.. since it's located in ReadOnlySpace and thus immutable. We could
extend this to other strings in RO-space but for now I want to avoid
too much movement. This bumps jetstream2/gbemu scores by ~30%.
Bug: v8:12790,v8:12161
Change-Id: I3fe10703e9ceca19c110d19c83143d811e090192
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3607994
Auto-Submit: Jakob Linke <jgruber@chromium.org>
Reviewed-by: Toon Verwaest <verwaest@chromium.org>
Commit-Queue: Toon Verwaest <verwaest@chromium.org>
Cr-Commit-Position: refs/heads/main@{#80188}
For lazy compilation there is one generic lazy compilation builtin that
spills all registers on the stack and then triggers the compilation of
the called function. Some of these registers may contain references.
If a GC was triggered during lazy compilation, the GC would have to
know which spill slots on the stack contain references.
This CL adds a check to guarantee that no GC can be triggered during
lazy compilation. Thereby it is not necessary for the GC to know which
spill slots contain references.
If successful, lazy compilation indeed does not allocate on the heap
and therefore cannot trigger a GC. However, when compilation fails, an
error objects needs to be allocated and thrown. This allocation may
trigger a GC, but that's not a problem, because the reference
parameters which may get corrupted by the GC will not be used anyways,
because the called function will never get executed after the failed
compilation.
R=clemensb@chromium.org
Fixes: v8:11366
Change-Id: Ic526d169d4e80ba83f517970ff234e669f854331
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3599474
Reviewed-by: Clemens Backes <clemensb@chromium.org>
Commit-Queue: Andreas Haas <ahaas@chromium.org>
Cr-Commit-Position: refs/heads/main@{#80187}
I cannot reproduce the original problem anymore. The test passes
nicely. If the test (wasm-js/limits) fails again, then we should
just disable it for those platforms where it is failing.
Fixed: v8:11577
Change-Id: I186da2ec2dd8fd518ce4034e8e3838376fefbc3b
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3605814
Reviewed-by: Clemens Backes <clemensb@chromium.org>
Commit-Queue: Andreas Haas <ahaas@chromium.org>
Cr-Commit-Position: refs/heads/main@{#80186}
.. since they may be written from background threads (for SP concurrent
codegen).
Bug: v8:12161
Fixed: chromium:1319857
Change-Id: I11860137ea4d79645821df69e61b62836f7c8283
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3605818
Auto-Submit: Jakob Linke <jgruber@chromium.org>
Reviewed-by: Leszek Swirski <leszeks@chromium.org>
Commit-Queue: Leszek Swirski <leszeks@chromium.org>
Cr-Commit-Position: refs/heads/main@{#80185}
... in DeserializeContexts.
Allocate the context before filling the scope info with the correct
variable name may cause DCHECK failure in TorqueGeneratedClassVerifiers::ScopeInfoVerify.
This CL changes the context variables serialization order from
"name, value, name, value" to "name, name... value, value...".
And only allocate and fill the context after setting up the correct
scope info.
Bug: v8:12832
Change-Id: I6e91aa867a528b7ffcae85de0f2364d0f9ea1bd7
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3607374
Reviewed-by: Marja Hölttä <marja@chromium.org>
Commit-Queue: 王澳 <wangao.james@bytedance.com>
Cr-Commit-Position: refs/heads/main@{#80184}
This is a reland of commit 8ba60b7a8e
Changes to original:
- Weaken DCHECK in MakeThin to allow direct transitions of shared
strings during deserialization. This is safe as the string to be
transitioned is freshly created and hasn't escaped the thread yet.
- To enable this, add has_active_deserializer() to LocalIsolate
- Shared thin string transitions are subject to the same layout changes
as non-shared thin string transitions, thus treat them equally when
checking if a map transition is safe.
Original change's description:
> [string] Non-transitioning shared strings
>
> Instead of transitioning shared strings to ThinString on
> internalization, use a forwarding table to the internalized string and
> store the index into the forwarding table in the string's hash field.
