Trying to optimize in such case breaks down the optimization, as we
end up with potentially non-eliminatable nodes that depend on the dead
IfTrue/IfFalse node.
Drive-by: Clean up dead nodes with {Kill()}.
Bug: v8:11510, chromium:1255354
Change-Id: Ia89fe6c243974c3c2abac6ad80bd4677a935f637
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3200073
Reviewed-by: Nico Hartmann <nicohartmann@chromium.org>
Commit-Queue: Manos Koukoutos <manoskouk@chromium.org>
Cr-Commit-Position: refs/heads/main@{#77211}
The checks for assignemnts to member during prefinalizers assumed the
slot has to live. It was assumed that if a slot is dead then we would
not be updating it.
Prefinalizers are allowed to touch dead objects and thus are techincally
allowed to write to dead slots. Such writes are usually redundant (the
object will be swept soon anyway) but are not always easy to get rid of.
Bug: chromium:1255152, v8:11749
Change-Id: I57e143abd53d434c3198616909c506eb70d8944b
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3199800
Reviewed-by: Michael Lippautz <mlippautz@chromium.org>
Commit-Queue: Omer Katz <omerkatz@chromium.org>
Cr-Commit-Position: refs/heads/main@{#77208}
When scanning for capture groups, we have to consider the case that the
current state is inside a character class. In that case skip everything
until the end of the current character class. Otherwise we would wrongly
count open brackets inside the character class as start of a capture
group.
Bug: chromium:1254704
Change-Id: I91d2177c464f7e507413d96216fe570253f17676
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3199871
Commit-Queue: Jakob Gruber <jgruber@chromium.org>
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Reviewed-by: Mathias Bynens <mathias@chromium.org>
Cr-Commit-Position: refs/heads/main@{#77204}
If processing the marking worklists found new ephemeron pairs, but
processing the existing ephemeron pairs didn't mark new objects, marking
would stop and the newly discovered ephemeron pairs would not be
processed. This can lead to a marked key with an unmarked value.
Bug: chromium:1252878
Change-Id: I0f158f6f64490f1f06961520b4ba57fa204bd867
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3199872
Commit-Queue: Omer Katz <omerkatz@chromium.org>
Reviewed-by: Michael Lippautz <mlippautz@chromium.org>
Cr-Commit-Position: refs/heads/main@{#77197}
When we generate identical signatures in the fuzzer,
we generate one function for each of the copies.
However, when these functions are added to WasmModulBuilder,
all will be assigned the same signature index.
Therefore, when ref.func tries to find a function corresponding
to a signature index, it will fail, despite a matching signature
existing in the module.
This CL fixes this issue by looking up functions by signature
over signature index.
Bug: v8:11954, chromium:1254387
Change-Id: Iac8d5444d4914d993da63d0630ca4d95e671630c
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3197711
Reviewed-by: Manos Koukoutos <manoskouk@chromium.org>
Reviewed-by: Thibaud Michaud <thibaudm@chromium.org>
Commit-Queue: Maria Tîmbur <mtimbur@google.com>
Cr-Commit-Position: refs/heads/main@{#77187}
The logic to locate the correct function to set a breakpoint in based
on script position was treating SharedFunctionInfo::EndPosition() as
inclusive rather than exclusive. There are various assumptions all over
the Debugger that seem to demand this treatment for the toplevel script.
But it's definitely wrong for function literals.
Fixed: chromium:1253277
Change-Id: I3421703673f4d78aee28e923e03e2fca24bc06ac
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3197715
Commit-Queue: Benedikt Meurer <bmeurer@chromium.org>
Commit-Queue: Kim-Anh Tran <kimanh@chromium.org>
Auto-Submit: Benedikt Meurer <bmeurer@chromium.org>
Reviewed-by: Kim-Anh Tran <kimanh@chromium.org>
Cr-Commit-Position: refs/heads/main@{#77186}
Previously we'd treat %_AsyncFunctionReject (and %AsyncFunctionReject)
as side-effect free (in async functions), but that's not correct, since
promise rejections have side-effects (at the very least triggering the
unhandled promise rejection machinery in the browser).
This required a minor refactoring as previously we'd classify functions
as side-effecting or not depending on whether they contain any calls to
side-effecting intrinsics, no matter whether this call is actually
executed or not. That would break REPL mode however if we'd generally
treat all async functions with %_AsyncFunctionReject intrinsic calls as
side-effecting, so instead of performing the intrinsic checks ahead of
time, we now perform the test at execution time.
