There doesn't seem to be any reason to use our custom hashmap here,
which has a more complicated interface.
Change-Id: Ib08c2e400a3cb402a5984b925034aac29750c2ec
Reviewed-on: https://chromium-review.googlesource.com/1019445
Reviewed-by: Yang Guo <yangguo@chromium.org>
Commit-Queue: Peter Marshall <petermarshall@chromium.org>
Cr-Commit-Position: refs/heads/master@{#52724}
When attributing ticks to a function, we first check if the current pc
matches a section within the dynamic library. If we find a match here,
then we don't continue looking within dynamically generated range
information, e.g. for JS functions and builtins.
This logic breaks when embedded builtins come into play. They live
within the libv8.so shared library, and are found when looking up
statics. But what we really want is to look up the dynamically
generated code-range, which contains more precise information.
In this CL, this case is detected by matching the found symbol name.
If it's the embedded blob, then we continue to dynamic lookup.
Bug: v8:6666
Change-Id: I7cea2cd4898f5a08381a071bdbc2f862b9c80880
Reviewed-on: https://chromium-review.googlesource.com/1023422
Commit-Queue: Jakob Gruber <jgruber@chromium.org>
Reviewed-by: Jaroslav Sevcik <jarin@chromium.org>
Cr-Commit-Position: refs/heads/master@{#52723}
This CL fixes the new {CodeReference} class to comply to the style
guide. It makes it a proper class, renames private fields to end in an
underscore and simplifies the union declaration.
R=ahaas@chromium.orgCC=herhut@chromium.org
Bug: v8:7570
Change-Id: I329bbc6fca1ba3c0cb34fb4e1179eb4fa9044e76
Reviewed-on: https://chromium-review.googlesource.com/1023414
Reviewed-by: Andreas Haas <ahaas@chromium.org>
Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
Cr-Commit-Position: refs/heads/master@{#52722}
For on-heap code, there is the nice feature of highlighting the current
pc on the "jco" gdb macro (calling {Code->Print}, {Code->Disassemble}
or {WasmCode::Disassemble}). For wasm code, this feature was missing so
far. This CL adds it.
R=ahaas@chromium.org
Change-Id: I0ee86d3c5cf9f42581f03c2ba4ec16b4c992e016
Reviewed-on: https://chromium-review.googlesource.com/1021517
Reviewed-by: Andreas Haas <ahaas@chromium.org>
Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
Cr-Commit-Position: refs/heads/master@{#52721}
For binary operations and unary operations, we were pinning too many
registers, thereby disallowing to reuse a src register for the result.
R=ahaas@chromium.org
Bug: v8:6600
Change-Id: Id4eea1e35d29c170e191249ef42d11ed1d284490
Reviewed-on: https://chromium-review.googlesource.com/1021818
Reviewed-by: Andreas Haas <ahaas@chromium.org>
Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
Cr-Commit-Position: refs/heads/master@{#52720}
Several functions on Array.prototype incorrectly threw a TypeError just
because their receiver was sealed or frozen.
Bug: v8:7677
Change-Id: I4ec38bfbf468f9bd676f1c0b341c8a50cf814f15
Reviewed-on: https://chromium-review.googlesource.com/1021870
Commit-Queue: Georg Neis <neis@chromium.org>
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Cr-Commit-Position: refs/heads/master@{#52718}
With certain ICU data bundles (such as the Node.js "small-icu"),
%GetDefaultICULocale() may return a more specific language tag (e.g.
"en-US") than what's available (e.g. "en"). In those cases, consider the
more specific language tag supported.
This CL also resolves the following Node.js issue:
https://github.com/nodejs/node/issues/15223
Bug: v8:7024
Cq-Include-Trybots: luci.v8.try:v8_linux_noi18n_rel_ng
Change-Id: Ifda0776b3418734d5caa8af4e50c17cda95add73
Reviewed-on: https://chromium-review.googlesource.com/668350
Commit-Queue: Daniel Ehrenberg <littledan@chromium.org>
Reviewed-by: Daniel Ehrenberg <littledan@chromium.org>
Cr-Commit-Position: refs/heads/master@{#52716}
This CL introduces an stack abstraction for writing frame descriptions.
