology peterb at corp.oaktech.com
Wang Mai Oak Technology
Chi-Chen Wu Promise Technology chiwu at promise.com
Ron Lin Promise Technology ronlin at promise.com
Farbod Falakfarsa Quantum
James McGrath Quantum jmcgrath at qntm.com
John Brooks Quantum jbrooks at asic.qntm.com
Gene Milligan Seagate gene_milligan at notes.seagate.com
John Masiewicz Seagate Technology masiewicz at notes.seagate.com
Stephen Finch Silicon Systems 5723283 at mcimail.com
Devon Worrell Western Digital worrell at dt.wdc.com
Tom Hanan Western Digital t_hanan at dt.wdc.com
Approval of Agenda
Item 7 was changed to ATA-3 Command Goals & Requirements.
The agenda was approved as amended.
Conner ATA Tagged Command Queuing 10/18/94
Quantum ATA Command Queuing 10/13/94 James McGrath
Western Digital Enhanced IDE Design Guidelines - Revision 1.0 - 10/18/94 - Hanan
Suggested Multi-thread & Queing Process - Hanan
ATA-3 Project Plan
Gene Milligan noted that ATA-3 is supposed to be delivered within 14 months.
He requested an analysis of the work items be done so as to insure that we
will meet that goal.
Tom Hanan presented a list of goals for ATA-3.
plug and play compatibility with existing systems
highest performance at cost point for single disk work station
ATAPI CD-ROM & Tape
PCI Bus Master DMA
Command/device vs command/channel
Commands/device vs. command/device
Jim McGrath asked if ATA-3 could address issues above the interface, ala CAM
for SCSI. Gene Milligan responded that there was ample precedence for such work
as exemplified by the SCSI CAM projects and the FDDI Network layer
(subsequently dropped due the availability of existing solutions for FDDI).
However if the committee decides such work is desirable an additional project
would be required and consideration for IEEE projects in related areas should
The term multi-thread was changed to overlapped commands. The multi-thread
term generally refers to the software driver feature not a peripherial.
Command queuing is yet another feature of a device.
There was considerable discussion over whether overlapped commands and command
queuing are two separate issues or part of the same issue.
The ATA-3 Study Group agreed to recommend that key elements of the project
plan should be:
Proposal cut off January 1995 ( later items slated for ATA-4)
Preliminary description of proposed items by 11/10/94
Concensus on approach and acceptance of a working document by March 1995.
Solidify the dpANS by July 1995.
Final edit by September 1995.
Initiate Letter Ballot for forwarding by November 1995.
ATA Suspended Issues List
The following list was supplied by Steve Finch:
a) Data reliability - CRC and Parity proposals - Adpatec, Cirrus, . All
interested parties should update their proposals. (X3T10/94-027r2)
b) Major clean-up/re-write of the ATA document. - Deleted in view of the ATA-2
c) Inclusion of ATAPI and CD-ROM support. - To be considered as items are
developed by the ATAPI project activity with this item being principaly liaison.
d) 20 MB/s transfer rates - tied in with termination issues. It was agreed
that higer transfer rates should be DMA and not PIO.
Hale Landis suggested that the addition of a DMA mode IDENTIFY DRIVE command
could eliminate the need for PIO mode commands for systems that wanted to
implement just one set of command types. He urged that optimizing on DMA only
hardware would simplify the drive and adapter silicon
e) Reset Pulse Width (X3T10/94-053r2) - need to resolve over the internet.
f) Secure Mode (for removable devices) (X3T10/94-087r0) - comments are
g) Format Issues. Resovled in ATA-2.
h) Termintation issues - tied in with item d).
i) Check Power Mode additional capability (X3T10/94-154r0) - comments if any
j) Multi-threading. (Command overlap and command queuing) - see further agenda
All prior proposals for ATA-3 that the proposer wishes to champion for
inclusion in this project are requested to re-posted on the ATA reflector to
insure that all participants have an opportunity to review nad comment on the
The study group recommended that all proposals for consideration ahead of ATA-4
need to be submitted by January 12, 1995.
The ATA-2 document will serve as the basis for the ATA-3 document.
ATA Command Queuing Proposals
See item 5.4.
Tom Hanan discussed the slides noted under document distribution. Is
compatibility with with existing adapters required for overlapped commands?
... for queuing? Most of the group felt that a good solid performance solution
should be the objective and if there are compatibility issues these are of
secondary importance for queueing. Not requiring host intervention in DMA data
transfers is a goal.
interrupts during data transfers
master/slave vs SCSI architecture
register for command tag
overlapped vs queuing support
List of topics to discuss:
arbitrated TF (DRQ, IRQ)
interrupts during DMA data transfer
register for command queueing
master/slave vs peer architecture
identify host hardware changes required
electrical (buffer, cable, drivers)
error reporting & management
degree of CPU involvment
effect of resets & aborts
sharing interrupt line
command by comand enabling of queing and arbitration
allow only simple que tags
Does the group want to proceed with an ATA command queueing and overlapped
commands? 9 in favor, 6 opposed. Larry Lamers noted his negative was due to
the fact that the ATA-3 project proposal does not specifically specify that
ATA-3 will include command queuing.
