Add missing kEnd opcode after each body. Also, avoid the macro.
This fix is needed for follow-up changes to the control transfer
computation.
R=ahaas@chromium.org
BUG=v8:5822
Change-Id: If2b4dbea831ec40939a2045701f3d13479331773
Reviewed-on: https://chromium-review.googlesource.com/485481
Reviewed-by: Andreas Haas <ahaas@chromium.org>
Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
Cr-Commit-Position: refs/heads/master@{#44794}
For now skip some WASM SIMD tests that fail when MIPS SIMD extension is
not available. Turn on these tests again when simd scalar lowering
mechanism supports all WASM SIMD operations.
BUG=
Review-Url: https://codereview.chromium.org/2829963003
Cr-Commit-Position: refs/heads/master@{#44791}
Consistently support calls to host-C-linkage functions with up to 9 arguments
from the simulator, and check that these limits aren't exceeded accidentally.
BUG=v8:6281
Review-Url: https://codereview.chromium.org/2825393003
Cr-Commit-Position: refs/heads/master@{#44790}
We already have an optimization to constant-fold access to an object's
prototype via the special __proto__ accessor (specified in appendix B).
We can use the same optimization to also constant-fold accesses to an
object's prototype via the official Object.getPrototypeOf function.
Also add the optimization for Reflect.getPrototypeOf, which is
equivalent for object inputs.
This is commonly used by Babel to implement various new language
features, for example subclassing and certain property lookups.
R=yangguo@chromium.org
BUG=v8:6292
Review-Url: https://codereview.chromium.org/2841463002
Cr-Commit-Position: refs/heads/master@{#44788}
.. replace all of them with prepared global object template.
+ bonus: wrap inspector related methods with inspector namespace.
BUG=none
R=dgozman@chromium.org
Review-Url: https://codereview.chromium.org/2832723004
Cr-Commit-Position: refs/heads/master@{#44785}
These can be synthesized from existing operations and scheduled for
better performance than if we have to generate blocks of instructions
that take many cycles to complete.
- Remove F32x4RecipRefine, F32x4RecipSqrtRefine. Clients are better off
synthesizing these from splats, multiplies and adds.
- Remove F32x4Div, F32x4Sqrt, F32x4MinNum, F32x4MaxNum. Clients are
better off synthesizing these or using the reciprocal approximations,
possibly with a refinement step.
LOG=N
BUG=v8:6020
Review-Url: https://codereview.chromium.org/2827143002
Cr-Commit-Position: refs/heads/master@{#44784}
Ensure source map is not stable if elements kind transitions are expected.
BUG=chromium:700733
Change-Id: Ie937e7064127250b1100109986c3e9b411fae1d6
Reviewed-on: https://chromium-review.googlesource.com/483442
Reviewed-by: Michael Starzinger <mstarzinger@chromium.org>
Reviewed-by: Jakob Kummerow <jkummerow@chromium.org>
Commit-Queue: Igor Sheludko <ishell@chromium.org>
Cr-Commit-Position: refs/heads/master@{#44780}
This also removes a broken optimization regarding immutable (i.e. const)
global variables. For now mutable and immutable global variables are
treated the same and hence copied during module initialization.
R=rossberg@chromium.org
TEST=mjsunit/asm/global-imports
BUG=v8:6279
Change-Id: I020fc12036dc534f5a62fb43f5c6fdb252314e62
Reviewed-on: https://chromium-review.googlesource.com/483360
Reviewed-by: Andreas Rossberg <rossberg@chromium.org>
Commit-Queue: Michael Starzinger <mstarzinger@chromium.org>
Cr-Commit-Position: refs/heads/master@{#44763}
The included test case illustrates the problem. It subtracts (16 << 27)
from another number. The Machine Operator Reducer would replace the
shift computation with 0x0000000080000000, and then change the subtract
to an add of -(0x0000000080000000), which is 0xffffffff80000000.
The instruction selector would determine that this value could be an
immediate, because it fits in 32 bits, so it would select the lea
instruction. Finally, the code generator would detect that the
immediate was less than 0, flip the sign and replace the add with a
subtract of 0x80000000. Because the x64 subtract instruction's
immediate field is 32 bits, the processor would interpret this as
0xffffffff80000000 instead of an unsigned value.
This change fixes the issue by making the CanBeImmediate check
explicitly compare against INT_MIN and INT_MAX. We disallow INT_MIN
as an immediate precisely because we cannot tell 0x0000000080000000
from 0xffffffff80000000 when truncated to 32 bits.
Bug: chromium:711203
Change-Id: Ie371b8ea290684a6bb723bae9c693a866f961850
Reviewed-on: https://chromium-review.googlesource.com/482448
Commit-Queue: Eric Holk <eholk@chromium.org>
Reviewed-by: Mircea Trofin <mtrofin@chromium.org>
Cr-Commit-Position: refs/heads/master@{#44758}
With recent CLs we always store maximum N async stack traces and when we reach limit we drop half of them.
