Mobile Intel® 915GM/PM/GME/GMS and 910GML/GMLE Express Chipset Specification Update April 2007 Document Number: 307167-012 Revision History NFORMATION IN THIS DOCUMENT IS PROVIDED IN CONNECTION WITH INTEL® PRODUCTS. NO LICENSE, EXPRESS OR IMPLIED, BY ESTOPPEL OR OTHERWISE, TO ANY INTELLECTUAL PROPERTY RIGHTS IS GRANTED BY THIS DOCUMENT. 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. The information here is subject to change without notice. Do not finalize a design with this information. The products described in this document may contain design defects or errors known as errata which may cause the product to deviate from published specifications. Current characterized errata are available on request. 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 order 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 Web Site. This document contains information on products in the design phase of development. Intel processor numbers are not a measure of performance. Processor numbers differentiate features within each processor family, not across different processor families. See http://www.intel.com/products/processor_number for details. Code Names are only for use by Intel to identify products, platforms, programs, services, etc. (“products”) in development by Intel that have not been made commercially available to the public, i.e., announced, launched or shipped. They are never to be used as “commercial” names for products. Also, they are not intended to function as trademarks. BunnyPeople, Celeron, Celeron Inside, Centrino, Centrino logo, Core Inside, FlashFile, i960, InstantIP, Intel, Intel logo, Intel386, Intel486, Intel740, IntelDX2, IntelDX4, IntelSX2, Intel Core, Intel Inside, Intel Inside logo, Intel. Leap ahead., Intel. Leap ahead. logo, Intel NetBurst, Intel NetMerge, Intel NetStructure, Intel SingleDriver, Intel SpeedStep, Intel StrataFlash, Intel Viiv, Intel vPro, Intel XScale, Itanium, Itanium Inside, MCS, MMX, Oplus, OverDrive, PDCharm, Pentium, Pentium Inside, skoool, Sound Mark, The Journey Inside, VTune, Xeon, and Xeon Inside are trademarks of Intel Corporation in the U.S. and other countries. *Other names and brands may be claimed as the property of others. Copyright © 2005–2007, Intel Corporation. All rights reserved. This device is protected by U.S. patent numbers 5,315,448 and 6,516,132, and other intellectual property rights. The use of Macrovision's copy protection technology in the device must be authorized by Macrovision and is intended for home and other limited pay-per-view uses only, unless otherwise authorized in writing by Macrovision. Devices incorporating Macrovision’s copy protection technology can only be sold or distributed to companies appearing on Macrovision’s list of “Authorized Buyers” at: www.macrovision.com. Reverse engineering or disassembly is prohibited. 2 Specification Update Revision History Contents Revision History ...................................................................................................................4 Preface ...............................................................................................................................5 Summary Tables of Changes ..................................................................................................6 Identification Information ....................................................................................................13 Errata ............................................................................................................................... 16 Specification Changes .........................................................................................................27 Specification Clarifications.................................................................................................... 30 Documentation Changes ......................................................................................................31 Specification Update 3 Revision History Revision History Revision Description -001 Initial public release -002 • Added Errata 25 – 27 Date March 2005 May 2005 • Added Specification Changes 2 – 5 • Added Specification Clarification 1 • Added Documentation Change 5 -003 June 2005 • Added Erratum 28 • Added Erratum and Component Marking Information for C1 Stepping -004 • Added Erratum 29 July 2005 -005 • Added Erratum 30 August 2005 • Updated Erratum 7 & 25 • Updated Specification Clarification 1 -006 • Spec Clarification: DVMT 3.0 support configurations September 2005 • Doc Change : LBKLT_CRTL signal description update -007 • Corrected S-Spec numbers for 915GMS & 910GML PB-free parts (SSpec numbers were swapped) -008 • Added Erratum 31 -009 • Updated Erratum 29 status September 2005 October 2005 November 2005 • Added C-2 stepping information -010 January 2006 • Added Erratum 32 • Doc Change : register bit definition changed for C0DRC0 • Doc Change : register bit definition changed for C0DRC1 -011 • Added Erratum 33 October 2006 -012 • Added 915GME SKU to most references of 915GM April 2007 • Added 910GMLE SKU to most references of 910GML § 4 Specification Update Preface Preface This document is an update to the specifications contained in the documents listed in the following Affected 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 Documents Document Title Document Location/Number Mobile Intel® 915 and 910 Express Chipset Family of Products Datasheet 305264 Nomenclature Errata are design defects or errors. Errata may cause the behavior of the Intel® 915GM/PM/GMS/GME and 910GML/GMLE Express Chipset GMCH 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. Note: Errata remain in the specification update throughout the product’s lifecycle, or until a particular stepping is no longer commercially available. Under these circumstances, errata removed from the specification update are archived and available upon request. Specification changes, specification clarifications and documentation changes are removed from the specification update when the appropriate changes are made to the appropriate product specification or user documentation (datasheets, manuals, etc.). § Specification Update 5 Summary Tables of Changes Summary Tables of Changes The following table indicates the Errata, Specification Changes, Specification Clarifications or Documentation Changes which apply to the listed Intel 915GM/PM/GMS/GME and 910GML/GMLE Express Chipset GMCH 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. Note: B1 and C1 stepping parts will not be produced for the 915GME and the 910GMLE products. This table uses the following notations: Codes Used in Summary Table Stepping X: Erratum, Specification Change or Clarification that applies to this stepping. (No mark) or (Blank Box): This erratum is fixed in listed stepping or specification change does not apply to listed stepping. Status Doc: Document change or update that will be implemented. Plan Fix: This erratum may be fixed in a future stepping of the product. Fixed: This erratum has been previously fixed. No Fix: There are no plans to fix this erratum. Shaded: This item is either new or modified from the previous version of the document. Row 6 Specification Update Summary Tables of Changes Steppings No 915PM 915GM/GME 915GMS C1 C2 B1 C1 Plans ERRATA C0 C1 C2 B1 C1 C2 1 X X X X X X No Fix Reported L0s Exit Latency Is Not Updated When PCI Express* Is Not Operating in Common Clock Mode 2 X X X X X X No Fix GMCH Will Not Identify Back to Back Malformed Packets 3 X X X X X X No Fix The GMCH Is Limited to Reporting Poisoned TLPs through Standard PCI Error Status Reporting Structures 4 X X X X X X No Fix Incorrect PCI Express Lane Transition after Receiving Several TS1 Packets 5 X X X X X X X X X X X No Fix DMI Link Egress Port Address Is Not Programmable 6 X X X X X X X X X X X No Fix DDR2 OCD Nonfunctional 7 X X X X X X No Fix PCI Express Graphics Initiated Snooped Reads to