Cleanup deoptimization code to make it architecture independent (in progress).

Cleanup deoptimization code to make it architecture independent (in progress).
Implement some optimizing code on ARM in order to debug the above cleanup using
inline_stack_frame_test (still skipped in the vm status file).

This is not quite working yet, because CompilerDeoptInfo::CreateDeoptInfo
makes assumptions about the stack layout that are not valid for ARM or MIPS.
This will be cleaned up in a following cl.

R=srdjan@google.com, zra@google.com

Committed: https://code.google.com/p/dart/source/detail?r=22386

[regis, 2013-05-03 00:48:58]

https://codereview.chromium.org/14812005/diff/1/runtime/vm/code_generator.cc
File runtime/vm/code_generator.cc (left):

https://codereview.chromium.org/14812005/diff/1/runtime/vm/code_generator.cc#...
runtime/vm/code_generator.cc:1571: + 3  // caller-fp, pc, pc-marker.
I think 3 was wrong for IA32 and X64, but it had no negative effect. It should
be 2 for IA32 and X64, unless I miss something.

[zra, 2013-05-03 15:59:59]

lgtm

[srdjan, 2013-05-03 17:16:18]

LGTM

https://codereview.chromium.org/14812005/diff/1/runtime/vm/code_generator.cc
File runtime/vm/code_generator.cc (left):

https://codereview.chromium.org/14812005/diff/1/runtime/vm/code_generator.cc#...
runtime/vm/code_generator.cc:1571: + 3  // caller-fp, pc, pc-marker.
On 2013/05/03 00:48:58, regis wrote:
> I think 3 was wrong for IA32 and X64, but it had no negative effect. It should
> be 2 for IA32 and X64, unless I miss something.

Thanks for fixing it.

https://codereview.chromium.org/14812005/diff/1/runtime/vm/code_patcher_arm.cc
File runtime/vm/code_patcher_arm.cc (right):

https://codereview.chromium.org/14812005/diff/1/runtime/vm/code_patcher_arm.c...
runtime/vm/code_patcher_arm.cc:43: ASSERT(start +
CallPattern::kFixedLengthInBytes <= target);
Currently there is always a patch jump before the lazy target, so that assert
will probably not help as the call should not overlap the jump neither. It may
help in future, if the code gets rearranged

https://codereview.chromium.org/14812005/diff/1/runtime/vm/constants_x64.h
File runtime/vm/constants_x64.h (right):

https://codereview.chromium.org/14812005/diff/1/runtime/vm/constants_x64.h#ne...
runtime/vm/constants_x64.h:106: static const int kPcSlotIndexFromSp = -1;
Add an ASCII drawing explaining the constants.

[regis, 2013-05-03 18:34:12]

Thanks!

https://codereview.chromium.org/14812005/diff/1/runtime/vm/code_patcher_arm.cc
File runtime/vm/code_patcher_arm.cc (right):

https://codereview.chromium.org/14812005/diff/1/runtime/vm/code_patcher_arm.c...
runtime/vm/code_patcher_arm.cc:43: ASSERT(start +
CallPattern::kFixedLengthInBytes <= target);
On 2013/05/03 17:16:18, srdjan wrote:
> Currently there is always a patch jump before the lazy target, so that assert
> will probably not help as the call should not overlap the jump neither. It may
> help in future, if the code gets rearranged

Agreed. It will not catch all issues.

https://codereview.chromium.org/14812005/diff/1/runtime/vm/constants_x64.h
File runtime/vm/constants_x64.h (right):

https://codereview.chromium.org/14812005/diff/1/runtime/vm/constants_x64.h#ne...
runtime/vm/constants_x64.h:106: static const int kPcSlotIndexFromSp = -1;
On 2013/05/03 17:16:18, srdjan wrote:
> Add an ASCII drawing explaining the constants.

Done.

[regis, 2013-05-03 18:39:50]

Committed patchset #2 manually as r22386 (presubmit successful).