Before: https://imgur.com/5BvJP9d.png
After: https://imgur.com/10FanNr.png
Fixed: chromium:1249275
Change-Id: Ib06f945ba21f1e06ee9b13a1363fad342464fd9a
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3197712
Commit-Queue: Benedikt Meurer <bmeurer@chromium.org>
Commit-Queue: Simon Zünd <szuend@chromium.org>
Auto-Submit: Benedikt Meurer <bmeurer@chromium.org>
Reviewed-by: Simon Zünd <szuend@chromium.org>
Cr-Commit-Position: refs/heads/main@{#77183}
We implement two optimizations for trap conditionals for patterns that
come up in wasm-gc.
In case of a Merge followed by a trap, where the path conditions of all
branches of the Merge contain the trap condition, we lift the trap into
the branches of the Merge.
In case of a Branch whose IfTrue branch is followed by a TrapIf with the
same condition, we replace it with the trap followed by the IfFalse
branch. Symmetrically for IfFalse and TrapUnless.
Bug: v8:7748
Change-Id: I43040aebe60eab7b2230fc3130e3b8250e8b2f45
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3190109
Reviewed-by: Nico Hartmann <nicohartmann@chromium.org>
Commit-Queue: Manos Koukoutos <manoskouk@chromium.org>
Cr-Commit-Position: refs/heads/main@{#77181}
Currently, it is possible to declare macros, builtins, etc., without
specifying a return type, in which case the return type is treated as
void. This is confusing; the code is more clear if we require the return
type to be specified.
Aside from src/torque, this change is almost entirely just adding
`: void` until the compiler is happy. However, two intrinsics in
src/builtins/torque-internal.tq have been corrected to declare an
appropriate return type. Those two intrinsics were only used in code
generated within the compiler after the type-checking phase, so we never
noticed that their return types were declared incorrectly.
Bug: v8:7793
Change-Id: Ib7df88678c25393a9e3eba389a6a1c4d9233dcbb
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3176502
Commit-Queue: Seth Brenith <seth.brenith@microsoft.com>
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Reviewed-by: Nico Hartmann <nicohartmann@chromium.org>
Cr-Commit-Position: refs/heads/main@{#77178}
Loop exits are only used during loop unrolling and are then removed, as
they cannot be handled by later optimization stages. Since unrolling
comes before inlining in the compilation pipeline, we should not emit
loop exits in inlined functions.
Bug: v8:12166
Change-Id: I28b3ebaf67c9e15b127eeb1a63906c4ecfd77480
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3195871
Reviewed-by: Clemens Backes <clemensb@chromium.org>
Commit-Queue: Manos Koukoutos <manoskouk@chromium.org>
Cr-Commit-Position: refs/heads/main@{#77175}
We currently could produce the error message 'not enough arguments on
the stack for block, expected 0 more'. This CL fixes this by printing
the available number of arguments and the needed number, and adds
DCHECKs to catch similar miscomputations in the future.
It also adds a new test that produced the broken error before, and
includes the expected failure message in a few more tests for
robustness.
R=manoskouk@chromium.org
Change-Id: Ia08863889ae36ae0a05d96d36e92295b7159a01e
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3194264
Reviewed-by: Manos Koukoutos <manoskouk@chromium.org>
Commit-Queue: Clemens Backes <clemensb@chromium.org>
Cr-Commit-Position: refs/heads/main@{#77167}
The test allocates a lot of wasm memories. This got a low slower after
https://crrev.com/c/3190476, because we can now allocate more than 102
memories, and do not explicitly trigger a GC any more to get rid of
unused memories.
We should figure out how to tell the GC about the external memory such
that the memories get collected earlier.
R=ahaas@chromium.org
Bug: v8:12076, v8:12278
Change-Id: I9b8795a9999a806380d86f22e751de2727942648
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3196131
Reviewed-by: Jakob Kummerow <jkummerow@chromium.org>
Commit-Queue: Clemens Backes <clemensb@chromium.org>
Cr-Commit-Position: refs/heads/main@{#77164}
The address space limit puts an arbitrary cap on the total reservation
size, thus limiting the total number of Wasm memories to around 100 on
64-bit systems.
Since the usable address space on 64 bit is much larger than the
1TB+4GB limit, this makes us reject code that we could otherwise just
execute.
This CL thus removes that limit completely.
See the linked issue for more discussion, including security
considerations.
