86h - EV6 Detected DCache Tag Parity error on initial probe
Indicator:All cpu error registers are zero
Description:The EV6 detected a Dcache tag paraity error during the initial tag probe of a load or store instruction. This error created a synchronous fault to the dfault PALcode entry point and was corrected. The error floow starts of treating the fault as a machine check. If the error happened in PALmode, the system halts. If no machine checks are in progress, we turn the fault in to a crd. If a machine check is in progress this error will be dismissed.
Analysis:All processor specific registers are zero.
Recovery:Corrected by PALcode.
Faulting FRU:

86h - EV6 detected ICache Tag or Data Parity Error
Indicator: EV6 -Pass 2.2 I_STAT<31:28> = 0010b for Tag ; 0100b for Data
EV6 Pass 3.0 I_STAT<29> = 1 for Tag or Data.
Description:The DECChip 21264 detected an ICache Tag or Data Parity read error. This EV6 Pass 2.2 error is uncorrectable and no PALcode action is performed other than to deliver the uncorrectable frame. On the EV6 Pass 3.0 error hardware posts this as correctable read data error.
Analysis:C_ADDR Contains the physical address of the octaword associated with the error.
Recovery:The EV6 Pass 2.2 hardware traps. The error is processed by the SRM Console PAL, UNIX and OpenVMS 660, and WNT Hal machine check handlers. The EV6 Pass 3.0 hardware traps and replays the instructions that were fetched during the error, then flushes the entire ICache to insure the re-fetched instructions do not come directly from the failing cache context. The SRM console PALcode, UNIX and OpenVMS 630 and WNT HAL machine check handlers process this error. Typically this error condition will be threshold processed by the UNIX, OpenVMS, and WNT filters and logged by the appropriate error handler.
Faulting FRU:

86h  - EV6 detected System Memory Single Bit ECC Error on DCache Fill
Indicator:DC_STAT<4:0> = 08h and EV6 C_STAT<4:0> = 03h for DSTREAM or
Description:A single bit System Memory ECC the DECChip 21264 detected correctable error during a DCache fill. The 630-machine check handler processes this error.
Analysis:C_ADDR Contains the physical address of the octaword associated with the error.
C_SYNDROME0 or C_SYNDROME1 will contain the ECC syndrome bits for the associated quadword.
Recovery:PALcode will scrub always the DCache block that contained the error. Typically this error condition will be processed by the Operating System filter and logged when the extended CRD filter exceeds filter threshold. System operation typically requires no recovery.
Faulting FRU:

86h - EV6 detected System Memory Single Bit ECC Error ICache Fill
Indicator: DC_STAT<4:0> = 08h and C_STAT<4:0> = 0Bh
Description:The DECChip 21264 detected a correctable a single bit System Memory ECC error during an ICache fill. The 630-machine check handler processes this error.
Analysis:C_ADDR Contains the physical address of the octaword associated with the error.
C_SYNDROME0 or C_SYNDROME1 will contain the ECC syndrome bits for the associated quadword.
Recovery:PALcode will scrub always the ICache or DCache block that contained the error. Typically this error condition will be processed by the Operating System filter and logged when the extended CRD filter exceeds filter threshold. System operation typically requires no recovery.
Faulting FRU:

86h - EV6 detected BCache Single Bit ECC Error on DCache Fill
Indicator:DC_STAT<4:0> = 08h and EV6 C_STAT<4:0> = 04h for DSTREAM
Description:A single bit ECC correctable error was detected by the DECChip 21264 during a DCache fill from BCache. The 630-machine check handler processes this error.
Analysis:C_ADDR Contains the physical address of the octaword associated with the error.
C_SYNDROME0 or C_SYNDROME1 will contain the ECC syndrome bits for the associated quadword.
Recovery:PALcode will scrub the BCache block that contained the error. Typically this error condition will be processed by the Operating System filter and logged if the filter threshold is exceeded by extended CRD filter. System operation typically requires no actual recovery.
Faulting FRU:

