On 6/17/13 3:13 PM, [email protected] wrote:
Updated Branches:
refs/heads/master 1cc22d69f -> 052da6724
TS-1684 - Added more ProxyAllocators for faster and local memory
allocation. Have seen a doubling in performance depending on the benchmark.
Details in the ticket.
Two things:
1) I added in the missing ProxyAllocator.cc file from the source code. I
don't know if that's what the intent was, cause we have the templates
(and see 2).
2) It seems one of the templates in I_ProxyAllocator.h isn't used any
more? See patch below, do we need that?
Cheers,
-- Leif
diff --git a/iocore/eventsystem/I_ProxyAllocator.h
b/iocore/eventsystem/I_ProxyAllocator.h
index a221d63..c578f7d 100644
--- a/iocore/eventsystem/I_ProxyAllocator.h
+++ b/iocore/eventsystem/I_ProxyAllocator.h
@@ -69,18 +69,6 @@ template<class C> inline C * thread_alloc_init(ClassAllocator<C>
&a, ProxyAlloca
return a.alloc();
}
-template<class C> inline void
-thread_freeup(ClassAllocator<C> &a, ProxyAllocator & l)
-{
- while (l.freelist) {
- C *v = (C *) l.freelist;
- l.freelist = *(C **) l.freelist;
- l.allocated--;
- a.free(v); // we could use a bulk free here
- }
- ink_assert(!l.allocated);
-}
-
void* thread_alloc(Allocator &a, ProxyAllocator &l);
void thread_freeup(Allocator &a, ProxyAllocator &l);