This is a reland of a528b19d2c
Updating the hash of the empty type argument vector in write-protected memory caused a crash in precomp mode.
The fix is to correctly initialize the hash so it does not need an update.
Original change's description:
> [VM/runtime] Fix computation of the hash of recursive types.
>
> When computing the hash of a type, only include the hashes of the type arguments corresponding to the type parameters of the type's class and not of the whole type argument vector, otherwise, the variable location of TypeRefs in the base class type arguments may result in different hashes.
>
> Add a regression test (this test was not failing in all configurations, since the issue becomes visible depending on the order in which types are loaded and finalized).
>
> This fixes issue https://github.com/dart-lang/sdk/issues/42954
>
> Small improvement: pick hash value 1 for a vector of dynamic types, so it can be cached and not recomputed repeatedly, which was the case with value 0.
>
> Change-Id: I3468952cbf61944a82cf8e00efbf4952bd831fbe
> Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/157563
> Reviewed-by: Alexander Markov <alexmarkov@google.com>
> Commit-Queue: Régis Crelier <regis@google.com>
Change-Id: I538b4538c69472e09b7bd745c6bd41023e1f154f
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/157781
Commit-Queue: Régis Crelier <regis@google.com>
Reviewed-by: Alexander Markov <alexmarkov@google.com>
Generally, methods that take a *TextBuffer pointer do not care how the
internal buffer is allocated, and so they could be used for either if
both were subclasses of a base class that contained the printing
methods. This CL makes that base class, and now TextBuffer and
ZoneTextBuffer now share the exact same set of methods for printing to
the internal buffer.
Since the base class is in platform, this does mean dropping the
overload of AddString for Dart String objects that was part of
ZoneTextBuffer. Instead, this CL just adds an intermediate call to
ToCString() for the small number of callers that used the overload,
keeping the printing interface the same for both.
In addition, one use of TextBuffer that then re-allocated the buffer
contents into the zone manually has been replaced with a ZoneTextBuffer
instead.
Change-Id: I438a085e7e20d55d93987fd7f36afd636f95955f
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/157741
Reviewed-by: Ryan Macnak <rmacnak@google.com>
Commit-Queue: Tess Strickland <sstrickl@google.com>
This reverts commit a528b19d2c.
Reason for revert: Failures in precompiled mode.
Original change's description:
> [VM/runtime] Fix computation of the hash of recursive types.
>
> When computing the hash of a type, only include the hashes of the type arguments corresponding to the type parameters of the type's class and not of the whole type argument vector, otherwise, the variable location of TypeRefs in the base class type arguments may result in different hashes.
>
> Add a regression test (this test was not failing in all configurations, since the issue becomes visible depending on the order in which types are loaded and finalized).
>
> This fixes issue https://github.com/dart-lang/sdk/issues/42954
>
> Small improvement: pick hash value 1 for a vector of dynamic types, so it can be cached and not recomputed repeatedly, which was the case with value 0.
>
> Change-Id: I3468952cbf61944a82cf8e00efbf4952bd831fbe
> Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/157563
> Reviewed-by: Alexander Markov <alexmarkov@google.com>
> Commit-Queue: Régis Crelier <regis@google.com>
TBR=rmacnak@google.com,alexmarkov@google.com,asiva@google.com,regis@google.com
Change-Id: I904e7043fa175d03bb9af56a9346b097b883c7a0
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/157700
Reviewed-by: Régis Crelier <regis@google.com>
Commit-Queue: Régis Crelier <regis@google.com>
When computing the hash of a type, only include the hashes of the type arguments corresponding to the type parameters of the type's class and not of the whole type argument vector, otherwise, the variable location of TypeRefs in the base class type arguments may result in different hashes.
Add a regression test (this test was not failing in all configurations, since the issue becomes visible depending on the order in which types are loaded and finalized).
This fixes issue https://github.com/dart-lang/sdk/issues/42954
Small improvement: pick hash value 1 for a vector of dynamic types, so it can be cached and not recomputed repeatedly, which was the case with value 0.
Change-Id: I3468952cbf61944a82cf8e00efbf4952bd831fbe
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/157563
Reviewed-by: Alexander Markov <alexmarkov@google.com>
Commit-Queue: Régis Crelier <regis@google.com>
Before null safety
try {
...
} catch (e) {
...
}
was translated to a try-catch block with 'dynamic' catch type.
VM has a special, more efficient handling of such catch-all try blocks.
Those try blocks were detected by comparing catch type with 'dynamic'.
With null safety front-end started to translate those try blocks
using non-nullable Object as a catch type. As a result, this disabled
all optimizations for catch-all try blocks in the VM.
