This is a Dart-tailored implementation of the "row-displacement
dispatch table" technique for closed-world instance calls:
All interface targets in the program are grouped into selectors such
that all targets that could potentially be called from the same call
site have the same selector (currently just grouped by name).
Each selector is assigned a selector offset such that offset + classid
is unique for all selector/classid combinations where the class
implements the selector.
At every instance call site that has an interface target (i.e. where
the static type of the receiver is not dynamic), the selector offset +
receiver classid is computed and used as index into a global table of
entry points.
If the receiver can be null (as determined by the front-end TFA and the
VM type propagation), a null check is inserted before the call.
An arguments descriptor is provided (only) for selectors that need it
(those which have type parameters or optional/named parameters).
The dispatch table calls don't need the monomorphic entry code, so for
functions that are only called via dispatch table calls (i.e. never
called dynamically), the monomorphic entry code is left out.
Some future improvements to the table dispatch implementation are
mentioned in https://github.com/dart-lang/sdk/issues/40188
The table dispatch flag is disabled by default in this commit. A
separate commit enables the flag.
Change-Id: Ic2911742b4a2c9a8d3bc7df60605454cbe4c0714
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/126648
Commit-Queue: Aske Simon Christensen <askesc@google.com>
Reviewed-by: Martin Kustermann <kustermann@google.com>
Reviewed-by: Alexander Markov <alexmarkov@google.com>
This reverts commit 9eb531bde4.
Reason for revert: Bots are red. Some tests are failing.
https://ci.chromium.org/p/dart/builders/ci.sandbox/vm-kernel-precomp-obfuscate-linux-release-x64/6039https://ci.chromium.org/p/dart/builders/ci.sandbox/vm-kernel-precomp-android-release-arm_x64/957
Original change's description:
> [vm] Add support for real unboxed floating point fields in AOT
>
> Non-nullable floating point fields (double, Float32x4, Float64x2)
> are fully unboxed in their classes.
>
> A bitmap for each class was added to the shared class table in order to keep
> track of the pointers of the classes. Since all classes in Flutter Gallery
> have less than 64 fields, the bitmap is represented by a 64 bit integer and
> fields whose offset is more than 64 words are not unboxed.
>
> The instance sizes and field offsets might change between target and host
> in cross-compilation, since the number of words used to store unboxed fields
> may differ.
>
> dart-aot Xeon
>
> SplayLatency -4.62%
> SplayHarderLatency -4.17%
> NavierStokes -2.20%
> Tracer 8.72%
> ParticleSystemPaint 2.90%
> NBodySIMD 8.35%
> NBody 25.59%
>
> With hack TFA to make doubles in Rect/Offset/Size classes in flutter non-nullable:
>
> flutter arm-v8:
>
> gallery total size: -1%
>
> matrix_utils_transform_rect_perspective -16.70% (less is better)
> matrix_utils_transform_rect_affine -31.82% (less is better)
> matrix_utils_transform_point_perspective -24.90% (less is better)
> matrix_utils_transform_point_affine) -27.26% (less is better)
> rrect_contains_bench -4.719% (less is better)
>
> Change-Id: I9ae09c9c3167d99f9efd071a92937aa51093fd1d
> Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/131824
> Commit-Queue: Victor Agnez Lima <victoragnez@google.com>
> Reviewed-by: Martin Kustermann <kustermann@google.com>
> Reviewed-by: Ryan Macnak <rmacnak@google.com>
> Reviewed-by: Samir Jindel <sjindel@google.com>
TBR=kustermann@google.com,rmacnak@google.com,sjindel@google.com,victoragnez@google.com
Change-Id: Ic73858f6adb7f55c4129d4f46ff4731b378cb634
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/134020
Reviewed-by: Zichang Guo <zichangguo@google.com>
Commit-Queue: Zichang Guo <zichangguo@google.com>
Non-nullable floating point fields (double, Float32x4, Float64x2)
are fully unboxed in their classes.
A bitmap for each class was added to the shared class table in order to keep
track of the pointers of the classes. Since all classes in Flutter Gallery
have less than 64 fields, the bitmap is represented by a 64 bit integer and
fields whose offset is more than 64 words are not unboxed.
