Maps in monomorphic Load, KeyedLoad, Store, KeyedStore, and CompareNil IC
stubs are treated as weak references by the marking visitor.

During generation of an IC stub with a weak map, the stub is appended to the
dependent code array of the map. When the map dies, all stubs in its dependent
code array are invalidated by setting embedded maps to undefined.

BUG=v8:2073
LOG=Y
TEST=cctest/test-heap/WeakMapInMonomorphic*IC
R=mstarzinger@chromium.org, verwaest@chromium.org

Review URL: https://codereview.chromium.org/188783003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@20679 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

This reverts commit r20676.

TBR=danno@chromium.org

Review URL: https://codereview.chromium.org/233013005

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@20678 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

R=rossberg@chromium.org

Review URL: https://codereview.chromium.org/234743003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@20677 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Setting the FPSCR flags is expensive on some CPUs. Get rid of repeated
setting of the FPSCR by relying on the correct default flags being set
when doing uint8 clamping. Also use vcvt_u32_f64 instead of vcvt_s32_f64,
which enables removing the check against zero (vcvt_u32_f64 will clamp to
zero).

To be on the safe side, add asserts to check that the VFP rounding mode
flags are set to default as expected.

This increases performance of a hot loop repeatedly setting
Uint8ClampedArray values on some CPUs by as much as a factor of 12.

BUG=v8:3253
LOG=N
R=jacob.bramley@arm.com, rmcilroy@chromium.org, ulan@chromium.org

Review URL: https://codereview.chromium.org/230473005

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@20676 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

R=hpayer@chromium.org

Review URL: https://codereview.chromium.org/234703003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@20675 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

The ranges are simply wrong after range analysis, and we should only rely on computed flags.

R=bmeurer@chromium.org

Review URL: https://codereview.chromium.org/234583005

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@20674 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Deferred error formatting on GC was a failed experiment. It has been
removed, except for this part.  See r13371.

R=mstarzinger@chromium.org

Review URL: https://codereview.chromium.org/233163004

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@20672 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

R=ulan@chromium.org

Review URL: https://codereview.chromium.org/225823003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@20669 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

R=svenpanne@chromium.org

Review URL: https://codereview.chromium.org/233673007

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@20667 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Avoid type assertion on object comparison in Hydrogen - the comparison is unreachable because of previous checks.

BUG=
R=yangguo@chromium.org

Review URL: https://codereview.chromium.org/232053004

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@20666 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

R=bmeurer@chromium.org

Review URL: https://codereview.chromium.org/232883002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@20665 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

BUG=360611
LOG=y
R=bmeurer@chromium.org

Review URL: https://codereview.chromium.org/225393005

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@20664 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

There is no definition for HArgumentsObject, so LDummyUse confuses the register allocator. I have recently made similar fix for HCapturedObject (see https://codereview.chromium.org/222283002/).

BUG=
R=mstarzinger@chromium.org

Review URL: https://codereview.chromium.org/226613007

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@20663 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

R=verwaest@chromium.org

Review URL: https://codereview.chromium.org/232533002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@20662 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Fix a couple of places were x8 and x9 are excluded from lists of saved
registers. These are caller-saved registers, so C code can corrupt them.

x8 and x9 were originally reserved for debug code in the ARM64 port, so
we didn't bother preserving them, but they are now normal allocatable
registers.

BUG=v8:3263
LOG=N
R=rmcilroy@chromium.org, ulan@chromium.org

Review URL: https://codereview.chromium.org/233373002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@20658 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Semantics of elements accessors are now preserved in all optimized code paths
through Array.push(). Previously it was possible to have inconsistent behavior
between optimized and unoptimized code, and there were cases where element
accessors were completely ingored.

R=verwaest@chromium.org

Review URL: https://codereview.chromium.org/232873002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@20655 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

During handlification of TransitionArray code, an error was introduced
in TransitionArray::CopyInsert because after creating a copy of a
TransitionArray, it may be that the array disappears during GC
because it is modified during the marking of the owning map.

