summaryrefslogtreecommitdiff
path: root/src/dev/x86
diff options
context:
space:
mode:
authorAndreas Sandberg <Andreas.Sandberg@ARM.com>2013-02-15 17:40:09 -0500
committerAndreas Sandberg <Andreas.Sandberg@ARM.com>2013-02-15 17:40:09 -0500
commitb904bd5437ead0dfc2c4c0977f3d29d63299c601 (patch)
treef7d324fe5c806338534c5e41e9b251d2e62a3132 /src/dev/x86
parent1eec115c31395e2819c073a1859d75eb5933dac2 (diff)
downloadgem5-b904bd5437ead0dfc2c4c0977f3d29d63299c601.tar.xz
sim: Add a system-global option to bypass caches
Virtualized CPUs and the fastmem mode of the atomic CPU require direct access to physical memory. We currently require caches to be disabled when using them to prevent chaos. This is not ideal when switching between hardware virutalized CPUs and other CPU models as it would require a configuration change on each switch. This changeset introduces a new version of the atomic memory mode, 'atomic_noncaching', where memory accesses are inserted into the memory system as atomic accesses, but bypass caches. To make memory mode tests cleaner, the following methods are added to the System class: * isAtomicMode() -- True if the memory mode is 'atomic' or 'direct'. * isTimingMode() -- True if the memory mode is 'timing'. * bypassCaches() -- True if caches should be bypassed. The old getMemoryMode() and setMemoryMode() methods should never be used from the C++ world anymore.
Diffstat (limited to 'src/dev/x86')
-rw-r--r--src/dev/x86/i82094aa.cc3
1 files changed, 1 insertions, 2 deletions
diff --git a/src/dev/x86/i82094aa.cc b/src/dev/x86/i82094aa.cc
index 54824c778..0692718bf 100644
--- a/src/dev/x86/i82094aa.cc
+++ b/src/dev/x86/i82094aa.cc
@@ -222,8 +222,7 @@ X86ISA::I82094AA::signalInterrupt(int line)
apics.push_back(selected);
}
}
- intMasterPort.sendMessage(apics, message,
- sys->getMemoryMode() == Enums::timing);
+ intMasterPort.sendMessage(apics, message, sys->isTimingMode());
}
}