Inline functions with different contexts in the optimizing code generator.

Inline functions with different contexts in the optimizing code generator.

Add HInlinedContext to Hydrogen, and use the context from an environment HValue in all places, not from the context slot in the frame.

BUG=
TEST=

[William Hesse, 2011-09-15 09:39:01]

Here is the change.  We can either enable inlining of builtins with this change,
or leave builtins as not inlinable, and just inline user functions with context
changes.

[Kevin Millikin (Chromium), 2011-09-15 11:17:24]

A couple of quick drive bys.

It might be simpler to just use an HConstant than introduce a hydrogen
instruction for this (unless we're getting some value out of it).

I also wonder what it would cost just to always use the context of the callee
rather than trying to catch cases that we can use the context of the caller.

http://codereview.chromium.org/7739018/diff/3001/src/arm/lithium-arm.cc
File src/arm/lithium-arm.cc (right):

http://codereview.chromium.org/7739018/diff/3001/src/arm/lithium-arm.cc#newco...
src/arm/lithium-arm.cc:1117: return instr->HasNoUses() ? NULL :
DefineAsRegister(new LInlinedContext);
Just UNREACHABLE, please, no untested dead code.

Otherwise it's too tempting to think deleting UNREACHABLE() will make everything
work.  That might not be true in three or six months.

http://codereview.chromium.org/7739018/diff/3001/src/arm/lithium-codegen-arm.cc
File src/arm/lithium-codegen-arm.cc (right):

http://codereview.chromium.org/7739018/diff/3001/src/arm/lithium-codegen-arm....
src/arm/lithium-codegen-arm.cc:2754: Register result =
ToRegister(instr->result());
This is dead code?

http://codereview.chromium.org/7739018/diff/3001/src/ia32/lithium-codegen-ia3...
File src/ia32/lithium-codegen-ia32.cc (right):

http://codereview.chromium.org/7739018/diff/3001/src/ia32/lithium-codegen-ia3...
src/ia32/lithium-codegen-ia32.cc:2541: __ mov(result, instr->closure());
Does something like

__ mov(result, Handle<Context>(instr->closure()->context()))

work?