Folding of bounded dynamic size allocations with const size allocations.

Folding of bounded dynamic size allocations with const size allocations.

R=hpayer@chromium.org

Committed: https://code.google.com/p/v8/source/detail?r=21675

[Igor Sheludko, 2014-06-03 08:46:40]

PTAL

[Hannes Payer (out of office), 2014-06-04 09:44:29]

https://codereview.chromium.org/301973014/diff/1/src/hydrogen-instructions.cc
File src/hydrogen-instructions.cc (right):

https://codereview.chromium.org/301973014/diff/1/src/hydrogen-instructions.cc...
src/hydrogen-instructions.cc:3837: HConstant* new_dominator_size_upper_bound_c =
I don't like the _c, can we have a proper name?

Moreover, the size variable names are very confusing and inconsistent. Can we
clean that up?

I think the start point is good:
dominator_size is the dominator size instruction, the new dominator size
instruction should be called new_dominator_size. When we take the int value of a
constant size instruction it could be called dominator_size_value...

https://codereview.chromium.org/301973014/diff/1/src/hydrogen-instructions.cc...
src/hydrogen-instructions.cc:4047: clear_next_map->InsertAfter(this);
Can you move
clear_next_map->ClearAllSideEffects();
clear_next_map->InsertAfter(this);
after the else case?

[Igor Sheludko, 2014-06-04 10:07:41]

https://codereview.chromium.org/301973014/diff/1/src/hydrogen-instructions.cc
File src/hydrogen-instructions.cc (right):

https://codereview.chromium.org/301973014/diff/1/src/hydrogen-instructions.cc...
src/hydrogen-instructions.cc:3837: HConstant* new_dominator_size_upper_bound_c =
On 2014/06/04 09:44:29, Hannes Payer wrote:
> I don't like the _c, can we have a proper name?
> 
> Moreover, the size variable names are very confusing and inconsistent. Can we
> clean that up?
> 
> I think the start point is good:
> dominator_size is the dominator size instruction, the new dominator size
> instruction should be called new_dominator_size. When we take the int value of
a
> constant size instruction it could be called dominator_size_value...

Done.

https://codereview.chromium.org/301973014/diff/1/src/hydrogen-instructions.cc...
src/hydrogen-instructions.cc:4047: clear_next_map->InsertAfter(this);
On 2014/06/04 09:44:29, Hannes Payer wrote:
> Can you move
> clear_next_map->ClearAllSideEffects();
> clear_next_map->InsertAfter(this);
> after the else case?

Done.

[Hannes Payer (out of office), 2014-06-04 11:05:52]

lgtm

[Igor Sheludko, 2014-06-04 13:39:10]

Committed patchset #2 manually as r21675 (presubmit successful).