This was not happening when there was no need to typecheck the entry.
Additional changes:
- Add tests with null table entries for typed and untyped function
tables.
- Allow AddIndirectFunctionTable in wasm-run-utils to specify table
type.
- Add possibility to define tables in test-gc.cc.
- Merge trapTableOutOfBounds with trapInvalidFunc.
- Use trapTableOutOfBounds in call_indirect as appropriate.
- Fix emission of table types in wasm-module-builder.cc.
Bug: v8:9495
Change-Id: I4a857ff4378e5a87dc0646d94b4c75635a43c55b
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2442622
Reviewed-by: Tobias Tebbi <tebbi@chromium.org>
Reviewed-by: Jakob Kummerow <jkummerow@chromium.org>
Commit-Queue: Manos Koukoutos <manoskouk@chromium.org>
Cr-Commit-Position: refs/heads/master@{#70311}
The interpreter is not an execution tier in production any more. It's
only used in tests.
Thus, remove {ExecutionTier::kInterpreter} and instead add a
{TestExecutionTier} that still has {kInterpreter}.
If needed (in {TestingModuleBuilder::execution_tier()}), we translate
back from {TestExecutionTier} to {ExecutionTier} (for {kLiftoff} and
{kTurboFan} only).
The {TraceMemoryOperation} method, which is shared between interpreter
and production code, now receives a {base::Optional<ExecutionTier>}, and
we will just pass en empty optional if called from the interpreter.
R=thibaudm@chromium.org
Bug: v8:10389
Change-Id: Ibe133b91e8dca6d6edbfaee5ffa0d7fe72ed6d64
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2335186
Reviewed-by: Thibaud Michaud <thibaudm@chromium.org>
Commit-Queue: Clemens Backes <clemensb@chromium.org>
Cr-Commit-Position: refs/heads/master@{#69260}
... that controls whether the TF graph zones should support compression.
Bug: v8:9923
Change-Id: Ifbe237b75e9c92e62eb32b69d6b3b1a818269b83
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2308347
Commit-Queue: Igor Sheludko <ishell@chromium.org>
Reviewed-by: Tobias Tebbi <tebbi@chromium.org>
Reviewed-by: Ross McIlroy <rmcilroy@chromium.org>
Cr-Commit-Position: refs/heads/master@{#69036}
I was looking around and couldn't find any tests for select with v128
params. Adding it for completion.
Change-Id: I9e3770e37eaddb23fc99649579b175536c1df844
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2308093
Reviewed-by: Bill Budge <bbudge@chromium.org>
Commit-Queue: Zhi An Ng <zhin@chromium.org>
Cr-Commit-Position: refs/heads/master@{#68951}
This allows the compiler to eliminate more unneeded branches. Since all
functions just do a lookup in a static table (either directly, or via
compiling a switch to such a lookup), they are also good candidates for
inlining, which is made possible by this change.
One DCHECK is removed instead of pulling in the inl header, which would
require more refactoring since the check is in a non-inl header.
R=thibaudm@chromium.orgTBR=jkummerow@chromium.org
Bug: v8:10576
Change-Id: If0fd25fd62c5f30b896fc67a5458a5ae475a6351
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2259944
Commit-Queue: Clemens Backes <clemensb@chromium.org>
Reviewed-by: Thibaud Michaud <thibaudm@chromium.org>
Cr-Commit-Position: refs/heads/master@{#68508}
The reference types wasm proposal dropped all subtyping. Subsequently,
the 'anyref' type was renamed to externref.
This changes all references of the *type* anyref to externref.
Additionally, the flag that permits this extension is renamed to
"reftypes" to mirror the proposal name.
Bug: v8:7748
Change-Id: Icf323f13b9660fd10540e65125af053fca3a03f9
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2232941
Commit-Queue: Clemens Backes <clemensb@chromium.org>
Reviewed-by: Kim-Anh Tran <kimanh@chromium.org>
Reviewed-by: Clemens Backes <clemensb@chromium.org>
Reviewed-by: Yang Guo <yangguo@chromium.org>
Cr-Commit-Position: refs/heads/master@{#68270}
Even in unreachable code, the targets of br_table have to have matching
types.