86h - EV6 detected BCache Single Bit ECC Error on ICache Fill
Indicator:DC_STAT<4:0> = 08h and EV6 C_STAT<4:0> = 0Ch for ISTREAM
Description:A single bit ECC correctable error was detected by the DECChip 21264 during an Istream fill from BCache. The 630-machine check handler processes this error.
Analysis:C_ADDR Contains the physical address of the octaword associated with the error.
C_SYNDROME0 or C_SYNDROME1 will contain the ECC syndrome bits for the associated quadword.
Recovery:PALcode will scrub the BCache block that contained the error. Typically this error condition will be processed by the Operating System filter and logged if the filter threshold is exceeded by extended CRD filter. System operation typically requires no actual recovery.
Faulting FRU:

9eh - EV6 detected DCache Tag Parity Error
Indicator:EV6 DC_STAT<1:0> = 00b and MM_STAT<31:0> = 00000400h
Description:The DECChip 21264 detected a DCache Tag Parity error. The 630-machine check handler processes this error and the EXC_ADDR is loaded with the PC of the instruction that triggered the error.
Analysis:C_ADDR Contains the physical address of the octaword associated with the error.
Recovery:EV6 PALcode will flush the error block using Evict Cache block instruction. The on-chip duplicate cache tag provides the correct victim address and cache coherency resulting in the system operation not requiring any actual recovery. The SRM console PAL, UNIX and OpenVMS 630 and WNT HAL machine check handlers process this error. Typically this error condition will be threshold processed by the UNIX, OpenVMS, and WNT filters and logged by the appropriate error handler.
Faulting FRU:CPU

90h - OS Detected Bugcheck Error
Indicator:
Description:The OS detected an abnormal occurrence. This is usually a software problem.
Analysis:
Recovery:Operating system specific.
Faulting FRU:

98h - EV6 detected Second DCache Store Data ECC Error
Indicator:DC_STAT<4:0> = 14h
Description:The DECChip 21264 detected a DCache store instruction ECC Error while DC_STAT was locked by a previous DCache store error. This error is processed by the 670 machine check handler and is typically fatal to the system especially if the CPU is the master and error recovery was in process during the time the second error was detected.
Analysis:C_ADDR Contains the physical address of the octaword associated with the error. Syndrome data is irrelevant.
Recovery:This error is only recoverable if CPU failover is operational within the system and total system context has been preserved.
Faulting FRU:CPU or memory

98h - EV6 detected DCache Tag Parity Error in Pipeline 0
Indicator:EV6 MM_STAT<31:0>=00000100 h AND DC_STAT<4:0> = 01h for pipeline 0
Description:The DECChip 21264 detected a DCache Tag Parity error on a memory reference. If the EV6 has already retired the memory reference, then EXC_ADDR is not available. If the memory reference has not been retired, the MBOX will load the EXC_ADDR with the PC address of the instruction that triggered the error. This error is processed by the 670-machine check handler and is typically fatal to the system especially if the CPU is the master.
Analysis:C_ADDR Contains the physical address of the octaword associated with the error.
Recovery:This error is only recoverable if CPU failover is operational within the system and system context has been preserved.
Faulting FRU:CPU or memory

98h - EV6 detected DCache Tag Parity Error in Pipeline 1
Indicator:EV6 MM_STAT<31:0>=00000100h and DC_STAT<4:0> = 02h for pipeline 1
Description:The DECChip 21264 detected a DCache Tag Parity error on a memory reference. If the EV6 has already retired the memory reference, then EXC_ADDR is not available. If the memory reference has not been retired, the MBOX will load the EXC_ADDR with the PC address of the instruction that triggered the error. This error is processed by the 670-machine check handler and is typically fatal to the system especially if the CPU is the master.
Analysis:C_ADDR Contains the physical address of the octaword associated with the error.
Recovery:This error is only recoverable if CPU failover is operational within the system and system context has been preserved.
Faulting FRU:CPU or memory

