This is a reland of 20f33823c9.
Win32 failures should be fixed by https://crrev.com/c/1120175.
Original change's description:
> [wasm] Enable Liftoff by default on ia32 and x64
>
> R=titzer@chromium.org, hablich@chromium.org
>
> Bug: v8:6600, chromium:787421
> Change-Id: Ia8ae56ddef3b27b0721d5a66ff19abe098a2c6ca
> Reviewed-on: https://chromium-review.googlesource.com/1109899
> Reviewed-by: Ben Titzer <titzer@chromium.org>
> Reviewed-by: Michael Hablich <hablich@chromium.org>
> Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#53965}
TBR=titzer@chromium.org, hablich@chromium.org
Bug: v8:6600, chromium:787421
Change-Id: Ia0fc1f152988a2df95d94d002b9ddfdc0a9c4abd
Reviewed-on: https://chromium-review.googlesource.com/1120205
Reviewed-by: Clemens Hammacher <clemensh@chromium.org>
Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54154}
Only use the "W" part (lower 32 bit) of the src register. Otherwise, we
can get results larger than 32.
R=ahaas@chromium.org
CC=rodolph.perfetta@arm.com
Bug: v8:7914, chromium:854011
Change-Id: I6329231e6cc0ae537c165b2d383fc5a14bd28ca3
Reviewed-on: https://chromium-review.googlesource.com/1122409
Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
Reviewed-by: Andreas Haas <ahaas@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54152}
This reverts commit 3dfaf8264f.
Reason for revert: Failures - https://ci.chromium.org/p/v8/builders/luci.v8.ci/V8%20Linux%20gcc%204.8/20394
Original change's description:
> [debug] liveedit in native
>
> Liveedit step-by-step:
> 1. calculate diff between old source and new source,
> 2. map function literals from old source to new source,
> 3. create new script for new_source,
> 4. mark literals with changed code as changed, all others as unchanged,
> 5. check that for changed literals there are no:
> - running generators in the heap,
> - non droppable frames (e.g. running generator) above them on stack.
> 6. mark the bottom most frame with changed function as scheduled for
> restart if any.
> 7. for unchanged functions:
> - deoptimize,
> - remove from cache,
> - update source positions,
> - move to new script,
> - reset feedback information and preparsed scope information if any,
> - replace any sfi in constant pool with changed one if any.
> 8. for changed functions:
> - deoptimize
> - remove from cache,
> - reset feedback information,
> - update all links from js functions to old shared with new one.
> 9. swap scripts.
>
> TBR=ulan@chromium.org
>
> Bug: v8:7862,v8:5713
> Cq-Include-Trybots: luci.chromium.try:linux_chromium_headless_rel;luci.chromium.try:linux_chromium_rel_ng;master.tryserver.blink:linux_trusty_blink_rel
> Change-Id: I8f6f6156318cc82d6f36d7ebc1c9f7d5f3aa1461
> Reviewed-on: https://chromium-review.googlesource.com/1105493
> Reviewed-by: Aleksey Kozyatinskiy <kozyatinskiy@chromium.org>
> Reviewed-by: Dmitry Gozman <dgozman@chromium.org>
> Reviewed-by: Yang Guo <yangguo@chromium.org>
> Commit-Queue: Aleksey Kozyatinskiy <kozyatinskiy@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#54146}
TBR=dgozman@chromium.org,ulan@chromium.org,yangguo@chromium.org,kozyatinskiy@chromium.org
Change-Id: I45df5b6f3abaf29e593c6ac11edefbd0177d0109
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Bug: v8:7862, v8:5713
Cq-Include-Trybots: luci.chromium.try:linux_chromium_headless_rel;luci.chromium.try:linux_chromium_rel_ng;master.tryserver.blink:linux_trusty_blink_rel
Reviewed-on: https://chromium-review.googlesource.com/1124159
Reviewed-by: Yang Guo <yangguo@chromium.org>
Commit-Queue: Yang Guo <yangguo@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54151}
This CL changes the NumberDictionary fast-path for Array.p.sort to
throw a TypeError when trying to write to a read-only property.
Previously, the fast-path simply bailed to the slow-path which could
swallow the TypeError by accident. I.e. because the fast-path could
leave the array in an inconsistent state that is already sorted.
Example:
let arr = new Array(10);
Object.defineProperty(arr, 0, {value: 2, writable: false});
Object.defineProperty(arr, 2, {value: 1, writable: false});
arr.sort();
The pre-processing step will move the value 1 to index 1: {0: 2, 1: 1}
When trying to swap those 2 values, the fast-path will write the 2 at
index 1, then try to write the 1 at index 0 and fail, bailing to the
slow-path. As the array looks like {0: 2, 1: 2} its already sorted
and the TypeError will not be thrown.
