Deal with filler object map pointers in the ool constant pool correctly.

Deal with filler object map pointers in the ool constant pool correctly.

This CL updates
StoreBuffer::FindPointersToNewSpaceOnPage such that it skips  constant pool array objects.  Constant Pool Arrays should never have pointers to the new space and might contain pointers to the FreeSpace map (e.g., due to code
being generated which needs to do a map-check on a FreeSpace object) which would incorrectly treated as a filler object if processed by FindPointersToNewSpaceOnPage().

R=hpayer@chromium.org

Committed: https://code.google.com/p/v8/source/detail?r=19771

[rmcilroy, 2014-02-27 18:36:53]

PTAL.

[Hannes Payer (out of office), 2014-03-03 13:56:13]

https://codereview.chromium.org/183553003/diff/1/src/store-buffer.cc
File src/store-buffer.cc (right):

https://codereview.chromium.org/183553003/diff/1/src/store-buffer.cc#newcode534
src/store-buffer.cc:534: if (o == constant_pool_array_map) {
I don't think a constant pool can contain old-to-new references, right? In that
case you could just skip the constant pool right away.

[rmcilroy, 2014-03-03 15:07:47]

https://codereview.chromium.org/183553003/diff/1/src/store-buffer.cc
File src/store-buffer.cc (right):

https://codereview.chromium.org/183553003/diff/1/src/store-buffer.cc#newcode534
src/store-buffer.cc:534: if (o == constant_pool_array_map) {
On 2014/03/03 13:56:13, Hannes Payer wrote:
> I don't think a constant pool can contain old-to-new references, right? In
that
> case you could just skip the constant pool right away.

You are right, skipping it entirely probably simpler. The one thing I was
considering was that this is called from VerifyPointers, so we might want to
ensure that we hadn't accidentally added new-space pointers into a
ConstantPoolArray, but since VerifyPointers only has a dummy ObjectSlotCallback
this probably isn't worth while anyway.

I've uploaded a change to just skip constant pool arrays.

[Hannes Payer (out of office), 2014-03-03 15:18:37]

Cool, lgtm with one nit

https://codereview.chromium.org/183553003/diff/20001/src/store-buffer.cc
File src/store-buffer.cc (right):

https://codereview.chromium.org/183553003/diff/20001/src/store-buffer.cc#newc...
src/store-buffer.cc:503: // looking for pointers into new space.  Within the
pages there may be dead
Please update this comment as well.

[rmcilroy, 2014-03-03 15:51:16]

Thanks for the review!

https://codereview.chromium.org/183553003/diff/20001/src/store-buffer.cc
File src/store-buffer.cc (right):

https://codereview.chromium.org/183553003/diff/20001/src/store-buffer.cc#newc...
src/store-buffer.cc:503: // looking for pointers into new space.  Within the
pages there may be dead
On 2014/03/03 15:18:37, Hannes Payer wrote:
> Please update this comment as well.

Done.

[rmcilroy, 2014-03-10 18:44:25]

Committed patchset #3 manually as r19771 (presubmit successful).