Also do the same for the unsafe version of StoreIndexed.
Similar to the inliner, we may replace calls in the specializers with
these non-definition instructions, so we make variants that allow an
unrelated value to be used for the old users of the call site value. (In
all cases, this ends up being null, since the original calls wouldn't
have returned a useable value anyway.)
Change-Id: I0befd54920c4c0e63e086b804860ee4c9a7d3767
Reviewed-on: https://dart-review.googlesource.com/c/89585
Reviewed-by: Martin Kustermann <kustermann@google.com>
Commit-Queue: Stevie Strickland <sstrickl@google.com>
This reverts commit a8777149a7.
Reason for revert: benchmarking infrastructure reports broken benchmarks on ARM.
Original change's description:
> [VM/Compiler] Revoke definition status of certain instructions.
>
> Change StoreIndexed, StoreIndexedUnsafe, and StoreInstanceField
> to be non-definition instructions. Fix up the flow graph builders
> so they no longer treat StoreIndexed and StoreIndexedUnsafe as
> pushing values on the stack that need to be dropped.
>
> We recently changed FlowGraphInliner::TryInlineRecognizedMethod
> to create a distinction between the last instruction to be inlined
> and the definition that should be used as the new result. Perform
> the same separation of replacement code and result in the call
> specializers.
>
> Change-Id: I66c7ab60a662dcca05092e47156d1d4ba809d51f
> Reviewed-on: https://dart-review.googlesource.com/c/88326
> Commit-Queue: Stevie Strickland <sstrickl@google.com>
> Reviewed-by: Martin Kustermann <kustermann@google.com>
TBR=kustermann@google.com,sstrickl@google.com
Change-Id: I9c666d80a69ba6a5c37b421aae1ed83502694f19
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Reviewed-on: https://dart-review.googlesource.com/c/89366
Reviewed-by: Vyacheslav Egorov <vegorov@google.com>
Commit-Queue: Vyacheslav Egorov <vegorov@google.com>
Change StoreIndexed, StoreIndexedUnsafe, and StoreInstanceField
to be non-definition instructions. Fix up the flow graph builders
so they no longer treat StoreIndexed and StoreIndexedUnsafe as
pushing values on the stack that need to be dropped.
We recently changed FlowGraphInliner::TryInlineRecognizedMethod
to create a distinction between the last instruction to be inlined
and the definition that should be used as the new result. Perform
the same separation of replacement code and result in the call
specializers.
Change-Id: I66c7ab60a662dcca05092e47156d1d4ba809d51f
Reviewed-on: https://dart-review.googlesource.com/c/88326
Commit-Queue: Stevie Strickland <sstrickl@google.com>
Reviewed-by: Martin Kustermann <kustermann@google.com>
- Introduce a slimmed down version of thread.h, which just depends on the
Zone and StackResource.
- Introduce a layering check that would prevent the coupling in the future.
This is the first step towards decoupling compiler from runtime.
There are multiple reasons to introduce the decoupling but the main
reason currently is to introduce a controlled surface through which
compiler reaches into runtime to catch any places where runtime word size
might influence the compiler and then enable building compiler that
targets 32-bit runtime but is embedded into a 64-bit runtime.
Issue https://github.com/dart-lang/sdk/issues/31709
Change-Id: Id63ebbaddca55dd097298e51c90d957a73fa476e
Reviewed-on: https://dart-review.googlesource.com/c/87182
Commit-Queue: Vyacheslav Egorov <vegorov@google.com>
Reviewed-by: Martin Kustermann <kustermann@google.com>
Rationale:
With the new induction variable bounds in place, this
is the first very low hanging fruit. When a loop bound
matches the array bounds check, it is redundant. This
removes AOT non-speculative bounds checks for the first
time. More improvements to follow!
Change-Id: I5dac885a1ef6fdb64b5844a0617d2fe9d800d9d6
Reviewed-on: https://dart-review.googlesource.com/c/87940
Commit-Queue: Aart Bik <ajcbik@google.com>
Reviewed-by: Vyacheslav Egorov <vegorov@google.com>
The inliner assumes that the last instruction replacing the original
call is a definition so it can replace all the uses of the call.
