[Interpreter] Adds support to fetch return value on break at return.

[Interpreter] Adds support to fetch return value on break at return.

Debugger fetches the return value of a function when we break at return.
Interpreter holds the return value in accumulator. This is not stored in a
specified location on stack and hence it is not possible to look it up from
stack similar to full-codegen or optimized frames. This cl adds support to
store the value of accumulator on debug breaks. The value of accumulator is
passed to the runtime function and is then stored in thread local data.

Also changes full-codegen implementation to match that of ignition.
The return value from full-codegen is also stored in thread local data.
The return value is fetched directly thread local data instead of
finding it by iterating over frames.

BUG=v8:4280, v8:4690
LOG=N

Committed: https://crrev.com/fb65527b75754bcf3b173f16f5d0b04a1c6d9b99
Cr-Commit-Position: refs/heads/master@{#35060}

[mythria, 2016-03-21 10:34:17]

Description was changed from

==========
[Interpreter] Adds support to fetch return value on break at return in
interpreter.

Interpreter holds the return value in accumulator. This is not stored in a
specified location on stack and hence it is not possible to look it up from
stack similar to full-codegen or optimized frames. This cl adds support to
store the value of accumulator on debug breaks. The value of accumulator is
passed to the runtime function and is then stored in thread local data.

BUG=v8:4280, v8:4690
LOG=N
==========

to

==========
[Interpreter] Adds support to fetch return value on break at return in
interpreter.

Debugger fetches the return value of a function when we break at return.
Interpreter holds the return value in accumulator. This is not stored in a
specified location on stack and hence it is not possible to look it up from
stack similar to full-codegen or optimized frames. This cl adds support to
store the value of accumulator on debug breaks. The value of accumulator is
passed to the runtime function and is then stored in thread local data.


BUG=v8:4280, v8:4690
LOG=N
==========

[mythria, 2016-03-21 10:34:28]

The CQ bit was checked by mythria@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-03-21 10:34:41]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1818873003/1
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1818873003/1

[commit-bot: I haz the power, 2016-03-21 10:54:19]

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

[commit-bot: I haz the power, 2016-03-21 10:54:20]

Dry run: Try jobs failed on following builders:
  v8_linux_arm64_rel on tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_linux_arm64_rel/builds/1...)

[mythria, 2016-03-21 15:50:20]

The CQ bit was checked by mythria@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-03-21 15:50:25]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1818873003/20001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1818873003/20001

[mythria, 2016-03-21 15:53:48]

Description was changed from

==========
[Interpreter] Adds support to fetch return value on break at return in
interpreter.

Debugger fetches the return value of a function when we break at return.
Interpreter holds the return value in accumulator. This is not stored in a
specified location on stack and hence it is not possible to look it up from
stack similar to full-codegen or optimized frames. This cl adds support to
store the value of accumulator on debug breaks. The value of accumulator is
passed to the runtime function and is then stored in thread local data.


BUG=v8:4280, v8:4690
LOG=N
==========

to

==========
[Interpreter] Adds support to fetch return value on break at return.

Debugger fetches the return value of a function when we break at return.
Interpreter holds the return value in accumulator. This is not stored in a
specified location on stack and hence it is not possible to look it up from
stack similar to full-codegen or optimized frames. This cl adds support to
store the value of accumulator on debug breaks. The value of accumulator is
passed to the runtime function and is then stored in thread local data.

Also changes full-codegen implementation to match that of ignition.
The return value from full-codegen is also stored in thread local data
instead of looking it up by iterating over frames.


BUG=v8:4280, v8:4690
LOG=N
==========

[mythria, 2016-03-21 15:55:29]

Description was changed from

==========
[Interpreter] Adds support to fetch return value on break at return.

Debugger fetches the return value of a function when we break at return.
Interpreter holds the return value in accumulator. This is not stored in a
specified location on stack and hence it is not possible to look it up from
stack similar to full-codegen or optimized frames. This cl adds support to
store the value of accumulator on debug breaks. The value of accumulator is
passed to the runtime function and is then stored in thread local data.

Also changes full-codegen implementation to match that of ignition.
The return value from full-codegen is also stored in thread local data
instead of looking it up by iterating over frames.


