This fixes the collection of function sizes (see comment #1 of
v8:6361). This was done by adding a wrapper function around
VerifyWasmCode() that updates UMA stats for function size (as well as
decode time when validating).
Bug: v8:6361
Change-Id: Ia2597db22cbed5b0429d9c8514e96cdea5d7323a
Reviewed-on: https://chromium-review.googlesource.com/600530
Commit-Queue: Karl Schimpf <kschimpf@chromium.org>
Reviewed-by: Bill Budge <bbudge@chromium.org>
Cr-Commit-Position: refs/heads/master@{#47147}
Collect compile time separately for asm.js and Wasm functions.
Note: Dependent on CL https://chromium-review.googlesource.com/c/598542
Bug: v8:6361
Change-Id: I77c43b663cb66261f17bca73db03ab1a4af5e11d
Reviewed-on: https://chromium-review.googlesource.com/598930
Commit-Queue: Karl Schimpf <kschimpf@chromium.org>
Reviewed-by: Bill Budge <bbudge@chromium.org>
Cr-Commit-Position: refs/heads/master@{#47146}
On advancing the iterator we need to reset the current object,
so that it can be lazily reloaded later on.
TBR=mlippautz@chromium.org
Bug: chromium:694255
Change-Id: If7ddd8670df9d11837f491503312919b55b451fe
Reviewed-on: https://chromium-review.googlesource.com/600687
Commit-Queue: Ulan Degenbaev <ulan@chromium.org>
Reviewed-by: Ulan Degenbaev <ulan@chromium.org>
Cr-Commit-Position: refs/heads/master@{#47142}
Make sure to fail PrepareTransfer when the transferables array contains
a non-ArrayBuffer, otherwise the function leaks a scheduled_exception.
Bug: chromium:736565
Change-Id: I64c2e09eb92720519c7bda2dca41749ff5ac9c8d
Reviewed-on: https://chromium-review.googlesource.com/599357
Commit-Queue: Ben Smith <binji@chromium.org>
Reviewed-by: Yang Guo <yangguo@chromium.org>
Cr-Commit-Position: refs/heads/master@{#47141}
We need to modify that module, e.g. to add signatures to the signature
maps. Hence it cannot be const.
R=titzer@chromium.orgCC=mtrofin@chromium.org
Change-Id: I261af5b4233a0b2ec8031a9cbe0cf9f826316ae0
Reviewed-on: https://chromium-review.googlesource.com/600428
Reviewed-by: Ben Titzer <titzer@chromium.org>
Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
Cr-Commit-Position: refs/heads/master@{#47138}
Change-Id: I42f51fd00b1a21a1ff7ef2379fbf81a42266a737
Reviewed-on: https://chromium-review.googlesource.com/599111
Reviewed-by: Georg Neis <neis@chromium.org>
Commit-Queue: Adam Klein <adamk@chromium.org>
Cr-Commit-Position: refs/heads/master@{#47137}
This fixes collection of function decoding time (See comment #20 of
v8:6361). Updated other path for decoding to also update UMA stat.
Bug: v8:6361
Change-Id: I8e6cb398aebd7847e064d3b9a979586ded118513
Reviewed-on: https://chromium-review.googlesource.com/598458
Commit-Queue: Karl Schimpf <kschimpf@chromium.org>
Reviewed-by: Bill Budge <bbudge@chromium.org>
Cr-Commit-Position: refs/heads/master@{#47136}
An optimization in li checks if a value has all their "n" LSBs set to
one, and their "64-n" MSBs set to zero. This is done by checking
base::bits::IsPowerOfTwo(value + 1), but this doesn't cover the case
when only MSB is zero, and other bits are one. Added another check that
covers the missing case, this also fixes Dsubu test.
BUG=
TEST=cctest/test-assembler-mips/Dsubu
Change-Id: I97b8fb4984af3b7f974ba901e3a1a9c89e925805
Reviewed-on: https://chromium-review.googlesource.com/598228
Reviewed-by: Ivica Bogosavljevic <ivica.bogosavljevic@imgtec.com>
Commit-Queue: Miran Karić <Miran.Karic@imgtec.com>
Cr-Commit-Position: refs/heads/master@{#47134}
This is a reland of 35c923cc10
Original change's description:
> [heap] Add support for atomic access to page flags.
>
> This patch renames AsAtomicWord to AsAtomicPointer and
> adds new AsAtomicWord that works with intptr_t.
