ARM: Pass arguments to keyed store IC in registers...

ARM: Pass arguments to keyed store IC in registers

The calling convention for keyed store IC on ARM is changed to receive the arguments value, key and receiver in registers r0, r1 and r2 instead of on the stack. When calling keyed store IC with a virtual frame the arguments are passed through the virtual frame and consumed by the call.

Changed the register usage in the IC code to postpone spilling the registers holding value, key and receiver to the stack until making  a call into the runtime system.

Runs all the tests with: 

  --special-command="@ --nofull-compiler"
  --special-command="@ --always-full-compiler"
  --special-command="@ --noenable-vfp3"

Committed: http://code.google.com/p/v8/source/detail?r=4668

[Erik Corry, 2010-05-17 19:15:14]

Looks good to me.  All those register changes made steam come out of my ears,
but it's nice to know you ran all the tests with all those flags.

http://codereview.chromium.org/2116003/diff/10001/11005
File src/arm/ic-arm.cc (right):

http://codereview.chromium.org/2116003/diff/10001/11005#newcode1332
src/arm/ic-arm.cc:1332: // Condition code from comparing key and array length is
still available.
Thanks for that comment!

http://codereview.chromium.org/2116003/diff/10001/11005#newcode1341
src/arm/ic-arm.cc:1341: __ add(r4, key, Operand(1 << (kSmiTagSize +
kSmiShiftSize)));
Operand(Smi::FromInt(1))

http://codereview.chromium.org/2116003/diff/10001/11005#newcode1368
src/arm/ic-arm.cc:1368: __ b(eq, &exit);
I think you might as well do a Ret(eq) here.

http://codereview.chromium.org/2116003/diff/10001/11005#newcode1555
src/arm/ic-arm.cc:1555: // r3: base pointer of external storage
Full stop here and lots of other places.

http://codereview.chromium.org/2116003/diff/10001/11007
File src/arm/virtual-frame-arm.cc (right):

http://codereview.chromium.org/2116003/diff/10001/11007#newcode344
src/arm/virtual-frame-arm.cc:344: EmitPop(r2);
This pop and the spillall should be swapped, since we plan to register allocate
r2 soon.

[Søren Thygesen Gjesse, 2010-05-18 06:56:52]

http://codereview.chromium.org/2116003/diff/10001/11005
File src/arm/ic-arm.cc (right):

http://codereview.chromium.org/2116003/diff/10001/11005#newcode1341
src/arm/ic-arm.cc:1341: __ add(r4, key, Operand(1 << (kSmiTagSize +
kSmiShiftSize)));
On 2010/05/17 19:15:15, Erik Corry wrote:
> Operand(Smi::FromInt(1))

Done.

http://codereview.chromium.org/2116003/diff/10001/11005#newcode1368
src/arm/ic-arm.cc:1368: __ b(eq, &exit);
On 2010/05/17 19:15:15, Erik Corry wrote:
> I think you might as well do a Ret(eq) here.

Done.

http://codereview.chromium.org/2116003/diff/10001/11005#newcode1555
src/arm/ic-arm.cc:1555: // r3: base pointer of external storage
On 2010/05/17 19:15:15, Erik Corry wrote:
> Full stop here and lots of other places.

Done.

http://codereview.chromium.org/2116003/diff/10001/11007
File src/arm/virtual-frame-arm.cc (right):

http://codereview.chromium.org/2116003/diff/10001/11007#newcode344
src/arm/virtual-frame-arm.cc:344: EmitPop(r2);
On 2010/05/17 19:15:15, Erik Corry wrote:
> This pop and the spillall should be swapped, since we plan to register
allocate
> r2 soon.

Done.