BUG=v8:4280, v8:4690
LOG=N
==========

to

==========
[Interpreter] Adds support to fetch return value on break at return.

Debugger fetches the return value of a function when we break at return.
Interpreter holds the return value in accumulator. This is not stored in a
specified location on stack and hence it is not possible to look it up from
stack similar to full-codegen or optimized frames. This cl adds support to
store the value of accumulator on debug breaks. The value of accumulator is
passed to the runtime function and is then stored in thread local data.

Also changes full-codegen implementation to match that of ignition.
The return value from full-codegen is also stored in thread local data.
The return value is fetched directly thread local data instead of 
finding it by iterating over frames.


BUG=v8:4280, v8:4690
LOG=N
==========

[mythria, 2016-03-21 16:24:28]

mythria@chromium.org changed reviewers:
+ rmcilroy@chromium.org, yangguo@chromium.org

[mythria, 2016-03-21 16:24:29]

For now, I implemented it only for x64 and ia32. If the implementation looks ok
I will port it to other architectures as well. 

PTAL.

https://codereview.chromium.org/1818873003/diff/20001/src/debug/debug.cc
File src/debug/debug.cc (right):

https://codereview.chromium.org/1818873003/diff/20001/src/debug/debug.cc#newc...
src/debug/debug.cc:1594: return *thread_local_.return_value_;
Returns the actual return value to avoid pushing and popping rax/eax in the
DebugBreakStub.

https://codereview.chromium.org/1818873003/diff/20001/src/debug/x64/debug-x64.cc
File src/debug/x64/debug-x64.cc (right):

https://codereview.chromium.org/1818873003/diff/20001/src/debug/x64/debug-x64...
src/debug/x64/debug-x64.cc:90: for (int i = SAVE_RESULT_REGISTER ? 1 : 0; i <
kNumJSCallerSaved; ++i) {
If this is too hacky, I will just push and pop rax when mode ==
SAVE_ERSULT_REGISTER.

[commit-bot: I haz the power, 2016-03-21 16:30:24]

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

[commit-bot: I haz the power, 2016-03-21 16:30:25]

Dry run: Try jobs failed on following builders:
  v8_linux_arm_rel on tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_linux_arm_rel/builds/15672)

[rmcilroy, 2016-03-22 11:32:35]

Approach looks good to me, with a couple of comments. I'll leave Yang to
approve.

https://codereview.chromium.org/1818873003/diff/20001/src/debug/debug.h
File src/debug/debug.h (right):

https://codereview.chromium.org/1818873003/diff/20001/src/debug/debug.h#newco...
src/debug/debug.h:685: // this value will be a hole.
/s/a hole/the hole/

https://codereview.chromium.org/1818873003/diff/20001/src/debug/ia32/debug-ia...
File src/debug/ia32/debug-ia32.cc (right):

https://codereview.chromium.org/1818873003/diff/20001/src/debug/ia32/debug-ia...
src/debug/ia32/debug-ia32.cc:90: for (int i = SAVE_RESULT_REGISTER ? 1 : 0; i <
kNumJSCallerSaved; ++i) {
This is relying on JSCallerSavedCode(0) being eax, this is a bit brittle - could
you make this more explicit.

https://codereview.chromium.org/1818873003/diff/20001/src/debug/x64/debug-x64.cc
File src/debug/x64/debug-x64.cc (right):

https://codereview.chromium.org/1818873003/diff/20001/src/debug/x64/debug-x64...
src/debug/x64/debug-x64.cc:90: for (int i = SAVE_RESULT_REGISTER ? 1 : 0; i <
kNumJSCallerSaved; ++i) {
On 2016/03/21 16:24:29, mythria wrote:
> If this is too hacky, I will just push and pop rax when mode ==
> SAVE_ERSULT_REGISTER.

Ditto. I'm not sure it's necessary to push / pop, but making this loop
explicitly check when the register is rax and mode is SAVE_RESULT_REGISTER.

