Always execute the implicit return if we fell off the function bytes.
This is not considered an additional "step" as it is not executing a
wasm instruction.
Otherwise, we might pause at an invalid position (one after the
function bytes).
R=titzer@chromium.org
BUG=v8:5822
Review-Url: https://codereview.chromium.org/2650293003
Cr-Commit-Position: refs/heads/master@{#42730}
The v8 waterfall currently doesn't run MIPS tests in the debug configuration,
so although there are tests that would have found them, they currently are
not running in the standard CI setup. A bug has been fixed to add the debug
configuration of MIPS & MIPS64, too.
Review-Url: https://codereview.chromium.org/2654263002
Cr-Commit-Position: refs/heads/master@{#42727}
Previously, when restarting a frame, we would rewrite all frames
between the debugger activation and the frame to restart to squash
them, and replace the return address with that of a builtin to
leave that rewritten frame, and restart the function by calling it.
We now simply remember the frame to drop to, and upon returning
from the debugger, we check whether to drop the frame, load the
new FP, and restart the function.
R=jgruber@chromium.org, mstarzinger@chromium.org
BUG=v8:5587
Review-Url: https://codereview.chromium.org/2636913002
Cr-Commit-Position: refs/heads/master@{#42725}
As Blink needs to set Symbol.toPrimitive, exposes the symbol in C++ APIs
as Symbol::GetToPrimitive.
BUG=chromium:680409
Review-Url: https://codereview.chromium.org/2657933003
Cr-Commit-Position: refs/heads/master@{#42724}
wasm::ErrorThrower doesn't actually throw exceptions, it just schedules them.
As a result, this exception isn't handled properly by code which expects
ValueDeserializer to actually throw. For instance, the unit tests use a
TryCatch to catch and handle expected exceptions in unit tests.
Before this patch, I see local unit test failures because a wasm decode test
schedules one, but it isn't caught (and instead causes Context::New to fail
at the beginning of the next test).
BUG=685713
Review-Url: https://codereview.chromium.org/2659483004
Cr-Commit-Position: refs/heads/master@{#42718}
Memory.Grow should detach the ArrayBuffer associated with the Mem object after Grow. Currently, when guard pages are enabled protection is changed to make more of the buffer accessible. This does not work for when the buffer should be detached after grow, because the memory object has a reference to the same buffer befor/after grow.
R=titzer@chromium.org, eholk@chromium.org
Review-Url: https://codereview.chromium.org/2653183003
Cr-Commit-Position: refs/heads/master@{#42717}
This is a flag useful for testing/perf investigations for asm-wasm.
We can dump the internal representation of a asm.js module (when asm-wasm is
enabled with --validate-asm) using --dump-wasm-module. We can't pick that
module afterwards because it contains non-compliant opcodes - used as
implementation detail of our asm-wasm pipeline.
This flag forces asm.js origin on the decoder.
BUG=
Review-Url: https://codereview.chromium.org/2656103003
Cr-Commit-Position: refs/heads/master@{#42716}
Chrome now uses the default_args system for specifying its own definitions of
these variables so the separate file in build_overrides is no longer necessary.
BUG=684096
Review-Url: https://codereview.chromium.org/2652263003
Cr-Commit-Position: refs/heads/master@{#42712}
There is a print in AstNumbering which needs to dereference the string
containing the function name, which clashes with the disallowed handle
reference scope used to allow ast-numbering to run off-thread.
This could be fixed by pushing the print out of this function, and
somehow propagating disable_crankshaft_reason out of the visitor, but in
reality this check will soon be removed anyway, and until it is this
function will be on the main thread, so we may as well just hack it.
Review-Url: https://codereview.chromium.org/2653953006
Cr-Commit-Position: refs/heads/master@{#42711}
This corrects the case when we need to allocate a
blocked register, but the blockage happens after a
use as an instruction input, and there's no place to
split before that.
BUG=v8:5888
Review-Url: https://codereview.chromium.org/2652153005
Cr-Original-Commit-Position: refs/heads/master@{#42706}
Committed: ca779b29a6
Review-Url: https://codereview.chromium.org/2652153005
Cr-Commit-Position: refs/heads/master@{#42710}
This CL adds --crankshaft and --no-always-opt flags to the tests that use
assertOptimized() and assertUnoptimized() respectively.
This CL also adds presubmit checks that ensure that tests have the proper
flags set.
