We had an optimization in Crankshaft where we would call into the
megamorphic handler stub directly if an inline cache was already
found to be megamorphic when it hit the optimizing compiler. This
way we could avoid the dispatch overhead when we know that there's
no point in checking for the other states anyways. However we somehow
missed to port this optimization to TurboFan.
Now this change introduces support to call into LoadIC_Megamorphic and
KeyedLoadIC_Megamorphic directly (plus the trampoline versions), which
saves quite a lot of overhead for the cases where the map/name pair is
found in the megamorphic stub cache, and it's quite a simple change. We
can later extend this to also handle the StoreIC and KeyedStoreIC cases
if that turns out to be beneficial.
This improves the score on the Octane/TypeScript test by around ~2%
and the TypeScript test in the web-tooling-benchmark by around ~4%. On
the ARES-6 Air test the steady state mean improves by 2-4%, and on the
ARES-6 ML test the steady state mean seems to also improve by 1-2%, but
that might be within noise.
On a micro-benchmark that just runs `o.x` in a hot loop on a set of 9
different objects, which all have `x` as the first property and are
all in fast mode, we improve by around ~30%, and are now almost on par
with JavaScriptCore.
Bug: v8:6344, v8:6936
Change-Id: Iaa4c6e34c37e78da217ee75f32f6acc95a834250
Reviewed-on: https://chromium-review.googlesource.com/1215623
Reviewed-by: Jaroslav Sevcik <jarin@chromium.org>
Reviewed-by: Jakob Kummerow <jkummerow@chromium.org>
Commit-Queue: Benedikt Meurer <bmeurer@chromium.org>
Cr-Commit-Position: refs/heads/master@{#55803}
This reverts commit ece86adc6b.
Reason for revert: Potential cause of auto-roller breakage https://ci.chromium.org/p/chromium/builders/luci.chromium.try/win10_chromium_x64_rel_ng/91864
Original change's description:
> [typedarray] Properly convert hole to undefined in TypedArray.from
>
> It used to call the old IterableToList, which had the wrong
> semantics for holes.
>
> Bug: v8:8133
> Change-Id: Idd5acd55a155bc43df7552135a44151bb2db38e9
> Reviewed-on: https://chromium-review.googlesource.com/1213204
> Reviewed-by: Peter Marshall <petermarshall@chromium.org>
> Commit-Queue: Georg Neis <neis@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#55745}
TBR=neis@chromium.org,petermarshall@chromium.org,dhai@google.com
# Not skipping CQ checks because original CL landed > 1 day ago.
Bug: v8:8133
Change-Id: I09b108e7844c598253fbbe02d705699c21308637
Reviewed-on: https://chromium-review.googlesource.com/1220286
Reviewed-by: Deepti Gandluri <gdeepti@chromium.org>
Commit-Queue: Deepti Gandluri <gdeepti@chromium.org>
Cr-Commit-Position: refs/heads/master@{#55802}
The fallback for {Pinsrd} and {Pextrd} for the non-avx and non-sse
variant clobbered the {xmm0} register. This CL fixes this by storing
the values on the stack and modifying them there instead.
The alternative would have been to pass in a scratch register. But this
path is not commonly used and we cannot express in the API whether the
scratch register is needed or not. So we would sometimes have to spill a
register to pass it as scratch register even though it is then unused.
R=titzer@chromium.org
CC=mstarzinger@chromium.org
Bug: v8:6600
Change-Id: Ieae53b892cc55eed4fcfa3d0e7f82f3e1afe72be
Reviewed-on: https://chromium-review.googlesource.com/1219633
Reviewed-by: Michael Starzinger <mstarzinger@chromium.org>
Reviewed-by: Ben Titzer <titzer@chromium.org>
Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
Cr-Commit-Position: refs/heads/master@{#55800}
In the CHECK_FLOAT_EQ and CHECK_DOUBLE_EQ wrappers, do also print hex
representations on failure. Otherwise, single bit flips might not be
visible in the output, like here:
Check failed: DoubleWrapper(x) == y (-2e+66 vs. -2e+66).