This change extends detection of catch-all try blocks to handle both
dynamic and Object as catch types.
Improves ParserCombinators benchmark with null safety 12x in JIT mode,
15x in AOT mode. This benchmark is now on par with legacy (pre-NNBD)
version.
Change-Id: I128aa1599d8a6f979fc2e8535d0f5c934bf3a5ba
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/157565
Reviewed-by: Régis Crelier <regis@google.com>
Commit-Queue: Alexander Markov <alexmarkov@google.com>
To make implicit getters/setters frameless in common cases we use graph
intrinsics. We do this for boxed as well as unboxed fields. We do this
only for simple cases (e.g. not if field guards or possible boxing are
involved).
It also removes the hand-written assembly intrinsics for the getters,
which avoids bugs if boxing-state of field is not the same as
boxing-state of setter/getter.
Issue https://github.com/dart-lang/sdk/issues/40876
Change-Id: Id3bd7842ead0278fee61d4a21c8661ce18dbccee
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/155900
Commit-Queue: Martin Kustermann <kustermann@google.com>
Reviewed-by: Alexander Markov <alexmarkov@google.com>
Reviewed-by: Aske Simon Christensen <askesc@google.com>
Reviewed-by: Tess Strickland <sstrickl@google.com>
- Makes Future.wait a recognised function, and asserts its chained
future, _future is allocated at a known index in the context.
- Adds logic to locate, extract the chained future during lazy async
stack unwinding.
- Adds tests for the Future.wait async case.
- Minor consistency nits, comments.
This change is similar to a previous CL, adding Future.timeout support:
https://dart-review.googlesource.com/c/sdk/+/152328
Change-Id: I7439750968595d25d7bbac0068ad64fcc891e176
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/155420
Commit-Queue: Clement Skau <cskau@google.com>
Reviewed-by: Martin Kustermann <kustermann@google.com>
Also relands the followup CLs:
"Perform non-covariant checks when dynamically invoking callables."
"Use AreValidArguments so that names are checked as well."
Original description of first CL:
This change only affects compilation when running in non-precompiled
mode with --no-lazy-dispatchers enabled.
Instead of always compiling in non-covariant checks, even for closures
not called dynamically, remove the non-covariant checks from the closure
and instead do the non-covariant checks for dynamic calls during the
NoSuchMethodForCallStub fallback by calling
Function::DoArgumentTypesMatch.
Adds two overloads for Function::DoArgumentTypesMatch, one which takes a
function type argument vector and one which takes neither an
instantiator type argument vector or a function type argument vector.
For the versions that are not explicitly passed a type argument vector,
an appropriate one is constructed using the arguments. If there is not
enough information in the arguments, then we fall back to assuming the
empty type argument vector for the instantiator case and instantiating
to bounds in the function type argument case.
Fixes Function::DoArgumentTypesMatch to handle generic functions and to
check arguments appropriately according to the active null safety mode.
For generic functions, the provided or resulting function type vector
has non-covariant checks performed against the type parameter bounds.
This change uncovered one test that was incorrectly passing in strong
mode, see https://github.com/dart-lang/sdk/issues/42688 for details.
Original description of second CL:
The VM only does this when the callable function does not expect dynamic
invocations. Otherwise, performing the checks would be redundant, as the
function body already contains the appropriate non-covariant checks.
Third CL had no additional description.
Cq-Include-Trybots: luci.dart.try:vm-kernel-precomp-linux-debug-x64-try,vm-dartkb-linux-release-x64-try,vm-kernel-reload-linux-release-x64-try, vm-kernel-reload-rollback-linux-debug-x64-try
Bug: https://github.com/dart-lang/sdk/issues/40813
Change-Id: I1a3e9c1865103a8d716e1cad814267caffaaadf2
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/154688
Reviewed-by: Martin Kustermann <kustermann@google.com>
(Assignment of libraries to loading units is already done in the kernel generation step.)
After compiling and before serializing, we walk the program and for each Code we assign its Instructions, CodeSourceMap and CompressedStackMap to the loading unit of that Code's defining library. Deduplication may cause Instructions, CodeSourceMaps and CompressedStackMaps to belong to more than one loading unit; in this case the objects are assigned to the root loading unit. Later they can be more precisely assigned to the dominating loading unit.
All objects except some Instructions, CodeSourceMaps and CompressedStackMaps belong to the root loading unit's snapshot. This snapshot is written like an unsplit snapshot, except that when serializing Code, we will write a reference to a stub or null when the Code's Instructions, CodeSourceMap or CompressedStackMap belongs to a non-root loading unit. The snapshots of non-root loading units contain these deferred objects and references to the corresponding Code objects to patch. The types of objects we defer (Instructions, CodeSourceMaps and CompressedStackMaps) usually represent 70+% of the snapshot size.