98h - EV6 detected Duplicate DCache Tag Parity Error
Indicator:C_STAT<4:0> = 00010b
Description:The DECChip 21264 detected a Duplicate DCache Tag Parity error. This error occurs when the EV6 CBOX performs a physically indexed read of eight locations in the tag array. This error is processed by the 670-machine check handler and is typically fatal to the system especially if the CPU is the master.
Analysis:C_ADDR Contains the physical address of the octaword associated with the error.
Recovery:This error is only recoverable if CPU failover is operational within the system and system context has been preserved.
Faulting FRU:CPU or memory

98h - EV6 detected Memory Double Bit ECC DCache Fill
Indicator:EV6 DC_STAT<4:0> = 01000b and (EV6 C_STAT<4:0> = 13h for DCache Fill from system memory
Description:The DECChip 21264 detected a BCache or System Memory Double Bit ECC during a DCache fill. This error is processed by the 670-machine check handler and is typically fatal to the system.
Analysis:C_ADDR Contains the physical address of the octaword associated with the error.
Recovery:This error is only recoverable if CPU failover is operational within the system and system context has been preserved. and/or the defective system memory page can be mapped out by the Operating System and proper execution can resume.
Faulting FRU:CPU or memory

98h - EV6 detected Memory Double Bit ECC ICache Fill
Indicator:EV6 DC_STAT<4:0> = 01000b and EV6 C_STAT<4:0> = 1Bh for ICache Fill from system memory
Description:The DECChip 21264 detected a BCache or System Memory Double Bit ECC during an ICache fill. This error is processed by the 670-machine check handler and is typically fatal to the system.
Analysis:C_ADDR Contains the physical address of the octaword associated with the error.
Recovery:This error is only recoverable if CPU failover is operational within the system and system context has been preserved and/or the defective system memory page can be mapped out by the Operating System and proper execution can resume.
Faulting FRU:CPU or memory

98h - EV6 detected BCache Double Bit ECC DCache Fill
Indicator:EV6 DC_STAT<4:0> = 01000b and EV6 C_STAT<4:0> = 14h for DCache Fill from BCache.
Description:The DECChip 21264 detected a BCache or System Memory Double Bit ECC error during a DCache fill. This error is processed by the 670-machine check handler and is typically fatal to the system.
Analysis:C_ADDR Contains the physical address of the octaword associated with the error.
Recovery:This error is only recoverable if CPU failover is operational within the system and system context has been preserved and/or the defective system memory page can be mapped out by the Operating System and proper execution can resume.
Faulting FRU:CPU or memory

98h - EV6 detected BCache Double Bit ECC ICache Fill
Indicator:EV6 DC_STAT<4:0> = 01000b and EV6 C_STAT<4:0> = 1Ch for ICache Fill from BCache
Description:The DECChip 21264 detected a BCache or System Memory Double Bit ECC error during an ICache fill. This error is processed by the 670-machine check handler and is typically fatal to the system.
Analysis:C_ADDR Contains the physical address of the octaword associated with the error.
Recovery:This error is only recoverable if CPU failover is operational within the system and system context has been preserved and/or the defective system memory page can be mapped out by the Operating System and proper execution can resume.
Faulting FRU:CPU or memory

98h - EV6 detected BCache Probe Hit Double Bit ECC Error
Indicator:EV6 C_STAT<4:1> = 1011b
Description:A double bit BCache ECC error was detected during EV6 execution of probe data transaction command. This error is processed by the PALcode 670 machine check handler and is typically fatal to the system.
Analysis:C_ADDR Contains the physical address of the octaword associated with the error.
Recovery:This error is only recoverable if CPU failover is operational within the system and BCache coherency in the system can be maintained for continued proper execution.
Faulting FRU:CPU or memory

