This adds a pointer to the shared function info into deoptimization data of an optimized code. Whenever the code is deoptimized, it clears the cache in the shared function info.
This fixes the problem when the optimized function dies in new space GC before the code is deoptimized due to code dependency and before the optimized code cache is cleared in old space GC (see mjsunit/regress/regress-343609.js).
This partially reverts r19603 because we need to be able to evict specific code from the optimized code cache.
BUG=343609
LOG=Y
TEST=mjsunit/regress/regress-343609.js
R=yangguo@chromium.org
Review URL: https://codereview.chromium.org/184923002
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@19635 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
This feature makes it possible to associate data with a script and get it back
when the script is compiled or when an event is handled. It was historically
used by Chromium Dev Tools, but not any more. It is not used by node.js.
Note: this has nothing to do with the preparse data, despite the confusing name.
The preparse data is passed as ScriptData*.
R=svenpanne@chromium.org
BUG=
Review URL: https://codereview.chromium.org/184403002
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@19616 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
We already have code that expects this, but incorrectly asserted that the
filler map case would never happen when allocation folding is turned on.
However, even folding has it's limits, bailing out of continued folding
when the object size grows too large. Therefore, it's a general problem
when verifying JSArray objects, that we might encounter a filler map
in elements().
Discovered by ClusterFuzz crbug 347903.
R=hpayer@chromium.org
LOG=N
BUG=347903
Review URL: https://codereview.chromium.org/184493002
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@19604 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
Preparsing is always maximally lazy (every function that can be lazy is preparsed
lazily), but Parser has more complicated laziness logic.
If we're going to parse eagerly, and we have preparse data from lazy preparsing,
we're gonna have a bad time. The symbol stream won't contain symbols inside lazy
functions, and when the Parser parses them eagerly, it will consume symbols from
the symbol stream, and everything will go wrong.
This bug was hidden because the symbol cache was not used for real (see
https://codereview.chromium.org/172753002/ ).
R=ulan@chromium.org
BUG=346207
LOG=Y
Review URL: https://codereview.chromium.org/177973002
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@19532 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
The Parser never used the symbol stream produced by the PreParser for anything
useful, due to a bug introduced 3.5 years ago by
https://codereview.chromium.org/3356010/diff/7001/src/parser.cc.
The bug is that calling Initialize on symbol_cache_ doesn't change its
length. So the length remains 0, and the "if" in Parser::LookupSymbol is always
true, and Parser::LookupCachedSymbol is never called and symbol_cache_ never
filled.
This bug also masked a bug that the symbol stream produced by PreParser doesn't
match what Parser wants to consume. The repro case is the following:
var myo = {if: 4}; print(myo.if);
PreParser doesn't log a symbol for the first "if", but in the corresponding
place, Parser consumes one symbol from the symbol stream. Since the consumed
symbols were never really used, this mismatch went unnoticed.
This CL also fixes that bug.
BUG=
R=ulan@chromium.org
Review URL: https://codereview.chromium.org/172753002
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@19505 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
Refresh the implementation of Symbols to catch up with what the
specification now mandates:
* The global Symbol() function manufactures new Symbol values,
optionally with a string description attached.
* Invoking Symbol() as a constructor will now throw.
* ToString() over Symbol values still throws, and
Object.prototype.toString() stringifies like before.
* A Symbol value is wrapped in a Symbol object either implicitly if
it is the receiver, or explicitly done via Object(symbolValue) or
(new Object(symbolValue).)
* The Symbol.prototype.toString() method no longer throws on Symbol
wrapper objects (nor Symbol values.) Ditto for Symbol.prototype.valueOf().
* Symbol.prototype.toString() stringifies as "Symbol("<description>"),
valueOf() returns the wrapper's Symbol value.
* ToPrimitive() over Symbol wrapper objects now throws.
Overall, this provides a stricter separation between Symbol values and
wrapper objects than before, and the explicit fetching out of the
description (nee name) via the "name" property is no longer supported
(by the spec nor the implementation.)
Adjusted existing Symbol test files to fit current, adding some extra
tests for new/changed behavior.
LOG=N
R=arv@chromium.org, rossberg@chromium.org, arv, rossberg
BUG=v8:3053
Review URL: https://codereview.chromium.org/118553003
Patch from Sigbjorn Finne <sigbjornf@opera.com>.