Unfortunately, we have a couple of instructions (StoreIndexed and
StoreInstanceField) that are not _really_ definitions.
Change FlowGraphInliner::TryInlineRecognizedMethod to create a
distinction between the last instruction to be inlined and the
definition that should be used as the new result. For inlines that end
in a StoreIndexed or StoreInstanceField, use the null constant as the
new result.
Further work towards a better fix for https://github.com/dart-lang/sdk/issues/35481
Change-Id: I5a9effe44388de9f03681565b1de2cf2c46e91c6
Reviewed-on: https://dart-review.googlesource.com/c/88960
Commit-Queue: Stevie Strickland <sstrickl@google.com>
Reviewed-by: Martin Kustermann <kustermann@google.com>
On arm64 Code.UncheckedEntryPoint() returned the wrong value because it
is based on the offset of the Label of FunctionEntryInstr.
That works on X64/ARM, but not on ARM64, because there the monomorphic
entry and frame setup is emitted in FlowGraphCompiler::CompileGraph().
Change-Id: I9c1b1be71997b83d76b385d5b811a665015c32b5
Reviewed-on: https://dart-review.googlesource.com/c/88708
Reviewed-by: Vyacheslav Egorov <vegorov@google.com>
Commit-Queue: Martin Kustermann <kustermann@google.com>
There might be a race between background compiler and mutator where
mutator changes guarded state of the field after Slot was created from
it.
A situation is possible where we have a clone of a field with its
guarded state set to unknown, however Slot::Get for this field returns
a Slot created from the previous clone of the same field with a known
guarded state. In this case we must add *old* clone from which the Slot
was created to guarded fields and not the new clone, because new clone
has no guarded state to begin with and thus
ParsedFunction::AddToGuardedFields(...) would simply ignore it.
Such slots with inconsistent guarded state that are not in the current
list of guarded fields arise due to unsuccessful inlining attempts.
If we built and discard the graph, then guarded fields associated with
that graph are also discarded. However the slot itself stays behind in
the compilation global cache.
TEST=vm/cc/SlotFromGuardedField
Bug: b/121271056
Change-Id: Ib6ee5ec4922c033c3a71bdc46da74ead47b1edd0
Reviewed-on: https://dart-review.googlesource.com/c/88575
Commit-Queue: Vyacheslav Egorov <vegorov@google.com>
Reviewed-by: Martin Kustermann <kustermann@google.com>
In some cases, we could replace a binding instruction like a
InstanceCall with a subgraph that would end with a non-binding
instruction like a StoreInstanceFieldExpr. When the inliner attempted
to replace the uses of the former, the lack of an SSA temp in the latter
would cause weird results.
Change the code to check that the new replacement has an SSA temp, and
if it doesn't, use the temp assigned to the null constant in the flow
graph.
Closes https://github.com/dart-lang/sdk/issues/35481
Change-Id: Id90717b6a906703fb442f9cd0cacc3bc14fb4887
Reviewed-on: https://dart-review.googlesource.com/c/88321
Commit-Queue: Stevie Strickland <sstrickl@google.com>
Reviewed-by: Martin Kustermann <kustermann@google.com>
In Dart2, the type relation "is more specific than" is no different than "is
subtype of". Since Dart1 is not supported anymore, type checking in the runtime
can be simplified and cleaned up.
Change-Id: I8e95b9e47146315c9a6f6c29b732ecbeb4e9ae97
Reviewed-on: https://dart-review.googlesource.com/c/88221
Commit-Queue: Régis Crelier <regis@google.com>
Reviewed-by: Siva Annamalai <asiva@google.com>
Rationale:
Finds general bounds on an induction variable by
looking at loop exits that use strict comparisons.
An intended client is bounds-check-elimination (AOT).
Change-Id: Icfd6481aad9669c15b78630e4322667cfce7c1ef
Reviewed-on: https://dart-review.googlesource.com/c/87485
Commit-Queue: Aart Bik <ajcbik@google.com>
Reviewed-by: Alexander Markov <alexmarkov@google.com>
CanonicalType is a leftover method/field from the time classes were keeping
track of their own canonical types and is still used for non-generic classes.