R=thibaudm@chromium.org
Bug: v8:10556
Change-Id: I2e85df3cb92f7910a6bcb5ac03927c424194660d
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2218062
Commit-Queue: Andreas Haas <ahaas@chromium.org>
Reviewed-by: Thibaud Michaud <thibaudm@chromium.org>
Cr-Commit-Position: refs/heads/master@{#68148}
Unreachable code may consume values from an empty stack as needed, known
as stack polymorphism. After consuming the values, the stack height
should still be 0, which was incorrectly handled by multi-value blocks.
R=ahaas@chromium.org
Bug: chromium:1085507
Change-Id: Ibf5f2d05bec0fbe029cfa66ee2d07540a370934a
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2218033
Reviewed-by: Andreas Haas <ahaas@chromium.org>
Commit-Queue: Thibaud Michaud <thibaudm@chromium.org>
Cr-Commit-Position: refs/heads/master@{#68070}
This adds a flag to {WasmCode} objects to store whether this code was
generated for debugging. This flag can be set for Liftoff code (in which
case the code will e.g. have an extended prologue for debugging), but it
can also be set for TurboFan, in case Liftoff bailed out when producing
the debugging code.
Having this flag allows us to remove the hack to pass the compilation
results to {OnFinishedUnits} just to check whether we actually wanted to
compile Liftoff functions.
Drive-by: Replace the {ReachedRecompilationTierField} by a
{MissingRecompilationField}, because all we need to know is if we are
still waiting for that function to get recompiled.
R=ahaas@chromium.org
Bug: v8:10330,v8:10410
Change-Id: Ia023df8955a60d9f5595a6cb2737e14d83baf716
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2142259
Commit-Queue: Clemens Backes <clemensb@chromium.org>
Reviewed-by: Andreas Haas <ahaas@chromium.org>
Cr-Commit-Position: refs/heads/master@{#67119}
In preparation for adding reference types, which need an additional
parameter to indicate the referenced type.
Bug: v8:7748
Change-Id: If4023f3d9c7f42ed603b69c43356d2e8b81a0daa
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2091471
Commit-Queue: Jakob Kummerow <jkummerow@chromium.org>
Reviewed-by: Clemens Backes <clemensb@chromium.org>
Cr-Commit-Position: refs/heads/master@{#66687}
This is a reland of 79398ab09d
Original change's description:
> [wasm] Further reduce the size of WasmCode
>
> Also, save dynamic allocations (plus their memory overhead).
> This is realized by storing the relocation information, source position
> table, and protected instruction information together in one "metadata"
> byte array.
> For each of the three components, we just store their size, such that
> the accessors can return the respecitive {Vector} views as before.
>
> This makes each WasmCode object 24 bytes smaller on 64-bit
> architectures. It also saves a few more bytes per code object because
> less padding is needed for the individual allocations, and each dynamic
> allocation comes with some constant memory overhead.
>
> Since the protected instructions will just be stored in a byte array
> now, some APIs are refactored to just return that byte array directly
> (instead of an array of {ProtectedInstructionData}). This also
> simplifies serialization and deserialization, and will allow for
> switching to a more compact representation in the future.
>
> Drive-by: Add some more checks to {Vector::cast} to protect against
> undefined behaviour.
>
> R=ahaas@chromium.org
>
> Bug: v8:10254
> Change-Id: I81ca847023841110e3e52cc402fcb0349325d7af
> Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2078545
> Reviewed-by: Andreas Haas <ahaas@chromium.org>
> Reviewed-by: Tobias Tebbi <tebbi@chromium.org>
> Commit-Queue: Clemens Backes <clemensb@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#66596}
Tbr: ahaas@chromium.org
Bug: v8:10254
Change-Id: Idcdcb4f13c3eb7a3f7fb5ef8a1229103ca0ae975
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2089934
Reviewed-by: Clemens Backes <clemensb@chromium.org>
Reviewed-by: Jakob Kummerow <jkummerow@chromium.org>
Commit-Queue: Clemens Backes <clemensb@chromium.org>
Cr-Commit-Position: refs/heads/master@{#66598}
This reverts commit 79398ab09d.
Reason for revert: Makes UBSan unhappy: https://ci.chromium.org/p/v8/builders/ci/V8%20Linux64%20UBSan/10186
Original change's description:
> [wasm] Further reduce the size of WasmCode
>
> Also, save dynamic allocations (plus their memory overhead).
> This is realized by storing the relocation information, source position
> table, and protected instruction information together in one "metadata"
> byte array.
> For each of the three components, we just store their size, such that
> the accessors can return the respecitive {Vector} views as before.
