This change includes two CLs by pliard@chromium.org:
1. http://codereview.chromium.org/9447052/ (Add CallOnce() and simple LazyInstance implementation):
Note that this implementation of LazyInstance does not handle global destructors (i.e. the lazy instances are never deleted).
This CL was initially reviewed on codereview.appspot.com:
http://codereview.appspot.com/5687064/
2. http://codereview.chromium.org/9455088/ (Remove static initializers in v8):
This CL depends on CL 9447052 (adding CallOnce and LazyInstance).
It is based on a patch sent by Digit.
With this patch applied, we have only one static initializer left (in atomicops_internals_x86_gcc.cc). This static initializer populates a structure used by x86 atomic operations. It seems that we can hardly remove it. If possible, it will be removed in a next CL.
This CL also modifies the presubmit script to check the number of static initializers.
BUG=v8:1859
Review URL: https://chromiumcodereview.appspot.com/9666052
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@11010 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
While enabling "-fstack-protector", compiler generates code in
function prologue and epilogue to do stack check. However, without
knowing that 'r1', 'r2' and 'r3' is used/destroyed in inline asm,
compiler assumes that 'r1', 'r2', or 'r3' can be used exclusively,
which leads to a core dump.
Fix to this is quite straightforward, just add clobber list to the
inlineasm.
BUG=None
TEST=manually checked the generated asm code,boot up chrome browser successfully with this modification
Review URL: https://chromiumcodereview.appspot.com/9618017
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@10977 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
This is a preparatory step for using the HashMap clas with Literal keys in the
full code generator. It removes some duplicated code already and removes the
need for 2 HashMaps at each use, which was overly complicated.
Removed one dead function on the way.
Review URL: https://chromiumcodereview.appspot.com/9639011
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@10973 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
1. The write barrier (RecordWriteStub) expects that pointer stored
points to an initialized object. Specifically, the map must be set
before it is stored.
2. The backing store for smi-only elements can only be reused for
double elements if it is in new-space. Otherwise, we need to allocate
a fresh one because the old one is in pointer-space and the new one
has to be in data-space.
BUG=117037
Review URL: https://chromiumcodereview.appspot.com/9633017
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@10968 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
All module expressions, and all variables that might refer to modules,
are assigned interfaces (module types) that are resolved using
unification. This is necessary to deal with the highly recursive
nature of ES6 modules, which does not allow any kind of bottom-up
strategy for resolving module names and paths.
Error messages are rudimental right now. Probably need to track
more information to make them nicer.
R=svenpanne@chromium.org
BUG=v8:1569
TEST=
Review URL: https://chromiumcodereview.appspot.com/9615009
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@10966 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
Previously, there were 1 or 2 calls to the runtime when accessors were changed
or set. This doesn't really work well with property attributes, leading to some
hacks and complicates things even further when trying to share maps in presence
of accessors. Therefore, the runtime entry now takes the full triple (getter,
setter, attributes), where the getter and/or the setter can be null in case they
shouldn't be changed.
For now, we do basically the same on the native side as we did before on the
JavaScript side, but this will change in future CLs, the current CL is already
large enough.
Note that object literals with a getter and a setter for the same property still
do 2 calls, but this is a little bit more tricky to fix and will be handled in a
separate CL.
Review URL: https://chromiumcodereview.appspot.com/9616016
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@10956 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
The heap profiler randomly crashed because of memory corruption caused
by unexpected heap objects layout changes occured between count and fill
passes. The changes lead the number of retainers counted on the first pass
did not match its number on the fill pass leading to the out of bounds
array access.
Besides that the mark bit scheme has been changed to a plain vector one in
dominators building algorithm. It is up to 4x faster because of smaller
memory access footprint.
BUG=
TEST=
Review URL: https://chromiumcodereview.appspot.com/9594020
Patch from Alexei Filippov <alexeif@chromium.org>.
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@10928 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
Port r10905 (2a997cf).
Original commit message:
Allow Crankshaft to inline ordered relational comparisons (<, >, <=, >=) that have undefined arguments in addition to double value arguments (rather than calling the generic Compare stub).
BUG=
TEST=
Review URL: https://chromiumcodereview.appspot.com/9583038
Patch from Daniel Kalmar <kalmard@homejinni.com>.
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@10925 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
Inlined strict mode functions (that are not called as methods) will get
their receiver reset to undefined. This should not happen when inlining
constructors.