BUG=v8:5890
Review-Url: https://codereview.chromium.org/2653753007
Cr-Commit-Position: refs/heads/master@{#42709}
Port f9367847b0
Port bf782ec512
Original Commit Message:
We can share almost all of the architecture-specific builtin code with super-call-with-spread.
Info to port-writers: The code in CheckSpreadAndPushToStack has changed slightly
from what was in Generate_ConstructWithSpread, in that we take the length of the
spreaded parameters from the JSArray rather than the FixedArray backing store.
R=petermarshall@chromium.org, joransiu@ca.ibm.com, jyan@ca.ibm.com, michael_dawson@ca.ibm.com
BUG=v8:5511
LOG=N
Review-Url: https://codereview.chromium.org/2655043004
Cr-Commit-Position: refs/heads/master@{#42708}
Reason for revert:
Introduces new crashers, e.g.
load("test/mjsunit/wasm/wasm-constants.js");
load("test/mjsunit/wasm/wasm-module-builder.js");
(function() {
var builder = new WasmModuleBuilder();
builder.addMemory(16, 32, false);
builder.addFunction("test", kSig_i_iii)
.addBodyWithEnd([
// body:
kExprI64Const, 0x42,
kExprI64Const, 0x7a,
kExprI64Ctz,
kExprI64Mul,
kExprI64Ctz,
kExprI64Const, 0x41,
kExprI64Ctz,
kExprI64Ctz,
kExprI64Shl,
kExprI64Const, 0x41,
kExprI64Ctz,
kExprI64Ctz,
kExprI64Shl,
kExprF32SConvertI64,
kExprUnreachable,
kExprEnd, // @20
])
.exportFunc();
var module = builder.instantiate();
module.exports.test(1, 2, 3);
})();
Original issue's description:
> [turbofan] Correct regalloc blocked register behavior
>
>
> This corrects the case when we need to allocate a
> blocked register, but the blockage happens after a
> use as an instruction input, and there's no place to
> split before that.
>
> BUG=v8:5888
>
> Review-Url: https://codereview.chromium.org/2652153005
> Cr-Commit-Position: refs/heads/master@{#42706}
> Committed: ca779b29a6TBR=bmeurer@chromium.org,jarin@chromium.org,mtrofin@chromium.org
# Skipping CQ checks because original CL landed less than 1 days ago.
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=v8:5888
Review-Url: https://codereview.chromium.org/2654993007
Cr-Commit-Position: refs/heads/master@{#42707}
This corrects the case when we need to allocate a
blocked register, but the blockage happens after a
use as an instruction input, and there's no place to
split before that.
BUG=v8:5888
Review-Url: https://codereview.chromium.org/2652153005
Cr-Commit-Position: refs/heads/master@{#42706}
Port d287c81969
Original Commit Message:
[RELAND with one change: until literal arrays are rooted in the outer
feedback vector (coming in the next days), the runtime-scope.cc change
is held off.]
When a function is declared in global scope, the closure is created
by the DeclareGlobals runtime service. It needs a pointer to the
literals array, already allocated in the feedback vector. This
fixes a bug where it's behavior wasn't in sync with CreateClosure,
which accepts the literals from the vector.
This enables a follow-on performance improvement in the CompileLazy
builtin.
R=mvstanton@chromium.org, joransiu@ca.ibm.com, jyan@ca.ibm.com, michael_dawson@ca.ibm.com
BUG=680637
LOG=N
Review-Url: https://codereview.chromium.org/2658053002
Cr-Commit-Position: refs/heads/master@{#42705}
The mentioned asserts did not work properly with interpreted and turbofanned functions.
To fix this issue %GetOptimizationStatus() now returns a set of flags instead of a single value.
This CL also adds more helper functions to mjsunit, like isNeverOptimize(), isAlwaysOptimize(),
isOptimized(fun), etc.
BUG=v8:5890
Review-Url: https://codereview.chromium.org/2654733004
Cr-Commit-Position: refs/heads/master@{#42703}
When testing turning --verify-csa off to generate better code for builtins, mips
started failing mksnapshot due to an assert in instruction-selection-mips.cc's
VisitBinop routine, which creates a buffer to hold InstructionOperand of size 4
that would be overflowed. This fix is somewhat speculative, assuming that either
the LHS or the RHS operand of a mips binary op can be an immediate (as opposed
to the current code which seems to have a code path where both the left and
right operands are added, leading to the buffer overflow). With this fix, the
assert doesn't fire and all of the mips tests run through successfully in debug
mode.
