Backends do not care about the concrete type, they only need to know the
"kind" (e.g. "ref" or "i32").
In order to prepare Liftoff to use the value kind instead of the
value type for all stored data, this CL moves the kind out of the
ValueType and makes it a top-level enum.
R=manoskouk@chromium.org
Bug: v8:11477
Change-Id: I489d6c5207e6ff1b66e2afbe78a156d66df27eb3
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2707169
Commit-Queue: Clemens Backes <clemensb@chromium.org>
Reviewed-by: Manos Koukoutos <manoskouk@chromium.org>
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Cr-Commit-Position: refs/heads/master@{#72896}
Also move it from post-mvp to mvp, since it is now in the proposal.
Bug: v8:11002
Change-Id: I711ee7a92e6937948c93e6028ef018188ea4c976
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2676937
Reviewed-by: Deepti Gandluri <gdeepti@chromium.org>
Commit-Queue: Zhi An Ng <zhin@chromium.org>
Cr-Commit-Position: refs/heads/master@{#72643}
Load lane instructions also need a v128 input.
Bug: chromium:1173488
Change-Id: I45e4c4f8fc93a5b3246ac4d1b07925b41cbe3e89
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2673275
Reviewed-by: Deepti Gandluri <gdeepti@chromium.org>
Commit-Queue: Zhi An Ng <zhin@chromium.org>
Cr-Commit-Position: refs/heads/master@{#72517}
This is a reland of commit 9c09c227b0.
The fix for gc stress failure is merged: https://crrev.com/c/2656857.
Original change's description:
> Bug: v8:11331
> Change-Id: Ie394ec841a1a1c4030c4f589eac2cee8a6a2a1f9
> Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2639033
> Reviewed-by: Georg Neis <neis@chromium.org>
> Reviewed-by: Deepti Gandluri <gdeepti@chromium.org>
> Commit-Queue: Zhi An Ng <zhin@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#72304}
Bug: v8:11331
Change-Id: I82f57b3fe5f0c456472aa7ce404703f34b73d17e
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2659511
Reviewed-by: Deepti Gandluri <gdeepti@chromium.org>
Reviewed-by: Georg Neis <neis@chromium.org>
Commit-Queue: Zhi An Ng <zhin@chromium.org>
Cr-Commit-Position: refs/heads/master@{#72465}
This CL implements the upcoming spec change:
https://github.com/tc39/proposal-regexp-match-indices/pull/49
A new JSRegExpResultWithIndices subclass is introduced with a separate map and
an extra slot for storing the indices. If /d is passed, exec() constructs a
JSRegExpResultWithIndices and eagerly builds indices.
The existing re-execution logic is removed.
Bug: v8:9548
Change-Id: Ic11853e7521017af5e8bd583c7b82bb672821132
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2616873
Commit-Queue: Shu-yu Guo <syg@chromium.org>
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Reviewed-by: Toon Verwaest <verwaest@chromium.org>
Cr-Commit-Position: refs/heads/master@{#72306}
This reverts commit 9c09c227b0.
Reason for revert: gc stress failures https://ci.chromium.org/ui/p/v8/builders/ci/V8%20Linux%20-%20arm64%20-%20sim%20-%20gc%20stress/20563/overview
Original change's description:
> [wasm-simd] Merge all any_true to v128.any_true
>
> In https://github.com/WebAssembly/simd/pull/423, all any_true
> instructions were removed, and replaced with a single v128.any_true.
>
> This patch removes all but v8x16.any_true, and renames it to
> v128.any_true.
>
> Bug: v8:11331
> Change-Id: Ie394ec841a1a1c4030c4f589eac2cee8a6a2a1f9
> Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2639033
> Reviewed-by: Georg Neis <neis@chromium.org>
> Reviewed-by: Deepti Gandluri <gdeepti@chromium.org>
> Commit-Queue: Zhi An Ng <zhin@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#72304}
TBR=neis@chromium.org,gdeepti@chromium.org,neis@google.com,zhin@chromium.org
Change-Id: I52dbf8de679059dd7b17908c1fe3ada0eb54ff84
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Bug: v8:11331
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2649240
Reviewed-by: Zhi An Ng <zhin@chromium.org>
Commit-Queue: Zhi An Ng <zhin@chromium.org>
Cr-Commit-Position: refs/heads/master@{#72305}
In https://github.com/WebAssembly/simd/pull/423, all any_true
instructions were removed, and replaced with a single v128.any_true.
This patch removes all but v8x16.any_true, and renames it to
v128.any_true.
