ARM: Don't require the receiver on the stack for load IC...

ARM: Don't require the receiver on the stack for load IC

Previously the receier was passed in both r0 and on the stack for a load IC. With this change the receiver is in r0 only.

Committed: http://code.google.com/p/v8/source/detail?r=4681

[Erik Corry, 2010-05-19 09:35:24]

LGTM

http://codereview.chromium.org/2119007/diff/1/2
File src/arm/codegen-arm.cc (right):

http://codereview.chromium.org/2119007/diff/1/2#newcode5455
src/arm/codegen-arm.cc:5455: __ Swap(r0, r1, ip);
We are always spilled here so there's no reason to do a 3-instruction swap.  
You could just do __ Move(r0, receiver) which includes the if().

http://codereview.chromium.org/2119007/diff/1/4
File src/arm/stub-cache-arm.cc (right):

http://codereview.chromium.org/2119007/diff/1/4#newcode1620
src/arm/stub-cache-arm.cc:1620: //  -- r0    : name
Surely r0 is receiver and r2 is name?

http://codereview.chromium.org/2119007/diff/1/5
File src/arm/virtual-frame-arm.cc (right):

http://codereview.chromium.org/2119007/diff/1/5#newcode524
src/arm/virtual-frame-arm.cc:524: // r0 and r1 contains the same value. Prefer a
state with r0 holding TOS.
Does it lint?

[Søren Thygesen Gjesse, 2010-05-19 10:37:37]

http://codereview.chromium.org/2119007/diff/1/2
File src/arm/codegen-arm.cc (right):

http://codereview.chromium.org/2119007/diff/1/2#newcode5455
src/arm/codegen-arm.cc:5455: __ Swap(r0, r1, ip);
On 2010/05/19 09:35:25, Erik Corry wrote:
> We are always spilled here so there's no reason to do a 3-instruction swap.  
> You could just do __ Move(r0, receiver) which includes the if().

Done.

http://codereview.chromium.org/2119007/diff/1/4
File src/arm/stub-cache-arm.cc (right):

http://codereview.chromium.org/2119007/diff/1/4#newcode1620
src/arm/stub-cache-arm.cc:1620: //  -- r0    : name
On 2010/05/19 09:35:25, Erik Corry wrote:
> Surely r0 is receiver and r2 is name?

r0 is receiver, and no name in any register.

http://codereview.chromium.org/2119007/diff/1/5
File src/arm/virtual-frame-arm.cc (right):

http://codereview.chromium.org/2119007/diff/1/5#newcode524
src/arm/virtual-frame-arm.cc:524: // r0 and r1 contains the same value. Prefer a
state with r0 holding TOS.
On 2010/05/19 09:35:25, Erik Corry wrote:
> Does it lint?

Almost, but not quite. Fixed.