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@19490 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
This addresses several TODOs:
- Push and Pop requests can be queued up so that arrays of Registers
can be pushed efficiently, with just one PrepareForPush/Pop.
- PushMultipleTimes now takes an Operand. This allows variable-length
arguments arrays to be initialized, for example.
- A NoUseRealAbortsScope has been added to Abort so that
AssertStackConsistency can be called from PrepareForPush without
introducing infinite recursion.
BUG=
R=rmcilroy@chromium.org, ulan@chromium.org
Review URL: https://codereview.chromium.org/170623002
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@19474 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
Code generation would fail when assembling a branch to a label that is bound
outside the immediate range of the instruction. A64 is sensitive to this, as the
various branching instructions have different ranges, going down to +-32KB for
TBZ/TBNZ. The MacroAssembler is augmented to handle branches to targets that
may exceed the immediate range of instructions.
When branching backward to a label exceeding the instruction range, the
MacroAssembler can simply tweak the generated code to use an unconditional
branch with a longer range. For example instead of
B(cond, &label);
the MacroAssembler can generate:
b(InvertCondition(cond), &done);
b(&label);
bind(&done);
Since the target is not known when the branch is emitted, forward branches uses
a different mechanism. The MacroAssembler keeps track of forward branches to
unbound labels. When the code generation approaches the end of the range of a
branch, a veneer is generated for the branch.
BUG=v8:3148
LOG=Y
R=ulan@chromium.org
Review URL: https://codereview.chromium.org/169893002
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@19444 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
Arithmetic right shifting is *not* division in two's complement
representation, only in one's complement. So we convert to one's
complement, shift, and go back to two's complement. By permutating the
last steps, one can get efficient branch-free code. This insight comes
from the paleozoic era of computer science, see the paper from 1976:
Guy Lewis Steele Jr.: "Arithmetic Shifting Considered Harmful"
ftp://publications.ai.mit.edu/ai-publications/pdf/AIM-378.pdf
This results in better and more correct code than our previous
"neg/shift/neg" dance.
LOG=y
BUG=v8:3151
R=bmeurer@chromium.org
Review URL: https://codereview.chromium.org/166793002
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@19434 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
Notes:
- We use simple recursion to keep track of how many "new" operators we have seen
and where.
- This makes the self-baked stack class PositionStack in parser.cc unnecessary.
- Now the logic is also unified between Parser and PreParser.
- This is a fixed version of r19386.
R=ulan@chromium.org
BUG=v8:3126
LOG=N
Review URL: https://codereview.chromium.org/168583008
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@19417 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
This addresses several TODOs:
- Push and Pop requests can be queued up so that arrays of Registers
can be pushed efficiently, with just one PrepareForPush/Pop.
- PushMultipleTimes now takes an Operand. This allows variable-length
arguments arrays to be initialized, for example.
- A NoUseRealAbortsScope has been added to Abort so that
AssertStackConsistency can be called from PrepareForPush without
introducing infinite recursion.
BUG=
R=ulan@chromium.org
Review URL: https://codereview.chromium.org/169533002
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@19403 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
Notes:
- We use simple recursion to keep track of how many "new" operators we have seen
and where.
- This makes the self-baked stack class PositionStack in parser.cc unnecessary.
- Now the logic is also unified between Parser and PreParser.
- It might have been a copy-paste artifact (ParseLeftHandSideExpression ->
ParseMemberWithNewPrefixesExpression) that the logic was so complicated
before.
R=ulan@chromium.org
BUG=v8:3126
LOG=N
Review URL: https://codereview.chromium.org/166943002
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@19386 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
The fix removes wrong simulates from the number branch of polymorphic
call/field access handling.
The change also fixes the same thing for polymorphic named field
access even thourgh the field access is probably safe in practice
(because it cannot deoptimize). It is better to keep all our simulates
in sync with full codegen.
R=jkummerow@chromium.org
BUG=
Review URL: https://codereview.chromium.org/166503002
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@19375 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
Notes:
- To be able to move the recursive descent functions to ParserBase one at a
time, we temporarily need routing functions from traits to Parser/PreParser,
since the recursive descent functions form a cyclic structure.
- PreParser used to always allow intrinsic syntax. After this CL, it depends on
allow_natives_syntax() which was already in ParserBase.
