S390: InstanceOfStub incorrectly interprets the hole as a prototype.

S390: InstanceOfStub incorrectly interprets the hole as a prototype.

Port 2aa070be4fd2960df98905b254f12ed801ef26cd

Original commit message:
    Repair this to match what the runtime correctly does, by first checking if the
    function is a constructor before we access the prototype.

R=mvstanton@chromium.org, joransiu@ca.ibm.com, michael_dawson@ca.ibm.com, mbrandy@us.ibm.com
BUG=

Committed: https://crrev.com/71d525a3b836d88bcff6d6414520228d6b3ece3d
Cr-Commit-Position: refs/heads/master@{#34872}

[john.yan, 2016-03-17 18:28:14]

PTAL

[john.yan, 2016-03-17 18:28:22]

The CQ bit was checked by jyan@ca.ibm.com to run a CQ dry run

[commit-bot: I haz the power, 2016-03-17 18:28:27]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1813003002/1
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1813003002/1

[MTBrandyberry, 2016-03-17 18:42:13]

lgtm

[JoranSiu, 2016-03-17 18:44:57]

On 2016/03/17 18:28:27, commit-bot: I haz the power wrote:
> Dry run: CQ is trying da patch. Follow status at
>  https://chromium-cq-status.appspot.com/patch-status/1813003002/1
> View timeline at
>  https://chromium-cq-status.appspot.com/patch-timeline/1813003002/1

Is it possible to use the TestUnderMask instruction here, instead of having to
load the byte into scratch register.  Also, we can probably add kNear on the
branch to the slow_case label.

[commit-bot: I haz the power, 2016-03-17 18:50:29]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2016-03-17 18:50:30]

Dry run: This issue passed the CQ dry run.

[john.yan, 2016-03-17 18:56:27]

On 2016/03/17 18:44:57, JoranSiu wrote:
> On 2016/03/17 18:28:27, commit-bot: I haz the power wrote:
> > Dry run: CQ is trying da patch. Follow status at
> >  https://chromium-cq-status.appspot.com/patch-status/1813003002/1
> > View timeline at
> >  https://chromium-cq-status.appspot.com/patch-timeline/1813003002/1
> 
> Is it possible to use the TestUnderMask instruction here, instead of having to
> load the byte into scratch register.  Also, we can probably add kNear on the
> branch to the slow_case label.

Nice suggestion! but the scratch reg is used afterwards for Object type
comparison, so tm won't be better here.

[john.yan, 2016-03-17 18:56:34]

The CQ bit was checked by jyan@ca.ibm.com

[commit-bot: I haz the power, 2016-03-17 18:56:36]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1813003002/1
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1813003002/1

[commit-bot: I haz the power, 2016-03-17 18:58:35]

Committed patchset #1 (id:1)

[commit-bot: I haz the power, 2016-03-17 19:00:04]

Description was changed from

==========
S390: InstanceOfStub incorrectly interprets the hole as a prototype.

Port 2aa070be4fd2960df98905b254f12ed801ef26cd

Original commit message:
    Repair this to match what the runtime correctly does, by first checking if
the
    function is a constructor before we access the prototype.

R=mvstanton@chromium.org, joransiu@ca.ibm.com, michael_dawson@ca.ibm.com,
mbrandy@us.ibm.com
BUG=
==========

to

==========
S390: InstanceOfStub incorrectly interprets the hole as a prototype.

Port 2aa070be4fd2960df98905b254f12ed801ef26cd

Original commit message:
    Repair this to match what the runtime correctly does, by first checking if
the
    function is a constructor before we access the prototype.

R=mvstanton@chromium.org, joransiu@ca.ibm.com, michael_dawson@ca.ibm.com,
mbrandy@us.ibm.com
BUG=

Committed: https://crrev.com/71d525a3b836d88bcff6d6414520228d6b3ece3d
Cr-Commit-Position: refs/heads/master@{#34872}
==========

[commit-bot: I haz the power, 2016-03-17 19:00:05]

Patchset 1 (id:??) landed as
https://crrev.com/71d525a3b836d88bcff6d6414520228d6b3ece3d
Cr-Commit-Position: refs/heads/master@{#34872}

[JoranSiu, 2016-03-17 19:16:56]

On 2016/03/17 18:56:27, john.yan wrote:
> On 2016/03/17 18:44:57, JoranSiu wrote:
> > On 2016/03/17 18:28:27, commit-bot: I haz the power wrote:
> > > Dry run: CQ is trying da patch. Follow status at
> > >  https://chromium-cq-status.appspot.com/patch-status/1813003002/1
> > > View timeline at
> > >  https://chromium-cq-status.appspot.com/patch-timeline/1813003002/1
> > 
> > Is it possible to use the TestUnderMask instruction here, instead of having
to
> > load the byte into scratch register.  Also, we can probably add kNear on the
> > branch to the slow_case label.
> 
> Nice suggestion! but the scratch reg is used afterwards for Object type
> comparison, so tm won't be better here.

Hmm, do you mean this line:
"CompareObjectType(function_prototype, scratch, scratch, MAP_TYPE);"

If you look at the implementation, both the 2nd and 3rd parm with 'scratch', are
clobbered.  scratch reg will contain the Map after the call.  I think we should
still add kNear to the branch.