Bug: v8:11331
Change-Id: Ie394ec841a1a1c4030c4f589eac2cee8a6a2a1f9
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2639033
Reviewed-by: Georg Neis <neis@chromium.org>
Reviewed-by: Deepti Gandluri <gdeepti@chromium.org>
Commit-Queue: Zhi An Ng <zhin@chromium.org>
Cr-Commit-Position: refs/heads/master@{#72304}
The inspector fuzzer is terminating the isolate after two seconds. At
this point, we can be in pretty much any state, and any further JS
execution would fail.
This CL fixes an issue where we got the termination signal when creating
a context for a regexp (while installing extensions).
There might be more places that need fixing, but with this CL the linked
issue does not reproduce locally any more, so it's a step forward.
R=szuend@chromium.org, bmeurer@chromium.org
Bug: chromium:1166549
Change-Id: I33b48205b71877aca6cfe5267f353fa899bfa05c
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2636153
Reviewed-by: Benedikt Meurer <bmeurer@chromium.org>
Reviewed-by: Simon Zünd <szuend@chromium.org>
Commit-Queue: Clemens Backes <clemensb@chromium.org>
Cr-Commit-Position: refs/heads/master@{#72156}
Exception ref is not part of the latest exception handling
proposal. Remove the corresponding value type, the type opcode and some
obsolete tests.
R=clemensb@chromium.org
Bug: v8:8091
Change-Id: I5e068ba3426f4b56b90ef056193acdd8cc8fe7d7
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2632599
Reviewed-by: Clemens Backes <clemensb@chromium.org>
Commit-Queue: Thibaud Michaud <thibaudm@chromium.org>
Cr-Commit-Position: refs/heads/master@{#72116}
The watchdog previously didn't terminate execution, it just prevented
the execution of additional tasks.
This CL fixes that by making {TaskRunner::Terminate} actually terminate
execution in the isolate.
It also adds a regression test for this.
R=szuend@chromium.org
Bug: chromium:1154412, chromium:1142437
Change-Id: Ic6638e8a5c37e8840a85651b4d4bea2ee0f71c43
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2622212
Reviewed-by: Simon Zünd <szuend@chromium.org>
Commit-Queue: Clemens Backes <clemensb@chromium.org>
Cr-Commit-Position: refs/heads/master@{#72031}
read_heap_type did not have knowledge of the module for which the heap
type was being decoded. As a result, callers of read_heap_type (or
read_value_type, which in turn calls read_heap_type) had to check after
the fact that a decoded indexed type (ref, ref null, or rtt) references
a type index within the module's bounds. This was not done consistently,
and was missing (at least) in DecodeLocals.
To avoid such problems in the future, this CL refactors read_heap_type
to accept a module and check the decoded index against it.
Changes:
- Add WasmModule argument to read_heap_type. Do so accordingly to all
its transitive callers (read_value_type, immediate arguments,
DecodeLocalDecls, DecodeValue/HeapType in unittests).
- Add index check to read_heap_type and emit an error for an
out-of-bounds index.
- Remove all other now-redundant index validations. Replace them with
decoder->ok() if needed (since read_heap_type will now emit an error).
- Fix error message in Validate for BlockTypeImmediate.
- In DecodeLocalDecls in unittests, pass an empty module to
DecodeLocalDecls in the main code.
- Add a unit test with an invalid index in local type declarations.
Bug: v8:9495
Change-Id: I4ed1204847db80f78b6ae85fa40d300cd2456295
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2569757
Reviewed-by: Jakob Kummerow <jkummerow@chromium.org>
Commit-Queue: Manos Koukoutos <manoskouk@chromium.org>
Cr-Commit-Position: refs/heads/master@{#71572}
The wasm fuzzer corpus is pretty outdated. The script that was used to
generate it did not work any more.
This CL updates the script, and runs it. This generates a fuzzer corpus
of 42011 wasm modules, compared to 15290 before. The new modules will
contain new features like SIMD and multi-value, which will be
interesting fuzzer inputs.
R=ahaas@chromium.org
Change-Id: Ic3df26930cb8c1c6e8d521597ceb06cc338c02ed
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2565512
Reviewed-by: Andreas Haas <ahaas@chromium.org>
Commit-Queue: Clemens Backes <clemensb@chromium.org>
Cr-Commit-Position: refs/heads/master@{#71537}
The auto-generated inspector fuzzer corpus seed files will overwrite the
'utils' class by a proxy which provides non-existing functions.
See https://crrev.com/c/2563552.