R=jgruber@chromium.org
Bug: v8:7382, v8:7907
Change-Id: I5d2f2d73478fdca066ce1048dcb2b8301751cb1f
Reviewed-on: https://chromium-review.googlesource.com/1122120
Commit-Queue: Simon Zünd <szuend@google.com>
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54150}
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.
Bug: chromium:856095
Change-Id: I0227f4ccbc6c61c8f5f7669a266ef7a64c6a9a43
Reviewed-on: https://chromium-review.googlesource.com/1117922
Reviewed-by: Camillo Bruni <cbruni@chromium.org>
Commit-Queue: Jakob Kummerow <jkummerow@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54149}
Also improve its code generation.
Change-Id: I4358500a66b0d21cdc8850a4e63986c4901d3cf4
Reviewed-on: https://chromium-review.googlesource.com/1118005
Commit-Queue: Jing Bao <jing.bao@intel.com>
Reviewed-by: Bill Budge <bbudge@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54147}
Liveedit step-by-step:
1. calculate diff between old source and new source,
2. map function literals from old source to new source,
3. create new script for new_source,
4. mark literals with changed code as changed, all others as unchanged,
5. check that for changed literals there are no:
- running generators in the heap,
- non droppable frames (e.g. running generator) above them on stack.
6. mark the bottom most frame with changed function as scheduled for
restart if any.
7. for unchanged functions:
- deoptimize,
- remove from cache,
- update source positions,
- move to new script,
- reset feedback information and preparsed scope information if any,
- replace any sfi in constant pool with changed one if any.
8. for changed functions:
- deoptimize
- remove from cache,
- reset feedback information,
- update all links from js functions to old shared with new one.
9. swap scripts.
TBR=ulan@chromium.org
Bug: v8:7862,v8:5713
Cq-Include-Trybots: luci.chromium.try:linux_chromium_headless_rel;luci.chromium.try:linux_chromium_rel_ng;master.tryserver.blink:linux_trusty_blink_rel
Change-Id: I8f6f6156318cc82d6f36d7ebc1c9f7d5f3aa1461
Reviewed-on: https://chromium-review.googlesource.com/1105493
Reviewed-by: Aleksey Kozyatinskiy <kozyatinskiy@chromium.org>
Reviewed-by: Dmitry Gozman <dgozman@chromium.org>
Reviewed-by: Yang Guo <yangguo@chromium.org>
Commit-Queue: Aleksey Kozyatinskiy <kozyatinskiy@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54146}
Removes JS version and creates a runtime functions for now to
interface with existing JS uses.
Bug: v8:5751
Cq-Include-Trybots: luci.v8.try:v8_linux_noi18n_rel_ng
Change-Id: I150701d338a0951a5e5da1aca667c65f941850d9
Reviewed-on: https://chromium-review.googlesource.com/1122024
Commit-Queue: Sathya Gunasekaran <gsathya@chromium.org>
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54145}
Currently we rely on tasks to invoke the second pass phantom callbacks.
This may accumulate phantom callbacks and make GCs ineffective if we
do not enter the message loop to run the tasks between the GCs.
Bug: v8:7912
Change-Id: I799c97ff99ed6967480bda24ea0bf1c6a7dd69be
Reviewed-on: https://chromium-review.googlesource.com/1122621
Commit-Queue: Ulan Degenbaev <ulan@chromium.org>
Reviewed-by: Clemens Hammacher <clemensh@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54144}
Adds Intl::IsObjectOfType method to do type checks. This will make it
easier to port the methods using the runtime type check calls as we
won't have to create a v8::string for type checks.
Bug: v8:5751
Cq-Include-Trybots: luci.v8.try:v8_linux_noi18n_rel_ng
Change-Id: I0babdc8709564be693ce808e2ef3ffef7b24ceec
Reviewed-on: https://chromium-review.googlesource.com/1121943
Commit-Queue: Sathya Gunasekaran <gsathya@chromium.org>
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54143}
This should restore the old behavior of giving low priority to
newly-added live ranges.
Bug: chromium:859021
Change-Id: If22c9a1d0897d82623eb141fa03c30110e68bfc4
Reviewed-on: https://chromium-review.googlesource.com/1122402
Reviewed-by: Stephan Herhut <herhut@chromium.org>
Commit-Queue: Tobias Tebbi <tebbi@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54142}
On Windows (32-bit), we need to emit explicit stack limit checks for
stack frames bigger than one page (4kB). This CL implements this by
emitting corresponding code at the end of Liftoff functions if needed.
