Inline Float32x4 constructors

Inline Float32x4 constructors

Committed: https://code.google.com/p/dart/source/detail?r=22055

[srdjan, 2013-04-25 20:19:25]

https://codereview.chromium.org/13936009/diff/1/runtime/vm/flow_graph_optimiz...
File runtime/vm/flow_graph_optimizer.cc (right):

https://codereview.chromium.org/13936009/diff/1/runtime/vm/flow_graph_optimiz...
runtime/vm/flow_graph_optimizer.cc:4761: void
ConstantPropagator::VisitFloat32x4Zero(Float32x4ZeroInstr* instr) {
Add SetValue(...)

https://codereview.chromium.org/13936009/diff/1/runtime/vm/intermediate_langu...
File runtime/vm/intermediate_language.h (right):

https://codereview.chromium.org/13936009/diff/1/runtime/vm/intermediate_langu...
runtime/vm/intermediate_language.h:4221:
Float32x4ConstructorInstr(MethodRecognizer::Kind op_kind,
Why do you need the op_kind (here and below). Do you expect other kinds than the
one used to currently generate the instruction?

https://codereview.chromium.org/13936009/diff/1/runtime/vm/intermediate_langu...
runtime/vm/intermediate_language.h:4282: StaticCallInstr* static_call)
Bad indent

https://codereview.chromium.org/13936009/diff/1/runtime/vm/intermediate_langu...
File runtime/vm/intermediate_language_arm.cc (right):

https://codereview.chromium.org/13936009/diff/1/runtime/vm/intermediate_langu...
runtime/vm/intermediate_language_arm.cc:1875: UNIMPLEMENTED();
bad indent

[Cutch, 2013-04-25 20:52:07]

https://codereview.chromium.org/13936009/diff/1/runtime/vm/flow_graph_optimiz...
File runtime/vm/flow_graph_optimizer.cc (right):

https://codereview.chromium.org/13936009/diff/1/runtime/vm/flow_graph_optimiz...
runtime/vm/flow_graph_optimizer.cc:4761: void
ConstantPropagator::VisitFloat32x4Zero(Float32x4ZeroInstr* instr) {
On 2013/04/25 20:19:25, srdjan wrote:
> Add SetValue(...)

Done.

https://codereview.chromium.org/13936009/diff/1/runtime/vm/intermediate_langu...
File runtime/vm/intermediate_language.h (right):

https://codereview.chromium.org/13936009/diff/1/runtime/vm/intermediate_langu...
runtime/vm/intermediate_language.h:4221:
Float32x4ConstructorInstr(MethodRecognizer::Kind op_kind,
On 2013/04/25 20:19:25, srdjan wrote:
> Why do you need the op_kind (here and below). Do you expect other kinds than
the
> one used to currently generate the instruction?

I left it there so that it would be easy to support other kinds of constructors.
This is very unlikely, so I've removed it from these three constructors.

https://codereview.chromium.org/13936009/diff/1/runtime/vm/intermediate_langu...
File runtime/vm/intermediate_language_arm.cc (right):

https://codereview.chromium.org/13936009/diff/1/runtime/vm/intermediate_langu...
runtime/vm/intermediate_language_arm.cc:1875: UNIMPLEMENTED();
On 2013/04/25 20:19:25, srdjan wrote:
> bad indent

Done.

[srdjan, 2013-04-25 21:10:47]

LGTM
Same comment for x64 as for ia32.

https://codereview.chromium.org/13936009/diff/7001/runtime/vm/intermediate_la...
File runtime/vm/intermediate_language_ia32.cc (right):

https://codereview.chromium.org/13936009/diff/7001/runtime/vm/intermediate_la...
runtime/vm/intermediate_language_ia32.cc:2916: XmmRegister v3 =
locs()->in(3).fpu_reg();
ASSERT result is same as v0.

https://codereview.chromium.org/13936009/diff/7001/runtime/vm/intermediate_la...
runtime/vm/intermediate_language_ia32.cc:2946: XmmRegister value =
locs()->out().fpu_reg();
ditto

https://codereview.chromium.org/13936009/diff/7001/runtime/vm/intermediate_la...
runtime/vm/intermediate_language_ia32.cc:2963: XmmRegister value =
locs()->out().fpu_reg();
ditto

[Cutch, 2013-04-25 22:03:17]

Committed patchset #4 manually as r22055 (presubmit successful).