R=szuend@chromium.org
Bug: chromium:1142437
Change-Id: If1e86617c4244f1b12fe007b5059b5a5f57454d5
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2565127
Reviewed-by: Simon Zünd <szuend@chromium.org>
Commit-Queue: Clemens Backes <clemensb@chromium.org>
Cr-Commit-Position: refs/heads/master@{#71515}
Scopes in V8 are used to guarantee one or more properties during its
lifetimes. If a scope is not named e.g MyClassScope(args) instead of
MyClassScope scope(args) it will get created and automatically destroyed
and therefore, being useless as a scope. This CL would produce a
compiling warning when that happens to ward off this developer error.
Follow-up to ccrev.com/2552415 in which it was introduced and
implemented for Guard classes.
Change-Id: Ifa0fb89cc3d9bdcdee0fd8150a2618af5ef45cbf
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2555001
Commit-Queue: Santiago Aboy Solanes <solanes@chromium.org>
Reviewed-by: Ulan Degenbaev <ulan@chromium.org>
Reviewed-by: Leszek Swirski <leszeks@chromium.org>
Reviewed-by: Michael Lippautz <mlippautz@chromium.org>
Reviewed-by: Jakob Kummerow <jkummerow@chromium.org>
Reviewed-by: Ross McIlroy <rmcilroy@chromium.org>
Reviewed-by: Tobias Tebbi <tebbi@chromium.org>
Cr-Commit-Position: refs/heads/master@{#71425}
Make compileAndRunWithOrigin accept the same six arguments as
inspector-test. This makes inspector tests more useful as seed for the
inspector fuzzer, and allows to run more inspector fuzzer outputs
directly in the inspector-test binary.
R=szuend@chromium.org
Bug: chromium:1142437
Change-Id: Ib9e9768c834204ff17a641e9d462400a139bf6b0
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2557507
Reviewed-by: Simon Zünd <szuend@chromium.org>
Commit-Queue: Clemens Backes <clemensb@chromium.org>
Cr-Commit-Position: refs/heads/master@{#71363}
The fuzzer is expected to generate a lot of syntax and runtime errors,
and the respective messages just flood the fuzzer output. By always
putting a {TryCatch} scope around the execution, we prevent those
messages from being printed.
At the same time, inspector tests need to properly propagate uncaught
exceptions in the backend to the inspector, and fail on uncaught
exceptions in the frontend.
This CL allows for all these behaviours by extending the
{CatchExceptions} enum and the {TryCatch} logic in the task runner.
Drive-by: Use {base::OS::ExitProcess} instead of the explicit
{fflush} and {_exit}.
R=szuend@chromium.org
Bug: chromium:1142437
Change-Id: Ic2cb3b0de2399d25bd8c53090575308cb0e09ab0
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2529135
Commit-Queue: Clemens Backes <clemensb@chromium.org>
Reviewed-by: Simon Zünd <szuend@chromium.org>
Cr-Commit-Position: refs/heads/master@{#71152}
This CL
* renames Name::hash_field field to raw_hash_field.
* all local variables that store raw_hash_field value are also renamed
to raw_hash_field where possible.
Bug: chromium:1133527, v8:11074
Change-Id: I17313f386110b33a64f629cc2b9d4afd1e06c6c0
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2471999
Reviewed-by: Peter Marshall <petermarshall@chromium.org>
Reviewed-by: Ulan Degenbaev <ulan@chromium.org>
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Reviewed-by: Toon Verwaest <verwaest@chromium.org>
Commit-Queue: Igor Sheludko <ishell@chromium.org>
Cr-Commit-Position: refs/heads/master@{#71114}
For the fuzzer it's unwise to exit on uncaught exceptions, as this
terminates the whole fuzzing process. Just ignore those exceptions
instead.
Drive-by: Fix a typo.
R=szuend@chromium.org
Bug: chromium:1142437
Change-Id: Ided1c0f35840c158f157acd8c0bb1c12ecf8a37f
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2526386
Commit-Queue: Clemens Backes <clemensb@chromium.org>
Reviewed-by: Simon Zünd <szuend@chromium.org>
Cr-Commit-Position: refs/heads/master@{#71059}
Instead of passing two bools to the {TaskRunner} constructor, pass to
enums. This makes the semantics more clear in the caller.
In the fuzzer, we actually *do not* want to catch exceptions. This
semantic fix will be done in a follow-up CL, such that this CL is a pure
refactoring.
R=szuend@chromium.org
Bug: v8:11074
Change-Id: I7f6df3a3f344524deb08db10b9317a6734b7ea42
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2526385
Reviewed-by: Simon Zünd <szuend@chromium.org>
Commit-Queue: Clemens Backes <clemensb@chromium.org>
Cr-Commit-Position: refs/heads/master@{#71056}
Joining the thread from the watchdog is problematic, since e.g.