>
> This way we don't need to handle concurrent string transitions that
> modify the underlying string data.
>
> During stop-the-world GC, live strings in the forwarding table are
> migrated to regular ThinStrings.
>
> Bug: v8:12007
> Change-Id: I6c6f3d41c6f644e0aaeafbf25ecec5ce0aa0d2d8
> Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3536647
> Reviewed-by: Dominik Inführ <dinfuehr@chromium.org>
> Reviewed-by: Jakob Linke <jgruber@chromium.org>
> Reviewed-by: Shu-yu Guo <syg@chromium.org>
> Reviewed-by: Igor Sheludko <ishell@chromium.org>
> Commit-Queue: Patrick Thier <pthier@chromium.org>
> Cr-Commit-Position: refs/heads/main@{#79801}
Bug: v8:12007
Change-Id: I022e5c4768b763a86bb28c9c82218c3b807371a0
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3571817
Reviewed-by: Shu-yu Guo <syg@chromium.org>
Reviewed-by: Igor Sheludko <ishell@chromium.org>
Reviewed-by: Dominik Inführ <dinfuehr@chromium.org>
Reviewed-by: Jakob Linke <jgruber@chromium.org>
Commit-Queue: Patrick Thier <pthier@chromium.org>
Cr-Commit-Position: refs/heads/main@{#80183}
This adds the implementation of the {AssemblerBufferCache} class.
PKU-protected memory is allocated via the {WasmCodeManager}, which has
access to the actual protection key.
R=thibaudm@chromium.org
Bug: v8:12809
Change-Id: Id26abd6f98248d5c646ae337ccb903d3e168bed1
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3593137
Reviewed-by: Thibaud Michaud <thibaudm@chromium.org>
Commit-Queue: Clemens Backes <clemensb@chromium.org>
Cr-Commit-Position: refs/heads/main@{#80181}
Other places use the {kSimd} bailout reason for bailouts because of
missing CPU support for SIMD. We should do the same when encountering
SIMD locals or parameters.
R=thibaudm@chromium.org
Change-Id: I7ea6ff32927a035cdafe437d581a79f67ff0b30f
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3605243
Reviewed-by: Thibaud Michaud <thibaudm@chromium.org>
Commit-Queue: Clemens Backes <clemensb@chromium.org>
Cr-Commit-Position: refs/heads/main@{#80180}
When passing anyref-typed things to Wasm, we cannot expect that
all functions are WasmExternalFunctions. Instead of adding a
relatively expensive type check to such calls, this patch disables
function unwrapping for anyref-typed values.
Fixed: v8:12789
Change-Id: Ied57187bac7fde0326634f7b4fc428ad21dc9c2f
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3605231
Auto-Submit: Jakob Kummerow <jkummerow@chromium.org>
Reviewed-by: Andreas Haas <ahaas@chromium.org>
Commit-Queue: Andreas Haas <ahaas@chromium.org>
Cr-Commit-Position: refs/heads/main@{#80179}
- For simplicity we call a builtin when allocating a number.
- Elision of boxing/unboxing nodes will be done in a followup CL.
Bug: v8:7700
Change-Id: Iec4422d84c6597d3369ab512a1662adb0f077c98
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3602514
Reviewed-by: Leszek Swirski <leszeks@chromium.org>
Commit-Queue: Victor Gomes <victorgomes@chromium.org>
Auto-Submit: Victor Gomes <victorgomes@chromium.org>
Cr-Commit-Position: refs/heads/main@{#80178}
.. when deoptimization occurs after the outermost loop containing the
loop that triggered OSR compilation. The reasoning is that the main
benefit of OSR'd code is speeding up the OSR'd loop; the speedup of
the OSR'd loop is assumed to be higher than deoptimization overhead.
This is a slightly modified version of crrev.com/c/3521361, credit
goes to tao.pan@intel.com for most of the investigation and
implementation work.