R=mstarzinger@chromium.org
Bug: v8:7908, v8:6600
Change-Id: Iacb3e7afdd433a4e68620d9230bd0ba473611da8
Reviewed-on: https://chromium-review.googlesource.com/1120175
Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
Reviewed-by: Michael Starzinger <mstarzinger@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54141}
This correctly serializes {RelocInfo::INTERNAL_REFERENCE} addresses in a
position-independent form, so that they can be properly relocated when
the code is deserialized again. We store the offset within the code in
the serialized stream.
R=clemensh@chromium.org
TEST=mjsunit/wasm/compiled-module-serialization
BUG=chromium:857049
Change-Id: Ie8c84ee67bdfc17a65faa159a21cc1f2a78ac924
Reviewed-on: https://chromium-review.googlesource.com/1122414
Reviewed-by: Clemens Hammacher <clemensh@chromium.org>
Commit-Queue: Michael Starzinger <mstarzinger@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54140}
Pass Isolate to ConsumedPreParsedScopeData::SetData (guaranteed to be
called on main thread) and use it to create the handle in
ConsumedPreParsedScopeData::GetDataForSkippableFunction, rather than
calling GetIsolate.
Bug: v8:7786
Change-Id: Ibd632bb57f35a921f37c620d77dd6dfdb1f092c6
Reviewed-on: https://chromium-review.googlesource.com/1088703
Commit-Queue: Dan Elphick <delphick@chromium.org>
Reviewed-by: Leszek Swirski <leszeks@chromium.org>
Reviewed-by: Marja Hölttä <marja@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54138}
In the process, create a shared array utility GetLengthProperty that fast-paths
accessing the length properties of JSArray.
Bug: v8:7793
Change-Id: I6d7f0007c162794773dc0fc3e8bf12b3adf12fa0
Reviewed-on: https://chromium-review.googlesource.com/1116221
Commit-Queue: Daniel Clifford <danno@chromium.org>
Reviewed-by: Tobias Tebbi <tebbi@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54133}
When parsing a numeric literal in a line like "a=0x0e+b|0;",
currently the scanner consumes the "e+" part (as it thinks
it's the start of an exponent).
In the ECMAScript lexical grammar HexIntegerLiteral cannot
contain exponents, which means the '+' character should be
parsed as a binary operator.
R=bradnelson@chromium.org
BUG=v8:7893
Change-Id: I97a0d4ea2ee1d38a3462efbfaef5eb87b8ea704b
Reviewed-on: https://chromium-review.googlesource.com/1116551
Reviewed-by: Michael Starzinger <mstarzinger@chromium.org>
Commit-Queue: Michael Starzinger <mstarzinger@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54132}
Adds a instance types for each HashTable subtype, rather than reusing
HASH_TABLE_TYPE. This allows us to check for these types by checking
the instance type directly, rather than by comparing against the map in
the root set.
Also, as a drive-by, do the same for SCRIPT_CONTEXT_TABLE_TYPE, and use
instance type checks rather than map checks for Context types.
This is a good general clean-up, but in particular is intended for
GetIsolate removal.
TBR=jarin@chromium.org
Bug: v8:7786
Change-Id: I420abdd12906dfa20c922e486ecdc657eb3c6ef0
Reviewed-on: https://chromium-review.googlesource.com/1114958
Reviewed-by: Leszek Swirski <leszeks@chromium.org>
Reviewed-by: Ulan Degenbaev <ulan@chromium.org>
Reviewed-by: Yang Guo <yangguo@chromium.org>
Commit-Queue: Leszek Swirski <leszeks@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54128}
Currently we don't check the result of AllocateUninitialized and just
continue execution, failing at the following memcpy.
Instead, check for a nullptr result from the embedder-provided allocator
and crash with OOM.
Bug: chromium:853706
Change-Id: I7a9b4e5d0ce3ef2b2cbb5fd3dc949ec2705bdfaa
Reviewed-on: https://chromium-review.googlesource.com/1113454
Reviewed-by: Ben Titzer <titzer@chromium.org>
Commit-Queue: Peter Marshall <petermarshall@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54127}
This change helps mostly to reduce register pressure. By default, we
compile LoadStackPointer into a move from the stack register into some
general purpose register and then later use that register. This might
trigger a spill to free up space, which is particularly costly in
loops.