{pthread_join} (the implementation of {Thread::Join} on POSIX systems)
has undefined behaviour if multiple threads try to join at the same
time. In practice, this leads to deadlocks.
Thus implement termination by just calling {TaskRunner::Terminate}, but
not {TaskRunner::Join}. This fixes the deadlocks in the inspector
fuzzer.
The inspector test binary is fixed simarly, even though there it seems
to not cause problems so far.
In both files, the {Terminate} function is inlined into callers because
it's only a single line now, with one to two users.
Also, replace the single fuzzer test (which is invalid javascript) by
two tests: One called "invalid" explicitly, still with invalid
javascript, and one empty file, which is valid input. That one
reproduced the deadlock.
R=szuend@chromium.org
Bug: chromium:1142437
Change-Id: I8fb98b0cdbf3ceff6af6849397e5da5a4e9acd3c
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2526384
Reviewed-by: Simon Zünd <szuend@chromium.org>
Commit-Queue: Clemens Backes <clemensb@chromium.org>
Cr-Commit-Position: refs/heads/master@{#71054}
Keep tasks in unique_ptrs, such that they are freed independent of
whether they have been executed or not.
R=szuend@chromium.org
Bug: chromium:1142437, v8:11107, v8:11074
Change-Id: Ia265df3187c724b63e0f576d33235c1bfa522c4f
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2517694
Commit-Queue: Clemens Backes <clemensb@chromium.org>
Reviewed-by: Simon Zünd <szuend@chromium.org>
Cr-Commit-Position: refs/heads/master@{#71005}
The inspector fuzzer is running into timeouts most of the time
currently, because the test explicitly needs to quit execution.
Make fuzzing more efficient by adding a watchdog thread which stop
execution after 2 seconds. This will still result in valid test cases,
i.e. everything that was executed within those two seconds will count as
covered code.
Drive-by: Slightly simplify the storage of task runners. No need to
clear the vector after termination.
R=szuend@chromium.org
Bug: chromium:1142437, chromium:1145285
Change-Id: I7b5fe7ddcbce731fbc3d74ee8c43f7249f34b918
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2520906
Commit-Queue: Clemens Backes <clemensb@chromium.org>
Reviewed-by: Simon Zünd <szuend@chromium.org>
Cr-Commit-Position: refs/heads/master@{#71002}
This adds a first simple version of the inspector fuzzer, which is a
stripped-down version of the inspector-test executable. The fuzzer
generates inputs which are compatible with inspector-test.
There are still memory leaks, and the fuzzer will probably run into
timeouts most of the time. Both of this will be addressed in follow-ups.
R=szuend@chromium.org, machenbach@chromium.org
Bug: chromium:1142437
Change-Id: I4d13da460f571d791a3642b0705a1f07b442c11b
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2505722
Commit-Queue: Clemens Backes <clemensb@chromium.org>
Reviewed-by: Simon Zünd <szuend@chromium.org>
Reviewed-by: Michael Achenbach <machenbach@chromium.org>
Cr-Commit-Position: refs/heads/master@{#70922}
This commit adds the 'l' (linear) RegExp flag (as in e.g. /asdf|123/l)
that forces execution in linear time. These regexps are handled by the
experimental engine. If the experimental engine cannot handle the
pattern, an exception is thrown on creation of the regexp.
The commit also adds a new global V8 flag and changes an existing one:
* --enable-experimental-engine, which turns on recognition of the RegExp
'l' flag. Previously this flag also caused all supported regexps to
be executed by the experimental engine; this is not the case anymore.
* --default-to-experimental-regexp-engine takes over the previous
semantics of --enable-experimental-regexp-engine: We execute all
supported regexps with the experimental engine.
Cq-Include-Trybots: luci.v8.try:v8_linux64_fyi_rel_ng
Bug: v8:10765
Change-Id: I5622a89b19404105e8be280d454e9fdd63c003b3
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2461244
Reviewed-by: Ulan Degenbaev <ulan@chromium.org>
Reviewed-by: Georg Neis <neis@chromium.org>
Reviewed-by: Simon Zünd <szuend@chromium.org>
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Commit-Queue: Martin Bidlingmaier <mbid@google.com>
Cr-Commit-Position: refs/heads/master@{#70892}
This is a reland of fbfa9bf4ec
The arm64 was missing proper codegen for CFI, thus sizes were off.
Original change's description:
> Reland "[deoptimizer] Change deopt entries into builtins"
>
> This is a reland of 7f58ced72e
>
> It fixes the different exit size emitted on x64/Atom CPUs due to
> performance tuning in TurboAssembler::Call. Additionally, add
> cctests to verify the fixed size exits.
