The embedder should ultimately be responsible for handling this since they
anyway give us a copy of the data. They can easily make sure that the chunks we
get do not have lonely bytes.
Cq-Include-Trybots: luci.chromium.try:linux_chromium_rel_ng
Change-Id: Ie862107bbbdd00c4d904fbb457a206c2fd52e5d0
Reviewed-on: https://chromium-review.googlesource.com/1127044
Reviewed-by: Ulan Degenbaev <ulan@chromium.org>
Reviewed-by: Marja Hölttä <marja@chromium.org>
Commit-Queue: Toon Verwaest <verwaest@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54262}
This introduces a second factory method for allocating new module
objects that are based on existing native modules which are potentially
shared between multiple Isolates.
R=clemensh@chromium.org
TEST=cctest/test-wasm-shared-engine
BUG=v8:7424
Change-Id: I8c74d821542d443a8ad2352cb77b84b445e21acb
Reviewed-on: https://chromium-review.googlesource.com/1126931
Commit-Queue: Michael Starzinger <mstarzinger@chromium.org>
Reviewed-by: Clemens Hammacher <clemensh@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54261}
Creates two different functions specialized for string and boolean
types.
Gets rid of several allocations and keeps the C++ <-> JS type
conversions to a minimum. Improves the API as we don't have to create
the fallback or V8 Strings unnecessarily.
Bug: v8:5751
Cq-Include-Trybots: luci.v8.try:v8_linux_noi18n_rel_ng
Change-Id: I788e43e6ef23f3e9144ff719f01d6334fe6cb9ce
Reviewed-on: https://chromium-review.googlesource.com/1126750
Commit-Queue: Sathya Gunasekaran <gsathya@chromium.org>
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Reviewed-by: Camillo Bruni <cbruni@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54259}
When assigning a register to a value that is used as a function
argument, we use a hint to suggest the corresponding register. However,
if the argument is also used after the call, the register will not be
free for the entire live range of the value. Hence we need to split the
live range.
To minimize the number of splits, we aim to choose a register with
maxium availability. This heuristic was implemented based on lifetime
positions with sub-instruction precision. In such a model, argument
registers typically have a shorter available time, as they need to
hold a value before a call, whereas all other registers are free until
right after the call, where they have been overwritten by the called
function. Hence, we typically chose a non-argument register, ignoring
the hint and creating an extra move.
This change moves the heuristic to instruction granularity, which
gives argument and other registers the same free time. We also now
prefer hinted registers if they have the same free time.
Change-Id: Ia8dd73b6c086d28859a836c42ea9ff8afce4c371
Reviewed-on: https://chromium-review.googlesource.com/1124852
Reviewed-by: Sigurd Schneider <sigurds@chromium.org>
Commit-Queue: Stephan Herhut <herhut@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54255}
This is a reland of 52a10e5081, after
eliminating an invalid assumption about maps.
TBR=jarin@chromium.orgTBR=mstarzinger@chromium.org
Original change's description:
> [turbofan] Rewrite CompilationDependencies
>
> Instead of installing code dependencies during graph reduction,
> install them after code generation.
>
> Bug: v8:7902, v8:7790
> Change-Id: I8a3798254abb5b9ec7c295a1592aeb6b51f24c7a
> Reviewed-on: https://chromium-review.googlesource.com/1119913
> Commit-Queue: Georg Neis <neis@chromium.org>
> Reviewed-by: Jaroslav Sevcik <jarin@chromium.org>
> Reviewed-by: Michael Starzinger <mstarzinger@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#54170}
Bug: v8:7902, v8:7790
Change-Id: I9cbaf98980379b9b17464af5952ec0c47e1cdc6f
Reviewed-on: https://chromium-review.googlesource.com/1126999
Reviewed-by: Georg Neis <neis@chromium.org>
Commit-Queue: Georg Neis <neis@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54254}
Previously we were getting the code object from the stack, so printed incorrect
position details for interpreted frames.
