Intel SL6NQ - Xeon 2.4 GHz/533MHz/512 KB CPU Processor 2.4GHz Specification

Intel

®

Xeon

®

Processor

Specification Update

January 2007

Notice: The Intel

®

Xeon

® processor may contain design defects or errors known as errata which may cause the product to deviate from published specifications. Current characterized errata are documented in this specification update.

Document Number: 249678-057

INFORMATION IN THIS DOCUMENT IS PROVIDED IN CONNECTION WITH INTEL

®

PRODUCTS. EXCEPT AS PROVIDED IN INTEL'S TERMS

AND CONDITIONS OF SALE FOR SUCH PRODUCTS, INTEL ASSUMES NO LIABILITY WHATSOEVER, AND INTEL DISCLAIMS ANY EXPRESS

OR IMPLIED WARRANTY RELATING TO SALE AND/OR USE OF INTEL PRODUCTS, INCLUDING LIABILITY OR WARRANTIES RELATING TO

FITNESS FOR A PARTICULAR PURPOSE, MERCHANTABILITY, OR INFRINGEMENT OF ANY PATENT, COPYRIGHT, OR OTHER

INTELLECTUAL PROPERTY RIGHT.

Intel products are not intended for use in medical, life saving, life sustaining, critical control or safety systems, or in nuclear facility applications.

Intel may make changes to specifications and product descriptions at any time, without notice.

Designers must not rely on the absence or characteristics of any features or instructions marked “reserved” or “undefined.” Intel reserves these for future definition and shall have no responsibility whatsoever for conflicts or incompatibilities arising from future changes to them.

Contact your local Intel sales office or your distributor to obtain the latest specifications and before placing your product order.

Copies of documents which have an ordering number and are referenced in this document, or other Intel literature may be obtained by calling

1-800-548-4725 or by visiting Intel's website at http://developer.intel.com/design/litcentr.

Intel®, the Intel® logo, Pentium®, Pentium® III Xeon™, Celeron, Intel® NetBurst™ and Intel® Xeon™ are trademarks or registered trademarks of

Intel® Corporation or its subsidiaries in the United States and other countries.

Copyright © 2001-2005, Intel Corporation. All rights reserved.

*Other names and brands may be claimed as the property of others.

2

Intel

®

Xeon

®

Processor Specification Update

Contents

Revision History ................................................................................................................. 5

Preface ............................................................................................................................... 9

Identification Information .................................................................................................. 10

Mixed Steppings in DP Systems ...................................................................................... 18

Summary Table of Changes............................................................................................. 20

Errata................................................................................................................................ 27

Specification Changes...................................................................................................... 51

Specification Clarifications ............................................................................................... 53

Documentation Changes.................................................................................................. 56

Intel

®

Xeon

®


3

4

Intel

®

Xeon

®


Revision History


Version

-001

-002

-003

-004

-005

-006

-007

-008

-009

-010

-011

-012

-013

-014

Description

• Initial release.

• Added errata P27-P28.

• Updated erratum P25.

• Added RCPPS, RCPSS, RSQRTPS and RSQRTSS instruction specification clarification.


• Added Unused outputs specification clarifications.

• Added errata P33-34.

• Production mark update to include 2D matrix.

• Added 2 GHz frequency, D0 stepping and FC-BGA package information.

• Added DP Platform Population Matrix.

• Updated Summary of Errata table with applicable “Fixed” & “No Fix” errata plans.

• Updated errata P32.

• Added Erratum P35.

• Updated Summary of Errata Table with latest errata status.

• Added Documentation Changes P1-P5.

• Added errata P36 and P37.

• Added Documentation Change P6.

• Updated Summary Tables with latest errata and documentation changes.



• Added new processor with Processor Signature=0F24h B0 Step.

• Added new S-spec processors to Processor ID and DP Matrix tables.

• Added new package markings (Figs 3, 4) for 512KB Cache processor.

• Corrected 3 mislabeled S-Spec Table parts.

• Added erratum P45 to Processor ID Table.



• Minor changes to “Fix” descriptions.

• Added PWRGOOD Specification Change.

• Added errata P48.

• Updated errata P11.

• Added Specification Changes P1-P3.


• Added new S-specs SL687 and SL65T to Processor ID table.

• Updated datasheet name to 2.40 GHz.

Date

May 2001

June 2001

June 2001

July 2001

August 2001

September 2001

October 2001

November 2001

December 2001

January 2002

January 2002

February 2002

April 2002

May 2002

Intel

®

Xeon

®


5


Version

-015

-016

-017

-018

-019

-020

-021

-022

-023

-024

-025

-026

-027

Description Date

• Added erratum P49.

• Updated errata P24, P40 status.

• Added Document Changes P1-P2.

• Added new erratum P50.

• Added new Documentation Changes P3-P12.

• Edited Summary of Errata Table erratum P40 to Plan Fix.

• Minor edits to processor markings.

• Edited DP Matrix table.

• Updated the Summary of Errata table w C1 Step info.

• Added errata P51, P52. Edited erratum P14.


• Added Specification Clarification P1.

• Added new C1 S-specs to Processor ID Info table.

• Edited erratum P35.

• Removed erratum previously numbered P52 as not applicable. Added new erratum P52.



• Added new Specification Change P1.

• Added new Specification Clarification P2.

• Deleted old Documentation Changes.

• Deleted old Specification Changes

• Added new Processor (Intel

®

Xeon

®

Processor with 533 MHz Front

Side Bus.

• Added new processor with Processor Signature=0F27h C1 Step.

• Added New S-specs to the processor ID table.

• All References to CPUID are now renamed Processor Signature.

• Added reference to IA-32 Intel

®

Architecture Software Developer’s

Manual.

• Added erratum 054-055.

• Added new errata P56, P57.

• Added boxed processors S-Specs for 400 MHz FSB with the frequencies: 1.80, 2, 2.20 and 2.40 GHz.

• Added new errata P58 and P59.

• Updated previous errata fixing status, P55.

• Updated the name of the datasheet reference.

• Added S-Spec #SL6RR, SL6GH.

• Updated previous errata P55 and P58.

• Added Specification Clarification P1-P4.

• Added Low Voltage Intel

®

Xeon

®

Processor Specifications.


• Updated Sections with Low Voltage Intel

®

Xeon

®

Processor

Datasheets.


June 2002

July 2002

August 2002

September 2002

October 2002

November 2002

December 2002

January 2003

February 2003

March 2003

April 2003

May 2003

June 2003

6

Intel

®

Xeon

®



-034

-035

-036

-039

-040

-041

-042

-043

-044

-045

-046

-047

-048

-049

-050

Version

-028

-029

-030

-031

-032

-033

-037

-038

Description

• Added new Processor Intel

®

Xeon

®

Processor with 1-MB L3 Cache with Processor Signature=0F25H (M0 Stepping).

• Added New S-specs to the processor ID table for 0F25H (M0 Stepping) and 0F29h (D1 Stepping).

• Added new processor with Processor Signature=0F29H (D1 Stepping).

• Updated DP Platform Population Matrix for the Intel

®

Xeon

®

Processor to include 0F25H and 0F29H.

• Removed Specification Clarification P3.

• Updated erratum P30 – Title Correction.

Date

July 2003

August 2003

September 2003

• Updated errata P54 and P56.


• Updated erratum P9, P22, P39.

• Added new Processor - Intel

®

Xeon

®

Processor with 1-MB L3 Cache at 3.20 GHz with Processor Signature=0F25H (M0 Stepping).

• Added New S-specs - Intel

®

Xeon

®

Processor with 1-MB L3 Cache at

3.20 GHz.


• Removed Specification Clarification P4.

• Added S-spec SL74T.




• Added new Processor Intel

®

Xeon

®

Processor with 2-MB L3 Cache with Processor Signature=0F25H (M0 Stepping).

October 2003

October 2003

November 2003

December 2003

February 2004


• Added Specification Change P1.




• Added S-spec numbers SL7D5, SL7DG, SL7D4, and SL7DF to

Processor ID table.

March 2004

April 2004

May 2004

June 2004

July 2004

• Added errata P70, P71, P72.

• Updated affected documents listed under Preface, Specification

Changes, Specification Clarifications, and Documentations Changes.



• Updated erratum 72 in the Errata Summary Table.







• Updated related documents list

• Added Erattum P82.

• Updated codes used in summary table.

August 2004

August 2004

September 2004

September 2004

October 2004

November 2004

December 2004

April 2005

July 2005

October 2005

Intel

®

Xeon

®


7


Version

-051

-052

-053

-054

-055

-056

-057

Description Date

• Updated erratum P53 and added erratum P83.

January 2006

• Updated links to Software Developers Manuals. Added s-spec SL8TJ.

March 2006

• Added S-specs SL8TK, SL8TL, SL8SE and SL8TH.

• Updated Summary Table of Changes.

• Updated the Software Developer Manual Name.

April 2006

October 2006

• Made changes to the DP Platform Population Matrix.

November 2006

December 2006

• Updated Summary Table of Changes.

• Updated Summary Table of Changes. January 2007

8

Intel

®

Xeon

®


Preface

Preface

This document is an update to the specifications contained in the documents listed in the following

Affected Documents/Related Documents table. It is a compilation of device and document errata and specification clarifications and changes, and is intended for hardware system manufacturers and for software developers of applications, operating system, and tools.

Information types defined in the Nomenclature section of this document are consolidated into this update document and are no longer published in other documents. This document may also contain information that has not been previously published.

Affected/Related Documents

Document Title Document Number

Intel

®

Xeon

®

Processor at 1.40 GHz, 1.50 GHz, 1.70 and 2 GHz Datasheet

Intel

®

Xeon

®

Processor with 512 KB L2 Cache at 1.80 GHz to 3.0 GHz Datasheet

Intel

®

Xeon

®

Processor with 533 MHz Front Side Bus at 2 GHz to 3.20 GHz Datasheet

Low Voltage Intel

®

Xeon

®

Processor at 1.60 GHz, 2.0 GHz, and 2.4 GHz Datasheet

Intel

®

64 and IA-32 Intel

®

Architectures Software Developer’s Manual, Volumes 1, 2A,

2B, 3A, and 3B

249665

298642

252135

273766

253665, 253666,

253667, 253668 and

253669, respectively

Nomenclature

S-Spec Number is a five-digit code used to identify products. Products are differentiated by their unique characteristics, e.g., core speed, L2 cache size, package type, etc., as described in the processor identification information table. Care should be taken to read all notes associated with each S-Spec number.

Errata are design defects or errors. Errata may cause the Intel

®

Xeon

®

processor’s behavior to deviate from published specifications. Hardware and software designed to be used with any given stepping must assume that all errata documented for that stepping are present on all devices.

Specification Changes are modifications to the current published specifications. These changes will be incorporated in the next release of the specifications.

Specification Clarifications describe a specification in greater detail or further highlight a specification’s impact to a complex design situation. These clarifications will be incorporated in the next release of the specifications.

Documentation Changes include typos, errors, or omissions from the current published specifications. These changes will be incorporated in the next release of the specifications.

Intel

®

Xeon

®


9

Identification Information


Intel

®

Xeon

®

Processor Markings, 256-KB Cache (603-pin

Interposer INT-mPGA Package)

Figure 1. Top Side Processor Marking

Intel® Xeon™ i(m) ©’01

2D Matrix

OR

Dynamic Laser

Mark Area

D0096109

0032

ATPO Mark

(8 Characters)

Serial Number

Mark

(4 digits)

Figure 2. Bottom Side Processor Marking

S-Spec

Country of

Assy

Dynamic Laser

Mark Area

Speed / Cache / Bus / Voltage

1700DP/256/400/1.7V

SL56N COSTA RICA

C0096109 - 0021

FPO – Serial #

(13 Characters)

Intel

®

Xeon

®


10


Intel

®

Xeon

®

Processor, 512-KB Cache, 400 and 533 MHz

FSB Markings, (603-pin Interposer INT-mPGA Package and

604-pin Fc-mPGA2 Package)

Figure 3. Top Side Processor Marking

Intel® Xeon™ i m c ‘02

2D Matrix Includes ATPO and

Serial Number (front end mark)

Figure 4. Bottom Side Processor Marking

S-Spec

Country of

Assy

Dynamic Laser

Mark Area

Speed / Cache / Bus / Voltage

1800DP/512/400/1.5V

SL5Z8 COSTA RICA

C0096109 - 0021

FPO – Serial #

(13 Characters)

Intel

®

Xeon

®


11


Figure 5. Example of Production Mark – Top View

Intel® Xeon™ i m c ‘01

Top View

Brand

Copyright information

2D Matrix Includes

ATPO and Serial

Number (front end mark)

Pin A1 Triangle

12

Figure 6. Example of Production Mark – Bottom View

Bottom View

1700DP/256L2/400/1.7V

SL56N COSTA RICA

C0096109-0021

Product Code with feature information:

Speed/cache/bus/ voltage

S-Spec and Assembly Site

FPO and Serial Number

(End of Line Mark)

Pin A1 Triangle

The Intel Xeon processor can be identified by the following values:

Family

1

1111

1111

1111

Model

0000

0001

0010

2

Brand ID

3

00001110

00001110

00001011

1. The Family corresponds to bits [11:8] of the EDX register after RESET, bits [11:8] of the EAX register after the Processor

Signature instruction is executed with a 1 in the EAX register, and the generation field of the Device ID register accessible through Boundary Scan.

2. The Model corresponds to bits [7:4] of the EDX register after RESET, bits [7:4] of the EAX register after the Processor Signature

2 instruction is executed with a 1 in the EAX register, and the model field of the Device ID register accessible through Boundary

Scan.

3. The Brand ID corresponds to bits [7:0] of the EBX register after the Processor Signature instruction is executed with a 1 in the

EAX register.

Intel

®

Xeon

®



Cache and TLB descriptor parameters are provided in the EAX, EBX, ECX and EDX registers after the Processor Signature instruction is executed with a 2 in the EAX register. Please refer to the

Intel Processor Identification and the Processor Signature Instruction Application Note (AP-485) for further information on the Processor Signature instruction.