- This CL also decouples (Pre)ParserTraits better from (Pre)Parser, passing more
information as parameters, so that the Traits don't need to get it from
(Pre)Parser.
R=ulan@chromium.org
BUG=v8:3126
LOG=N
Review URL: https://codereview.chromium.org/163333003
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@19374 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
This patch generalizes Object.observe callbacks and promise resolution into a FIFO queue called a "microtask queue".
It also exposes new V8 API which exposes the microtask queue to the embedder. In particular, it allows the embedder to
-schedule a microtask (EnqueueExternalMicrotask)
-run the microtask queue (RunMicrotasks)
-control whether the microtask queue is run automatically within V8 when the last script exits (SetAutorunMicrotasks).
R=dcarney@chromium.org, rossberg@chromium.org, dcarney, rossberg, svenpanne
BUG=
Review URL: https://codereview.chromium.org/154283002
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@19344 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
(Second try, with fixes. First try: https://codereview.chromium.org/149913006/ )
The long-term goal is to move all recursive descent functions from Parser and
PreParser into ParserBase, but first they need to be unified.
Notes:
- The functions moved in this CL: ParseIdentifier, ParseIdentifierName,
ParseIdentifierNameOrGetOrSet, ParseIdentifierOrStrictReservedWord.
- IOW, this CL removes Parser::ParseIdentifier and PreParser::ParseIdentifier
and adds ParserBase::ParseIdentifier, etc.
- Error reporting used to require virtual funcs; now error reporting is moved to
the Traits too, and ParserBase no longer needs to be virtual.
- I had to move PreParser::Identifier out of the PreParser class, because
otherwise PreParserTraits cannot use it in a typedef.
BUG=v8:3126
LOG=N
R=mstarzinger@chromium.org
Review URL: https://codereview.chromium.org/158913003
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@19265 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
The long-term goal is to move all recursive descent functions from Parser and
PreParser into ParserBase, but first they need to be unified.
Notes:
- The functions moved in this CL: ParseIdentifier, ParseIdentifierName,
ParseIdentifierNameOrGetOrSet, ParseIdentifierOrStrictReservedWord.
- IOW, this CL removes Parser::ParseIdentifier and PreParser::ParseIdentifier
and adds ParserBase::ParseIdentifier, etc.
- Error reporting used to require virtual funcs; now error reporting is moved to
the Traits too, and ParserBase no longer needs to be virtual.
- I had to move PreParser::Identifier out of the PreParser class, because
otherwise PreParserTraits cannot use it in a typedef.
BUG=v8:3126
LOG=N
R=mstarzinger@chromium.org
Review URL: https://codereview.chromium.org/149913006
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@19230 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
ES6 makes the Function object properties "length" and "name"
configurable; switch the implementation over to follow that.
Doing so exposed a problem in the handling of non-writable, but
configurable properties backed by foreign callback accessors
internally. As an optimization, if such an accessor property is
re-defined with a new value, its setter was passed the new value
directly, keeping the property as an accessor property. However, this
is not correct should the property be non-writable, as its setter will
then simply ignore the updated value. Adjust the enabling logic for
this optimization accordingly, along with adding a test.
LOG=N
R=rossberg@chromium.org, rossberg
BUG=v8:3045
Review URL: https://codereview.chromium.org/116083006
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@19200 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
Still relevant parts of the original commit message:
Unify paren handling in Parser and PreParser.
It is only needed in (Pre)Parser::ParseExpressionOrLabelledStatement for not
recognizing parenthesized identifiers as labels and in
(Pre)Parser::ParseSourceElements for not recognizing a parenthesized string as
directive prologue.
Parser Expressions don't keep track of whether they're parenthesized, so
PreParser Expressions shouldn't either.
BUG=3126
LOG=N
R=ulan@chromium.org
Review URL: https://codereview.chromium.org/148323011
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@19153 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
It is only needed in (Pre)Parser::ParseExpressionOrLabelledStatement for not
recognizing parenthesized identifiers as labels and in
(Pre)Parser::ParseSourceElements for not recognizing a parenthesized string as
directive prologue.
Parser Expressions don't keep track of whether they're parenthesized, so
PreParser Expressions shouldn't either.
In addition, add helper funcs for checking if an Expression is identifier or a
given identifier. (PreParser Expressions can do this.)