BUG=v8:7916
Change-Id: I2f87584117d88b7db3f3b9bdbfe793c4d3e33fe9
Reviewed-on: https://chromium-review.googlesource.com/1126313
Reviewed-by: Toon Verwaest <verwaest@chromium.org>
Commit-Queue: Ross McIlroy <rmcilroy@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54253}
This CL is a manual revert of 3e78711c21.
Adding the fast path did not result in any performance improvement,
and actually decreased DataView performance in the performance test.
Change-Id: I149670088b07c31e44d93ed65505254ec0caf8ba
Reviewed-on: https://chromium-review.googlesource.com/1126925
Reviewed-by: Michael Stanton <mvstanton@chromium.org>
Commit-Queue: Théotime Grohens <theotime@google.com>
Cr-Commit-Position: refs/heads/master@{#54251}
This is a first set of test cases for sharing an {WasmEngine} and the
contained {WasmCode} between multiple Isolates. Currently this can only
be done using internal API methods on the Isolate, an external API that
is usable by embedders does not exist yet.
R=clemensh@chromium.org
TEST=cctest/test-wasm-shared-engine
BUG=v8:7424
Change-Id: I35541a76b5aceec4519e3a46e6a9ef4d01cad22b
Reviewed-on: https://chromium-review.googlesource.com/1126382
Reviewed-by: Clemens Hammacher <clemensh@chromium.org>
Commit-Queue: Michael Starzinger <mstarzinger@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54248}
MaybeObjectHandle already has an API for making a weak pointer.
BUG=v8:7308
Change-Id: I9f8390e0dffec850a24391e7f50a455b86b217f5
Reviewed-on: https://chromium-review.googlesource.com/1125922
Reviewed-by: Michael Starzinger <mstarzinger@chromium.org>
Commit-Queue: Marja Hölttä <marja@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54247}
In the WebAssembly fuzzers we detect infinite loops with the
interpreter: if the interpreter does not finish after a finite number
of steps, we do not execute the compiled code. However, we cannot
redirect the start function to the interpreter in the fuzzer, and
therefore we cannot detect infinite loops in the start function. With
this CL we avoid the problem completely by not instantiating a module
in the fuzzer which has a start function. Note that the module still
gets compiled.
R=clemensh@chromium.org
Bug: chromium:858914
Change-Id: Icbbe9a003544918d5267cdd1d9405b21bb681133
Reviewed-on: https://chromium-review.googlesource.com/1126766
Commit-Queue: Andreas Haas <ahaas@chromium.org>
Reviewed-by: Clemens Hammacher <clemensh@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54246}
Replace Is<oddball>(GetIsolate()) calls with a no-parameter version
that goes through ReadOnlyRoots, and add a version that takes a
ReadOnlyRoots if that is available in the parent (but Isolate isn't).
Also opportunistically clean up a few places where ReadOnlyRoots are
available but we still pass in an Isolate parameter.
TBR=yangguo@chromium.org
Bug: v8:7786
Cq-Include-Trybots: luci.chromium.try:linux_chromium_rel_ng
Change-Id: Icc0b8a7d8f9c6b84c0ec8fe771fcfb75c9dc5468
Reviewed-on: https://chromium-review.googlesource.com/1126302
Reviewed-by: Leszek Swirski <leszeks@chromium.org>
Reviewed-by: Camillo Bruni <cbruni@chromium.org>
Reviewed-by: Michael Starzinger <mstarzinger@chromium.org>
Reviewed-by: Dan Elphick <delphick@chromium.org>
Commit-Queue: Leszek Swirski <leszeks@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54245}
Mips and mips64 generate non-position-independent trampolines once code
exceeds 128K in size. This is a quick fix for failing bots, disabling
the ten largest builtins.
