This patch finally allows running and passing tests with vector-based
Load and KeyedLoad ICs.

BUG=
R=jkummerow@chromium.org

Review URL: https://codereview.chromium.org/767743002

Cr-Commit-Position: refs/heads/master@{#25800}

port r24319.

original commit message:

  Replace OStream with std::ostream.

BUG=
R=weiliang.lin@intel.com

Review URL: https://codereview.chromium.org/635203002

Patch from Chunyang Dai <chunyang.dai@intel.com>.

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@24447 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

It has been turned on by default for a long time, and hydrogenized BinaryOpStubs actually depend on it being turned on.

BUG=v8:3487
LOG=n
R=ishell@chromium.org

Review URL: https://codereview.chromium.org/630023002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@24415 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Review URL: https://codereview.chromium.org/618643002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@24319 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

   The test mjsunit/regress/regress-sqrt compares the result of Math.sqrt function
   when using full-compiler and crankshaft compiler seperately. But according to
   glibc bug fixing(https://sourceware.org/bugzilla/show_bug.cgi?id=14032). The
   glibc implementation of std::sqrt() (It is invoked in the generated code when
   full-compiler is used.) will change since glibc 2.19.

   In order to keep consistence of Math.sqrt translation in crankshaft compiler
   and the pass of mjsunit/regress/regress-sqrt. we translate the Math.sqrt func
   by calling the runtime function.

BUG=
R=weiliang.lin@intel.com

Review URL: https://codereview.chromium.org/606403002

Patch from Chunyang Dai <chunyang.dai@intel.com>.

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@24271 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

port r24161.

original commit message:
  Refactor bailout reasons and disable optimization in more cases.

BUG=
R=weiliang.lin@intel.com

Review URL: https://codereview.chromium.org/597913002

Patch from Chunyang Dai <chunyang.dai@intel.com>.

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@24190 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

BUG=
R=weiliang.lin@intel.com

Review URL: https://codereview.chromium.org/579713002

Patch from Chunyang Dai <chunyang.dai@intel.com>.

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@24102 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

BUG=
R=ishell@chromium.org

Review URL: https://codereview.chromium.org/551803005

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@23969 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

port r23778.

original commit message:

  Get CallInterfaceDescriptor directly from CodeStub.

BUG=
R=weiliang.lin@intel.com

Review URL: https://codereview.chromium.org/562983003

Patch from Jing Bao <jing.bao@intel.com>.

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@23894 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

port r23773.

original commit message:

  Initialize CodeStubInterfaceDescriptor in the constructor.

BUG=
R=weiliang.lin@intel.com

Review URL: https://codereview.chromium.org/563133002

Patch from Jing Bao <jing.bao@intel.com>.

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@23890 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

port r23772.

original commit message:

  To aid vector-based load ic work, we need to be able to handle
  the megamorphic load case in hydrogen. A simple approach is to
  wrap the probe activity in a hydrogen instruction.

  The instruction is novel in that it always tail-calls away.

BUG=
R=weiliang.lin@intel.com

Review URL: https://codereview.chromium.org/564683002

Patch from Jing Bao <jing.bao@intel.com>.

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@23889 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

port r23744.

original commit message:

  Do not cache CodeStubInterfaceDescriptor on the isolate.

BUG=
R=weiliang.lin@intel.com

Review URL: https://codereview.chromium.org/560793002

Patch from Jing Bao <jing.bao@intel.com>.

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@23853 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Tweak LConstantD and LStoreKeyed to avoid fp register spilling

BUG=
R=weiliang.lin@intel.com

Review URL: https://codereview.chromium.org/545673003

Patch from Jing Bao <jing.bao@intel.com>.

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@23710 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

port r23639

original commit message:

  Make concrete classes for individual call descriptors. The
  ic-convention classes that hold register specifications are
  merged into these new call descriptor classes, which should
  represent a final home for that information.

BUG=
R=weiliang.lin@intel.com

Review URL: https://codereview.chromium.org/544943002

Patch from Jing Bao <jing.bao@intel.com>.