98h - EV6 detected BCache Tag Parity Error
Indicator:EV6 C_STAT<4:0> = 00001b
Description:The DECChip 21264 detected a BCache Tag Parity error. Direct load, store or indirect probe references can cause this error. This error is processed by the 670-machine check handler and is always fatal to the system.
Analysis:C_ADDR Contains the physical address of the octaword associated with the error.
Recovery:This error is only recoverable if CPU failover is operational within the system and BCache coherency in the system can be maintained for continued proper execution.
Faulting FRU:CPU or memory

204h - System ( 212xx PChip ) detected Single bit Correctable ECC Error
Indicator:P0_CTL<18>= 1 and DIR0<62> = 1 and P0_ERROR<11>= 1 and P0_ERROR<51> = 0 from PCHIP0 or
P1_CTL<18>= 1 and DIR0<61> = 1 and P1_ERROR<11>= 1 and P1_ERROR<51> = 0 from PCHIP1
Description:A system single bit correctable DMA or S/G read error from PCI memory space has been detected by PChip 0 or 1. The 620-machine check handler processes this error.
Analysis: If PChip 0 error
P0_ERROR<63:56>, SYN, will contain the ECC Syndrome of the error P0_ERROR<55:52>, CMD, will contain the type of DMA operation. Refer to Table 5-27 for the definitions
P0_ERROR<50:16>, ADDR, will contain the system address bits <34:3> of the erroneous quadword.
If PChip 1 error then
P1_ERROR<63:56>, SYN, will contain the ECC Syndrome of the error P1_ERROR<55:52>, CMD, will contain the type of DMA operation. Refer to Table 5-28 for the definitions
P1_ERROR<50:16>, ADDR, will contain the system address bits <34:3> of the erroneous quadword.
Recovery:The PChip corrects the ECC data error within the longword (for single PCI access) or quadword (for dual-access cycle). Typically this error condition will be processed by the Operating System Memory CRD filter and the system operation typically requires no recovery unless CRD filtering thresholds have been exceeded. In this case the operating system should map out the related memory page and log the appropriate error event.
Faulting FRU:Error detected in device or option and corrected during DMA reads.

202h - System detected PCI Address Parity Error (SERR#)
Indicator:DIR0<62> = 1 and P0_ERROR<1> = 1 and P0_ERROR<51> = 0 from PCHIP0 or DIR0<61> = 1 and P1_ERROR<1> = 1and P1_ERROR<51> = 0 from PCHIP1
Description:PChip 0 or 1 has detected a PCI address parity error when acting as a master. This is processed through the 660-error handler.
Analysis: If PChip 0 error
P0_ERROR<55:52>, CMD, will contain the PCI Command of the transaction that caused the error.
P0_ERROR<47:18>, ADDR, will contain address bits <31:2> of the starting PCI Address of the transaction that caused the error.
P0_ERROR<17:16> will be 00b if not a DAC operation; 01b if via DAC SG to Window 3; and 1xb via monster window.
If PChip 1 error then
P1_ERROR<55:52>, CMD, will contain the PCI Command of the transaction that caused the error.
P1_ERROR<47:18>, ADDR, will contain address bits <31:2> of the starting PCI Address of the transaction that caused the error.
P1_ERROR<17:16> will be 00b if not a DAC operation; 01b if via DAC SG to Window 3; and 1xb via monster window.
Recovery:There is no recovery for this defective PCI target transaction.
Faulting FRU:Option

202h - System detected PCI Write Data Parity Error (PERR#)
Indicator:DIR0<62> = 1 and P0_ERROR<2> = 1 and P0_ERROR<51> = 0 from PCHIP0 or DIR0<61> = 1 and P1_ERROR<2> = 1and P1_ERROR<51> = 0 from PCHIP1
Description:PChip 0 or 1 detected a PCI write data parity error when acting as a master. This is processed through the 660-error handler.
Analysis: If PChip 0 error
P0_ERROR<55:52>, CMD, will contain the PCI Command of the transaction that caused the error.
P0_ERROR<47:18>, ADDR, will contain address bits <31:2> of the starting PCI Address of the transaction that caused the error.
P0_ERROR<17:16> will be 00b if not a DAC operation; 01b if via DAC SG to Window 3; and 1xb via monster window.
If PChip 1 error then
P1_ERROR<55:52>, CMD, will contain the PCI Command of the transaction that caused the error.
P1_ERROR<47:18>, ADDR, will contain address bits <31:2> of the starting PCI Address of the transaction that caused the error.
P1_ERROR<17:16> will be 00b if not a DAC operation; 01b if via DAC SG to Window 3; and 1xb via monster window.
Recovery:There is no recovery for this defective PCI target transaction.
Faulting FRU:Option