TBR=machenbach@chromium.org
Bug: v8:7882, v8:7920
Change-Id: I027ff8452f927f890773210fb99853adf4934bbc
Reviewed-on: https://chromium-review.googlesource.com/1126923
Commit-Queue: Jakob Gruber <jgruber@chromium.org>
Reviewed-by: Michael Achenbach <machenbach@chromium.org>
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54244}
This is a port of e5630ea9:
"[builtins,x64] pc-relative builtin-to-builtin calls".
Bug: v8:6666
Change-Id: Id53bf45c6a4fedb1c972ced72984b6673808ee96
Reviewed-on: https://chromium-review.googlesource.com/1126249
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Commit-Queue: Georgia Kouveli <georgia.kouveli@arm.com>
Cr-Commit-Position: refs/heads/master@{#54243}
This is a reland of 2265a852b4
Original change's description:
> [snapshot] Reduce serialization chunk size to 4K.
>
> This change may slow down deserialization of the snapshot but may result in less fragmentation.
>
> Bug: v8:7887
> Change-Id: Id8bb9f1a561b08b7ae0f10b80aa77bc00eb23172
> Reviewed-on: https://chromium-review.googlesource.com/1125722
> Reviewed-by: Yang Guo <yangguo@chromium.org>
> Commit-Queue: Hannes Payer <hpayer@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#54198}
TBR=hpayer@chromium.org
Bug: v8:7887
Change-Id: If03f241a80341cb053190c5bc0451fe9a032e1a8
Reviewed-on: https://chromium-review.googlesource.com/1126899
Reviewed-by: Yang Guo <yangguo@chromium.org>
Commit-Queue: Yang Guo <yangguo@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54241}
Now that ReadOnlyRoots is used everywhere to access the read-only roots,
this makes the Heap accessors for such roots private.
It also adds tests that the roots reachable from ReadOnlyRoots are all
in RO_SPACE as well as tests that the roots still publicly accessible
from Heap are not in RO_SPACE. There's a white list in the file for
the few roots where the root pointer itself can change. (For instance
materialized_objects points to empty_fixed_array to start with before
before later pointing to a mutable array).
Also fixes up new use of heap->empty_fixed_array() in elements.cc added
since I cleaned it up.
Bug: v8:7786
Change-Id: I9ac7985c9f85910b5b22d2f9f559dfd04d43ed44
Reviewed-on: https://chromium-review.googlesource.com/1126252
Reviewed-by: Hannes Payer <hpayer@chromium.org>
Commit-Queue: Dan Elphick <delphick@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54240}
We can actually prepare everything in the background, all that remains
to do in the finisher task (on the main thread) is actually adding the
code to the {NativeModule}.
As a next step, even that should happen in the background.
R=mstarzinger@chromium.org
Bug: v8:6600
Change-Id: I570f99a9aa7dc7e324046da36cca9b4297f1bc5e
Reviewed-on: https://chromium-review.googlesource.com/1126391
Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
Reviewed-by: Michael Starzinger <mstarzinger@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54239}
The stack of the wasm interpreter was stored in a zone. However, when
the stack grows, the previous memory of the stack cannot be reused.
With this CL we allocate the stack on the C++ heap to remove this memory
leak.
R=clemensh@chromium.org
Bug: chromium:856594
Change-Id: Idce22c5c1732f7097fc4281c73174892c95f49e2
Reviewed-on: https://chromium-review.googlesource.com/1126301
Commit-Queue: Andreas Haas <ahaas@chromium.org>
Reviewed-by: Clemens Hammacher <clemensh@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54238}
Fixes V8 correctness failure when there's a proxy in the global object
prototype chain and unsuccessful attempt is made to access a property.
Bug: chromium:849024
Change-Id: I829e1a6c038982b7c7a77f8bdefb61facb4614f0
Reviewed-on: https://chromium-review.googlesource.com/1124446
Commit-Queue: Maya Lekova <mslekova@chromium.org>
Reviewed-by: Igor Sheludko <ishell@chromium.org>
Reviewed-by: Benedikt Meurer <bmeurer@chromium.org>
Reviewed-by: Georg Neis <neis@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54237}
The code target vector is used by backends whenever initial code
creation uses immediates that cannot hold a code pointer: An index into
the vector is used instead, and the vector contains the corresponding
code object. This CL makes the code target vector available on all
platforms.