Current promise collected event requires creating weak handle:
- it takes time,
- it consumes memory.
Since async task id distribution for promises is uniform (each new promise has last_async_task_id + 1 as an id) our hash map is good enough to handle any amount of async task ids, following time of executing 1 000 000 000 of lookups:
- for empty hash map: 1.45 seconds,
- for hash map with one entry: 14.95 seconds
- 1024 entries: 15.03 seconds
- 1024 * 1024 entries: 14.82 seconds
- 1024 * 1024 * 1024: 17.9 seconds
BUG=v8:6189
R=dgozman@chromium.org,yangguo@chromium.org
Review-Url: https://codereview.chromium.org/2819423005
Cr-Commit-Position: refs/heads/master@{#44750}
For MIPS64, many load/store operations from/to memory emit more then
one instruction. This is the reason for moving them from assembler to
macro-assembler.
TEST=
BUG=
Review-Url: https://codereview.chromium.org/2829073002
Cr-Commit-Position: refs/heads/master@{#44746}
So that we can delete object properties without a runtime call.
The builtin implements a few fast paths (for now only deletion
of dictionary properties), and calls the runtime for all other
cases.
Review-Url: https://codereview.chromium.org/2810363003
Cr-Commit-Position: refs/heads/master@{#44740}
- Adds unary Reverse shuffles (swizzles): S32x2Reverse, S16x4Reverse,
S16x2Reverse, S8x8Reverse, S8x4Reverse, S8x2Reverse. Reversals are
done within the sub-vectors that prefix the opcode name, e.g. S8x2
reverses the 8 consecutive pairs in an S8x16 vector.
- Adds binary Zip (interleave) left and right half-shuffles to return a
single vector: S32x4ZipLeft, S32x4ZipRightS16x8ZipLeft, S16x8ZipRight,
S8x16ZipLeft, S8x16ZipRight.
- Adds binary Unzip (de-interleave) left and right half shuffles to return
a single vector: S32x4UnzipLeft, S32x4UnzipRight, S16x8UnzipLeft,
S16x8UnzipRight, S8x16UnzipLeft, S8x16UnzipRight.
- Adds binary Transpose left and right half shuffles to return
a single vector: S32x4TransposeLeft, S32x4TransposeRight,
S16x8TransposeLeft, S16xTransposeRight, S8x16TransposeLeft,
S8x16TransposeRight.
- Adds binary Concat (concatenate) byte shuffle: S8x16Concat #bytes to
paste two vectors together.
LOG=N
BUG=v8:6020
Review-Url: https://codereview.chromium.org/2801183002
Cr-Commit-Position: refs/heads/master@{#44734}
There's no reason to keep track, for a preparsed function itself,
whether that function calls eval. All that matters is that the ancestor
scopes are marked as having an inner scope which calls eval. The function
will have its "calls eval" bit persisted if/when it's fully parsed.
The only "behavioral" change in this patch is the removal of a DCHECK.
Bug: v8:6092
Change-Id: I17e396c8a265030fe0ad941707e4a97972e6650b
Reviewed-on: https://chromium-review.googlesource.com/481223
Commit-Queue: Adam Klein <adamk@chromium.org>
Reviewed-by: Marja Hölttä <marja@chromium.org>
Cr-Commit-Position: refs/heads/master@{#44732}
let/const declarations in "standard" C-style for-loops have
some complex desugaring to accommodate the case where loop
loop variables may be captured. This slows down the baseline
performance of for-loops with let variables.
This change attempts to avoid this desugaring if it's known that
the loop variable is not captured at any point. A side effect of
this change is that let/const loop variables, when not captured
within the loop body, are not necessarily shown in the debugger,
similar to other stack-allocated vars.
BUG=v8:4762, v8:5460
R=marja@chromium.org, adamk@chromium.org, yangguo@chromium.org
Change-Id: I8dbe545a12c086f675972bdba60c94998268311a
Reviewed-on: https://chromium-review.googlesource.com/472247
Commit-Queue: Caitlin Potter <caitp@igalia.com>
Reviewed-by: Yang Guo <yangguo@chromium.org>
Reviewed-by: Adam Klein <adamk@chromium.org>
Reviewed-by: Marja Hölttä <marja@chromium.org>
Cr-Commit-Position: refs/heads/master@{#44731}
When asked for a module that previously failed to compile or
instantiate, the embedder necessarily has to signal failure. In this
case, we expect an exception to be scheduled, which we will rethrow.
BUG=v8:1569
Review-Url: https://codereview.chromium.org/2827733002
Cr-Commit-Position: refs/heads/master@{#44729}
Move JSOrdinaryHasInstance lowering to JSNativeContextSpecialization,
which was previously mostly done in JSTypedLowering (for no reason).
