Implement long jump in ARM and MIPS simulators.

Implement long jump in ARM and MIPS simulators.
Implement error and exception handler stubs on ARM.
Hook up simulator for object tests.
Enable codegen and object tests on ARM.

Committed: https://code.google.com/p/dart/source/detail?r=21680

[siva, 2013-04-18 02:07:12]

LGTM with one comment regarding re-factoring.

https://codereview.chromium.org/14309004/diff/1/runtime/vm/exceptions.cc
File runtime/vm/exceptions.cc (right):

https://codereview.chromium.org/14309004/diff/1/runtime/vm/exceptions.cc#newc...
runtime/vm/exceptions.cc:254: #endif
We not factor this code after we check and see if it is possible to unify the
stubs.

The check if (raw_stacktrace == NULL) for deciding whether to call the
ExceptionHandler or ErrorHandler seems a little unreadable.

[zra, 2013-04-18 15:45:25]

lgtm with question

https://codereview.chromium.org/14309004/diff/1/runtime/vm/stub_code_arm.cc
File runtime/vm/stub_code_arm.cc (right):

https://codereview.chromium.org/14309004/diff/1/runtime/vm/stub_code_arm.cc#n...
runtime/vm/stub_code_arm.cc:1489: __ ldr(R1, Address(SP, 0));  // Error object.
Stacktrace object?

https://codereview.chromium.org/14309004/diff/1/runtime/vm/stub_code_arm.cc#n...
runtime/vm/stub_code_arm.cc:1491: __ mov(SP, ShifterOperand(R1));  // Stack
pointer.
Is R1 the stack pointer here? I thought it was already the stacktrace object.

[regis, 2013-04-18 16:18:27]

Thanks!

I will try to merge the stubs in a following cl on all architectures.

-- Regis

https://codereview.chromium.org/14309004/diff/1/runtime/vm/exceptions.cc
File runtime/vm/exceptions.cc (right):

https://codereview.chromium.org/14309004/diff/1/runtime/vm/exceptions.cc#newc...
runtime/vm/exceptions.cc:254: #endif
On 2013/04/18 02:07:12, siva wrote:
> We not factor this code after we check and see if it is possible to unify the
> stubs.
> 
> The check if (raw_stacktrace == NULL) for deciding whether to call the
> ExceptionHandler or ErrorHandler seems a little unreadable.

Yes, I will try to factor this code after making sure we can set one more
register without adverse effect. Actually, the stub could also test for NULL,
and not set the register in that case.
I have added a comment about testing for NULL.

https://codereview.chromium.org/14309004/diff/1/runtime/vm/stub_code_arm.cc
File runtime/vm/stub_code_arm.cc (right):

https://codereview.chromium.org/14309004/diff/1/runtime/vm/stub_code_arm.cc#n...
runtime/vm/stub_code_arm.cc:1489: __ ldr(R1, Address(SP, 0));  // Error object.
On 2013/04/18 15:45:25, zra wrote:
> Stacktrace object?

Done.

https://codereview.chromium.org/14309004/diff/1/runtime/vm/stub_code_arm.cc#n...
runtime/vm/stub_code_arm.cc:1491: __ mov(SP, ShifterOperand(R1));  // Stack
pointer.
On 2013/04/18 15:45:25, zra wrote:
> Is R1 the stack pointer here? I thought it was already the stacktrace object.

Good catch! This stub is only used when running on the hardware, so I did not
notice the clobbering yet :-)

[regis, 2013-04-18 16:23:49]

Committed patchset #2 manually as r21680 (presubmit successful).