>
> This makes each WasmCode object 24 bytes smaller on 64-bit
> architectures. It also saves a few more bytes per code object because
> less padding is needed for the individual allocations, and each dynamic
> allocation comes with some constant memory overhead.
>
> Since the protected instructions will just be stored in a byte array
> now, some APIs are refactored to just return that byte array directly
> (instead of an array of {ProtectedInstructionData}). This also
> simplifies serialization and deserialization, and will allow for
> switching to a more compact representation in the future.
>
> Drive-by: Add some more checks to {Vector::cast} to protect against
> undefined behaviour.
>
> R=ahaas@chromium.org
>
> Bug: v8:10254
> Change-Id: I81ca847023841110e3e52cc402fcb0349325d7af
> Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2078545
> Reviewed-by: Andreas Haas <ahaas@chromium.org>
> Reviewed-by: Tobias Tebbi <tebbi@chromium.org>
> Commit-Queue: Clemens Backes <clemensb@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#66596}
TBR=jkummerow@chromium.org,ahaas@chromium.org,clemensb@chromium.org,tebbi@chromium.org
Change-Id: Id80aa82cfce8942879031032b322ee66855b5600
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Bug: v8:10254
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2089933
Reviewed-by: Clemens Backes <clemensb@chromium.org>
Commit-Queue: Clemens Backes <clemensb@chromium.org>
Cr-Commit-Position: refs/heads/master@{#66597}
Also, save dynamic allocations (plus their memory overhead).
This is realized by storing the relocation information, source position
table, and protected instruction information together in one "metadata"
byte array.
For each of the three components, we just store their size, such that
the accessors can return the respecitive {Vector} views as before.
This makes each WasmCode object 24 bytes smaller on 64-bit
architectures. It also saves a few more bytes per code object because
less padding is needed for the individual allocations, and each dynamic
allocation comes with some constant memory overhead.
Since the protected instructions will just be stored in a byte array
now, some APIs are refactored to just return that byte array directly
(instead of an array of {ProtectedInstructionData}). This also
simplifies serialization and deserialization, and will allow for
switching to a more compact representation in the future.
Drive-by: Add some more checks to {Vector::cast} to protect against
undefined behaviour.
R=ahaas@chromium.org
Bug: v8:10254
Change-Id: I81ca847023841110e3e52cc402fcb0349325d7af
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2078545
Reviewed-by: Andreas Haas <ahaas@chromium.org>
Reviewed-by: Tobias Tebbi <tebbi@chromium.org>
Commit-Queue: Clemens Backes <clemensb@chromium.org>
Cr-Commit-Position: refs/heads/master@{#66596}
Most function signatures are created once and never changed. Hence pass
them as const pointer. This makes it clear in function signatures that
these parameters will not be modified.
This also avoids a few ugly const_casts where we were passing pointers
to constexpr FunctionSigs via non-const pointers.
R=jkummerow@chromium.org
Bug: v8:10155
Change-Id: Ieb658ab5582bff276f76babdaf7ddb8f72bd4790
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2072739
Reviewed-by: Jakob Kummerow <jkummerow@chromium.org>
Commit-Queue: Clemens Backes <clemensb@chromium.org>
Cr-Commit-Position: refs/heads/master@{#66478}
This removes the --experimental-wasm-sat-f2i-conversions flag. This
feature is shipped since v7.5.
R=ahaas@chromium.org
Change-Id: I354d9528be40caac77cd4e41adcd39d013448339
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1958009
Reviewed-by: Andreas Haas <ahaas@chromium.org>
Commit-Queue: Clemens Backes <clemensb@chromium.org>
Cr-Commit-Position: refs/heads/master@{#65384}
Fix a TODO from Ben to change the macro argument order to match the
actual order in wasm code.
After this fix, we can remove the individual {WASM_CALL_INDIRECT[0-5]}
macros and implement them via a common variadic macro.
Also, rename {WASM_CALL_INDIRECT_TABLE0} to {WASM_CALL_INDIRECT_TABLE}.
The name was confusing, because this macro explictly allows to set a
table index different from 0. Thus, just drop the "0" in the name.
The individual test changes were done via a vim macro, to avoid manual
errors.
