The WebAssembly spec is not fully deterministic: the sign bit of NaN
can be arbitrary. This sign bit can be observed by several WebAssembly
opcodes. In the testcase the sign bit of NaN makes the difference
between terminating code and an infinite loop.
In the libfuzzer fuzzer we have to prevent infinite loops ourselves.
At the moment we do this by only execute generated code of WebAssembly
modules for which the interpretation of the code ends in a limited
number of steps. With the non-determinism described above we cannot
guarantee the absence of infinite loops with this method. Therefore
we stop now to execute generated code of WebAssembly modules for which
we observe possible non-determinism in the interpreter.
R=clemensh@chromium.org
Bug: chromium:863829
Change-Id: I461d67df87d672bed25d6c915ba7ea5134cb5890
Reviewed-on: https://chromium-review.googlesource.com/1141945
Reviewed-by: Clemens Hammacher <clemensh@chromium.org>
Commit-Queue: Andreas Haas <ahaas@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54541}
This CL fixes a bug where a fast-path was used on non-extensible
objects.
R=jgruber@chromium.org
Bug: chromium:865264,chromium:865285
Change-Id: Ie14c95b383a65576799c71576a5c0f9f8e1c29ca
Reviewed-on: https://chromium-review.googlesource.com/1142766
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Commit-Queue: Simon Zünd <szuend@google.com>
Cr-Commit-Position: refs/heads/master@{#54539}
This is a reland of 7f67cbd4d1
Original change's description:
> [turbofan] More brokerization in JSCreateLowering.
>
> Brokerized ReduceJSCreateEmptyLiteralObject and added the scope
> for ReduceJSCreateLiteralArrayOrObject.
>
> Bug: v8:7790
> Change-Id: Ife34a6b610678a3fe24152151cf343400ee515bd
> Reviewed-on: https://chromium-review.googlesource.com/1140306
> Commit-Queue: Jaroslav Sevcik <jarin@chromium.org>
> Reviewed-by: Georg Neis <neis@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#54507}
Bug: v8:7790
Change-Id: Ia79ff9ef49c727155a9c476268234e56d2cc9fcb
Reviewed-on: https://chromium-review.googlesource.com/1142984
Commit-Queue: Georg Neis <neis@chromium.org>
Reviewed-by: Georg Neis <neis@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54538}
The previous change to futex.js will spawn too many workers in d8 in
some test configurations, which will throw an error.
This CL works around that by spawning fewer workers for the
Atomics.notify tests.
TBR=adamk@chromium.org
Bug: v8:7338
Change-Id: I0f3583781e5352b4d6672d43a087dc56d920122d
Reviewed-on: https://chromium-review.googlesource.com/1142895
Reviewed-by: Ben Smith <binji@chromium.org>
Commit-Queue: Ben Smith <binji@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54536}
Turbofan support for property loads from module namespace objects
has been tested by the test/js-perf-tests/Modules/basic-namespace
benchmark, but so far not by the mjsunit suite. This CL adds such
a test.
This is a follow-up to 8d7379c066.
Change-Id: I3c4183d761693199e6bc8740b812279efcd791a0
Reviewed-on: https://chromium-review.googlesource.com/1142594
Reviewed-by: Adam Klein <adamk@chromium.org>
Commit-Queue: Jakob Kummerow <jkummerow@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54535}
At the May 2018 TC39 meeting, they decided to rename Atomics.wake to
Atomics.notify. This change adds Atomics.notify as an alias, but does
not remove Atomics.wake, which will be removed later.
This allows for embedders to use either name to prevent
breaking tests. When the tests are switched over, we can remove
Atomics.wake.
Bug: v8:7883
Change-Id: If057ebff162bde975c6e1b60d83a4662f144e81f
Reviewed-on: https://chromium-review.googlesource.com/1142290
Commit-Queue: Ben Smith <binji@chromium.org>
Reviewed-by: Adam Klein <adamk@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54534}
- Modifies LinkageAllocator to understand aliasing on ARM.
- Adds ability to allocate SIMD registers too. Before, these would
default to stack allocated.
- Modifies WasmCompiler to be platform-independent.
Bug: v8:7754
Change-Id: I0c4355a44a4f409053b51ff675521a465e38aeb8
Reviewed-on: https://chromium-review.googlesource.com/1141114
Reviewed-by: Michael Starzinger <mstarzinger@chromium.org>
Commit-Queue: Bill Budge <bbudge@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54533}
Avoid the C-call to switch the flag, just store to the address
directly. Since js-to-wasm wrappers are still isolate dependent,
we just store the address of the thread-local flag in the isolate
(in ThreadLocalTop) and update it if threads are switched.