Bug: v8:6666
Change-Id: I2026acd6a77a3465fe2285b84d47866c3b794855
Reviewed-on: https://chromium-review.googlesource.com/1118885
Commit-Queue: Sigurd Schneider <sigurds@chromium.org>
Reviewed-by: Michael Starzinger <mstarzinger@chromium.org>
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54236}
If the first object in the space already exceeds the target chunk size,
we would leave the first chunk empty. That violates some assumptions
later when we allocate for deserialization.
R=hpayer@chromium.org, jgruber@chromium.orgTBR=hpayer@chromium.org
Bug: v8:7887
Change-Id: Iee8147fe1205bb6b1c893d48acde1099b5032a14
Reviewed-on: https://chromium-review.googlesource.com/1126763
Reviewed-by: Hannes Payer <hpayer@chromium.org>
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Commit-Queue: Yang Guo <yangguo@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54235}
This just fixes the text for the '--liftoff' and '--trace-liftoff'
flags to avoid calling Liftoff "experimental".
R=ahaas@chromium.org
Bug: v8:6600
Change-Id: Iff9032070fe0e18ee77debf376f5453d1a61480e
Reviewed-on: https://chromium-review.googlesource.com/1126114
Reviewed-by: Andreas Haas <ahaas@chromium.org>
Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54233}
After a certain number of iterations in the fixpoint
iteration switch to a linear algorithm. This
algorithm uses a key-to-values map for weak collections
contents to avoid checking all EphemeronHashTables
again.
Bug: chromium:844008
Change-Id: I044fede5911e0a780d088d1ba2bb5343317d9b7a
Reviewed-on: https://chromium-review.googlesource.com/1105835
Commit-Queue: Dominik Inführ <dinfuehr@google.com>
Reviewed-by: Ulan Degenbaev <ulan@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54232}
This is just derived from is_wasm_memory.
Change-Id: I2f77fb5e32e325c51de9af4228ca33313c21abc6
Reviewed-on: https://chromium-review.googlesource.com/1126107
Reviewed-by: Michael Starzinger <mstarzinger@chromium.org>
Commit-Queue: Peter Marshall <petermarshall@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54230}
Only pass around the unadulterated value identifier in the VisitResult class
until the very last moment before code generation, at which point the
declaration that was used to originally define the value is used to generate the
correct final source code string in the context of a l-value or r-value.
Bug: v8:7793
Change-Id: Ifd0c0d245b2eb65c7f3ddb1ad4c87ee235c54a82
Reviewed-on: https://chromium-review.googlesource.com/1125063
Commit-Queue: Daniel Clifford <danno@chromium.org>
Reviewed-by: Tobias Tebbi <tebbi@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54229}
Port 34225a6afb
Original Commit Message:
For spread calls with arrays with double elements but zero length,
we skip the box-as-heapnumber step; so in this corner case the
Call builtin sees a FixedDoubleArray, which is fine because it
doesn't read any of the raw double values from it.
This patch doesn't change the implementation, it only updates the
assert to match reality.
R=jkummerow@chromium.org, joransiu@ca.ibm.com, michael_dawson@ca.ibm.com
BUG=
LOG=N
Change-Id: Iafa52f21fb0bbee5656fdfd6c5f3a50894ff683f
Reviewed-on: https://chromium-review.googlesource.com/1126212
Reviewed-by: Michael Dawson <michael_dawson@ca.ibm.com>
Commit-Queue: Junliang Yan <jyan@ca.ibm.com>
Cr-Commit-Position: refs/heads/master@{#54227}
This DCHECK does not hold for compaction spaces that may go over the limit
with large semi-spaces and small old generation size maxima.