>
> Slot recording uses atomic page flag accessors.
>
> BUG=chromium:694255
>
> Change-Id: I1c692813244b41320182e9eea50462d1802fcd98
> Reviewed-on: https://chromium-review.googlesource.com/597688
> Commit-Queue: Ulan Degenbaev <ulan@chromium.org>
> Reviewed-by: Michael Lippautz <mlippautz@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#47086}
Bug: chromium:694255
Change-Id: I36780ff4001e068815d4be1e16cd06f1a4f98d13
Reviewed-on: https://chromium-review.googlesource.com/599909
Reviewed-by: Michael Lippautz <mlippautz@chromium.org>
Commit-Queue: Ulan Degenbaev <ulan@chromium.org>
Cr-Commit-Position: refs/heads/master@{#47131}
This DCHECK is incorrect in no-snapshot builds because we can run
garbage collection while deserializing the snapshot.
Bug: v8:6559, v8:6557
Change-Id: I0b8456ee19443f796889656952b236ff40b1e2fb
Reviewed-on: https://chromium-review.googlesource.com/600367
Reviewed-by: Ulan Degenbaev <ulan@chromium.org>
Commit-Queue: Maya Lekova <mslekova@google.com>
Cr-Commit-Position: refs/heads/master@{#47130}
This mitigates the problem of blocking on the main thread when the
platform is unable to execute background tasks in a timely manner.
Bug: v8:6671
Change-Id: I741d4b7594e8d62721dad32cbfb19551ffacd0c3
Reviewed-on: https://chromium-review.googlesource.com/599528
Commit-Queue: Michael Lippautz <mlippautz@chromium.org>
Reviewed-by: Ulan Degenbaev <ulan@chromium.org>
Cr-Commit-Position: refs/heads/master@{#47126}
The predicate in question used to report true on both, Crankshaft and
TurboFan code. It has hence become obsolete and can be replaced by the
existing {Code::is_turbofanned} predicate. This also frees up a bit in
the second kind-specific bit field.
R=jarin@chromium.org
BUG=v8:6408
Change-Id: I204d7dd78a639c752c9749fd305c7006c6b6aca3
Reviewed-on: https://chromium-review.googlesource.com/599868
Reviewed-by: Jaroslav Sevcik <jarin@chromium.org>
Commit-Queue: Michael Starzinger <mstarzinger@chromium.org>
Cr-Commit-Position: refs/heads/master@{#47125}
For determining whether unaligned memory accessed is allowed, it is
actually sufficient to pass the MachineRepresentation instead of the
MachineType.
As the MachineRepresentation is an enum, this allows to store the set
of unsupported types as an EnumSet instead of std::vector.
R=mstarzinger@chromium.org
Change-Id: Ibe44717bae3a6034bae683057e80bd40d92e70ae
Reviewed-on: https://chromium-review.googlesource.com/598790
Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
Reviewed-by: Michael Starzinger <mstarzinger@chromium.org>
Cr-Commit-Position: refs/heads/master@{#47123}
Bug: 749486
The feature is off by default, and could be turned on via
`v8_enable_csa_write_barrier = true`. With this CL, only x64 uses this
feature
Change-Id: Ie024f08b7d796a4cc4d55285dc9fe796780f0e53
Reviewed-on: https://chromium-review.googlesource.com/588891
Reviewed-by: Yang Guo <yangguo@chromium.org>
Reviewed-by: Tobias Tebbi <tebbi@chromium.org>
Reviewed-by: Ulan Degenbaev <ulan@chromium.org>
Commit-Queue: Albert Mingkun Yang <albertnetymk@google.com>
Cr-Commit-Position: refs/heads/master@{#47122}
Currently, EnumSet cannot be used with scoped enums (declared as "enum
class" or "enum struct"), as these cannot be implicitly casted to their
underlying integral type. This CL changes this by adding a static cast.
Plus drive-by fixes.
R=mstarzinger@chromium.org
Change-Id: I3aa2ef78e896b2734f1e5695237efc8a1130703c
Reviewed-on: https://chromium-review.googlesource.com/598789
Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
Reviewed-by: Michael Starzinger <mstarzinger@chromium.org>
Cr-Commit-Position: refs/heads/master@{#47120}
Don't hold a pointer to parse_info in compilation_info, and instead explicitly
add the fields needed in compiation_info. The intention is to make ParseInfo
only actually needed for parsing, and eventually make it possible to compile
with only a CompileInfo.