Memory That Are Fast Dispatched Could Result in Incorrect Data Being Returned 8 X X X X X X No Fix PCI Express Common Mode Voltage Noise Immediately Following Receiver Detect sequence 9 X X X X X X No Fix GMCH Does Not Ignore a PCI Express Null Packet Specification Update C0 910GML/GMLE C2 7 Summary Tables of Changes Steppings No 915PM 915GM/GME B1 C1 C2 10 X X X X X X No Fix Data Payload Byte Count Supplied during an unsupported Upstream Configuration Read Is Not 4 Bytes 11 X X X X X X No Fix PCI Express Replay Timer Register Default Setting Is Incorrect 12 X X X X X X No Fix E_SMERR Bit Set Incorrectly Fixed PCI Express Lane 3 Bit Errors Occur on a Small Percentage of GMCH B-1 Units on certain Boards. PCI Express x16 Interface Only Supported on C0 or Later Stepping No Fix Pixel Discoloration Seen When Intel® Dual Frequency Graphics Technology (Intel® DFGT) Is Enabled Fixed System Hang with DDR-333 Memory When Intel® Rapid Memory Power Management Is Enabled during C2/C3/C4 Fixed Display becomes a solid Color with Intel® S2DDT Enabled (Only supported on 915GM, 915GMS & 915GME) No Fix PCI Express Scrambling X 14 X X 16 17 X X X X X X X X X X X X X X X C2 X X B1 X X X C1 ERRATA C2 X C1 Plans C1 13 C0 910GML/GMLE C0 15 8 915GMS X X C2 X X Specification Update Summary Tables of Changes Steppings No 915PM C0 C1 915GM/GME C2 18 B1 C1 C2 X 915GMS C0 C1 910GML/GMLE C2 X B1 C1 Plans ERRATA C2 X Fixed Macrovision Failure for 480p / 576p Modes 19 X X X X X X X X X X X X No Fix Thermometer Read Register is NonFunctional with Negative Calibration Offset 20 X X X X X X X X X X X X No Fix Potential Electrical Spec Violation If a SO-DIMM Is Only Populated in Channel B X X X X X X X X X No Fix LVDS Panel Power Down Timing Violation during System Reset X X X No Fix PCI Express SKP/InitFCx Contention Fixed SMVREF High Current Draw 21 22 X 23 X X X X 24 X X X X X X X X X X No Fix Mobile Intel 915GM/GMS/GME and 910GML/GMLE Express Chipset Graphics Clock Crossing Issue X X X X X X NoFix Mobile Intel 915GM/PM/GMS/GM E and 910GML/GMLE Express Chipset Memory Refresh Queue NoFix Packet Dropped When Replay Timer Expires and Replay Is in Progress NoFix LOCK to non-DRAM Memory Flag (Register Dev 0, Fun 0, Offset C8, Bit 9) Is Getting Asserted 25 X X X X X X 26 X X X X X X 27 X X X X X X Specification Update X X X X X X X 9 Summary Tables of Changes Steppings No 915PM C0 C1 915GM/GME C2 915GMS 910GML/GMLE B1 C1 C2 C0 C1 C2 B1 C1 C2 28 X X X X X X X X X 29 X X* X X* X X* Plans ERRATA NoFix Mobile Intel 915GM/GMS/GME and 910GML/GMLE Express Chipset LVDS – CRT Switching Fixed Mobile Intel 915GM/GMS/GME and 910GML/GMLE Express Chipset LVDS Vos *C1 units with FPO date code of WW49 or later are not affected by the erratum. 30 X X X X X X NoFix Mobile Intel 915PM/GM/GME Express Chipset x1 False Detect 31 X X X X X X NoFix Mobile Intel 915PM/GM/GME Express Chipset PCI Express Ztx_diff_dc Impedance Violation 32 X X X X X X X X X X X X No Fix Mobile Intel 915PM/GM/GMS/GM E and 910GML/GMLE Express Chipset DRAM Clock to CKE Power-up Timing X X X X X X X X X X No Fix Mobile Intel 915PM/GM/GMS/GM E and 910GML/GMLE Express Chipset SMRAM D_CLS Bit 33 X X NOTE: Number 10 B1 and C1 stepping parts will not be produced for the 915GME and the 910GMLE products. SPECIFICATION CHANGES 1 Graphics Render and Display Clock Frequency Combination Support 2 Intel 915GM Express Chipset Host/Memory/Graphics Clock Support Update 3 Intel 915GM Express Chipset Host/Memory/Graphics Clock Support Table 11-1 Update Specification Update Summary Tables of Changes Number SPECIFICATION CHANGES 4 Mobile Intel 915GM/GMS/GME and 910GML/GMLE Express Chipset Display and Render Clock Support Updates 5 Mobile Intel 910GML/GMLE Express Chipset Memory Channel Organization Support Update Number SPECIFICATION CLARIFICATIONS 1 Slow Transfers for LPC Devices 2 DVMT 3.0 Support Configurations Specification Update 11 Summary Tables of Changes Number DOCUMENTATION CHANGES 1 DRAM Enhanced Addressing Update 2 Section 5.2.15 C0DRC1 – Channel 0 DRAM Controller Mode 1 3 Section 5.2.31 DCC – DRAM Channel Control 4 Section 2.2.33 GCFGC – Graphics Clock Frequency and Gating Control 5 Section 5.2.12 C0DRT1 – Channel 0 DRAM Timing Register 1 6 Section 2.5.3 – LBKLT_CRTL Signal Description Update 7 Section 17.2.14 C0DRC0 – Channel 0 DRAM Controller Mode 0 8 Section 17.2.15 C0DRC1 – Channel 0 DRAM Controller Mode 1 § 12 Specification Update Identification Information Identification Information Component Identification via Programming Interface The Intel 915PM/GM/GMS/GME and 910GML/GMLE Express Chipset family may be identified by the following register contents: 1 2 Stepping Vendor ID Device ID Revision Number3 B1 8086h 2590h 03 C0 8086h 2590h 03 C1 and C2 8086h 2590h 04 NOTES: 1. The Vendor ID corresponds to bits 15:0 of the Vendor ID Register located at offset 00– 01h in the PCI function 0 configuration spaces. 2. The Device ID corresponds to bits 15:0 of the Device ID Register located at offset 02– 03h in the PCI function 0 configuration spaces. 3. The Revision Number corresponds to bits 7:0 of the Revision ID Register located at offset 08h in the PCI function 0 configuration spaces. Specification Update 13 Identification Information Component Marking Information The Intel 915PM/GM/GMS/GME and 910GML/GMLE Express Chipset family may be identified by the following component markings: Product 14 MM# or S-Spec# Stepping Notes Mobile Intel 915GM Express Chipset 865926 B1 Mobile Intel 915GM Express Chipset 865971 B1 Mobile Intel 910GML Express Chipset 865930 B1 Mobile Intel 910GML Express Chipset 865923 B1 Mobile Intel 915PM Express Chipset 867706 C0 Mobile Intel 915PM Express Chipset 867713 C0 Mobile Intel 915GMS Express Chipset 867708 C0 Mobile Intel 915GMS Express Chipset 867712 C0 Mobile Intel 915GM Express Chipset SL8G2 C1 and C2 Mobile Intel 915GM Express Chipset SL8G6 C1 and C2 Pb-free Mobile Intel 915GME Express Chipset SLA9K C1 and C2 Pb-free Mobile Intel 915PM Express Chipset SL8G3 C1 and C2 Mobile Intel 915PM Express Chipset SL8G7 C1 and C2 Mobile Intel 915GMS Express Chipset SL8G4 C1 and C2 Mobile Intel 915GMS Express Chipset SL8G9 C1 and C2 Mobile Intel 910GML Express Chipset SL8G5 C1 and C2 Mobile Intel 910GML Express Chipset SL8G8 C1 and C2 Pb-free Mobile Intel 910GMLE Express Chipset SLA9L C1 and C2 Pb-free Pb-free Pb-free Pb-free Pb-free Pb-free Pb-free Specification Update Identification Information Figure 1. Mobile 915PM/GM/GMS/GME & 910GML/GMLE Express Chipset Package Markings Die Grp1 Line1: Product Name Grp1 Line2: S-Spec# Grp1 Line3: FPO# Grp2 Line1: INTEL (M)(C) ‘YY Decoding date code from FPO#: 55494578 Denotes Work Week #49 § Specification Update 15 Errata Errata 1. Reported L0s Exit Latency Is Not Updated When PCI Express* Is Not Operating in Common Clock Mode Problem: When PCI Express is operating with separate reference clocks, L0s exit latency may be greater than the setting in the L0s Exit Latency register. Implication: If the PCI Express link is operating in non-common clock mode, the actual L0s exit latency may be longer than advertised. In this situation the link will likely enter the Recovery state before transitioning into the normal L0 state. Workaround: System BIOS can program the appropriate Exit Latency and advertised N_FTS value if it detects that the downstream device is not using the common reference clock (indicated in the Slot Clock Configuration bit 12 of the device’s Link Status register). Status: For the steppings affected, see the Summary Tables of Changes. 2. GMCH Will Not Identify Back-to-Back Malformed Packets Problem: If the GMCH receives two back-to-back malformed packets, the second malformed packet is not trapped or logged. Implication: The GMCH will not log or identify the second malformed packet. However, the 1st malformed TLP is logged, and is considered a Fatal Error. Link behavior is not guaranteed at that point whether a 2nd malformed TLP is detected or not. Workaround: None Status: For the steppings affected, see the Summary Tables of Changes. 