BUG=3126
LOG=N
R=ulan@chromium.org
Review URL: https://codereview.chromium.org/150103004
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@19140 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
We used to have error messages which provide context, like "Variable name may
not be eval or arguments in strict mode", but for other illegal words we only
have non-context specific error messages like "Unexpected reserved word".
Providing the context makes the code unnecessarily complex, since every
individual place must remember to check for eval or arguments. This CL produces
a unified error message ("Unexpected eval or arguments in strict mode"), and puts
the error reporting to (Pre)Parser::ParseIdentifier.
Notes:
- The module feature is so experimental, that I decided to not allow "eval" or
"arguments" as module-related identifiers in the strict mode (even though this
check wasn't there before).
- Unfortunately, there were some inconsistencies, since it was the
responsibility of the caller of ParseIdentifier to check "eval" and "arguments"
and some places didn't have the check for no good reason. This CL is supposed to
keep backward compatibility and *not* introduce any new errors.
- ECMA allows "eval" and "arguments" as labels even in strict mode. (Syntax:
"LabelledStatement: Identifier : Statement", and no strict mode restrictions on
Identifier are listed.)
- Tests which compare error message strings will fail, and need to be updated.
BUG=3126
LOG=N
R=ulan@chromium.org
Review URL: https://codereview.chromium.org/152813005
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@19112 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
We need a way to assert that for a given source code snippet, an error *is*
produced or *is not* produced. Otherwise we might accidentally create new
errors or start accepting code which was previously not accepted. Just checking
that Parser and PreParser produce the same result doesn't cut it.
BUG=3126
LOG=N
R=ulan@chromium.org
Review URL: https://codereview.chromium.org/154243005
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@19107 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
This is what I think is a better solution to the "external strings in
old pointer space" problem. Basically, it is an issue because GC scans
all fields of objects in old pointer space and if the cached address
of the backing store is unaligned, it looks like a heap object, boom.
The solution here is to use short external strings when we externalize
a string in old pointer space, and when the address is unaligned.
Short external strings don't cache the address, so GC has no issues.
BUG=268686
LOG=Y
R=dcarney@chromium.org
Review URL: https://codereview.chromium.org/146183006
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@19093 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
In addition:
- Fix: PreParser used to report an unexpected token one token too late when
ParsePrimaryExpression failed.
- Unified identifier handling (PreParser::GetIdentifier is now like Parser::GetIdentifier).
- Fix: PreParser used to produce "unexpected_token YIELD" errors when Parser
produced "unexpected_token_identifier"; fixed PreParser to match Parser.
BUG=3126
LOG=N
R=ulan@chromium.org
Review URL: https://codereview.chromium.org/151103006
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@19082 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
This contains the following fixes:
- We had strict_reserved_word and unexpected_strict_reserved, which one to use
was totally mixed in Parser and PreParser. Removed strict_reserved_word.
- When we saw a strict future reserved word when expecting something completely
different (such as "(" in "function foo interface"), Parser reports unexpected
identifier, whereas PreParser used to report unexpected strict reserved
word. Fixed PreParser to report unexpected identifier too.
- Unified parser and preparser error locations when the name of a function is a
strict reserved word. Now both point to the name.
BUG=3126
LOG=N
R=ulan@chromium.org
Review URL: https://codereview.chromium.org/149253010
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@19067 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
BUG=v8:2943
LOG=Y
TEST=Unit tests for "real life" use cases, edge cases, various types of normalization.
==========================
This is identical to the previous CL
https://codereview.chromium.org/40133004/
with two differences:
* Added a dummy implementation of String.prototype.normalize to be used when v8 is compiled without intl support
* Rebased the the test files for webkit. That was the only reason for the previous failure (and revert).
Thank you,
Mihai
R=svenpanne@chromium.org
Review URL: https://codereview.chromium.org/68133016
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18972 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
The webkit/function-apply-aliased.js test fails on simulators (both MIPS
and ARM) as the printed output does not match to the expected. The
failing test forces a stack overflow exception and the ToString()
operation of the exception object fails because of an other stack
overflow and returns an empty string.
The problem is that on hardware a common JS and C stack is used so the
stack overflow can be caught in C functions also while on simulator
separated JS and C stacks are used.
This patch adds a "sim" condition to test .status files to skip tests
only on simulator.