202h - System detected PCI Delayed Completion Retry Time Out Error (DCRTO)
Indicator:DIR0<62> = 1 and P0_ERROR<3> = 1 and P0_ERROR<51> = 0 from PCHIP0 or DIR0<61> = 1 and P1_ERROR<3> = 1and P1_ERROR<51> = 0 from PCHIP1
Description:A PCI delayed completion retry time out error has been detected by a target PChip 0 or 1. This is processed through the 660-error handler.
Analysis: If PChip 0 error
P0_ERROR<55:52>, CMD, will contain the PCI Command of the transaction that caused the error.
P0_ERROR<47:18>, ADDR, will contain address bits <31:2> of the starting PCI Address of the transaction that caused the error.
P0_ERROR<17:16> will be 00b if not a DAC operation; 01b if via DAC SG to Window 3; and 1xb via monster window.
If PChip 1 error then
P1_ERROR<55:52>, CMD, will contain the PCI Command of the transaction that caused the error.
P1_ERROR<47:18>, ADDR, will contain address bits <31:2> of the starting PCI Address of the transaction that caused the error.
P1_ERROR<17:16> will be 00b if not a DAC operation; 01b if via DAC SG to Window 3; and 1xb via monster window.
Recovery:There is no recovery for this defective PCI master transaction.
Faulting FRU:Option

202h - System detected PCI Invalid Scatter/Gather Page Table Entry Error (SGTE)
Indicator:DIR0<62> = 1 and P0_ERROR<4> = 1 and P0_ERROR<51> = 0 from PCHIP0 or DIR0<61> = 1 and P1_ERROR<4> = 1and P1_ERROR<51> = 0 from PCHIP1
Description:A PCI invalid scatter/gather page table entry error has been detected by a target PChip 0 or 1. This is processed through the 660-error handler.
Analysis: If PChip 0 error
P0_ERROR<55:52>, CMD, will contain the PCI Command of the transaction that caused the error.
P0_ERROR<47:18>, ADDR, will contain address bits <31:2> of the starting PCI Address of the transaction that caused the error.
P0_ERROR<17:16> will be 00b if not a DAC operation; 01b if via DAC SG to Window 3; and 1xb via monster window.
If PChip 1 error then
P1_ERROR<55:52>, CMD, will contain the PCI Command of the transaction that caused the error.
P1_ERROR<47:18>, ADDR, will contain address bits <31:2> of the starting PCI Address of the transaction that caused the error.
P1_ERROR<17:16> will be 00b if not a DAC operation; 01b if via DAC SG to Window 3; and 1xb via monster window.
Recovery:There is no recovery for this defective PCI master transaction.
Faulting FRU:Option

202h - System detected PCI Address/Command Parity Error (APE)
Indicator:DIR0<62> = 1 and P0_ERROR<5> = 1 and P0_ERROR<51> = 0 from PCHIP0 or DIR0<61> = 1 and P1_ERROR<5> = 1and P1_ERROR<51> = 0 from PCHIP1
Description:A target PChip 0 or 1 detected a PCI address/command cycle parity error. This is processed through the 660-error handler.
Analysis: If PChip 0 error
P0_ERROR<55:52>, CMD, will contain the PCI Command of the transaction that caused the error.
P0_ERROR<47:18>, ADDR, will contain address bits <31:2> of the starting PCI Address of the transaction that caused the error.
P0_ERROR<17:16> will be 00b if not a DAC operation; 01b if via DAC SG to Window 3; and 1xb via monster window.
If PChip 1 error then
P1_ERROR<55:52>, CMD, will contain the PCI Command of the transaction that caused the error.
P1_ERROR<47:18>, ADDR, will contain address bits <31:2> of the starting PCI Address of the transaction that caused the error.
P1_ERROR<17:16> will be 00b if not a DAC operation; 01b if via DAC SG to Window 3; and 1xb via monster window.
Recovery:There is no recovery for this defective PCI master transaction.
Faulting FRU:Option