3. The GMCH Is Limited to Reporting Poisoned TLPs through Standard PCI Error Status Reporting Structures Problem: The GMCH does not set the Non-Fatal Error Detected status bit, in the PCI Express Device Status register when a poisoned TLP is received. Implication: Future OS’s (that comprehend PCI Express error reporting) will not be notified via standard PCI Express mechanisms when a poisoned TLP is received. Workaround: Standard PCI error status reporting must be used for Poisoned TLP reporting. The reception of Poisoned TLP is reported by hardware setting the Detected Parity Error bit in Device 1, secondary status register, and if so enabled by additionally setting the Master Data Parity Error bit in the same register. Status: 16 For the steppings affected, see the Summary Tables of Changes. Specification Update Errata 4. Incorrect PCI Express Lane Transition after Receiving Several TS1 Packets Problem: If the GMCH receives several TS1 packets with Link and Lane numbers set to PAD, after 4 µs it will time out and transition into configuration state instead of going directly to the Detect state as it should. However, the link will still transition to the Detect state after timing out of Configuration. Implication: The GMCH will experience longer latency when transitioning to Detect state. Workaround: None Status: For the steppings affected, see the Summary Tables of Changes. 5. DMI Link Egress Port Address Is Not Programmable Problem: The PCI SIG approved ECR 04 to allow future system software (e.g., operating system) to discover the link structure of the Root Complex. One of the registers in the GMCH that “points” from the DMI port to the ICH cannot be programmed correctly. Implication: There is no impact on platform functionality. ECR’s do not retroactively apply to the current PCI Express* 1.0a Specification and no existing software understands the Root Topology discovery structures. These structures are implemented in the GMCH only to aid future software development. Such software will need to comprehend the incorrect pointer Workaround: None Status: For the steppings affected, see the Summary Tables of Changes. 6. DDR2 OCD Nonfunctional Problem: During BIOS initialization the GMCH will not be able to adjust and set the DDR2-DRAM Device DQ/DQS/DQS# buffer impedance. Implication: Adjustable DQ/DQS/DQS# buffer impedance settings will not be programmed into the DDR2-DRAM Devices, and the DRAM device OCD default output characteristics will be used instead. Workaround: None Status: For the steppings affected, see the Summary Tables of Changes. 7. PCI Express* Graphics Initiated Snooped Reads to Memory That Are Fast Dispatched Could Result in Incorrect Data Being Returned Problem: This has only occurred at 1.05-V Core and 533-MHz FSB. GMCH could provide incorrect data for a PCI Express Graphics initiated memory read. Implication: Could cause a data miscompare or system hang. Workaround: It is possible for system BIOS to contain a workaround. Contact your Intel field representative for more details. Status: For the steppings affected, see the Summary Tables of Changes. Specification Update 17 Errata 8. PCI Express Common Mode Voltage Noise Immediately following Receiver Detect Sequence Problem: The PCI Express Common Mode Voltage is not stable immediately after Receiver Detect Sequence when entering Polling.Active from Detect.Active states. Implication: Common Mode Voltage noise may result in bit errors early in Polling.Active state. May result in additional training time before transitioning on to Polling.Configuration. Performance impact expected to be minimal. Workaround: None Status: For the steppings affected, see the Summary Tables of Changes. 9. GMCH Does Not Ignore a PCI Express Null Packet Problem: If the GMCH receives a PCI Express Null packet, it should drop the packet and not perform sequence number checking or respond with any Ack or Nak DLLP. The issue is that the GMCH still performs sequence number checks for Null packets and may respond with an ACK or NAK depending on the result of the check. Implication: GMCH may send ACK or NAK DLLPs in response to a Null packet if the sequence number is not as expected. This may cause unnecessary retries, but Null packets are expected to be rare on the link. Workaround: None Status: For the steppings affected, see the Summary Tables of Changes. 10. Data Payload Byte Count Supplied during an Unsupported Upstream Configuration Read Is Not 4 Bytes Problem: During configuration reads to unsupported PCI Express configuration space, the byte count for data payload is not 4. Implication: Data payload byte count is 5 and not the expected 4. Workaround: Do not perform unsupported upstream PCI Express configuration cycles. Status: For the steppings affected, see the Summary Tables of Changes. 11. PCI Express Replay Timer Register Default Setting Is Incorrect Problem: The default Replay timer value is too aggressive and does not properly account for L0s exit latency or internal pipeline delays. As a result the GMCH may initiate a replay sooner than expected. Implication: The GMCH may initiate a replay sooner than expected. This should not happen as long as the downstream device complies with the PCI Express Ack Latency requirement. Workaround: It is possible for BIOS to contain a workaround. Contact your Intel Field Representative for more details. Status: 18 For the steppings affected, see the Summary Tables of Changes. Specification Update Errata 12. E_SMERR Bit Set Incorrectly Problem: The E_SMERR bit may be incorrectly set when performing valid accesses to SMM space. Implication: If this bit is used by the SMI handler to determine cache line flushes, unnecessary cache line flushes may occur when in SMM mode. A slight performance impact to the SMI handler may result from unnecessary cache line flushes. Workaround: None Status: For the steppings affected, see the Summary Tables of Changes. 13. PCI Express* Lane 3 Bit Errors Occur on a Small Percentage of GMCH B-1 Units on Certain Boards Problem: On systems with External Graphics, a small percentage of GMCH B-1 units may experience bit-errors on PCI Express lane 3 on certain boards. To date, we have only seen these bit errors on internal test (SV) boards at nominal core voltage (1.05 V). Implication: A significant level of bit errors could lead to link retraining, link down, or hang conditions. Intel is currently trying to understand the variables that affect the bit error rate. All samples shipped outside of Intel have been screened for this issue and Intel has not seen any system hangs with screened parts. PCI Express x16 Graphics not supported on Mobile Intel 915GM/GME B1 stepping. Workaround: None Status: For the steppings affected, see the Summary Tables of Changes. 14. Pixel Discoloration Seen When Intel® Dual Frequency Graphics Technology (DFGT) Is Enabled Problem: Pixel discoloration seen when Intel Dual Frequency Graphics Technology (Intel DFGT) is enabled. Root caused to the internal render clocks getting misaligned during the frequency switching. Implication: Pixel discoloration seen on display. Workaround: Disable the DFGT feature through VBT. Status: For the steppings affected, see the Summary Tables of Changes. 15. System Hang with DDR-333 Memory When Intel® Rapid Memory Power Management Is Enabled during C2/C3/C4 Problem: DDR333 systems may not exit self refresh state when memory self refresh is enabled during C2/C3/C4. Implication: System may hang. Workaround: None. Status: For the steppings affected, see the Summary Tables of Changes. Specification Update 19 Errata 16. Display becomes a Solid Color with Intel® Smart 2D Display Technology (Intel® S2DDT) Enabled Problem: If Memory Self-Refresh during C3 (C3SR) occurs when the Intel Smart 2D Display Technology (Intel S2DDT) recompresses at the start of a frame, the graphics engine may use the uncompressed frame instead of the recompressed frame. The issue may occurs if the system is left idle (>20 minutes) with Intel S2DDT. Implication: Display becomes a solid color or may cause a system hang. Workaround: None Status: For the steppings affected, see the Summary Tables of Changes. 