diff options
author | Andreas Sandberg <Andreas.Sandberg@arm.com> | 2012-11-02 11:32:02 -0500 |
---|---|---|
committer | Andreas Sandberg <Andreas.Sandberg@arm.com> | 2012-11-02 11:32:02 -0500 |
commit | dc01535c7e800189c4cb46cab470f4b958be94c7 (patch) | |
tree | 031c5f3b6d8c3e43f493e2efbca413b32359f134 /src/python/m5/simulate.py | |
parent | 196397fea4e25f097b4a1624ee16fbbbc1c45b64 (diff) | |
download | gem5-dc01535c7e800189c4cb46cab470f4b958be94c7.tar.xz |
python: Rename doDrain()->drain() and make it do the right thing
There is no point in exporting the old drain() method in
Simulate.py. It should only be used internally by doDrain(). This
patch moves the old drain() method into doDrain() and renames
doDrain() to drain().
Diffstat (limited to 'src/python/m5/simulate.py')
-rw-r--r-- | src/python/m5/simulate.py | 47 |
1 files changed, 24 insertions, 23 deletions
diff --git a/src/python/m5/simulate.py b/src/python/m5/simulate.py index 4c6076581..ca2cc4a9c 100644 --- a/src/python/m5/simulate.py +++ b/src/python/m5/simulate.py @@ -157,28 +157,29 @@ atexit.register(stats.dump) # register our C++ exit callback function with Python atexit.register(internal.core.doExitCleanup) -# This loops until all objects have been fully drained. -def doDrain(root): - all_drained = drain(root) - while (not all_drained): - all_drained = drain(root) - -# Tries to drain all objects. Draining might not be completed unless -# all objects return that they are drained on the first call. This is -# because as objects drain they may cause other objects to no longer -# be drained. +# Drain the system in preparation of a checkpoint or memory mode +# switch. def drain(root): - all_drained = False - dm = internal.drain.createDrainManager() - unready_objs = sum(obj.drain(dm) for obj in root.descendants()) - # If we've got some objects that can't drain immediately, then simulate - if unready_objs > 0: - dm.setCount(unready_objs) - simulate() - else: - all_drained = True - internal.drain.cleanupDrainManager(dm) - return all_drained + # Try to drain all objects. Draining might not be completed unless + # all objects return that they are drained on the first call. This + # is because as objects drain they may cause other objects to no + # longer be drained. + def _drain(): + all_drained = False + dm = internal.drain.createDrainManager() + unready_objs = sum(obj.drain(dm) for obj in root.descendants()) + # If we've got some objects that can't drain immediately, then simulate + if unready_objs > 0: + dm.setCount(unready_objs) + simulate() + else: + all_drained = True + internal.drain.cleanupDrainManager(dm) + return all_drained + + all_drained = _drain() + while (not all_drained): + all_drained = _drain() def resume(root): for obj in root.descendants(): obj.drainResume() @@ -187,7 +188,7 @@ def checkpoint(dir): root = objects.Root.getInstance() if not isinstance(root, objects.Root): raise TypeError, "Checkpoint must be called on a root object." - doDrain(root) + drain(root) print "Writing checkpoint" internal.core.serializeAll(dir) resume(root) @@ -197,7 +198,7 @@ def changeMemoryMode(system, mode): raise TypeError, "Parameter of type '%s'. Must be type %s or %s." % \ (type(system), objects.Root, objects.System) if system.getMemoryMode() != mode: - doDrain(system) + drain(system) system.setMemoryMode(mode) else: print "System already in target mode. Memory mode unchanged." |