DeclarationType is more general and can be used where CanonicalType is used.
The field canonical_type_ is renamed to declaration_type_ and now caches the
DeclarationType.
Change-Id: I585c7bc2d3ae760a4a0e783bc77fb8d8efdbf682
Reviewed-on: https://dart-review.googlesource.com/c/87482
Commit-Queue: Régis Crelier <regis@google.com>
Reviewed-by: Alexander Markov <alexmarkov@google.com>
If the --use-bare-instructions flag is enabled we will:
* Make call sites load the target directly from the pool (instead of
the code object) - this saves one instruction (and an indirect load)
* Ensure the object pool will have direct entry addresses by:
- Letting the clustered snapshot reader change any StubCode::UnlinkedCall()
in the object pool by it's monomorphic entry
- Change the code patcher to patch SwitchableCalls by writing the
monomorphic entry into the pool (instead of the code object)
Issue https://github.com/dart-lang/sdk/issues/33274
Change-Id: I4e41fc8e4461bde477cc559a6a4fccaaf3a350b5
Reviewed-on: https://dart-review.googlesource.com/c/86160
Commit-Queue: Martin Kustermann <kustermann@google.com>
Reviewed-by: Ryan Macnak <rmacnak@google.com>
Reviewed-by: Vyacheslav Egorov <vegorov@google.com>
This is the final CL which adds a new --use-bare-instructions flag to
the VM.
If this flag is set during AOT compilation, we will:
* Build one global object pool (abbr: GOP) which all code objects
share. This gop will be stored in the object store. The PP register
is populated in the enter dart stub and it is restored when
returning from native calls.
* Gets rid of the CODE_REG/PP slots from the dart frames. Instead the
compiled code uses the global object pool, which is always in PP.
* Starts emitting pc-relative calls for calls between two dart
functions or when invoking a stub.
Limitation: We only emit pc-relative calls between two code objects
in the same isolate (this is because the image writer is writing
instruction objects for vm-isolate/main-isolate seperately)
* We do compile-time relocation of those static calls after the
precompiler has finished its work, but before writing the snapshot.
This patches all the instruction objects with pc-relative calls to
have the right .text distance.
* We emit a sorted list of code objects in ObjectStore::reverse_code_table,
which will be used by the AOT runtime to go back from PC to Code
objects (where all metadata, e.g. stack maps, catch entry moves, pc
descriptors are available).
Issue https://github.com/dart-lang/sdk/issues/33274
Change-Id: I6c5dd2b1571e3a889b27e804a24c2986c71e03b6
Reviewed-on: https://dart-review.googlesource.com/c/85769
Commit-Queue: Martin Kustermann <kustermann@google.com>
Reviewed-by: Ryan Macnak <rmacnak@google.com>
Reviewed-by: Vyacheslav Egorov <vegorov@google.com>
Rationale:
Having a literal null checked by CheckNull with
a subsequent Redefinition resulted in some strange
unboxing that crashed AOT (fix courtesy Slava!).
Note that we still have some ambiguity around
adding/removing environments from instructions
that may deoptimize, but this change fixes all
prior related DartFuzz failures.
https://github.com/dart-lang/sdk/issues/35335
Change-Id: Ifb50d8cddf93e57758b2bbb83ad397ea281e9307
Reviewed-on: https://dart-review.googlesource.com/c/87280
Reviewed-by: Vyacheslav Egorov <vegorov@google.com>
Commit-Queue: Aart Bik <ajcbik@google.com>
In Dart2, mixin application classes are generated by kernel and type resolution
is performed by kernel. In Dart1, using the VM parser, these tasks were done
during class and type finalization.
Change-Id: I927b1446d906ffd76ed9591c96e0d974df09f3bf
Reviewed-on: https://dart-review.googlesource.com/c/87025
Commit-Queue: Régis Crelier <regis@google.com>
Reviewed-by: Ryan Macnak <rmacnak@google.com>
- We can now skip checks on tearoffs of functions called only via tearoffs.
- We don't always insert a check when intrinsifying the an unchecked call to set indexed.
- We only build the unchecked entry-point of regular methods when inlining to reduce code size.