[mythria, 2016-03-22 12:08:09]

Thanks, Fixed.

https://codereview.chromium.org/1818873003/diff/20001/src/debug/debug.h
File src/debug/debug.h (right):

https://codereview.chromium.org/1818873003/diff/20001/src/debug/debug.h#newco...
src/debug/debug.h:685: // this value will be a hole.
On 2016/03/22 11:32:34, rmcilroy wrote:
> /s/a hole/the hole/

Done.

https://codereview.chromium.org/1818873003/diff/20001/src/debug/ia32/debug-ia...
File src/debug/ia32/debug-ia32.cc (right):

https://codereview.chromium.org/1818873003/diff/20001/src/debug/ia32/debug-ia...
src/debug/ia32/debug-ia32.cc:90: for (int i = SAVE_RESULT_REGISTER ? 1 : 0; i <
kNumJSCallerSaved; ++i) {
On 2016/03/22 11:32:34, rmcilroy wrote:
> This is relying on JSCallerSavedCode(0) being eax, this is a bit brittle -
could
> you make this more explicit.

Thanks, I did not realize we could check if a register is eax. Done.

https://codereview.chromium.org/1818873003/diff/20001/src/debug/x64/debug-x64.cc
File src/debug/x64/debug-x64.cc (right):

https://codereview.chromium.org/1818873003/diff/20001/src/debug/x64/debug-x64...
src/debug/x64/debug-x64.cc:90: for (int i = SAVE_RESULT_REGISTER ? 1 : 0; i <
kNumJSCallerSaved; ++i) {
On 2016/03/22 11:32:34, rmcilroy wrote:
> On 2016/03/21 16:24:29, mythria wrote:
> > If this is too hacky, I will just push and pop rax when mode ==
> > SAVE_ERSULT_REGISTER.
> 
> Ditto. I'm not sure it's necessary to push / pop, but making this loop
> explicitly check when the register is rax and mode is SAVE_RESULT_REGISTER.

Done.

[Yang, 2016-03-22 12:22:58]

https://codereview.chromium.org/1818873003/diff/20001/src/debug/arm64/debug-a...
File src/debug/arm64/debug-arm64.cc (right):

https://codereview.chromium.org/1818873003/diff/20001/src/debug/arm64/debug-a...
src/debug/arm64/debug-arm64.cc:114: if (mode == SAVE_RESULT_REGISTER) __
Pop(x0);
I thought we are going to use the return value of the runtime function to
restore x0. What we are doing now is using push/pop to persist x0, and passing
the-hole to debug break, which returns it.

Instead, we should push x0 if mode == SAVE_RESULT_REGISTER, and the-hole
otherwise. Runtime_DebugBreak would get the return value for the
Return_DebugBreak.

https://codereview.chromium.org/1818873003/diff/20001/src/debug/debug.cc
File src/debug/debug.cc (right):

https://codereview.chromium.org/1818873003/diff/20001/src/debug/debug.cc#newc...
src/debug/debug.cc:571: thread_local_.return_value_ = args.at<Object>(0);
Please introduce a method to set and retrieve the return value, and use it in
Runtime_DebugBreak.

https://codereview.chromium.org/1818873003/diff/20001/src/debug/debug.cc#newc...
src/debug/debug.cc:1594: return *thread_local_.return_value_;
On 2016/03/21 16:24:29, mythria wrote:
> Returns the actual return value to avoid pushing and popping rax/eax in the
> DebugBreakStub.

I forgot about this part. For the interpreter, the return value of this function
is the after-break-target. For FCG, we store it in a memory location.

Let's split Runtime_DebugBreak into two runtime functions. One to cater to FCG.
It would set the return value before calling Debug::Break, and return the return
value. The other caters to Ignition, it sets the return value before calling
Debug::Break, and returns the debug break bytecode handler.

Let's also move the code in SetAfterBreakTarget for the interpreter case into
the runtime function and only call SetAfterBreakTarget for FCG.