>
> Original change's description:
> > [deoptimizer] Change deopt entries into builtins
> >
> > While the overall goal of this commit is to change deoptimization
> > entries into builtins, there are multiple related things happening:
> >
> > - Deoptimization entries, formerly stubs (i.e. Code objects generated
> > at runtime, guaranteed to be immovable), have been converted into
> > builtins. The major restriction is that we now need to preserve the
> > kRootRegister, which was formerly used on most architectures to pass
> > the deoptimization id. The solution differs based on platform.
> > - Renamed DEOPT_ENTRIES_OR_FOR_TESTING code kind to FOR_TESTING.
> > - Removed heap/ support for immovable Code generation.
> > - Removed the DeserializerData class (no longer needed).
> > - arm64: to preserve 4-byte deopt exits, introduced a new optimization
> > in which the final jump to the deoptimization entry is generated
> > once per Code object, and deopt exits can continue to emit a
> > near-call.
> > - arm,ia32,x64: change to fixed-size deopt exits. This reduces exit
> > sizes by 4/8, 5, and 5 bytes, respectively.
> >
> > On arm the deopt exit size is reduced from 12 (or 16) bytes to 8 bytes
> > by using the same strategy as on arm64 (recalc deopt id from return
> > address). Before:
> >
> > e300a002 movw r10, <id>
> > e59fc024 ldr ip, [pc, <entry offset>]
> > e12fff3c blx ip
> >
> > After:
> >
> > e59acb35 ldr ip, [r10, <entry offset>]
> > e12fff3c blx ip
> >
> > On arm64 the deopt exit size remains 4 bytes (or 8 bytes in same cases
> > with CFI). Additionally, up to 4 builtin jumps are emitted per Code
> > object (max 32 bytes added overhead per Code object). Before:
> >
> > 9401cdae bl <entry offset>
> >
> > After:
> >
> > # eager deoptimization entry jump.
> > f95b1f50 ldr x16, [x26, <eager entry offset>]
> > d61f0200 br x16
> > # lazy deoptimization entry jump.
> > f95b2b50 ldr x16, [x26, <lazy entry offset>]
> > d61f0200 br x16
> > # the deopt exit.
> > 97fffffc bl <eager deoptimization entry jump offset>
> >
> > On ia32 the deopt exit size is reduced from 10 to 5 bytes. Before:
> >
> > bb00000000 mov ebx,<id>
> > e825f5372b call <entry>
> >
> > After:
> >
> > e8ea2256ba call <entry>
> >
> > On x64 the deopt exit size is reduced from 12 to 7 bytes. Before:
> >
> > 49c7c511000000 REX.W movq r13,<id>
> > e8ea2f0700 call <entry>
> >
> > After:
> >
> > 41ff9560360000 call [r13+<entry offset>]
> >
> > Bug: v8:8661,v8:8768
> > Change-Id: I13e30aedc360474dc818fecc528ce87c3bfeed42
> > Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2465834
> > Commit-Queue: Jakob Gruber <jgruber@chromium.org>
> > Reviewed-by: Ross McIlroy <rmcilroy@chromium.org>
> > Reviewed-by: Tobias Tebbi <tebbi@chromium.org>
> > Reviewed-by: Ulan Degenbaev <ulan@chromium.org>
> > Cr-Commit-Position: refs/heads/master@{#70597}
>
> Tbr: ulan@chromium.org, tebbi@chromium.org, rmcilroy@chromium.org
> Bug: v8:8661,v8:8768,chromium:1140165
> Change-Id: Ibcd5c39c58a70bf2b2ac221aa375fc68d495e144
> Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2485506
> Reviewed-by: Jakob Gruber <jgruber@chromium.org>
> Reviewed-by: Tobias Tebbi <tebbi@chromium.org>
> Commit-Queue: Jakob Gruber <jgruber@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#70655}
Tbr: ulan@chromium.org, tebbi@chromium.org, rmcilroy@chromium.org
Bug: v8:8661
Bug: v8:8768
Bug: chromium:1140165
Change-Id: I471cc94fc085e527dc9bfb5a84b96bd907c2333f
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2488682
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Commit-Queue: Jakob Gruber <jgruber@chromium.org>
Cr-Commit-Position: refs/heads/master@{#70672}
This is a reland of 7f58ced72e
It fixes the different exit size emitted on x64/Atom CPUs due to
performance tuning in TurboAssembler::Call. Additionally, add
cctests to verify the fixed size exits.
