Return Object* instead of MaybeObject* from runtime calls.

Return Object* instead of MaybeObject* from runtime calls.

R=jkummerow@chromium.org

Committed: https://code.google.com/p/v8/source/detail?r=20888

[Yang, 2014-04-17 12:48:18]

This depends on

https://codereview.chromium.org/240493007/
https://codereview.chromium.org/240883003/
https://codereview.chromium.org/240393003/

Platform ports for x64, arm, arm64 and mips for the CEntryStub are still missing
and will trickle in.

[Jakob Kummerow, 2014-04-17 14:20:11]

Looks good so far.

https://codereview.chromium.org/240053010/diff/30001/src/ia32/code-stubs-ia32.cc
File src/ia32/code-stubs-ia32.cc (right):

https://codereview.chromium.org/240053010/diff/30001/src/ia32/code-stubs-ia32...
src/ia32/code-stubs-ia32.cc:2596: // NOTE: Invocations of builtins may return
failure objects instead
I think this is outdated ;-)

https://codereview.chromium.org/240053010/diff/30001/src/runtime.cc
File src/runtime.cc (right):

https://codereview.chromium.org/240053010/diff/30001/src/runtime.cc#newcode2579
src/runtime.cc:2579: Runtime_OptimizeObjectForAddingMultipleProperties) {
nit: fits on one line now

[Yang, 2014-04-22 11:28:54]

https://codereview.chromium.org/240053010/diff/90001/src/arm64/code-stubs-arm...
File src/arm64/code-stubs-arm64.cc (right):

https://codereview.chromium.org/240053010/diff/90001/src/arm64/code-stubs-arm...
src/arm64/code-stubs-arm64.cc:1513: // TODO(jbramley): If the first call attempt
succeeds in the common case (as
@jbramley: now that we don't have a retry loop anymore, this should be doable. I
made a modest attempt, but for some reason it didn't work out, and I punted on
this. Maybe you want to have a go at this?

[jbramley, 2014-04-22 11:37:03]

https://codereview.chromium.org/240053010/diff/90001/src/arm64/code-stubs-arm...
File src/arm64/code-stubs-arm64.cc (right):

https://codereview.chromium.org/240053010/diff/90001/src/arm64/code-stubs-arm...
src/arm64/code-stubs-arm64.cc:1513: // TODO(jbramley): If the first call attempt
succeeds in the common case (as
On 2014/04/22 11:28:55, Yang wrote:
> @jbramley: now that we don't have a retry loop anymore, this should be doable.
I
> made a modest attempt, but for some reason it didn't work out, and I punted on
> this. Maybe you want to have a go at this?

Yep, I'll have a look. We're gradually working through TODOs in the arm64 port
so they'll all get looked at eventually.

[Jakob Kummerow, 2014-04-22 12:22:22]

LGTM.

https://codereview.chromium.org/240053010/diff/90001/src/arm/code-stubs-arm.cc
File src/arm/code-stubs-arm.cc (right):

https://codereview.chromium.org/240053010/diff/90001/src/arm/code-stubs-arm.c...
src/arm/code-stubs-arm.cc:1506: // Called from JavaScript; parameters are on
stack as if calling JS function
nit: trailing full stop please

https://codereview.chromium.org/240053010/diff/90001/src/ia32/code-stubs-ia32.cc
File src/ia32/code-stubs-ia32.cc (right):

https://codereview.chromium.org/240053010/diff/90001/src/ia32/code-stubs-ia32...
src/ia32/code-stubs-ia32.cc:2619:
Immediate(ExternalReference::isolate_address(masm->isolate())));
nit: s/masm->isolate()/isolate/ here...

https://codereview.chromium.org/240053010/diff/90001/src/ia32/code-stubs-ia32...
src/ia32/code-stubs-ia32.cc:2627: __ cmp(eax,
masm->isolate()->factory()->the_hole_value());
...and here...

https://codereview.chromium.org/240053010/diff/90001/src/ia32/code-stubs-ia32...
src/ia32/code-stubs-ia32.cc:2633: // Check for failure result.
nit: s/failure/exception/ to avoid confusion? (If you change it, rename the
"failure_returned" label too.)

https://codereview.chromium.org/240053010/diff/90001/src/ia32/code-stubs-ia32...
src/ia32/code-stubs-ia32.cc:2639: Isolate::kPendingExceptionAddress,
masm->isolate());
...and here.

https://codereview.chromium.org/240053010/diff/90001/src/x64/code-stubs-x64.cc
File src/x64/code-stubs-x64.cc (right):

https://codereview.chromium.org/240053010/diff/90001/src/x64/code-stubs-x64.c...
src/x64/code-stubs-x64.cc:2429: //      On entry to code generated by
GenerateCore, it must hold
"GenerateCore" doesn't exist any more.

https://codereview.chromium.org/240053010/diff/90001/src/x64/code-stubs-x64.c...
src/x64/code-stubs-x64.cc:2516: __ cmpq(r14, pending_exception_operand);
I think the x32 port requires a "cmpp" here ("p" as in "pointer size").

[Yang, 2014-04-22 12:38:16]

Thanks for the review. I uploaded a new patch set addressing the comments.
Landing...

[Yang, 2014-04-22 12:51:27]

Committed patchset #8 manually as r20888 (presubmit successful).