LOG=N
BUG=v8:3124
R=jkummerow@chromium.org, plind44@gmail.com
Review URL: https://codereview.chromium.org/139233005
Patch from Balazs Kilvady <kilvadyb@homejinni.com>.
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18959 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
version is passing all the existing test + a bunch of new tests
(packaged in the change list, too).
The patch extends the SlotRef object to describe captured and duplicated
objects. Since the SlotRefs are not independent of each other anymore,
there is a new SlotRefValueBuilder class that stores the SlotRefs and
later materializes the objects from the SlotRefs.
Note that unlike the previous implementation of SlotRefs, we now build
the SlotRef entries for the entire frame, not just the particular
function. This is because duplicate objects might refer to previous
captured objects (that might live inside other inlined function's part
of the frame).
We also need to store the materialized objects between other potential
invocations of the same arguments object so that we materialize each
captured object at most once. The materialized objects of frames live
in the new MaterielizedObjectStore object (contained in Isolate),
indexed by the frame's FP address. Each argument materialization (and
deoptimization) tries to lookup its captured objects in the store before
building new ones. Deoptimization also removes the materialized objects
from the store. We also schedule a lazy deopt to be sure that we always
get rid of the materialized objects and that the optmized function
adopts the materialized objects (instead of happily computing with its
captured representations).
Concerns:
- Is the FP address the right key for a frame? (Note that deoptimizer's
representation of frame is different from the argument object
materializer's one - it is not easy to find common ground.)
- Performance is suboptimal in several places, but a quick local run of
benchmarks does not seem to show a perf hit. Examples of possible
improvements: smarter generation of SlotRefs (build other functions'
SlotRefs only for captured objects and only if necessary), smarter
lookup of stored materialized objects.
- Ideally, we would like to share the code for argument materialization
with deoptimizer's materializer. However, the supporting data structures
(mainly the frame descriptor) are quite different in each case, so it
looks more like a separate project.
Thanks for any feedback.
R=danno@chromium.org, mstarzinger@chromium.org
LOG=N
BUG=
Committed: https://code.google.com/p/v8/source/detail?r=18918
Review URL: https://codereview.chromium.org/103243005
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18936 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
mostly to make sure that it is going in the right direction. The current
version is passing all the existing test + a bunch of new tests
(packaged in the change list, too).
The patch extends the SlotRef object to describe captured and duplicated
objects. Since the SlotRefs are not independent of each other anymore,
there is a new SlotRefValueBuilder class that stores the SlotRefs and
later materializes the objects from the SlotRefs.
Note that unlike the previous implementation of SlotRefs, we now build
the SlotRef entries for the entire frame, not just the particular
function. This is because duplicate objects might refer to previous
captured objects (that might live inside other inlined function's part
of the frame).
We also need to store the materialized objects between other potential
invocations of the same arguments object so that we materialize each
captured object at most once. The materialized objects of frames live
in the new MaterielizedObjectStore object (contained in Isolate),
indexed by the frame's FP address. Each argument materialization (and
deoptimization) tries to lookup its captured objects in the store before
building new ones. Deoptimization also removes the materialized objects
from the store. We also schedule a lazy deopt to be sure that we always
get rid of the materialized objects and that the optmized function
adopts the materialized objects (instead of happily computing with its
captured representations).
Concerns:
- Is there a simpler/more correct way to store the already-materialized
objects? (At the moment there is a custom root reference to JSArray
containing frames' FixedArrays with their captured objects.)
- Is the FP address the right key for a frame? (Note that deoptimizer's
representation of frame is different from the argument object
materializer's one - it is not easy to find common ground.)
- Performance is suboptimal in several places, but a quick local run of
benchmarks does not seem to show a perf hit. Examples of possible
improvements: smarter generation of SlotRefs (build other functions'
SlotRefs only for captured objects and only if necessary), smarter
lookup of stored materialized objects.
- Ideally, we would like to share the code for argument materialization
with deoptimizer's materializer. However, the supporting data structures
(mainly the frame descriptor) are quite different in each case, so it
looks more like a separate project.
Thanks for any feedback.
R=mstarzinger@chromium.org, danno@chromium.org
LOG=N
BUG=
Review URL: https://codereview.chromium.org/103243005
git-svn-id: http://v8.googlecode.com/svn/branches/bleeding_edge@18918 ce2b1a6d-e550-0410-aec6-3dcde31c8c00