R=mstarzinger@chromium.org
Bug: v8:10021
Change-Id: I9f0f31511c5c6e20a0b07524bf75fe9cf1598eba
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1940265
Reviewed-by: Michael Starzinger <mstarzinger@chromium.org>
Commit-Queue: Clemens Backes <clemensb@chromium.org>
Cr-Commit-Position: refs/heads/master@{#65242}
This brings our constants back in line with the changed spec text. We
already use kExprTableGet and kExprTableSet, but for locals and globals
we still use the old wording.
This renaming is mostly mechanical.
PS1 was created using:
ag -l 'kExpr(Get|Set|Tee)Local' src test | \
xargs -L1 sed -E 's/kExpr(Get|Set|Tee)Local\b/kExprLocal\1/g' -i
PS2 contains manual fixes.
R=mstarzinger@chromium.org
Bug: v8:9810
Change-Id: I1617f1b2a100685a3bf56218e76845a9481959c5
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1847354
Reviewed-by: Michael Starzinger <mstarzinger@chromium.org>
Commit-Queue: Clemens Backes <clemensb@chromium.org>
Cr-Commit-Position: refs/heads/master@{#64161}
This CL refactors the type-checking for br_table instructions.
Originally, we iterated over all targets of br_table and checked
if the values on the stack match the types expected by the
target's signature. However, this caused problems with type
checking unreachable br_table instructions where some stack
values are unavailable. According to the anyref proposal, the
expected type of br_table is the greatest lower bound of
all its targets. With the existing implementation, the expected
types were the types of the first target.
With this CL, we first calculate the expected types of br_table,
and only then inspect the stack if matching values are available.
R=titzer@chromium.org
Bug: v8:7581
Change-Id: I12208323bda88c363e28ffb0e002d59ef9a6b9d8
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1649791
Commit-Queue: Andreas Haas <ahaas@chromium.org>
Reviewed-by: Clemens Hammacher <clemensh@chromium.org>
Cr-Commit-Position: refs/heads/master@{#62354}
This removes two default switch cases for unsupported opcodes, and
replaces them by explicit lists. This makes it easy to see what is
currently not supported in Liftoff. In a follow-up CL, each bailout
will be associated with a category to track which features currently
cause Liftoff to bailout.
This change also makes Liftoff crash (in UNREACHABLE) if invoked with
asm.js code. Hence, change the asm.js tests to not test Liftoff. In
production, we do not invoke Liftoff for asm.js anyway.
R=mstarzinger@chromium.org
Change-Id: I971c6146ed325103d14008c0e67a973a47a35bc2
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1634909
Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
Reviewed-by: Michael Starzinger <mstarzinger@chromium.org>
Cr-Commit-Position: refs/heads/master@{#61911}
The instruction is the same as the existing {select} instruction with
type. Both inputs must be in a sub-type relationship with the type
specified in the type instruction.
R=clemensh@chromium.org
Bug: v8:7581
Change-Id: Ibead6cd0253210828c8114336ea0942e6cbd6126
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1631413
Commit-Queue: Andreas Haas <ahaas@chromium.org>
Reviewed-by: Clemens Hammacher <clemensh@chromium.org>
Cr-Commit-Position: refs/heads/master@{#61886}
This CL was generated by an automatic clang AST rewriter using this
matcher expression:
callExpr(
callee(
cxxMethodDecl(
hasName("operator->"),
ofClass(isSameOrDerivedFrom("v8::internal::Object"))
)
),
argumentCountIs(1)
)
The "->" at the expression location was then rewritten to ".".
R=jkummerow@chromium.orgTBR=mstarzinger@chromium.org,verwaest@chromium.org,yangguo@chromium.org
Bug: v8:9183, v8:3770
No-Try: true
No-Tree-Checks: true
Change-Id: I0a7ecabdeafe51d0cf427f5280af0c7cab96869e
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1624209
Reviewed-by: Clemens Hammacher <clemensh@chromium.org>
Reviewed-by: Jakob Kummerow <jkummerow@chromium.org>
Reviewed-by: Toon Verwaest <verwaest@chromium.org>
Commit-Queue: Yang Guo <yangguo@chromium.org>
Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
Cr-Commit-Position: refs/heads/master@{#61764}
Up until now, we cached export wrappers per export index. With the
anyref proposal potentially many more functions will need export
wrappers, e.g. any function that is stored in a table, and any
function accessed by the new ref.func instruction.
With this CL, we change the caching scheme an do the caching per
signature. Thereby we can guarantee that any export wrapper which
potentially exists can be stored in the cache.
