Author: Armin Rigo <ar...@tunes.org>
Branch: stackroot-speedup-2
Changeset: r75715:41114edfebad
Date: 2015-02-04 21:39 +0100
http://bitbucket.org/pypy/pypy/changeset/41114edfebad/
Log: I think this was a bug, and this is a fix. Theoretical thinking
though.
diff --git a/rpython/memory/gc/incminimark.py b/rpython/memory/gc/incminimark.py
--- a/rpython/memory/gc/incminimark.py
+++ b/rpython/memory/gc/incminimark.py
@@ -375,6 +375,10 @@
# the nursery.
self.pinned_objects_in_nursery = 0
#
+ # This flag is set if the previous minor collection found at least
+ # one pinned object alive.
+ self.any_pinned_object_kept = False
+ #
# Keeps track of old objects pointing to pinned objects. These objects
# must be traced every minor collection. Without tracing them the
# referenced pinned object wouldn't be visited and therefore collected.
@@ -1489,8 +1493,9 @@
# The following counter keeps track of alive and pinned young objects
# inside the nursery. We reset it here and increace it in
# '_trace_drag_out()'.
- any_pinned_objects_in_nursery = (self.pinned_objects_in_nursery > 0)
+ any_pinned_object_from_earlier = self.any_pinned_object_kept
self.pinned_objects_in_nursery = 0
+ self.any_pinned_object_kept = False
#
# Before everything else, remove from 'old_objects_pointing_to_young'
# the young arrays.
@@ -1514,7 +1519,7 @@
# are copied out or flagged. They are also added to the list
# 'old_objects_pointing_to_young'.
self.nursery_surviving_size = 0
- self.collect_roots_in_nursery(any_pinned_objects_in_nursery)
+ self.collect_roots_in_nursery(any_pinned_object_from_earlier)
#
# visit all objects that are known for pointing to pinned
# objects. This way we populate 'surviving_pinned_objects'
@@ -1650,7 +1655,7 @@
def _visit_old_objects_pointing_to_pinned(self, obj, ignore):
self.trace(obj, self._trace_drag_out, obj)
- def collect_roots_in_nursery(self, any_pinned_objects_in_nursery):
+ def collect_roots_in_nursery(self, any_pinned_object_from_earlier):
# we don't need to trace prebuilt GcStructs during a minor collect:
# if a prebuilt GcStruct contains a pointer to a young object,
# then the write_barrier must have ensured that the prebuilt
@@ -1661,10 +1666,12 @@
else:
callback = IncrementalMiniMarkGC._trace_drag_out1
#
- # Note a subtlety: if the nursery contains pinned objects right
- # now, we can't use the "is_minor=True" optimization. We really
- # need to walk the complete stack to be sure we still see them.
- use_jit_frame_stoppers = not any_pinned_objects_in_nursery
+ # Note a subtlety: if the nursery contains pinned objects "from
+ # earlier", i.e. created earlier than the previous minor
+ # collection, then we can't use the "is_minor=True" optimization.
+ # We really need to walk the complete stack to be sure we still
+ # see them.
+ use_jit_frame_stoppers = not any_pinned_object_from_earlier
#
self.root_walker.walk_roots(
callback, # stack roots
@@ -1852,6 +1859,7 @@
self.surviving_pinned_objects.append(
llarena.getfakearenaaddress(obj - size_gc_header))
self.pinned_objects_in_nursery += 1
+ self.any_pinned_object_kept = True
return
else:
# First visit to an object that has already a shadow.
_______________________________________________
pypy-commit mailing list
pypy-commit@python.org
https://mail.python.org/mailman/listinfo/pypy-commit