[mythria, 2016-03-22 14:36:02]

The CQ bit was checked by mythria@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-03-22 14:36:12]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1818873003/60001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1818873003/60001

[commit-bot: I haz the power, 2016-03-22 14:37:21]

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

[commit-bot: I haz the power, 2016-03-22 14:37:22]

Dry run: Try jobs failed on following builders:
  v8_linux_gcc_compile_rel on tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_linux_gcc_compile_rel/bu...)
  v8_linux_rel_ng on tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_linux_rel_ng/builds/3336)

[mythria, 2016-03-22 14:49:52]

The CQ bit was checked by mythria@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-03-22 14:50:01]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1818873003/80001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1818873003/80001

[mythria, 2016-03-22 15:11:44]

Thanks Yang. Fixed them.

PTAL.

Thanks,
Mythri

https://codereview.chromium.org/1818873003/diff/20001/src/debug/arm64/debug-a...
File src/debug/arm64/debug-arm64.cc (right):

https://codereview.chromium.org/1818873003/diff/20001/src/debug/arm64/debug-a...
src/debug/arm64/debug-arm64.cc:114: if (mode == SAVE_RESULT_REGISTER) __
Pop(x0);
On 2016/03/22 12:22:57, Yang wrote:
> I thought we are going to use the return value of the runtime function to
> restore x0. What we are doing now is using push/pop to persist x0, and passing
> the-hole to debug break, which returns it.
> 
> Instead, we should push x0 if mode == SAVE_RESULT_REGISTER, and the-hole
> otherwise. Runtime_DebugBreak would get the return value for the
> Return_DebugBreak. 

Done. I am sorry, did not port it to arm earlier.

https://codereview.chromium.org/1818873003/diff/20001/src/debug/debug.cc
File src/debug/debug.cc (right):

https://codereview.chromium.org/1818873003/diff/20001/src/debug/debug.cc#newc...
src/debug/debug.cc:1594: return *thread_local_.return_value_;
On 2016/03/22 12:22:58, Yang wrote:
> On 2016/03/21 16:24:29, mythria wrote:
> > Returns the actual return value to avoid pushing and popping rax/eax in the
> > DebugBreakStub.
> 
> I forgot about this part. For the interpreter, the return value of this
function
> is the after-break-target. For FCG, we store it in a memory location.
> 
> Let's split Runtime_DebugBreak into two runtime functions. One to cater to
FCG.
> It would set the return value before calling Debug::Break, and return the
return
> value. The other caters to Ignition, it sets the return value before calling
> Debug::Break, and returns the debug break bytecode handler.
> 
> Let's also move the code in SetAfterBreakTarget for the interpreter case into
> the runtime function and only call SetAfterBreakTarget for FCG.

Done.

[commit-bot: I haz the power, 2016-03-22 15:12:45]

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

[commit-bot: I haz the power, 2016-03-22 15:12:46]

Dry run: This issue passed the CQ dry run.

[Yang, 2016-03-23 14:22:30]

Thanks! LGTM with two more small comments.

https://codereview.chromium.org/1818873003/diff/80001/src/debug/debug.h
File src/debug/debug.h (right):

https://codereview.chromium.org/1818873003/diff/80001/src/debug/debug.h#newco...
src/debug/debug.h:527: Handle<Object> get_return_value() { return
thread_local_.return_value_; }
The convention for getter would be return_value() without the "get" prefix.

https://codereview.chromium.org/1818873003/diff/80001/src/runtime/runtime-deb...
File src/runtime/runtime-debug.cc (right):

https://codereview.chromium.org/1818873003/diff/80001/src/runtime/runtime-deb...
src/runtime/runtime-debug.cc:43: isolate->debug()->Break(args, it.frame());
let's not pass args to Debug::Break anymore. It's not used there. Probably some
historic artifact :)

[mythria, 2016-03-23 15:29:12]

Thanks, Done. I will port it to other architectures as well.

https://codereview.chromium.org/1818873003/diff/80001/src/debug/debug.h
File src/debug/debug.h (right):

https://codereview.chromium.org/1818873003/diff/80001/src/debug/debug.h#newco...
src/debug/debug.h:527: Handle<Object> get_return_value() { return
thread_local_.return_value_; }
On 2016/03/23 14:22:30, Yang wrote:
> The convention for getter would be return_value() without the "get" prefix.