Add new logic to the lowering to constant-fold OrdinaryHasInstance
checks when the map of the left-hand side and the "prototype" of the
right-hand side is known. This address the performance issue with the
(base) class constructors generated by Babel, i.e.:
function _classCallCheck(instance, Constructor) {
if (!(instance instanceof Constructor)) {
throw new TypeError("Cannot call a class as a function");
}
}
var C = function C() { _classCallCheck(this, C); };
for
class C {}
Also ensure that a known constructor being used inside an instanceof
get's a proper initial map on-demand.
BUG=v8:6275
R=mstarzinger@chromium.org
Review-Url: https://codereview.chromium.org/2827013002
Cr-Commit-Position: refs/heads/master@{#44727}
When adding or overwriting properties of an object, the generic
keyed store stub must check if that property's name might have
an associated protector (e.g. the ArraySpeciesProtector) and
take the slow path if so to ensure that the protector is updated
as needed.
BUG=v8:6269
Review-Url: https://codereview.chromium.org/2821213004
Cr-Commit-Position: refs/heads/master@{#44726}
1) Function recording conditions need to be consistent (this same condition is used above)
2) byte is not wide enough for storing the backing store size.
Bugs uncovered by the existing tests with the flag on.
BUG=v8:5516
Change-Id: Iec6aff0cf1858afe1083526e4ada9a8eca08f062
Reviewed-on: https://chromium-review.googlesource.com/481320
Commit-Queue: Marja Hölttä <marja@chromium.org>
Reviewed-by: Daniel Vogelheim <vogelheim@chromium.org>
Cr-Commit-Position: refs/heads/master@{#44724}
This includes a fastpath in the ElementsAccessor for the source
array being a JSArray with FastSmi or FastDouble packed kinds. This
is probably a pretty common usage, where an array is passed in as
a way of initializing the TypedArray at creation (as there is not other
syntax to do this). e.g. new Float64Array([1.0, 1.0, 1.0]) for some
sort of vector application.
BUG= v8:5977
Change-Id: Ice4ad9fc29f56b1c4b0b30736a1330efdc289003
Reviewed-on: https://chromium-review.googlesource.com/465126
Reviewed-by: Camillo Bruni <cbruni@chromium.org>
Commit-Queue: Peter Marshall <petermarshall@chromium.org>
Cr-Commit-Position: refs/heads/master@{#44722}
In particular the Babel generated code for simple constructors is
of interest here, since that's what we hit in the wild most often
currently.
BUG=v8:6275
R=yangguo@chromium.org
Review-Url: https://codereview.chromium.org/2822203004
Cr-Commit-Position: refs/heads/master@{#44719}
This new API sets a native data property on an object
directly, as Template::SetNativeDataProperty does.
It is similar to Object::SetAccessor, but properties
set by SetNativeDataProperty without kReadOnly flag
can be replaced.
Bug:chromium:617892
Change-Id: I32973f7190906d76be6802da9a0489edce0bd93e
Reviewed-on: https://chromium-review.googlesource.com/479474
Reviewed-by: Jochen Eisinger <jochen@chromium.org>
Commit-Queue: Hitoshi Yoshida <peria@chromium.org>
Cr-Commit-Position: refs/heads/master@{#44716}
This fixes source position tracking within the "increment" expression of
a for-statement. The old {StashCode} method was inherently incompatible
with side tables like the source position table, as it would leave them
untouched while mutating the bytecode stream. It was hence trimmed down
to {DeleteCode}.
R=bradnelson@chromium.org
BUG=v8:6127
Change-Id: I7a5ff60cd5334208c44b165c8b54144d9ae83209
Reviewed-on: https://chromium-review.googlesource.com/480301
Commit-Queue: Michael Starzinger <mstarzinger@chromium.org>
Reviewed-by: Brad Nelson <bradnelson@chromium.org>
Cr-Commit-Position: refs/heads/master@{#44714}
The typing rule for JSCreateArguments must properly declare rest
parameters as arrays and only consider sloppy and strict arguments
objects as Type::OtherObject.
TBR=jarin@chromium.org
BUG=v8:6262,chromium:712802
Review-Url: https://codereview.chromium.org/2828573004
Cr-Commit-Position: refs/heads/master@{#44712}
Functions that take mandatory parameters plus a number of optional
parameters, that need to be materialized as an Array are quite common.
The simplest possible case of this is essentially:
function foo(mandatory, ...args) { return args; }
Babel translates this to something like:
function foo(mandatory) {
"use strict";
for (var _len = arguments.length,
args = Array(_len > 1 ? _len - 1 : 0),
_key = 1; _key < _len; _key++) {
args[_key - 1] = arguments[_key];
}
return args;
}
The key to great performance here is to make sure that we don't
materialize the (unmapped) arguments object in this case, plus that we
have some kind of fast-path for the Array constructor and the
initialization loop.