Previously, we managed the output offset by hand, which is verbose and
error prone. This CL introduced FrameWriter, which offers a stack
abstraction with push operations for constructing the output frame
descriptions.
The abstraction is only applied to DoComputeBuiltinContinuation;
following CLs will apply it to the other DoCompute* functions.
Bug: v8:7679
Change-Id: Ia6e34de1ed63ba9245e2a08945b1e0548562ed43
Reviewed-on: https://chromium-review.googlesource.com/1019143
Reviewed-by: Jaroslav Sevcik <jarin@chromium.org>
Commit-Queue: Sigurd Schneider <sigurds@chromium.org>
Cr-Commit-Position: refs/heads/master@{#52714}
The function allocating objects in the test can be inlined in the middle
of the run. All allocations after inlining are currently not accounted.
This patch sums up allocations of the function and its outer function.
The difference between counts is now about 4%-6% (down from 15%).
Bug: chromium:834832
Change-Id: Iad071bd5bf53bb3527c9cb24d0a9ea38618c833c
Reviewed-on: https://chromium-review.googlesource.com/1021734
Reviewed-by: Ali Ijaz Sheikh <ofrobots@google.com>
Commit-Queue: Ulan Degenbaev <ulan@chromium.org>
Cr-Commit-Position: refs/heads/master@{#52713}
This CL adds StoreFixedTypedArrayElementFromTagged to the CSA. This
method takes an array element in the form of a Number/BigInt, and
converts and stores it appropriately.
For BigInts, an existing method was refactored slightly to allow its
usage.
The added method is used in the upcoming Torque implementation
of TypedArray.p.sort.
R=jgruber@chromium.org
Bug: v8:7382
Change-Id: I5135de0eff96eb2048aaca73de327a027c1faef8
Reviewed-on: https://chromium-review.googlesource.com/1021083
Commit-Queue: Simon Zünd <szuend@google.com>
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Reviewed-by: Georg Neis <neis@chromium.org>
Cr-Commit-Position: refs/heads/master@{#52711}
This CL adds a test for each typed element kind where the array to sort
consists of some max/min/zero elements.
When providing a custom compare function, the upcoming torque version
of TypedArray.p.sort needs to convert array elements to Number/BigInt
and back. The tests check the edge cases for that conversion.
R=jgruber@chromium.org
Bug: v8:7382
Change-Id: Ia85ca343f62ece208acdeb1595e94f17ce12b713
Reviewed-on: https://chromium-review.googlesource.com/1021080
Commit-Queue: Simon Zünd <szuend@google.com>
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Cr-Commit-Position: refs/heads/master@{#52710}
The existing signature is problematic for two reasons:
1. The void* -> V cast is invalid if sizeof(V) < sizeof(void*)
2. It's impossible to distinguish between a returned value of 0 and
nullptr, designating failure.
Bug: v8:6666
Change-Id: I71e8fc9119256c24a15b5bb73438f024f1af4f88
Reviewed-on: https://chromium-review.googlesource.com/1018466
Reviewed-by: Peter Marshall <petermarshall@chromium.org>
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Commit-Queue: Jakob Gruber <jgruber@chromium.org>
Cr-Commit-Position: refs/heads/master@{#52708}
Embed a hash of the embedded blob at serialization-time, and verify it
when loading the blob.
Drive-by: Keep the size of the blob trimmed, and only page-align for
dynamic page allocation and freeing during mksnapshot.
Bug: v8:6666
Change-Id: I10b6064f4f2847d4cc400f6f1bd6961194985b8f
Reviewed-on: https://chromium-review.googlesource.com/1019144
Commit-Queue: Jakob Gruber <jgruber@chromium.org>
Reviewed-by: Peter Marshall <petermarshall@chromium.org>
Cr-Commit-Position: refs/heads/master@{#52707}
This implement float to i64/u64 conversions on ia32 and x64.
These conversions emit a C call on ia32, and are implemented using
native instructions on x64.