17. PCI Express Scrambling Problem: While entering the Recovery state, 915PM/GM/GME PCI Express discrete graphics port stops scrambling two symbols before the first TS (training sequence). Implication: When these non-scrambled symbols are received by the endpoint, the de-scrambler of the endpoint will observe two symbols of random data. The first symbol of TS1 will reset the endpoint’s de-scrambler so that the endpoint should recognize the TS1 and TS2 ordered-sets being transmitted and move into the Recovery state as planned. There is no system level impact if the endpoint is PCI Express Specification 1.0a compliant in ignoring the random data. Workaround: None Problem: For the steppings affected, see the Summary Tables of Changes. 18. Macrovision Failure for 480p / 576p Modes Problem: GMCH does not encode Macrovision (content protection) properly for 480p and 578p progressive modes and does not meet the Macrovision specification for these modes. 480p and 576p progressive modes will not pass MacroVision certification. Implication: No DVD playback supported/available for these modes, other progressive scan modes are not affected. Workaround: None Status: 20 For the steppings affected, see the Summary Tables of Changes. Specification Update Errata 19. Thermometer Read Register is Non-Functional with Negative Calibration Offset Problem: When initializing the Thermometer counter, a negative offset will cause an overflow condition, causing the counter to stop. The GMCH will not update the Thermometer Read Register (TRR) register with the correct temperature trending values and the TRR will display “FF”. All other thermal sensor features will still work (throttling, trip points and interrupts) since these do not use the thermometer mode. Implication: Thermometer Read Register (TRR) is not functional with negative Thermal Calibration Offset value (TCO). Workaround: Please contact your Intel representative for the latest information. Status: For the steppings affected, see the Summary Tables of Changes. 20. Potential Electrical Spec Violation if an SO-DIMM Is Only Populated in Channel B Problem: When DDR or DDR2 memory devices are present on channel B, but not channel A, memory control logic only looks at channel A and incorrectly performs Rcomp value update. This could cause an electrical spec violation internal to the chip. Implication: System with 915PM/GM/GME, or 910GML/GMLE may hang or experience other anomalous system behavior. No impact on system with 915GMS. Workaround: It is possible for BIOS to contain a workaround. Contact your Intel Field Representative for more details. Status: For the steppings affected, see the Summary Tables of Changes. 21. LVDS Panel Power Down Timing Violation during System Reset Problem: During system reset, there is insufficient time for handshake between ICH and GMCH LVDS logic. As a result, timing from panel backlight enable going low to LVDS data going low (TX) and timing from LVDS data going low to panel VCC enable going low (T3) do not match the programmed values. Panel backlight enable (LBKLT_EN), panel Vcc enable (LVDD_EN) and LVDS data lines go low at the same time. Implication: No system level issues have been observed. Workaround: None Status: For the steppings affected, see the Summary Tables of Changes. Specification Update 21 Errata 22. PCI Express SKP/InitFCx Contention Problem: During GMCH PCI Express initialization, if a SKP is being transmitted immediately before an InitFCx DLLP, then a partial InitFCx may be transmitted. Implication: A slight delay (less than 100 ns) may occur during link initialization. Device may report correctable error. InitFCx will automatically be repeated. Workaround: None Status: For the steppings affected, see the Summary Tables of Changes 23. SMVREF High Current Draw Problem: When SMOCDCOMP[1:0] are connected to ground (either directly or through a pulldown resistor), SMVREF[1:0] may draw significantly more current than specified. If the corresponding voltage supply is not capable of handling the resulting current, the SMVREF[1:0] voltage will droop below specification. Implication: If SMVREF[1:0] droops below specification, the system may hang. Failure observed during S3 only. Workaround: Leave SMOCDCOMP[1:0] as NC (No Connect) for both DDR and DDR2 designs. This limits the maximum current draw on SMVREF[1:0] to 200 µA (i.e. 100uA/pin). In addition, use a voltage supply that can supply at least 200 µA (total) for SMVREF[1:0]. Status: For the steppings affected, see the Summary Tables of Changes. 24. Mobile Intel 915GM/GMS/GME and 910GML/GMLE Express Chipset Graphics Clock Crossing Issue Problem: While attempting to launch or exit an application that uses the overlay engine (e.g. media playback software running a video clip), the system may hang (blue screen) or screen resolution may change to VGA mode. This event may occur when a signal pulse is missed inside the chip due to synchronization of the graphics render and display clocks while graphics overlay mode is being turned off. Implication: Failure is difficult to reproduce because this event only occurs under a very rare but valid timing condition where graphics render and display clocks are aligned while the graphics overlay is being turned off. System will hang or screen resolution may change to VGA mode when launching or exiting an application that uses the overlay engine. Mobile Intel 915GM/GMS/GME and 910GML/GMLE platforms that use render clock at 166 MHz are not affected by this sighting. Workaround: It is possible for BIOS to contain a workaround. Contact your Intel Field Representative for more details. Status: 22 For the steppings affected, see the Summary Tables of Changes. Specification Update Errata 25. Mobile Intel 915PM/GM/GMS/GME and 910GML/GMLE Express Chipset Memory Refresh Queue Problem: When thermal based throttling is enabled, memory refresh commands get queued up inside MCH/GMCH. As a result, MCH/GMCH may not perform memory refresh commands at a frequency needed by memory devices. Thermal based throttling happens only when either using the on-die memory thermal sensors or the external thermal sensor. Implication: When not enough refresh commands are issued to system memory to retain memory content, data loss may occur. System is more susceptible to this issue when memory devices approach temperature limit. Workaround: It is possible for system BIOS to contain a workaround. Contact your Intel field representative for more details. Status: For affected steppings, see the Summary Tables of Changes. 26. Packet Dropped When Replay Timer Expires and Replay Is in Progress Problem: When a packet replay is in progress on the PCI Express Port, and if some but not all of the packets to be replayed are acknowledged and the replay timer expires on the same clock cycle as the replay start of the first unacknowledged packet, the next packet in the replay buffer may be sent with an old sequence number. That packet is seen by receiver side as a duplicate and subsequently dropped. Note: This has only been reproduced in a synthetic test environment. Implication: Anomalous behavior may result if all of the above conditions are met. Workaround: None Status: For the steppings affected, see the Summary Tables of Changes. 