The instance sizes and field offsets might change between target and host
in cross-compilation, since the number of words used to store unboxed fields
may differ.
dart-aot Xeon
SplayLatency -4.62%
SplayHarderLatency -4.17%
NavierStokes -2.20%
Tracer 8.72%
ParticleSystemPaint 2.90%
NBodySIMD 8.35%
NBody 25.59%
With hack TFA to make doubles in Rect/Offset/Size classes in flutter non-nullable:
flutter arm-v8:
gallery total size: -1%
matrix_utils_transform_rect_perspective -16.70% (less is better)
matrix_utils_transform_rect_affine -31.82% (less is better)
matrix_utils_transform_point_perspective -24.90% (less is better)
matrix_utils_transform_point_affine) -27.26% (less is better)
rrect_contains_bench -4.719% (less is better)
Change-Id: I9ae09c9c3167d99f9efd071a92937aa51093fd1d
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/131824
Commit-Queue: Victor Agnez Lima <victoragnez@google.com>
Reviewed-by: Martin Kustermann <kustermann@google.com>
Reviewed-by: Ryan Macnak <rmacnak@google.com>
Reviewed-by: Samir Jindel <sjindel@google.com>
Dynamic invocation forwarders have the same recognized kind
as the method they are forwarding to.
That causes us to inline the recognized method and not the
dyn: forwarder itself.
It is not safe for user-implementable types. This is triggered
in some of the existing tests when unboxed integer fields are
allowed in AOT.
Issue https://github.com/dart-lang/sdk/issues/37737
Change-Id: Id518c6b13ccd08998ec78384e55c97c64546098a
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/133561
Commit-Queue: Victor Agnez Lima <victoragnez@google.com>
Reviewed-by: Martin Kustermann <kustermann@google.com>
This CL is a code cleanup landed separately from https://dart-review.googlesource.com/c/sdk/+/129081 to make the CL size smaller.
Refactorings:
* Split up runtime/vm/compiler/ffi.{cc,h} in multiple files.
* Move FrameRebase to compiler/ffi and prevent it from modifying Locations.
* Remove use of <algorithm>.
Change-Id: Ice2a96470a3242a731255383b9647fabab107ac5
Cq-Include-Trybots: luci.dart.try:vm-ffi-android-debug-arm-try,vm-ffi-android-debug-arm64-try,app-kernel-linux-debug-x64-try,vm-kernel-linux-debug-ia32-try,vm-kernel-win-debug-x64-try,vm-kernel-win-debug-ia32-try,vm-kernel-precomp-linux-debug-x64-try,vm-dartkb-linux-release-x64-abi-try,vm-kernel-precomp-android-release-arm64-try,vm-kernel-asan-linux-release-x64-try,vm-kernel-linux-release-simarm-try,vm-kernel-linux-release-simarm64-try,vm-kernel-precomp-android-release-arm_x64-try,vm-kernel-precomp-obfuscate-linux-release-x64-try,dart-sdk-linux-try,analyzer-analysis-server-linux-try,analyzer-linux-release-try,front-end-linux-release-x64-try,vm-kernel-precomp-win-release-x64-try
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/133224
Commit-Queue: Daco Harkes <dacoharkes@google.com>
Reviewed-by: Martin Kustermann <kustermann@google.com>
This reverts commit 5f81efb4a3.
Reason for revert: 3xH bot is green again.
Original change's description:
> Revert "[vm/aot] Only patch call-sites to monomorphic entry-points when the receiver is proven to be a heap object."
>
> This reverts commit ab2026af45.
>
> Reason for revert: Flutter 3xH is broken on Golem.
>
> Original change's description:
> > [vm/aot] Only patch call-sites to monomorphic entry-points when the receiver is proven to be a heap object.
> >
> > Flutter Gallery ARMv8: -0.8% total, -1.3% instructions
> > Flutter Gallery ARMv7: -0.7% total, -1.2% instructions
> >
> > There are some performance regressions, in four categories:
> >
> > 1. Arithmetic on num types (e.g. MinLib, MaxLib):
> > We believe arithmetic on num types is rare in practice, and can be optimized later
> > by specializing the call-sites.
