This allowed the following changes:
- heap profiler now makes one pass less over the heap.
- HeapEntriesMap does not allocate EntryInfo per each entry.
- there's no need for an extra pass to set indexes before serialization.
As a result snapshot taking time has reduced up to 2x times.
Review URL: https://chromiumcodereview.appspot.com/10353010
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@11531 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
This makes back pointers in the map transition tree explicit by having
accurate back pointers throughout the lifetime of maps instead of
establishing and destroying back pointers before and after each marking
phase. This is a prerequisite for being able to clear map transitions
during incremental marking.
R=vegorov@chromium.org
BUG=v8:1465
Review URL: https://chromiumcodereview.appspot.com/10381053
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@11528 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
tells us that a map can transition to another map when we are generating
code for load operations. This may cause us to deopt if the same routine
is seeing different maps caused by branching in constructors. If so, I
have a different change that is around 100 times more complicated that
lets us generated Crankshaft code for negative lookups.
Review URL: https://chromiumcodereview.appspot.com/10306010
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@11524 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
Port r11492 (d14ada19)
Original commit message:
Fix register clobbering in LoadIC for interceptors.
This fixes a corner-case where the receiver register was clobbered by
LoadICs for interceptors and inlined followup code still relied on the
receiver to be intact in case of prototype changes.
BUG=
TEST=
Review URL: https://chromiumcodereview.appspot.com/10315016
Patch from Akos Palfi <palfia@homejinni.com>.
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@11506 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
Port r11491 (705d40cc)
Original commit message:
Implement clearing of CompareICs.
This allows CompareICs to be cleared during garbage collection to avoid
cross-context garbage retention through maps stored in CompareIC stubs
for the KNOWN_OBJECTS state.
BUG=
TEST=
Review URL: https://chromiumcodereview.appspot.com/10342024
Patch from Akos Palfi <palfia@homejinni.com>.
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@11505 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
Port r11454 (72c662fc)
Original commit message:
Reduce size of LIR instruction by one word and remove dead code.
Until now we always recorded two deoptimization environments for instructions
that are marked as calls. We actually don't need two for all LIR
instructions except one (LInstanceOfKnownGlobal) where there is a lazy
deoptimization point in deferred code.
This change remove on of them and uses one virtual function instead
to make LInstanceOfKnownGlobal work as before.
Additionally, this change removes an unused predicate save_doubles_ from LIR
instructions and removes some helper functions that are used only in one place.
BUG=
TEST=
Review URL: https://chromiumcodereview.appspot.com/10233019
Patch from Akos Palfi <palfia@homejinni.com>.
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@11501 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
AccessorPairs can now contain map transitions, which is similar to our current
handling of CONSTANT_FUNCTION/CONSTANT_TRANSITION, but generalized to a pair for
holding info about the getter and the setter. This way we can achieve map
sharing for objects with accessor properties, which is a prerequisite for making
them fast via inlining. We fall back to the previous way of handling accessor
properties when sharing is not possible or we don't handle a special case.
Note: When an exisiting accessor property is redefined we could in principle
move the AccessorPair out of the descriptor into the object itself (again just
like the way we do something similar for CONSTANT_FUNCTION/CONSTANT_TRANSITION),
but this would require a new property kind for holding a pair of values. Perhaps
we can implement this later, but for now this hopefully rare case is handled
like before, losing map sharing and potentially creating more maps than strictly
necessary.
Review URL: https://chromiumcodereview.appspot.com/10238005
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@11496 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
Port r11470 (621f96c0)
Original commit message:
Fix LFastLiteral to check boilerplate elements kind.
Adds a missing check that the elements kind of the boilerplate object
still has the expected elements kind, unoptimized code can transition
the boilerplate. Corner cases might cause the optimized code to be
reentered again.
BUG=
TEST=
Review URL: https://chromiumcodereview.appspot.com/10271018
Patch from Akos Palfi <palfia@homejinni.com>.
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@11474 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
Until now we always recorded two deoptimization environments for instructions
that are marked as calls. We actually don't need two for all LIR
instructions except one (LInstanceOfKnownGlobal) where there is a lazy
deoptimization point in deferred code.
This change remove on of them and uses one virtual function instead
to make LInstanceOfKnownGlobal work as before.
Additionally, this change removes an unused predicate save_doubles_ from LIR
instructions and removes some helper functions that are used only in one place.
Review URL: https://chromiumcodereview.appspot.com/10035021
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@11454 ce2b1a6d-e550-0410-aec6-3dcde31c8c00