- Now "inner_scope_uses_eval_" is also set of scopes that call eval themselves.
- AllowLazyParsing doesn't check force_eager_compilation_ anymore.
- Both inner_scope_uses_eval_ and force_eager_compilation_ are propagated
outwards immediately when set.
BUG=v8:5209
Review-Url: https://codereview.chromium.org/2269603002
Cr-Commit-Position: refs/heads/master@{#38797}
(40B -> 32B, due to alignment rules. Should be neutral on 32b platforms.)
BUG=v8:4947
Review-Url: https://codereview.chromium.org/2256183005
Cr-Commit-Position: refs/heads/master@{#38794}
Now that -0.0 and 0.0 are ordered by Float64Max and Float64Min, these
two operator are both associative and commutative for all numbers. The
operators would not be associative and commutative for NaNs, but
neither JavaScript nor WebAssembly distinguish between NaNs nowadays.
R=jarin@chromium.org
Review-Url: https://codereview.chromium.org/2263163002
Cr-Commit-Position: refs/heads/master@{#38791}
CallPrinter currently has only a single use, namely to print the call
site when throwing CalledNonCallable (ConstructedNonConstructable)
errors. This ensures that unicode call sites such as "あいう"() are
printed correctly by using an IncrementalStringBuilder internally.
BUG=637167
Review-Url: https://codereview.chromium.org/2265073002
Cr-Commit-Position: refs/heads/master@{#38790}
Disable inlining of Cpp to a direct CEntryStub call when a call would
require argument adaption, i.e. when argument adaption is enabled for
the given function and the actual argument count differs from the formal
parameter count.
This is intended to be a temporary fix until we either disable argument
adaption for all Cpp builtins or add adaption logic to inlined Cpp
builtins.
BUG=chromium:639752
Review-Url: https://codereview.chromium.org/2266893002
Cr-Commit-Position: refs/heads/master@{#38788}
Fix getter/setter typo so we use the correct object in DCHECK.
BUG=
Review-Url: https://codereview.chromium.org/2255713002
Cr-Commit-Position: refs/heads/master@{#38787}
In theory the flag could otherwise be set on a blockscope that might be finalized (removed) before PropagateScopeInfo can propagate the flag around. In that case we'd lose the flag.
BUG=v8:5209
Review-Url: https://codereview.chromium.org/2267683002
Cr-Commit-Position: refs/heads/master@{#38781}
This is the only user of outer_scope_calls_sloppy_eval. Inlining it simplifies scope analysis.
BUG=v8:5209
Review-Url: https://codereview.chromium.org/2263123002
Cr-Commit-Position: refs/heads/master@{#38779}
Adds InterpreterCompilationJob as a sub-class of
CompilationJob, to enable off-thread bytecode
generation. Currently only used in
Interpreter::MakeBytecode.
As part of this change, CompilationJob is modified
to make it less specific to optimized compilation,
renaming the phases as follows:
- CreateGraph -> PrepareJob
- OptimizeGraph -> ExecuteJob
- GenerateCode -> FinalizeJob
RegisterWeakObjectsInOptimizedCode is also moved out
of CompilationJob and instead becomes a static function
on Compiler.
BUG=v8:5203
Committed: https://crrev.com/1fb6a7e697e8bc5b4af51647553741f966e00cdc
Committed: https://crrev.com/785990e9fc0dd9a9d963d25d0bed2909165e4ca9
Committed: https://crrev.com/d7c6195c4c5cdc080caa74dfe2ae9ecab69bea73
Review-Url: https://codereview.chromium.org/2240463002
Cr-Original-Original-Original-Commit-Position: refs/heads/master@{#38662}
Cr-Original-Original-Commit-Position: refs/heads/master@{#38668}
Cr-Original-Commit-Position: refs/heads/master@{#38725}
Cr-Commit-Position: refs/heads/master@{#38778}
This makes us able to get rid of dependencies to parser.h from places
which only need the ParseInfo, and also gets rid of the curious Parser
<-> Compiler circular dependency.