R=jkummerow@chromium.org, rsesek@chromium.org
Bug: v8:12076
Change-Id: I1f61511d68efdab1f8cef4e09c0a39fc1d6fed60
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3190476
Reviewed-by: Jakob Kummerow <jkummerow@chromium.org>
Commit-Queue: Clemens Backes <clemensb@chromium.org>
Cr-Commit-Position: refs/heads/main@{#77161}
It's confusing that we have CSA_CHECK and CSA_ASSERT and it's not
clear from the names that the former works in release mode and the
latter only in debug mode.
Renaming CSA_ASSERT to CSA_DCHECK makes it clear what it does. So now
we have CSA_CHECK and CSA_DCHECK and they're not confusing.
This also renames assert() in Torque to dcheck().
Bug: v8:12244
Change-Id: I6f25d431ebc6eec7ebe326b6b8ad3a0ac5e9a108
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3190104
Reviewed-by: Nico Hartmann <nicohartmann@chromium.org>
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Reviewed-by: Igor Sheludko <ishell@chromium.org>
Reviewed-by: Shu-yu Guo <syg@chromium.org>
Commit-Queue: Marja Hölttä <marja@chromium.org>
Cr-Commit-Position: refs/heads/main@{#77160}
The V8 Inspector was sending an additional frame as part of async stack
traces for async functions, which pointed to the first executed `await`
in the async function. This is leaking an implementation detail of how
(and more precisely when) the inspector decides to collect this stack
trace. From the users perspective the async part of the stack trace is
supposed to capture what happened _prior to the task_ - meaning in case
of async functions: What lead to the execution of the async function.
This is reflected by the fact that the DevTools front-end (and the V8
Inspector itself) performs post-processing on these async call stacks,
removing the misleading top frame from it. But this post-processing is
not applied consistently to all async stack traces (i.e. the Console
message stack traces don't get this), and potentially also not applied
consistently across consumers of the Chromium debugger backend.
Instead the V8 Inspector now removes the top frame itself and thus
reports `await` consistently with how other async tasks are reported to
debugger front-ends.
Note: This preserves backwards compatibility with old versions of
devtools-frontend, which do post-processing (for the Call Stack) only on
async stack traces marked with "async function", while we now mark these
async stack traces with "await" instead (aligned with what the front-end
is using as user visibile string anyways in the Call Stack section, and
this matching will be updated in a separate follow up CL to look for
"await" instead of "async function").
Before: https://imgur.com/kIrWcIc.png
After: https://imgur.com/HvZGqiP
Fixed: chromium:1254259
Bug: chromium:1229662
Change-Id: I57ce051a28892177b6b96221f083ae957f967e52
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3193535
Auto-Submit: Benedikt Meurer <bmeurer@chromium.org>
Commit-Queue: Benedikt Meurer <bmeurer@chromium.org>
Reviewed-by: Kim-Anh Tran <kimanh@chromium.org>
Cr-Commit-Position: refs/heads/main@{#77157}
Due to MIPS64 ISA feature, 32-bit values should be sign-extended
in 64-bit registers, no matter it's signed or unsigned.
Besides, LoongArch64 also has this feature, and a similar change
has been made before loong64 port's land in V8. This CL also make
a small fix for loong64.
Change-Id: Ib284662931082365f727925af61781e3653debc8
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3193595
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Commit-Queue: Liu yu <liuyu@loongson.cn>
Cr-Commit-Position: refs/heads/main@{#77154}
This fixes a long-standing TODO to disallow importing receivers that
have "toString" or "valueOf" patched. Calling those methods could have
observable side effects, so allowing that would require bigger
refactorings to ensure that we only call each such function exactly once
per import, and in the right order.
Since this use case is rare, we just forbid importing such receivers.
R=jkummerow@chromium.org
Bug: chromium:1248677
Change-Id: I99bbd7db950ec3c7ac9cc1f59e8c476688e7d7b6
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3190475
Reviewed-by: Jakob Kummerow <jkummerow@chromium.org>
Commit-Queue: Clemens Backes <clemensb@chromium.org>
Cr-Commit-Position: refs/heads/main@{#77149}
.. and refactor js-regexp.h.
- Hide the generic DataAt/SetDataAt accessors and replace them by
dedicated accessors. Use the common lower_case naming scheme for
these.
- Shuffle around definitions in js-regexp.h s.t. they are in a
meaningful order.
- Dedupe the source/flags accessors - these fields are stored both
on the instance and on the data array. We keep only accessors for
the instance. Previously, these were disambiguated through naming
oddities (e.g. Pattern() returned data->source).
