The heavy version is for x86 and x64.  The light version
is for ARM and MIPS.  Remove the elements_ array from the
virtual frame in the light version.  More simplifications
to come, followed by light register allocation.
Review URL: http://codereview.chromium.org/1164002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4272 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Since we add more type (StringType, PrimitiveType) the name
NumberInfo does not make sense anymore.

Review URL: http://codereview.chromium.org/1207006

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4268 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

This re-applies r4220 and r4233, which was reverted in r4254 due to a bug. This bug has now been fixed, with the only change being line 2884 changed from

  __ SmiTag(left_side->reg());

to

  __ SmiTag(operand->reg());

Added a regression test.

BUG=http://crbug.com/39160
TEST=test/mjsunit/regress/regress-crbug-39160.js

Review URL: http://codereview.chromium.org/1251009

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4261 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

This reverts svn revisons r4220 and r4233.

Review URL: http://codereview.chromium.org/1279001

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4254 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Inline floating point compare instead of calling the stub when the following conditions are met:
  * Code is in a loop
  * Compare is not a for loop condition
  * Compare is not an equal comparison

This inlined code handles heap number to heap number and heap number to smi compare. It can also handle smi to smi compare, but whenever there is a chance of comparing two smis the smi compare is inlined before the inlined floating point compare. Support for non SSE2 hardware is included.

A new set of variants of the compare stub without the floating point comparison code is called if the inline comapre fails due to the operands not beeing heap numbers or smis.

The virtual frame has been extended with a branch taking two live results to be carried through to the destination. This makes this change much simpler as the inlined code have two live results in registers and a number of bailouts.

CompareStub::GetName needs to be updated as well. I will do that as a separate change.

Also inlined equality check if both operands can't be NaN. This can only provide positive equals if it is the same object.
Review URL: http://codereview.chromium.org/1117011

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4220 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Makes NumberInfo into a real class.
Fix bug where NumberInfo was lost in ToRegister.
Allow 17 bits in safe Smis instead of 16.
Review URL: http://codereview.chromium.org/668151

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@4046 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Introducing a virtual-frame-inl.h file containing some platform-independent
virtual frame function which are small enough to be inlined.

Removed unnecessary #include of virtual-frame.h from register-allocator-inl.h
and added the necessary explicit includes in a number of files.

Review URL: http://codereview.chromium.org/660104

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3962 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

These two asserts in ComputeEntryFrame are wrong since the
virtual frame already knows how to deal with the number type
information of copy elements: We do not store type 
information with copy elements. Instead the backing element
contains the type information.


Review URL: http://codereview.chromium.org/652044

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3927 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Each frame element gets a new attribute with number type information. A frame element can be: 

- smi 
- heap number 
- number (i.e. either of the above) 
- or something else. 

The type information is propagated along with all virtual frame operations. 
Results popped from the frame carry the number information with them.

Two optimizations in the code generator make use of the new 
information: 
- GenericBinaryOpSyub omits map checks if input operands are numbers. 
- Boolean conversion for numbers: Emit inline code for converting a number (smi or heap number) to boolean. Do not emit call to ToBoolean stub in this case.


Review URL: http://codereview.chromium.org/545007

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@3861 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

overhead on the ARM platform.

Because virtual frames are always in memory, there is no work to be
done for control flow except in the case of breaking or continuing
when there is extra statement state on the stack.

Review URL: http://codereview.chromium.org/165230

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2655 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

The complexity of static type inference never really paid off, and as an
analysis it didn't fit well with codegen frame propagation.

Review URL: http://codereview.chromium.org/146077


git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2262 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

class.  Manually manage off-frame results at the two sites where they
were used.

Review URL: http://codereview.chromium.org/125048

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2157 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

branches, and bind with more than two arguments).  Remove unneeded
variable (the "watermark" on virtual frames).

Review URL: http://codereview.chromium.org/125045

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2153 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

to the platform-specific reserved registers.  They are always in use
for their intended purpose, cannot appear in the virtual frame, and
can be freely used without allocation in the code generator.

Review URL: http://codereview.chromium.org/113837

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2061 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

This issue was raised by Brett Wilson while reviewing my changelist for readability. Craig Silverstein (one of C++ SG maintainers) confirmed that we should declare one namespace per line. Our way of namespaces closing seems not violating style guides (there is no clear agreement on it), so I left it intact.

Review URL: http://codereview.chromium.org/115756


git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2038 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

assignment operator) and add an AddAll method to lists instead.
Review URL: http://codereview.chromium.org/115705

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2033 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

macro assembler pointers and all derived state.

Review URL: http://codereview.chromium.org/115564

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2011 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Review URL: http://codereview.chromium.org/113582

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@2000 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

two words: there is no reason to keep a pointer to the current code
generator and macro assembler in the JumpTarget.
Review URL: http://codereview.chromium.org/113458

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1987 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

M    src/parser.cc


git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1975 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1973 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Review URL: http://codereview.chromium.org/115402

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1969 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

JumpTarget class in favor of using other internal state.

Review URL: http://codereview.chromium.org/113456

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1966 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Add virtual destructor to jump targets to make compiler happy.
Review URL: http://codereview.chromium.org/113396

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1951 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Review URL: http://codereview.chromium.org/115350

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1950 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

encoding the values in one word and by using an indirection table for
handles.

This reduces compilation time by roughly 10% and we should be able to make the slow case equality checking of frame elements faster as well.
Review URL: http://codereview.chromium.org/115347

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1949 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

- Compacted the VirtualFrame representation.
-> Improved the compiler speed with 10%

Review URL: http://codereview.chromium.org/115345

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1947 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

entry labels where it is statically safe.

Review URL: http://codereview.chromium.org/115296

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1943 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Review URL: http://codereview.chromium.org/113329

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1930 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

that sets the static type on elements in the entry frame.  Instead set
the static type as early as possible when computing elements.

Because the elements are shared with the reaching frames, this has the
side effect of sometimes changing the static type of reaching frame
elements.  They will be changed anyway in VirtualFrame::PrepareMergeTo
immediately after, so this is safe.

Review URL: http://codereview.chromium.org/113258

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1928 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

immediately when putting a copy in the entry frame, rather than as
part of a separate pass after fully constructing the entry fraem.

Review URL: http://codereview.chromium.org/113198

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1907 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

over the frame elements to find registers used (before allocating
undetermined elements) and another pass to set the frame's register
indices (after allocating all elements).
Review URL: http://codereview.chromium.org/114018

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1906 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Uses static string type to optimize string additions.


git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1675 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

constants and copies below the watermark on backward jump targets
(that's the whole point!).
Review URL: http://codereview.chromium.org/60044

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1670 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

counting the reference to the return value and passing it to the
return label.  This requires threading it through try/catch and
try/finally.  The return value is loaded into eax more lazily than
before.

Also, perform some related refactoring of jump targets.
Review URL: http://codereview.chromium.org/56172

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1669 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Review URL: http://codereview.chromium.org/42565

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1605 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

frame elements.
Review URL: http://codereview.chromium.org/53005

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1593 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

destructor to no longer invalidate the result.

Review URL: http://codereview.chromium.org/50051

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1560 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

forward declarations.

Review URL: http://codereview.chromium.org/42389

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1559 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

been made of the element.  Set it when copying an element, clear it
when writing to an element marked copied and discovering that it is
not copied.

The copied flags on entry frames are exact.  Preparing a frame for
merging set the copied flags to be exact (to increase the likelihood
of matching the entry frame or another frame).
Review URL: http://codereview.chromium.org/50005

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@1546 ce2b1a6d-e550-0410-aec6-3dcde31c8c00