Author: Armin Rigo <ar...@tunes.org>
Branch:
Changeset: r60:39d1f410f3ce
Date: 2013-06-03 10:59 +0200
http://bitbucket.org/pypy/stmgc/changeset/39d1f410f3ce/
Log: Saved the day by moving the cache check inline in 4 cpu
instructions.
diff --git a/c3/doc-stmgc.txt b/c3/doc-stmgc.txt
--- a/c3/doc-stmgc.txt
+++ b/c3/doc-stmgc.txt
@@ -158,18 +158,25 @@
references an object that is not outdated already; and it needs to
record the pointer in the "read set" of the current transaction.
-The first check is easy, and can be implemented by checking a flag in
-the header of the copy. In all the common cases, this flag is not set,
-and no actual call needs to be done.
+The first check is easy, and could be implemented by checking a flag in
+the header of the copy. But the recording in the read set is a bit more
+annoying. We need to maintain a thread-local *set* of all accessed
+objects, but we don't care about the order or recording the occasional
+duplicate. Moreover we don't need to record the private objects; but we
+do need all other protected objects, as well as public objects. The
+best approach is probably to have a quick check "is it definitely
+recorded already?" inline, and do the call if the check fails. It needs
+careful design to be done in only a few CPU instructions, but it should
+be possible.
-The recording in the read set is a bit more annoying. We need to
-maintain a thread-local *set* of all accessed objects, but we don't care
-about the order or recording the occasional duplicate. Moreover we
-don't need to record the private objects; but we do need all other
-protected objects, as well as public objects. The best approach is
-probably to have a quick check "is it definitely recorded already?"
-inline, and do the call if the check fails. It needs careful design to
-be done in only a few CPU instructions, but it should be possible.
+(Code like this compiles to 4 instructions in the fast path:
+
+ __thread char *read_barrier_cache; /* thread-local cache of 64KB */
+
+ if (__builtin_expect(*(gcptr *)(read_barrier_cache + (((long)x) & 65535))
+ != x, 0))
+ x = call_read_barrier(x);
+)
The case of the write barrier is similar to the first half of the read
barrier, but differs in the check we need to do. We need to do a call
_______________________________________________
pypy-commit mailing list
pypy-commit@python.org
http://mail.python.org/mailman/listinfo/pypy-commit