R=ahaas@chromium.org
Bug: v8:6600
Change-Id: I5b97a74d336e196598b29d407a3d06405b74ee14
Reviewed-on: https://chromium-review.googlesource.com/1014114
Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
Reviewed-by: Andreas Haas <ahaas@chromium.org>
Cr-Commit-Position: refs/heads/master@{#52705}
To implement builtin continuations called from the deoptimizer,
we need to better support writing builtins declared with javascript
binding in Torque. This CL adds fixed number of argument support.
So you can declare in Torque, something like:
builtin javascript Foo(context: Context, receiver: Object, bar: Object):
Object {
...
}
Formerly, this would give you an error because we only supported
javascript bindings with a varargs array.
Bug: v8:7672
Change-Id: I5b5b25bdbbd5e054049c39dd2f1a4c606472dcd5
Reviewed-on: https://chromium-review.googlesource.com/1018941
Commit-Queue: Michael Stanton <mvstanton@chromium.org>
Reviewed-by: Daniel Clifford <danno@chromium.org>
Cr-Commit-Position: refs/heads/master@{#52704}
The convert operation only write an output value if the conversion
succeeded. Thus, we always initialize the buffer before calling the
conversion function.
R=ahaas@chromium.org
Change-Id: Ide230a1e608205f9067349db08adde6a90b31d6f
Reviewed-on: https://chromium-review.googlesource.com/1021377
Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
Reviewed-by: Andreas Haas <ahaas@chromium.org>
Cr-Commit-Position: refs/heads/master@{#52703}
Instead of passing multiple pointers to input and output, or to two
input values, just pass one pointer which holds all inputs and where
the output is written.
This also reduces the size of generated Turbofan graphs, since only one
stack slot is needed and less arguments are passed to the call.
It also fixes undefined behaviour, since we were passing a pointer e.g.
as {uint64_t*}, but accessed it using {ReadUnalignedValue}. Now we pass
an Address, which does not have any alignment constraints.
R=ahaas@chromium.org
Bug: v8:3770, v8:6600
Change-Id: I54ef80b7e27f77587a9062560c0b3e01d6593e6d
Reviewed-on: https://chromium-review.googlesource.com/1019147
Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
Reviewed-by: Andreas Haas <ahaas@chromium.org>
Cr-Commit-Position: refs/heads/master@{#52702}
This makes testing with random gc stress use the status file entrees
for gc fuzzing as well, since many test cases not suitable for one, are
not suitable for the other.
This also skips two more tests that rely on assert(Un)Optimized, which
is unreliable with gc fuzzing.
TBR=sigurds@chromium.org
Change-Id: I33a617b251d5cf65cf6e486d07ec55cde050b8ab
Reviewed-on: https://chromium-review.googlesource.com/1021082
Commit-Queue: Michael Achenbach <machenbach@chromium.org>
Reviewed-by: Michael Achenbach <machenbach@chromium.org>
Cr-Commit-Position: refs/heads/master@{#52701}
This is relevent for when the only "git" in your PATH is git.bat (from
depot_tools). I'd guess this is pretty common for Googlers.
Bug: v8:5960
Change-Id: I35bc49c6054afed20481ed408cfd02b7a4c346c8
Reviewed-on: https://chromium-review.googlesource.com/1019340
Commit-Queue: agrieve <agrieve@chromium.org>
Reviewed-by: Michael Achenbach <machenbach@chromium.org>
Cr-Commit-Position: refs/heads/master@{#52698}
The memory metric samples memory usage immediately after forcing GC via
LowMemoryNotification. This makes the metric sensitive to the unmapper
tasks timing.
This patch forces eager unmapping in CollectAllAvailableGarbage.
It also forces eager unmapping of non-regular chunks at the beginning
of Mark-Compact to avoid accumulation of non-regular chunks.
Bug: chromium:833291, chromium:826384
Change-Id: Iddf02cd4ab8613385d033899d29525fe6ee47fdd
Reviewed-on: https://chromium-review.googlesource.com/1017102
Commit-Queue: Ulan Degenbaev <ulan@chromium.org>
Reviewed-by: Michael Lippautz <mlippautz@chromium.org>
Cr-Commit-Position: refs/heads/master@{#52696}
The heuristic is no longer needed now that we have concurrent marking.