Also IWYUd where necessary.
BUG=
Review-Url: https://codereview.chromium.org/2268513002
Cr-Commit-Position: refs/heads/master@{#38777}
This fixes the uncommon fallback when Compiler::CompileOptimized fails
creating optimized code and also the underlying SharedFunctionInfo is
not compiled. We make sure not to uselessly regenerate the bytecode.
R=rmcilroy@chromium.org
BUG=chromium:639753
Review-Url: https://codereview.chromium.org/2261123002
Cr-Commit-Position: refs/heads/master@{#38776}
AstGraphBuilder overrides Visit(Expression*) to ensure that even if there is a
stack overflow, a value still gets produced. However, if there was no stack
overflow in the overriden function, but calling
AstVisitor<AstGraphBuilder>::Visit(expr) pushes us over the stack limit, then
the stack overflow check in that function will return without visiting the
expression, and the result will never get pushed.
To fix this, we add a new VisitNoStackOverflowCheck function which avoids the
inner stack check, and call that instead.
Since this depends on the size of C++ stack frames, there is no reliable test
I can add, however regress-635429.js exibits this behavior after
https://codereview.chromium.org/2240463002/ lands.
Review-Url: https://codereview.chromium.org/2262703002
Cr-Commit-Position: refs/heads/master@{#38774}
This detects loops with integer decrements.
Drive-by fix: Add lower bounds to lower_bounds
zone vector instead of upper_bounds.
BUG=
Review-Url: https://codereview.chromium.org/2260153002
Cr-Commit-Position: refs/heads/master@{#38772}
Now it just relies on NeedsContext and hence is guaranteed to be in sync.
BUG=v8:5209
Review-Url: https://codereview.chromium.org/2262133002
Cr-Commit-Position: refs/heads/master@{#38771}
Rebuilding (after touching certain files) is crazy slow because
includes are out of control.
Fixing it:
- Don't include stuff in headers unless necessary.
- Include the stuff you need, not some other stuff that happens to include the
stuff you need.
BUG=v8:5294
Review-Url: https://codereview.chromium.org/2258603003
Cr-Commit-Position: refs/heads/master@{#38770}
This adds cached versions of the DeoptimizeIf, DeoptimizeUnless and
Deoptimize operators that occur most often in various benchmarks.
R=jarin@chromium.org
Review-Url: https://codereview.chromium.org/2266743002
Cr-Commit-Position: refs/heads/master@{#38769}
parser and the preparser, so that they contain the same set of methods,
with the same signatures. It mainly flags some traits methods as const.
It also contains a small cosmetic change in the definition of CHECK_OK.
R=adamk@chromium.org, marja@chromium.org
BUG=
LOG=N
Review-Url: https://codereview.chromium.org/2258123002
Cr-Commit-Position: refs/heads/master@{#38767}
This patch simply reorders the elements of classes ParserTraits
and PreParserTraits (mainly method headers), so that in both
classes the elements are defined in the same order. This is
useful for easily verifying the consistency between these two
classes.
R=adamk@chromium.org, marja@chromium.org
BUG=
LOG=N
Review-Url: https://codereview.chromium.org/2264483003
Cr-Commit-Position: refs/heads/master@{#38766}
This patch subsumes CoverInitializedNameProduction to create an ObjectLiteralProduction which is now used to report the duplicate proto error as well.
This patch also changes ObjectLiteralChecker::CheckProperty
to record an ObjectLiteralProduction error instead of
bailing out immediately. Once we realize that we're in a
pattern, we rewind the error, otherwise we report the
error.
BUG=v8:5121
Review-Url: https://codereview.chromium.org/2255353002
Cr-Commit-Position: refs/heads/master@{#38764}
Also removed the "StandardProductions" enum value as its name is unhelpful
in deciding when to use it (and it only had a single reference).
R=gsathya@chromium.org
Review-Url: https://codereview.chromium.org/2260183003
Cr-Commit-Position: refs/heads/master@{#38763}
Currently when redefining eval or arguments in non-simple parameter list and
destructuring binding, V8 doesn't throw any error, this patch fixes it.