Table 1. Intel

®

Xeon

®

Processor Identification and Package Information (Sheet 1 of 5)

S-Spec

Number

SL4WX

Core

Stepping

C1

Processor

Signature

0F0Ah

Speed

Core/Front

Side Bus

(GHz/MHz)

1.40/400

L2 Size

(Kbytes)

256-KB

L3 Size

(Kbytes)

Processor

Interposer

Revision

B0

Package and

Revision

SL56G

SL4WY

SL4ZT

SL56N

SL56H

SL5TD

SL5U6

SL5TE

SL5U7

SL5TH

SL5U8

SL5Z8

SL622

SL5Z9

SL623

C1

C1

C1

C1

C1

D0

D0

D0

D0

D0

D0

B0

B0

B0

B0

0F0Ah

0F0Ah

0F0Ah

0F0Ah

0F0Ah

0F12h

0F12h

0F12h

0F12h

0F12h

0F12h

0F24h

0F24h

0F24h

0F24h

1.40/400

1.50/400

1.50/400

1.70/400

1.70/400

1.50/400

1.50/400

1.70/400

1.70/400

2/400

2/400

1.80/400

1.80/400

2/400

2/400

256-KB

256-KB

256-KB

256-KB

256-KB

256-KB

256-KB

256-KB

256-KB

256-KB

256-KB

512-KB

512-KB

512-KB

512-KB

B0

B0

B0

B0

B0

C0

C0

C0

C0

C0

C0

01

01

01

01

603-pin micro-PGA interposer with 31 mm

OLGA rev 2.0


OLGA rev 2.0


OLGA rev 2.0


OLGA rev 2.0


OLGA rev 2.0


OLGA rev 2.0


FC-BGA package


FC-BGA package


FC-BGA package.0


FC-BGA package


FC-BGA package 0


FC-BGA package


FC-BGA package


FC-BGA package


FC-BGA package


FC-BGA package

Notes

1, 4

1, 2, 4

1, 4

1, 2, 4

1, 4

1, 2, 4

1, 4

1, 2, 3,

4

1,4

1, 2, 3,

4

1, 3, 4

1, 2, 3,

4

1

1, 2

1

1, 2

Intel

®

Xeon

®


13


Table 1. Intel

®

Xeon

®


S-Spec

Number

SL5ZA

Core

Stepping

B0

Processor

Signature

0F24h

Speed

Core/Front

Side Bus

(GHz/MHz)

2.20/400

L2 Size

(Kbytes)

512-KB

L3 Size

(Kbytes)

Processor

Interposer

Revision

01

Package and

Revision

SL624

SL687

SL65T

SL6EL

SL6JX

SL6EM

SL6JY

SL6EN

SL6JZ

SL6EP

SL6K2

SL6EQ

SL6K3

SL6M7

SL6MS

SL6NP

SL6NQ

B0

B0

B0

C1

C1

C1

C1

C1

C1

C1

C1

C1

C1

C1

C1

C1

C1

0F24h

0F24h

0F24h

0F27H

0F27H

0F27H

0F27H

0F27H

0F27H

0F27H

0F27H

0F27H

0F27H

0F27H

0F27H

0F27H

0F27H

2.20/400

2.40/400

2.40/400

1.80/400

1.80/400

2/400

2/400

2.20/400

2.20/400

2.40/400

2.40/400

2.60/400

2.60/400

2.80/400

2.80/400

2/533

2.40/533

512-KB

512-KB

512-KB

512-KB

512-KB

512-KB

512-KB

512-KB

512-KB

512-KB

512-KB

512-KB

512-KB

512-KB

512-KB

512-KB

512-KB

01

01

01

01

01

01

01

01

01

01

01

01

01

01

01

01

01


FC-BGA package


FC-BGA package


FC-BGA package


FC-BGA package


FC-BGA package


FC-BGA package


FC-BGA package


FC-BGA package


FC-BGA package


FC-BGA package


FC-BGA package


FC-BGA package


FC-BGA package


FC-BGA package


FC-BGA package


FC-BGA package

604-pin micro-PGA interposer with 42.5 mm

FC-PGA2 package


FC-PGA2 package

Notes

1

1, 2

1,2

2

2

2

2

2

2, 5

2, 5

14

Intel

®

Xeon

®



Table 1. Intel

®

Xeon

®


S-Spec

Number

SL6NR

Core

Stepping

C1

Processor

Signature

0F27H

Speed

Core/Front

Side Bus

(GHz/MHz)

2.66/533

L2 Size

(Kbytes)

512-KB

L3 Size

(Kbytes)

Processor

Interposer

Revision

01

Package and

Revision

SL6NS

SL6RQ

SL6GD

SL6GF

SL6GG

SL6GH

SL6RR

SL6GV

SL6GV

SL6W3

SL6YS

SL6W6

SL6YT

SL6W7

SL6YU

SL6W8

SL6YV

SL6W9

SL6YW

SL6WA

SL6YX

SL6WB

SL6YY

SL6XK

SL6XL

C1

C1

C1

C1

C1

C1

C1

C1

D1

D1

D1

D1

D1

D1

D1

D1

D1

0F27H

0F27H

0F27H

0F27H

0F27H

0F27H

0F27H

0F27H

0F29H

0F29H

0F29H

0F29H

0F29H

0F29H

0F29H

0F29H

0F29H

2.8/533

2/533

2.40/533

2.66/533

2.8/533

3.06/533

1.60/400

1.60/400

1.80/400

2/400

2.20/400

2.40/400

2.60/400

2.8/0400

3/400

1.60/400

2/400

512-KB

512-KB

512-KB

512-KB

512-KB

512-KB

512-KB

512-KB

512-KB

512-KB

512-KB

512-KB

512-KB

512-KB

512-KB

512-KB

512-KB

01

01

01

01

01

01

01

01

01

01

01

01

01

01

01

01

01


FC-PGA2 package


FC-PGA2 package


FC-PGA2 package


FC-PGA2 package


FC-PGA2 package


FC-PGA2 package


FC-PGA2 package


FC-PGA2 package


FC-PGA2 package


FC-BGA package


FC-BGA package


FC-BGA package


FC-BGA package


FC-BGA package


FC-BGA package


FC-BGA package


FC-PGA2 package


FC-PGA2 package

Notes

2, 5

2, 5

5

5

5

5

5, 6

2, 6

7

7

7

3, 6

2

7

3

2

3

2

3

2

3

2

3

2

3

2

Intel

®

Xeon

®


15


Table 1. Intel

®

Xeon

®


S-Spec

Number

SL74T

Core

Stepping

D1

Processor

Signature

0F29H

Speed

Core/Front

Side Bus

(GHz/MHz)

2.40/533

L2 Size

(Kbytes)

512-KB

L3 Size

(Kbytes)

Processor

Interposer

Revision

01

Package and

Revision

SL6VK

SL6YM

SL6VL

SL6YN

SL6VM

SL6NR

SL6VN

SL6YQ

SL6VP

SL6YR

SL73K

SL72C

SL73L

SL72D

SL73M

SL72E

SL73N

SL72F

SL73P

SL72G

SL73Q

SL72Y

SL7AE

SL7BW

D1

D1

D1

D1

D1

M0

M0

M0

M0

M0

M0

M0

M0

0F29H

0F29H

0F29H

0F29H

0F29H

0F25H

0F25H

0F25H

0F25H

0F25H

0F25H

0F25H

0F25H

2/533

2.40/533

2.66/533

2.80/533

3.06/533

2/533

2.40/533

2.66/533

2.80/533

3.06/533

3.2/533

3.2/533

2.80/533

512-KB

512-KB

512-KB

512-KB

512-KB

512-KB

512-KB

512-KB

512-KB

512-KB

512-KB

512-KB

1-MB

1-MB

2-MB

01

01

01

01

01

01

01

01

01

01

01

01

01


FC-PGA2 package

604-pin micro-PGA

Interposer with 42.5 mm

FC-PGA2 package


FC-PGA2 package


FC-PGA2 package


FC-PGA2 package


FC-PGA2 package


FC-PGA2 package


FC-PGA2 package


FC-PGA2 package


FC-PGA2 package


FC-PGA2 package


FC-PGA2 package


FC-PGA2 package

604-pin micro-PGA SL7D5

SL7DG

SL7D4

SL7DF

SL8TJ

M0

L0

0F25H

0F29h

2.40/533

0F29H

512-KB

512-KB

2.40/533

1-MB

1-MB

512-KB

01

01

604-pin micro-PGA

SL8TK L0 0F29H 2.80/533 512-KB 01


FC-PGA2 package


FC-PGA2 package

Notes

7

2

2

2

5

2

2, 6, 8

6, 8

2, 6, 8

6, 8

2, 6, 9

6

6

2, 6,8

6

2, 6, 8

7

5

2

5

2

5

2

5

2

5, 6

2

2

16

Intel

®

Xeon

®



Table 1. Intel

®

Xeon

®


S-Spec

Number

SL8TL

Core

Stepping

L0

Processor

Signature

0F29H

Speed

Core/Front

Side Bus

(GHz/MHz)

2.40/533

L2 Size

(Kbytes)

512-KB

L3 Size

(Kbytes)

Processor

Interposer

Revision

01

Package and

Revision

SL8SE

SL8TH

L0

L0

0F29H

0F29H

2/400

1.60/400

512-KB

512-KB

01

01


FC-PGA2 package


FC-PGA2 package


FC-PGA2 package

Notes

5

2

7

7

NOTES:

1. The Intel

®

Xeon

®

processor listed here is installed onto a micro pin grid array (mPGA) interposer. The overall processor package is called INT-mPGA.

2. These parts are Intel boxed processors.

3. FC-BGA packaging maintains form, fit, and functionality when compared to OLGA packaging. Users may notice a color change.

4. These parts require the inputs from A20M#, IGNNE#, LINT[1]/NMI and LINT[0]/INTR pins during RESET to set the correct core to bus frequency ratio.

5. These parts are the Intel

®

Xeon

®

Processor with 533 MHz Front Side Bus.

6. These parts have a VID of 1.525V.

7. These parts are the Low Voltage Intel

®

Xeon


®

Xeon

®


®

Xeon

®

®

Processor.

Processor with 1-MB L3 Cache.

Processor with 2-MB L3 Cache.

Intel

®

Xeon

®


17

Mixed Steppings in DP Systems


Intel Corporation fully supports mixed steppings of Intel Xeon processors. The following list and processor matrix describes the requirements to support mixed steppings:

•

Mixed steppings are only supported with processors that have identical family numbers as indicated by the Processor Signature instruction. The Intel Xeon processor is available with two different Model numbers as indicated by the Processor Signature. Please refer to Table 2 for details regarding inclusion of processors with mixed Processor Signature/Core steppings.

•

While Intel has done nothing to specifically prevent processors operating at differing frequencies from functioning within a multiprocessor system, there may be uncharacterized errata that exist in such configurations. Intel does not support such configurations. In mixed stepping systems, all processors must operate at identical frequencies (i.e., the highest frequency rating commonly supported by all processors).

•

While there are no known issues associated with the mixing of processors with differing cache sizes in a multiprocessor system, and Intel has done nothing to specifically prevent such system configurations from operating, Intel does not support such configurations since there may be uncharacterized errata that exist. In mixed stepping systems, all processors must be of the same cache size.

•

While Intel believes that certain customers may wish to perform validation of system configurations with mixed frequency or cache sizes, and that those efforts are an acceptable option to our customers, customers would be fully responsible for the validation of such configurations.

•

Intel requires that the proper microcode update be loaded on each processor operating in a multiprocessor system. Any processor that does not have the proper microcode update loaded is considered by Intel to be operating out of specification.

•

The workarounds identified in this and following specification updates must be properly applied to each processor in the system. Certain errata are specific to the multiprocessor environment and are identified in Table 2 found at the end of this section. Errata for all processor steppings will affect system performance if not properly worked around. Also see

Table 1 for additional details on which processors are affected by specific errata.

•

In mixed stepping systems, the processor with the lowest feature-set, as determined by the

Processor Signature Feature Bytes, must be the bootstrap processor (BSP). In the event of a tie in feature-set, the tie should be resolved by selecting the BSP as the processor with the lowest stepping as determined by the Processor Signature instruction.

In the following processor matrix, “NI” indicates that there are currently no known issues associated with mixing these steppings. A number indicates that a known issue has been identified as listed in the table following the matrix. “X” indicates the processors cannot be mixed. A dual processor system using mixed processor steppings must assure that errata are addressed appropriately for each processor.

Intel

®

Xeon

®


18


Table 2. DP Platform Matrix for the Intel

®

Xeon

®

Processor

2

Processor

Signature/Core

Stepping

0F0Ah/C1

0F12h/D0

0F24h/B0

0F27h/C1

0F29h/D1

0F25h/M0

0F29h/L0

0F0Ah/C1

NI

Note 1

X

X

X

X

X

0F12h/D0

Note 1

NI

X

X

X

X

X

0F24h/B0

NI

NI

NI

X

X

NI

NI

0F27h/C1

NI

NI

NI

X

X

NI

NI

0F29h/D1 0F25h/M0

3

0F25h/M0

4

NI

NI

NI

X

X

NI

NI

NI

NI

NI

X

X

NI

NI

X

NI

X

X

X

X

X

NOTES:

1. Some of these processors are affected by errata, which may affect the features an MP system is able to support. See the

Table 1 for details on which processors are affected by these errata.

2. This Matrix also applies to the Intel

®

Xeon

®

Processor with 533 MHz Front Side Bus, Low Voltage Intel

®

Xeon

®

Processor,

Intel

®

Xeon

®

Processor with 1-MB L3 cache, and Intel

®

Xeon

®

Processor with 2-MB L3 cache.

3. This only applies to 0F25h stepping without L3 cache.

4. This only applies to 0F25h stepping with L3 cache.

Intel

®

Xeon

®


19

Summary Table of Changes


The following table indicates the Specification Changes, Errata, Specification Clarifications or

Documentation Changes, which apply to the listed MCH steppings. Intel intends to fix some of the errata in a future stepping of the component, and to account for the other outstanding issues through documentation or Specification Changes as noted. This table uses the following notations:

Codes Used in Summary Table

X: Erratum, Specification Change or Clarification that applies to the given processor stepping.

(No mark) or (Blank Box): This erratum is fixed in listed stepping or specification change does not apply to listed stepping.

Doc:

Plan Fix:

Fixed:

Document change or update that will be implemented.

This erratum may be fixed in a future of the product.

This erratum has been previously fixed.

No Fix:

PKG:

AP

There are no plans to fix this erratum.

This column refers to errata on the Intel Xeon processor substrate.

APIC-related erratum.

Change bar to left of table row indicates this item is either new or modified from the previous version of this document.