For cctests which use wasm-run-utils, we don't know the size of the
cache anymore ahead of time. However, we assume that no more than
5 signatures will be used in any cctest. If this assumption is not
true, we can just adjust the number.
The cache is now accessed in all code paths where we need an export
wrapper.
Bug: chromium:962850
Change-Id: I32df60dfa7801d1e71f7d837da091f388198af1f
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1615247
Commit-Queue: Andreas Haas <ahaas@chromium.org>
Reviewed-by: Clemens Hammacher <clemensh@chromium.org>
Cr-Commit-Position: refs/heads/master@{#61752}
Our {Vector} template provides both {start} and {begin} methods. They
return exactly the same value. Since the {begin} method is needed for
iteration, and is also what standard containers provide, this CL
switches all uses of the {start} method to use {begin} instead.
Patchset 1 was auto-generated by using this clang AST matcher:
callExpr(
callee(
cxxMethodDecl(
hasName("start"),
ofClass(hasName("v8::internal::Vector")))
),
argumentCountIs(0))
Patchset 2 was created by running clang-format. Patchset 3 then
removes the now unused {Vector::start} method.
R=jkummerow@chromium.orgTBR=mstarzinger@chromium.org,yangguo@chromium.org,verwaest@chromium.org
Bug: v8:9183
Change-Id: Id9f01c92870872556e2bb3f6d5667463b0e3e5c6
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1587381
Reviewed-by: Jakob Kummerow <jkummerow@chromium.org>
Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
Cr-Commit-Position: refs/heads/master@{#61081}
The {code_table_} in {NativeModule} is protected by the
{allocation_mutex_}. The {code} and {code_table} accessors did not
acquire this lock though.
This CL removes the unsafe {code_table} accessor, renames {code} to
{GetCode} and protects it by a lock.
R=mstarzinger@chromium.org
Bug: v8:9112
Change-Id: Id2df68460b4c10291a49b4016b9574e02744e8b9
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1561315
Reviewed-by: Jakob Kummerow <jkummerow@chromium.org>
Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
Cr-Commit-Position: refs/heads/master@{#60779}
Locks for compilation state callbacks and for the native module are
again taken one after the other. As a consequence, publishing compiled
Wasm code again happens in parallel. Compile times are now comparable to
before lazy hints were enabled.
Bug: chromium:949050
Change-Id: I45c52254d046de080938bd131fd3ed8116660bef
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1552787
Commit-Queue: Frederik Gossen <frgossen@google.com>
Reviewed-by: Michael Starzinger <mstarzinger@chromium.org>
Reviewed-by: Clemens Hammacher <clemensh@chromium.org>
Cr-Commit-Position: refs/heads/master@{#60646}
When running wasm tests, the interpreter previously used a static
collection of function indexes stored in WasmTable to perform
call_indirect calls internal to that module. This has the wrong behavior
if the table is changed (via WasmTableObject::Set, `table.copy`, or
`table.init`).
This CL changes the cctests to always generate an intepreter entry for
all functions, and stores those entries in the dispatch table. This
allows us to use the same execution path as for non-testing code.
The interpreter entry compiler needed to be changed to support
multi-value returns too, since a 64-bit integer return value may be
lowered to two 32-bit integer returns.
Bug: v8:9016
Change-Id: I277df21ffde5c2eee0b691fcc9bab2b1a43eeffc
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1531137
Commit-Queue: Andreas Haas <ahaas@chromium.org>
Reviewed-by: Andreas Haas <ahaas@chromium.org>
Cr-Commit-Position: refs/heads/master@{#60380}
This prepares a refactoring to add and publish compilation results in
batches. For this, we need to separate the two phases, so that we can
lock the module, allocate all the code space, release the lock, copy
the code, lock the module, publish the code, and release the lock
again.
In particular, this CL does the following:
1) It removes the {AddOwnedCode} method. The functionality of creating
the {WasmCode} and memcpy'ing the instruction into that is done in
the other {Add*Code} methods. Adding to {owned_code_} is done in
{PublishCode}.
2) {PublishInterpreterEntry} is now functionally equivalent to
{PublishCode}, so it's removed.
3) After {AddCode}, the caller has to call {PublishCode}. In a
follow-up CL, this will be called in batches (first {AddCode} them
all, then {PublishCode} them all).
4) {AddCompiledCode} now assumes that the {WasmCompilationResult}
succeeded. Otherwise, the caller should directly call {SetError} on
the {CompilationState}.
