Allow esi to be an allocatable register on IA32.

Allow esi to be an allocatable register on IA32.

Make esi available to the register allocator rather than dedicating it
permanently to the context.

The context is still passed in register esi to JavaScript and to the runtime
as part of the calling convention.  Because some stubs might end up calling
JS or the runtime, it is also conservatively passed to stubs.

Roughly half the calls have been modified to use the context as an input
value in fixed register esi.  The other half are marked as calls or deferred
code so esi is spilled and can be explicitly set.

It is no longer necessary to restore the context to esi after a call that
might change it.

Committed: http://code.google.com/p/v8/source/detail?r=6713

[fschneider, 2011-02-09 13:41:53]

http://codereview.chromium.org/6452001/diff/1/src/hydrogen-instructions.h
File src/hydrogen-instructions.h (right):

http://codereview.chromium.org/6452001/diff/1/src/hydrogen-instructions.h#new...
src/hydrogen-instructions.h:1181: // (see HCall).
Maybe it would be less effort to land the refactoring of PushArguments first -
all this code will be greatly cleaned up afterwards.

http://codereview.chromium.org/6452001/diff/1/src/ia32/lithium-codegen-ia32.cc
File src/ia32/lithium-codegen-ia32.cc (right):

http://codereview.chromium.org/6452001/diff/1/src/ia32/lithium-codegen-ia32.c...
src/ia32/lithium-codegen-ia32.cc:377: if (!adjusted) {
Is there a convenient way to assert that 

"!adjusted || there is a fixed use of the context in esi"?

http://codereview.chromium.org/6452001/diff/1/src/ia32/lithium-ia32.cc
File src/ia32/lithium-ia32.cc (right):

http://codereview.chromium.org/6452001/diff/1/src/ia32/lithium-ia32.cc#newcod...
src/ia32/lithium-ia32.cc:1237: LOperand* context = UseFixed(instr->context(),
esi);
I'm not sure if it is worth mentioning the context as input operand. It seems
that we have to reload it from the stack before most calls anyway. Is there a
big performance impact of always reloading it from the stack before a call as
part of CallCode or CallRuntime?

[fschneider, 2011-02-09 15:23:31]

LGTM as intermediate step and let's land the refactoring of the hydrogen call
instructions code asap.

http://codereview.chromium.org/6452001/diff/1/src/ia32/lithium-codegen-ia32.cc
File src/ia32/lithium-codegen-ia32.cc (right):

http://codereview.chromium.org/6452001/diff/1/src/ia32/lithium-codegen-ia32.c...
src/ia32/lithium-codegen-ia32.cc:2231: __ mov(esi, Operand(ebp,
StandardFrameConstants::kContextOffset));
Should this function for now also take an additional argument  like CallCode and
CallRuntime?