Change-Id: I3d53c8b095f0d59621ff779608438f7fa5e8c92a
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3193534
Auto-Submit: Jakob Gruber <jgruber@chromium.org>
Commit-Queue: Camillo Bruni <cbruni@chromium.org>
Reviewed-by: Maya Lekova <mslekova@chromium.org>
Reviewed-by: Camillo Bruni <cbruni@chromium.org>
Reviewed-by: Patrick Thier <pthier@chromium.org>
Cr-Commit-Position: refs/heads/main@{#77138}
This is a reland of 7d849870ff
Original change's description:
> [regexp] Reorganize and deduplicate in the regexp parser
>
> The parser is organized in a somewhat tricky way s.t. it can be
> hard to map the implementation back to the specified grammar.
>
> In particular, the logic for CharacterClassEscape, ClassEscape,
> and CharacterEscape was implemented twice - once inside a character
> class, once outside.
>
> This CL refactors related logic to have only a single implementation.
>
> As a drive-by, fix one related inconsistency related to \k inside
> a character class.
>
> Fixed: v8:10602
> Change-Id: I5858840159694fa6f8d1aa857027db80754e3dfd
> Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3178966
> Reviewed-by: Mathias Bynens <mathias@chromium.org>
> Commit-Queue: Jakob Gruber <jgruber@chromium.org>
> Cr-Commit-Position: refs/heads/main@{#77114}
Fixed: v8:10602,chromium:1253976
Change-Id: I9e7cc6a34d3be06e1a68895775aa50b0eee78c57
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3193531
Reviewed-by: Mathias Bynens <mathias@chromium.org>
Commit-Queue: Mathias Bynens <mathias@chromium.org>
Auto-Submit: Jakob Gruber <jgruber@chromium.org>
Cr-Commit-Position: refs/heads/main@{#77135}
Because there can be fewer function signatures than there are
functions a new ForceAddSignature function was added which does
not deduplicate.
Bug: v8:11954
Change-Id: Ib028de47962dd1e797ead1b1f05a4e6815029487
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3186431
Commit-Queue: Maria Tîmbur <mtimbur@google.com>
Reviewed-by: Thibaud Michaud <thibaudm@chromium.org>
Reviewed-by: Manos Koukoutos <manoskouk@chromium.org>
Cr-Commit-Position: refs/heads/main@{#77126}
This reverts commit 7d849870ff.
Reason for revert: Will block roll. Broke error message tests upstream:
https://ci.chromium.org/ui/p/v8/builders/ci/V8%20Blink%20Win/6635/overview
Original change's description:
> [regexp] Reorganize and deduplicate in the regexp parser
>
> The parser is organized in a somewhat tricky way s.t. it can be
> hard to map the implementation back to the specified grammar.
>
> In particular, the logic for CharacterClassEscape, ClassEscape,
> and CharacterEscape was implemented twice - once inside a character
> class, once outside.
>
> This CL refactors related logic to have only a single implementation.
>
> As a drive-by, fix one related inconsistency related to \k inside
> a character class.
>
> Fixed: v8:10602
> Change-Id: I5858840159694fa6f8d1aa857027db80754e3dfd
> Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3178966
> Reviewed-by: Mathias Bynens <mathias@chromium.org>
> Commit-Queue: Jakob Gruber <jgruber@chromium.org>
> Cr-Commit-Position: refs/heads/main@{#77114}
Change-Id: Ic7404d6c9f0e6ea51e8cd8f1ab672856dca0c637
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3190692
Auto-Submit: Shu-yu Guo <syg@chromium.org>
Commit-Queue: Rubber Stamper <rubber-stamper@appspot.gserviceaccount.com>
Bot-Commit: Rubber Stamper <rubber-stamper@appspot.gserviceaccount.com>
Cr-Commit-Position: refs/heads/main@{#77125}
When handling the 'end' opcode, there was a confusion with the value of
"unreachable". This DCHECK assumes the value of "unreachable" before
closing the scope, but was placed after the value is potentially updated
to reflect the state after the 'end' opcode.
R=clemensb@chromium.org
Bug: chromium:1251845
Change-Id: Iea526e6485fa705f6ff4077f54a53204cc70a926
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3190102
Reviewed-by: Clemens Backes <clemensb@chromium.org>
Commit-Queue: Thibaud Michaud <thibaudm@chromium.org>
Cr-Commit-Position: refs/heads/main@{#77116}
The parser is organized in a somewhat tricky way s.t. it can be
hard to map the implementation back to the specified grammar.