R=titzer@chromium.org
Change-Id: I2521706aedc6ff81c0dbb25259230f8e29ce9a3e
Reviewed-on: https://chromium-review.googlesource.com/1219630
Reviewed-by: Ben Titzer <titzer@chromium.org>
Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
Cr-Commit-Position: refs/heads/master@{#55799}
- Update those counters from space
- Add fast path for move memory
Bug: chromium:845409
Change-Id: Icd72e551df2422a635801fb4e31588073f81544e
Reviewed-on: https://chromium-review.googlesource.com/1219707
Reviewed-by: Ulan Degenbaev <ulan@chromium.org>
Commit-Queue: Michael Lippautz <mlippautz@chromium.org>
Cr-Commit-Position: refs/heads/master@{#55798}
This CL adds a contextual variable that tracks lint errors and will
cause the Torque compiler to abort if there were any.
R=tebbi@chromium.org
Bug: v8:7793
Change-Id: Ia7c2b5b6c7cb12d489b12596afa84f933fc3f337
Reviewed-on: https://chromium-review.googlesource.com/1219388
Reviewed-by: Tobias Tebbi <tebbi@chromium.org>
Commit-Queue: Simon Zünd <szuend@google.com>
Cr-Commit-Position: refs/heads/master@{#55797}
This improves the error message if we accidentally call this method on
a CPU which does not support it.
R=mstarzinger@chromium.org
Bug: v8:6600
Change-Id: If67bf9e95ab0e71d832a833a70b6ba748b53296e
Reviewed-on: https://chromium-review.googlesource.com/1219631
Reviewed-by: Michael Starzinger <mstarzinger@chromium.org>
Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
Cr-Commit-Position: refs/heads/master@{#55796}
To make the changes in base.tq work, there were 2 changes needed on
the C++ side:
- calls to "FromConstexpr" are generated by the compiler for
implicit conversions.
- type switch is desugared and uses "Cast"
R=jgruber@chromium.org, tebbi@chromium.org
Change-Id: I085f1a393f93e501e6bbcaeacb0d6568259a4714
Reviewed-on: https://chromium-review.googlesource.com/1219629
Commit-Queue: Simon Zünd <szuend@google.com>
Reviewed-by: Tobias Tebbi <tebbi@chromium.org>
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Cr-Commit-Position: refs/heads/master@{#55794}
This checks the type signature during import against the expected
signature. For this the {WasmExceptionObject} now contains a serialized
version of the signature.
R=clemensh@chromium.org
TEST=mjsunit/wasm/exceptions-import
BUG=v8:8091
Change-Id: I5a34ef87eccf4d2ed3a784620796ec009623fd90
Reviewed-on: https://chromium-review.googlesource.com/1219509
Reviewed-by: Clemens Hammacher <clemensh@chromium.org>
Commit-Queue: Michael Starzinger <mstarzinger@chromium.org>
Cr-Commit-Position: refs/heads/master@{#55789}
There was a bug in WebAssembly.instantiate in the case where a CSP
disallows WebAssembly compilation. In this case the promise returned by
WebAssembly.instantiate was rejected immediately because of the CSP,
but then compilation was started anyways, and the promise was resolved
after compilation for a second time, which caused the crash. With this
CL we do not start compilation if CSP disallows WebAssembly compilation.
R=clemensh@chromium.org
Bug: chromium:881978
Change-Id: Iffdb3e02c3006eb7f86211ab197f81cf20438f0e
Reviewed-on: https://chromium-review.googlesource.com/1219706
Commit-Queue: Andreas Haas <ahaas@chromium.org>
Reviewed-by: Clemens Hammacher <clemensh@chromium.org>
Cr-Commit-Position: refs/heads/master@{#55788}
Use v8_executable instead of executable for
bytecode_builtins_list_generator so that coverage link flags should be
picked up too.
Bug: v8:8068
Change-Id: I260e4de3485a78d3b3a3455441837e1a5936aa25
Reviewed-on: https://chromium-review.googlesource.com/1219326
Reviewed-by: Michael Starzinger <mstarzinger@chromium.org>
Commit-Queue: Dan Elphick <delphick@chromium.org>
Cr-Commit-Position: refs/heads/master@{#55787}
This enables cctest, unittests, fuzzer and inspector on Android.
The cctest suite requires extra resource-fetching logic for the
bytecode-generator expectation files.
Bug: chromium:866862
Cq-Include-Trybots: luci.v8.try:v8_android_arm64_n5x_rel_ng
Change-Id: If3da853a62c047388476a7f38e32e64e2859f186
Reviewed-on: https://chromium-review.googlesource.com/1213208
Commit-Queue: Michael Achenbach <machenbach@chromium.org>
Reviewed-by: Ross McIlroy <rmcilroy@chromium.org>
Reviewed-by: Sergiy Byelozyorov <sergiyb@chromium.org>
Cr-Commit-Position: refs/heads/master@{#55784}
This reverts commit 6afe7d1815.