This change also simplifies the test suite to reuse the same closures
into which constructors get inlined and use gc() to force V8 to forget
collected type feedback.
R=vegorov@chromium.org
TEST=mjsunit/compiler/inline-construct
Review URL: https://chromiumcodereview.appspot.com/9597017
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@10920 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
This is essential for the upcoming map sharing with accessors, and we aleady
have enough machinery to handle the holes now. Furthermore, use "To" template in
2 nearby functions, and made naming a bit more consistent.
In a nutshell: This CL should have no visible effect at all at the moment.
Famous last words...
Review URL: https://chromiumcodereview.appspot.com/9594013
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@10919 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
Main change from the original CL: Call::ComputeTarget does not use IsProperty
anymore, because this would potentially need a holder, which we don't have
here. Using Map::LookupInDescriptors with a NULL holder is a bit fishy in
general, because one has to be *extremely* careful when using its LookupResult.
The original CL made Chrome's NetInternalsTest.netInternalsTourTabs browser test
fail, but it's a mystery how this could happen: We should never reach
Call::ComputeTarget via Call::RecordTypeFeedback with a CALLBACKS property,
because we never consider calls to them monomorphic, which is in turn because of
the stub cache leaving them in the pre-monomorphic state. Therefore, I don't
have a clue how to write a regression test for this...
As an additional tiny bonus, the --trace-opt output for deoptimizations has been
improved.
Review URL: https://chromiumcodereview.appspot.com/9584003
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@10906 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
The old code used a separate HToInt32 instruction which had a wrong register
constraint for the input register which caused wrong result when the stored value
is used after a typed array store. (UseRegister instead of UseTempRegister) when no
SSE3 is available.
This change fixes it by replacing HToInt32 with the corresponding HChange
instruction which has correct register contraints.
TEST=mjsunit/compiler/regress-toint32.js
Review URL: https://chromiumcodereview.appspot.com/9565007
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@10891 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
Port r10881 (0d25c61e).
Original commit message:
Generates inlined code for object allocation specific to the initial map
of the given constructor function. Also forces completion of inobject
slack tracking while crankshafting to finalize instance size of these
objects.
BUG=
TEST=
Review URL: https://chromiumcodereview.appspot.com/9569008
Patch from Daniel Kalmar <kalmard@homejinni.com>.
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@10890 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
Port r10794 (654fe910).
Original commit message:
Only JSObject enumerables with enum cache (fast case properties, no interceptors, no enumerable properties on the prototype) are supported.
HLoadKeyedGeneric with keys produced by for-in enumeration are recognized and rewritten into direct property load by index. For this enum-cache was extended to store property indices in a separate array (see handles.cc).
New hydrogen instructions:
- HForInPrepareMap: checks for-in fast case preconditions and returns map that contains enum-cache;
- HForInCacheArray: extracts enum-cache array from the map;
- HCheckMapValue: map check with HValue map instead of immediate;
- HLoadFieldByIndex: load fast property by it's index, positive indexes denote in-object properties, negative - out of object properties;
Changed hydrogen instructions:
- HLoadKeyedFastElement: added hole check suppression for loads from internal FixedArrays that are knows to have no holes inside.
BUG=
TEST=
Review URL: https://chromiumcodereview.appspot.com/9453009
Patch from Daniel Kalmar <kalmard@homejinni.com>.
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@10865 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
Port r10849 (b0fe79c).
Also included: Fixed a bug in GenerateRecordCallTarget.
This bug prevented certain functions from being registered as Monomorphic
and thus prevented them from being inlined using the new system (b0fe79c).
BUG=
TEST=
Review URL: https://chromiumcodereview.appspot.com/9511002
Patch from Daniel Kalmar <kalmard@homejinni.com>.
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@10862 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
This CL is a step towards removing ZoneObject's new operator without a Zone
parameter, which uses Isolate::Current. For e.g. the bulletben benchmark, this
CL reduces the number of calls to this new operator by roughly 120k, but we are
still left with 780k calls from other sites...
Review URL: https://chromiumcodereview.appspot.com/9487010
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@10860 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
With transitions in AccessorPairs, it is not enough to look at the PropertyType
alone to decide whether we look at a property or not: For objects with
JavaScript accessors, we have to look into the AccessorPair itself and see if
one of its 2 parts is actually a JavaScript accessor. Therefore, a predicate
with a PropertyType argument alone doesn't make sense anymore, we might need the
associated value, too.
