Dell PowerEdge 7250 - Manual (Page 14)

SR870BN4 Machine Check Error Handling SR870BN4 Error Reference Guide
Revision 1.0 8
There are two types of machine check events: local and global. A local MCA is when an
individual processor enters machine check. Some examples of local machine checks
include a Distributed Translation Lookaside Buffer (DTLB) data parity error, or when the
processor consumes data with an uncorrectable error.
A machine check is global when all processors enter machine check. On the
SR870BN4 platform, the method used to get all processors into machine check are the
BINIT# and BERR# signals. When a processor takes a local machine check, it may
escalate the error to a global machine check to transition other processors to a known
state and/or for error containment. For example, the processor may assert BINIT# in
response to a transaction time-out event.
The SR870BN4 platform does not assert BINIT#, only BERR#. BERR# is asserted for
platform fatal errors and when an uncorrectable error is detected on outbound data.
For more information on the SR870BN4 implementation of machine check error
handling, refer to the SR870BN4 SAL Error Handling Specification.
4.2 Error Reporting
SR870BN4 machine check error handling allows enhanced error reporting of processor and
platform errors. These errors are prioritized and signaled to system hardware and software.
System software (PAL/SAL) provides well-defined APIs for application software to acquire
information about system errors in the form of standard data structures. These errors are
logged to non-volatile storage and/or made available for consumption by application software
during runtime. These errors are in the MCA records and they are based on the Itanium™
System Abstraction Layer Specification Rev 3.0.
On the SR870BN4, based on the MCA records, system events related to Field Replaceable
Units (FRUs) are logged in the BMC SEL. Each MCA record results in the generation of one or
more corresponding BMC SEL events. In addition, an auxiliary log entry event will be logged
corresponding to each MCA record. The SEL messages are IPMI 1.5-compliant platform event
messages.
The following rules are applied to the translation of SAL 3.0 MCA records to IPMI 1.5-compliant
platform event messages :
Table 1. SAL 3.0 MCA Records
MCA SAL Record Section Type SEL Event : Sensor Type SEL event: Event Data Bytes
Processor Processor IERR SMBIOS Type 4 0-based index
Error Severity
PCI Bus PERR/SERR Critical Interrupt
PERR
SERR PCI Bus number
PCI Bus Other Errors Critical Interrupt
Bus Correctable error
Bus Uncorrectable error None