Reason for revert: Failures (-2e+66 vs. -2e+66): https://ci.chromium.org/p/v8/builders/luci.v8.ci/V8%20Linux%20-%20debug/22148
Original change's description:
> [Liftoff] Implement f32.copysign and f64.copysign
>
> These are two of the few missing instructions. This CL implements them
> for ia32 and x64, and bails out on other platforms.
> On x64, we are using the BTR instruction since we cannot have 64-bit
> immediates.
>
> Drive-by: Fix naming of existing bt/bts instructions on x64.
>
> R=titzer@chromium.org
>
> Bug: v8:6600
> Change-Id: Ib8532ca811160cd61f4ba7c06b04ce093861c872
> Reviewed-on: https://chromium-review.googlesource.com/1174383
> Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
> Reviewed-by: Ben Titzer <titzer@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#55780}
TBR=titzer@chromium.org,clemensh@chromium.org
Change-Id: I4377c13346b42b65e8db04cbd15fc2f906113f65
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Bug: v8:6600
Reviewed-on: https://chromium-review.googlesource.com/1219446
Reviewed-by: Clemens Hammacher <clemensh@chromium.org>
Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
Cr-Commit-Position: refs/heads/master@{#55783}
These are two of the few missing instructions. This CL implements them
for ia32 and x64, and bails out on other platforms.
On x64, we are using the BTR instruction since we cannot have 64-bit
immediates.
Drive-by: Fix naming of existing bt/bts instructions on x64.
R=titzer@chromium.org
Bug: v8:6600
Change-Id: Ib8532ca811160cd61f4ba7c06b04ce093861c872
Reviewed-on: https://chromium-review.googlesource.com/1174383
Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
Reviewed-by: Ben Titzer <titzer@chromium.org>
Cr-Commit-Position: refs/heads/master@{#55780}
The macro has been deprecated since 2016, and it keeps confusing me, so
let's just remove it completely from the code base.
R=leszeks@chromium.orgTBR=mstarzinger@chromium.org, verwaest@chromium.org, jgruber@chromium.org
Bug: v8:8015
Change-Id: Ibe1122fd9d2624bc94873d9c51dc8499c54a04fd
Reviewed-on: https://chromium-review.googlesource.com/1209322
Reviewed-by: Clemens Hammacher <clemensh@chromium.org>
Reviewed-by: Michael Starzinger <mstarzinger@chromium.org>
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Reviewed-by: Leszek Swirski <leszeks@chromium.org>
Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
Cr-Commit-Position: refs/heads/master@{#55779}
This adds the ability to import exception into a module at instantiation
time. Only a {WasmExceptionObject} that has been exported by another
module instance can be imported, all other values are rejected.
Note that currently there is no signature check being performed to make
sure the imported exception matches the expected type. Also the identity
of imported exceptions is not yet preserved.
Furthermore the engine does not yet match thrown exception objects on a
global level across modules. Hence imported exceptions will (wrongly)
behave as completely new types within the module.
R=clemensh@chromium.org
TEST=mjsunit/wasm/exceptions-import,unittests/WasmModuleVerifyTest
BUG=v8:8091
Change-Id: If247762b949a1ba4a87d13bc3e790a45dbc67815
Reviewed-on: https://chromium-review.googlesource.com/1216402
Commit-Queue: Michael Starzinger <mstarzinger@chromium.org>
Reviewed-by: Clemens Hammacher <clemensh@chromium.org>
Cr-Commit-Position: refs/heads/master@{#55777}
This patch removes unnecessary copies and adds reserves to vectors that
are filled in a loop afterwards. Fixing clang-tidy warning.
Bug: v8:8015
Change-Id: I4e13c0445a9760e09ef03a62ae48be622ebecc6b
Reviewed-on: https://chromium-review.googlesource.com/1209783
Reviewed-by: Michael Starzinger <mstarzinger@chromium.org>
Commit-Queue: Florian Sattler <sattlerf@google.com>
Cr-Commit-Position: refs/heads/master@{#55776}
Manually roll ICU to 7ca3ffa to pick up an upstream fix for quarter handling.
Remove the hack that prevent unexpected behavior in ICU
and remove the skip of failing tests.