27. LOCK to non-DRAM Memory Flag (Register Dev 0, Fun 0, Offset C8, Bit 9) Is Getting Asserted Problem: A CPU lock cycle request is unintentionally being recognized as request to a nonsystem memory destination. Implication: The GMCH may incorrectly flag an error for a valid lock cycle that targets DRAM. A System Error (SERR) may be generated if enabled by System BIOS. Note: The default setting for ERRCMD[9] Bus 0 Device 0 Offset CAh is to disable this reporting. Workaround: Do not enable or change default setting of ERRCMD[9] Bus 0 Device 0 Offset CAh (SERR reporting for Lock cycles to non-DRAM Memory) Status: For the steppings affected, see the Summary Tables of Changes. Specification Update 23 Errata 28. Mobile Intel 915GM/GMS/GME and 910GML/GMLE Express Chipset LVDS – CRT Switching Problem: 915GM/GMS/GME and 910GML/GMLE LVDS interface may hang while switching between LVDS and CRT outputs. Implication: Once LVDS interface hangs a system reset is required. Workaround: Fixed in VBIOS, Please refer to release notes for VBIOS version 1220 or Newer. Status: For the steppings affected, see the Summary Tables of Changes. 29. Mobile Intel 915GM/GMS/GME and 910GML/GMLE Express Chipset LVDS Vos Problem: On some Mobile Intel 915GM/GMS/GME and 910GML/GMLE Express devices, one or more LVDS lanes may exceed Intel’s Vos specification limits. LVDS Interface: Functional Operating Range (VCC= 2.5V ±5%) Symbol Parameter Min Nom Max Unit VOS(Tested) Offset Voltage 0.700 1.25 1.600 V VOS (Spec) Offset Voltage 1.125 1.25 1.375 V NOTES: Table reflects Intel tested Vos specification limits and the Mobile Intel 915GM/GMS/GME & 910GML/GMLE specification for Vos values. Intel may make changes to specifications, release dates and product descriptions from time to time. Implication: LVDS Panels sensitive to Vos range may not detect transmitted LVDS data. Potential visual anomalies on LVDS display panels may occur, although no visual anomalies have been reported due to Vos levels in the above tested range. Workaround: None Status: 24 For the steppings affected, see the Summary Tables of Changes. Specification Update Errata 30. Mobile Intel 915PM/GM/GME Express Chipset x1 False Detect Problem: 915PM/GM/GME may falsely detect the presence of a PCIe endpoint while operating in PCIe Low Power Mode. • False detect may occur for 915PM/GM/GME systems supporting x1 only endpoints. • False detect may occur on PCIe* lanes 1 through 15 for a graphics endpoint when switching from x16 to x1. Implication: Systems may become inoperable when an endpoint is operating in x1 mode. • The scenario occurs because false detection of a PCIe endpoint occurs on one or more lanes resulting in the GMCH’s PCIe interface erroneously entering and looping continuously in Polling.Compliance. • When an endpoint is operating in x1 mode, the scenario may occur at boot or during any subsequent attempt to retrain the link. Workaround: For systems requiring x1 PCIe link operation: • Systems operating in x1 mode only: It is possible for system BIOS to contain a workaround. Contact your Intel field representative for more details. • Systems requiring run-time switching between x16 and x1 operation: A graphics driver workaround has been defined. Please contact your graphic’s controller vendor for driver status. • Use the erratum title when contacting graphic’s controller vendor for driver status. • For questions pertaining to the erratum or workaround please contact your Intel representative. Status: For the steppings affected, see the Summary Tables of Changes. 31. Mobile Intel 915PM/GM/GME Express Chipset PCI Express Ztx_diff_dc Impedance Violation Problem: 915PM/GM violates the PCI Express Ztx_diff_dc max impedance specification. Implication: None • No failures have been observed due to the violation of the PCI Express Ztx_diff_dc impedance specification. • PCI Express Tx_eye signal integrity specifications are still met. Workaround: None Status: For affected steppings, see the Summary Tables of Changes. Specification Update 25 Errata 32. Mobile Intel 915PM/GM/GMS/GME and 910GML/GMLE Express Chipset DRAM Clock to CKE Power-up Timing Problem: During memory power-up and initialization, the timing between DRAM clock stabilization to CKE going high is observed to be minimum of 35 ns against the JEDEC spec of 200 µs. Implication: None. No functional failures have been observed. • Intel has characterized the timing and shared the data with major DRAM suppliers. Intel has determined and major DRAM suppliers agree that DRAM devices need < 35 ns. This erratum should not cause memory-clock functionality or timing related issues. Please refer to latest Intel DRAM spec Addendum for power-up and initialization timing requirements available at http://developer.intel.com/technology/memory/#Specs Workaround: None. Status: For affected steppings, see the Summary Tables of Changes. 33. Mobile Intel 915PM/GM/GMS/GME and 910GML/GMLE Express Chipset SMRAM D_CLS Bit Problem: Data and stack which residing in Extended SMRAM (TSEG/HSEG) is inaccessible if D_CLS bit (Bus 0, Device 0, Function 0, Register 9Dh, Bit 5) is set. Implication: May result in system hang. Workaround: It is possible for system BIOS to contain a workaround. Contact your Intel field representative for more details. Status: No Fix. For steppings affected, see the Summary Tables of Changes. § 26 Specification Update Specification Changes Specification Changes 1. Graphics Core Clock Frequency Update The following section replaces the Mobile Intel® 915 and 910 Express Chipset Family of Products Datasheet, Section 10.14.3. 10.14.3.1 Intel 915GM/GME Express Chipset Host/Memory/Graphics Clock Support Table 10-16a. Intel 915GM/GME Express Chipset Graphics Clock Frequency Support at 1.05 V Host Memory 2D Display clock 3D Render Clock 400 MHz DDR 333 200 166 400 MHz DDR 333 213 200 533 MHz DDR 333 190 166 533 MHz DDR 333 213 190 400 MHz / 533 MHz DDR2 400 / DDR2 533 200 160 400 MHz / 533 MHz DDR2 400 / DDR2 533 213 200 NOTE: Mobile Intel 915GM/ 915GME / 910GML / 910GMLE Express Chipset will remove support for 133 MHz for both graphics render and display clock due to graphics clock crossing issue (Errata #24). Table 10-16b. Intel 915GM/GME Express Chipset Graphics Clock Frequency Support at 1.5 V Host Memory 2D Display clock 3D Render Clock 400 MHz / 533 MHz DDR2 400 / DDR2 533 200 160 400 MHz / 533 MHz DDR2 400 / DDR2 533 213 200 400 MHz / 533 MHz DDR2 400 / DDR2 533 333 333 NOTE: Specification Update Mobile Intel 915GM/GME/ 910GML/GMLE does not support DDR333 if chipset core voltage is designed for 1.5 V. 27 Specification Changes 10.14.3.2 Intel 915GMS Express Chipset Host/Memory/Graphics Clock Support Table 10-17. Intel 915GMS Express Chipset Graphics Clock Frequency Support at 1.05 V Host Memory 2D Display Clock 3D Render Clock 400 MHz DDR2 400 152 133 400 MHz DDR2 400 200 160 10.14.3.3 Intel 910GML/GMLE Express Chipset Host/Memory/Graphics Clock Support Table 10-18. Intel 910GML/GMLE Express Chipset Graphics Clock Frequency Support at 1.05 V Host 2. Memory 2D Display Clock 3D Render Clock 400 MHz DDR 333 200 166 400 MHz DDR2 400 200 160 Intel 915GM/GME Express Chipset Host/Memory/Graphics Clock Support Update Table 10-16 and Table 10-7 incorrectly state support for Display clocks for 400 MHz / DDR333 and 533 MHz/DDR333 frequency combinations. These combinations will result with a Display clock of 222 MHz. See Specification Change Item #1 above and Documentation Change #4 for updated information in grey shaded text. 3. Mobile Intel 915GM/GMS/GME and 910GML/GMLE Express Chipset Display and Render Clock Support Update The following section replaces the Mobile Intel® 915 and 910 Express Chipset Family of Products Datasheet, Graphics Clock Frequency Feature section, on page 22. ⎯ 2D Display core frequency at 190/200 or 213/222 MHz @ Vcc=1.05 V depending on the host/memory configurations ⎯ 3D Render core frequency at 160/166 or 190/200 MHz @ Vcc=1.05 V depending on the host/memory configurations ⎯ 2D Display core frequency at 200, 213 or 333 MHz @ Vcc=1.5 V depending on the host/memory configurations ⎯ 3D Render core frequency at 166, 200 or 333 MHz @ Vcc=1.5 V depending on the host/memory configurations 28 