Improve representation inference

Improve representation inference

- Extend HValue interface to allow splitting observed input representation (comming from type feedback) from required input representation (dictated by instruction itself). Currently all instructions except for bitwise binary operations have this representations match. For bitwise binary operations hydrogen builder unconditionaly forces Integer32 representation for those operations that have Double type feedback. Thus causing representation inference to incorrectly count such uses as Integer32 instead of Double. This change also prepares for more fine grained type feedback for inputs of binary operations.

- For phies that are not convertable to Integer32 discard direct and indirect use count of Integer32 type to avoid propagation of these uses to connected phies. 

R=jkummerow@chromium.org
BUG=v8:2096
Committed: https://code.google.com/p/v8/source/detail?r=11737

[Jakob Kummerow, 2012-06-07 15:38:05]

LGTM with nits.

Medium-term, I want to move special handling of observed input representations
from HBitwiseBinaryOperation to HBinaryOperation, but this is a good first step.

https://chromiumcodereview.appspot.com/10540049/diff/3001/src/hydrogen.cc
File src/hydrogen.cc (right):

https://chromiumcodereview.appspot.com/10540049/diff/3001/src/hydrogen.cc#new...
src/hydrogen.cc:2217: //         push information about values which can't be
converted to integer
nit: indentation

https://chromiumcodereview.appspot.com/10540049/diff/3001/src/hydrogen.cc#new...
src/hydrogen.cc:2233: //         sum up the non-phi use counts of all connected
phis.
nit: indentation

[Vyacheslav Egorov (Google), 2012-06-08 09:23:10]

Thanks for the review! Landed.

Of course storage for observed input representation should be moved to HBinaryOp
when there will be something to store in it. I did not want to waste space in
the IR.