Bug: v8:8151,v8:7869
Cq-Include-Trybots: luci.v8.try:v8_linux_noi18n_rel_ng
Change-Id: Ic9a56679bf1eb9dc18c739161838d518fd664d6f
Reviewed-on: https://chromium-review.googlesource.com/1214522
Reviewed-by: Jungshik Shin <jshin@chromium.org>
Reviewed-by: Sathya Gunasekaran <gsathya@chromium.org>
Commit-Queue: Frank Tang <ftang@chromium.org>
Cr-Commit-Position: refs/heads/master@{#55774}
Moves some functions that used to be simple, but now do more complex logic
out of the inlined headers into .cc to avoid having to export all the
functions they depend on as V8_PRIVATE_EXPORT for tests. Also mark them
as V8_PRIVATE_EXPORT for tests along with a couple in ast.h
BUG=v8:8041
Change-Id: I6e94ca160cd3d84a0d1b099167a7b9862f08b122
Reviewed-on: https://chromium-review.googlesource.com/1216642
Reviewed-by: Leszek Swirski <leszeks@chromium.org>
Reviewed-by: Adam Klein <adamk@chromium.org>
Commit-Queue: Ross McIlroy <rmcilroy@chromium.org>
Cr-Commit-Position: refs/heads/master@{#55772}
This inserts unreachable node after uncoditional deopt in bit-to-word
conversion and wires it as an input to the dead-value.
This fixes a problem, where a floating dead-value was inserted by a change
of bit-to-word (which always fails because bit cannot be converted to word).
Without the unrachable node (which this CL inserts in the effect chain after
deopt), the dead value was scheduled before the uncoditional deoptimization
and crash at runtime.
Unfortunately, I do not know how to construct a test that does not end up in
an infinite loop.
Bug: chromium:878805
Change-Id: Ia03060949f6a9b914807f5614fadcf2271911998
Reviewed-on: https://chromium-review.googlesource.com/1196663
Reviewed-by: Tobias Tebbi <tebbi@chromium.org>
Commit-Queue: Jaroslav Sevcik <jarin@chromium.org>
Cr-Commit-Position: refs/heads/master@{#55770}
This CL adds a new "Call" stub that can be used by builtins that will
call the same JS call-back function often (e.g. compare function in
Array.p.sort). The checks have to be done upfront once, but can then
be omitted.
R=jgruber@chromium.org
Bug: v8:7861
Change-Id: Id6e4ca27c3d488a7b1f708cbcb4cbe6cc382513e
Reviewed-on: https://chromium-review.googlesource.com/1208574
Commit-Queue: Simon Zünd <szuend@google.com>
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Reviewed-by: Camillo Bruni <cbruni@chromium.org>
Cr-Commit-Position: refs/heads/master@{#55769}
The ICU API for maximizing and minimizing a locale ID takes ICU format locale id as an input
so that a BCP 47 language tag must be converted to the corresponding ICU locale id.
Bug: v8:7982
Cq-Include-Trybots: luci.v8.try:v8_linux_noi18n_rel_ng
Change-Id: I1cb1dacbf057bbc8bb8beb9b62d1ec4becd82624
Reviewed-on: https://chromium-review.googlesource.com/1150934
Reviewed-by: Sathya Gunasekaran <gsathya@chromium.org>
Reviewed-by: Jungshik Shin <jshin@chromium.org>
Commit-Queue: Frank Tang <ftang@chromium.org>
Cr-Commit-Position: refs/heads/master@{#55768}
In many cases, the condition already explains the assertion and a string
message clutters the code without adding value. This wrapper macro is
actually very useful in such cases.
Change-Id: I44f27c1c9255f5fc80453c92565f07f8f3cbacae
Reviewed-on: https://chromium-review.googlesource.com/1216183
Reviewed-by: Michael Starzinger <mstarzinger@chromium.org>
Reviewed-by: Clemens Hammacher <clemensh@chromium.org>
Commit-Queue: Jakob Gruber <jgruber@chromium.org>
Cr-Commit-Position: refs/heads/master@{#55767}
The motivation behind this change is that the WasmArgumentsAdaptor
builtin needed too many registers on ia32 and was blocking work on
adding a root register.