Specification Update Specification Changes The following section replaces the Mobile Intel® 915 and 910 Express Chipset Family of Products Datasheet, Graphics Clock Frequency Feature section, on page 25. ⎯ 2D Display core frequency at 152 or 200 MHz @ Vcc=1.05 V depending on the host/memory configurations ⎯ 3D Render core frequency at 133 or 160 MHz @ Vcc=1.05 V depending on the host/memory configurations The following section replaces the Mobile Intel® 915 and 910 Express Chipset Family of Products Datasheet, Graphics Clock Frequency Feature section, on page 26. ⎯ 2D Display core frequency at 200 MHz @ Vcc=1.05 V depending on the host/memory configurations ⎯ 3D Render core frequency at 160/166 MHz @ Vcc=1.05 V depending on the host/memory configurations 4. Mobile Intel 910GML/GMLE Express Chipset Memory Channel Organization Support Update The following section replaces the Mobile Intel® 915 and 910 Express Chipset Family of Products Datasheet in the Mobile Intel 910GML Express Chipset Feature section, on page 26 and in Section 10.2. ⎯ Memory channel organization support for DDR is: ⎯ Single Channel Mode ⎯ Memory channel organization support for DDR2 is: ⎯ Dual Channel Asymmetric Mode only Note: Dual Channel Symmetric (Interleave) Mode not supported on Mobile Intel 910GML/GMLE Express Chipset The following bullet replaces the Mobile Intel® 915 and 910 Express Chipset Family of Products Datasheet, Section 10.2, System Memory Controller. ⎯ Dual Channel Symmetric (Interleave) Mode not supported for Mobile Intel 910GML/GMLE Express Chipset. § Specification Update 29 Specification Clarifications Specification Clarifications 1. Slow DMA Transfers for LPC Devices 8-bit DMA transfers to some LPC devices, may incur latencies, that result in a slowed down transfer rate. Longer latencies may occur when the platform becomes idle and the GMCH dynamically asserts HCPUSLP# for Processor C2 states. For example, a 4-Mbps FIR device may transmit faster than the resulting 8-bit DMA transfer rate resulting in data retransmission and lower FIR throughput. So as not to impede the transfer rate of certain LPC Devices, the dynamic assertion of HCPUSLP# by GMCH during Processor C2 states can be disabled. Contact your Intel field representative for more details on how to disable HCPUSLP# assertion during C2 states. 2. Dynamic Video Memory Technology (DVMT) 3.0 Support Configurations The following note should be added to Section 1.1 of the Mobile Intel® 915 and 910 Express Chipset Family of Products Datasheet: Please contact your local Intel field representative for additional information regarding DVMT 3.0 support configurations. § 30 Specification Update Documentation Changes Documentation Changes 1. DRAM Enhanced Addressing Update The following section is inserted into the Mobile Intel® 915 and 910 Express Chipset Family of Products Datasheet Chapter 10, Section 10.2.1.3 10.2.1.3 DRAM Address Mapping Enhanced Addressing swaps the most significant bit controlling one of the Sx_BS lines with bit 18 (which normally controls row address bit 2). Without Enhanced Addressing, rank bits are the most significant two bits of the address. With Enhanced Addressing, the rank bits are always bits 19 and 20. Table 10-8 DRAM Device Configurations – Dual Channel Asymmetric Mode w/Enh. Addr. Swap (0) Technology (Mb) 256 256 512 512 512 512 1024 1024 1024 1024 13 13 13 13 13 14 14 14 13 14 Column bits 9 10 10 11 10 10 10 11 10 10 Bank bits 2 2 2 2 2 2 2 2 3 3 Width (b) 16 8 16 8 16 8 16 8 16 8 8192 8192 8192 8192 8192 16384 16384 16384 8192 16384 512 1024 1024 2048 1024 1024 1024 2048 1024 1024 Banks 4 4 4 4 4 4 4 4 8 8 Page Size (KB) 4 8 8 16 8 8 8 16 8 8 Devices per rank 4 8 4 8 4 8 4 8 4 8 Rank Size (MB) 128 256 256 512 256 512 512 1024 512 1024 Depth (M) 16 32 32 64 32 64 64 128 64 128 Addr bits [n:0] 26 27 27 28 27 28 28 29 28 29 Available in DDR yes yes yes yes no no yes yes no no Available in DDRII yes yes no no yes yes no No yes yes Row bits Rows Columns Host Address bit Mem Addr-bit 31 - - - - - - - - - - 30 - - - - - - - r3 - r3 29 - - - r3 - r3 r3 r 13 r3 r 13 Specification Update 31 Documentation Changes 32 28 - r3 r3 r 11 r3 r 13 r 13 r 11 r 11 r 11 27 r3 r 12 r 12 r 12 r 12 r 12 r 12 r 12 r 12 r 12 26 r 10 r 10 r 10 r 10 r 10 r 10 r 10 r 10 r 10 r 10 25 r9 r9 r9 r9 r9 r9 r9 r9 r9 r9 24 r8 r8 r8 r8 r8 r8 r8 r8 r8 r8 23 r7 r7 r7 r7 r7 r7 r7 r7 r7 r7 22 r6 r6 r6 r6 r6 r6 r6 r6 r6 r6 21 r5 r5 r5 r5 r5 r5 r5 r5 b2 b2 20 r4 r4 r4 r4 r4 r4 r4 r4 r4 r4 19 s0 s0 s0 s0 s0 s0 s0 s0 s0 s0 18 b1 b1 b1 b1 b1 b1 b1 b1 b1 b1 17 r1 r1 r1 r1 r1 r1 r1 r1 r1 r1 16 r0 r0 r0 r0 r0 r0 r0 r0 r0 r0 15 r 11 r 11 r 11 r2 r 11 r 11 r 11 r2 r5 r5 14 r 12 r2 r2 b0 r2 r2 r2 b0 r2 r2 13 r2 b0 b0 c 11 b0 b0 b0 c 11 b0 b0 12 b0 c9 c9 c9 c9 c9 c9 c9 c9 c9 11 c8 c8 c8 c8 c8 c8 c8 c8 c8 c8 10 c7 c7 c7 c7 c7 c7 c7 c7 c7 c7 9 c6 c6 c6 c6 c6 c6 c6 c6 c6 c6 8 c5 c5 c5 c5 c5 c5 c5 c5 c5 c5 7 c4 c4 c4 c4 c4 c4 c4 c4 c4 c4 6 c3 c3 c3 c3 c3 c3 c3 c3 c3 c3 5 c2 c2 c2 c2 c2 c2 c2 c2 c2 c2 4 c1 c1 c1 c1 c1 c1 c1 c1 c1 c1 3 c0 c0 c0 c0 c0 c0 c0 c0 c0 c0 Specification Update Documentation Changes Table 10-9 DRAM Device Configurations –Dual Channel Interleaved Mode w/ Enh. Addr. Swap (0) Technology (Mb) Row bits 256 256 512 512 512 512 1024 1024 1024 1024 13 13 13 13 13 14 14 14 13 14 Column bits 9 10 10 11 10 10 10 11 10 10 Bank bits 2 2 2 2 2 2 2 2 3 3 Width (b) 16 8 16 8 16 8 16 8 16 8 8192 8192 8192 8192 8192 16384 16384 16384 8192 16384 512 1024 1024 2048 1024 1024 1024 2048 1024 1024 Banks 4 4 4 4 4 4 4 4 8 8 Page Size (KB) 4 8 8 16 8 8 8 16 8 8 Devices per rank 4 8 4 8 4 8 4 8 4 8 Rank Size (MB) 128 256 256 512 256 512 512 1024 512 1024 Depth (M) 16 32 32 64 32 64 64 128 64 128 Addr bits [n:0] 26 27 27 28 27 28 28 29 28 29 Available in DDR yes yes yes yes no no yes Yes no no Available in DDRII yes yes no no yes yes no No yes yes Rows Columns Host Address bit Mem Addr-bit 31 - - - - - - - r3 - r3 30 - - - r3 - r3 r3 r 13 r3 r 13 29 - r3 r3 r 11 r3 r 13 r 13 r 11 r 11 r 11 28 r3 r 12 r 12 r 12 r 12 r 12 r 12 r 12 r 12 r 12 27 r 10 r 10 r 10 r 10 r 10 r 10 r 10 r 10 r 10 r 10 26 r9 r9 r9 r9 r9 r9 r9 r9 r9 r9 25 r8 r8 r8 r8 r8 r8 r8 r8 r8 r8 24 r7 r7 r7 r7 r7 r7 r7 r7 r7 r7 23 r6 r6 r6 r6 r6 r6 r6 r6 r6 r6 22 r5 r5 r5 r5 r5 r5 r5 r5 b2 b2 21 r4 r4 r4 r4 r4 r4 r4 r4 r4 r4 20 s0 s0 s0 s0 s0 s0 s0 s0 s0 s0 19 b1 b1 b1 b1 b1 b1 b1 b1 b1 b1 18 r1 r1 r1 r1 r1 r1 r1 r1 r1 r1 17 r0 r0 r0 r0 r0 r0 r0 r0 r0 r0 16 r 11 r 11 r 11 r2 r 11 r 11 r 11 r2 r5 r5 15 r 12 r2 r2 b0 r2 r2 r2 b0 r2 r2 Specification Update 33 Documentation Changes 2. 14 r2 b0 b0 c 11 b0 b0 b0 c 11 b0 b0 13 b0 c9 c9 c9 c9 c9 c9 c9 c9 c9 12 c8 c8 c8 c8 c8 c8 c8 c8 c8 c8 11 c7 c7 c7 c7 c7 c7 c7 c7 c7 c7 10 c6 c6 c6 c6 c6 c6 c6 c6 c6 c6 9 c5 c5 c5 c5 c5 c5 c5 c5 c5 c5 8 c4 c4 c4 c4 c4 c4 c4 c4 c4 c4 7 c3 c3 c3 c3 c3 c3 c3 c3 c3 c3 6 h h H h h h h H h h 5 c2 c2 c2 c2 c2 c2 c2 c2 c2 c2 4 c1 c1 c1 c1 c1 c1 c1 c1 c1 c1 3 c0 c0 c0 c0 c0 c0 c0 c0 c0 c0 Section 5.2.15 C0DRC1 – Channel 0 DRAM Controller Mode 1 The following register bits definition replaces the Mobile Intel® 915 and 910 Express Chipset Family of Products Datasheet, Section 5.2.15 5.2.15 C0DRC1––Channel 0 DRAM Controller Mode 1 PCI Device: MCHBAR Address Offset: 124h Default Value: 00000000h Access: RO, R/W Size: 32 bits Bit Access and Default Description Swap Mode Enable: R/W 31 0b R/W 0 = Swap mode disabled. Dram Address Map Follows the standard address map described in the EDS. 1 = Swap mode enabled. This bit can be enabled if all dimm are either single sided or symmetrically populated. Reserved 30 0b Address Swap Mode: R/W 29 00 b 34 0 = Swap Enabled for Bank Selects and Rank Selects. 