R=ahaas@chromium.org, mstarzinger@chromium.org
Bug: chromium:862123, v8:5277
Change-Id: I9e8a40094f11a8b3ba6701dfa7fa026a2d052cb1
Reviewed-on: https://chromium-review.googlesource.com/1136299
Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
Reviewed-by: Andreas Haas <ahaas@chromium.org>
Reviewed-by: Michael Starzinger <mstarzinger@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54531}
Move brackets enclosing declarator in a couple of V8_DEPRECATE_SOON
cases so that the attribute is correctly placed.
I.e. should come after the class keyword and after a function signature
but before the body.
Also adds suppressions for a case where a replacement method calls
the deprecated method it replaces.
Cq-Include-Trybots: luci.chromium.try:linux_chromium_rel_ng
Change-Id: I490ff07a13552cf335186d64ca061f21ac4e8afc
Reviewed-on: https://chromium-review.googlesource.com/1140593
Reviewed-by: Yang Guo <yangguo@chromium.org>
Commit-Queue: Dan Elphick <delphick@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54527}
ReadOnlyRoots means that some added Isolate parameters are no longer
needed. So, we can remove them.
This patch was generated mostly automatically with a bespoke tool.
Bug: v8:7786
Bug: v8:7754
Cq-Include-Trybots: luci.v8.try:v8_linux_noi18n_rel_ng
Change-Id: Ia44fd2a66652253f780e3674bf7fb431caef0493
Reviewed-on: https://chromium-review.googlesource.com/1136305
Commit-Queue: Leszek Swirski <leszeks@chromium.org>
Reviewed-by: Michael Starzinger <mstarzinger@chromium.org>
Reviewed-by: Dan Elphick <delphick@chromium.org>
Reviewed-by: Benedikt Meurer <bmeurer@chromium.org>
Reviewed-by: Yang Guo <yangguo@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54526}
For a script '()=>42', the anonymous arrow function has both start and
end position the same as the script function itself. This causes issues
when sorting the SourcePositionEvents of the function, in two ways:
* If the start positions are the same, we should order by *furthest*
end position to ensure the stack is in the right order
* If both start and end are the same, we need to order by function
literal id to make sure that start order and end order are inversed.
Also, MapLiterals assumes that start+end position uniquely identifies a
function, which is false in this case, so we process the top-level
script function separately in MapLiterals.
Change-Id: I2b2185dc2825018b7ea44c7d0918238e9b1dd972
Reviewed-on: https://chromium-review.googlesource.com/1141741
Commit-Queue: Leszek Swirski <leszeks@chromium.org>
Reviewed-by: Yang Guo <yangguo@chromium.org>
Reviewed-by: Aleksey Kozyatinskiy <kozyatinskiy@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54525}
This CL moves Array.p.fill from JavaScript to a C++ builtin. It has
a generic slow-path and fast-paths implemented via ElementsAccessor in
elements.cc.
R=cbruni@chromium.org
Bug: v8:7624
Change-Id: I8820e1195d2cd9b41c254058923ad9875aab067c
Reviewed-on: https://chromium-review.googlesource.com/1131130
Commit-Queue: Simon Zünd <szuend@google.com>
Reviewed-by: Yang Guo <yangguo@chromium.org>
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54522}
This is no longer necessary since we removed the debug context.
R=jgruber@chromium.org
Bug: v8:5530
Change-Id: Ibb9df3a1f139ee076296faedb80204e7fcc23197
Reviewed-on: https://chromium-review.googlesource.com/1134746
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Commit-Queue: Yang Guo <yangguo@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54521}
This helps debugging unexpected traps, as you can figure out which
instruction triggered the trap.
R=mstarzinger@chromium.org
Change-Id: I61735f14e2838ace195f6b84b555b9ddfc06aa0f
Reviewed-on: https://chromium-review.googlesource.com/1140296
Reviewed-by: Michael Starzinger <mstarzinger@chromium.org>
Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54515}
Convert all remaining uses (in api.cc) to use private local
implementations. These local uses all appear inside deprecated
functions.
Bug: v8:7786
Cq-Include-Trybots: luci.chromium.try:linux_chromium_rel_ng
Change-Id: I86bcafd4c8a737da32fcb8ab275ec708632f9e39
Reviewed-on: https://chromium-review.googlesource.com/1140319
Commit-Queue: Dan Elphick <delphick@chromium.org>
Reviewed-by: Leszek Swirski <leszeks@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54514}
Strictly speaking there are some left in api.cc, but they are in
deprecated functions with non-deprecated alternatives.