Bug: chromium:855960
Change-Id: I39eea4c974b94c170e7314471fb1c50f9a1a2b45
Reviewed-on: https://chromium-review.googlesource.com/1126392
Reviewed-by: Michael Lippautz <mlippautz@chromium.org>
Commit-Queue: Hannes Payer <hpayer@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54226}
In future the RO_SPACE root accessors in Heap will become private, so
instead convert them all to use ReadOnlyRoots.
Bug: v8:7786
Change-Id: I315e63a30ca6f3077f18eb1a5004daefba9bc673
Reviewed-on: https://chromium-review.googlesource.com/1125929
Commit-Queue: Dan Elphick <delphick@chromium.org>
Reviewed-by: Michael Starzinger <mstarzinger@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54223}
In future the RO_SPACE root accessors in Heap will become private, so
instead convert them all to use ReadOnlyRoots.
Bug: v8:7786
Change-Id: Ib3c45c1023d76bec5e1f4bc8f971062880b6c53f
Reviewed-on: https://chromium-review.googlesource.com/1126240
Reviewed-by: Michael Starzinger <mstarzinger@chromium.org>
Commit-Queue: Dan Elphick <delphick@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54221}
Changes SharedFunctionInfo to store a function_token_offset, relative
to the start_position, instead of the full function_token_position.
This enables us to reduce both FunctionTokenPosition and
ExpectedNofProperties to 16 bits each, saving 32 bits per SFI.
BUG=chromium:818642,chromium:783853
TBR=yangguo@chromium.org
Cq-Include-Trybots: luci.chromium.try:linux_chromium_rel_ng
Change-Id: I45aefcec605c1da502053c23c73564ceaed6c9b5
Reviewed-on: https://chromium-review.googlesource.com/1122982
Commit-Queue: Ross McIlroy <rmcilroy@chromium.org>
Reviewed-by: Ross McIlroy <rmcilroy@chromium.org>
Reviewed-by: Yang Guo <yangguo@chromium.org>
Reviewed-by: Ulan Degenbaev <ulan@chromium.org>
Reviewed-by: Sathya Gunasekaran <gsathya@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54220}
This CL creates a new Operator called LoadDataViewElement, similar to
LoadTypedArray, for DataView getters.
This operator will be used as a wrapper around all the computations
that DataViews need to do when loading values, due to the endianness
parameter of DataView loads.
Change-Id: Ie67d63c9669142e539a5c8d7ae82dc1018ce5858
Reviewed-on: https://chromium-review.googlesource.com/1125928
Commit-Queue: Théotime Grohens <theotime@google.com>
Reviewed-by: Jaroslav Sevcik <jarin@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54217}
In future the RO_SPACE root accessors in Heap will become private, so
instead convert them all to use ReadOnlyRoots.
Bug: v8:7786
Cq-Include-Trybots: luci.v8.try:v8_linux_noi18n_rel_ng
Change-Id: I4a5fb4bc70169874d003d9c31f1713f52e2bba93
Reviewed-on: https://chromium-review.googlesource.com/1126118
Reviewed-by: Leszek Swirski <leszeks@chromium.org>
Commit-Queue: Dan Elphick <delphick@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54216}
In future the RO_SPACE root accessors in Heap will become private, so
instead convert them all to use ReadOnlyRoots.
Bug: v8:7786
Change-Id: I631fa1e44d2b9998a5567c96c6977955ac86d5a1
Reviewed-on: https://chromium-review.googlesource.com/1125931
Reviewed-by: Yang Guo <yangguo@chromium.org>
Commit-Queue: Dan Elphick <delphick@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54215}
In future the RO_SPACE root accessors in Heap will become private, so
instead convert them all to use ReadOnlyRoots.
Bug: v8:7786
Change-Id: I6a7fb375411e379681637469bb3fb360635e6bbe
Reviewed-on: https://chromium-review.googlesource.com/1126111
Reviewed-by: Igor Sheludko <ishell@chromium.org>
Commit-Queue: Dan Elphick <delphick@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54214}
