Hello Curtis Dunham, Andreas Sandberg,
I'd like you to do a code review. Please visit
https://gem5-review.googlesource.com/5054
to review the following change.
Change subject: mem: Ignore clean requests in the abstract memory
......................................................................
mem: Ignore clean requests in the abstract memory
Systems with atomic cores and the fastmem option enabled bypass the
whole memory system and access the abstract memory directly. Cache
maintenance operations which would be normally handled before the
point of unification/coherence should be ignored by the abstract
memory.
Change-Id: I696cdd158222e5fd67f670cddbcf2efbbfd5eca4
Reviewed-by: Curtis Dunham <[email protected]>
Reviewed-by: Andreas Sandberg <[email protected]>
---
M src/mem/abstract_mem.cc
1 file changed, 6 insertions(+), 7 deletions(-)
diff --git a/src/mem/abstract_mem.cc b/src/mem/abstract_mem.cc
index e05296c..5b461c9 100644
--- a/src/mem/abstract_mem.cc
+++ b/src/mem/abstract_mem.cc
@@ -396,13 +396,12 @@
bytesRead[pkt->req->masterId()] += pkt->getSize();
if (pkt->req->isInstFetch())
bytesInstRead[pkt->req->masterId()] += pkt->getSize();
- } else if (pkt->isInvalidate()) {
+ } else if (pkt->isInvalidate() || pkt->isClean()) {
+ assert(!pkt->isWrite());
+ // in a fastmem system invalidating and/or cleaning packets
+ // can be seen due to cache maintenance requests
+
// no need to do anything
- // this clause is intentionally before the write clause: the only
- // transaction that is both a write and an invalidate is
- // WriteInvalidate, and for the sake of consistency, it does not
- // write to memory. in a cacheless system, there are no WriteInv's
- // because the Write -> WriteInvalidate rewrite happens in the
cache.
} else if (pkt->isWrite()) {
if (writeOK(pkt)) {
if (pmemAddr) {
@@ -416,7 +415,7 @@
bytesWritten[pkt->req->masterId()] += pkt->getSize();
}
} else {
- panic("unimplemented");
+ panic("Unexpected packet %s", pkt->print());
}
if (pkt->needsResponse()) {
--
To view, visit https://gem5-review.googlesource.com/5054
To unsubscribe, or for help writing mail filters, visit
https://gem5-review.googlesource.com/settings
Gerrit-Project: public/gem5
Gerrit-Branch: master
Gerrit-MessageType: newchange
Gerrit-Change-Id: I696cdd158222e5fd67f670cddbcf2efbbfd5eca4
Gerrit-Change-Number: 5054
Gerrit-PatchSet: 1
Gerrit-Owner: Nikos Nikoleris <[email protected]>
Gerrit-Reviewer: Andreas Sandberg <[email protected]>
Gerrit-Reviewer: Curtis Dunham <[email protected]>
_______________________________________________
gem5-dev mailing list
[email protected]
http://m5sim.org/mailman/listinfo/gem5-dev