Each Specification Update item is prefixed with a capital letter to distinguish the product. The key below details the letters that are used in Intel’s microprocessor Specification Updates:

A = Intel

®

Pentium

®

II processor

B = Mobile Intel

®

Pentium

®

II processor

C = Intel

®

Celeron

®

processor

D = Dual-Core Intel

®

Xeon

®

Processor 2.8 GHz

E = Intel

®

Pentium

®

III processor

F = Intel

®

Pentium

®

processor Extreme Edition and Intel

®

Pentium

®

D processor

G = Intel

®

Pentium

®

III Xeon™ processor

H = Mobile Intel

®

Celeron

®

processor at 466/433/400/366/333/300 and 266 MHz

J = 64-bit Intel

®

Xeon

®

processor MP with 1MB L2 cache

K = Mobile Intel

®

Pentium

®

III processor

L = Intel

®

Celeron

®

D processor

M = Mobile Intel

®

Celeron

®

processor

N = Intel

®

Pentium

®

4 processor

O = Intel

®

Xeon

®

processor MP

P = Intel

®

Xeon

®

processor

Q = Mobile Intel

®

Pentium

®

4 processor supporting Hyper-Threading Technology on 90-nm process technology

R = Intel

®

Pentium

®

4 processor on 90 nm process

S = 64-bit Intel

®

Xeon

®

processor with 800 MHz system bus (1 MB and 2 MB L2 cache versions)

20

Intel

®

Xeon

®



T = Mobile Intel

®

Pentium

®

4 processor-M

U = 64-bit Intel

®

Xeon

®

processor MP with up to 8MB L3 Cache

V = Mobile Intel

®

Celeron

®

processor on .13 Micron Process in Micro-FCPGA Package

W = Intel

®

Celeron

®

M processor

X = Intel

®

Pentium

®

M processor on 90nm process with 2-MB L2 Cache

Y = Intel

®

Pentium

®

M processor

Z = Mobile Intel

®

Pentium

®

4 processor with 533 MHz system bus

AA = Intel

®

Pentium

® processor Extreme Edition and Intel

®

Pentium

®

D processor on 65 nm process

AB = Intel

®

Pentium

®

4 processor on 65 nm process

AC = Intel

®

Celeron

®

Processor in 478 Pin Package

AG= Dual-Core Intel

®

Xeon

®

Processor 5100 Series

AH= Intel

®

Core™2 Duo mobile Processor

AI= Intel

®

Core™2 Extreme Processor X6800 and Intel

®

Core™2 Duo Desktop Processor

E6000 Sequence

AJ = Quad-Core Intel® Xeon® Processor 5300 Series

AL= Dual-Core Intel

®

Xeon

®


AP= Dual-Core

®

Xeon

®


Note:

The Specification Updates for the Pentium

®

processor, Pentium

®

Pro processor, and other Intel products do not use this convention.

Intel

®

Xeon

®


21


Errata (Sheet 1 of 4)

No.

P1

C1/

0F0Ah

X

D0/

0F12h

X

B0/

0F24h

C1/

0F27h

D1/

0F29h

M0/

0F25h

L0/

0F29h

Plans

Fixed

P2

P3

P4

P5

P6

P7

P8

P9

P10

P11

P12

P13

P14

P15

P16

P17

P18

P19

P20

P21

P22

P23

P24

P25

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

No Fix

No Fix

No Fix

No Fix

Fixed

Fixed

Fixed

No Fix

No Fix

No Fix

Fixed

Fixed

Fixed

No Fix

No Fix

Fixed

Fixed

Fixed

Fixed

Fixed

No Fix

Fixed

Fixed

Fixed

Errata

UC Code in same line as write back (WB) data may lead to data corruption

Transaction is not retried after BINIT#

Invalid opcode 0FFFh requires a ModRM byte

When in no-fill mode (CR0.CD=1) the memory type of large (PSE-4M and PAE-2M) pages are wrongly forced to uncacheable

Processor may hang due to speculative page walks to nonexistent system memory

Writing a performance counter may result in an incorrect counter value

Performance counter may contain incorrect value after being stopped

REP MOV instruction with overlapping source and destination may result in data corruption

Memory type of the load lock different from its corresponding store unlock

Machine check architecture error reporting and recovery may not work as expected

Debug mechanisms may not function as expected

Processor may live-lock if PDEs or PTEs are in UC space

Thermal status log bit may not be set when the thermal control circuit is active

Processor may timeout waiting for a device to respond after 0.67 seconds

Cascading of performance counters does not work correctly when forced overflow is enabled

EMON event counting of x87 loads may not work as expected

Simultaneous code breakpoint and uncorrectable error results in processor hang

Software controlled clock modulation using a

12.5% or 25% duty cycle may cause the processor to hang

RFO with ECC error may result in incorrect data

Speculative page-fault may cause livelock

PAT index MSB may be calculated incorrectly

System bus interrupt messages without data and which receive a Hard failure response may hang the processor

SQRTPD and SQRTSD may return QNaN indefinite instead of negative zero

Bus invalidate line requests that returns unexpected data may result in L1 cache corruption

Multiprocessor boot protocol may not complete with an IOQ depth of one

22

Intel

®

Xeon

®




No.

P26

P27

P28

P29

P30

P31

P32

P33

P34

P35

P36

P37

P38

P39

P40

P41

P42

P43

P44

P45

P46

C1/

0F0Ah

X

D0/

0F12h

X

B0/

0F24h

X

C1/

0F27h

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

D1/

0F29h

X

M0/

0F25h

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

L0/

0F29h

X

X

X

X

X

X

X

X

X

X

Plans Errata

No Fix

Fixed

No Fix

Fixed

No Fix

Fixed

No Fix

No Fix

No Fix

Fixed

Fixed

Fixed

Fixed

The processor signals page-fault exception (#pf) instead of alignment check exception (#ac) on an unlocked cmpxchg8b instruction

Incorrect data may be returned when page tables are located in write combining (WC) memory

FSW may not be completely restored after page-fault on FRSTOR or FLDENV instructions

Write combining (WC) load may result in an unintended address on system bus

Processor provides a 4-byte store unlock after an

8-byte load lock

Multiple accesses to the same S-state L2 cache line and ECC error combination may result in loss of cache coherency

IA32_MC0_ADDR and IA32_MC0_MISC registers will contain invalid or stale data following a data, address, or response parity error

When the processor is in the system management mode (SMM), debug registers may be fully writeable

Associated counting logic must be configured when using event selection control (ESCR) MSR

Livelock may occur when bus parking is disabled

CR2 May be incorrect or an incorrect page-fault error code may be pushed on to stack after execution of an LSS instruction

Buffer on resistance may exceed specification

Instruction pointer stored on stack may become invalid

No Fix

Fixed

Fixed

Fixed

Shutdown and IERR# may result due to a machine check exception on a Hyper-threading Technology enabled processor

Hyper-Threading Technology enabled processors may hang in the presence of extensive self-modifying code

Global bit incorrectly set for secondary logical processors in ITLB

Machine check exception (MCE) observed on DP platforms

Plan Fix

BPM[5:3]# VIL does not meet specification

No Fix

Processor may hang under certain frequencies and

12.5% STPCLK# duty cycle

No Fix

Fixed

System may hang if a fatal cache error causes bus write line (BWL) transaction to occur to the same cache line address as an outstanding bus read line

(BRL) or bus read-invalidate line (BRIL)

L2 cache may contain stale data in the exclusive state

Intel

®

Xeon

®


23



No.

P47

P48

P49

P50

P51

P52

P53

P54

P55

P56

P57

P58

P59

P60

P61

P62

P63

P64

P65

P66

P67

C1/

0F0Ah

X

D0/

0F12h

X

B0/

0F24h

X

C1/

0F27h

X

D1/

0F29h

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

M0/

0F25h

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

L0/

0F29h

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

Plans Errata

No Fix

No Fix

Fixed

No Fix

Fixed

Fixed

No Fix

Re-mapping the APIC base address to a value less than or equal to 0xDC001000 may cause I/O and special cycle failure

Erroneous BIST result found in EAX register after reset

Processor does not flag #GP on non-zero write to certain MSRs

Simultaneous assertion of A20M# and INIT# may result in incorrect data fetch

Processor does not respond to break requests from ITP

Glitches on address and data strobe signals may cause system shutdown

A write to an APIC Register Sometimes May

Appear to Have Not Occured

Plan Fix STPCLK# signal assertion under certain conditions may cause a system hang

Plan Fix

Store to load data forwarding may result in switched data bytes

Plan Fix

ITP cannot continue single step execution after the first breakpoint

No Fix

Parity error in the L1 cache may cause the processor to hang

Plan Fix

The TCK input in the test access port (TAP) is sensitive to low clock edge rates and prone to noise coupling onto TCK's rising or falling edges

No Fix

Disabling a local APIC disables both logical processor on a Hyper-Threading Technology enabled processor

No Fix

Using STPCLK and executing code from very slow memory could lead to a system hang

Plan Fix

Simultaneous cache line eviction from L2 and L3 caches may result in the write back of stale data

No Fix

The state of the resume flag (RF flag) in a task-state segment (TSS) may be incorrect

No Fix

Plan Fix

Simultaneous page-faults at similar page offsets on both logical processors of an Hyper-Threading

Technology enabled processor may cause application failure

No Fix

Changes to CR3 register do not fence pending instruction page

No Fix

No Fix

A 16-bit address wrap resulting from a near branch

(jump or call) may cause an incorrect address to be reported to the #GP exception handler

Locks and SMC detection may cause the processor to temporarily hang

Incorrect debug exception (#DB) may occur when a data breakpoint is set on a FP instruction

24

Intel

®

Xeon

®




No.

P68

P69

P70

P71

P72

P73

P74

P75

P76

P77

P78

P79

P80

P81

P82

P83

C1/

0F0Ah

D0/

0F12h

B0/

0F24h

C1/

0F27h

D1/

0F29h

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

M0/

0F25h

X

L0/

0F29h

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

Plans Errata

No Fix

No Fix

No Fix

Incorrect duty cycle is chosen when On-Demand

Clock Modulation is enabled in a processor supporting Hyper-Threading Technology

Plan Fix

Memory aliasing of pages as uncacheable memory type and write back (WB) may hang the system

Plan Fix

A timing marginality in the Instruction Decoder unit may cause an unpredictable application behavior and/or system hang

No Fix

Modified cache line eviction from L2 cache may result in write back of stale data xAPIC may not report some illegal vector errors

No Fix

Missing Stop Grant Acknowledge special bus cycle may cause a system hang

Machine check exceptions may not update

Last-Exception Record MSRs (LERs)

No Fix Stores to page tables may not be visible to page walks for subsequent loads without serializing or invalidating the page table entry

Plan Fix A timing marginality in the Arithmetic Logic Unit

(ALU) may cause indeterminate behavior

No Fix With Trap Flag (TF) asserted, FP instruction that triggers an unmasked FP exception may take single step trap before retirement of instruction

No Fix PDE/PTE loads and continuous locked updates to the same cache line may cause a system livelock

No Fix

No Fix

No Fix

No Fix

No Fix

Branch Trace Store (BTS) and Precise Event

Based Sampling (PEBS) may update memory outside the BTS/PREBS buffer

Memory Ordering Failure may occur with Snoop

Filtering Third-Party Agents after Issuing and completing a BWIL (Bus Write Invalidate Line) or

BLW (Bus Locked Write) transaction

Control Register 2 (CR2) can be updated during a

REP MOVS/STOS instruction with Fast Strings enabled

Writing the Local Vector Table (LVT) when an Interrupt is Pending May Cause an

Unexpected Interrupt

The Processor May Report a #TS Instead of a #GP

Fault

Intel

®

Xeon

®


25


Specification Changes

No.

SPECIFICATION CHANGES

P1 Context ID feature added to processor signature instruction feature Flags/IA32_MISC_Enable registers

Specification Clarifications

No.

SPECIFICATION CLARIFICATIONS

P1 Maximum ITCC specification correction

P2 Specification Clarification with respect to Time-Stamp Counter

Documentation Changes

No.

DOCUMENTATION CHANGES

None for this revision of the Specification Update.

26

Intel

®

Xeon

®


Errata

Errata

P1

Problem:

Implication:

Workaround:

Status:

UC Code in same line as write back (WB) data may lead to data corruption

This erratum occurs when both code (being accessed as uncacheable [UC] or write combining

[WC]) and data (being accessed as write back [WB]) are placed in the same cache line. The UC fetch will cause the processor to self-snoop and generate an implicit WB. The data supplied by this implicit WB may be corrupted due to the way the processor is currently handling self-modifying code.

UC code located in the same cache line as WB data may lead to data corruption.

UC or WC code should not be located in the same 64-byte cache line as any location that is being stored to with WB data.

For the steppings effected, see the Summary Table of Changes.

P2

Problem:

Implication:

Workaround:

Status:

P3

Problem:

Implication:

Workaround:

Status:

Transaction is not retried after BINIT#

If the first transaction of a locked sequence receives a HITM# and DEFER# during the snoop phase it should be retried and the locked sequence restarted. However, if BINIT# is also asserted during this transaction, the transaction will not be retried.

When this erratum occurs, locked transactions will not be retried.

None at this time.


Invalid opcode 0FFFh requires a ModRM byte

Some invalid opcodes require a ModRM byte and other following bytes, while others do not. The invalid opcode 0FFFh did not require a ModRM in previous generation microprocessors such as

Pentium II or Pentium III processors, but it is required in the Intel Xeon processor

The use of an invalid opcode 0FFFh without the ModRM byte may result in a page or limit fault on the Intel Xeon processor.

To avoid this erratum use ModRM byte with invalid 0FFFh opcode.

For the steppings affected, see the Summary Table of Changes.

P4

Problem:

Implication:

Workaround:

Status:

When in no-fill mode (CR0.CD=1) the memory type of large (PSE-4M and

PAE-2M) pages are wrongly forced to uncacheable

When the processor is operating in no-fill mode (CR0.CD=1), the page miss hardware incorrectly forces the memory type of large (PSE-4M and PAE-2M) pages to UC memory type regardless of the MTRR settings. By forcing the memory type of these pages to UC, load operations, which should hit valid data in the L1 cache, are forced to load the data from system memory. Some applications will lose the performance advantage associated with the caching permitted by other memory types.

This erratum may result in some performance degradation when using no-fill mode with large pages.

None at this time.

