Inline recognized List factory in the flow graph optimizer.

Inline recognized List factory in the flow graph optimizer.

The call to the native List factory is lowered into IL instructions if
the length is known to be a valid smi.

This is mostly performance neutral. The acutal allocation now takes place
in the array allocation stub instead of the intrinsic code (and the runtime
List_allocate in case the fast case path fails).

This is a preparation for enabling allocation sinking for arrays and will
be extended to handle type list allocation as well. This way the allocation
site is explicitly represented as a CreateArrayInstr, instead of just being
a static call.

Another benefit is that this allows to simplify the special handling of
recognized factory calls in the optimizer once all array types are handled
this way.

R=johnmccutchan@google.com, srdjan@google.com

Committed: https://code.google.com/p/dart/source/detail?r=32194

[srdjan, 2014-01-30 17:55:23]

lgtm

https://codereview.chromium.org/138523004/diff/1/runtime/vm/flow_graph_builde...
File runtime/vm/flow_graph_builder.cc (right):

https://codereview.chromium.org/138523004/diff/1/runtime/vm/flow_graph_builde...
runtime/vm/flow_graph_builder.cc:2177: Bind(new
ConstantInstr(Smi::ZoneHandle(Smi::New(node->length()))));
Indent 4 spaces

https://codereview.chromium.org/138523004/diff/1/runtime/vm/flow_graph_optimi...
File runtime/vm/flow_graph_optimizer.cc (right):

https://codereview.chromium.org/138523004/diff/1/runtime/vm/flow_graph_optimi...
runtime/vm/flow_graph_optimizer.cc:3702: }
In another CL maybe: especially in checked mode we should inline Map factory so
that the type arguments are known when allocating concrete Map implementation
and thus the type arguments are canonicalized at compile time (example: new
Map<String, int>())

https://codereview.chromium.org/138523004/diff/1/runtime/vm/intermediate_lang...
File runtime/vm/intermediate_language.cc (right):

https://codereview.chromium.org/138523004/diff/1/runtime/vm/intermediate_lang...
runtime/vm/intermediate_language.cc:2922:
!num_elements->BoundConstant().IsSmi()) {
I think num_elements->BoundConstant must be a Smi, you could assert, but this
way is safe as well.

[Cutch, 2014-01-30 19:07:36]

lgtm

[Florian Schneider, 2014-01-31 12:43:33]

https://codereview.chromium.org/138523004/diff/1/runtime/vm/flow_graph_builde...
File runtime/vm/flow_graph_builder.cc (right):

https://codereview.chromium.org/138523004/diff/1/runtime/vm/flow_graph_builde...
runtime/vm/flow_graph_builder.cc:2177: Bind(new
ConstantInstr(Smi::ZoneHandle(Smi::New(node->length()))));
On 2014/01/30 17:55:23, srdjan wrote:
> Indent 4 spaces

Done.

https://codereview.chromium.org/138523004/diff/1/runtime/vm/flow_graph_optimi...
File runtime/vm/flow_graph_optimizer.cc (right):

https://codereview.chromium.org/138523004/diff/1/runtime/vm/flow_graph_optimi...
runtime/vm/flow_graph_optimizer.cc:3702: }
On 2014/01/30 17:55:23, srdjan wrote:
> In another CL maybe: especially in checked mode we should inline Map factory
so
> that the type arguments are known when allocating concrete Map implementation
> and thus the type arguments are canonicalized at compile time (example: new
> Map<String, int>())

Agree. This can definitely be extended to other known factory methods.

https://codereview.chromium.org/138523004/diff/1/runtime/vm/intermediate_lang...
File runtime/vm/intermediate_language.cc (right):

https://codereview.chromium.org/138523004/diff/1/runtime/vm/intermediate_lang...
runtime/vm/intermediate_language.cc:2922:
!num_elements->BoundConstant().IsSmi()) {
On 2014/01/30 17:55:23, srdjan wrote:
> I think num_elements->BoundConstant must be a Smi, you could assert, but this
> way is safe as well.

For now, I'll keep this (more general) version since it should not affect the
result.

[Florian Schneider, 2014-01-31 13:26:58]

Committed patchset #2 manually as r32194 (presubmit successful).