1 = Swap Enabled for Bank Selects only. Specification Update Documentation Changes Bit Access and Default RO Description Reserved 28:27 0b Swap Address Selection: Single Channel and Asymm-Stacked Dual Channel: R/W 26:24 000 b Mode Rank Select Bank Select 0 Bank Select 1 Bank Select 2 Dimm Select 000 A19 std A18 A21 A20 001 A20 std A19 A22 A21 010 A18 std A17 A20 A19 011 A17 Std A16 A19 A18 Others Rsvd. Rsvd. Rsvd. Rsvd. Recommended Priority: 0, 1, 2 Dual Channel Interleaved Mode Rank Select Bank Select 0 Bank Select 1 Bank Select 2 Dimm Select 000 A20 std A19 A22 A21 001 A21 Std A20 A23 A22 010 A19 std A18 A21 A20 011 A18 std A17 A20 A19 Others Rsvd. Rsvd. Rsvd. Rsvd. Recommended Priority: 2, 3, 0 Specification Update 35 Documentation Changes 3. Section 5.2.31 DCC – DRAM Channel Control The following register bits definition replaces the Mobile Intel® 915 and 910 Express Chipset Family of Products Datasheet; Section 5.2.31 Section 5.2.31 DCC—DRAM Channel Control MMIO Range: MCHBAR Address Offset: 200h Default Value: 00000000h Access: RO, R/W Size: 32 bits This register controls how the DRAM channels work together. It affects how the CxDRB registers are interpreted and allows them to steer transactions to the correct channel. Bit Access and Default 10 R/W 0b Description Channel XOR Randomization Disable (CXRDIS): When enabled, the DRAM Controller will try to spread page accesses evenly among the channels by including more address bits in the choice for which channel holds the requested address. 0: Channel XOR Randomization is enabled. CXRSEL’s bit will be XORed with CHSEL’s bit to pick the channel for a given address. 1: Channel XOR Randomization is disabled 9 R/W 0b Channel XOR Randomization Control Bit (CXRSEL): When Channel XOR Randomization is in use (see CXRDIS), the channel select will be randomized by XORing it with either host address bit 11 or 17. 0: Bit 11 will be XORed with the Channel Select bit 1: Bit 17 will be XORed with the Channel Select bit 36 Specification Update Documentation Changes 4. Section 7.2.33 GCFGC – Graphics Clock Frequency and Gating Control The following register bits definition replaces the Mobile Intel® 915 and 910 Express Chipset Family of Products Datasheet; Section 7.2.33 Section 7.2.33 GCFGC—Graphics Clock Frequency and Gating Control PCI Device 2 Function: 0 Address Offset F0h Default Value 0000h Access: RO, R/W Size: 16 bits Bit Access and Default 15 R/W 0b Description Graphics Display Clock Low Frequency Override (for 915GMS only): 0 = Do not use Graphics cdclk low frequency override 1 = Use Graphics cdclk = 152 MHz low frequency NOTE: This bit is reserved for 915GM, 910GML and 910GMLE 14 R/W Reserved 0b 13 R/W GFX GVL Low Frequency Enable: 0b 0 = Do not Use GFX GVL low frequency target for Render Clock. 1 = Use GFX GVL low frequency target for Render Clock. 12 R/W GFX GVL Low Frequency Target: 0b 0 = 133 MHz (Default Value) 1 = Reserved 11 R/W Gate Core Render Clock (GCRC): 0b 0: Core render clock (crclk) is running 1: Core render clock (crclk) is gated 10 Specification Update R/W Asynchronously Change Core Render Clock (ACCRC): 0b A 0 to 1 transition on this bit will immediately load new pre- and postdivider values for the crclk and crx2clk. Writing 1 to 1, 1 to 0, and 0 to 0 have no effect. 37 Documentation Changes Description Bit Access and Default 9 R/W Gate Core Display Clock (GCRC): 0b 0: Core display clock (cdclk) is running 1: Core display clock (cdclk) is gated 8 7 R/W Asynchronously Change Core Display Clock (ACCDC): 0b A 0 to 1 transition on this bit will immediately load new pre- and postdivider values for the cdclk. Writing 1 to 1, 1 to 0, and 0 to 0 have no effect. R/W Core Display Low Frequency Enable: 0b 0 = Do not Use low frequency target for Display Clock. 1 = Use low frequency target (152 MHz) for Display Clock. (For 915GMS only) NOTE: This bit is reserved for 915GM / 915GME and 910GM / 910GMLE. 6:4 R/W Graphics Core Display Clock Select: 000 b 000 = 190/200 MHz (Intel 915GM / 915GMS / 915GME and Intel 910GML / 910GMLE) 001 = 213/222 MHz (Intel 915GM / 915GME ) 010 = Reserved 011 = Reserved 100 = 333 MHz (Intel 915GM/GME @ 1.5 V only) 101 = Reserved 110 = Reserved 111 = Reserved 3 RO Reserved 0b 2:0 R/W Graphics Core Render Clock Select: 000 b 000 = 160/166 MHz (Intel 915GM / 915GMS / 915GME and Intel 910GML / 910GMLE) 001 = 190/200 MHz (Intel 915GM / 915GME) 010 = Reserved 011 = Reserved 100 = 333 MHz (Intel 915GM/GME @ 1.5 V only) 101 = Reserved Others = Reserved 38 Specification Update Documentation Changes 5. Section 5.2.12 C0DRT1 – Channel 0 DRAM Timing Register 1 The following register bits definition replaces the Mobile Intel® 915 and 910 Express Chipset Family of Products Datasheet; Section 5.2.12 Section 5.2.12 C0DRT1—Channel 0 DRAM Timing Register 1 PCI Device: MCHBAR Address Offset: 114h Default Value: 00006111h Access: RO, R/W Size: 32 bits Bit 31:30 Description Access and Default RO Reserved 00 b Read to Pre-charge (tRTPC): These bits control the number of clocks that are inserted between a read command to a row pre-charge command to the same rank. 29:28 27:24 Specification Update R/W Encoding 00 b 00: BL/2 01: Reserved 10: Reserved 11: Reserved R/W tRP Reserved 0h 39 Documentation Changes Bit Description Access and Default Activate to Precharge Delay (tRAS): This bit controls the number of DRAM clocks for tRAS. Minimum recommendations are beside their corresponding encodings. 0000: Reserved 0001: Reserved 0010: Reserved 0011: Reserved 0100: 4 clocks 0101: 5 clocks 0110: 6 clocks 0111: 7 clocks 23:20 R/W 6h (DDR333) 1000: 8 clocks 1001: 9 clocks (DDR 2 400) 1010: 10 clocks 1011: 11 clocks 1100: 12 clocks (DDR 2 533) 1101: 13 clocks 1110: 14 clocks 1111: 15 clocks Recommended values: 7h DDR 333 9h DDR 2 400 Ch DDR 2 533 19 18 RO Reserved 0b R/W Reserved 0b Activate to Activate Delay: 17 R/W 0b Control Act to Act delay between the different banks of the same rank. Trr is specified in “ns”. 10 ns for 2-KB page size and 7.5 ns for 1-KB page 0 = 2 Clock 1 = 3 Clock 40 Specification Update Documentation Changes Bit Access and Default Description Pre-All to Activate Delay (tRPALL). 16 R/W 0b This is applicable only to 8 bank architectures. Must be set to 1 if any Rank is populated with 8 bank device technology. 0: tRPALL = tRP 1: tRPALL = tRP + 1 Refresh Cycle Time (tRFC). Refresh cycle time is measured from a Refresh command (REF) until the first Activate command (ACT) to the same rank, required to perform a read or write. DDR 2 tRFC spec 15:11 R/W 01100 b tRFC 256 Mb 512 Mb 1 Gb DDR2 400 (5 ns) 75 ns = 15 clks 105 ns = 21 clks 127.5 ns = 26 clks DDR2 533 (3.75 ns) 75 ns = 20 clks 105 ns = 28 clks 127.5 ns = 34clks DDR 1 tRFC spec tRFC 64 Mb -512 Mb DDR 333 (6 ns) 75 ns = 13 clks 00000b – 11111b 1 Gb 120 ns = 20 clks Zero Clocks to Thirty-one Clocks respectively Actual clocks period depends on DDR clock frequency. BIOS should round up. If the required clock count exceeds as allowed by this register, the bios should set this register to the max value and set corresponding bits in SDBUP. 10 RO Reserved CASB Latency (tCL): This value is programmable on DDR 2 SO-DIMM’s. The value programmed here must match the CAS Latency of every DDR 2 SO-DIMM in the system. 9:8 R/W 01 b Encoding DDR CL DDR2 CL 00: 3 5 01: 2.5 4 (DDR2 533) 10: Reserved 3 (DDR2 400) 11: Reserved Reserved NOTE: Only the above recommended DDR2 timing values have been validated. 7 Specification Update RO Reserved 41 Documentation Changes Bit Description Access and Default DRAM RASB to CASB Delay (tRCD): This bit controls the number of clocks inserted between a row activate command and a read or write command to that row. Encoding 6:4 R/W 001 b DDR tRCD DDR2 tRCD 000: 2 2 clocks 001: 3 3 clocks (DDR2 400) 010: 4 4 clocks (DDR2 533) 011: 5 5 clocks 100 - 111: Reserved Reserved NOTE: Only the above recommended DDR2 timing values have been validated. 3 RO Reserved DRAM RASB Precharge (tRP): This bit controls the number of clocks that are inserted between a row precharge command and an activate command to the same rank. Encoding 2:0 R/W 001 b DDR tRP DDR2 tRP 000: 2 2 Clocks 001: 3 3 Clocks (DDR2 400) 010: 4 4 Clocks (DDR2 533) 011: 5 5 Clocks 100 - 111: Reserved NOTE: Only the above recommended DDR2 timing values have been validated. 