For the steppings affected, see the Summary Table of Changes

Intel

®

Xeon

®


27

Errata

P5

Problem:

Implication:

Workaround:

Status:

Processor may hang due to speculative page walks to nonexistent system memory

A load operation issued speculatively by the processor that misses the data translation lookaside buffer (DTLB) results in a page walk. A branch instruction older than the load retires so that this load operation is now in the mispredicted branch path. Due to an internal boundary condition, in some instances the load is not canceled before the page walk is issued.

The page miss handler (PMH) starts a speculative page walk for the load and issues a cacheable load of the page directory entry (PDE). This PDE load returns data that points to a page table entry in uncacheable (UC) memory. The PMH issues the PTE Load to UC space, which is issued on the front side bus. No response comes back for this load PTE operation since the address is pointing to system memory, which does not exist.

This load to non-existent system memory causes the processor to hang because other bus requests are queued up behind this UC PTE load, which never gets a response. If the load was accessing valid system memory, the speculative page-walk would successfully complete and the processor would continue to make forward progress.

Processor may hang due to speculative page walks to non-existent system memory.

Page directories and page tables in UC memory space must point to system memory that exists.


P6

Problem:

Implication:

Workaround:

Status:

Writing a performance counter may result in an incorrect counter value

Accessing a performance counter also enables the counter input so that writing one half of the counter can cause the other half to increment. When a performance counter is written and the event counter for the event being monitored is non-zero, the performance counter will be incremented by the value on that event counter. Because the upper eight bits of the performance counter are not written at the same time as the lower 32 bits, the increment due to the non-zero event counter may cause a carry to the upper bits such that the performance counter contains a value higher than what was written. The worst-case error caused by this can be about 4 billion counts.

When this erratum occurs, the performance counter will contain a different value from that which was written.

If the performance counter is set to select a null event and the counter configuration control register

(CCCR) for that counter has its compare bit set to zero, before the performance counter is written, this erratum will not occur.


P7

Problem:

Implication:

Workaround:

Status:

Performance counter may contain incorrect value after being stopped

If a performance counter is stopped on the precise internal clock cycle where the intermediate carry from the lower 32 bits of the counter to the upper eight bits occurs, the intermediate carry is lost.

When this erratum occurs the performance counter may contain a value about 4 billion (

2 than it should.

32) less

Since this erratum does not occur if the performance counters are read when running, a possible workaround is to read the counter before stopping it.


P8

Problem:

REP MOV instruction with overlapping source and destination may result in data corruption

When fast strings are enabled and a REP MOV instruction is used to move a string and the source and destination strings overlap by 56 bytes or less, data corruption may occur.

28

Intel

®

Xeon

®


Errata

Implication:

Workaround:

Status:

When this erratum occurs, data corruption may occur.

It is possible for the BIOS to contain a workaround for this erratum.


P9

Problem:

Implication:

Workaround:

Status:

Memory type of the load lock different from its corresponding store unlock

A use-once protocol is employed to ensure that the processor in a multi-agent system may access data that is loaded into its cache on a RFO operation at least once before it is snooped out by another agent. This protocol is necessary to avoid a multi-agent livelock scenario in which the processor cannot gain ownership of a line and modify it before that data is snooped out by another agent. In the case of this erratum, split load lock instructions incorrectly trigger the use-once protocol. A load lock operation accesses data that splits across a page boundary with both pages of

WB memory type. The use-once protocol activates and the memory type for the split halves get forced to UC. Since use-once does not apply to stores, the store unlock instructions go out as WB memory type. The full sequence on the bus is: locked partial read (UC), partial read (UC), partial write (WB), locked partial write (WB). The use-once protocol should not be applied to load locks.

When this erratum occurs, the memory type of the load lock will be different than the memory type of the store unlock operation. This behavior (load locks and store unlocks having different memory types) does not introduce any functional failures such as system hangs or memory corruption.

None at this time.


P10

Problem:

Machine check architecture error reporting and recovery may not work as expected

When the processor detects errors it should attempt to report and/or recover from the error. In the situations described below, the processor does not report and/or recover from the error(s) as intended.

•

When a transaction is deferred during the snoop phase and subsequently receives a Hard

Failure response, the transaction should be removed from the bus queue so that the processor may proceed. Instead, the transaction is not properly removed from the bus queue, the bus queue is blocked, and the processor will hang.

•

When a hardware prefetch results in an uncorrectable tag error in the L2 cache,

MC0_STATUS.UNCOR and MC0_STATUS.PCC are set but no machine check exception

(MCE) is signaled. No data loss or corruption occurs because the data being prefetched has not been used. If the data location with the uncorrectable tag error is subsequently accessed, an

MCE will occur. However, upon this MCE, or any other subsequent MCE,.the information for that error will not be logged because MC0_STATUS.UNCOR has already been set and the

MCA status registers will not contain information about the error which caused the MCE assertion but instead will contain information about the prefetch error event.

•

When the reporting of errors is disabled for machine check architecture (MCA) Bank 2 by setting all MC2_CTL register bits to 0, uncorrectable errors should be logged in the

IA32_MC2_STATUS register but no machine-check exception should be generated.

Uncorrectable loads on bank 2, which would normally be logged in the IA32_MC2_STATUS register, are not logged.

•

When one half of a 64 byte instruction fetch from the L2 cache has an uncorrectable error and the other 32 byte half of the same fetch from the L2 cache has a correctable error, the processor will attempt to correct the correctable error but cannot proceed due to the uncorrectable error.

When this occurs the processor will hang.

•

When an L1 cache parity error occurs, the cache controller logic should write the physical address of the data memory location that produced that error into the IA32_MC1_ADDR

REGISTER (MC1_ADDR). In some instances of a parity error on a load operation that hits

Intel

®

Xeon

®


29

Errata

30 the L1 cache, the cache controller logic may write the physical address from a subsequent load or store operation into the IA32_MC1_ADDR register.

•

When an error exists in the tag field of a cache line such that a read for ownership (RFO) issued by the processor hits multiple tag fields in the L2 cache (the correct tag and the tag with the error) and the accessed data also has a correctable error, the processor will correctly log the multiple tag match error but will hang when attempting to execute the machine check exception handler.

•

If a memory access receives a machine check error on both 64 byte halves of a 128-byte L2 cache sector, the IA32_MC0_STATUS register records this event as multiple errors, i.e., the valid error bit and the overflow error bit are both set indicating that a machine check error occurred while the results of a previous error were in the error-reporting bank. The

IA32_MC1_STATUS register should also record this event as multiple errors but instead records this event as only one correctable error.

•

The overflow bit should be set to indicate when more than one error has occurred. The overflow bit being set indicates that more than one error has occurred. Because of this erratum, if any further errors occur, the MCA overflow bit will not be updated, thereby incorrectly indicating only one error has been received.

•

If an I/O instruction (IN, INS, REP INS, OUT, OUTS, or REP OUTS) is being executed, and if the data for this instruction becomes corrupted, the processor will signal a MCE. If the instruction is directed at a device that is powered down, the processor may also receive an assertion of SMI#. Since MCEs have higher priority, the processor will call the MCE handler, and the SMI# assertion will remain pending. However, while attempting to execute the first instruction of the MCE handler, the SMI# will be recognized and the processor will attempt to execute the SMM handler. If the SMM handler is successfully completed, it will attempt to restart the I/O instruction, but will not have the correct machine state due to the call to the

MCE handler. This can lead to failure of the restart and shutdown of the processor.

•

If PWRGOOD is deasserted during a RESET# assertion causing internal glitches, the MCA registers may latch invalid information.

•

If RESET# is asserted, then deasserted, and reasserted, before the processor has cleared the

MCA registers, then the information in the MCA registers may not be reliable, regardless of the state or state transitions of PWRGOOD.

•

If MCERR# is asserted by one processor and observed by another processor, the observing processor does not log the assertion of MCERR#. The MCE handler called upon assertion of

MCERR# will not have any way to determine the cause of the MCE.

•

The Overflow Error bit (bit 62) in the IA32_MC0_STATUS register indicates, when set, that a machine check error occurred while the results of a previous error were still in the error reporting bank (i.e. The Valid bit was set when the new error occurred). If an uncorrectable error is logged in the error-reporting bank and another error occurs, the overflow bit will not be set.

•

The MCA Error Code field of the IA32_MC0_STATUS register gets written by a different mechanism than the rest of the register. For uncorrectable errors, the other fields in the

IA32_MC0_STATUS register are only updated by the first error. Any further errors that are detected will update the MCA Error Code field without updating the rest of the register, thereby leaving the IA32_MC0_STATUS register with stale information.

•

When a speculative load operation hits the L2 cache and receives a correctable error, the

IA32_MC1_Status Register may be updated with incorrect information. The

IA32_MC1_Status Register should not be updated for speculative loads.

•

The processor should only log the address for L1 parity errors in the IA32_MC1_Status register if a valid address is available. If a valid address is not available, the Address Valid bit in the IA32_MC1_Status register should not be set. In instances where an L1 parity error

Intel

®

Xeon

®


Errata

Implication:

Workaround:

Status:

occurs and the address is not available because the linear to physical address translation is not complete or an internal resource conflict has occurred, the Address Valid bit is incorrectly set.

•

The processor may hang when an instruction code fetch receives a hard failure response from the system bus. This occurs because the bus control logic does not return data to the core, leaving the processor empty. IA32_MC0_STATUS MSR does indicate that a hard fail response occurred.

•

The processor may hang when the following events occur and the machine check exception is enabled, CR4.MCE=1. A processor that has it's STPCLK# pin asserted will internally enter the

Stop Grant State and finally issue a Stop Grant Acknowledge special cycle to the bus. If an uncorrectable error is generated during the Stop Grant process it is possible for the Stop Grant special cycle to be issued to the bus before the processor vectors to the machine check handler.

Once the chipset receives its last Stop Grant special cycle it is allowed to ignore any bus activity from the processors. As a result, processor accesses to the machine check handler may not be acknowledged, resulting in a processor hang.

The processor is unable to correctly report and/or recover from certain errors.

None at this time.


P11

Problem:

Implication:

Workaround:

Status:

Debug mechanisms may not function as expected

Certain debug mechanisms may not function as expected on the processor. The cases are as follows:

•

When the following conditions occur: 1) An FLD instruction signals a stack overflow or underflow, 2) the FLD instruction splits a page-boundary or a 64 byte cache line boundary, 3) the instruction matches a debug register on the high page or cache line respectively, and 4) the

FLD

has a stack fault and a memory fault on a split access, the processor will only signal the stack fault and the debug exception will not be taken.

•

When a data breakpoint is set on the ninth and/or tenth byte(s) of a floating point store using the Extended Real data type, and an unmasked floating point exception occurs on the store, the break point will not be captured.

•

When any instruction has multiple debug register matches, and any one of those debug registers is enabled in DR7, all of the matches should be reported in DR6 when the processor goes to the debug handler. This is not true during a REP instruction. As an example, during execution of a REP MOVSW instruction the first iteration a load matches DR0 and DR2 and sets DR6 as FFFF0FF5h. On a subsequent iteration of the instruction, a load matches only

DR0. The DR6 register is expected to still contain FFFF0FF5h, but the processor will update

DR6 to FFFF0FF1h.

•

A Data breakpoint that is set on a load to uncacheable memory may be ignored due to an internal segment register access conflict. In this case the system will continue to execute instructions, bypassing the intended breakpoint. Avoiding having instructions that access segment descriptor registers e.g. LGDT, LIDT close to the UC load, and avoiding serialized instructions before the UC load will reduce the occurrence of this erratum.

Certain debug mechanisms do not function as expected on the processor.

None at this time.


Intel

®

Xeon

®


31

Errata

P12

Problem:

Implication:

Workaround:

Status:

Processor may live-lock if PDEs or PTEs are in UC space

The processor may livelock under the following boundary conditions:

•

The page-directory entries (PDEs) or page-table entries (PTEs) are in uncacheable (UC) space.

•

An instruction fetch misses the ITLB resulting in a page walk.

•

This instruction fetch is immediately followed by a store that splits a page boundary.

When this erratum occurs, the processor will livelock. This erratum was found using random instruction testing and has not been observed with commercial software.



P13

Problem:

Implication:

Workaround:

Status:

Thermal status log bit may not be set when the thermal control circuit is active

Bit 1 of the IA32_THERM_STATUS register (Thermal Status Log) is a sticky bit designed to be set to '1' if the thermal control circuit (TCC) has been active since either the previous processor reset or software cleared this bit. If TCC is active and the Thermal Status Log bit is cleared by a processor reset or by software, it will remain clear (set to ‘0’) as long as the TCC remains active.

Once TCC deactivates, the next activation of the TCC will set the Thermal Status Log bit.

When this erratum occurs, the Thermal Status Log bit will be cleared (set to ‘0’) although the thermal control circuit is active.

None at this time.


P14

Problem:

Implication:

Workaround:

Status:

Processor may timeout waiting for a device to respond after 0.67 seconds

The PCI 2.1 target initial latency specification allows two seconds for a device to respond during initialization-time. The processor may timeout after only approximately 0.67 seconds. When the processor times out it will hang with IERR# asserted. PCI devices that take longer than 0.67 seconds to initialize may not be initialized properly.

System may hang with IERR# asserted.

None at this time.


P15

Problem:

Implication:

Workaround:

Status:

Cascading of performance counters does not work correctly when forced overflow is enabled

The performance counters are organized into pairs. When the CASCADE bit of the CCCR is set, a counter that overflows will continue to count in the other counter of the pair. The FORCE_OVF bit forces the counters to overflow on every non-zero increment. When the FORCE_OVF bit is set, the counter overflow bit will be set but the counter no longer cascades.

The performance counters do not cascade when the FORCE_OVF bit is set.

None at this time.


P16

Problem:

EMON event counting of x87 loads may not work as expected

If a performance counter is set to count x87 loads and floating-point (FP) exceptions are unmasked, the FPU Operand (Data) Pointer (FDP) may become corrupted.

32

Intel

®

Xeon

®


Errata

Implication:

Workaround:

Status:

When this erratum occurs, FPU Operand (Data) Pointer (FDP) may become corrupted.

This erratum will not occur with floating point exceptions masked. If FP exceptions are unmasked, then performance counting of x87 loads should be disabled.


P17

Problem:

Implication:

Workaround:

Status:

Simultaneous code breakpoint and uncorrectable error results in processor hang

If an instruction fetch results in an uncorrectable error and there is also a debug breakpoint at this address, the processor will hang and the uncorrectable error will not be logged in the machine check registers.

When this erratum occurs the processor will hang.

None at this time.