Apart from changes made using tooling, this also modifies
FieldType::AsClass to return Map* rather than Handle<Map> and converts
its call sites to create the Handle when they need it - currently
several sites immediately dereference the Handle.
Also marks WasmDebugInfo as NeverReadOnlySpaceObject so GetIsolate and
GetHeap remain usable.
Bug: v8:7786
Cq-Include-Trybots: luci.chromium.try:linux_chromium_rel_ng
Change-Id: I7ea5048f97f140c757f651712b8c33a5c7e0ebc1
Reviewed-on: https://chromium-review.googlesource.com/1140302
Reviewed-by: Adam Klein <adamk@chromium.org>
Reviewed-by: Yang Guo <yangguo@chromium.org>
Reviewed-by: Leszek Swirski <leszeks@chromium.org>
Reviewed-by: Michael Starzinger <mstarzinger@chromium.org>
Commit-Queue: Dan Elphick <delphick@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54513}
This removes two pointers to Histograms from the memory tracker. These
histograms are stored as part of the Isolate and their lifetime is also
coupled to the Isolate. We cannot bind the pointers but need to pass
them (or the Isolate) as a parameter instead.
R=clemensh@chromium.org
BUG=v8:7424
Change-Id: I6b141b924bd858234641d6603a25fcb08cdf40e3
Reviewed-on: https://chromium-review.googlesource.com/1140312
Commit-Queue: Michael Starzinger <mstarzinger@chromium.org>
Reviewed-by: Clemens Hammacher <clemensh@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54512}
Now we can finally get rid of Map::weak_cell_cache!
BUG=v8:7308
Change-Id: I87a06509bf638bf6833ea2ba1eca525fb4b15df1
Reviewed-on: https://chromium-review.googlesource.com/1128882
Commit-Queue: Marja Hölttä <marja@chromium.org>
Reviewed-by: Ulan Degenbaev <ulan@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54511}
The BodyDescriptor of an object should use its aligned size.
Change-Id: If743ca130b3cb97c4f25054db6dc887d88fc5e32
Reviewed-on: https://chromium-review.googlesource.com/1140309
Reviewed-by: Ulan Degenbaev <ulan@chromium.org>
Commit-Queue: Leszek Swirski <leszeks@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54510}
Brokerized ReduceJSCreateEmptyLiteralObject and added the scope
for ReduceJSCreateLiteralArrayOrObject.
Bug: v8:7790
Change-Id: Ife34a6b610678a3fe24152151cf343400ee515bd
Reviewed-on: https://chromium-review.googlesource.com/1140306
Commit-Queue: Jaroslav Sevcik <jarin@chromium.org>
Reviewed-by: Georg Neis <neis@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54507}
We try to prevent side effects by forbidding running any JavaScript
when we get property from node object.
In case of object node it is possible that by calling property we force
internal object initialization which may force creation of new context,
this initialization can not be made with forbided JavaScript and at the
same time is side effect free.
As workaround we can warmup dom objects first and then generate
description.
R=dgozman@chromium.org
Bug: chromium:827585
Cq-Include-Trybots: master.tryserver.blink:linux_trusty_blink_rel
Change-Id: Ifd2c6317ffd5cb3822d2a2eedf3d0b0f36a201f1
Reviewed-on: https://chromium-review.googlesource.com/1041078
Reviewed-by: Dmitry Gozman <dgozman@chromium.org>
Commit-Queue: Aleksey Kozyatinskiy <kozyatinskiy@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54505}
This patch makes `d8` recognize files with the `.mjs` extension as
modules instead of classic scripts. This change can be tested by saving
the following JavaScript program as both `module.mjs` and as
`script.js`:
console.log(this === undefined ? 'strict' : 'sloppy');
Then, run these files in `d8` without passing the `--module` flag:
$ d8 module.mjs
strict
$ d8 script.js
sloppy
The use of `.mjs` matches not just Google’s recommendation [1] but also
the current modules implementation in Node.js [2].
[1] https://developers.google.com/web/fundamentals/primers/modules
[2] https://nodejs.org/api/esm.html
Bug: v8:7950
Change-Id: I8f39420dc24a5eedd7e88d3b1aa48207ebfeff6e
Reviewed-on: https://chromium-review.googlesource.com/1140314
Commit-Queue: Mathias Bynens <mathias@chromium.org>
Reviewed-by: Georg Neis <neis@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54502}
Struct are bundles of value types. They are essentially just shorthand
for passing around a group of individually defined values.
