- Remember the fact that an object has been added to the

- Remember the fact that an object has been added to the
  store buffer by a bit in the header.
- This bit can be used to filter out redundant additions
  into the store buffer making the dedup sets not needed.
- Remove the need for a HashSet when remembering old to
  new references.

Committed: https://code.google.com/p/dart/source/detail?r=22393

[Ivan Posva, 2013-04-26 21:31:44]

This CL still needs some cleanups, but I would like to get initial feedback
about the approach.

Still missing:
- MIPS implementation
- Removal of dedup sets.
- Removal of commented out code.

Thanks,
-Ivan

[siva, 2013-04-27 00:49:37]

https://codereview.chromium.org/14307013/diff/1/runtime/vm/raw_object.h
File runtime/vm/raw_object.h (right):

https://codereview.chromium.org/14307013/diff/1/runtime/vm/raw_object.h#newco...
runtime/vm/raw_object.h:226: kWatchedBit = 4,
The WatchedBit is only used during GC right so we could swap it with kFreeBit.

[Vyacheslav Egorov (Google), 2013-04-27 19:34:46]

I am not sure I understand why all changes with FreeBit and friends?

https://codereview.chromium.org/14307013/diff/1/runtime/vm/stub_code_ia32.cc
File runtime/vm/stub_code_ia32.cc (right):

https://codereview.chromium.org/14307013/diff/1/runtime/vm/stub_code_ia32.cc#...
runtime/vm/stub_code_ia32.cc:967: __ movl(ECX, FieldAddress(EAX,
Object::tags_offset()));
I wonder if doing testl with memory operand before pushing EDX/ECX would save us
something noticeable. (it would save pushing and poping definitely)

[Ivan Posva, 2013-05-03 20:59:58]

Committed patchset #3 manually as r22393 (presubmit successful).

[Ivan Posva, 2013-05-03 21:01:37]

https://codereview.chromium.org/14307013/diff/1/runtime/vm/stub_code_ia32.cc
File runtime/vm/stub_code_ia32.cc (right):

https://codereview.chromium.org/14307013/diff/1/runtime/vm/stub_code_ia32.cc#...
runtime/vm/stub_code_ia32.cc:967: __ movl(ECX, FieldAddress(EAX,
Object::tags_offset()));
On 2013/04/27 19:34:46, Vyacheslav Egorov (Google) wrote:
> I wonder if doing testl with memory operand before pushing EDX/ECX would save
us
> something noticeable. (it would save pushing and poping definitely)

Different CL.