P18

Problem:

Implication:

Workaround:

Status:

Software controlled clock modulation using a 12.5% or 25% duty cycle may cause the processor to hang

Per the ACPI 1.0b specification, processor clock modulation may be controlled via a processor register (IA32_THERM_CONTROL). The On-Demand Clock Modulation Duty Cycle is controlled by bits 3:1. If these bits are set to a duty cycle of 12.5% or 25%, the processor may hang while attempting to execute a FP instruction. In this failure, the last instruction pointer (LIP) is pointing to a FP instruction whose instruction bytes are in UC space and which takes an exception

16 (FP error exception). The processor stalls trying to fetch the bytes of the faulting FP instruction and those following it. This processor hang is caused by interactions between the thermal control circuit and FP event handler.

When the clock modulation is set to 12.5% or 25% duty cycle, the processor will go into a sleep state from which it fails to return.

Use a duty cycle other than 12.5% or 25%.


P19

Problem:

Implication:

Workaround:

Status:

RFO with ECC error may result in incorrect data

This erratum occurs as the result of the following conditions:

•

A Request for Ownership (RFO) generates a correctable error.

•

In the process of correcting the error, a locked RFO (LRFO) is issued that uses the same internal buffer as the previous RFO.

•

Another processor issues a snoop to the same address as the LRFO.

An internal boundary condition exists which may prevent the LRFO from completing correctly causing the snoop to receive incorrect data. Intel has not been able to reproduce this erratum with commercial software.

When this erratum occurs, data corruption may result.

None at this time.


P20

Problem:

Speculative page-fault may cause livelock

If the processor detects a page-fault, which is corrected before the operating system page-fault handler can be called (e.g., a second processor or DMA activity modifies the page tables and the

Intel

®

Xeon

®


33

Errata

Implication:

Workaround:

Status:

corrected page tables are left in a non-accessed or non-modified state) the processor may livelock.

Intel has not been able to reproduce this erratum with commercial software.

This erratum occurs in systems where page tables are being modified by other processors. If this erratum is encountered, the processor will livelock resulting in a system hang or operating system failure.



P21

Problem:

Implication:

Workaround:

Status:

PAT index MSB may be calculated incorrectly

When Mode B or Mode C paging support is enabled and all of the following events occur:

•

A page walk returns the page directory entry (PDE) for a large page from memory.

•

A subsequent page walk returns the page table entry (PTE) for a 4k page from memory and the page attribute table (PAT) upper index bit in this PTE is set to 1b.

It is possible that the PAT upper index bit in the PTE is incorrectly ignored and assumed to be

0b. The result is that the memory type in the PAT that should have come from the corresponding PAT index [4-7] incorrectly comes from PAT index [0-3].

If an operating system has programmed the PAT in an asymmetrical fashion i.e. PAT[0-3] is different from PAT[4-7] then an incorrect memory type may be used.

None at this time.


P22

Problem:

Implication:

Workaround:

Status:

System bus interrupt messages without data and which receive a Hard failure response may hang the processor

When a system bus agent (processor or chipset) issues an interrupt transaction without data onto the system bus and the transaction receives a HardFailure response, a potential processor hang can occur. The processor, which generates an inter-processor interrupt (IPI) that receives the

HardFailure response, will still log the MCA error event cause as HardFailure, even if the APIC causes a hang. Other processors, which are true targets of the IPI, will also hang on hardfail-without-data, but will not record an MCA HardFailure event as the cause. If a HardFailure response occurs on a system bus interrupt message with data, the APIC will complete the operation so as not to hang the processor.

The processor may hang.

None at this time.


P23

Problem:

Implication:

Workaround:

Status:

SQRTPD and SQRTSD may return QNaN indefinite instead of negative zero

When DAZ mode is enabled, and a SQRTPD or SQRTSD instruction has a negative denormal operand, the instruction will return a QNaN indefinite when the specified response should be zero.

When this erratum occurs, the instruction will return a QNaN indefinite when a zero is expected.

Ensure that negative denormals are not used as operands to the SQRTPD or SQRTSD instructions when DAZ mode is enabled. Software could enable FTZ mode to ensure that negative denormals are not generated by computation prior to execution of the SQRTPD or SQRTSD instructions.


34

Intel

®

Xeon

®


Errata

P24

Problem:

Bus invalidate line requests that returns unexpected data may result in L1 cache corruption

When a bus invalidate line (BIL) request receives unexpected data from a deferred reply, and a store operation write combines to the same address, there is a small window where the L1 cache is corrupt, and loads can retire with this corrupted data. This erratum occurs in the following scenario:

•

A RFO transaction is issued by the processor and hits a line in shared state in the L2 cache.

•

The RFO is then issued on the system bus as a 0 length read-invalidate (BIL), since it doesn’t need data, just ownership of the cache line.

•

This transaction is deferred by the chipset.

•

At some later point, the chipset sends a deferred reply for this transaction with an implicit write-back response. For this erratum to occur, no snoop of this cache line can be issued between the BIL and the deferred reply.

•

The processor issues a write-combining store to the same cache line while data is returning to the processor. This store straddles an 8-byte boundary.

Note:

Due to an internal boundary condition, a time window exists where the L1 cache contains corrupt data, which could be accessed by a load.

Implication:

Workaround:

The L1 cache may contain corrupted data. No known commercially available chipsets trigger the failure conditions.

The chipset could issue a BIL (snoop) to the deferred processor to eliminate the failure conditions.

Status:


P25

Problem:

Implication:

Workaround:

Status:

Multiprocessor boot protocol may not complete with an IOQ depth of one

When the in-order queue (IOQ) depth is managed by the chipset to be one entry deep, the system may hang during the multiprocessor boot protocol. This hang occurs when the chipset drives BNR# in such a way that the processors are continually throttled off the bus then released to access the bus in alternating cycles which never allows the multiprocessor boot protocol to complete execution.

The system may hang during the multiprocessor boot protocol.

If the chipset drives BNR# in such a way that the processors are continually throttled off the bus then released to access the bus in alternating cycles, do not use IOQ de-pipelining.


P26

Problem:

Implication:

Workaround:

Status:

The processor signals page-fault exception (#pf) instead of alignment check exception (#ac) on an unlocked cmpxchg8b instruction

If a page-fault exception (#pf) and alignment check exception (#ac) both occur for an unlocked cmpxc8b

instruction, then #PF will be flagged.

Software that depends the #AC before the #PF will be affected since #PF is signaled in this case.

Remove the software’s dependency on the fact that #AC has precedence over #PF. Alternately, correct the page-fault in the page-fault handler and then restart the faulting instruction.


Intel

®

Xeon

®


35

Errata

P27

Problem:

Implication:

Workaround:

Status:

Incorrect data may be returned when page tables are located in write combining (WC) memory

If page directories and/or page tables are located in WC memory, speculative loads to cacheable memory may complete with incorrect data.

Cacheable loads to memory mapped using page tables located in WC memory may return incorrect data. Intel has not been able to reproduce this erratum with commercially available software.

Do not place page directories and/or page tables in WC memory.


P28

Problem:

Implication:

Workaround:

Status:

FSW may not be completely restored after page-fault on FRSTOR or

FLDENV instructions

If the FPU operating environment or FPU state (operating environment and register stack) being loaded by an FLDENV or FRSTOR instruction wraps around a 64-Kbyte or 4-Gbyte boundary and a

#PF or segment limit fault (#GP or #SS) occurs on the instruction near the wrap boundary, the upper byte of the FPU status word (FSW) might not be restored. If the fault handler does not restart program execution at the faulting instruction, stale data may exist in the FSW.

When this erratum occurs, stale data will exist in the FSW.

Ensure that the FPU operating environment and FPU state do not cross 64-Kbyte or 4-Gbyte boundaries. Alternately, ensure that the page-fault handler restarts program execution at the faulting instruction after correcting the paging problem.


P29

Problem:

Implication:

Workaround:

Status:

Write combining (WC) load may result in an unintended address on system bus

When the processor performs a speculative WC load, down the path of a mispredicted branch, and the address happens to match a valid UC address translation with the DTLB, an unintended

UnCacheable load operation may be sent out on the system bus.

When this erratum occurs, an unintended load may be sent on system bus. Intel has only encountered this erratum during pre-silicon simulation.

It is possible for the BIOS to contain a workaround for this erratum for some steppings of the processor.


P30

Problem:

Implication:

Workaround:

Status:

Processor provides a 4-byte store unlock after an 8-byte load lock

When the processor is in the page address extension (PAE) mode and detects the need to set the

Access and/or Dirty bits in the page directory or page table entries, the processor sends an 8 byte load lock onto the system bus. A subsequent 8 byte store unlock is expected, but instead a 4 byte store unlock occurs. Correct data is provided since only the lower bytes change, however external logic monitoring the data transfer may be expecting an 8-byte store unlock.

No known commercially available chipsets are affected by this erratum.

None at this time.

For the steppings affected, see the Summary Table of Changes

36

Intel

®

Xeon

®


Errata

P31

Problem:

Implication:

Workaround:

Status:

Multiple accesses to the same S-state L2 cache line and ECC error combination may result in loss of cache coherency

When a RFO cycle has a 64 bit address match with an outstanding read hit on a line in the L2 cache which is in the S-state AND that line contains an ECC error, the processor should recycle the RFO until the ECC error is handled. Due to this erratum, the processor does not recycle the RFO and attempt to service both the RFO and the read hit at the same time.

When this erratum occurs, cache may become incoherent.

None at this time.


P32

Problem:

Implication:

Workaround:

Status:

IA32_MC0_ADDR and IA32_MC0_MISC registers will contain invalid or stale data following a data, address, or response parity error

If the processor experiences a data, address, or response parity error, the ADDRV and MISCV bits of the IA32_MC0_STATUS register are set, but the IA32_MC0_ADDR and IA32_MC0_MISC registers are not loaded with data regarding the error.

When this erratum occurs, the IA32_MC0_ADDR and IA32_MC0_MISC registers will contain invalid or stale data.

Ignore any information in the IA32_MC0_ADDR and IA32_MC0_MISC registers after a data or response parity error.


P33

Problem:

Implication:

Workaround:

Status:

When the processor is in the system management mode (SMM), debug registers may be fully writeable

When in system management mode (SMM), the processor executes code and stores data in the

SMRAM space. When the processor is in this mode and writes are made to DR6 and DR7, the processor should block writes to the reserved bit locations. Due to this erratum, the processor may not block these writes. This may result in invalid data in the reserved bit locations.

Reserved bit locations within DR6 and DR7 may become invalid.

Software may perform a read/modify/write when writing to DR6 and DR7 to ensure that the value in the reserved bits is maintained.


P34

Problem:

Implication:

Workaround:

Associated counting logic must be configured when using event selection control (ESCR) MSR

ESCR MSRs allow software to select specific events to be counted, with each ESCR usually associated with a pair of performance counters. ESCRs may also be used to qualify the detection of at-retirement events that support precise-event-based sampling (PEBS). A number of performance metrics that support PEBS require a 2nd ESCR to tag uops for the qualification of at-retirement events. (The first ESCR is required to program the at-retirement event.) Counting is enabled via counter configuration control registers (CCCR) while the event count is read from one of the associated counters. When counting logic is configured for the subset of at-retirement events that require a 2nd ESCR to tag uops, at least one of the CCCRs in the same group of the 2nd ESCR must be enabled.

If no CCCR/counter is enabled in a given group, the ESCR in that group that is programmed for tagging uops will have no effect. Hence a subset of performance metrics that require a 2nd ESCR for tagging uops may result in 0 count.

Ensure that at least one CCCR/counter in the same group as the tagging ESCR is enabled for those performance metrics that require 2 ESCRs and tagging uops for at-retirement counting.

Intel

®

Xeon

®


37

Errata

Status:


P35

Problem:

Implication:

Workaround:

Status:

Livelock may occur when bus parking is disabled

A livelock may occur when processor bus parking is disabled, and when (1) the processor is the symmetric owner of the bus with one internal request pending, and (2) the processor observes the assertion of BPRI#, BNR# or a full IOQ. In this scenario, the processor bus interface unit assumes that the assertion of ADS# is not required, deasserts BREQ, and, as a result, relinquishes bus ownership without issuing the pending request. If the BPRI#, BNR# or full IOQ pattern continues coincident with the arbitration phase of the processor that still has only one outstanding internal request, livelock may occur. Assertion of bus parking, any change to the regular pattern of BPRI# or BNR# assertion noted above, or the arrival of a second internal transaction will release the processor from the livelock condition

.

This erratum may result in a livelock.

This erratum can be avoided by enabling bus parking. The deassertion of signal A15# during the active-to- inactive edge of RESET# will enable bus parking.


P36

Problem:

Implication:

Workaround:

Status:

CR2 May be incorrect or an incorrect page-fault error code may be pushed on to stack after execution of an LSS instruction

Under certain timing conditions, the internal load of the selector portion of the LSS instruction may complete with potentially incorrect speculative data before the load of the offset portion of the address completes. The incorrect data is corrected before the completion of the LSS instruction but the value of CR2 and the error code pushed on the stack are reflective of the speculative state. Intel has not observed this erratum with commercially available software.

When this erratum occurs, the contents of CR2 may be off by two, or an incorrect page-fault error code may be pushed onto the stack.



P37

Problem:

Implication:

Workaround:

Status:

Buffer on resistance may exceed specification

The datasheet specifies the resistance range for buffer on resistance (RON) for the AGTL+ buffer as 5 to 11 ohms. Due to this erratum, RON may be as high as 12 ohms.

The RON value affects the voltage level of the signals when the buffer is driving the signal low. A higher RON may adversely affect the system's ability to meet specifications such as VIL. As the system design also affects margin to specification, designs may or may not have sufficient margin to function properly with an increased RON. System designers should evaluate whether a particular system is affected by this erratum. Designs that follow the recommendations in the Intel

®

Xeon

®

Processor and Intel

®

860 Chipset Platform Design Guide are not expected to be affected.

No workaround is necessary for systems with margin sufficient to accept a higher R

ON

.


P38

Problem:

Instruction pointer stored on stack may become invalid

The instruction pointer stored on the stack my become invalid due to an internal boundary condition which may exist on a Hyper-Threading Technology (HT Technology) enabled processors. The following sequence of events must occur in order to encounter this erratum:

•

One logical processor executes the WRMSR instruction with incorrect data causing a general protection fault.