Thanks, Done.

https://codereview.chromium.org/1818873003/diff/80001/src/runtime/runtime-deb...
File src/runtime/runtime-debug.cc (right):

https://codereview.chromium.org/1818873003/diff/80001/src/runtime/runtime-deb...
src/runtime/runtime-debug.cc:43: isolate->debug()->Break(args, it.frame());
On 2016/03/23 14:22:30, Yang wrote:
> let's not pass args to Debug::Break anymore. It's not used there. Probably
some
> historic artifact :)

Done.

[mythria, 2016-03-23 16:59:50]

The CQ bit was checked by mythria@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-03-23 16:59:59]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1818873003/120001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1818873003/120001

[commit-bot: I haz the power, 2016-03-23 17:03:04]

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

[commit-bot: I haz the power, 2016-03-23 17:03:05]

Dry run: Try jobs failed on following builders:
  v8_linux64_rel_ng on tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_linux64_rel_ng/builds/3384)

[Yang, 2016-03-24 05:28:43]

lgtm if tests pass.

[mythria, 2016-03-24 09:11:56]

The CQ bit was checked by mythria@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-03-24 09:12:03]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1818873003/140001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1818873003/140001

[commit-bot: I haz the power, 2016-03-24 09:42:59]

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

[commit-bot: I haz the power, 2016-03-24 09:43:00]

Dry run: This issue passed the CQ dry run.

[mythria, 2016-03-24 10:10:05]

mythria@chromium.org changed reviewers:
+ balazs.kilvady@imgtec.com, jyan@ca.ibm.com, weiliang.lin@intel.com

[mythria, 2016-03-24 10:10:06]

weiliang.lin@ Could you please take a look at the x87 port.

balazs.kilvady@ Could you please take a look at mips and mips64 ports.

jyan@ could you please take a look at ppc and s390 ports. 

Thanks,
Mythri

[mythria, 2016-03-24 15:09:07]

The CQ bit was checked by mythria@chromium.org

[mythria, 2016-03-24 15:09:07]

The patchset sent to the CQ was uploaded after l-g-t-m from yangguo@chromium.org
Link to the patchset: https://codereview.chromium.org/1818873003/#ps160001
(title: "Port to ppc,mips,mips64,x87,s390")

[commit-bot: I haz the power, 2016-03-24 15:09:14]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1818873003/160001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1818873003/160001

[commit-bot: I haz the power, 2016-03-24 15:14:41]

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

[commit-bot: I haz the power, 2016-03-24 15:14:43]

Try jobs failed on following builders:
  v8_presubmit on tryserver.v8 (JOB_FAILED,
http://build.chromium.org/p/tryserver.v8/builders/v8_presubmit/builds/12862)

[rmcilroy, 2016-03-24 15:16:49]

lgtm on interpreter

[mythria, 2016-03-24 15:18:01]

The CQ bit was checked by mythria@chromium.org

[commit-bot: I haz the power, 2016-03-24 15:18:12]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1818873003/160001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1818873003/160001

[john.yan, 2016-03-24 15:25:32]

On 2016/03/24 10:10:06, mythria wrote:
> weiliang.lin@ Could you please take a look at the x87 port.
> 
> balazs.kilvady@ Could you please take a look at mips and mips64 ports.
> 
> jyan@ could you please take a look at ppc and s390 ports. 
> 
> Thanks,
> Mythri

lgtm on s390 port.

mbrandy@us.ibm.com Could you plz take a look at ppc port?

[MTBrandyberry, 2016-03-24 15:39:47]

On 2016/03/24 15:25:32, john.yan wrote:
> On 2016/03/24 10:10:06, mythria wrote:
> > weiliang.lin@ Could you please take a look at the x87 port.
> > 
> > balazs.kilvady@ Could you please take a look at mips and mips64 ports.
> > 
> > jyan@ could you please take a look at ppc and s390 ports. 
> > 
> > Thanks,
> > Mythri
> 
> lgtm on s390 port.
> 
> mailto:mbrandy@us.ibm.com Could you plz take a look at ppc port?

lgtm for ppc.

