Intel
®
81341 and 81342 I/O Processors
December 2007
Developer’s Manual
Order Number: 315037-002US
621
DDR SDRAM Memory Controller—Intel
®
81341 and 81342
7.5
ECC Interrupts/Error Conditions
The DMCU has two conditions which require intervention from the Intel XScale
®
microarchitecture. When a single-bit error is detected during a read cycle, the DMCU
corrects the data returned but software needs to fix the error in the memory array.
When a multi-bit error is detected, the core decides how to handle the condition. For all
ECC errors, the DMCU records the requester of the transaction resulting in the error in
DELOGx[23:16] and interrupts the core.
When the DMCU detects an ECC error during a read or write cycle
23
, DMCISR[0] or
DMCISR[1] is set to 1. Whenever the DMCU toggles one of the DMCISR bits from 0 to
1, an interrupt is generated to the core.
shows how the DMCU responds to error conditions.
Note:
When ECC reporting is enabled with DECCR[1] or DECCR[0] and an ECC error occurs,
DMCISR[1] or DMCISR[0] is set and DELOGx/DEARx logs the error in addition to
actions.
23.Any error condition during a write cycle actually occurs while performing the read portion of a
read-modify-write on a partial write. See
Section 7.3.4.1, “DDR ECC Generation” on page 608
for
details.
Table 371. DMCU Error Response
Error Type
DMCU Action
Single-Bit during a read or write Fix Error (when ECC error correction enabled in the DECCR)
Multi-bit during a read
Target Abort the Internal Bus or direct port transaction.
Multi-bit during a write
New ECC is generated with bad data and written to DDR SDRAM array. Data
location is no longer valid.
Notes:
1.
The ECC Enable bit in the DECCR must be set in order for these actions to occur.