Original change's description:
> [deoptimizer] Change deopt entries into builtins
>
> While the overall goal of this commit is to change deoptimization
> entries into builtins, there are multiple related things happening:
>
> - Deoptimization entries, formerly stubs (i.e. Code objects generated
> at runtime, guaranteed to be immovable), have been converted into
> builtins. The major restriction is that we now need to preserve the
> kRootRegister, which was formerly used on most architectures to pass
> the deoptimization id. The solution differs based on platform.
> - Renamed DEOPT_ENTRIES_OR_FOR_TESTING code kind to FOR_TESTING.
> - Removed heap/ support for immovable Code generation.
> - Removed the DeserializerData class (no longer needed).
> - arm64: to preserve 4-byte deopt exits, introduced a new optimization
> in which the final jump to the deoptimization entry is generated
> once per Code object, and deopt exits can continue to emit a
> near-call.
> - arm,ia32,x64: change to fixed-size deopt exits. This reduces exit
> sizes by 4/8, 5, and 5 bytes, respectively.
>
> On arm the deopt exit size is reduced from 12 (or 16) bytes to 8 bytes
> by using the same strategy as on arm64 (recalc deopt id from return
> address). Before:
>
> e300a002 movw r10, <id>
> e59fc024 ldr ip, [pc, <entry offset>]
> e12fff3c blx ip
>
> After:
>
> e59acb35 ldr ip, [r10, <entry offset>]
> e12fff3c blx ip
>
> On arm64 the deopt exit size remains 4 bytes (or 8 bytes in same cases
> with CFI). Additionally, up to 4 builtin jumps are emitted per Code
> object (max 32 bytes added overhead per Code object). Before:
>
> 9401cdae bl <entry offset>
>
> After:
>
> # eager deoptimization entry jump.
> f95b1f50 ldr x16, [x26, <eager entry offset>]
> d61f0200 br x16
> # lazy deoptimization entry jump.
> f95b2b50 ldr x16, [x26, <lazy entry offset>]
> d61f0200 br x16
> # the deopt exit.
> 97fffffc bl <eager deoptimization entry jump offset>
>
> On ia32 the deopt exit size is reduced from 10 to 5 bytes. Before:
>
> bb00000000 mov ebx,<id>
> e825f5372b call <entry>
>
> After:
>
> e8ea2256ba call <entry>
>
> On x64 the deopt exit size is reduced from 12 to 7 bytes. Before:
>
> 49c7c511000000 REX.W movq r13,<id>
> e8ea2f0700 call <entry>
>
> After:
>
> 41ff9560360000 call [r13+<entry offset>]
>
> Bug: v8:8661,v8:8768
> Change-Id: I13e30aedc360474dc818fecc528ce87c3bfeed42
> Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2465834
> Commit-Queue: Jakob Gruber <jgruber@chromium.org>
> Reviewed-by: Ross McIlroy <rmcilroy@chromium.org>
> Reviewed-by: Tobias Tebbi <tebbi@chromium.org>
> Reviewed-by: Ulan Degenbaev <ulan@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#70597}
Tbr: ulan@chromium.org, tebbi@chromium.org, rmcilroy@chromium.org
Bug: v8:8661,v8:8768,chromium:1140165
Change-Id: Ibcd5c39c58a70bf2b2ac221aa375fc68d495e144
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2485506
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Reviewed-by: Tobias Tebbi <tebbi@chromium.org>
Commit-Queue: Jakob Gruber <jgruber@chromium.org>
Cr-Commit-Position: refs/heads/master@{#70655}
This reverts commit 7f58ced72e.
Reason for revert: Segfaults on Atom_x64 https://ci.chromium.org/p/v8-internal/builders/ci/v8_linux64_atom_perf/5686?
Original change's description:
> [deoptimizer] Change deopt entries into builtins
>
> While the overall goal of this commit is to change deoptimization
> entries into builtins, there are multiple related things happening:
>
> - Deoptimization entries, formerly stubs (i.e. Code objects generated
> at runtime, guaranteed to be immovable), have been converted into
> builtins. The major restriction is that we now need to preserve the
> kRootRegister, which was formerly used on most architectures to pass
> the deoptimization id. The solution differs based on platform.
> - Renamed DEOPT_ENTRIES_OR_FOR_TESTING code kind to FOR_TESTING.
> - Removed heap/ support for immovable Code generation.
> - Removed the DeserializerData class (no longer needed).
> - arm64: to preserve 4-byte deopt exits, introduced a new optimization
> in which the final jump to the deoptimization entry is generated
> once per Code object, and deopt exits can continue to emit a
> near-call.
> - arm,ia32,x64: change to fixed-size deopt exits. This reduces exit
> sizes by 4/8, 5, and 5 bytes, respectively.