Bare instructions mode must be disabled when splitting because we cannot have PC-relative calls between loading units. Later we can re-enable this for calls within loading units.
Broken: Compactor probably crashes we can now have an unbounded number of image pages and the compactor assumes a fixed number.
Embedder's guide:
At compile-time, gen_snapshot should be passed --loading_unit_manifest with a path, which will enable splitting and output a mapping from loading unit ids to snapshot output paths.
At runtime, sometime during isolate startup, an embedder should call Dart_SetDeferredLoadHandler, probably near an existing call to Dart_SetLibraryTagHandler. The callback is given a loading unit id, and should eventually call Dart_DeferredLoadComplete[Error].
Bug: https://github.com/dart-lang/sdk/issues/41974
Change-Id: Ib597eb87c8cd634416d5ee1f00629c5550aebb00
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/152427
Commit-Queue: Ryan Macnak <rmacnak@google.com>
Reviewed-by: Alexander Markov <alexmarkov@google.com>
This reverts commit 556f0fb5c8 and fixes merge conflict.
Revert "[vm] Perform non-covariant checks when dynamically invoking callables." as a collateral.
This reverts commit 7a6287bb92.
Revert "[vm] Use AreValidArguments so that names are checked as well." as a collateral
This reverts commit 90c683018f.
Change-Id: Iee2c5e91c39d19dca7a1e640c103443c95e6c4f9
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/154705
Reviewed-by: Alexander Aprelev <aam@google.com>
Commit-Queue: Alexander Aprelev <aam@google.com>
This CL adds support for unboxing of return values of getters and
therefore completes the work on unboxing support for normal members
(closures are still missing).
As with existing unboxing support for methods and setters, we utilize
TFA information to proof a getter will always return non-nullable
int/double.
We then make such a getter return unboxed int/double. If there are
dynamic calls to the getter a dyn:get:* forwarder will be created
which performs boxing of the return value.
Overall this reduces RX by eliminating BoxInt64 instructions. It
sometimes increases metadata due to more dyn:get:* function objects.
=> On our main size benchmark targets show no significant change.
Issue https://github.com/dart-lang/sdk/issues/40876
Change-Id: If7450ef7e5e3fc9c2e0eaa6b86ffa817699a7e17
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/154329
Commit-Queue: Martin Kustermann <kustermann@google.com>
Reviewed-by: Alexander Markov <alexmarkov@google.com>
Reviewed-by: Tess Strickland <sstrickl@google.com>
The VM only does this when the callable function does not expect dynamic
invocations. Otherwise, performing the checks would be redundant, as the
function body already contains the appropriate non-covariant checks.
Cq-Include-Trybots: luci.dart.try:vm-kernel-precomp-linux-release-x64-try,vm-dartkb-linux-release-x64-try
Bug: https://github.com/dart-lang/sdk/issues/40813
Change-Id: I27b5f4067247c6b1b76ba57951a1d91455c2eeb5
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/154463
Commit-Queue: Tess Strickland <sstrickl@google.com>
Reviewed-by: Martin Kustermann <kustermann@google.com>
This change only affects compilation when running in non-precompiled
mode with --no-lazy-dispatchers enabled.
Instead of always compiling in non-covariant checks, even for closures
not called dynamically, remove the non-covariant checks from the closure
and instead do the non-covariant checks for dynamic calls during the
NoSuchMethodForCallStub fallback by calling
Function::DoArgumentTypesMatch.
Adds two overloads for Function::DoArgumentTypesMatch, one which takes a
function type argument vector and one which takes neither an
instantiator type argument vector or a function type argument vector.
For the versions that are not explicitly passed a type argument vector,
an appropriate one is constructed using the arguments. If there is not
enough information in the arguments, then we fall back to assuming the
empty type argument vector for the instantiator case and instantiating
to bounds in the function type argument case.
Fixes Function::DoArgumentTypesMatch to handle generic functions and to
check arguments appropriately according to the active null safety mode.
For generic functions, the provided or resulting function type vector
has non-covariant checks performed against the type parameter bounds.
This change uncovered one test that was incorrectly passing in strong
mode, see https://github.com/dart-lang/sdk/issues/42688 for details.