BUG=v8:5203
Change-Id: Iecd39245e44c218874401c3991eeaf3ceef2816f
Reviewed-on: https://chromium-review.googlesource.com/595738
Reviewed-by: Michael Starzinger <mstarzinger@chromium.org>
Reviewed-by: Andreas Haas <ahaas@chromium.org>
Commit-Queue: Ross McIlroy <rmcilroy@chromium.org>
Cr-Commit-Position: refs/heads/master@{#47119}
The {alignment} parameter in {UnalignedLoadSupported} and
{UnalignedStoreSupported} is never being used.
This CL removes it from several interfaces and users.
R=mstarzinger@chromium.org
Change-Id: Iad5a638b7ed358beeed42f0ea7308dc04f1ed23e
Reviewed-on: https://chromium-review.googlesource.com/598788
Reviewed-by: Michael Starzinger <mstarzinger@chromium.org>
Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
Cr-Commit-Position: refs/heads/master@{#47115}
Since frames.h no longer defines basic register types like RegList, it
is no longer necessary to include it in the macro assemblers.
Next step: split out frame-constants.h from frames.h so that it will be
possible to get frame constants without include the stackwalking logic,
which needs objects.h.
R=mstarzinger@chromium.org
Bug:
Change-Id: Ia12d3c8a8d46a73106c3c90bcb4b470c85f1eaa7
Reviewed-on: https://chromium-review.googlesource.com/597788
Commit-Queue: Ben Titzer <titzer@chromium.org>
Reviewed-by: Yang Guo <yangguo@chromium.org>
Reviewed-by: Michael Starzinger <mstarzinger@chromium.org>
Cr-Commit-Position: refs/heads/master@{#47114}
The ScheduledErrorThrower is also needed in the wasm-async fuzzer so I
moved the implementation from wasm-js.cc to wasm-api.[h|cc].
R=clemensh@chromium.org
Bug: chromium:749838
Change-Id: I49d7438d1ec0281285ce0c64ba462c22001be08e
Reviewed-on: https://chromium-review.googlesource.com/591447
Commit-Queue: Andreas Haas <ahaas@chromium.org>
Reviewed-by: Clemens Hammacher <clemensh@chromium.org>
Cr-Commit-Position: refs/heads/master@{#47112}
Rename OPTIONAL_ACCESSOR to OPTIONAL_ACCESSORS (to be in line with
ACCESSORS and DECL_OPTIONAL_ACCESSORS), and make OPTIONAL_ACCESSORS also
define the ACCESSORS (to be in line with DECL_OPTIONAL_ACCESSORS).
Drive-by: Remove DECL_OPTIONAL_GETTER; just define the setter too.
R=titzer@chromium.org
Change-Id: Id41a652153d7a1e2a15a5f69c6f52b96298f9dc1
Reviewed-on: https://chromium-review.googlesource.com/598868
Reviewed-by: Ben Titzer <titzer@chromium.org>
Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
Cr-Commit-Position: refs/heads/master@{#47111}
This removes the obsolete {Crankshaft} factory method as it returns the
same configuration as the {Turbofan} factory by now. We now consistently
use {RegisterConfiguration::Default} everywhere.
R=jkummerow@chromium.org
BUG=v8:6408
Change-Id: I6be25774aa6714ef4dc1ef6856bb6dbc95593a29
Reviewed-on: https://chromium-review.googlesource.com/597858
Reviewed-by: Jakob Kummerow <jkummerow@chromium.org>
Commit-Queue: Michael Starzinger <mstarzinger@chromium.org>
Cr-Commit-Position: refs/heads/master@{#47109}
This begins splitting up the Deserializer class into
{Object,Partial,Startup}Deserializer. For now, all functionality remains in
the Deserializer base clase, to be refactored in future CLs. Empty .cc files
are added here to avoid having to touch build files again.
Bug: v8:6624
Change-Id: If563e03492991bd55c91cd2e09312c0a26aaab2c
Reviewed-on: https://chromium-review.googlesource.com/598067
Reviewed-by: Yang Guo <yangguo@chromium.org>
Commit-Queue: Jakob Gruber <jgruber@chromium.org>
Cr-Commit-Position: refs/heads/master@{#47107}
This reverts commit 062bb7d487.