>
> On arm the deopt exit size is reduced from 12 (or 16) bytes to 8 bytes
> by using the same strategy as on arm64 (recalc deopt id from return
> address). Before:
>
> e300a002 movw r10, <id>
> e59fc024 ldr ip, [pc, <entry offset>]
> e12fff3c blx ip
>
> After:
>
> e59acb35 ldr ip, [r10, <entry offset>]
> e12fff3c blx ip
>
> On arm64 the deopt exit size remains 4 bytes (or 8 bytes in same cases
> with CFI). Additionally, up to 4 builtin jumps are emitted per Code
> object (max 32 bytes added overhead per Code object). Before:
>
> 9401cdae bl <entry offset>
>
> After:
>
> # eager deoptimization entry jump.
> f95b1f50 ldr x16, [x26, <eager entry offset>]
> d61f0200 br x16
> # lazy deoptimization entry jump.
> f95b2b50 ldr x16, [x26, <lazy entry offset>]
> d61f0200 br x16
> # the deopt exit.
> 97fffffc bl <eager deoptimization entry jump offset>
>
> On ia32 the deopt exit size is reduced from 10 to 5 bytes. Before:
>
> bb00000000 mov ebx,<id>
> e825f5372b call <entry>
>
> After:
>
> e8ea2256ba call <entry>
>
> On x64 the deopt exit size is reduced from 12 to 7 bytes. Before:
>
> 49c7c511000000 REX.W movq r13,<id>
> e8ea2f0700 call <entry>
>
> After:
>
> 41ff9560360000 call [r13+<entry offset>]
>
> Bug: v8:8661,v8:8768
> Change-Id: I13e30aedc360474dc818fecc528ce87c3bfeed42
> Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2465834
> Commit-Queue: Jakob Gruber <jgruber@chromium.org>
> Reviewed-by: Ross McIlroy <rmcilroy@chromium.org>
> Reviewed-by: Tobias Tebbi <tebbi@chromium.org>
> Reviewed-by: Ulan Degenbaev <ulan@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#70597}
TBR=ulan@chromium.org,rmcilroy@chromium.org,jgruber@chromium.org,tebbi@chromium.org
# Not skipping CQ checks because original CL landed > 1 day ago.
Bug: v8:8661,v8:8768,chromium:1140165
Change-Id: I3df02ab42f6e02233d9f6fb80e8bb18f76870d91
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2485504
Reviewed-by: Jakob Gruber <jgruber@chromium.org>
Commit-Queue: Jakob Gruber <jgruber@chromium.org>
Cr-Commit-Position: refs/heads/master@{#70649}
While the overall goal of this commit is to change deoptimization
entries into builtins, there are multiple related things happening:
- Deoptimization entries, formerly stubs (i.e. Code objects generated
at runtime, guaranteed to be immovable), have been converted into
builtins. The major restriction is that we now need to preserve the
kRootRegister, which was formerly used on most architectures to pass
the deoptimization id. The solution differs based on platform.
- Renamed DEOPT_ENTRIES_OR_FOR_TESTING code kind to FOR_TESTING.
- Removed heap/ support for immovable Code generation.
- Removed the DeserializerData class (no longer needed).
- arm64: to preserve 4-byte deopt exits, introduced a new optimization
in which the final jump to the deoptimization entry is generated
once per Code object, and deopt exits can continue to emit a
near-call.
- arm,ia32,x64: change to fixed-size deopt exits. This reduces exit
sizes by 4/8, 5, and 5 bytes, respectively.
On arm the deopt exit size is reduced from 12 (or 16) bytes to 8 bytes
by using the same strategy as on arm64 (recalc deopt id from return
address). Before:
e300a002 movw r10, <id>
e59fc024 ldr ip, [pc, <entry offset>]
e12fff3c blx ip
After:
e59acb35 ldr ip, [r10, <entry offset>]
e12fff3c blx ip
On arm64 the deopt exit size remains 4 bytes (or 8 bytes in same cases
with CFI). Additionally, up to 4 builtin jumps are emitted per Code
object (max 32 bytes added overhead per Code object). Before:
9401cdae bl <entry offset>
After:
# eager deoptimization entry jump.
f95b1f50 ldr x16, [x26, <eager entry offset>]
d61f0200 br x16
# lazy deoptimization entry jump.
f95b2b50 ldr x16, [x26, <lazy entry offset>]