Change-Id: I2e6a647409130385cda03906e734a3c161d56ab4
Cq-Include-Trybots: luci.dart.try:vm-kernel-win-release-x64-try, vm-kernel-optcounter-threshold-linux-release-x64-try, vm-kernel-precomp-linux-debug-x64-try, vm-kernel-precomp-linux-release-simarm-try, vm-kernel-precomp-linux-release-simarm64-try, vm-kernel-precomp-linux-release-x64-try, vm-kernel-precomp-win-release-x64-try
Reviewed-on: https://dart-review.googlesource.com/c/86562
Commit-Queue: Samir Jindel <sjindel@google.com>
Reviewed-by: Vyacheslav Egorov <vegorov@google.com>
Rationale:
This avoids rewriting
v1 <- Constant(#1)
v2 <- BoxInt64:1402(v1)
v3 <- UnboxInt64(v2)
into the non-existing
v3 <- UnboxedIntConverter(tagged->int64, v1)
Note:
Arguably the input was incorrect (v1 should have been
an unboxed constant), but we allow type sloppy expressions
at a lot of places that are repaired later on, as for
this case after the fix. See the bug for more details.
https://github.com/dart-lang/sdk/issues/35371
Change-Id: I719124ba150a1c72fe6819a1edd8380847fd82c2
Reviewed-on: https://dart-review.googlesource.com/c/86823
Reviewed-by: Ryan Macnak <rmacnak@google.com>
Commit-Queue: Aart Bik <ajcbik@google.com>
In Dart2, type bounds checking is either performed by the common front-end or
by explicitly generated code, but not by type finalization or runtime anymore,
as it was done in Dart1.
Consequently, the class BoundedType is not needed anymore, and malbounded or
malformed types are not seen by the runtime either.
Change-Id: I5d6e4c68d153d6730fa7ff7f6d9dcfa611299c16
Reviewed-on: https://dart-review.googlesource.com/c/86687
Commit-Queue: Régis Crelier <regis@google.com>
Reviewed-by: Ryan Macnak <rmacnak@google.com>
Reviewed-by: Samir Jindel <sjindel@google.com>
Rationale:
After discussion with Slava, it turns out that the ASSERT
taht calls IsRepresentable() actually violates the truncating()
property that was assumed when the IsRepresentable() function was
only used in executable code. The simplest solution, therefore,
is to simply accept the uint32 case too.
https://github.com/dart-lang/sdk/issues/35326
Change-Id: I56a05c1bc0b985019f7bd2fb04afdef8708ef7ea
Reviewed-on: https://dart-review.googlesource.com/c/86251
Reviewed-by: Alexander Markov <alexmarkov@google.com>
Reviewed-by: Vyacheslav Egorov <vegorov@google.com>
Commit-Queue: Aart Bik <ajcbik@google.com>
This change extends/fixes the exiting "pattern recognition" which tries
to recognize the pattern
v2 <- LoadClassIdInstr(v1)
BranchIf v2 == IntegerConstant(cid)
Furthermore we start inlining the recognized `ClassID.getID` method very
early in the pipeline. This allows the VM to recognize the above
pattern and insert redefinitions before the actual inlining pass.
Furthermore we special-case two very hot methods in utf8 decoding by
manually having two loops, one of which is guarded by a class-id check
against the _Uint8ArrayView class, which is most common. (In the future
we would like to unify the typed data layouts so we no longer need to
use `ClassId.getID`, thereby also allowing non core library code to use
this).
This improves dart-aot by
* 31%+ for a protobuf decoding benchmark we care about
Issue https://github.com/dart-lang/sdk/issues/31954
Change-Id: I7181bbf096aabe303634fd3b2bff9cc96d69719c
Reviewed-on: https://dart-review.googlesource.com/c/85443
Reviewed-by: Vyacheslav Egorov <vegorov@google.com>
Commit-Queue: Martin Kustermann <kustermann@google.com>
This reverts commit d7cf6959bb.