202h - System detected PCI Target Abort Error (TA)
Indicator:DIR0<62> = 1 and P0_ERROR<6> = 1 and P0_ERROR<51> = 0 from PCHIP0 or DIR0<61> = 1 and P1_ERROR<6> = 1and P1_ERROR<51> = 0 from PCHIP1
Description:PChip 0 or 1 detected a PCI target abort when acting as a master. This is processed through the 660-error handler.
Analysis: If PChip 0 error
P0_ERROR<55:52>, CMD, will contain the PCI Command of the transaction that caused the error.
P0_ERROR<47:18>, ADDR, will contain address bits <31:2> of the starting PCI Address of the transaction that caused the error.
P0_ERROR<17:16> will be 00b if not a DAC operation; 01b if via DAC SG to Window 3; and 1xb via monster window.
If PChip 1 error then
P1_ERROR<55:52>, CMD, will contain the PCI Command of the transaction that caused the error.
P1_ERROR<47:18>, ADDR, will contain address bits <31:2> of the starting PCI Address of the transaction that caused the error.
P1_ERROR<17:16> will be 00b if not a DAC operation; 01b if via DAC SG to Window 3; and 1xb via monster window.
Recovery:There is no recovery for this defective PCI target transaction.
Faulting FRU:Option

202h - System detected PCI Bus Read Data Parity Error -(RDPE)
Indicator:DIR0<62> = 1 and P0_ERROR<7> = 1 and P0_ERROR<51> = 0 from PCHIP0 or DIR0<61> = 1 and P1_ERROR<7> = 1and P1_ERROR<51> = 0 from PCHIP1
Description:PChip 0 or 1 detected a PCI read data parity error when acting as a master. This is processed through the 660-error handler.
Analysis: If PChip 0 error
P0_ERROR<55:52>, CMD, will contain the PCI Command of the transaction that caused the error.
P0_ERROR<47:18>, ADDR, will contain address bits <31:2> of the starting PCI Address of the transaction that caused the error.
P0_ERROR<17:16> will be 00b if not a DAC operation; 01b if via DAC SG to Window 3; and 1xb via monster window.
If PChip 1 error then
P1_ERROR<55:52>, CMD, will contain the PCI Command of the transaction that caused the error.
P1_ERROR<47:18>, ADDR, will contain address bits <31:2> of the starting PCI Address of the transaction that caused the error.
P1_ERROR<17:16> will be 00b if not a DAC operation; 01b if via DAC SG to Window 3; and 1xb via monster window.
Recovery:There is no recovery for this defective PCI target transaction.
Faulting FRU:Option

202h - System detected PCI No Device Select Error (NDS)
Indicator:DIR0<62> = 1 and P0_ERROR<8> = 1 and P0_ERROR<51> = 0 from PCHIP0 or DIR0<61> = 1 and P1_ERROR<8> = 1and P1_ERROR<51> = 0 from PCHIP1
Description:A PCI no device select error has been detected by a master PChip 0 or 1 and the transaction has been terminated by PChip master abort. This is processed through the 660-error handler.
Analysis: If PChip 0 error
P0_ERROR<55:52>, CMD, will contain the PCI Command of the transaction that caused the error.
P0_ERROR<47:18>, ADDR, will contain address bits <31:2> of the starting PCI Address of the transaction that caused the error.
P0_ERROR<17:16> will be 00b if not a DAC operation; 01b if via DAC SG to Window 3; and 1xb via monster window.
If PChip 1 error then
P1_ERROR<55:52>, CMD, will contain the PCI Command of the transaction that caused the error.
P1_ERROR<47:18>, ADDR, will contain address bits <31:2> of the starting PCI Address of the transaction that caused the error.
P1_ERROR<17:16> will be 00b if not a DAC operation; 01b if via DAC SG to Window 3; and 1xb via monster window.
Recovery:There is no recovery for this defective PCI target transaction.
Faulting FRU:Option