> >
> > 2. Lack of types (e.g. MeshDecompression, ImagingGaussianBlur, Crypto*):
> > These (and similar benchmarks) are written in a very untyped style, which is not
> > representative of Dart 2 style. We've confirmed that the regressions disappear if
> > the benchmarks are annotated with appropriate types.
> >
> > 3. Megamorphic calls which are specialized in the benchmark (e.g. DeltaBlue):
> > DeltaBlue uses string interpolation in one place, so the toString() method in the
> > string interpolation helper is monomorphic for _Smis. However, in a realistic programs,
> > string interpolation is used much more frequently and with different types, so it's very
> > unlikely that this call would be monomorphic. When string interpolation is removed, there's
> > a remaining 1.3% regression on DeltaBlue which is not yet accounted for.
> >
> > 4. Noisy benchmarks
> > The remaining regressions are on benchmarks which show frequent jitter historically.
> >
> > There are also 1-4% improvements on many benchmarks (Golem's "noise" analysis is hiding
> > many of them, look at the graphs).
> >
> > For the reasons above, we believe the regressions justified by the code size improvement.
> >
> > Change-Id: Ic9b281d4383a6111de9d6f44347976ffa61a6ca6
> > Cq-Include-Trybots:luci.dart.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-bare-linux-release-simarm64-try
> > Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/128667
> > Commit-Queue: Samir Jindel <sjindel@google.com>
> > Reviewed-by: Ryan Macnak <rmacnak@google.com>
> > Reviewed-by: Martin Kustermann <kustermann@google.com>
>
> TBR=kustermann@google.com,rmacnak@google.com,alexmarkov@google.com,sjindel@google.com
>
> Change-Id: I411aae4b230f2a08146ad3bf3e7a10aa6592db0e
> No-Presubmit: true
> No-Tree-Checks: true
> No-Try: true
> Cq-Include-Trybots: luci.dart.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-bare-linux-release-simarm64-try
> Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/131660
> Reviewed-by: Samir Jindel <sjindel@google.com>
> Commit-Queue: Samir Jindel <sjindel@google.com>
TBR=kustermann@google.com,rmacnak@google.com,alexmarkov@google.com,sjindel@google.com
# Not skipping CQ checks because original CL landed > 1 day ago.
Change-Id: I36e4e9d598033fcc934c2cdf5b061aa11255d3b2
Cq-Include-Trybots: luci.dart.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-bare-linux-release-simarm64-try
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/131837
Reviewed-by: Samir Jindel <sjindel@google.com>
Commit-Queue: Samir Jindel <sjindel@google.com>
- Update the flow graph and block order in the call specializer after
round trip serialization. This ensures that compiler passes using the
call specializer operate on the deserialized flow graph rather than
the original one.
- Add 2 instead of 1 to the highest seen SSA temp index after
deserialization to obtain the current SSA temp index. This makes room
for values that expand into two SSA slots during some optimizations,
e.g. unboxed integers on 32-bit platforms.
The changes expose an issue with early round-trip serialization,
described in https://github.com/dart-lang/sdk/issues/40120
Change-Id: I21e22fa1bab595fb5764ab72b599a3c663b4ff47
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/128405
Commit-Queue: Aske Simon Christensen <askesc@google.com>
Reviewed-by: Teagan Strickland <sstrickl@google.com>
This reverts commit ab2026af45.
Reason for revert: Flutter 3xH is broken on Golem.
Original change's description:
> [vm/aot] Only patch call-sites to monomorphic entry-points when the receiver is proven to be a heap object.
>
> Flutter Gallery ARMv8: -0.8% total, -1.3% instructions
> Flutter Gallery ARMv7: -0.7% total, -1.2% instructions
>
> There are some performance regressions, in four categories:
>
> 1. Arithmetic on num types (e.g. MinLib, MaxLib):
> We believe arithmetic on num types is rare in practice, and can be optimized later
> by specializing the call-sites.