Bug: chromium:834371
Change-Id: I8ca3eaacdab618f690d8007aff66713260ace19f
Reviewed-on: https://chromium-review.googlesource.com/1017123
Reviewed-by: Michael Lippautz <mlippautz@chromium.org>
Commit-Queue: Ulan Degenbaev <ulan@chromium.org>
Cr-Commit-Position: refs/heads/master@{#52693}
These functions are now unused within V8, as we need tighter control in
mksnapshot for creating embedded.cc.
Embedders should switch to using SnapshotCreator directly.
Bug: v8:6666
Cq-Include-Trybots: luci.chromium.try:linux_chromium_rel_ng
Change-Id: Id231b3d1fdf50e06c9278f13d095186364264b86
Reviewed-on: https://chromium-review.googlesource.com/1019442
Reviewed-by: Yang Guo <yangguo@chromium.org>
Commit-Queue: Jakob Gruber <jgruber@chromium.org>
Cr-Commit-Position: refs/heads/master@{#52692}
The handling of wasm breakpoints was prone to forget previously set
breakpoints when inserting new ones. In particular, adding breakpoints
in reverse order or adding more than 4 breakpoints would fail.
Change-Id: I94f314e86bdf9b53a4170ce1b6b47339b7cb7848
Reviewed-on: https://chromium-review.googlesource.com/1019302
Reviewed-by: Clemens Hammacher <clemensh@chromium.org>
Commit-Queue: Stephan Herhut <herhut@chromium.org>
Cr-Commit-Position: refs/heads/master@{#52691}
ScavengeExternalString and ScavengeExternalOneByteString clear GC flags,
which confuses concurrent sweeping that was started from another test.
Tests that mutate flags should not be threaded.
Bug: v8:7671
Change-Id: I08656d06fe85ff45baca685ebe5982528aad774c
Reviewed-on: https://chromium-review.googlesource.com/1019102
Reviewed-by: Peter Marshall <petermarshall@chromium.org>
Commit-Queue: Ulan Degenbaev <ulan@chromium.org>
Cr-Commit-Position: refs/heads/master@{#52690}
This problem was discovered by mvstanton@'s forEach Torque patch. In the absense
of test coverage for Torque, his patch will serve as the regression test case
for this bug when it lands.
Change-Id: Ic77446a8e46168928da221e6eb18753dd6478c87
Reviewed-on: https://chromium-review.googlesource.com/1018763
Reviewed-by: Michael Stanton <mvstanton@chromium.org>
Commit-Queue: Daniel Clifford <danno@chromium.org>
Cr-Commit-Position: refs/heads/master@{#52688}
Module names in torque are allowed to have underscores but not dashes.
To stay consistent with C++ file naming conventions, the underscores
in module names are replaced by dashes for file names.
Example:
module typed_array {} would now generate:
builtins-typed-array-from-dsl-gen.(cc|h)
instead of:
builtins-typed_array-from-dsl-gen.(cc|h)
R=danno@chromium.org
Change-Id: Iff42d7b9b5f65c378ee30f9d884ab6a3a3cd42a7
Reviewed-on: https://chromium-review.googlesource.com/1016460
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Commit-Queue: Daniel Clifford <danno@chromium.org>
Cr-Commit-Position: refs/heads/master@{#52686}
'ref' builds might not yet have BigInt support. Instead of crashing
when the benchmark file is loaded, the benchmark now fails during
setup.
R=jgruber@chromium.org
Bug: chromium:834654
Change-Id: I2c03aeba4b6af947061504c369b5228d7b344307
Reviewed-on: https://chromium-review.googlesource.com/1019100
Commit-Queue: Simon Zünd <szuend@google.com>
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Cr-Commit-Position: refs/heads/master@{#52685}
This refactoring is a big step towards separating Turbofan-related code
from backend independent code. This will allow us to include way less
headers from "src/compiler" at various places.
The {WasmCompilationUnit} contained information for Turbofan
compilation, and for Liftoff compilation. This CL tears this apart, such
that {WasmCompilationUnit} holds backend-independent information, plus
a pointer to either {LiftoffCompilationUnit} or
{TurbofanWasmCompilationUnit}. These pointers are opaque, so that
{function-compiler.h}, defining {WasmCompilationUnit}, does not need to
include any Turbofan specific or Liftoff specific headers.