Bug: v8:12161
Change-Id: Ie729dd5d1df9c7f529a1cf1b9471bb60ce76c41a
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3607988
Reviewed-by: Leszek Swirski <leszeks@chromium.org>
Commit-Queue: Jakob Linke <jgruber@chromium.org>
Cr-Commit-Position: refs/heads/main@{#80177}
Table.set has two arguments, the table index and the value. Therefore
Table.set was defined with a length of 2. However, the value argument is
optional, so the length should actually be 1.
Change-Id: Ica2ea13a8e78c974cb011df2b5dc99f8e7eb4bcd
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3398496
Reviewed-by: Thibaud Michaud <thibaudm@chromium.org>
Commit-Queue: Andreas Haas <ahaas@chromium.org>
Cr-Commit-Position: refs/heads/main@{#80176}
When setting a range for a newly allocated lab, consider adjacent cards.
If either is young, don't mark it as kMixed.
Bug: chromium:1029379
Change-Id: If7d1d920dd5769679de68800eae61f3a8dc1eb17
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3584116
Reviewed-by: Michael Lippautz <mlippautz@chromium.org>
Commit-Queue: Anton Bikineev <bikineev@chromium.org>
Cr-Commit-Position: refs/heads/main@{#80175}
This reverts commit 5d235def26.
Reason for revert: Speculative revert because of https://ci.chromium.org/ui/p/v8/builders/ci/V8%20Linux%20-%20gc%20stress/38153/overview
Original change's description:
> [heap] Store size with invalidated object
>
> When updating pointers during a full GC, a page might not be swept
> already. In such cases there might be invalid objects and slots recorded
> in free memory. Updating tagged slots in free memory is fine even though
> it is superfluous work.
>
> However, the GC also needs to calculate the size of potentially dead
> invalid objects in order to be able to check whether a slot is within
> that object. But since that object is dead, its map might be dead as
> well which makes size calculation impossible on such objects. The CL
> changes this to cache the size of invalid objects. A follow-up CL will
> also check the marking bit of invalid objects.
>
> Bug: v8:12578, chromium:1316289
> Change-Id: Ie773d0862a565982957e0dc409630d76552d1a32
> Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3599482
> Reviewed-by: Michael Lippautz <mlippautz@chromium.org>
> Reviewed-by: Jakob Linke <jgruber@chromium.org>
> Reviewed-by: Patrick Thier <pthier@chromium.org>
> Commit-Queue: Dominik Inführ <dinfuehr@chromium.org>
> Cr-Commit-Position: refs/heads/main@{#80169}
Bug: v8:12578, chromium:1316289
Change-Id: I6949412c5d6e1aa15718d027043d9528137a60a0
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3605812
Auto-Submit: Nico Hartmann <nicohartmann@chromium.org>
Owners-Override: Nico Hartmann <nicohartmann@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@{#80173}
When updating pointers during a full GC, a page might not be swept
already. In such cases there might be invalid objects and slots recorded
in free memory. Updating tagged slots in free memory is fine even though
it is superfluous work.
However, the GC also needs to calculate the size of potentially dead
invalid objects in order to be able to check whether a slot is within
that object. But since that object is dead, its map might be dead as
well which makes size calculation impossible on such objects. The CL
changes this to cache the size of invalid objects. A follow-up CL will
also check the marking bit of invalid objects.
Bug: v8:12578, chromium:1316289
Change-Id: Ie773d0862a565982957e0dc409630d76552d1a32
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3599482
Reviewed-by: Michael Lippautz <mlippautz@chromium.org>
Reviewed-by: Jakob Linke <jgruber@chromium.org>
Reviewed-by: Patrick Thier <pthier@chromium.org>
Commit-Queue: Dominik Inführ <dinfuehr@chromium.org>
Cr-Commit-Position: refs/heads/main@{#80169}
This is a reland of commit 91da38831d
Fixed: Use an X register for JumpIfCodeTIsMarkedForDeoptimization
on arm64.