>
> 2. Lack of types (e.g. MeshDecompression, ImagingGaussianBlur, Crypto*):
> These (and similar benchmarks) are written in a very untyped style, which is not
> representative of Dart 2 style. We've confirmed that the regressions disappear if
> the benchmarks are annotated with appropriate types.
>
> 3. Megamorphic calls which are specialized in the benchmark (e.g. DeltaBlue):
> DeltaBlue uses string interpolation in one place, so the toString() method in the
> string interpolation helper is monomorphic for _Smis. However, in a realistic programs,
> string interpolation is used much more frequently and with different types, so it's very
> unlikely that this call would be monomorphic. When string interpolation is removed, there's
> a remaining 1.3% regression on DeltaBlue which is not yet accounted for.
>
> 4. Noisy benchmarks
> The remaining regressions are on benchmarks which show frequent jitter historically.
>
> There are also 1-4% improvements on many benchmarks (Golem's "noise" analysis is hiding
> many of them, look at the graphs).
>
> For the reasons above, we believe the regressions justified by the code size improvement.
>
> Change-Id: Ic9b281d4383a6111de9d6f44347976ffa61a6ca6
> Cq-Include-Trybots:luci.dart.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-bare-linux-release-simarm64-try
> Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/128667
> Commit-Queue: Samir Jindel <sjindel@google.com>
> Reviewed-by: Ryan Macnak <rmacnak@google.com>
> Reviewed-by: Martin Kustermann <kustermann@google.com>
TBR=kustermann@google.com,rmacnak@google.com,alexmarkov@google.com,sjindel@google.com
Change-Id: I411aae4b230f2a08146ad3bf3e7a10aa6592db0e
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Cq-Include-Trybots: luci.dart.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-bare-linux-release-simarm64-try
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/131660
Reviewed-by: Samir Jindel <sjindel@google.com>
Commit-Queue: Samir Jindel <sjindel@google.com>
Flutter Gallery ARMv8: -0.8% total, -1.3% instructions
Flutter Gallery ARMv7: -0.7% total, -1.2% instructions
There are some performance regressions, in four categories:
1. Arithmetic on num types (e.g. MinLib, MaxLib):
We believe arithmetic on num types is rare in practice, and can be optimized later
by specializing the call-sites.
2. Lack of types (e.g. MeshDecompression, ImagingGaussianBlur, Crypto*):
These (and similar benchmarks) are written in a very untyped style, which is not
representative of Dart 2 style. We've confirmed that the regressions disappear if
the benchmarks are annotated with appropriate types.
3. Megamorphic calls which are specialized in the benchmark (e.g. DeltaBlue):
DeltaBlue uses string interpolation in one place, so the toString() method in the
string interpolation helper is monomorphic for _Smis. However, in a realistic programs,
string interpolation is used much more frequently and with different types, so it's very
unlikely that this call would be monomorphic. When string interpolation is removed, there's
a remaining 1.3% regression on DeltaBlue which is not yet accounted for.
4. Noisy benchmarks
The remaining regressions are on benchmarks which show frequent jitter historically.
There are also 1-4% improvements on many benchmarks (Golem's "noise" analysis is hiding
many of them, look at the graphs).
For the reasons above, we believe the regressions justified by the code size improvement.
Change-Id: Ic9b281d4383a6111de9d6f44347976ffa61a6ca6
Cq-Include-Trybots:luci.dart.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-bare-linux-release-simarm64-try
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/128667
Commit-Queue: Samir Jindel <sjindel@google.com>
Reviewed-by: Ryan Macnak <rmacnak@google.com>
Reviewed-by: Martin Kustermann <kustermann@google.com>
This change introduces a base class for all kinds of instance calls
(InstanceCallInstr and PolymorphicInstanceCallInstr for now).
This allows us to have common code which works with both of them
uniformly and remove forwarding from PolymorphicInstanceCallInstr to
InstanceCallInstr.
Change-Id: If39aed06a52b52dbe6ec7bff3e4e3db246e67c7d
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/130935
Commit-Queue: Alexander Markov <alexmarkov@google.com>
Reviewed-by: Teagan Strickland <sstrickl@google.com>
Reviewed-by: Martin Kustermann <kustermann@google.com>
This also fixes the implementation of 'null is T' in legacy vs opted-in library
by making it independent of strong/weak mode.