38

Intel

®

Xeon

®


Errata

Implication:

Workaround:

Status:

•

Simultaneously, an event that requires micro-architectural synchronization among the two logical processors occurs on the second logical processor. This event may cause an invalid instruction pointer to be stored on the ring 0 stack during the transition to GP fault handler on the first logical processor.

The instruction pointer stored on the stack may be invalid, potentially causing errors during execution of or return from the GP fault handler.



P39

Problem:

Implication:

Workaround:

Status:

Shutdown and IERR# may result due to a machine check exception on a

Hyper-threading Technology enabled processor

When a machine check exception (MCE) occurs due to an internal error, both logical processors on a HT Technology enabled processor normally vector to the MCE handler. However, if one of the logical processors is in the “Wait-for-SIPI” state, that logical processor will not have an MCE handler and will shut down and assert IERR#.

A processor with a logical processor in the “Wait for SIPI” state will shut down when an MCE occurs on the other thread.

None at this time.


P40

Problem:

Implication:

Workaround:

Status:

Hyper-Threading Technology enabled processors may hang in the presence of extensive self-modifying code

For multiprocessor platforms, in which HT Technology enabled processors are executing extensive self modifying code, and branch trace messages are enabled on at least one logical processor, the system may hang. In this scenario, a processor executing within 1K of code being written to by another processor may attempt to end this flow, thereby resulting in a hang.

When this erratum occurs the system will hang.



P41

Problem:

Global bit incorrectly set for secondary logical processors in ITLB

Due to a boundary condition in the translation look-aside buffer logic, the global bit information in the TLB entry for a mapping belonging to the first logical processor can overwrite the global bit information for a mapping belonging to the second logical processor. This occurs in the following scenario:

•

The first logical processor misses the ITLB resulting in a page walk.

•

The second logical processor also misses the ITLB and generates a page walk.

Note:

In certain timing scenarios within the processor, the leftover global bit information from the first logical processor may overwrite the second logical processor.

Implication:

When this erratum occurs, if the Page global bit for the second logical processor is overwritten with a 0b, this will result in performance degradation for the first logical processor. If the page global bit is incorrectly changed from a 0 to 1, this erratum may result in software failures.

Workaround:

Status:

It is possible for BIOS code to contain a workaround for this erratum.


Intel

®

Xeon

®


39

Errata

P42

Problem:

Implication:

Workaround:

Status:

Machine check exception (MCE) observed on DP platforms

A system bus address parity error may be signaled if two processors run at odd core frequency to system Bus-ratios (17:1, 19:1, etc.) on DP processor platforms. This address parity error signaling issue does not occur if the processors run at even bus-ratios.

A MCE may be observed on DP platforms.

The system BIOS should ensure that processors run at even core frequency to system bus-ratios

(16:1, 18:1, etc.).


P43

Problem:

Implication:

Workaround:

Status:

BPM[5:3]# VIL does not meet specification

The VIL for BPM[5:3]# is specified as 0.9 * GTLREF [V]. Due to this erratum the VIL for these signals is 0.9 * GTLREF -.100 [V].

The processor requires a lower input voltage than specified to recognize a low voltage on the

BPM[5:3]# signals.

When intending to drive the BPM[5:3]# signals low, ensure that the system provides a voltage lower than (0.9 * GTLREF -.100 [V]).


P44

Problem:

Implication:

Workaround:

Status:

Processor may hang under certain frequencies and 12.5% STPCLK# duty cycle

If a system deasserts STPCLK# at a 12.5% duty cycle, and the processor is running below 2 GHz, and the processor thermal control circuit (TCC) on-demand clock modulation is active, the processor may hang. This erratum does not occur under the automatic mode of the TCC.

When this erratum occurs, the processor will hang.

If use of the on-demand mode of the processor's TCC is desired in conjunction with STPCLK# modulation, then assure that STPCLK# is not asserted at a 12.5% duty cycle.


P45

Problem:

Implication:

Workaround:

Status:

System may hang if a fatal cache error causes bus write line (BWL) transaction to occur to the same cache line address as an outstanding bus read line (BRL) or bus read-invalidate line (BRIL)

A processor internal cache fatal data ECC error may cause the processor to issue a bus write line

(BWL) transaction to the same cache line address as an outstanding bus read line (BRL) or bus read-invalidate line (BRIL). As it is not typical behavior for a single processor to have a BWL and a BRL/BRIL concurrently outstanding to the same address, this may represent an unexpected scenario to system logic within the chipset.

The processor may not be able to fully execute the machine check handler in response to the fatal cache error if system logic does not ensure forward progress on the system bus under this scenario.

System logic should ensure completion of the outstanding transactions. Note that during recovery from a fatal data ECC error, memory image coherency of the BWL with respect to BRL/BRIL transactions is not important. Forward progress is the primary requirement.


P46

Problem:

L2 cache may contain stale data in the exclusive state

If a cache line (A) is in Modified (M) state in the write-combining (WC) buffers and in the Invalid

(I) state in the L1 cache and it's adjacent sector (B) is in the Invalid (I) state and the following scenario occurs:

40

Intel

®

Xeon

®


Errata

Implication:

Workaround:

Status:

1. A read to B misses in the L2 cache and allocates cache line B and its associated second-sector pre-fetch into an almost full bus queue.

2. A BRL to cache line B completes with HIT# and fills data in Shared (S) state.

3. The bus queue full condition causes the prefetch to cache line A to be canceled, cache line A will remain M in the WC buffers and I in the L2 while cache line B will be in the S state. Then, if the further conditions occur: a. Cache line A is evicted from the WC buffers to the bus queue which is still almost full.

b. A hardware prefetch RFO to cache line B, hits the S state in the L2 and observes cache line A in the I state, allocates both cache lines.

c. An RFO to cache line A completes before the WC buffers write modified data back, filling the L2 with stale data.

d. The write-back from the WC Buffers completes leaving stale data, for cache line A, in the

Exclusive (E) state in the L2 cache.

Stale data may be consumed leading to unpredictable program execution. Intel has not been able to reproduce this erratum in commercial software.

It is possible for BIOS to contain a workaround for this erratum.


P47

Problem:

Implication:

Workaround:

Status:

Re-mapping the APIC base address to a value less than or equal to

0xDC001000 may cause I/O and special cycle failure

Re-mapping the APIC base address from its default can cause conflicts with either I/O or special cycle bus transactions.

Either I/O or special cycle bus transactions can be redirected to the APIC, instead of appearing on the front side bus.

Use any APIC base addresses above 0xDC001000 as the relocation address.


P48

Problem:

Implication:

Workaround:

Status:

Erroneous BIST result found in EAX register after reset

The processor may show an erroneous built-in self test (BIST) result in the EAX register bit 0 after reset.

When this erratum occurs, an erroneous BIST failure will be reported in the EAX register bit 0, however this failure can be ignored since it is not accurate.

It is possible for BIOS to workaround this issue by masking off bit 0 in the EAX register where

BIST results are written.


P49

Problem:

Implication:

Workaround:

Status:

Processor does not flag #GP on non-zero write to certain MSRs

When a non-zero write occurs to the upper 32 bits of IA32_CR_SYSENTER_EIP or

IA32_CR_SYSENTER_ESP, the processor should indicate a general protection fault by flagging

#GP. Due to this erratum, the processor does not flag #GP.

The processor unexpectedly does not flag #GP on a non-zero write to the upper 32 bits of

IA32_CR_SYSENTER_EIP or IA32_CR_SYSENTER_ESP. No known commercially available operating system has been identified to be affected by this erratum.

None at this time.


Intel

®

Xeon

®


41

Errata

P50

Problem:

Implication:

Workaround:

Status:

Simultaneous assertion of A20M# and INIT# may result in incorrect data fetch

If A20M# and INIT# are simultaneously asserted by software, followed by a data access to the

0xFFFFFXXX memory region, with A20M# still asserted, incorrect data will be accessed. With

A20M# asserted, an access to 0xFFFFFXXX should result in a load from physical address

0xFFEFFXXX. However, in the case of A20M# and INIT# being asserted together, the data load will actually be from the physical address 0xFFFFFXXX. Code accesses are not affected by this erratum.

Processor may fetch incorrect data, resulting in BIOS failure.

Deasserting and reasserting A20M# prior to the data access will workaround this erratum.


P51

Problem:

Implication:

Workaround:

Status:

Processor does not respond to break requests from ITP

On power-up and low-power state transitions, the processor's TAP circuitry may remain in the tap- logic-reset (TLR) state.

The ITP is unable to cause a break on reset in the processor, which may prevent the loading of processor and chipset registers, or affect the ability to debug from cold boot and low power transitions

None at this time.


P52

Problem:

Implication:

Workaround:

Status:

Glitches on address and data strobe signals may cause system shutdown

When the Processor Signature instruction is executed with EAX = 2 on a processor without HT

Technology or with HT Technology disabled via power on configuration, it should return a value of 51h in EAX[15:8] to indicate that the instruction translation lookaside buffer (ITLB) has 128 entries. On a processor with HT Technology enabled, the processor should return 50h (64 entries).

Due to this erratum, the Processor Signature instruction always returns 50h (64 entries).

Software may incorrectly report the number of ITLB entries. Operation of the processor is not affected.

None at this time.


P53

Problem:

Implication:

Workaround:

Status:

A write to an APIC Register Sometimes May Appear to Have Not Occured

With respect to the retirement of instructions, stores to the uncacheable memory-based APIC register space are handled in a non-synchronized way. For example if an instruction that masks the interrupt flag, e.g. CLI, is executed soon after an uncacheable write to the Task Priority Register

(TPR) that lowers the APIC priority, the interrupt masking operation may take effect before the actual priority has been lowered. This may cause interrupts whose priority is lower than the initial

TPR, but higher than the final TPR, to not be serviced until the interrupt enabled flag is finally set, i.e. by STI instruction. Interrupts will remain pending and are not lost.

In this example the processor may allow interrupts to be accepted but may delay their service.

This non-synchronization can be avoided by issuing an APIC register read after the APIC register write. This will force the store to the APIC register before any subsequent instructions are executed. No commercial operating system is known to be impacted by this erratum.

For the steppings affected, see the Summary Table of Changes.

42

Intel

®

Xeon

®


Errata

P54

Problem:

Implication:

Workaround:

Status:

STPCLK# signal assertion under certain conditions may cause a system hang

The assertion of STPCLK# signal before a logical processor awakens from the “Wait-for-SIPI” state for the first time, may cause a system hang. A processor supporting HT Technology may fail to initialize appropriately, and may not issue a Stop Grant Acknowledge special bus cycle in response to the second STPCLK# assertion.

When this erratum occurs in an HT Technology enabled system, it may cause a system hang.

BIOS should initialize the second thread of the processor supporting HT Technology prior to

STPCLK# assertion. Additionally, it is possible for the BIOS to contain a workaround for this erratum.


P55

Problem:

Implication:

Workaround:

Status:

Store to load data forwarding may result in switched data bytes

If in a short window, after an instruction that updates a segment register has executed but not yet retired, there is a load occurring to an address that matches a recent previous store operation but the data size is smaller than the size of the store, the resulting data forwarded from the store to the load may have some of the lower bytes switched.

If this erratum occurs, the processor may execute with incorrect data.



P56

Problem:

Implication:

Workaround:

Status:

ITP cannot continue single step execution after the first breakpoint

ITP will not continue in single step execution after the first software breakpoint. ITP is unable to reset the resume flag (RF) bit in the EFLAGS register.

The processor will break at the instruction breakpoint address instead of single stepping.

Execution after the break will continue if DR7 bit 1 (Global Breakpoint Enable) is manually cleared.


P57

Problem:

Implication:

Workaround:

Status:

Parity error in the L1 cache may cause the processor to hang

If a locked operation accesses a line in the L1 cache that has a parity error, it is possible that the processor may hang while trying to evict the line.

If this erratum occurs, it may result in a system hang. Intel has not observed this erratum with any commercially available software.

None at this time.


P58

Problem:

Implication:

Workaround:

The TCK input in the test access port (TAP) is sensitive to low clock edge rates and prone to noise coupling onto TCK's rising or falling edges

TCK is susceptible to double clocking when low amplitude noise occurs on TCK edge, while it is crossing the receiver's transition region. TAP failures tend to increase with increases in background system noise.

This only impacts JTAG/TAP accesses to the processor. Other bus accesses are not affected.

To minimize the effects of this issue, reduce noise on the TCK-net at the processor relative to ground, and position TCK relative to BCLK to minimize the TAP error rate. Decreasing rise times to under 800ps reduced the failure rate but does not stop all failures.

Intel

®

Xeon

®


43

Errata

Status:


P59

Problem:

Implication:

Workaround:

Status:

Disabling a local APIC disables both logical processor on a

Hyper-Threading Technology enabled processor

Disabling a local APIC on one logical processor of a HT Technology enabled processor by clearing bit 11 of the IA32_APIC_BASE MSR will effectively disable the local APIC on the other logical processor.

Disabling a local APIC on one logical processor prevents the other logical processor from sending or receiving interrupts. Multiprocessor specification compliant BIOSs and multiprocessor operating systems typically leave all local APICs enabled preventing any end-user visible impact from this erratum.

Do not disable the local APICs in a HT Technology enabled processor.


P60

Problem:

Implication:

Workaround:

Status:

Using STPCLK and executing code from very slow memory could lead to a system hang

The system may hang when the following conditions are met :

1. Periodic STPCLK mechanism is enabled via the chipset .

2. HT Technology is enabled.

3. One logical processor is waiting for an event (i.e., hardware interrupt).

4. The other logical processor executes code from very slow memory such that every code fetch is deferred long enough for the STPCLK to be reasserted.

If this erratum occurs, the processor will go into and out of the sleep state without making forward progress, since the logical processor will not be able to service any pending event. This erratum has not been observed in any commercial platform running commercial software.

None at this time.


P61

Problem:

Implication:

Workaround:

Status:

Simultaneous cache line eviction from L2 and L3 caches may result in the write back of stale data

If a cache line is evicted simultaneously from both the L2 and L3 caches, and the internal bus queues are full, an older L3 eviction may be allowed to remain in an internal queue entry. If in a narrow timing window an external snoop is generated the data from the older eviction may be used to respond to the external snoop.

In the event that this erratum occurs the contents of memory will be incorrect. This may result in application, operating system or system failure.

BIOS may contain a workaround for this erratum.


P62

Problem:

Implication:

The state of the resume flag (RF flag) in a task-state segment (TSS) may be incorrect

After executing a JMP instruction to the next (or other) task through a hardware task switch, it is possible for the state of the RF flag (in the EFLAGS register image) to be incorrect.