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@23708 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

port r23560.

original commit message:

  InterfaceDescriptor becomes CallInterfaceDescriptor.

  There was no difference between these two classes in a hierarchical relationship.

BUG=
R=weiliang.lin@intel.com

Review URL: https://codereview.chromium.org/530903002

Patch from Chunyang Dai <chunyang.dai@intel.com>.

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@23576 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

This adds context deoptimization to Turbofan and Crankshaft (also submitted separately as https://codereview.chromium.org/515723004/).

The second patchset removes the deoptimization/continuation block from calls.

BUG=
R=bmeurer@chromium.org

Review URL: https://codereview.chromium.org/522873002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@23547 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

port r23391.

original commit message:
  Move register conventions out of the IC classes.
  A change to a convention shouldn't require recompilation of ic.h/.cc.

BUG=
R=weiliang.lin@intel.com

Review URL: https://codereview.chromium.org/513533003

Patch from Chunyang Dai <chunyang.dai@intel.com>.

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@23428 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

This way we don't clash with the ASSERT* macros
defined by GoogleTest, and we are one step closer
to being able to replace our homegrown base/ with
base/ from Chrome.

R=jochen@chromium.org, svenpanne@chromium.org

Review URL: https://codereview.chromium.org/430503007

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22812 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

R=svenpanne@chromium.org
TEST=mozilla (with new --always-opt)

Review URL: https://codereview.chromium.org/423093009

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22803 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

port r22709.

This patch only covers the code change of orginal X87 port.
For the changes under ./src/compiler/ and ./tools/gyp/v8.gyp file,
I will submit it in another seperate patch.

BUG=
R=weiliang.lin@intel.com

Review URL: https://codereview.chromium.org/430423002

Patch from Chunyang Dai <chunyang.dai@intel.com>.

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22778 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

port r22665.

original commit message:
 Inline Math.fround in optimized code.

BUG=
R=weiliang.lin@intel.com

Review URL: https://codereview.chromium.org/429143003

Patch from Chunyang Dai <chunyang.dai@intel.com>.

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22691 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

port r22500.

original commit message:

  Introduce FLAG_vector_ics.

  When FLAG_vector_ics is true, then AST nodes that use Load and KeyedLoad ICs
  will allocate a type vector slot to store feedback information. Full codegen
  will emit a load of the slot into a register if the flag is on.

  Support is incomplete, right now the IC doesn't know how to use the feedback
  slot.

BUG=
R=mvstanton@chromium.org

Review URL: https://codereview.chromium.org/405343002

Patch from Chunyang Dai <chunyang.dai@intel.com>.

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22542 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

port r22448

original commit message:
  StubCallInterfaceDescriptor and CallInterfaceDescriptor are unified under a base class InterfaceDescriptor.

  Handling of the context register had to be massaged to effect the unification. This will make it easier
  to call hydrogen code stubs directly from crankshaft.

BUG=
R=weiliang.lin@intel.com

Review URL: https://codereview.chromium.org/399313005

Patch from Chunyang Dai <chunyang.dai@intel.com>.

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22459 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

port r22328

original commit message:

  Use a register spec for StoreIC and KeyedStoreIC.

  This continues refactoring already applied for LoadIC in r22035 (https://code.google.com/p/v8/source/detail?r=22035).

BUG=
R=weiliang.lin@intel.com

Review URL: https://codereview.chromium.org/382123003

Patch from Chunyang Dai <chunyang.dai@intel.com>.

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22343 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

port r22232

original commit message:

This is a mostly mechanical CL (more than 90% Emacs macros and
query-replace-regexp) moving FILE*/StringStream*-based APIs to
OStream-based APIs. There are a few places where this had to stop,
otherwise the CL would be even bigger, but this can easily and
incrementally cleaned up later.

BUG=
R=weiliang.lin@intel.com

Review URL: https://codereview.chromium.org/379493002

Patch from Chunyang Dai <chunyang.dai@intel.com>.