5) {PublishCode} is now the chokepoint for installing code to the code
table, the owned code vector, the jump table, and setting interpreter
redirections. It replaces previous direct calls to {InstallCode} or
explicitly adding to {owned_code_}.
6) Increasing the {generated_code_size_} counter is now done in
{AllocateForCode}, which is the chokepoint for allocating space for
generated code. This way, we will only increase this counter once
once we allocate in batches.
R=titzer@chromium.org
Bug: v8:8916
Change-Id: I71e02e3a838f21797915cee3ebd373804fb12237
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/1530817
Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
Reviewed-by: Ben Titzer <titzer@chromium.org>
Cr-Commit-Position: refs/heads/master@{#60369}
A small refactoring to reduce complexity. It also moves the call to
{RegisterTrapHandlerData} out of the mutex to reduce the time in this
critical section.
R=mstarzinger@chromium.org
Bug: v8:8916
Change-Id: I644f03db6099ebef22b2e33b607a2dc038b36423
Reviewed-on: https://chromium-review.googlesource.com/c/1478196
Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
Reviewed-by: Michael Starzinger <mstarzinger@chromium.org>
Cr-Commit-Position: refs/heads/master@{#59907}
These offsets are now transported via CodeDesc.
Bug: v8:8758
Change-Id: If7485c62878bc81a55c9b4ca8ce3578dab2376ce
Reviewed-on: https://chromium-review.googlesource.com/c/1447717
Commit-Queue: Jakob Gruber <jgruber@chromium.org>
Reviewed-by: Michael Starzinger <mstarzinger@chromium.org>
Reviewed-by: Clemens Hammacher <clemensh@chromium.org>
Cr-Commit-Position: refs/heads/master@{#59315}
This CL changes the usage pattern from
FOR_XXX_VALUES(i) { Use(*i); }
to
FOR_XXX_VALUES(i) { Use(i); }
which is way more intuitive.
Note that the replacement in the uses was done via regular expression,
so it's purely mechanical. In two locations I removed unneeded braces
around the macro, because they confused clang-format.
I plan to do more cleanups (remove redundant assignments within the
FOR_XXX_VALUES body) in a follow-up CL.
R=mstarzinger@chromium.org
Bug: v8:8562
Change-Id: I4329bfcf34e5b077d19b50f4204ceb3b4340fe61
Reviewed-on: https://chromium-review.googlesource.com/c/1449615
Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
Reviewed-by: Michael Starzinger <mstarzinger@chromium.org>
Cr-Commit-Position: refs/heads/master@{#59287}
Numeric conversions are defined behavior iff the value is in the
range of what the target type can represent.
Bug: v8:3770
Change-Id: Ic6f2276c64cb39345a45d8e37e604c28ecca34c2
Reviewed-on: https://chromium-review.googlesource.com/c/1436216
Commit-Queue: Jakob Kummerow <jkummerow@chromium.org>
Reviewed-by: Yang Guo <yangguo@chromium.org>
Reviewed-by: Georg Neis <neis@chromium.org>
Cr-Commit-Position: refs/heads/master@{#59144}
Anyref parameters can exist across GC runs. Therefore the GC has to
know where anyref parameters are on the stack so that it can mark them
in its marking phase, and update them in the compaction phase.
Already in a previous CL we grouped all anyref parameters so that they
can be found more easily in a stack frame, see
https://crrev.com/c/1371827. In this CL we implement the stack scanning
itself.
Note that anyref parameters are not scanned while iterating over the
caller's frame (to which they actually belong), but while iterating
over the callee's frame. The reason is that with tail-calls, only the
callee knows how many tagged stack parameters (aka anyref parameters)
there are.
R=mstarzinger@chromium.orgalso-by=mstarzinger@chromium.org
Bug: v8:7581
Change-Id: I7a41ce11d06c0d420146fdb0bb8d5606f28824d7
Reviewed-on: https://chromium-review.googlesource.com/c/1424955
Commit-Queue: Andreas Haas <ahaas@chromium.org>
Reviewed-by: Michael Starzinger <mstarzinger@chromium.org>
Cr-Commit-Position: refs/heads/master@{#59099}
Mostly signed integer overflows, and a few cases of double
division by zero (which is defined by IEEE-754 to return
Infinity (or NaN for 0/0) but is UB in C++).