The RF flag is normally used for code breakpoint management during debug of an application. It is not typically used during normal program execution. Code breakpoints or single step debug

44

Intel

®

Xeon

®


Errata

Workaround:

Status:

behavior in the presence of hardware task switches, therefore, may be unpredictable as a result of this erratum. This erratum has not been observed in commercially available software.

None at this time.


P63

Problem:

Implication:

Workaround:

Status:

Changes to CR3 register do not fence pending instruction page

When software writes to the CR3 register, it is expected that all previous/outstanding code, data accesses and page walks are completed using the previous value in CR3 register. Due to this erratum, it is possible that a pending instruction page walk is still in progress, resulting in an access

(to the PDE portion of the page table) that may be directed to an incorrect memory address.

The results of the access to the PDE will not be consumed by the processor so the return of incorrect data is benign. However, the system may hang if the access to the PDE does not complete with data (e.g., infinite number of retries

).

It is possible for the BIOS to have a workaround for this erratum

.


P64

Problem:

Implication:

Workaround:

Status:

Simultaneous page-faults at similar page offsets on both logical processors of an Hyper-Threading Technology enabled processor may cause application failure

An incorrect value of CR2 may be presented to one of the logical processors of an HT Technology enabled processor if a page access fault is encountered on one logical processor in the same clock cycle that the other logical processor also encounters a page-fault. Both accesses must cross the same 4 byte aligned offset for this erratum to occur. Only a small percentage of such simultaneous accesses are vulnerable. The vulnerability of the alignment for any given fault is dependent on the state of other circuitry in the processor. Additionally, a third fault from an access that occurs sequentially after one of these simultaneous faults has to be pending at the time of the simultaneous faults. This erratum is caused by a one-cycle hole in the logic that controls the timing by which a logical processor is allowed to access an internal asynchronous fault address register. The end result is that the value of CR2 presented to one logical processor may be corrupted.

The operating system is likely to terminate the application that generated an incorrect value of

CR2.

An operating system or page management software can significantly reduce the already small possibility of encountering this failure by restarting or retrying the faulting instruction and only terminate the application on a subsequent failures of the same instruction. It is possible for BIOS to contain a workaround for this erratum.


P65

Problem:

Implication:

Workaround:

Status:

A 16-bit address wrap resulting from a near branch (jump or call) may cause an incorrect address to be reported to the #GP exception handler

If a 16-bit application executes a branch instruction that causes an address wrap to a target address outside of the code segment, the address of the branch instruction should be provided to the general protection exception handler. It is possible that, as a result of this erratum, that the general protection handler may be called with the address of the branch target.

A 16-bit software environment that is affected by this erratum will see that the address reported by the exception handler points to the target of the branch rather than the address of the branch instruction.

None at this time.


Intel

®

Xeon

®


45

Errata

P66

Problem:

Implication:

Workaround:

Status:

Locks and SMC detection may cause the processor to temporarily hang

The processor may temporarily hang in an HT Technology enabled system if one logical processor executes a synchronization loop that includes one or more bus locks and is waiting for release by the other logical processor. If the releasing logical processor is executing instructions that are within the detection range of the self modifying code (SMC) logic, then the processor may be locked in the synchronization loop until the arrival of an interrupt or other event.

If this erratum occurs in an HT Technology enabled system, the application may temporarily stop making forward progress. Intel has not observed this erratum with any commercially available software.

None at this time.

For the steppings affected, see the Summary of Table of Changes.

P67

Problem:

Implication:

Workaround:

Status:

Incorrect debug exception (#DB) may occur when a data breakpoint is set on a FP instruction

The default Microcode Floating Point Event Handler routine executes a series of loads to obtain data about the FP instruction that is causing the FP event. If a data breakpoint is set on the instruction causing the FP event, the load in the microcode routine will trigger the data breakpoint resulting in a debug exception (#DB).

An incorrect #DB may occur if data breakpoint is placed on an FP instruction. Intel has not observed this erratum with any commercially available software or system.

None at this time.


P68

Problem:

Implication:

Workaround:

Status:

Modified cache line eviction from L2 cache may result in write back of stale data

It is possible for a modified cache line to be evicted from the L2 cache just prior to another update to the same line by software. In rare circumstances, the processor may accrue two bus queue entries that have the same address but have different data. If an external snoop is generated in a narrow timing window, the data from the older eviction may be used to respond to the external snoop.

In the event that this erratum occurs, the contents of memory will be incorrect. This may result in application, operating system, or system failure.



P69

Problem:

Implication:

Workaround:

Status:

xAPIC may not report some illegal vector errors

The local xAPIC has an error status register, which records all errors. The bit 6 (the Receive Illegal

Vector bit) of this register, is set when the local xAPIC detects an illegal vector in a received message. When an illegal vector error is received on the same internal clock that the error status register is being written (due to a previous error), bit 6 does not get set and illegal vector errors are not flagged.

The xAPIC may not report some illegal vectors errors when they occur at approximately the same time as other xAPIC errors. The other xAPIC errors will continue to be reported.

None at this time.


46

Intel

®

Xeon

®


Errata

P70

Problem:

Implication:

Workaround:

Status:

Incorrect duty cycle is chosen when On-Demand Clock Modulation is

enabled in a processor supporting Hyper-Threading Technology

When a processor supporting HT Technology enables On-Demand Clock modulation on both logical processors, the processor is expected to select the lowest duty cycle of the two potentially different values. When one logical processor enters the AUTOHALT state, the duty cycle implemented should be unaffected by the halted logical processor. Due to this erratum, the duty cycle is incorrectly chosen to be the higher duty cycle of both logical processors.

Due to this erratum, higher duty cycle may be chosen when the On-Demand Clock modulation is enabled on both logical processors.

None at this time.


P71

Problem:

Implication:

Workaround:

Status:

Memory aliasing of pages as uncacheable memory type and write back

(WB) may hang the system

When a page is being accessed as either UC or WC and WB, under certain bus and memory timing conditions, the system may loop in a continual sequence of UC fetch, implicit WB, and Request

For Ownership (RFO) retries.

This erratum has not been observed in any commercially available operating system or application.

The aliasing of memory regions, a condition necessary for this erratum to occur, is documented as being unsupported in the IA-32 Intel

®

Architecture Software Developer’s Manual, Volume 3,

Section 10.12.4 However, if this erratum occurs the system may hang.

The pages should not be mapped as either UC or WC and WB at the same time.


P72

Problem:

Implication:

Workaround:

Status:

P73

Problem:

Implication:

Workaround:

Status:

A timing marginality in the Instruction Decoder unit may cause an unpredictable application behavior and/or system hang

A timing marginality may exist in the clocking of the instruction decoder unit which leads to a circuit slowdown in the read path from the Instruction Decode PLA circuit. This timing marginality may not be visible for some period of time.

When this erratum occurs, an incorrect instruction stream may be executed resulting in an unpredictable application behavior and/or system hang.

It is possible for the BIOS to contain a workaround for this erratum. BIOS must load the microcode update during the BIOS POST time prior to memory initialization.


Missing Stop

Grant Acknowledge special bus cycle may cause a system

hang

If a Stop Grant Acknowledge special bus cycle is deferred by the processor for a period of time long enough for the chipset to de-assert and then re-assert STPCLK# signal, a processor supporting

HT Technology may fail to detect the de-assertion and re-assertion of STPCLK# signal. When this occurs, the processor will not issue a Stop Grant Acknowledge special bus cycle in response to the second STPCLK# assertion.

When this erratum occurs in an HT Technology enabled system, it may cause a system hang.



Intel

®

Xeon

®


47

Errata

P74

Problem:

Implication:

Workaround:

Status:

Machine check exceptions may not update Last-Exception Record MSRs

(LERs)

The Last-Exception Record MSRs (LERs) may not get updated when machine check exceptions

(MCE) occur.

When this erratum occurs, the LER may not contain information relating to the MCE. They will contain information relating to the exception prior to the MCE .

None at this time.


P75

Problem:

Implication:

Workaround:

Status:

Stores to page tables may not be visible to page walks for subsequent loads without serializing or invalidating the page table entry

Under rare timing circumstances, a page table load on behalf of a programmatically younger memory access may not get data from a programmatically older store to the page table entry if there is not a fencing operation or page translation invalidate operation between the store and the younger memory access. Refer to the IA-32 Intel

®

Architecture Software Developer’s Manual for the correct way to update page tables. Software that conforms to the IA-32 Intel

®

Architecture

Software Developer’s Manual will operate correctly.

If the guidelines in the IA-32 Intel

®

Architecture Software Developer’s Manual are not followed, stale data may be loaded into the processor's translation lookaside buffer (TLB) and used for memory operations. This erratum has not been observed with any commercially available software.

The guidelines in the IA-32 Intel

®

Architecture Software Developer’s Manual should be followed.


P76

Problem:

Implication:

Workaround:

Status:

A timing marginality in the Arithmetic Logic Unit (ALU) may cause indeterminate behavior

A timing marginality may exist in the clocking of the ALU which leads to a slowdown in the speed of the circuit’s operation. This could lead to incorrect behavior of the ALU.

When this erratum occurs, unpredictable application behavior and/or system hang may occur.



P77

Problem:

Implication:

Workaround:

Status:

With Trap Flag (TF) asserted, FP instruction that triggers an unmasked FP exception may take single step trap before retirement of instruction

If an FP instruction generates an unmasked exception with the EFLAGS.TF=1, it is possible for external events to occur, including a transition to a lower power state. When resuming from the lower power state, it may be possible to take the single step trap before the execution of the original

FP

instruction completes.

A Single Step trap will be taken when not expected.

None at this time.


P78

Problem:

PDE/PTE loads and continuous locked updates to the same cache line may cause a system livelock

In a multiprocessor configuration, if one processor is continuously doing locked updates to a cache line that is being accessed by another processor doing a page table walk, the page table walk may not complete.

48

Intel

®

Xeon

®


Errata

Implication:

Workaround:

Status:

Due to this erratum, the system may livelock until some external event interrupts the locked update. Intel has not observed this erratum with any commercially available software.

None at this time.


P79

Problem:

Branch Trace Store (BTS) and Precise Event Based Sampling (PEBS) may update memory outside the BTS/PREBS buffer

If the BTS/PREBS buffer is defined such that:

•

The difference between BTS/PREBS buffer base and BTS/PREBS absolute maximum is not an integer multiple of the corresponding record sizes.

•

BTS/PREBS absolute maximum is less than a record size from the end of the virtual address space.

•

The record that would cross BTS/PREBS absolute maximum will also continue past the end of the virtual address space.

Implication:

Workaround:

Status:

A BTS/PREBS record can be written that will wrap at the 4G boundary (IA-32) or 2^64 boundary

(Intel

®

Extended Memory 64 Technology (Intel

®

EM64T) mode), and write memory outside of the

BTS/PREBS buffer.

Software that uses BTS/PREBS near the 4G boundary (IA-32) or 2^64 boundary (Intel EM64T mode), and defines the buffer such that it does not hold an integer multiple of records can update memory outside the BTS/PREBS buffer.

Define BTS/PREBS buffer such that BTS/PREBS absolute maximum minus BTS/PREBS buffer base is integer multiple of the corresponding record sizes as recommended in the IA-32 Intel

®

Architecture Software Developer’s Manual, Volume 3.


P80

Problem:

Implication:

Memory Ordering Failure may occur with Snoop Filtering Third-Party

Agents after Issuing and completing a BWIL (Bus Write Invalidate Line) or

BLW (Bus Locked Write) transaction

Under limited circumstances, the processors may, after issuing and completing a BWIL or BLW transaction, retain data from the addressed cache line in shared state even though the specification requires complete invalidation. This data retention may also occur when a BWIL transaction’s self-snooping yields HITM snoop results.

A system may suffer memory ordering failures if its central agent incorporates coherence sequencing which depends on full self-invalidation of the cache line associated with (1) BWIL and

BLW transactions, or (2) all HITM snoop results without regard to the transaction type and snoop results’ source.

Workaround:

1. The central agent can issue a bus cycle that causes a cache line to be invalidated (Bus Read

Invalidate Line (BRIL) or BWIL transaction) in response to a processor-generated BWIL (or

BLW) transaction to insure complete invalidation of the associated cache line. If there are no intervening processor-originated transactions to that cache line, the central agent’s invalidating snoop will get a clean snoop result.

Or

2. Snoop filtering central agents can: a. Not use processor-originated BWIL or BLW transactions to update their snoop filter information, or

Intel

®

Xeon

®


49

Errata

Status:

b. Update the associated cache line state information to shared state on the originating bus

(rather than invalid state) in reaction to a BWIL or BLW.


P81

Problem:

Implication:

Workaround:

Status:

Control Register 2 (CR2) can be updated during a REP MOVS/STOS instruction with

Fast Strings enabled

Under limited circumstances while executing a REP MOVS/STOS string instruction, with fast strings enabled, it is possible for the value in CR2 to be changed as a result of an interim paging event, normally invisible to the user. Any higher priority architectural event that arrives and is handled while the interim paging event is occurring may see the modified value of CR2.

The value in CR2 is correct at the time that an architectural page fault is signaled. Intel has not observed this erratum with any commercially available software.

None identified.


P82

Problem:

Implication:

Workaround:

Status:

Writing the Local Vector Table (LVT) when an Interrupt is Pending May

Cause an Unexpected Interrupt

If a local interrupt is pending when the LVT entry is written, an interrupt may be taken on the new interrupt vector even if the mark bit is set.

An interrupt may immediately be generated with the new vector when a LVT entry is written, even if the new LVT entry has the mask bit set. If there is no Interrupt Service Routing (ISR) set up for that vector the system will GP fault. If the ISR does not do an End of Interrupt (EOI) the bit for the vector will be left set in the in-service register and mask all interrupts at the same or lower priority.

Any vector programmed into an LVT entry must have an ISR associated with it, even if that vector was programmed as masked. This ISR routine must do an EOI to clear any unexpected interrupts that may occur. The ISR associated with the spurious vector does not generate an EOI, therefore the spurious vector should not be used when writing the LVT.


P83

Problem:

Implication:

Workaround:

Status:

The Processor May Report a #TS Instead of a #GP Fault

A jump to a busy TSS (Task-State Segment) may cause a #TS (invalid TSS exception) instead of a

#GP fault (general protection exception).