Original change's description:
> [osr] Use the new OSR cache
>
> This CL switches over our OSR system to be based on the feedback
> vector osr caches.
>
> - OSRing to Sparkplug is fully separated from OSR urgency. If
> SP code exists, we simply jump to it, no need to maintain an
> installation request.
> - Each JumpLoop checks its dedicated FeedbackVector cache slot.
> If a valid target code object exists, we enter it *without*
> calling into runtime to fetch the code object.
> - Finally, OSR urgency still remains as the heuristic for
> requesting Turbofan OSR compile jobs. Note it no longer has a
> double purpose of being a generic untargeted installation
> request.
>
> With the new system in place, we can remove now-unnecessary
> hacks:
>
> - Early OSR tierup is replaced by the standard OSR system. Any
> present OSR code is automatically entered.
> - The synchronous OSR compilation fallback is removed. With
> precise installation (= per-JumpLoop-bytecode) we no longer
> have the problem of 'getting unlucky' with JumpLoop/cache entry
> mismatches. Execution has moved on while compiling? Simply spawn
> a new concurrent compile job.
> - Remove the synchronous (non-OSR) Turbofan compile request now
> that we always enter available OSR code as early as possible.
> - Tiering into Sparkplug no longer messes with OSR state.
>
> Bug: v8:12161
> Change-Id: I0a85e53d363504b7dac174dbaf69c03c35e66700
> Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3596167
> Commit-Queue: Jakob Linke <jgruber@chromium.org>
> Auto-Submit: Jakob Linke <jgruber@chromium.org>
> Reviewed-by: Leszek Swirski <leszeks@chromium.org>
> Commit-Queue: Leszek Swirski <leszeks@chromium.org>
> Cr-Commit-Position: refs/heads/main@{#80147}
Bug: v8:12161
Change-Id: Ib3597cf1d99cdb5d0f2c5ac18e311914f376231d
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3606232
Auto-Submit: Jakob Linke <jgruber@chromium.org>
Reviewed-by: Leszek Swirski <leszeks@chromium.org>
Commit-Queue: Leszek Swirski <leszeks@chromium.org>
Cr-Commit-Position: refs/heads/main@{#80167}
The size of flag is now 16 bits.
Bug: v8:12161
Change-Id: I5db5e05171281f27cce739c7b76e1d4b9ebf20b9
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3602236
Reviewed-by: Jakob Linke <jgruber@chromium.org>
Commit-Queue: Yu Liu <liuyu@loongson.cn>
Cr-Commit-Position: refs/heads/main@{#80165}
Make the file names consistent on "shadow-realm" (i.e. singular, with a
dash).
Bug: v8:11989
Change-Id: Id0a6f417fd9b53b9f7ddf9677da7396fa2481af6
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3606392
Reviewed-by: Adam Klein <adamk@chromium.org>
Commit-Queue: Shu-yu Guo <syg@chromium.org>
Cr-Commit-Position: refs/heads/main@{#80164}
Currently the Isolate is gotten off of the object that the operation is
being performed on. Shared objects return the shared Isolate, which is
incorrect as it shouldn't be used to run JS, nor does it have
HandleScopes open. Plumb the executing Isolate through.
Bug: v8:12547
Change-Id: I3d960751c798ac657a6122598154e36d9d504c31
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3606489
Reviewed-by: Adam Klein <adamk@chromium.org>
Commit-Queue: Shu-yu Guo <syg@chromium.org>
Cr-Commit-Position: refs/heads/main@{#80163}
This CL extends BoundedPageAllocator with PageFreeingMode parameter
which controls how pages should be freed: by setting permissions to
kNoAccess (preferred) or by discarding pages (Apple Silicon specific
behavior for RWX pages). The latter mode allows to ensure that once
pages are configured with RWX permissions they are never reconfigured
to anything else again.
The new mode will be used in a follow-up CL.
