Support for hardware dump collection from OCMB
This commit introduces the capability to collect hardware dumps from
Odyssey-based OCMB chips in addition to existing processor dump.
This commit addresses the need for additional failure data across
both processors and their associated OCMB chips.
Key changes include:
- Enabled hardware dump collection specifically for Odyssey OCMB chips.
- Optimized the collection process to occur only when the chip clock
is on to minimize overall dump collection time.
- Sequential dump collection from OCMB chips associated with each
processor.
Change-Id: Iea70912f9fef9c337ba5ca7c508728586be25af8
Signed-off-by: Dhruvaraj Subhashchandran <dhruvaraj@in.ibm.com>
diff --git a/dump/dump_utils.cpp b/dump/dump_utils.cpp
index 2689b50..3e0105e 100644
--- a/dump/dump_utils.cpp
+++ b/dump/dump_utils.cpp
@@ -60,8 +60,10 @@
void requestSBEDump(const uint32_t failingUnit, const uint32_t eid,
SBETypes sbeType)
{
- lg2::info("Requesting Dump PEL({EID}) chip position({FAILINGUNIT})", "EID",
- eid, "FAILINGUNIT", failingUnit);
+ lg2::info(
+ "Requesting Dump PEL({EID}) chip({CHIPTYPE}) position({FAILINGUNIT})",
+ "EID", eid, "CHIPTYPE", sbeTypeAttributes.at(sbeType).chipName,
+ "FAILINGUNIT", failingUnit);
auto path = sbeTypeAttributes.at(sbeType).dumpPath.c_str();
constexpr auto interface = "xyz.openbmc_project.Dump.Create";