Machine Check Code; Eccc - Prime Computer 50 Series Service Manual

Corrective Maintenance Prime Proprietary
TABLE 6-25: LOGREC DISK ERROR (Cont.)
ERROR DESCRIPTION
RETRIES=xx yy The number of retries attempted. If less than ten,
the operation was completed successfully, yy
will be recovered. If equal to ten, and the
error could not be corrected by ECC, yy is Uncor-
rectable. If an ECC error has been successfully
corrected by the software, yy is WORD NUMBER -
'xx, which gives that word number in the record
and the 32 bit correction pattern used. If
RETRIES equals ten, the error was unrecovered.
6.2.6.2 LOGREC Machine Checks
The possible LOGREC Machine Checks, their formats and descriptions are
presented in Table 6-26. For further information refer to the Machine
Check Handling Guide in this chapter.
TABLE 6-26: LOGREC MACHINE CHECKS
ERROR DESCRIPTION
MACHINE CHECK:xx If DSWPARITY is NOT present, an encoding of the
Machine Check code is presented.
. DSWSTAT=xx Diagnostic Status Word (32 bits long) that
contains a comprehensive status of the system at
the time the error was detected.
DSWRMA=xx Diagnostic Status Word Memory Address Register (32
bits) contains the address in the memory address
register.
DSWPB=xx Diagnostic Status Word Procedure Base that contains
the address where the fault was detected.
DSWPARITY=xx Diagnostic Status Word Parity (32 bits) reports
board and signal name.
6.2.6.3 LOGREC Memory Parity
Memory parity errors are either correctable (ECCC) or uncorrectable
(ECCU). ECCC errors are recorded in the LOGREC file during PRIMOS
operation. ECCU errors are forcing the system to a Halt during PRIMOS
operation, at location 4/276 octal (segment 4, location 276). This
shows up as octal 277 in the address lights. The LOGREC memory parity
format and a detailed EBCCC Handling Guide follow.
070-C 6 - 38