Reason for revert: This causes crashes on the vm-obfuscate builder at https://ci.chromium.org/p/dart/builders/luci.dart.ci.sandbox/vm-kernel-precomp-obfuscate-linux-release-x64
like
FAILED: dartkp-dart_precompiled release_x64 corelib_2/linked_hash_map_test
Expected: Pass
Actual: Crash
Unexpected compile error.
--- Command "vm_compile_to_kernel" (took 11.000189s):
DART_CONFIGURATION=ReleaseX64 /b/s/w/ir/cache/builder/sdk/pkg/vm/tool/gen_kernel --aot --platform=out/ReleaseX64/vm_platform_strong.dill -o /b/s/w/ir/cache/builder/sdk/out/ReleaseX64/generated_compilations/dartkp/tests_corelib_2_linked_hash_map_test/out.dill /b/s/w/ir/cache/builder/sdk/tests/corelib_2/linked_hash_map_test.dart --packages=/b/s/w/ir/cache/builder/sdk/.packages -Ddart.developer.causal_async_stacks=true
exit code:
0
--- Command "precompiler" (took 583ms):
DART_CONFIGURATION=ReleaseX64 out/ReleaseX64/gen_snapshot --snapshot-kind=app-aot-assembly --assembly=/b/s/w/ir/cache/builder/sdk/out/ReleaseX64/generated_compilations/dartkp/tests_corelib_2_linked_hash_map_test/out.S --obfuscate --sync-async --ignore-unrecognized-flags --packages=/b/s/w/ir/cache/builder/sdk/.packages /b/s/w/ir/cache/builder/sdk/out/ReleaseX64/generated_compilations/dartkp/tests_corelib_2_linked_hash_map_test/out.dill
exit code:
-6
stderr:
Warning: This VM has been configured to obfuscate symbol information which violates the Dart standard.
See dartbug.com/30524 for more information.
===== CRASH =====
version=2.2.0-edge.83712405657ff736033380cb24b0bd5a62fc3692 (Tue Nov 27 16:56:07 2018 +0000) on "linux_x64"
si_signo=Segmentation fault(11), si_code=1, si_addr=0xd1400050e5f
Dumping native stack trace for thread 2b71
[0x000056001fc426c7] Unknown symbol
[0x000056001fc426c7] Unknown symbol
[0x000056001fe5763d] Unknown symbol
[0x000056001fe4c99c] Unknown symbol
[0x000056001fe6166a] Unknown symbol
[0x000056001fe59c0c] Unknown symbol
Original change's description:
> [VM] Inline ClassID.getID() eagerly, extend pattern matching logic to recognize it, use it to special case ascii decoding
>
> This change extends/fixes the exiting "pattern recognition" which tries
> to recognize the pattern
>
> v2 <- LoadClassIdInstr(v1)
> BranchIf v2 == IntegerConstant(cid)
>
> Furthermore we start inlining the recognized `ClassID.getID` method very
> early in the pipeline. This allows the VM to recognize the above
> pattern and insert redefinitions before the actual inlining pass.
>
> Furthermore we special-case two very hot methods in utf8 decoding by
> manually having two loops, one of which is guarded by a class-id check
> against the _Uint8ArrayView class, which is most common. (In the future
> we would like to unify the typed data layouts so we no longer need to
> use `ClassId.getID`, thereby also allowing non core library code to use
> this).
>
> This improves dart-aot by
> * 31%+ for a protobuf decoding benchmark we care about
>
>
> Issue https://github.com/dart-lang/sdk/issues/31954
>
> Change-Id: Ia567b92b7e76ff28eda1726deaafda32732ed8f5
> Reviewed-on: https://dart-review.googlesource.com/c/85281
> Reviewed-by: Johnni Winther <johnniwinther@google.com>
> Reviewed-by: Vyacheslav Egorov <vegorov@google.com>
> Reviewed-by: Jenny Messerly <jmesserly@google.com>
> Commit-Queue: Martin Kustermann <kustermann@google.com>
TBR=vegorov@google.com,kustermann@google.com,jmesserly@google.com,johnniwinther@google.com,sra@google.com
Change-Id: I912b0768c32cbb00297ce48db29dbdbea44c14fa
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Reviewed-on: https://dart-review.googlesource.com/c/85441
Reviewed-by: William Hesse <whesse@google.com>
Commit-Queue: William Hesse <whesse@google.com>
This change extends/fixes the exiting "pattern recognition" which tries
to recognize the pattern
v2 <- LoadClassIdInstr(v1)
BranchIf v2 == IntegerConstant(cid)
Furthermore we start inlining the recognized `ClassID.getID` method very
early in the pipeline. This allows the VM to recognize the above
pattern and insert redefinitions before the actual inlining pass.
