VM: Clean up generated code for NoSuchMethod invocation of closures.

VM: Clean up generated code for NoSuchMethod invocation of closures.

Remove unnecessary ic-data argument from the runtime function that
invokes noSuchMethod for closures.  This reduces the size of generated
code for closures.

R=vegorov@google.com

Committed: https://code.google.com/p/dart/source/detail?r=39804

[Florian Schneider, 2014-09-02 12:01:31]

fschneider@google.com changed reviewers:
+ vegorov@google.com

[Vyacheslav Egorov (Google), 2014-09-02 22:28:04]

https://codereview.chromium.org/517383005/diff/1/runtime/vm/stub_code_arm.cc
File runtime/vm/stub_code_arm.cc (right):

https://codereview.chromium.org/517383005/diff/1/runtime/vm/stub_code_arm.cc#...
runtime/vm/stub_code_arm.cc:1169: // Push IC data object.
comment is obsolete, update

https://codereview.chromium.org/517383005/diff/1/runtime/vm/stub_code_arm.cc#...
runtime/vm/stub_code_arm.cc:1177: const intptr_t kNumArgs = 3;
this 3 is related to the number of bits set in the list above. I wonder if this
can be unified somehow.

https://codereview.chromium.org/517383005/diff/1/runtime/vm/stub_code_arm.cc#...
runtime/vm/stub_code_arm.cc:1182: __ LeaveStubFrame();
something is off here: three things were pushed and you dropped three things.
where does the result come from?

https://codereview.chromium.org/517383005/diff/1/runtime/vm/stub_code_arm64.cc
File runtime/vm/stub_code_arm64.cc (right):

https://codereview.chromium.org/517383005/diff/1/runtime/vm/stub_code_arm64.c...
runtime/vm/stub_code_arm64.cc:1255: // Push IC data object.
comment is obsolete.

https://codereview.chromium.org/517383005/diff/1/runtime/vm/stub_code_arm64.c...
runtime/vm/stub_code_arm64.cc:1268: __ Pop(R0);  // Get result into R0.
something is off here with amount of pushed args / dropped arguments and the
result.

https://codereview.chromium.org/517383005/diff/1/runtime/vm/stub_code_x64.cc
File runtime/vm/stub_code_x64.cc (right):

https://codereview.chromium.org/517383005/diff/1/runtime/vm/stub_code_x64.cc#...
runtime/vm/stub_code_x64.cc:1194: 
I think all the code after CallRuntime is dead because
kInvokeNoSuchMethodFunctionRuntimeEntry always throws. Just kill it on all
platforms?

[Florian Schneider, 2014-09-03 09:05:01]

https://codereview.chromium.org/517383005/diff/1/runtime/vm/stub_code_arm.cc
File runtime/vm/stub_code_arm.cc (right):

https://codereview.chromium.org/517383005/diff/1/runtime/vm/stub_code_arm.cc#...
runtime/vm/stub_code_arm.cc:1169: // Push IC data object.
On 2014/09/02 22:28:04, Vyacheslav Egorov (Google) wrote:
> comment is obsolete, update

Done.

https://codereview.chromium.org/517383005/diff/1/runtime/vm/stub_code_arm.cc#...
runtime/vm/stub_code_arm.cc:1177: const intptr_t kNumArgs = 3;
On 2014/09/02 22:28:04, Vyacheslav Egorov (Google) wrote:
> this 3 is related to the number of bits set in the list above. I wonder if
this
> can be unified somehow.

It's a little tricky to generate the push-list - maybe some macro-magic could
help out, but I'd rather do this in a separate CL. IP (=null) contains the slot
for the result

https://codereview.chromium.org/517383005/diff/1/runtime/vm/stub_code_arm.cc#...
runtime/vm/stub_code_arm.cc:1182: __ LeaveStubFrame();
On 2014/09/02 22:28:04, Vyacheslav Egorov (Google) wrote:
> something is off here: three things were pushed and you dropped three things.
> where does the result come from?

I'm sure PushArgumentsArray is the missing piece here.

https://codereview.chromium.org/517383005/diff/1/runtime/vm/stub_code_arm64.cc
File runtime/vm/stub_code_arm64.cc (right):

https://codereview.chromium.org/517383005/diff/1/runtime/vm/stub_code_arm64.c...
runtime/vm/stub_code_arm64.cc:1255: // Push IC data object.
On 2014/09/02 22:28:04, Vyacheslav Egorov (Google) wrote:
> comment is obsolete.

Done.

https://codereview.chromium.org/517383005/diff/1/runtime/vm/stub_code_arm64.c...
runtime/vm/stub_code_arm64.cc:1268: __ Pop(R0);  // Get result into R0.
On 2014/09/02 22:28:04, Vyacheslav Egorov (Google) wrote:
> something is off here with amount of pushed args / dropped arguments and the
> result. 

I'm sure PushArgumentsArray is the missing piece here.

https://codereview.chromium.org/517383005/diff/1/runtime/vm/stub_code_x64.cc
File runtime/vm/stub_code_x64.cc (right):

https://codereview.chromium.org/517383005/diff/1/runtime/vm/stub_code_x64.cc#...
runtime/vm/stub_code_x64.cc:1194: 
On 2014/09/02 22:28:04, Vyacheslav Egorov (Google) wrote:
> I think all the code after CallRuntime is dead because
> kInvokeNoSuchMethodFunctionRuntimeEntry always throws. Just kill it on all
> platforms?

Only Object::noSuchMethod throws, otherwise it can do whatever the user decides
to do.

[Florian Schneider, 2014-09-03 12:17:28]

https://codereview.chromium.org/517383005/diff/1/runtime/vm/stub_code_x64.cc
File runtime/vm/stub_code_x64.cc (right):

https://codereview.chromium.org/517383005/diff/1/runtime/vm/stub_code_x64.cc#...
runtime/vm/stub_code_x64.cc:1194: 
On 2014/09/03 09:05:00, Florian Schneider wrote:
> On 2014/09/02 22:28:04, Vyacheslav Egorov (Google) wrote:
> > I think all the code after CallRuntime is dead because
> > kInvokeNoSuchMethodFunctionRuntimeEntry always throws. Just kill it on all
> > platforms?
> 
> Only Object::noSuchMethod throws, otherwise it can do whatever the user
decides
> to do.

You're right that in the case of closures, it always throws.

[Vyacheslav Egorov (Google), 2014-09-03 12:34:33]

lgtm

[Florian Schneider, 2014-09-03 12:51:40]

Committed patchset #3 (id:40001) manually as r39804 (presubmit successful).