BUG=v8:5201
LOG=N
Review-Url: https://codereview.chromium.org/2185223002
Cr-Commit-Position: refs/heads/master@{#38762}
This is only needed to set a flag on the SharedFunctionInfo generated for a function literal. Hence we only need it in 1 scope, and only read it once. Recursing in that case will be more efficient than always tracking it.
BUG=v8:5209
Review-Url: https://codereview.chromium.org/2261693002
Cr-Commit-Position: refs/heads/master@{#38761}
Record which asm.js stdlib members are used and add a check that NaN is actually correctly set. Other stdlib members to be added in a later change.
Also add a stdlib argument to Wasm.instantiateModuleFromAsm, in preparation for that function to be replaced by normal asm.js instantiation.
BUG= https://bugs.chromium.org/p/v8/issues/detail?id=4203
LOG=N
R=jpp@chromium.org,titzer@chromium.org
Review-Url: https://codereview.chromium.org/2251433002
Cr-Commit-Position: refs/heads/master@{#38760}
While they have not been observed to slow down real-world use cases,
some blink_layout microbenchmarks feel better with these methods
disabled. In order to be concervative at the launch time, lift
the CFI defense for these methods.
8/10 of these methods will become much faster when an optimization
proposed in https://crbug.com/638056 -- we only need to load vptr
once (before the loop) and have a single CFI check instead of
a check per iteration.
BUG=638056,634139
Review-Url: https://codereview.chromium.org/2258003002
Cr-Commit-Position: refs/heads/master@{#38759}
This isn't the most elegant fix, but I'd prefer to not rework the logic
right now. What happens is:
- Most parts of the Scanner use nullptr to mean, no literal buffer.
- The bookmarking logic may end up with a state where there's a non-nullptr
literal buffer, but it's empty. (length 0)
- These are functionally equivalent, so there's no 'real' bug.
- But it makes it hard to reason. This patch hence checks for length-0
literal buffers, and uses nullptr instead.
R=marja@chromium.org
BUG=chromium:639191 v8:4947
Review-Url: https://codereview.chromium.org/2258073003
Cr-Commit-Position: refs/heads/master@{#38756}
Version 0 dense arrays cannot be deserialized by current Chromium, which
suggests that this is not necessary.
BUG=chromium:148757
Review-Url: https://codereview.chromium.org/2256413002
Cr-Commit-Position: refs/heads/master@{#38754}
Implement Neg_d and Neg_s in macro-assembler.
Floating point negate instructions are removed in release 6.
On r2, these instructoin do not change the sign of NaN operands.
TEST=cctest/test-run-wasm/RunWasmCompiled_Float32Neg, cctest/test-run-wasm/RunWasmCompiled_Float64Neg
BUG=
Review-Url: https://codereview.chromium.org/2256963003
Cr-Commit-Position: refs/heads/master@{#38749}
This fixes the self-healing mechanism for closures in the interpreter
entry trampoline not that bytecode can be preserved even when baseline
code is already available.
R=rmcilroy@chromium.org
TEST=cctest/test-compiler/IgnitionEntryTrampolineSelfHealing
BUG=chromium:638225
Review-Url: https://codereview.chromium.org/2257143002
Cr-Commit-Position: refs/heads/master@{#38747}
Unify the representation selection rules for NumberAdd/Subtract and
SpeculativeNumberAdd/Subtract wrt. Int32Add/Sub selection. We can
safely use Int32Add/Sub as long as the inputs are in the safe additive
integer range and the output is either truncated to Word32 or provably
in Signed32 or Unsigned32 range.
R=jarin@chromium.org
Review-Url: https://codereview.chromium.org/2253293005
Cr-Commit-Position: refs/heads/master@{#38746}
Changes the control flow builder classes to make use of the
BytecodeLabels helper class.
BUG=v8:4280
LOG=n
Review-Url: https://codereview.chromium.org/2254493002
Cr-Commit-Position: refs/heads/master@{#38744}
