Remove skip_static_calls_ as it uses an obsolete way to check for uncalled static calls. Will be re…

Remove skip_static_calls_ as it uses an obsolete way to check for uncalled static calls. Will be replaced by ICData tracking.
Add two different PCDescriptors to differentiate between optimized and unoptimized static calls (the former loads args. descr, the later loads ICData before calling).
Populate static call's ic_data field when optimizing.

R=asiva@google.com

Committed: https://code.google.com/p/dart/source/detail?r=24433

[siva, 2013-06-25 21:21:05]

lgtm

https://codereview.chromium.org/17723002/diff/33001/runtime/vm/code_patcher_i...
File runtime/vm/code_patcher_ia32.cc (right):

https://codereview.chromium.org/17723002/diff/33001/runtime/vm/code_patcher_i...
runtime/vm/code_patcher_ia32.cc:246: uword return_address, const Code& code,
ICData* ic_data_result) {
why not keep the param name consistent with GetInstanceCallAt
(i.e use ic_data instead of ic_data_result)
Here and elsewhere too.

https://codereview.chromium.org/17723002/diff/33001/runtime/vm/code_patcher_x...
File runtime/vm/code_patcher_x64.cc (right):

https://codereview.chromium.org/17723002/diff/33001/runtime/vm/code_patcher_x...
runtime/vm/code_patcher_x64.cc:237: }
This code seems pretty similar in all the target versions. Could they be
consolidated to be target independent?

https://codereview.chromium.org/17723002/diff/33001/runtime/vm/object.cc
File runtime/vm/object.cc (right):

https://codereview.chromium.org/17723002/diff/33001/runtime/vm/object.cc#newc...
runtime/vm/object.cc:8258: ASSERT(kind == PcDescriptors::kUnoptStaticCall);
This assert seems superfluous considering there is an
if (kind == kIcCall || kin kUnoptStatiCall) above

[srdjan, 2013-06-25 21:26:08]

Thanks!

https://codereview.chromium.org/17723002/diff/33001/runtime/vm/code_patcher_i...
File runtime/vm/code_patcher_ia32.cc (right):

https://codereview.chromium.org/17723002/diff/33001/runtime/vm/code_patcher_i...
runtime/vm/code_patcher_ia32.cc:246: uword return_address, const Code& code,
ICData* ic_data_result) {
On 2013/06/25 21:21:05, siva wrote:
> why not keep the param name consistent with GetInstanceCallAt
> (i.e use ic_data instead of ic_data_result)
> Here and elsewhere too.

Discussed offline. Leaving as it is, since we need a different ic_data object
inside the function.

https://codereview.chromium.org/17723002/diff/33001/runtime/vm/code_patcher_x...
File runtime/vm/code_patcher_x64.cc (right):

https://codereview.chromium.org/17723002/diff/33001/runtime/vm/code_patcher_x...
runtime/vm/code_patcher_x64.cc:237: }
On 2013/06/25 21:21:05, siva wrote:
> This code seems pretty similar in all the target versions. Could they be
> consolidated to be target independent?

Next CL.

https://codereview.chromium.org/17723002/diff/33001/runtime/vm/object.cc
File runtime/vm/object.cc (right):

https://codereview.chromium.org/17723002/diff/33001/runtime/vm/object.cc#newc...
runtime/vm/object.cc:8258: ASSERT(kind == PcDescriptors::kUnoptStaticCall);
On 2013/06/25 21:21:05, siva wrote:
> This assert seems superfluous considering there is an
> if (kind == kIcCall || kin kUnoptStatiCall) above

Removed

[srdjan, 2013-06-25 21:33:08]

Committed patchset #5 manually as r24433 (presubmit successful).