Operation systems that access a busy TSS may get invalid TSS fault instead of a #GP fault. Intel has not observed this erratum with any commercially available software.

None identified.


50

Intel

®

Xeon

®


Specification Changes


P1

There are no new Specification Changes for this month.

The Specification Changes listed in this section apply to the following documents:

•

Intel

®

Xeon

®

Processor at 1.40 GHz, 1.50 GHz, 1.70 and 2 GHz Datasheet

(Order Number 249665)

•

Intel

®

Xeon

®



•

Intel

®

Xeon

®

Processor with 533 MHz Front Side Bus at 2 GHz to 3.20 GHz Datasheet (Order

Number 252135)

•

Low Voltage Intel

®

Xeon

®

Processor at 1.60 GHz to 2.4 GHz Datasheet


•

Intel

®


®

Architectures Software Developer’s Manual, Volumes 1, 2A, 2B,

3A, and 3B (Order Numbers 253665, 253666, 253667, and 253668, respectively)

All Specification Changes will be incorporated into a future version of the appropriate Intel

®

Xeon

®

processor documentation.

Context ID feature added to processor signature instruction feature

Flags/IA32_MISC_Enable registers

IA32_MISC_ENABLE register, bit 24 status has changed from Reserved to the following definition:

IA32_MISC_ENABLE – Miscellaneous Enables Register, bit # 24

MSR Address:

Default Value:

01A0h Accessed as a Qword

High Dword XXXX XXXXh

Access:

Type:

Low Dword XXXX XXXX XXXX XXXX XXXX XX00 X0X0 0001b

Read/Write

Shared

IA32_MISC_ENABLE is a 64-bit register accessed only when referenced as a Qword through a

RDMSR or WRMSR instruction.

Bit 24 of the IA32_MISC_ENABLE status has changed from Reserved to the following:

24

Bit Descriptions

L1 Data Cache Context Mode (R/W). When set to a ‘1’ this bit places the L1 Data Cache into shared mode. When set to a ‘0’ (default) this bit places the L1 Data Cache into adaptive mode.

When this bit is set to a ‘0’, adaptive mode, the Page Directory Base Register contained in CR3 must be identical across all logical processors.

Note: If the Context ID feature flag, ECX[10], is not set to a ‘1’ after executing Processor SIgnature

Instruction with EAX = 1, then this feature is not supported and BIOS must not alter the contents of this bit location.

In the Processor Signature Instruction function 1 feature information, bit 10 of ECX register

(ECX[10]) has been assigned as a flag to identify “Context ID feature”. The status has changed from Reserved to the following:

Intel

®

Xeon

®


51


10

ECX [Bits] Descriptions of Feature Flag Value

Context ID. A value of ‘1’ indicates the L1 data cache mode can be set to either adaptive mode or shared mode. A value of ‘0’ this feature is not supported. See definition of the

IA32_MISC_ENABLE MSR Bit 24 (L1 Data Cache Context Mode) for more details.

52

Intel

®

Xeon

®


Specification Clarifications


P1

P2

The Specification Clarifications listed in this section apply to the following documents:

•

Intel

®

Xeon

®

Processor at 1.40 GHz, 1.50 GHz, 1.70 and 2 GHz Datasheet


•

Intel

®

Xeon

®



•

Intel

®

Xeon

®


Number 252135)

•


®

Xeon

®



•

Intel

®


®



All Specification Clarifications will be incorporated into a future version of the appropriate Intel

Xeon processor documentation.

Maximum ITCC specification correction

The maximum I

TCC specification has been corrected. This correction will be shown in the Voltage and Current Specification Table # 6 and note # 7 in the Intel

®

Xeon

®

Cache at 1.80 GHz to 2.80 GHz Datasheet, Low Voltage Intel

®

Xeon

Processor with 512 KB L2

®

Processor at 1.60 GHz, 2.0

GHz, and 2.4 GHz Datasheet and, Intel

®

Xeon

®

Processor with 533 MHz Front Side Bus at 2 GHz to 3.20 GHz Datasheet.

Current table and note:

Symbol

I

TCC

Parameter

ICC TCC Active

Core Freq

All Freq

Min Typ Max

X

VID Unit

A

Notes

XX

Note 7: The maximum instantaneous current the processor will draw while the thermal control circuit is active as indicated by the assertion of PROCHOT#.

It should state:

Symbol

I

TCC

Parameter

ICC TCC Active

Core Freq

All Freq

Min Typ Max

ICC

VID Unit

A

Notes

XX

Note 7: The maximum instantaneous current that the processor will draw while the thermal control circuit is active as indicated by the assertion of PROCHOT# is the same as the maximum I CC for the processor.

Average Icc does drop when PROCHOT# is active, as long as the Thermal Monitor is active.

Specification Clarification with respect to Time-Stamp Counter

In the “Debugging and Performance Monitoring” chapter (Section 15.8, Section 15.10.9 and

Section 15.10.9.3) of the IA-32 Intel

®

Architecture Software Developer’s Manual, Volume 3:

System Programming Guide, the Time-Stamp Counter definition has been updated to include support for the future processors. This change will be incorporated in the next revision of the IA-32

Intel

®

Architecture Software Developer’s Manual.

Intel

®

Xeon

®


53


15.8

Time-Stamp Counter

The IA-32 architecture (beginning with the Pentium

®

processor) defines a time-stamp counter mechanism that can be used to monitor and identify the relative time occurrence of processor events. The counter’s architecture includes the following components:

•

TSC flag — A feature bit that indicates the availability of the time-stamp counter. The counter is available in an IA-32 processor implementation if the function CPUID.1:EDX.TSC[bit 4] =

1.

•

IA32_TIME_STAMP_COUNTER MSR (called TSC MSR in P6 family and Pentium processors) — The MSR used as the counter.

•

RDTSC instruction — An instruction used to read the time-stamp counter.

•

TSD flag — A control register flag is used to enable or disable the time-stamp counter

(enabled if CR4.TSD[bit 2] = 1).

The time-stamp counter (as implemented in the P6 family, Pentium, Pentium M, Pentium 4, and

Intel

®

Xeon

®

processors) is a 64-bit counter that is set to 0 following a RESET of the processor.

Following a RESET, the counter will increment even when the processor is halted by the HLT instruction or the external STPCLK# pin. Note that the assertion of the external DPSLP# pin may cause the time-stamp counter to stop.

Members of the processor families increment the time-stamp counter differently:

•

For Pentium M processors (family [06H], models [09H, 0DH]); for Pentium 4 processors,

Intel Xeon processors (family [0FH], models [00H, 01H, or 02H]); and for P6 family processors: the time-stamp counter increments with every internal processor clock cycle. The internal processor clock cycle is determined by the current core-clock to bus-clock ratio. Intel

SpeedStep

®

technology transitions may also impact the processor clock.

•

For Pentium 4 processors, Intel Xeon processors (family [0FH], models [03H and higher]): the time-stamp counter increments at a constant rate. That rate may be set by the maximum core-clock to bus-clock ratio of the processor or may be set by the frequency at which the processor is booted. The specific processor configuration determines the behavior. Constant

TSC behavior ensures that the duration of each clock tick is uniform and supports the use of the TSC as a wall clock timer even if the processor core changes frequency. This is the architectural behavior moving forward.

Note:

To determine average processor clock frequency, Intel recommends the use of Performance

Monitoring logic to count processor core clocks over the period of time for which the average is required. See Section 15.10.9 and Appendix A in this manual for more information.

The RDTSC instruction reads the time-stamp counter and is guaranteed to return a monotonically increasing unique value whenever executed, except for a 64-bit counter wraparound. Intel guarantees that the time-stamp counter will not wraparound within 10 years after being reset. The period for counter wrap is longer for Pentium 4, Intel Xeon, P6 family, and Pentium processors.

Normally, the RDTSC instruction can be executed by programs and procedures running at any privilege level and in virtual-8086 mode. The TSD flag allows use of this instruction to be restricted to programs and procedures running at privilege level 0. A secure operating system would set the TSD flag during system initialization to disable user access to the time-stamp counter. An operating system that disables user access to the time-stamp counter should emulate the instruction through a user-accessible programming interface.

The RDTSC instruction is not serializing or ordered with other instructions. It does not necessarily wait until all previous instructions have been executed before reading the counter. Similarly, subsequent instructions may begin execution before the RDTSC instruction operation is performed.

54

Intel

®

Xeon

®



The RDMSR and WRMSR instructions read and write the time-stamp counter, treating the time-stamp counter as an ordinary MSR (address 10H). In the Pentium 4, Intel Xeon, and P6 family processors, all 64-bits of the time-stamp counter are read using RDMSR (just as with RDTSC).

When WRMSR is used to write the time-stamp counter on processors before family [0FH], models

[03H, 04H]: only the low order 32-bits of the time-stamp counter can be written (the high-order 32 bits are cleared to 0). For family [0FH], models [03H, 04H]: all 64 bits are writeable.

15.10.9

Counting Clocks

The count of cycles, also known as clockticks, forms a the basis for measuring how long a program takes to execute. Clockticks are also used as part of efficiency ratios like cycles per instruction

(CPI). Processor clocks may stop ticking under circumstances like the following:

•

The processor is halted when there is nothing for the CPU to do. For example, the processor may halt to save power while the computer is servicing an I/O request. When Hyper-Threading

Technology is enabled, both logical processors must be halted for performance-monitoring counters to be powered down.

•

The processor is asleep as a result of being halted or because of a power-management scheme.

There are different levels of sleep. In the some deep sleep levels, the time-stamp counter stops counting.

There are three ways to count processor clock cycles to monitor performance. These are:

•

Non-halted clockticks — Measures clock cycles in which the specified logical processor is not halted and is not in any power-saving state. When Hyper-Threading Technology is enabled, this these ticks can be measured on a per-logical-processor basis.

•

Non-sleep clockticks — Measures clock cycles in which the specified physical processor is not in a sleep mode or in a power-saving state. These ticks cannot be measured on a logical-processor basis.

•

Time-stamp counter — Some processor models permit clock cycles to be measured when the physical processor is not in deep sleep (by using the time-stamp counter and the RDTSC instruction). Note that such ticks cannot be measured on a per-logical-processor basis. See

Section 10.8 for detail on processor capabilities.

The first two methods use performance counters and can be set up to cause an interrupt upon overflow (for sampling). They may also be useful where it is easier for a tool to read a performance counter than to use a time stamp counter (the timestamp counter is accessed using the RDTSC instruction).

For applications with a significant amount of I/O, there are two ratios of interest:

•

Non-halted CPI — Non-halted clockticks/instructions retired measures the CPI for phases where the CPU was being used. This ratio can be measured on a logical-processor basis when

Hyper-Threading Technology is enabled.

•

Nominal CPI — Time-stamp counter ticks/instructions retired measures the CPI over the duration of a program, including those periods when the machine halts while waiting for I/O.

15.10.9.3

Incrementing the Time-Stamp Counter

The time-stamp counter increments when the clock signal on the system bus is active and when the sleep pin is not asserted. The counter value can be read with the RDTSC instruction.

The time-stamp counter and the non-sleep clockticks count may not agree in all cases and for all processors. See Section 10.8 for more information on counter operation.

Intel

®

Xeon

®


55

Documentation Changes

Documentation Changes

There are no new Documentation Changes for this month.

Note: Documentation changes for Intel

®


®

Architectures Software Developer’s

Manual, volumes 1, 2A, 2B, 3A, and 3B will be posted in the separate document IA-32 Intel®

Architecture and Intel® Extended Memory 64 Technology Software Developer’s Manual

Documentation Changes. Follow the link below to become familiar with this file.

http://developer.intel.com/design/pentium4/specupdt/252046.htm

The Documentation Changes listed in this section apply to the following documents:

•

Intel

®

Xeon

®

Processor at 1.40 GHz, 1.50 GHz, 1.70 and 2 GHz Datasheet


•

Intel

®

Xeon

®



•

Intel

®

Xeon

®


Number 252135)

•


®

Xeon

®



•

Intel

®


®



All Documentation Changes will be incorporated into a future version of the appropriate Intel

Xeon processor documentation.

56

Intel

®

Xeon

®


Intel SL6NQ - Xeon 2.4 GHz/533MHz/512 KB CPU Processor 2.4GHz Specification

Intel

®

Xeon

®

Processor

Specification Update

January 2007

Contents

Revision History

Preface

Affected/Related Documents

Nomenclature

Identification Information

Intel

Xeon

Processor Markings, 256-KB Cache (603-pin

Interposer INT-mPGA Package)

Intel

Xeon

Processor, 512-KB Cache, 400 and 533 MHz

FSB Markings, (603-pin Interposer INT-mPGA Package and

604-pin Fc-mPGA2 Package)

Bottom View

Mixed Steppings in DP Systems

Summary Table of Changes

Codes Used in Summary Table

Errata (Sheet 1 of 4)

Errata (Sheet 2 of 4)

Errata (Sheet 3 of 4)

Errata (Sheet 4 of 4)

Specification Changes

Specification Clarifications

Documentation Changes

Errata

P1

UC Code in same line as write back (WB) data may lead to data corruption

P2

P3

Transaction is not retried after BINIT#

Invalid opcode 0FFFh requires a ModRM byte

P4

When in no-fill mode (CR0.CD=1) the memory type of large (PSE-4M and

PAE-2M) pages are wrongly forced to uncacheable

P5

Processor may hang due to speculative page walks to nonexistent system memory

P6

Writing a performance counter may result in an incorrect counter value

P7

Performance counter may contain incorrect value after being stopped

P8

REP MOV instruction with overlapping source and destination may result in data corruption

P9

Memory type of the load lock different from its corresponding store unlock

P10

Machine check architecture error reporting and recovery may not work as expected

P11

Debug mechanisms may not function as expected

P12

Processor may live-lock if PDEs or PTEs are in UC space

P13

Thermal status log bit may not be set when the thermal control circuit is active

P14

Processor may timeout waiting for a device to respond after 0.67 seconds

P15

Cascading of performance counters does not work correctly when forced overflow is enabled

P16

EMON event counting of x87 loads may not work as expected

P17

Simultaneous code breakpoint and uncorrectable error results in processor hang

P18

Software controlled clock modulation using a 12.5% or 25% duty cycle may cause the processor to hang

P19

RFO with ECC error may result in incorrect data

P20

Speculative page-fault may cause livelock

P21

PAT index MSB may be calculated incorrectly

P22

System bus interrupt messages without data and which receive a Hard failure response may hang the processor