Exception management in the queued environment. Jim McGrath expounded his
view. Only simple queue tags should be allowed. Sending a command with the
same tag value aborts both commands. Error reporting is the same as in ATA-2.
Soft resets do not abort the queue. An error does not stop processing of the
Does the group favor using only simple queue tags? 13-0
Does the group favor having the queue cleared on error or reset? 13-0
Does the group favor having no provision to abort commands? 14-2
Does the group favor having a hardware reset put the device in non-queue mode?
(note - a software reset does not)
Does the group favor fixing the sleep mode? 13-0
Does the group favor eliminating the sleep mode? 1-12
Queue tag register location: There are four ways to do this.
a) redefine task file (e.g., error register)
issues-problems with a command that completes in error and in multiple driver
b) use non-disk register (e.g., 3f3)
issues-may be in use by other folks
c) use reserved hard disk bits
issues-historical uses make this difficult to predict their use
d) use word access method (e.g., use upper half of 3f6)
issues-how does the drive know it is in word mode? also 16-bit access in
PCMCIA may be a problem.
Without opposition items b) and c) were dropped from consideration.
Does the group favor redefining a task file for the queue tag (similar to
Conner proposal)? 4-8-8
Does the group favor using word access for the queue tag? 9-4-8
Does the group favor not making a decision at this time? 13-2-3
Enabling queuing on a command by command basis. Should the mixed driver
environment be addressed? If so, how should we do this?
Should all drivers know ATA-3? 8-4
Should a mix of drivers for ATA-2 and ATA-3 be allowed? 6-8
Does the group favor requiring ATA-3 accept queued and non-queued commands?
Does the group favor arbitration?
Are the devices autonomous relying on the host to resolve contention or is
there a mechanism for the devices to resolve bus usage (i.e., arbitration)?
John Masciewicz suggested that each company that wants to make a proposal for
queuing develop one and bring it back to the next meeting and be allowed
one-hour to present that proposal and its merits. There was general agreement
to follow this path.
Steve suggested that the slave drive use PDIAG for interrupt. This works if a
bridge chip handles the interrupts. Can a bridge chip handle a two signal
interrupt? Defining this as a functional requirement has compatibilitiy issues.
Agree that if it is host based that it must be simple enough for a bridge to do
Data prefetching across the ATA interface
Prefetching can be a problem with ATAPI devices because they are not a fixed
length transfers. Handle in ATAPI.
Device 1 only on ATA
Desired by ATAPI to allow only one device as device 1 so that timeout need not
be waited for and for applications where it is desired to keep the ATAPI and
Disk Device on a separate bus. Should the registers be shadowed for the
non-existent devices? This will be looked at in the ATA-2 working group as
part of Hale Landiss comments.
20 MB/s and greater ATA transfer rates
DMA only was agreed to.
Call for Patents
Gene Milligan requested that anyone aware of andy patents required fort the
proposals be disclosed early in accordance with the ANSI patent policy.
The Secure Mode proposal involves patents pointed out by Pete McLean and he
stated that a letter has been submitted by Maxtor. He also mentioned an IBM
patent and Dan Colgrove agreed to pursue that possibility.
Review of the document is continueing. A two day working group meeting is set
for Thursday and Friday. The target is to attempt a forwarding at the January
Tom Hanan reported that an ATAPI meeting is planned for 11/29/94 to deal with
command queuing and SFF-8020 issues. Tom plans to not submit ATAPI to X3T10
until the project proposal is approved.
1) Gene Milligan to investigate the appropriateness of command queuing in the
ATA-3 project proposal.
2) Larry lamers to get document numbers for later inclusion in next minutes.
3) All participants are requested to post existing proposals for ATA-3 on the
4) All partcipants to submit ATA-3 proposals for consideration in preliminary
concept form by 11/10 and final form by January 12, 1995.
November 10, 1994 at Palm Springs, CA. Host: Western Digital.
November 30, 1994 at Irvine, CA. Host: Western Digital.
The meeting adjourned at 5:00 PM on October 19, 1994.
Gene Milligan -- Gene_Milligan at notes.seagate.com
Seagate Technology - 920 Disc Drive - Scotts Valley, CA 95066 USA
Main Phone 408-438-6550 - Email Problems postmaster at notes.seagate.com
Technical Support: BBS 408-438-8771 Fax 408-438-8137 Voice 408-438-8222
### OGATE Version 8 message trace and attachment information:
### MsgFileName: m:\mgate\outbound\1078.MSG
### Org Date: 10-26-94 09:37:59 PM
### From: Gene Milligan at SEAGATE
### To: scsi @ wichitaks.ncr.com @ internet
### Subject: Minutes of X3T10 ATA -3 Working Group #1 October 19, 1994
### Attachments: none
More information about the T10