Bug: v8:6666
Change-Id: I9d39af96e8520acc092f5ef4b7f0a159fe4e3b12
Reviewed-on: https://chromium-review.googlesource.com/1209788
Commit-Queue: Jakob Gruber <jgruber@chromium.org>
Reviewed-by: Michael Starzinger <mstarzinger@chromium.org>
Cr-Commit-Position: refs/heads/master@{#55761}
To support all possible cases, we must limit the number of register
args for TFS builtins on ia32 to 3. Out of the 6 allocatable
registers, esi is taken as the context register and ebx is the root
register. One register must remain available to store the jump/call
target. Thus 3 registers remain for arguments.
The reason this applies to TFS builtins specifically is because this
becomes relevant for builtins used as targets of Torque function
pointers (which must have a register available to store the target).
Bug: v8:6666
Change-Id: I17d9450cc29c983ddaffc2deb36f45c1c414e166
Reviewed-on: https://chromium-review.googlesource.com/1209287
Commit-Queue: Jakob Gruber <jgruber@chromium.org>
Reviewed-by: Igor Sheludko <ishell@chromium.org>
Reviewed-by: Jaroslav Sevcik <jarin@chromium.org>
Reviewed-by: Michael Starzinger <mstarzinger@chromium.org>
Cr-Commit-Position: refs/heads/master@{#55759}
Previously the builtins table had a value for every single
OperandScale/Bytecode combination regardless of whether it was valid.
This change makes it so that only valid bytecode handlers are stored in
the builtins table. This prevents placeholders being serialized into the
snapshot (and embedded into the binary) saving 9KB in
CODE_SPACE/OLD_SPACE and 2.5KB in the embedded data as well as 66
entries in the builtins table.
To do this, it generates a new header file bytecodes-builtins-list.h
which is created from the BYTECODE_LIST and OPERAND_SCALE_LIST macros.
Since list macros cannot be used to conditionally generate elements in
the C-preprocessor, this is done by generator executable, compiled from
interpreter/generate-flat-headers.cc.
Additionally the generator creates the flat bytecode list so that it is
transposed from the previous result, i.e. the results are grouped by
bytecode and then operand scale rather than operand scale then bytecode.
This should give better locality for commonly used bytecodes and may
allow less commonly used ExtraWide bytecodes to never be mapped into
memory at all.
The cost to storing the handlers densely is that looking up a handler
now requires a binary search through the builtins table, but this should
only happen during debugging. It is also fixable at least for non-wide
handlers and could be improved for wide ones if the need arises.
Bug: v8:8068
Cq-Include-Trybots: luci.chromium.try:linux_chromium_rel_ng
Change-Id: Iaad22a952e2858f508030c5ddc082f91bf59f667
Reviewed-on: https://chromium-review.googlesource.com/1209304
Commit-Queue: Dan Elphick <delphick@chromium.org>
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Reviewed-by: Ross McIlroy <rmcilroy@chromium.org>
Cr-Commit-Position: refs/heads/master@{#55757}
- Provide MapData::SerializeDescriptors method for serializing the whole
descriptor array.
- Trigger this in JSObjectData::SerializeAsBoilerplate.
- Further make things more consistent across the broker.
Bug: v8:7790
Change-Id: Ie6499da8857f7c6561f7c44922aeffcea4876be7
Reviewed-on: https://chromium-review.googlesource.com/1199102
Commit-Queue: Georg Neis <neis@chromium.org>
Reviewed-by: Jaroslav Sevcik <jarin@chromium.org>
Reviewed-by: Maya Lekova <mslekova@chromium.org>
Cr-Commit-Position: refs/heads/master@{#55756}
- Provide getters for all member variables.
- Add missing flags for cycle detection.
- Be more consistent in stylistic matters.
Bug: v8:7790
Change-Id: I87c3901b203acc4254ec48f5cc895b2e68ff32aa
Reviewed-on: https://chromium-review.googlesource.com/1196431
Commit-Queue: Georg Neis <neis@chromium.org>
Reviewed-by: Jaroslav Sevcik <jarin@chromium.org>
Reviewed-by: Maya Lekova <mslekova@chromium.org>
Cr-Commit-Position: refs/heads/master@{#55755}
Issues that caused failure of this test have been resolved with commit
https://chromium-review.googlesource.com/c/v8/v8/+/1213183, so it can
be re-enabled.
Change-Id: I441998e5e63fce7a7e718b593c8e58a71841b78e
Reviewed-on: https://chromium-review.googlesource.com/1215168
Reviewed-by: Clemens Hammacher <clemensh@chromium.org>
Commit-Queue: Sreten Kovacevic <skovacevic@wavecomp.com>
Cr-Commit-Position: refs/heads/master@{#55754}