R=ahaas@chromium.org, titzer@chromium.org, mstarzinger@chromium.org
Bug: v8:7570, v8:6600
Change-Id: I024c9a23508ee1b4b3cbe1d068c8e785d732daca
Reviewed-on: https://chromium-review.googlesource.com/1016640
Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
Reviewed-by: Andreas Haas <ahaas@chromium.org>
Cr-Commit-Position: refs/heads/master@{#52684}
Since movf and movt instructions were removed in r6, sel_s and
sel_d instructions need to be used instead with some preparations.
Change-Id: Ia6a2fda7d3d79ada1ae1ec4649793efd2466f79b
Reviewed-on: https://chromium-review.googlesource.com/1016910
Reviewed-by: Ivica Bogosavljevic <ivica.bogosavljevic@mips.com>
Commit-Queue: Sreten Kovacevic <sreten.kovacevic@mips.com>
Cr-Commit-Position: refs/heads/master@{#52683}
This also adds a javascript operator JSCreateObject and an
associated TFS stub that handles Object.create in cases
where only a prototype, but no additional properties are
provided.
Bug: v8:7250
Change-Id: Ib1fd529a10a553c3718222356319bd6ccffbdf30
Reviewed-on: https://chromium-review.googlesource.com/1013576
Commit-Queue: Sigurd Schneider <sigurds@chromium.org>
Reviewed-by: Benedikt Meurer <bmeurer@chromium.org>
Cr-Commit-Position: refs/heads/master@{#52680}
This reverts commit f52b4b3b81.
Reason for revert: Makes benchmarks time out:
https://crbug.com/834655
Original change's description:
> [compiler] Temporarily turn some DCHECKs into CHECKs.
>
> This turns most DCHECKs in the register allocator code into CHECKs,
> which hopefully will tell us more about the crashes in crbug 831822.
>
> This CL will eventually be reverted again.
>
> Bug: chromium:831822
> Change-Id: I123ed507949ecab50bb0aaaf9e91978c9c2d8d65
> Reviewed-on: https://chromium-review.googlesource.com/1016762
> Commit-Queue: Georg Neis <neis@chromium.org>
> Reviewed-by: Sigurd Schneider <sigurds@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#52673}
TBR=neis@chromium.org,sigurds@chromium.org
Change-Id: Ic2b37a3ce04516c5871b801015153ce84e622e90
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Bug: chromium:831822
Reviewed-on: https://chromium-review.googlesource.com/1018860
Reviewed-by: Michael Achenbach <machenbach@chromium.org>
Commit-Queue: Michael Achenbach <machenbach@chromium.org>
Cr-Commit-Position: refs/heads/master@{#52679}
Since external strings are used for things like source strings, we
should tenure them from creation.
Change-Id: I226ab9036836d76d8c17ed168ad97d7f0f824278
Reviewed-on: https://chromium-review.googlesource.com/1006961
Commit-Queue: Leszek Swirski <leszeks@chromium.org>
Reviewed-by: Hannes Payer <hpayer@chromium.org>
Reviewed-by: Yang Guo <yangguo@chromium.org>
Reviewed-by: Michael Lippautz <mlippautz@chromium.org>
Cr-Commit-Position: refs/heads/master@{#52677}
The Cvtqui2ss and Cvtqui2sd methods did overwrite the {src} register and
the given {tmp} register.
This CL fixes this to avoid the overwrite of the {src} register, and to
use the kScratchRegister instead of an explicitly passed tmp register.
It also adds u32 -> f32/f64 conversions.
All these methods can take either a register or an operand as src. This
sometimes saves a mov operation.
R=neis@chromium.org, ahaas@chromium.org
Change-Id: I912a2a3b9d9c356f04ce51869a01c6fd11c76dd3
Reviewed-on: https://chromium-review.googlesource.com/1017121
Reviewed-by: Andreas Haas <ahaas@chromium.org>
Reviewed-by: Georg Neis <neis@chromium.org>
Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
Cr-Commit-Position: refs/heads/master@{#52676}