Furthermore we special-case two very hot methods in utf8 decoding by
manually having two loops, one of which is guarded by a class-id check
against the _Uint8ArrayView class, which is most common. (In the future
we would like to unify the typed data layouts so we no longer need to
use `ClassId.getID`, thereby also allowing non core library code to use
this).
This improves dart-aot by
* 31%+ for a protobuf decoding benchmark we care about
Issue https://github.com/dart-lang/sdk/issues/31954
Change-Id: Ia567b92b7e76ff28eda1726deaafda32732ed8f5
Reviewed-on: https://dart-review.googlesource.com/c/85281
Reviewed-by: Johnni Winther <johnniwinther@google.com>
Reviewed-by: Vyacheslav Egorov <vegorov@google.com>
Reviewed-by: Jenny Messerly <jmesserly@google.com>
Commit-Queue: Martin Kustermann <kustermann@google.com>
Rationale:
Replacing the simple induction analysis in the range analyzer
with the new framework for finding induction variables removes
a lot of code and even saves some memory in the IR (for phis).
Note that this is only the first phase, i.e. using the new
framework for the same behavior in the range analyzer. In the
second phase, the new framework will be fully exploited to
compute more accurate ranges in many more cases.
Change-Id: Iae5bd5a155ef8592558b5bdb2388c5aa9da64e59
Reviewed-on: https://dart-review.googlesource.com/c/84500
Reviewed-by: Vyacheslav Egorov <vegorov@google.com>
Commit-Queue: Aart Bik <ajcbik@google.com>
If we inline something inside a loop which might throw, the throwing blocks
are currently spliced in the middle of the loop. This CL moves those blocks
to the very end.
This improves a number of typed data benchmarks in dart-aot mode by 5-10%.
Issue https://github.com/dart-lang/sdk/issues/31954
Change-Id: I5dc86291240d8dac61798ff873ffa7205edc0007
Reviewed-on: https://dart-review.googlesource.com/c/85263
Commit-Queue: Martin Kustermann <kustermann@google.com>
Reviewed-by: Vyacheslav Egorov <vegorov@google.com>
A field/function annotated with this pragma must be guaranteed to not
return `null` at runtime.
Make use of this non-nullable annotation in the VM's type propagator.
Annotates the "_TypedListView._typedData" field to ensure the VM knows it
returns a non-nullable _TypedListView.
Furthermore annotates methods on the integer implementation. Those particular
methods are recognized methods with a "dynamic" return type. This caused
the type propagator to use CompileType::Dynamic() as result type. Since a
previous CL started to only utilize the annotated type if it is better than
"dynamic" more integer operations got handled in-line, though with null-checks.
Annotating those methods to return non-null improves the in-line handling of
integer operations.
This improves dart-aot
On arm7hf:
SHA256: +5%, SHA: +6%, JsonObjectRoundTrip: +7%, ...
On arm8:
SHA1: +28%, MD5: +25%, SHA256: +15%, TypedData.Int16ListViewBench: +18.5%, StringInterpolation: +18%, ...
Issue https://github.com/dart-lang/sdk/issues/31954
Issue https://github.com/dart-lang/sdk/issues/35154
Change-Id: Ia4263a37241a36c9dc35e8a48893297effa6f4b2
Reviewed-on: https://dart-review.googlesource.com/c/84421
Commit-Queue: Martin Kustermann <kustermann@google.com>
Reviewed-by: Vyacheslav Egorov <vegorov@google.com>
Reviewed-by: Alexander Markov <alexmarkov@google.com>
Until now a typed data view was first constructed and then afterwards the
constructor body validated that the offsetInBytes/length are non-null
and in range.