Struct types are declared like this:
struct A {
x: Smi;
y: int32;
}
and can be constructed explicitly like this:
A{0, 0}
Structs can be used wherever other types are used (e.g. variables,
parameters, return values) except for parameter/return types of
builtins and runtime functions.
Struct use field access notation to set/get their values like this:
let a: A = A{0, 0};
let b: Smi = a.x;
a.y = 0;
Change-Id: I9fd36a6514c37882831256a49a50809c5db75b56
Reviewed-on: https://chromium-review.googlesource.com/1122133
Commit-Queue: Daniel Clifford <danno@chromium.org>
Reviewed-by: Tobias Tebbi <tebbi@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54501}
i32 stack parameters can be loaded by Turbofan as 64-bit value, hence
they would not be zero extended. If this loaded value is then passed to
Liftoff (which assumes zero-extended i32 values), we could use it for
memory accesses, which would be out of bounds.
R=mstarzinger@chromium.org
Bug: chromium:864509, v8:6600
Change-Id: I0f45a269b1fb1c2befc2e6bc660c559a88323767
Reviewed-on: https://chromium-review.googlesource.com/1140168
Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
Reviewed-by: Michael Starzinger <mstarzinger@chromium.org>
Reviewed-by: Clemens Hammacher <clemensh@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54500}
This reverts commit f5a8352b0f.
Reason for revert: Performance issues
Original change's description:
> [embedded-builtins] Enable on all arches except x86 for benchmarks
>
> This CL enables embedded builtins to get benchmark feedback. We need
> this feedback to identify and address remaining performance problems.
>
> Bug: v8:6666
> Change-Id: I8f77f218e656b55ddabe1236eb2a1d14a5ac6233
> Reviewed-on: https://chromium-review.googlesource.com/1105834
> Reviewed-by: Jakob Gruber <jgruber@chromium.org>
> Commit-Queue: Sigurd Schneider <sigurds@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#53836}
TBR=sigurds@chromium.org,jgruber@chromium.org
# Not skipping CQ checks because original CL landed > 1 day ago.
Bug: v8:6666
Change-Id: I0e0897eefa069b0b9ad2dd56b2ffc3e3617f9258
Reviewed-on: https://chromium-review.googlesource.com/1139974
Commit-Queue: Sigurd Schneider <sigurds@chromium.org>
Reviewed-by: Sigurd Schneider <sigurds@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54498}
This changes the ARM64-specific {TurboAssembler::AssertSpAligned} helper
to not generate calls to the {Abort} builtin. It is needed to ensure all
WebAssembly runtime stubs (e.g. {WasmGrowMemory}) are independent of the
Isolate. In general calling the {Abort} builtin without a valid frame
being present will produce bogus debug messages anyways. Hence we just
unconditionally use traps for the debug code in question.
R=sigurds@chromium.org
Change-Id: I93eb87e8b87209da8506c9b28e2c800950d1118a
Reviewed-on: https://chromium-review.googlesource.com/1140170
Commit-Queue: Michael Starzinger <mstarzinger@chromium.org>
Reviewed-by: Sigurd Schneider <sigurds@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54496}
We used to have an optimized version for ToString on number nodes
which was allocating an object on the heap, therefore
preventing this code from being executed on the compiler thread.
Octane benchmark results show insignificant increase in performance
(< 0.5%) without this optimization - see
https://docs.google.com/spreadsheets/d/1MC5NrMoMSsqxZqw0ojoZvomBb7q2EOt1S0sFoJ8ld2c/edit#gid=1732639373
which leads to the conclusion we can safely remove the optimization for now.
Bug: v8:7790
Change-Id: Ia1d53608f8d10ba20e0ff57cccb34583655382c6
Reviewed-on: https://chromium-review.googlesource.com/1139063
Commit-Queue: Maya Lekova <mslekova@chromium.org>
Reviewed-by: Jaroslav Sevcik <jarin@chromium.org>
Reviewed-by: Georg Neis <neis@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54495}
Also moves ObjectVerify to GlobalHandles::CopyGlobal from
V8::CopyPersistent (which was the only caller) so it can get hold of an
Isolate*.
Bug: v8:7786
Cq-Include-Trybots: luci.chromium.try:linux_chromium_rel_ng
Change-Id: I0758bf6e431bf6e617244741ab2e1583a3566b20
Reviewed-on: https://chromium-review.googlesource.com/1140295
Reviewed-by: Leszek Swirski <leszeks@chromium.org>
Commit-Queue: Dan Elphick <delphick@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54493}