Reason for revert: <INSERT REASONING HERE>
Original change's description:
> [regexp] Limit ATOM regexps to single-character patterns
>
> There's an inherent trade-off when deciding between ATOM and IRREGEXP
> regexps: IRREGEXP is faster at runtime for all but trivial single-character
> patterns, while ATOM regexps have a lower memory overhead.
>
> This CL is intended to help investigate impact on benchmarks and real-world
> code - if something tanks, it's easy to revert, otherwise it can be a first
> step towards a possible removal of ATOM regexps.
>
> Bug: v8:6633
> Change-Id: Ia41d8eb28d33952735562d3d4127202746a6ac4e
> Reviewed-on: https://chromium-review.googlesource.com/589435
> Reviewed-by: Yang Guo <yangguo@chromium.org>
> Commit-Queue: Jakob Gruber <jgruber@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#47081}
TBR=yangguo@chromium.org,jgruber@chromium.org
Change-Id: I8655bc4055af5d593f507e16918b434ff45f5379
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Bug: v8:6633
Reviewed-on: https://chromium-review.googlesource.com/599547
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Commit-Queue: Jakob Gruber <jgruber@chromium.org>
Cr-Commit-Position: refs/heads/master@{#47106}
Previously TryToName bailed out for Oddball keys (i.e. when passing true
or false as the key), which meant that for example the generic
KeyedLoadIC would always bail out to the %KeyedGetProperty runtime
function. But handling Oddball keys is fairly easy, since every oddball
value carries it's unique string representation. Adding just this case
to the CodeStubAssembler::TryToName method boosts this simple
micro-benchmark by a factor of 4x:
const n = 1e7;
const obj = {};
const key = true;
console.time('foo');
for (let i = 0; i < n; ++i) {
if (obj[key] === undefined) {
obj[key] = key;
}
}
console.timeEnd('foo');
It also shows on the ARES-6 ML benchmark and on several Speedometer
tests, where objects are being used as dictionaries and the developers
rely on the implicit ToString conversions on the property accesses.
In the ARES-6 ML benchmark, the number of calls to %KeyedGetProperty
is reduced by 137,758.
Bug: v8:6278, v8:6344, v8:6670
Change-Id: Iaa965e30be4c247682a67ec09543655df9b761d2
Reviewed-on: https://chromium-review.googlesource.com/599527
Reviewed-by: Jaroslav Sevcik <jarin@chromium.org>
Commit-Queue: Benedikt Meurer <bmeurer@chromium.org>
Cr-Commit-Position: refs/heads/master@{#47105}
This is a reland of 470a10015d
Original change's description:
> Start migration of try/throw/catch to match proposal.
>
> This CL does the first baby steps on moving the current (experimental)
> exception handling to match that of the WebAssembly proposal.
>
> It does the following:
>
> 1) Use exception tags instead of integers.
>
> 2) Only handle empty exception signatures (i.e. no values associated
> with the exception tag.
>
> 3) Only handle one catch clause.
>
> 4) Be sure to rethrow the exception if the exception tag does not match.
>
> Note: There are many things that need to be fixed, and are too
> numerous to list here. However, the code should have TODO's on each
> missing parts of the implementation.
>
> Also note that the code currently doesn't handle nested catch blocks,
> nor does it change the throw value being an integer. Rather, the
> integer value is still being thrown, and currently is the exception
> tag. Therefore, we don't build an exception object. This is the reason
> why this CL doesn't handle exceptions that pass values.
>
> Also, the current implementation still can't handle multiple modules
> because tag resolution (between) modules has not be implemented yet.
>
> Bug: v8:6577
> Change-Id: Id6d08b641b3c42d1eec7d4db582f2dab35406114
> Reviewed-on: https://chromium-review.googlesource.com/591910
> Reviewed-by: Brad Nelson <bradnelson@chromium.org>
> Commit-Queue: Karl Schimpf <kschimpf@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#47087}
Bug: v8:6577
Change-Id: I41c3309827c292cb787681a95aaef7cf9b931835
Reviewed-on: https://chromium-review.googlesource.com/598968
Reviewed-by: Michael Lippautz <mlippautz@chromium.org>
Reviewed-by: Brad Nelson <bradnelson@chromium.org>
Commit-Queue: Brad Nelson <bradnelson@chromium.org>
Cr-Commit-Position: refs/heads/master@{#47100}