6. LBKLT_CRTL Signal Description Update The LBKLT_CRTL signal description in Section 2.5.3 should be replaced with the following: LBKLT_CRTL O HVCMOS Panel Backlight Brightness Control: Panel brightness control. This signal is also called VARY_BL in the CPIS specification and is used as the PWM Clock input signal. The accuracy of the PWM duty cycle of LBKLT_CRTL signal for any given value will be within ±20 ns. 42 Specification Update Documentation Changes 7. Section 17.2.14 C0DRC0 – Channel 0 DRAM Controller Mode 0 The following register bit definition replaces the Mobile Intel® 915PM/GM/GME/GMS and 910GML/GMLE Express Chipset EDS Volume 2.0 (Doc #17139); Section 17.2.14. Section 17.2.14 Mode 0 C0DRC0—Channel 0 DRAM Controller PCI Device: MCHBAR Address Offset: 120h Default Value: 40002801h Access: RO, R/W Size: 32 bits Bit 31:30 Access & Default RO Description Reserved 01 b Initialization Complete (IC): 29 28 R/W 0b This bit is used for communication of software state between the memory controller and the BIOS. BIOS sets this bit to 1 after initialization of the DRAM memory array is complete. R/W Reserved 0b Active SDRAM Ranks: Implementations may use this field to limit the maximum number of SDRAM ranks that may be active at once. 0000: All ranks allowed to be in the active state 27:24 R/W 0001: One Rank 0h 0010: Two Ranks 0011: Three Ranks Others: Reserved. If this field is set to a non-zero value, then bits CXDRT2(4:0) should be set to the minimum value as described by the formula, else the system hangs. 23:22 21:20 Specification Update R/W Reserved 00 b R/W Reserved 00 b 43 Documentation Changes Bit 19:18 17 16 Access & Default RO Description Reserved 00 b R/W Reserved 0b R/W Reserved 0b CMD Copy Enable: In a single channel mode, the CMD pins (MA, BS, RAS, CAS, WE) on both channels are driven and are physical copies of each other. 15 R/W 0b Setting this bit disables the CMD pins on channel B. Having the additional copy of CMD pins helps reduce loading on these pins, since in a two SO-DIMM system, each copy can be routed up to separate SO-DIMM. In a single DIMM system, the second copy can be disabled to eliminate unnecessary toggling of these pins. If this bit needs to be set, BIOS should do that before memory init sequence. This bit should not be set in a dual channel system 14 13:12 11 RO Reserved 0b R/W Reserved 10 b RO Reserved 1b Refresh Mode Select (RMS): This field determines whether refresh is enabled and, if so, at what rate refreshes will be executed. 10:8 R/W 000 b 000: Refresh disabled 001: Refresh enabled. Refresh interval 15.6 µs 010: Refresh enabled. Refresh interval 7.8 µs Other: 7 44 RO Reserved Reserved 0b Specification Update Documentation Changes Bit Access & Default 6:4 R/W Mode Select (SMS): 000 b These bits select the special operational mode of the DRAM interface. The special modes are intended for initialization at power up. Description 000: Post Reset state – When the GMCH exits reset (power-up or otherwise), the mode select field is cleared to 000. During any reset sequence, while power is applied and reset is active, the GMCH deasserts all DRAM CLK and CKE signals. After internal reset is deasserted, DRAM CLK and CKE signals remain deasserted until this field is written to a value different than “000”. On this event, DRAM CLKs are enabled and CKE signals remain deasserted for a minumum of 35 ns before CKE signals are asserted. During suspend, GMCH internal signal triggers DRAM controller to flush pending commands and enter all ranks into Self-Refresh mode. As part of resume sequence, GMCH will be reset – which will clear this bit field to “000” and maintain all DRAM CLK and CKE signals deasserted. After internal reset is deasserted, DRAM CLK and CKE signals remain deasserted until this field is written to a value different than “000”. On this event, DRAM CLKs are enabled and CKE signals remain deasserted for a minumum of 35 ns before CKE signals are asserted. During entry to other low power states (C3, S1), GMCH internal signal triggers DRAM controller to flush pending commands and enter all ranks into Self-Refresh mode. During exit to normal mode, GMCH signal triggers DRAM controller to exit Self-Refresh and resume normal operation without S/W involvement. 001: NOP Command Enable – All CPU cycles to DRAM result in a NOP command on the DRAM interface. 010: All Banks Pre-charge Enable – All CPU cycles to DRAM result in an “all banks precharge” command on the DRAM interface. 011: Mode Register Set Enable – All CPU cycles to DRAM result in a “mode register” set command on the DRAM interface. Host address lines are mapped to DRAM address lines in order to specify the command sent. Host address lines [12:3] are mapped to MA[9:0], and HA[13] is mapped to MA[11]. 101: Reserved 110: CBR Refresh Enable – In this mode all CPU cycles to DRAM result in a CBR cycle on the DRAM interface 111: 3 Specification Update R/W Normal operation Reserved 0b 45 Documentation Changes Bit Access & Default Description Burst Length (BL): 2 R/W 0b The burst length is the number of QWORDS returned by a SO-DIMM per read command, when not interrupted. This bit is used to select the DRAM controller’s Burst Length operation mode. It must be set to match to the behavior of the SO-DIMM. 0: Burst Length of 4 1: Burst Length of 8 DRAM Type (DT): Used to select between supported SDRAM types. 1:0 8. RO 00: Reserved 01 b 01: Dual Data Rate (DDR) 10: Dual Data Rate 2 (DDR 2) 11: Reserved SDRAM SDRAM Section 17.2.15 C0DRC1 – Channel 0 DRAM Controller Mode 1 The following register bit definition replaces the Mobile Intel® 915PM/GM/GME/GMS and 910GML/GMLE Express Chipset EDS Volume 2.0 (Doc #17139); Section 17.2.15. PCI Device: MCHBAR Address Offset: 124h Default Value: 00000000h Access: RO, R/W Size: 32 bits Bit Access & Default R/W Description Reserved 31 0b R/W Reserved 30 0b R/W Reserved 29:28 00 b RO Reserved 27 0b 46 Specification Update Documentation Changes Bit Access & Default R/W Description Reserved 26:24 000 b RO Reserved 23:20 00 h CKE Tri-state Enable Per Rank: Bit 16 corresponds to rank 0 Bit 17 corresponds to rank 1 R/W Bit 18 corresponds to rank 2 0h Bit 19 corresponds to rank3 19:16 0 = CKE is not tri-stated. 1 = CKE is tri-stated. This is set only if the Rank is physically not populated. RO Reserved 15:13 000 b CS# Tri-state Enable (CSBTRIEN): R/W 12 0b When set to a 1, the DRAM controller will tri-state CS# when the corresponding CKE is deasserted. 0: Address Tri-state Disabled 1: Address Tri-state Enabled Address Tri-state Enable (ADRTRIEN): R/W 11 0b R/W 10 When set to a 1, the DRAM controller will tri-state the MA, CMD, and CSB (CSB if lines only when all CKEs are deasserted. CKEs deassert based on Idle timer or max rank count control. 0: Address Tri-state Disabled 1: Address Tri-state Enabled Reserved 0b RO Reserved 9 0b Specification Update 47 Documentation Changes Bit Description Access & Default DRAM Channel IO-Buffers Activate: This bit is cleared to 0 during reset and remains inactive until it is set to 1 by BIOS. In addition, this bit can be cleared and set during debug procedures. R/W 8 0b While 0, the DRAM controller core logic forces the state of the IO-buffers in this channel to “reset” or “preset”, depending on the specific buffer type. The buffers’ state flops (counters, pointers, etc) are forced into their initial state. The core logic receive FIFO read pointer is forced into its initial state. Output signal groups are driven to their initial state (tristated). Refer to the following tables for the specific state. While 1, the DRAM controller core logic enables the DRAM IO-buffers in this channel to operate normally. R/W Reserved 7 0b R/W Reserved 6 0b RO Reserved 5 0b 4 R/W Reserved RO Reserved 3:2 00 b R/W Reserved 1 0b R/W Reserved 0 0b § 48 Specification Update