This CL moves the range check to the call sites of the constructor,
thereby guarenteeing that we will never initialize the
_TypedListView.{offsetInBytes,length} fields with non-_Smi values. (The
views are constructed only within the "dart:typed_data" library. Neither the
embedder nor the runtime construct them.)
Further more this CL removes the redundant _defaultIfNull() call, since
the call sites already handled the `length == null` case.
This CL also starts annotating _TypedListView.{offsetInBytes,length}
with @pragma("vm:exact-result-type", "dart:core#_Smi") to ensure any
[LoadFieldInstr] will have the right type attached, thereby avoiding any
null-checks and mint-handling.
This improves dart-aot
On arm7hf:
MD5: +38%, SHA256: +87%, SHA1: +89% (plus various typed data microbenchmarks)
JsonParseDefaultReviver/StringBuffer/StringIdiomatic/JsonParseCustomReviver: -5-8% (probably due to not inlining static calls, will find out)
On arm8:
MD5: +6.5%, SHA256: +12%, SHA1: 3.6%, JsonUtf8RoundTrip: 8% (plus various typed data microbenchmarks)
DeltaBlue: -6.7% (probably due to not inlining static calls, will find out)
Issue https://github.com/dart-lang/sdk/issues/35154
Issue https://github.com/dart-lang/sdk/issues/31954
Change-Id: I37c822e6879f5a2d17fd9650a68cf2eee4326b01
Reviewed-on: https://dart-review.googlesource.com/c/84241
Commit-Queue: Martin Kustermann <kustermann@google.com>
Reviewed-by: Alexander Markov <alexmarkov@google.com>
Reviewed-by: Vyacheslav Egorov <vegorov@google.com>
Currently the @pramga('vm:exact-result-type') annotation only works if
the function is a recognized method. This change changes that to make
the VM just look if a function has the annotation (no matter if it's
also in the list of recognized methods or not).
Furthermore this CL lets the type propgagator use
@pragma('vm:exact-result-type') annotations to narrow the [CompileType]
set on [LoadFieldInstr]s.
Since the @pragma is a general feature, this CL moves the
`Function::FindPragma()` to `Library::FindPragma` (which is where any
other metadata lookup happens). We also let the `FindPragma` accept any
of Class/Function/Field objects.
Furthermore the `FindPragma()` function is fixed to handle the case
when the evaluation of the metadata results in an error.
In this case we simply claim to not have found a pragma annotation.
Issue https://github.com/dart-lang/sdk/issues/31954
Change-Id: If03f566e334cd53549985823ee3dd6b5e9672969
Reviewed-on: https://dart-review.googlesource.com/c/85163
Commit-Queue: Martin Kustermann <kustermann@google.com>
Reviewed-by: Vyacheslav Egorov <vegorov@google.com>
This is a reland of 1a4cff1bd4
Original change's description:
> [VM] Remove unused flags
>
> Remove the following flags which are always ON
> --strong
> --reify_generic_functions
>
> and these flags which are not applicable anymore
> --error-on-bad-override
> --error-on-bad-type
> --enable-type-checks
>
> Change-Id: I6e1aeb68e663953f4ae49ca94ea0daa87b661900
> Reviewed-on: https://dart-review.googlesource.com/c/79431
> Commit-Queue: Siva Annamalai <asiva@google.com>
> Reviewed-by: Alexander Markov <alexmarkov@google.com>
> Reviewed-by: Régis Crelier <regis@google.com>
Change-Id: Ib4452c22fb05eed38820708abb648939e03b1df0
Reviewed-on: https://dart-review.googlesource.com/c/84923
Reviewed-by: Siva Annamalai <asiva@google.com>
Remove the following flags which are always ON
--strong
--reify_generic_functions
and these flags which are not applicable anymore
--error-on-bad-override
--error-on-bad-type
--enable-type-checks
Change-Id: I6e1aeb68e663953f4ae49ca94ea0daa87b661900
Reviewed-on: https://dart-review.googlesource.com/c/79431
Commit-Queue: Siva Annamalai <asiva@google.com>
Reviewed-by: Alexander Markov <alexmarkov@google.com>
Reviewed-by: Régis Crelier <regis@google.com>
Ryan Macnak