Bug: v8:12797
Change-Id: I3277f56ea6fee9c9b38b1682e68c22e66e9a02a4
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3606228
Reviewed-by: Michael Lippautz <mlippautz@chromium.org>
Commit-Queue: Igor Sheludko <ishell@chromium.org>
Cr-Commit-Position: refs/heads/main@{#80162}
Young GlobalHandle nodes were processed during post processing which
required a logic to defend against recursive GCs. This was originally
designed in a delayed way as the list itself was only treated as an
optimization and could thus be lazily cleared.
Since this still happens in the atomic pause and every node needs to
be visited, there's no advantage over just keeping it correct at all
times.
Bug: chromium:1319213
Change-Id: I199fc0be5e4b6ed5cbb60cf3b1452e37108cd4f8
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3605281
Reviewed-by: Dominik Inführ <dinfuehr@chromium.org>
Commit-Queue: Michael Lippautz <mlippautz@chromium.org>
Cr-Commit-Position: refs/heads/main@{#80161}
Make sure that Call nodes are only added to the graph after their
arguments are processed. Previously we were already doing this for the
undefined constant, but forgot to also do it for tagging nodes.
Now rather than trying to add those nodes before creating the Call node,
we create the Call node without adding it to the graph, then add it only
after setting up its inputs.
Bug: v8:7700
Change-Id: Id8c4c381f42fdd3c86d19d0fa2eb57163771060b
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3605248
Commit-Queue: Leszek Swirski <leszeks@chromium.org>
Auto-Submit: Leszek Swirski <leszeks@chromium.org>
Reviewed-by: Victor Gomes <victorgomes@chromium.org>
Commit-Queue: Victor Gomes <victorgomes@chromium.org>
Cr-Commit-Position: refs/heads/main@{#80159}
Correct use the lazy deopt's deopt index in safepoints, instead of the
index in the translation array.
Bug: v8:7700
Change-Id: I4a99ceb8bb54f2fba5bee0e6ddd6629b44439ce1
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3605609
Commit-Queue: Leszek Swirski <leszeks@chromium.org>
Auto-Submit: Leszek Swirski <leszeks@chromium.org>
Reviewed-by: Victor Gomes <victorgomes@chromium.org>
Commit-Queue: Victor Gomes <victorgomes@chromium.org>
Cr-Commit-Position: refs/heads/main@{#80157}
Invoking a JSBoundFunction is slightly slower than a normal
JSFunction (since we don't have to first jump to the
target_function).
The closure steps in SourceTextModule ExecuteAsyncModule is
controlled by the engine so it is better to create dedicated
context slots for the captured values.
Change-Id: I8163fc4b302d6d22906e578164470c9e28e768e0
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3584601
Reviewed-by: Camillo Bruni <cbruni@chromium.org>
Reviewed-by: Toon Verwaest <verwaest@chromium.org>
Commit-Queue: Chengzhong Wu <legendecas@gmail.com>
Cr-Commit-Position: refs/heads/main@{#80155}
gcc version 8.4 is currently throwing the following compilation
error:
```
error: variable 'first' set but not used
```
Change-Id: I25e1197b8deaa0089adee2e736448976705c7497
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3602060
Reviewed-by: Nico Hartmann <nicohartmann@chromium.org>
Commit-Queue: Milad Farazmand <mfarazma@redhat.com>
Cr-Commit-Position: refs/heads/main@{#80153}
Once we start protecting assembler buffers via PKU, the background
compile threads need to enable write permissions during compilation in
order to access them.
This CL adds that, with a little trick to allow passing {nullptr} to the
{CodeSpaceWriteScope} for configurations where we do not actually use
the provided {NativeModule}. This is needed because the background
compile threads do not keep the {NativeModule} alive during compilation,
so they also should not pass the potentially dangling pointer to the
{CodeSpaceWriteScope}.
R=thibaudm@chromium.org
Bug: v8:12809
Change-Id: I1a1d4d41239a6b4497f8a12fb252d7da76e72e1a
Cq-Include-Trybots: luci.v8.try:v8_linux64_ubsan_rel_ng
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3593117
Reviewed-by: Thibaud Michaud <thibaudm@chromium.org>
Commit-Queue: Clemens Backes <clemensb@chromium.org>
Cr-Commit-Position: refs/heads/main@{#80152}
Additionally this CL adds some #if V8_ENABLE_WEBASSEMBLY
that were missing.