Change-Id: I886233f890b7833f873fc24773f621add7cf0588
Reviewed-on: https://chromium-review.googlesource.com/1104351
Commit-Queue: Stephan Herhut <herhut@chromium.org>
Reviewed-by: Sigurd Schneider <sigurds@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54126}
This makes sure the reverse tag translation of direct call targets to
respective call tags is properly performed. Otherwise all direct call
end up being deserialized to call the function with index '0'. Ooops!
R=clemensh@chromium.org
TEST=mjsunit/wasm/compiled-module-serialization
BUG=chromium:857049
Change-Id: I37c1ee72b000daec87efdeed08d60a067b1a1b0c
Reviewed-on: https://chromium-review.googlesource.com/1120256
Reviewed-by: Clemens Hammacher <clemensh@chromium.org>
Commit-Queue: Michael Starzinger <mstarzinger@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54124}
Provide a more complete BigInt API.
Bug: v8:7712
Cq-Include-Trybots: luci.chromium.try:linux_chromium_rel_ng
Change-Id: Ic8562d616f3125deabdf8b52c7019b191bef0e07
Reviewed-on: https://chromium-review.googlesource.com/1101198
Commit-Queue: Yang Guo <yangguo@chromium.org>
Reviewed-by: Jakob Kummerow <jkummerow@chromium.org>
Reviewed-by: Yang Guo <yangguo@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54122}
Adds the builtin Trace and IsTraceCategoryEnabled functions
exposed via extra bindings. These are intended to use by
embedders to allow basic trace event support from JavaScript.
```js
isTraceCategoryEnabled('v8.some-category')
trace('e'.charCodeAt(0), 'v8.some-category',
'Foo', 0, { abc: 'xyz'})
```
Bug: v8:7851
Change-Id: I7bfb9bb059efdf87d92a56a0aae326650730c250
Reviewed-on: https://chromium-review.googlesource.com/1103294
Commit-Queue: Yang Guo <yangguo@chromium.org>
Reviewed-by: Yang Guo <yangguo@chromium.org>
Reviewed-by: Fadi Meawad <fmeawad@chromium.org>
Reviewed-by: Camillo Bruni <cbruni@chromium.org>
Reviewed-by: Benedikt Meurer <bmeurer@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54121}
This CL adds simd select, addHoriz, shuffle, anyTrue and all true to the
interpreter. It also gets rid of SIMD_COMPILED_AND_LOWERED_TEST and
SIMD_COMPILED_TEST macros.
R=gdeepti@chromium.org
BUG=v8:6020
Change-Id: I44abbcaddc3223a95c79ccc65ae9c6bf1a911c5d
Reviewed-on: https://chromium-review.googlesource.com/1119258
Commit-Queue: Aseem Garg <aseemgarg@chromium.org>
Reviewed-by: Deepti Gandluri <gdeepti@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54116}
Liftoff needs to emit the same code, thus move it to a place where it
can be reused.
Also, switch the _MSC_VER switch to V8_OS_WIN. It is unclear if
_MSC_VER would also be set in clang. V8_OS_WIN seems to fit better.
Drive-by: Reorder the code to save one jmp in the loop (loop inversion).
R=mstarzinger@chromium.org
Bug: v8:7908, v8:6600
Change-Id: Ia76aa4dad3f96e179a9daad8c1551b1aab499878
Reviewed-on: https://chromium-review.googlesource.com/1120174
Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
Reviewed-by: Michael Starzinger <mstarzinger@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54113}
This CL fixes the NumberDictionary fast-path in Array.p.sort, when
storing to a read-only property that was never read from.
R=jgruber@chromium.org
Bug: v8:7907
Change-Id: I2b772fb5b1619a94a7d239ba4417ecb7902a167c
Reviewed-on: https://chromium-review.googlesource.com/1119910
Commit-Queue: Simon Zünd <szuend@google.com>
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54109}
Since the wire bytes are now stored on the C++ heap, referenced
directly from the {NativeModule}, there is no need any more for the
additional copy in the {CompilationState}.
R=mstarzinger@chromium.org
Bug: v8:7868
Change-Id: I3ca25981b37bc14037a83a199354b70b4e52f8c7
Reviewed-on: https://chromium-review.googlesource.com/1118886
Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
Reviewed-by: Michael Starzinger <mstarzinger@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54107}
Just a refactoring to make clear that we are talking about function
names. Note that there are also names for locals inside functions,
which we currently don't use.
Drive-by: Remove style-guide violation by {WasmModule::names_} field.
R=mstarzinger@chromium.org
Bug: v8:7754
Change-Id: I9c47ea01893f128e1716be01032adfaf006ae28a
Reviewed-on: https://chromium-review.googlesource.com/1118271
Reviewed-by: Michael Starzinger <mstarzinger@chromium.org>
Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54105}