Includes updates to type propagation and call specializer.
Change-Id: I2d347ef277e08cef779faf51a972e863244ef095
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/130922
Commit-Queue: Régis Crelier <regis@google.com>
Reviewed-by: Alexander Markov <alexmarkov@google.com>
Reviewed-by: Liam Appelbe <liama@google.com>
NNBDMode does not reflect the semantics (legacy or nnbd) of type tests anymore.
Instead, the semantics are derived from the value of the strong mode flag.
This required one function to be specialized, namely the implementation of
'null is Type' in weak mode in an opted-in library, which still requires
nnbd semantics although run in weak mode.
Relevant changes are in object.h and object.cc:
- methods NNBD_NullIsInstanceOf and NNBD_IsTopType are new.
- methods IsNullType, IsTopType, and IsNeverType are modified.
Change-Id: I36cd43d93d2cfabd110cbcc6b26487a583bb089d
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/130444
Commit-Queue: Régis Crelier <regis@google.com>
Reviewed-by: Alexander Markov <alexmarkov@google.com>
This does not yet include checking of required named parameters.
Also, inlined type testing code, type testing stubs, and type propagation are
not yet updated.
Change-Id: If6d8996635474e349d28b7fd207d724838d08881
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/129922
Commit-Queue: Régis Crelier <regis@google.com>
Reviewed-by: Alexander Markov <alexmarkov@google.com>
When inserting Unbox instructions in SelectRepresentations pass, their
speculative mode depends on the instruction which uses corresponding
value. Phi instructions were using default kGuardInputs mode, which may
cause insertion of speculative Unbox instructions in AOT mode.
Such Unbox instructions may need deoptimization if type of an argument
changes during optimizations. Generating code for such Unbox
instructions causes assertion failure (in debug mode), or crash in
release mode.
The fix is to change speculative mode of Phi instructions in AOT, so
they won't cause speculative Unbox instructions.
Fixes https://github.com/dart-lang/sdk/issues/39747
Change-Id: Ic9c9957875fd775fe1c02f3666829f26dad64937
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/130063
Reviewed-by: Samir Jindel <sjindel@google.com>
Commit-Queue: Alexander Markov <alexmarkov@google.com>
The pc-relative calls were recorded with pc-offsets pointing to the start of the
call instruction, where as code-based calls were recorded with
pc-offsets pointing to the next instruction (i.e. return address).
This is inconsistent and caused us to hit an assert in
`Code::set_static_calls_target_table`. The assert was benign, but it is
good to maintain the uniqueness guarantee in the static calls table, so
we'll unify the encoding to use offsets to the instruction after the
call in both cases.
Closes https://github.com/dart-lang/sdk/issues/39811
Change-Id: Id0305befd78f09ed0b0e100f39641bca9e764442
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/129717
Reviewed-by: Ben Konyi <bkonyi@google.com>
Commit-Queue: Ben Konyi <bkonyi@google.com>
This change hides InstanceCallInstr inside PolymorphicInstanceCallInstr
and makes sure that PolymorphicInstanceCallInstr owns its arguments
(which would become critical once call arguments become instruction
inputs).
Also, PolymorphicInstanceCallInstr now extends TemplateDartCall which
makes it possible to avoid duplication of methods related to all calls.
Issue: https://github.com/dart-lang/sdk/issues/39788
Change-Id: Ie3d4ff46a8e99a0988ba88a88ca9a60be8503ce0
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/129307
Commit-Queue: Alexander Markov <alexmarkov@google.com>
Reviewed-by: Martin Kustermann <kustermann@google.com>
The check of not removing 64-bit->32-bit truncating conversions was in the wrong place.