Change-Id: I7c840ebfa800ad62f0b6f075c997bd1989a007e2
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3596169
Reviewed-by: Victor Gomes <victorgomes@chromium.org>
Reviewed-by: Thibaud Michaud <thibaudm@chromium.org>
Commit-Queue: Andreas Haas <ahaas@chromium.org>
Reviewed-by: Clemens Backes <clemensb@chromium.org>
Cr-Commit-Position: refs/heads/main@{#80151}
The bit is immutable and set during construction of the meta data. The
main thread needs to query this bit while the concurrent marker may be
running.
Bug: chromium:1319210
Change-Id: I811d8048a61a3c682245b34302b6078d91f9662a
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3599489
Reviewed-by: Dominik Inführ <dinfuehr@chromium.org>
Commit-Queue: Michael Lippautz <mlippautz@chromium.org>
Cr-Commit-Position: refs/heads/main@{#80149}
This reverts commit 91da38831d.
Reason for revert: https://ci.chromium.org/ui/p/v8/builders/ci/V8%20Linux64%20-%20arm64%20-%20sim%20-%20pointer%20compression%20-%20builder/21150/overview
Original change's description:
> [osr] Use the new OSR cache
>
> This CL switches over our OSR system to be based on the feedback
> vector osr caches.
>
> - OSRing to Sparkplug is fully separated from OSR urgency. If
> SP code exists, we simply jump to it, no need to maintain an
> installation request.
> - Each JumpLoop checks its dedicated FeedbackVector cache slot.
> If a valid target code object exists, we enter it *without*
> calling into runtime to fetch the code object.
> - Finally, OSR urgency still remains as the heuristic for
> requesting Turbofan OSR compile jobs. Note it no longer has a
> double purpose of being a generic untargeted installation
> request.
>
> With the new system in place, we can remove now-unnecessary
> hacks:
>
> - Early OSR tierup is replaced by the standard OSR system. Any
> present OSR code is automatically entered.
> - The synchronous OSR compilation fallback is removed. With
> precise installation (= per-JumpLoop-bytecode) we no longer
> have the problem of 'getting unlucky' with JumpLoop/cache entry
> mismatches. Execution has moved on while compiling? Simply spawn
> a new concurrent compile job.
> - Remove the synchronous (non-OSR) Turbofan compile request now
> that we always enter available OSR code as early as possible.
> - Tiering into Sparkplug no longer messes with OSR state.
>
> Bug: v8:12161
> Change-Id: I0a85e53d363504b7dac174dbaf69c03c35e66700
> Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3596167
> Commit-Queue: Jakob Linke <jgruber@chromium.org>
> Auto-Submit: Jakob Linke <jgruber@chromium.org>
> Reviewed-by: Leszek Swirski <leszeks@chromium.org>
> Commit-Queue: Leszek Swirski <leszeks@chromium.org>
> Cr-Commit-Position: refs/heads/main@{#80147}
Bug: v8:12161
Change-Id: I4a6955f4f20b6f3b13e98d5600c7c6a5205915bc
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3605608
Auto-Submit: Nico Hartmann <nicohartmann@chromium.org>
Owners-Override: Nico Hartmann <nicohartmann@chromium.org>
Reviewed-by: Nico Hartmann <nicohartmann@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@{#80148}
This CL switches over our OSR system to be based on the feedback
vector osr caches.
- OSRing to Sparkplug is fully separated from OSR urgency. If
SP code exists, we simply jump to it, no need to maintain an
installation request.
- Each JumpLoop checks its dedicated FeedbackVector cache slot.
If a valid target code object exists, we enter it *without*
calling into runtime to fetch the code object.