202h - System detected PCI Retry Time Out Error (RTO)
Indicator:DIR0<62> = 1 and P0_ERROR<9> = 1 and P0_ERROR<51> = 0 from PCHIP0 or DIR0<61> = 1 and P1_ERROR<9> = 1and P1_ERROR<51> = 0 from PCHIP1
Description:A PCI retry time out error has been detected by a master PChip 0 or 1 after 2**24 retry responses. This is processed through the 660-error handler.
Analysis: If PChip 0 error
P0_ERROR<55:52>, CMD, will contain the PCI Command of the transaction that caused the error.
P0_ERROR<47:18>, ADDR, will contain address bits <31:2> of the starting PCI Address of the transaction that caused the error.
P0_ERROR<17:16> will be 00b if not a DAC operation; 01b if via DAC SG to Window 3; and 1xb via monster window.
If PChip 1 error then
P1_ERROR<55:52>, CMD, will contain the PCI Command of the transaction that caused the error.
P1_ERROR<47:18>, ADDR, will contain address bits <31:2> of the starting PCI Address of the transaction that caused the error.
P1_ERROR<17:16> will be 00b if not a DAC operation; 01b if via DAC SG to Window 3; and 1xb via monster window.
Recovery:There is no recovery for this defective PCI target transaction.
Faulting FRU:Option

202h - System detected Uncorrectable ECC Error (UECC)
Indicator:P0_CTL<18>= 1 and DIR0<62> = 1 and P0_ERROR<10> = 1 and P0_ERROR<51> = 0 from PCHIP0 or
P1_CTL<18>= 1 and DIR0<61> = 1and P1_ERROR<10> = 1 and P1_ERROR<51> = 0 from PCHIP1
Description:An uncorrectable system ECC error has been detected by PChip 0 and/or 1. This is processed through the 660-error handler.
Analysis: If PChip 0 error
P0_ERROR<63:56>, SYN, will contain the ECC Syndrome of the error P0_ERROR<55:52>, CMD, will contain the type of DMA operation. Refer to Table 5-27 for the definitions
P0_ERROR<50:16>, ADDR, will contain the system address bits <34:3> of the erroneous quadword.
If PChip 1 error then
P1_ERROR<63:56>, SYN, will contain the ECC Syndrome of the error P1_ERROR<55:52>, CMD, will contain the type of DMA operation. Refer to Table 5-28 for the definitions
P1_ERROR<50:16>, ADDR, will contain the system address bits <34:3> of the erroneous quadword.
Recovery:Typically there is no recovery for this defect within the PCI memory space unless the operating system is capable of swapping out the erred memory map page(s) and still able to process PCI transactions properly.
Faulting FRU:Option

202h - System detected Non-Existent Memory Reference Error
Indicator:DIR0<63> = 1 and MISC<28>= 1
Description:The CChip has detected an asynchronous non-existing memory reference. This error is processed through the 660-error handler.
Analysis:
Recovery:There is no recovery for this error with the exception.
Faulting FRU:Option

206 - System detected Fan Failure Error
Indicator:CChip DRIR<48>, DIM0<48>, DIR0<48>
Description:A system fan correctable error has been detected. This error is processed through the PAL code 680-error handler.
Analysis:None
Recovery:Because the system does not have redundant fans, the OS should do what is necessary to save the current system state, log the error and shut the power supply off in a reason amount of time. If the power is not removed from the system severe damage may occur.
Faulting FRU:Fan