Things are complicated by the fact that the holder in a LookupResult can be
NULL, so we must be careful to retrieve its value only when it is really
needed. To achieve the needed call-by-name semantics, a new Entry is introduced,
which is basically a closure over a DescriptorArray and an index into this array
(C++0x to the rescue!). GCC is clever enough to inline this class, so we pay no
runtime penalty for this abstraction.
It's all a bit ugly, but this is caused by the current structure of Descriptor,
DescriptorArray and LookupResult: Things would be much easier if DescriptorArray
were, well, an array of Descriptors, and LookupResult were a 'Maybe Descriptor'
(in Haskell-terms).
Review URL: https://chromiumcodereview.appspot.com/9466047
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@10847 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
This patch changes the signature of the v8::HeapGraphNode::GetRetainedSize method, but it's not used in Chromium, and it should be easy for other clients (if any) to adjust to this change.
BUG=none
TEST=none
Review URL: https://chromiumcodereview.appspot.com/9466014
Patch from Alexei Filippov <alexeif@chromium.org>.
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@10846 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
As of dominators and retained sizes calculation take quite small time now
comparing to the main passes, it is worth to exclude these from progress
indicator. Now the indicator smoothly runs to 100%, while previously
it ran to 50% and then instantly jumped to 100%.
BUG=none
TEST=none
Review URL: https://chromiumcodereview.appspot.com/9465010
Patch from Alexei Filippov <alexeif@chromium.org>.
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@10831 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
This allows elements of the non-strict arguments object to be redefined
with custom attributes and still maintain an alias into the context.
Such a slow alias is maintained by placing a special marker into the
dictionary backing store of the arguments object.
R=rossberg@chromium.org
BUG=v8:1772
TEST=test262,mjsunit/object-define-property
Review URL: https://chromiumcodereview.appspot.com/9460004
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@10827 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
As of dominators and retained sizes calculation take quite small time now
comparing to the main passes, it is worth to exclude these from progress
indicator. Now the indicator smoothly runs to 100%, while previously
it ran to 50% and then instantly jumped to 100%.
BUG=none
TEST=none
Review URL: https://chromiumcodereview.appspot.com/9463008
Patch from Alexei Filippov <alexeif@chromium.org>.
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@10823 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
Replace timestamps with affected bool vector. Timestamps could cause
some entries marked as affected on iteration i, to be recalculated
twice on iterations i and i+1. Which is redundant.
BUG=none
TEST=none
Review URL: https://chromiumcodereview.appspot.com/9467002
Patch from Alexei Filippov <alexeif@chromium.org>.
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@10821 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
The previous code relied on the tricky global invariant that there is no map
sharing when accessor properties are involved (or in other words: that
TransformToFastProperties is dumb enough :-). Although this is not a real
problem with the current code, this assumption breaks when map sharing in fast
mode is enabled, so we defensively copy an AccessorPair.
Review URL: https://chromiumcodereview.appspot.com/9430048
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@10810 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
This refactors the way we (re)define elements to perform normalization
and attribute updating at a much deeper level, thereby removing some
bogus special cases in upper runtime layers.
Most element setters take an indicator flag that distinguishes between
setting and defining. Setting of an element causes attributes to remain
unchanged, writability to be checked and callbacks to be called.
Defining of an element causes attributes to be updated and callbacks to
be overridden. The same approach could be taken for properties.
R=svenpanne@chromium.org
BUG=v8:1772
TEST=test262,test262/15.2.3.6-4-333-11
Review URL: https://chromiumcodereview.appspot.com/9443014
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@10808 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
The old HashMap class had an explicit member to determine the allocation
policy. The template version matches the approach used already for
lists.
Cleanup some include dependencies and unnecessary forward declarations.
Cleanup some dead code from isolate.h and replace some HEAP macros
with GetHeap().
Review URL: https://chromiumcodereview.appspot.com/9372106
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@10806 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
It is achieved by:
1. skipping entries those dominators have already reached root.
2. processing only entries those retainers have changed their
dominators and skipping other entries.
3. removing extra memory indirection by making the dominators array
contain entry indices instead of entries themselves.
The dominators building time has dropped from ~4000 ms to ~200 ms
on gmail.com heap snapshot.
