changeset b6b4d41ba9b9 in /z/repo/gem5
details: http://repo.gem5.org/gem5?cmd=changeset;node=b6b4d41ba9b9
description:
cache: don't allow dirty data in the i-cache
removes the optimization that forwards an exclusive copy to a requester
on a
read, only for the i-cache. this optimization isn't necessary because we
typically won't be writing to the i-cache.
diffstat:
src/mem/cache/cache_impl.hh | 6 ++++--
1 files changed, 4 insertions(+), 2 deletions(-)
diffs (23 lines):
diff -r 8423aa8c2216 -r b6b4d41ba9b9 src/mem/cache/cache_impl.hh
--- a/src/mem/cache/cache_impl.hh Fri Jul 27 16:08:04 2012 -0400
+++ b/src/mem/cache/cache_impl.hh Fri Jul 27 16:08:04 2012 -0400
@@ -172,7 +172,7 @@
// on ReadExReq we give up our copy unconditionally
tags->invalidateBlk(blk);
} else if (blk->isWritable() && !pending_downgrade
- && !pkt->sharedAsserted()) {
+ && !pkt->sharedAsserted() && !pkt->req->isInstFetch()) {
// we can give the requester an exclusive copy (by not
// asserting shared line) on a read request if:
// - we have an exclusive copy at this level (& below)
@@ -180,7 +180,9 @@
// signaling another read request
// - no other cache above has a copy (otherwise it
// would have asseretd shared line on request)
-
+ // - we are not satisfying an instruction fetch (this
+ // prevents dirty data in the i-cache)
+
if (blk->isDirty()) {
// special considerations if we're owner:
if (!deferred_response && !isTopLevel) {
_______________________________________________
gem5-dev mailing list
[email protected]
http://m5sim.org/mailman/listinfo/gem5-dev