Fixes: https://github.com/dart-lang/sdk/issues/39885
Fixes: https://github.com/flutter/flutter/issues/47454
Change-Id: Ic32a13cbf7ec622692cdca86b3237dae1c0f1cef
Cq-Include-Trybots: luci.dart.try:vm-ffi-android-debug-arm-try,vm-ffi-android-debug-arm64-try,app-kernel-linux-debug-x64-try,vm-kernel-linux-debug-ia32-try,vm-kernel-win-debug-x64-try,vm-kernel-win-debug-ia32-try,vm-kernel-precomp-linux-debug-x64-try,vm-dartkb-linux-release-x64-abi-try,vm-kernel-precomp-android-release-arm64-try,vm-kernel-asan-linux-release-x64-try,vm-kernel-linux-release-simarm-try,vm-kernel-linux-release-simarm64-try,vm-kernel-precomp-android-release-arm_x64-try,vm-kernel-precomp-obfuscate-linux-release-x64-try,dart-sdk-linux-try,analyzer-analysis-server-linux-try,analyzer-linux-release-try,front-end-linux-release-x64-try
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/129284
Commit-Queue: Daco Harkes <dacoharkes@google.com>
Reviewed-by: Martin Kustermann <kustermann@google.com>
After dab60b485c we decide to use shared
slow path for BoxInt64Instr based on FLAG_use_bare_instructions.
However, even in --use_bare_instructions mode we might be compiling
something before VM isolate stubs are replaced with isolate-specific
stubs. In such case we shouldn't use shared slow path as
BoxInt64Instr::EmitNativeCode generates PC-relative call to the stub.
This change fixes the assertion
../../runtime/vm/compiler/backend/il_arm.cc: 4650: error: expected: !stub.InVMIsolateHeap()
version=2.8.0-edge.dab60b485ca4544d58970d9be299d5aaf78cf9e5 (Thu Dec 19 17:29:12 2019 +0000) on "linux_simarm"
thread=28698, isolate=isolate(0x5640171cd800)
pc 0x000056401446060c fp 0x00007ffec2747190 dart::Profiler::DumpStackTrace(void*)
pc 0x00005640141d23b2 fp 0x00007ffec2747270 dart::Assert::Fail(char const*, ...)
pc 0x00005640145fa85d fp 0x00007ffec2747300 out/DebugSIMARM_X64/gen_snapshot+0xa1285d
pc 0x00005640145b35b6 fp 0x00007ffec2747470 dart::FlowGraphCompiler::VisitBlocks()
pc 0x00005640145bfd94 fp 0x00007ffec27474a0 dart::FlowGraphCompiler::CompileGraph()
pc 0x000056401454aa07 fp 0x00007ffec2747af0 dart::PrecompileParsedFunctionHelper::Compile(dart::CompilationPipeline*)
pc 0x000056401454cf3f fp 0x00007ffec2747d00 out/DebugSIMARM_X64/gen_snapshot+0x964f3f
pc 0x000056401454765e fp 0x00007ffec2747da0 dart::Precompiler::CompileFunction(dart::Precompiler*, dart::Thread*, dart::Zone*, dart::Function const&)
pc 0x0000564014470016 fp 0x00007ffec2747de0 out/DebugSIMARM_X64/gen_snapshot+0x888016
pc 0x000056401446d72c fp 0x00007ffec2747e70 dart::ProgramVisitor::VisitClasses(dart::Visitor<dart::Class>*)
pc 0x000056401446d9f3 fp 0x00007ffec2747f00 dart::ProgramVisitor::VisitFunctions(dart::Visitor<dart::Function>*)
pc 0x000056401453e474 fp 0x00007ffec27480c0 dart::Precompiler::DoCompileAll()
pc 0x000056401453e245 fp 0x00007ffec27484b0 dart::Precompiler::CompileAll()
pc 0x00005640149c60eb fp 0x00007ffec2748570 Dart_Precompile
pc 0x00005640141a7a5f fp 0x00007ffec27486f0 dart::bin::main(int, char**)
-- End of DumpStackTrace
in debug mode.
Change-Id: I5095e8def7b3d8067423455f32bce5aea1e0c381
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/129141
Reviewed-by: Samir Jindel <sjindel@google.com>
Commit-Queue: Alexander Markov <alexmarkov@google.com>
BoxInt64Instr::MakeLocationSummary pretended that it is using shared
slow path, but it was not actually used in --no-use-bare-instructions mode.