R=verwaest@chromium.org

Review URL: https://codereview.chromium.org/228483005

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@20654 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Reason: test failure with --noenable-sse3.

R=dcarney@chromium.org
TBR=dcarney@chromium.org

Review URL: https://codereview.chromium.org/232933003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@20653 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

BUG=
R=dcarney@chromium.org, yangguo@chromium.org

Review URL: https://codereview.chromium.org/230693003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@20652 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

BUG=
R=mvstanton@chromium.org

Review URL: https://codereview.chromium.org/231283006

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@20651 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

TEST=cctest/test-types
R=rossberg@chromium.org

Review URL: https://codereview.chromium.org/232913002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@20650 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

R=bmeurer@chromium.org
BUG=

Review URL: https://codereview.chromium.org/232843002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@20649 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

This reverts r20646.

TBR=verwaest@chromium.org

BUG=

Review URL: https://codereview.chromium.org/232903002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@20648 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

When there is no type feedback yet, ComputeReceiverTypes
should still populate the SmallMapList when the receiver
is a HConstant.

BUG=
R=verwaest@chromium.org

Review URL: https://codereview.chromium.org/230363003

Patch from Petka Antonov <p.antonov@partner.samsung.com>.

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@20646 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Due to the SSA vs. SSI difference, we are only allowed to use the
flags computed during range analysis, not the ranges themselves. For
the case at hand, there is no such flag, so the condition is simply
remvoed.

BUG=361608
LOG=y
R=bmeurer@chromium.org

Review URL: https://codereview.chromium.org/232553004

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@20645 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

R=yangguo@chromium.org

Review URL: https://codereview.chromium.org/231103002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@20644 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

BUG=
R=bmeurer@chromium.org, mstarzinger@chromium.org

Review URL: https://codereview.chromium.org/232593003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@20640 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Add tests for TypeImpl::Of(), TypeImpl::NowOf() and
TypeImpl::NowContains(). Improves the implementation of
TypeImpl::NowIs() to match that of TypeImpl::NowContains().

Mark test-types with NO_VARIANTS to speedup testing, since
the variants do not affect the type system at all.

Also improve test coverage for types.

TEST=cctest/test-types
R=rossberg@chromium.org

Review URL: https://codereview.chromium.org/230673002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@20639 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

BUG=
R=jarin@chromium.org

Review URL: https://codereview.chromium.org/232413003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@20638 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

R=verwaest@chromium.org

Review URL: https://codereview.chromium.org/231013002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@20636 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

R=verwaest@chromium.org

Review URL: https://codereview.chromium.org/222133003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@20635 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

R=verwaest@chromium.org

Review URL: https://codereview.chromium.org/228073004

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@20634 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

R=yangguo@chromium.org

Review URL: https://codereview.chromium.org/230733003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@20631 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

R=hpayer@chromium.org

Review URL: https://codereview.chromium.org/230833002

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@20630 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

R=verwaest@chromium.org

Review URL: https://codereview.chromium.org/228333003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@20628 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

This reverts r20603.

BUG=

Review URL: https://codereview.chromium.org/230863003

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@20626 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

R=yangguo@chromium.org

Review URL: https://codereview.chromium.org/229943006

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@20621 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

ClusterFuzz test is triggering assertions for dead code. This fix issues
HDeoptimize instruction when it finds out that the compare instruction
is dead (because of previous checks).

R=yangguo@chromium.org
BUG=359491
LOG=N

Review URL: https://codereview.chromium.org/228883005

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@20620 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

R=ishell@chromium.org

Review URL: https://codereview.chromium.org/229373007

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@20619 ce2b1a6d-e550-0410-aec6-3dcde31c8c00

Reland of r20606: Further ElementsAccessor handlification (GetKeyForIndex(), GetCapacity(), GetType() and GetAttributes()).
It was not related to test failures.

R=yangguo@chromium.org

Review URL: https://codereview.chromium.org/230373007

git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@20616 ce2b1a6d-e550-0410-aec6-3dcde31c8c00