BUG=none
TEST=none
Review URL: https://chromiumcodereview.appspot.com/9372105
Patch from Alexei Filippov <alexeif@chromium.org>.
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@10799 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
Modify PreProcessOsrEntry to work with OSR entries that have non-empty expression stack.
Modify graph builder to take for-in state from environment instead of directly referencing emitted instructions.
Extend %OptimizeFunctionOnNextCall with an argument to force OSR to make writing OSR tests easier: %OptimizeFunctionOnNextCall(f, "osr").
R=fschneider@chromium.org
TEST=test/mjsunit/compiler/optimized-for-in.js
Review URL: https://chromiumcodereview.appspot.com/9431030
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@10796 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
Only JSObject enumerables with enum cache (fast case properties, no interceptors, no enumerable properties on the prototype) are supported.
HLoadKeyedGeneric with keys produced by for-in enumeration are recognized and rewritten into direct property load by index. For this enum-cache was extended to store property indices in a separate array (see handles.cc).
New hydrogen instructions:
- HForInPrepareMap: checks for-in fast case preconditions and returns map that contains enum-cache;
- HForInCacheArray: extracts enum-cache array from the map;
- HCheckMapValue: map check with HValue map instead of immediate;
- HLoadFieldByIndex: load fast property by it's index, positive indexes denote in-object properties, negative - out of object properties;
Changed hydrogen instructions:
- HLoadKeyedFastElement: added hole check suppression for loads from internal FixedArrays that are knows to have no holes inside.
R=fschneider@chromium.org
BUG=
TEST=
Review URL: https://chromiumcodereview.appspot.com/9425045
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@10794 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
This CL is an intermediate step only, in the end we need to have a single
DefineOrRedefineAccessorProperty call for a single Object.defineProperty
call. Currently we can end up making two such calls, making the necessary access
checks extremely ugly and hard (impossible?) to get right for complete spec
conformance.
The bulk of the change is quite mechanical:
* Prepare an AccessorPair *before* we add it to our data structures,
eliminating the previous voodoo-like threading of a placeholder.
* The previous item makes it possible to activate our check that we do not
share AccessorPairs by accident.
* Split a monster method into 2 quite unrelated methods.
* Use templated To method in a few places.
Review URL: https://chromiumcodereview.appspot.com/9428026
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@10788 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
Port r10700 (c976fbd5).
Original commit message:
This change enables optimization of top-level and eval-code. For this to work, it adds
support for declaring global variables in optimized code.
At the same time it disables the eager generation of deoptimization support data
in the full code generator (originally introduced in
r10040). This speeds up initial compilation and saves
memory for functions that won't be optimized. It requires
recompiling the function with deoptimization
support when we decide to optimize it.
BUG=
TEST=
Review URL: https://chromiumcodereview.appspot.com/9418006
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@10773 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
The bulk of this CL is purely mechanical: Make the CONVERT_FOO macros more
uniform by always using an index instead of an object. Apart from this, it
includes a few minor changes like using CONVERT_SMI_ARG_CHECKED a bit more or
introducing a new macro for PropertyDetails. Nothing spectacular, just something
sitting on my disk for quite some time now...
Review URL: https://chromiumcodereview.appspot.com/9395075
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@10772 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
A GC in the access check callbacks NamedSecurityCallback and IndexedSecurityCallback
violates the contract about these callbacks.
Added a EXTERNAL VMState scope around the call to FailedAccessCheckCallback to be
consistent with the other callback invocations.
BUG=v8:1952
TEST=cctest/test-api/TurnOnAccessCheckAndRecompile
Review URL: https://chromiumcodereview.appspot.com/9425048
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@10764 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
* Turned all uses of 'const' into 'var'.
* Turned all uses of local 'function' into 'var'.
* Added a couple of missing toplevel 'var' declarations.
One consequence is that the properties on the builtin object are no longer
non-writable, and I had to adapt one test. Is that a problem?
Unfortunately, we cannot actually switch the library scripts to strict mode
by default, because that makes observable things like poisoned .caller properties
for library functions.
Also removed dead flag code in Compiler::Compile.
R=yangguo@chromium.org
BUG=
TEST=
Review URL: https://chromiumcodereview.appspot.com/9415010
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@10758 ce2b1a6d-e550-0410-aec6-3dcde31c8c00