Implement IC for storing to dictionary case objects....

Implement IC for storing to dictionary case objects.

The IC stub is completely generic, so there will only be one such stub
in the system.

Added a new overloaded version of the macro assembler RecordWrite
method for cases where we have the address we store to computed up
front.


Committed: http://code.google.com/p/v8/source/detail?r=4991

[Erik Corry, 2010-06-30 10:35:15]

LGTM.

I think it belongs in the documentation of the RecordWrite macros in the .h
files whether or not they filter out Smis.

http://codereview.chromium.org/2804029/diff/1/4
File src/arm/macro-assembler-arm.cc (right):

http://codereview.chromium.org/2804029/diff/1/4#newcode390
src/arm/macro-assembler-arm.cc:390: Register scratch) {
In the .h file the register arguments are called object, address and value. 
Here they are called object, address and scratch.

http://codereview.chromium.org/2804029/diff/1/8
File src/ia32/ic-ia32.cc (right):

http://codereview.chromium.org/2804029/diff/1/8#newcode154
src/ia32/ic-ia32.cc:154: // storage. This function may fail to load a property
eventhough it is
eventhough -> even though

http://codereview.chromium.org/2804029/diff/1/16
File src/x64/ic-x64.cc (right):

http://codereview.chromium.org/2804029/diff/1/16#newcode210
src/x64/ic-x64.cc:210: // storage. This function may fail to store a property
eventhough it
eventhough

http://codereview.chromium.org/2804029/diff/1/16#newcode221
src/x64/ic-x64.cc:221: Register r0,
I find the names r0 and r1 very confusing considering that r12 and r13 are real
registers, but these are arguments.

[Mads Ager (chromium), 2010-06-30 12:27:08]

http://codereview.chromium.org/2804029/diff/1/4
File src/arm/macro-assembler-arm.cc (right):

http://codereview.chromium.org/2804029/diff/1/4#newcode390
src/arm/macro-assembler-arm.cc:390: Register scratch) {
On 2010/06/30 10:35:15, Erik Corry wrote:
> In the .h file the register arguments are called object, address and value. 
> Here they are called object, address and scratch.

Updated the header file to call it scratch as well. We do not filter out smi
stores on ARM which is why it is just a scratch register and not necessarily the
value being stored.

I am adding a comment about filtering out smis to the header files as well.

http://codereview.chromium.org/2804029/diff/1/8
File src/ia32/ic-ia32.cc (right):

http://codereview.chromium.org/2804029/diff/1/8#newcode154
src/ia32/ic-ia32.cc:154: // storage. This function may fail to load a property
eventhough it is
On 2010/06/30 10:35:15, Erik Corry wrote:
> eventhough -> even though

Done.

http://codereview.chromium.org/2804029/diff/1/16
File src/x64/ic-x64.cc (right):

http://codereview.chromium.org/2804029/diff/1/16#newcode210
src/x64/ic-x64.cc:210: // storage. This function may fail to store a property
eventhough it
On 2010/06/30 10:35:15, Erik Corry wrote:
> eventhough
> 

Done.

http://codereview.chromium.org/2804029/diff/1/16#newcode221
src/x64/ic-x64.cc:221: Register r0,
On 2010/06/30 10:35:15, Erik Corry wrote:
> I find the names r0 and r1 very confusing considering that r12 and r13 are
real
> registers, but these are arguments.

Agreed. Changed to scratch0 and scratch1.