This microbenchmark ensures that we have decent performance even in the
case where the assignment to args is polymorphic, i.e. the arguments
have seen different elements kinds, starting with FAST_HOLEY_ELEMENTS
and then FAST_HOLEY_SMI_ELEMENTS.
R=yangguo@chromium.org
BUG=v8:6262
Review-Url: https://codereview.chromium.org/2823343004
Cr-Commit-Position: refs/heads/master@{#44709}
It succeeds on my machine, at least, so marking pass/fail to avoid
unnecessary noise until the attached bug is resolved.
TBR=machenbach@chromium.org
Bug: v8:6252
Change-Id: I146eee0c464896e42058745096e929d221c04347
Reviewed-on: https://chromium-review.googlesource.com/480988
Reviewed-by: Adam Klein <adamk@chromium.org>
Cr-Commit-Position: refs/heads/master@{#44704}
When --harmony-strict-legacy-accessor-builtins is enabled, it brings
V8's behavior in line with the spec and more recent versions of
SpiderMonkey and JSC:
- No implicit receiver coercion
- Attempting to redefine a non-configurable property throws
Bug: v8:5070
Change-Id: I82b927538604136c0c550e19bcc606fbfea1377e
Reviewed-on: https://chromium-review.googlesource.com/478312
Reviewed-by: Daniel Ehrenberg <littledan@chromium.org>
Commit-Queue: Adam Klein <adamk@chromium.org>
Cr-Commit-Position: refs/heads/master@{#44703}
What will we get:
- console would be included into snapshot and allow us to reduce time that we spent in contextCreated function (~5 times faster),
- it allows us to make further small improvement of console methods, e.g. we can implement super quick return from console.assert if first argument is true,
- console calls are ~ 15% faster.
CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_trusty_blink_rel;master.tryserver.chromium.linux:linux_chromium_rel_ng
BUG=v8:6175
R=dgozman@chromium.org
Review-Url: https://codereview.chromium.org/2785293002
Cr-Original-Original-Original-Commit-Position: refs/heads/master@{#44353}
Committed: 55905f85d6
Review-Url: https://codereview.chromium.org/2785293002
Cr-Original-Original-Commit-Position: refs/heads/master@{#44355}
Committed: cc74ea0bc4
Review-Url: https://codereview.chromium.org/2785293002
Cr-Original-Commit-Position: refs/heads/master@{#44416}
Committed: f5dc738cda
Review-Url: https://codereview.chromium.org/2785293002
Cr-Commit-Position: refs/heads/master@{#44702}
+ additional fixes uncovered by bug, and addressed remaining feedback
from original CL (https://codereview.chromium.org/2806073002/).
Note that the regression test differs slightly from the bug reported one,
in that it catches the RangeError which will eventually be thrown due
to call stack size being exceeded.
BUG=chromium:712569
Review-Url: https://codereview.chromium.org/2825073002
Cr-Commit-Position: refs/heads/master@{#44700}
This switches the parser to use token positions (i.e. {Position})
instead of stream positions (i.e. {GetPosition}) everywhere. Access to
the latter is being removed as it is unsupported when the scanner is in
rewind state anyways. This prevents "skipping" a token when seeking.
R=bradnelson@chromium.org
BUG=v8:6127
Change-Id: I9c13dd20a981061a2bccc4fb57e5c57d2a64ac5c
Reviewed-on: https://chromium-review.googlesource.com/480300
Reviewed-by: Brad Nelson <bradnelson@chromium.org>
Commit-Queue: Brad Nelson <bradnelson@chromium.org>
Cr-Commit-Position: refs/heads/master@{#44699}
- separated V8StackTraceImpl and AsyncStackTrace,
- V8Debugger owns all AsyncStackTrace and cleanup half of them when limit is reached (first created - first cleaned),
- V8StackTraceImpl, AsyncStackTrace and async-task-related tables in V8Debugger have weak reference to other async stack traces.
- async tasks are cleared with related async stacks.
BUG=v8:6189
R=dgozman@chromium.org
Review-Url: https://codereview.chromium.org/2816043006
Cr-Original-Commit-Position: refs/heads/master@{#44670}
Committed: 1bca73bc83
Review-Url: https://codereview.chromium.org/2816043006
Cr-Commit-Position: refs/heads/master@{#44694}
The feature is not quite ready for getting ClusterFuzzed.
BUG=v8:5516
Change-Id: I90a42f950727c8ecf46cb2987c9a459b2ba1f5a7
Reviewed-on: https://chromium-review.googlesource.com/480400
Commit-Queue: Marja Hölttä <marja@chromium.org>
Reviewed-by: Daniel Vogelheim <vogelheim@chromium.org>
Cr-Commit-Position: refs/heads/master@{#44693}