Bug: v8:3770
Change-Id: Id92725b0ac57cb357978124a3dc6f477430bc97d
Reviewed-on: https://chromium-review.googlesource.com/c/1403133
Commit-Queue: Jakob Kummerow <jkummerow@chromium.org>
Reviewed-by: Clemens Hammacher <clemensh@chromium.org>
Cr-Commit-Position: refs/heads/master@{#58696}
This avoids creating an on-heap copy for import wrappers by directly
adding the {WasmCode} into the native heap instead. It reduces
compilation time as well as useless GC pressure.
R=clemensh@chromium.org
BUG=v8:8423
Change-Id: Ia063523834c963591027c7d1ed78b795d24907bf
Reviewed-on: https://chromium-review.googlesource.com/c/1335566
Commit-Queue: Michael Starzinger <mstarzinger@chromium.org>
Reviewed-by: Andreas Haas <ahaas@chromium.org>
Reviewed-by: Clemens Hammacher <clemensh@chromium.org>
Cr-Commit-Position: refs/heads/master@{#57511}
This CL implements several things needed for full SIMD BE support in WASM:
* Global variables are now kept in little-endian order as per specification
* Added support for SIMD on BE in wasm interpreter
* Fixed several tests that didn't work on BE because input or output
data were not using LE in-memory layout
Change-Id: I4542d13d09fd276e15b0fc39f02e4a58831f65e4
Reviewed-on: https://chromium-review.googlesource.com/1160484
Commit-Queue: Ivica Bogosavljevic <ibogosavljevic@wavecomp.com>
Reviewed-by: Bill Budge <bbudge@chromium.org>
Reviewed-by: Aseem Garg <aseemgarg@chromium.org>
Cr-Commit-Position: refs/heads/master@{#55075}
This moves the entire code generation phase (including code emission
into the native module) into the background task. The code manager is
fully thread safe by now and there are no Isolate-bound steps anymore.
The only step remaining on the foreground task is publishing the fully
finished code to other threads via {NativeModule::PublishCode}.
R=clemensh@chromium.org
BUG=v8:7921
Change-Id: Ia64c6ce945aabd071b26e61ef8d397fb7727a038
Reviewed-on: https://chromium-review.googlesource.com/1135004
Commit-Queue: Michael Starzinger <mstarzinger@chromium.org>
Reviewed-by: Clemens Hammacher <clemensh@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54425}
This CL removes the friendship between {NativeModule} and
{NativeModuleSerializer}/{NativeModuleDeserializer}.
Instead, it adds a new public method ({AddDeserializedCode}) which is
being called from the deserializer.
Drive-by: Unify the argument order to {AddCode}, {AddOwnedCode} and
{WasmCode}.
R=mstarzinger@chromium.org
Bug: chromium:856938
Change-Id: I88943c90c45650e21ae6bc17395a17f86319c046
Reviewed-on: https://chromium-review.googlesource.com/1117075
Reviewed-by: Michael Starzinger <mstarzinger@chromium.org>
Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54084}
We currently store the protected instructions per code object in a
{std::unique_ptr<std::vector<ProtectedInstructionData>>}. This wastes
memory, because it requires two heap allocations, plus the vector might
over-allocate (and it currently does, because it is filled dynamically
during compilation).
This CL changes that to store the protected instructions in an
{OwnedVector}. This requires one copy after generating the list of
{ProtectedInstructionData} in an {std::vector} during compilation, but
saves memory afterwards.
R=mstarzinger@chromium.org
Bug: chromium:856938
Change-Id: Ie290a17dc32f27fbbfe0c000a52297181c954550
Reviewed-on: https://chromium-review.googlesource.com/1116701
Reviewed-by: Michael Starzinger <mstarzinger@chromium.org>
Commit-Queue: Clemens Hammacher <clemensh@chromium.org>
Cr-Commit-Position: refs/heads/master@{#54052}
This changes the WebAssembly pipeline to no longer expect source
position tables for {WasmCode} to be allocated on the GC'ed heap.
R=clemensh@chromium.org
BUG=v8:7721
Change-Id: Ib2c6e3d0840e47b83809f60519c0d1b94af186af
Reviewed-on: https://chromium-review.googlesource.com/1109686
Commit-Queue: Michael Starzinger <mstarzinger@chromium.org>
Reviewed-by: Clemens Hammacher <clemensh@chromium.org>
Cr-Commit-Position: refs/heads/master@{#53961}