[weiliang.lin2, 2016-03-24 15:43:55]

On 2016/03/24 10:10:06, mythria wrote:
> weiliang.lin@ Could you please take a look at the x87 port.
> 
> balazs.kilvady@ Could you please take a look at mips and mips64 ports.
> 
> jyan@ could you please take a look at ppc and s390 ports. 
> 
> Thanks,
> Mythri

lgtm on x87

[commit-bot: I haz the power, 2016-03-24 17:13:46]

Description was changed from

==========
[Interpreter] Adds support to fetch return value on break at return.

Debugger fetches the return value of a function when we break at return.
Interpreter holds the return value in accumulator. This is not stored in a
specified location on stack and hence it is not possible to look it up from
stack similar to full-codegen or optimized frames. This cl adds support to
store the value of accumulator on debug breaks. The value of accumulator is
passed to the runtime function and is then stored in thread local data.

Also changes full-codegen implementation to match that of ignition.
The return value from full-codegen is also stored in thread local data.
The return value is fetched directly thread local data instead of 
finding it by iterating over frames.


BUG=v8:4280, v8:4690
LOG=N
==========

to

==========
[Interpreter] Adds support to fetch return value on break at return.

Debugger fetches the return value of a function when we break at return.
Interpreter holds the return value in accumulator. This is not stored in a
specified location on stack and hence it is not possible to look it up from
stack similar to full-codegen or optimized frames. This cl adds support to
store the value of accumulator on debug breaks. The value of accumulator is
passed to the runtime function and is then stored in thread local data.

Also changes full-codegen implementation to match that of ignition.
The return value from full-codegen is also stored in thread local data.
The return value is fetched directly thread local data instead of 
finding it by iterating over frames.


BUG=v8:4280, v8:4690
LOG=N
==========

[commit-bot: I haz the power, 2016-03-24 17:13:47]

Committed patchset #9 (id:160001)

[commit-bot: I haz the power, 2016-03-24 17:15:15]

Description was changed from

==========
[Interpreter] Adds support to fetch return value on break at return.

Debugger fetches the return value of a function when we break at return.
Interpreter holds the return value in accumulator. This is not stored in a
specified location on stack and hence it is not possible to look it up from
stack similar to full-codegen or optimized frames. This cl adds support to
store the value of accumulator on debug breaks. The value of accumulator is
passed to the runtime function and is then stored in thread local data.

Also changes full-codegen implementation to match that of ignition.
The return value from full-codegen is also stored in thread local data.
The return value is fetched directly thread local data instead of 
finding it by iterating over frames.


BUG=v8:4280, v8:4690
LOG=N
==========

to

==========
[Interpreter] Adds support to fetch return value on break at return.

Debugger fetches the return value of a function when we break at return.
Interpreter holds the return value in accumulator. This is not stored in a
specified location on stack and hence it is not possible to look it up from
stack similar to full-codegen or optimized frames. This cl adds support to
store the value of accumulator on debug breaks. The value of accumulator is
passed to the runtime function and is then stored in thread local data.

Also changes full-codegen implementation to match that of ignition.
The return value from full-codegen is also stored in thread local data.
The return value is fetched directly thread local data instead of
finding it by iterating over frames.

BUG=v8:4280, v8:4690
LOG=N

Committed: https://crrev.com/fb65527b75754bcf3b173f16f5d0b04a1c6d9b99
Cr-Commit-Position: refs/heads/master@{#35060}
==========

[commit-bot: I haz the power, 2016-03-24 17:15:17]

Patchset 9 (id:??) landed as
https://crrev.com/fb65527b75754bcf3b173f16f5d0b04a1c6d9b99
Cr-Commit-Position: refs/heads/master@{#35060}

[balazs.kilvady, 2016-03-24 17:31:40]

On 2016/03/24 10:10:06, mythria wrote:
> balazs.kilvady@ Could you please take a look at mips and mips64 ports.
MIPS, MIPS64 looks good, thanks for the port.

[Michael Achenbach, 2016-03-24 18:17:20]

A revert of this CL (patchset #9 id:160001) has been created in
https://codereview.chromium.org/1834733002/ by machenbach@chromium.org.

The reason for reverting is: [Sheriff] Seems to break nosnap debug:
https://build.chromium.org/p/client.v8/builders/V8%20Linux%20-%20nosnap%20-%2....