- Finally, OSR urgency still remains as the heuristic for
requesting Turbofan OSR compile jobs. Note it no longer has a
double purpose of being a generic untargeted installation
request.
With the new system in place, we can remove now-unnecessary
hacks:
- Early OSR tierup is replaced by the standard OSR system. Any
present OSR code is automatically entered.
- The synchronous OSR compilation fallback is removed. With
precise installation (= per-JumpLoop-bytecode) we no longer
have the problem of 'getting unlucky' with JumpLoop/cache entry
mismatches. Execution has moved on while compiling? Simply spawn
a new concurrent compile job.
- Remove the synchronous (non-OSR) Turbofan compile request now
that we always enter available OSR code as early as possible.
- Tiering into Sparkplug no longer messes with OSR state.
Bug: v8:12161
Change-Id: I0a85e53d363504b7dac174dbaf69c03c35e66700
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3596167
Commit-Queue: Jakob Linke <jgruber@chromium.org>
Auto-Submit: Jakob Linke <jgruber@chromium.org>
Reviewed-by: Leszek Swirski <leszeks@chromium.org>
Commit-Queue: Leszek Swirski <leszeks@chromium.org>
Cr-Commit-Position: refs/heads/main@{#80147}
This introduces an AssemblerBufferCache class which will (later) cache
the backing store of AssemblerBuffers. This is needed for PKU-protected
assembler buffers, which are expensive to allocate and deallocate.
For now, the AssemblerBufferCache does not do any caching, this will be
added in a follow-up CL.
R=thibaudm@chromium.org
Bug: v8:12809
Change-Id: I4a7ccff49c9930584a9fcda8899cfe38cfc61419
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3593136
Reviewed-by: Thibaud Michaud <thibaudm@chromium.org>
Commit-Queue: Clemens Backes <clemensb@chromium.org>
Cr-Commit-Position: refs/heads/main@{#80145}
This assumes that Phi nodes never point to untagged values and
TemporaryRegisters are always general registers.
Bug: v8:7700
Change-Id: I74a6c43ff9f1ba87dd258e90a193f683d666b8ec
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3598883
Reviewed-by: Leszek Swirski <leszeks@chromium.org>
Commit-Queue: Victor Gomes <victorgomes@chromium.org>
Cr-Commit-Position: refs/heads/main@{#80144}
The `set disable-randomization off` command may fail on some platforms,
such as the `rr` debugger. We can just ignore the error and carry on.
Change-Id: I9b8dae183a9852178a3d3411172bf3aef173c995
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3602519
Auto-Submit: Pierre Langlois <pierre.langlois@arm.com>
Reviewed-by: Clemens Backes <clemensb@chromium.org>
Commit-Queue: Clemens Backes <clemensb@chromium.org>
Cr-Commit-Position: refs/heads/main@{#80143}
We need to use MachineRepresentation to properly distinguish
the types in compiler::UnallocatedOperand.
Bug: v8:7700
Change-Id: I4273512a00290bb85b09aeb3788643e346be03f7
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3602515
Reviewed-by: Leszek Swirski <leszeks@chromium.org>
Auto-Submit: Victor Gomes <victorgomes@chromium.org>
Commit-Queue: Victor Gomes <victorgomes@chromium.org>
Cr-Commit-Position: refs/heads/main@{#80140}
For call_direct, feedback (for the first level of inlining) only
records call counts, not the statically-known target. So to be
able to make feedback for potential additional inlining levels
available, the feedback processor must look to the wire bytes
to extract the call targets. Without feedback and hence unknown
call counts, such multi-level inlining would not happen.
Bug: v8:12166
Change-Id: I84ca58019e927a8bf9dad4e4aceddd341f945c04
Reviewed-on: https://chromium-review.googlesource.com/c/v8/v8/+/3579105
Reviewed-by: Manos Koukoutos <manoskouk@chromium.org>
Commit-Queue: Jakob Kummerow <jkummerow@chromium.org>
Cr-Commit-Position: refs/heads/main@{#80139}