As a result, BoxInt64Instr slow path pushed only live registers, but
stack map was generated for all registers. This caused various crashes,
mostly in GC.
The fix is to correctly create LocationSummary with kCallOnSlowPath
instead of kCallOnSharedSlowPath if we're not going to use shared
slow path (if --no-use-bare-instructions is specified).
Fixes https://github.com/dart-lang/sdk/issues/39853
Fixes https://github.com/dart-lang/sdk/issues/39854
Change-Id: Iaba7cd7d88143f2621a007c8559b932bb171de4f
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/129004
Reviewed-by: Victor Agnez Lima <victoragnez@google.com>
Reviewed-by: Martin Kustermann <kustermann@google.com>
Commit-Queue: Alexander Markov <alexmarkov@google.com>
This change stops seting unbox_numeric_fields flag to false in AOT.
It also sets the speculative mode to be non speculative for store
unboxed fields instructions in AOT, since unboxing is done based on TFA.
Change-Id: Ic7a6391f8569e22eea13128251d83d109b06083c
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/127884
Commit-Queue: Victor Agnez Lima <victoragnez@google.com>
Reviewed-by: Samir Jindel <sjindel@google.com>
Reviewed-by: Martin Kustermann <kustermann@google.com>
This CL is responsible for inlining Float32x4 and Float64x2 arithmetic
operations in order to reduce their run-time cost.
Null-checks were required in order to allow the unbox instructions to
be non-speculative.
CheckNullInstr was modified to throw ArgumentError when checking
arguments for those operations.
Change-Id: I18b36d96a36011647480f301eb6dc6bb478f59a1
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/128589
Commit-Queue: Victor Agnez Lima <victoragnez@google.com>
Reviewed-by: Martin Kustermann <kustermann@google.com>
Reviewed-by: Samir Jindel <sjindel@google.com>
In the current compiler pipeline, PushArgument instructions can
end up outliving the calls to which they are an argument in some
circumstances. For example, the PushArgument is left in the graph
if it also appears in the environments of interleaving instructions.
When the dead code eliminator calculates liveness of a definition,
it ignores appearances in PushArgument instructions. Instead, it
looks for uses of the definition either directly or via
PushArgument in the arguments of calls and environments of other
instructions in the graph. An separate pass run before DCE
removes environments from instructions, so a definition used only
in PushArguments that have outlived their calls is considered dead.
The current DCE removes dead definitions but leaves PushArgument
instructions be. For orphaned PushArguments of dead definitions,
this means after DCE they now reference definitions no longer in
the graph.
This change instead checks the definitions referenced by
PushArgument instructions for liveness. If a PushArgument
references a dead definition, then the DCE removes the PushArgument
instruction, as its call must have already been removed (else
the definition would still be live), and environments, and thus
environmental uses, have already been removed from the graph.
Bug: https://github.com/dart-lang/sdk/issues/39767
Change-Id: I74c8e92a0d93f2aae42f8ff6f06ee70899b2f794
Cq-Include-Trybots: luci.dart.try:vm-kernel-precomp-linux-debug-x64-try
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/128403
Reviewed-by: Alexander Markov <alexmarkov@google.com>
Commit-Queue: Teagan Strickland <sstrickl@google.com>
If the dst_type of an AssertAssignable is a type parameter, the AssertAssignable
implementation will load the value of the type parameter using
the (instantiator or function) type arguments. It will then call the
type testing stub (TTS) of that type.
If the TTS is not exhaustive (e.g. because `T = X<..>` wher `X` is
implemented), it can fall back to the slower SubTypeTestCache implementation.
Right now the STC fallback will get the loaded value of the type
parameter for `dst_type` instead of the type parameter. Doing so is
incorrect.
=> This CL ensures we preserve dst_type = TypeParameter for the STC
fallback.
Issue https://github.com/dart-lang/sdk/issues/39716
Change-Id: Id5b318ed4940b4d749f6ab8b4c2a4efc295d9dde
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/127888
Auto-Submit: Martin Kustermann <kustermann@google.com>
Commit-Queue: Martin Kustermann <kustermann@google.com>
Reviewed-by: Régis Crelier <regis@google.com>
Victor Lima