In particular, the logic for CharacterClassEscape, ClassEscape,
and CharacterEscape was implemented twice - once inside a character
class, once outside.
This CL refactors related logic to have only a single implementation.
As a drive-by, fix one related inconsistency related to \k inside
a character class.
Fixed: v8:10602
Change-Id: I5858840159694fa6f8d1aa857027db80754e3dfd
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3178966
Reviewed-by: Mathias Bynens <mathias@chromium.org>
Commit-Queue: Jakob Gruber <jgruber@chromium.org>
Cr-Commit-Position: refs/heads/main@{#77114}
The intent of the RISC-V ISA is that 32-bit C values are stored sign extended in registers, even for unsigned types.
So we skip cctest case RunLoadStoreZeroExtend64/RunUnalignedLoadStoreZeroExtend64 due to sign extend uint32
Change-Id: Icfe727916b1c04aad5681902ec4782cc98906964
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3184560
Reviewed-by: Ji Qiu <qiuji@iscas.ac.cn>
Reviewed-by: Michael Achenbach <machenbach@chromium.org>
Commit-Queue: Ji Qiu <qiuji@iscas.ac.cn>
Cr-Commit-Position: refs/heads/main@{#77112}
Several of the cctest/test-bytecode-generator/* tests used to rely
on the __COUNTER__ C++ compiler macro, which could cause spurious
test failures due to mismatched expectations when usages of that
macro were added or removed elsewhere in the code. This patch switches
to a custom, controllable static counter variable instead, to make
the tests' behaviors self-contained.
Bug: v8:12244
Change-Id: I4d875b6095e4ee4ef91933ab8c60c89054d1894a
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3186836
Reviewed-by: Camillo Bruni <cbruni@chromium.org>
Commit-Queue: Jakob Kummerow <jkummerow@chromium.org>
Cr-Commit-Position: refs/heads/main@{#77111}
These tests test allocation failure and time out on TSAN.
Bug: v8:11852
Change-Id: I81e0af3ab8130c05488c6a117dd1c6ef12aba33b
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3188071
Auto-Submit: Shu-yu Guo <syg@chromium.org>
Commit-Queue: Adam Klein <adamk@chromium.org>
Reviewed-by: Adam Klein <adamk@chromium.org>
Cr-Commit-Position: refs/heads/main@{#77104}
After the parameter processing, the arraybuffer may have been detached.
TypedArray copyWithin/fill should throw in that condition. TypedArray
includes should return false if the search element is not undefined.
Change-Id: If507d0efa1dafbe3dcefcd368e5ea27406bb3df8
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3144315
Reviewed-by: Shu-yu Guo <syg@chromium.org>
Commit-Queue: Shu-yu Guo <syg@chromium.org>
Cr-Commit-Position: refs/heads/main@{#77103}
Prepare the Torque compiler to generate Kythe artifacts to be consumed
by CodeSearch.
Drive-by changes.
* Extend SourcePosition by an offset in the input string, as this is
required by the Kythe graph.
* Correctly set missing identifier positions in Declarations.
Bug: v8:12261
Change-Id: Ida0a4a562c99f58ab924ddde36f3146f3d3fd415
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3181102
Commit-Queue: Nico Hartmann <nicohartmann@chromium.org>
Reviewed-by: Seth Brenith <seth.brenith@microsoft.com>
Cr-Commit-Position: refs/heads/main@{#77099}
Don't really have a good name for this inner TryCatch.
Bug: v8:12244
Change-Id: I4c1353bcd98e78e49d0354e285249302a8ee0dd8
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3183522
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Commit-Queue: Zhi An Ng <zhin@chromium.org>
Cr-Commit-Position: refs/heads/main@{#77097}
When cloning objects using spread and update properties (e.g.
obj = {...o, x: 0}), we wrongly used the setter for the update argument
if one was set.
This CL changes the behaviour such that all arguments following the
spread are treated as dynamic arguments.
Bug: chromium:1251366
Change-Id: I76a6d02606dca0faa0a256f465834d85d3df4f6f
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3178969
Commit-Queue: Patrick Thier <pthier@chromium.org>
Reviewed-by: Shu-yu Guo <syg@chromium.org>
Reviewed-by: Igor Sheludko <ishell@chromium.org>
Reviewed-by: Marja Hölttä <marja@chromium.org>
Cr-Commit-Position: refs/heads/main@{#77079}