R=ishell@chromium.orgTBR=dusan.m.milosavljevic@gmail.com
Review-Url: https://codereview.chromium.org/2647283009
Cr-Commit-Position: refs/heads/master@{#42701}
The instance type of an object cannot change, only the concrete map
might. So when searching for an instance type witness, we don't need
to pay attention to potentially side-effecting nodes.
R=mstarzinger@chromium.org
Review-Url: https://codereview.chromium.org/2652893011
Cr-Commit-Position: refs/heads/master@{#42699}
This makes sure that static guarantees about object maps are not used
accross operations on the effect chain that might trigger a map change.
Such operations are missing the {Operator::kNoWrite} property.
R=bmeurer@chromium.org
TEST=mjsunit/regress/regress-crbug-685506
BUG=chromium:685506
Review-Url: https://codereview.chromium.org/2653273004
Cr-Commit-Position: refs/heads/master@{#42697}
According to the WebAssembly spec no arithmetic operation should ever
return a signalling NaN. With the constant folding in V8, however, it
was possible that some arithmetic operations were elided, and if the
input of the arithmetic operation was a signalling NaN, then also the
result was the same signalling NaN. This CL removes some constant
folding optimizations and adjusts others so that even with constant
folding the result of an arithmetic operation is never a signalling NaN.
R=titzer@chromium.org, rossberg@chromium.org, bmeurer@chromium.org
Review-Url: https://codereview.chromium.org/2647353007
Cr-Commit-Position: refs/heads/master@{#42694}
We compare ES5, ES6 and a Babel transpilation of the ES6 test.
BUG=v8:5894
Review-Url: https://codereview.chromium.org/2655063003
Cr-Commit-Position: refs/heads/master@{#42689}
The reference map was only recorded when a frame was entered for the
runtime call, but it is also needed when the frame already exists.
R=titzer@chromium.org
Review-Url: https://codereview.chromium.org/2655243002
Cr-Commit-Position: refs/heads/master@{#42687}
- Declaring a variable called "this" for preparsed functions was unnecessary;
DeclarationScope ctor already adds the variable.
- "arguments" for preparsed scopes need to be declared after parsing the
function, like it's done in the parser.
- Now arguments_ can be the dummy variable, so adapted code to it.
- A previous refactoring CL ( https://codereview.chromium.org/2638333002 ) was
incomplete; it had added ParserBase::ParseFunctionBody but
PreParser::ParseFunction didn't call it. This CL completes that work. This is
needed for getting "arguments" declared properly for preparsed functions.
- AllocateVariablesRecursively is already called for preparsed scopes (without
this CL, that is), and it bails out early. However, before the bailout it used
to dcheck num_stack_slots_ == 0; that is no longer true since we've done scope
analysis for preparsed scopes.
- Test fix: we cannot have any lazy inner functions in the test, except the
topmost lazy inner function. Such functions would also be lazy in the parser
case, and the parser would just throw away their variables. Then the test
tries to verify the preparsed data against the scopes without variables and fails.
- Disabled a test w/ a sloppy block function, will get that working again in the
upcoming CLs.
BUG=v8:5516
Review-Url: https://codereview.chromium.org/2655623005
Cr-Commit-Position: refs/heads/master@{#42685}
The InterpreterTester class cobbles together a JSFunction from
a manually created bytecode and feedback vector. However it's
fragile against design changes in the way literal arrays and
feedback vectors are handled. It's better to let it hand in
a feedback vector metadata object, and allow the system to
create the vector from that.
BUG=v8:5456
Review-Url: https://codereview.chromium.org/2652893010
Cr-Commit-Position: refs/heads/master@{#42684}
With creation frame we can show additional information with description of each async stack trace, which could help user to understand where promises were chained.
At least in case of Promise.resolve().then(foo1).then(foo2) we would be able to show following stack trace for break in foo2 callback:
foo2 (test.js:14:2)
-- Promise.resolve (test.js:29:14)--
-- Promise.resolve (test.js:28:14)--
promiseThen (test.js:30:2)
More details: https://docs.google.com/document/d/1u19N45f1gSF7M39mGsycJEK3IPyJgIXCBnWyiPeuJFE
BUG=v8:5738
R=dgozman@chromium.org,gsathya@chromium.org
Review-Url: https://codereview.chromium.org/2648873002
Cr-Commit-Position: refs/heads/master@{#42682}