d61f0200 br x16
# the deopt exit.
97fffffc bl <eager deoptimization entry jump offset>
On ia32 the deopt exit size is reduced from 10 to 5 bytes. Before:
bb00000000 mov ebx,<id>
e825f5372b call <entry>
After:
e8ea2256ba call <entry>
On x64 the deopt exit size is reduced from 12 to 7 bytes. Before:
49c7c511000000 REX.W movq r13,<id>
e8ea2f0700 call <entry>
After:
41ff9560360000 call [r13+<entry offset>]
Bug: v8:8661,v8:8768
Change-Id: I13e30aedc360474dc818fecc528ce87c3bfeed42
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2465834
Commit-Queue: Jakob Gruber <jgruber@chromium.org>
Reviewed-by: Ross McIlroy <rmcilroy@chromium.org>
Reviewed-by: Tobias Tebbi <tebbi@chromium.org>
Reviewed-by: Ulan Degenbaev <ulan@chromium.org>
Cr-Commit-Position: refs/heads/master@{#70597}
Rename AddSaturate and SubSaturate to the shorter version, AddSat and
SubSat, following the spec.
Bug: v8:10946,v8:10933
Change-Id: Idf74b3a1eb2e2f6d4e37d2b8e5fa6d96ea090db4
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2436615
Reviewed-by: Tobias Tebbi <tebbi@chromium.org>
Reviewed-by: Jakob Kummerow <jkummerow@chromium.org>
Reviewed-by: Bill Budge <bbudge@chromium.org>
Commit-Queue: Zhi An Ng <zhin@chromium.org>
Cr-Commit-Position: refs/heads/master@{#70549}
Fuzzers are executed in their own process, so instead of resetting flags
after execution, we can just keep the flag values.
This CL introduces a shared function to enable all staged features,
without ever resetting the value. This fixes a data race.
R=ahaas@chromium.org
Bug: v8:10979
Change-Id: I82ea35b887841850edd8b394a3644cf8df1e3bf8
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2449969
Commit-Queue: Clemens Backes <clemensb@chromium.org>
Reviewed-by: Andreas Haas <ahaas@chromium.org>
Cr-Commit-Position: refs/heads/master@{#70320}
CodeKind::OPTIMIZED_CODE -> TURBOFAN
Kinds are now more fine-grained and distinguish between TF, TP, NCI.
CodeKind::STUB -> DEOPT_ENTRIES_OR_FOR_TESTING
Code stubs (like builtins, but generated at runtime) were removed from
the codebase years ago, this is the last remnant. This kind is used
only for deopt entries (which should be converted into builtins) and
for tests.
Change-Id: I67beb15377cb60f395e9b051b25f3e5764982e93
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2440335
Auto-Submit: Jakob Gruber <jgruber@chromium.org>
Commit-Queue: Mythri Alle <mythria@chromium.org>
Reviewed-by: Mythri Alle <mythria@chromium.org>
Cr-Commit-Position: refs/heads/master@{#70234}
These instructions were changed from "s8x16" to "i8x16" prefixes in
https://github.com/WebAssembly/simd/pull/321.
This CL updates all V8 code, including arch-specific code, to match.
Bug: v8:10946, v8:10933
Change-Id: I26ef9ad77571f94501d42c1d65f57380fd507f3d
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2432068
Reviewed-by: Zhi An Ng <zhin@chromium.org>
Commit-Queue: Adam Klein <adamk@chromium.org>
Cr-Commit-Position: refs/heads/master@{#70143}
This unifies {max_initial_mem_pages} and {max_maximum_mem_pages} into
{max_mem_pages}.
The {CompilationEnv} constructor was incorrectly using the former
instead of the latter anyway. This did not really matter though, since
they typically have the same value.
Also, there is not a single test that sets --wasm-max-mem-pages-growth.
R=manoskouk@chromium.orgCC=jkummerow@chromium.org
Bug: v8:10949
Change-Id: Ib7ab9b4c239d50b72013087eda5a214829c90369
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2426619
Reviewed-by: Manos Koukoutos <manoskouk@chromium.org>
Commit-Queue: Clemens Backes <clemensb@chromium.org>
Cr-Commit-Position: refs/heads/master@{#70114}
Instantiating a module that contains a function (exported) with a v128
in its signature is fine, but then later calling it will trap.
So v128 values are technically not callable from JS, but we can give it
a default argument of 0, and will later trap anyway. This is useful when
fuzzers generate functions with v128 in the signature of the main
function that we then later try to call.
Bug: chromium:1129068
Change-Id: I93f239a0355b8059e25b8bd5f1274d151d71ee11
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/2419657
Commit-Queue: Zhi An Ng <zhin@chromium.org>
Reviewed-by: Clemens Backes <clemensb@chromium.org>
Cr-Commit-Position: refs/heads/master@{#70038}