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22259 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

port r22103.

original commit message:
On arm, arm64 and x64 there is a different register specification between LoadIC and KeyedLoadIC.

It would be nicer if these are the same, allowing some key optimizations.

BUG=
R=weiliang.lin@intel.com

Review URL: https://codereview.chromium.org/366583002

Patch from Chunyang Dai <chunyang.dai@intel.com>.

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22122 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

port r22035

original commit message:
  Use IC register definitions in platform files.

BUG=
R=weiliang.lin@intel.com

Review URL: https://codereview.chromium.org/356133005

Patch from Chunyang Dai <chunyang.dai@intel.com>.

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@22047 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

R=rossberg@chromium.org

Review URL: https://codereview.chromium.org/333013002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21894 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Port r21684.

Origin message:
   Preliminary support for block contexts in hydrogen.

   Patch from Steven Keuchel <keuchel@chromium.org>;

   BUG=v8:2198
   LOG=N
   TEST=mjsunit/harmony/block-let-crankshaft.js
R=ulan@chromium.org

Review URL: https://codereview.chromium.org/315233002

Patch from Chunyang Dai <chunyang.dai@intel.com>.

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21730 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

- this avoids using relative include paths which are forbidden by the style guide
- makes the code more readable since it's clear which header is meant
- allows for starting to use checkdeps

BUG=none
R=jkummerow@chromium.org, danno@chromium.org
LOG=n

Review URL: https://codereview.chromium.org/304153016

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21625 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Port r21589 (a9f3228)

BUG=
R=danno@chromium.org

Review URL: https://codereview.chromium.org/309043002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21611 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

TEST=mjsunit/array-shift,mjsunit/array-shift2,mjsunit/array-shift3
R=yangguo@chromium.org

Review URL: https://codereview.chromium.org/308793010

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21592 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

port r21578 (dcf13aa)

Original commit message:
- Move HType to it's own file.
- Add HType::HeapObject and some other useful types.
- Get rid of the broken and useless HType::NonPrimitive.
- Introduce HType::FromType() to convert from HeapType to HType.
- Also add unit tests for HType.
- Fix types in Crankshaft.

BUG=

Review URL: https://codereview.chromium.org/307903002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21579 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

port r21509 (78a54b0)

Original commit message:
This way an HStoreNamedField instruction can never deoptimize
itself, which is another important step towards a working
store elimination.

BUG=
R=danno@chromium.org

Review URL: https://codereview.chromium.org/306483013

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21547 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Support x87-only platform (ia32 without SSE)

R=danno@chromium.org

Review URL: https://codereview.chromium.org/293743005

Patch from Weiliang Lin <weiliang.lin@intel.com>.

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21469 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

R=ulan@chromium.org

Review URL: https://codereview.chromium.org/296113008

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21468 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

R=jkummerow@chromium.org, ulan@chromium.org

Review URL: https://codereview.chromium.org/296993002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21465 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

This is preparation for pending hydrogen stub work that needs to access memory using KeyedLoad/KeyedStore operations where the base offset used for the accesses are is the the default (e.g. the size of an FixedArray header for FixedArrays or zero for external arrays).

R=mvstanton@chromium.org

Review URL: https://codereview.chromium.org/295913009

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21426 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

This reverts commit r21383 for breaking the Mozilla tests.

TBR=hpayer@chromium.org

Review URL: https://codereview.chromium.org/292993003

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21386 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

There's actually no need to have the transition as part of the HStoreNamedField instruction. In fact, it is cleaner and faster to generate a separate HStoreNamedField for the transition map. This will also help to eliminate map stores with store elimination, as well as reduce register pressure for transitioning stores on ia32.

R=hpayer@chromium.org

Review URL: https://codereview.chromium.org/295743002

git-svn-id: https://v8.googlecode.com/svn/branches/bleeding_edge@21383 ce2b1a6d-e550-0410-aec6-3dcde31c8c00