Cq-Include-Trybots: luci.dart.try:vm-kernel-precomp-linux-release-x64-try,vm-dartkb-linux-release-x64-try
Change-Id: I5030a46b356778448b51a243f16406eacf1c04bf
Bug: https://github.com/dart-lang/sdk/issues/40813
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/153202
Commit-Queue: Tess Strickland <sstrickl@google.com>
Reviewed-by: Régis Crelier <regis@google.com>
Reviewed-by: Martin Kustermann <kustermann@google.com>
We already make the callsites use dyn:* selectors for normal method
calls as well as setters, this is doing the same thing for getters.
Once callsites use dyn:get:* various pieces in runtime need to be
adjusted to accomodate for that (e.g. NSM handling, etc).
A follow-up CL will then start actually generating dyn:* getters in
certain situations.
Issue https://github.com/dart-lang/sdk/issues/40876
Change-Id: If219603bc0b8eb119edd08b211a8897d21ec0fb6
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/154320
Commit-Queue: Martin Kustermann <kustermann@google.com>
Reviewed-by: Tess Strickland <sstrickl@google.com>
The precompiler has a notion of when a function should get its code
generated, when it should retain a functions types. This information
is used to prune function objects after AOT compilation.
This CL introduces the same pruning logic which we do for regular
functions as well as closure functions to invocation dispatchers.
Change-Id: I5c8a75f0cc822c71e876b8e2c9648d9976745895
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/154002
Commit-Queue: Martin Kustermann <kustermann@google.com>
Reviewed-by: Tess Strickland <sstrickl@google.com>
Right now we assign class ids to top-level classes, abstract classes as
well as concrete classes. All of them have allocated from a 16-bit pool
of ids. The VM FATAL()s once it hits that limit.
Customers who run very large programs (significant amount of generated
code) on the Dart VM have started to hit this 16-bit class limit.
Concrete classes can have instances in the heap. Our current heap layout
only allows 16-bit class ids to be encoded in the header word. To avoid
increasing the size of heap objects or shrinking the size of the identity
hash code to 16-bit we keep class ids in object headers to be 16-bit.
Abstract classes cannot have instances in the heap. Though their class
ids are encoded in type objects. Furthermore we sort classes in
AOT/AppJIT mode to perform fast class-id range checks. To avoid impacting
this optimization we treat abstract classes the same way as concrete
classes.
Top-level classes cannot have instances in the heap. Their class ids are
only used in the runtime code, for example for hot-reload as well as
part of the service protocol.
=> We can allocate class ids outside the 16-bit range for top-level
classes, thereby freeing a significant amount of space in the 16-bit
range.
This CL does exactly that: We change classid_t to be int32_t. The
ClassLayout::id_ can now be assigned ids outside 16-bit range for
top-level classes. To do this we keep dart classes and top level classes
as separate arrays in the ClassTable.
Issue https://github.com/dart-lang/sdk/issues/42533
See also b/160229360
Change-Id: I6710a644e7b0ab2d4f4c792bef8e1f91cb117421
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/153607
Commit-Queue: Martin Kustermann <kustermann@google.com>
Reviewed-by: Ryan Macnak <rmacnak@google.com>
Loading unit metadata is expanded to create an object representing each loading unit, and each library to reference the loading unit it is a part of.
To be used at compile-time to help assign objects to snapshots.
To be used at runtime to track which units have been loading and the base objects for cross-snapshot references.
Bug: https://github.com/dart-lang/sdk/issues/41974
Change-Id: Icd3fa95ffd5949bd8f7f84a6f750c4c515de0dba
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/151865
Commit-Queue: Ryan Macnak <rmacnak@google.com>
Reviewed-by: Alexander Markov <alexmarkov@google.com>
https://dart-review.googlesource.com/c/sdk/+/145591 introduced a try
catch into FFI calls to call ExitHandleScope on the exception path.
However, we only need this try-catch if we actually need to exit the
handle scope on the exception path, which is not the case if we have
no handles in the signature. So this CL makes the try catch optional.
This speeds up ffi calls without handles (tested on JIT x64):
FfiCall.Uint8x01(RunTime): 206.4801280066068 us.
->
FfiCall.Uint8x01(RunTime): 203.7240782236708 us.
Also adds a test that checks that an exception can still be propagated
with Dart_PropagateError from native code when the FFI trampoline has
no try catch.
Change-Id: I9fac7078381c60fb8055b64fff29ea364fbc948f
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-msan-linux-release-x64-try,vm-kernel-precomp-msan-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,vm-kernel-mac-debug-x64-try,vm-precomp-ffi-qemu-linux-release-arm-try,vm-kernel-nnbd-linux-debug-x64-try,analyzer-nnbd-linux-release-try,front-end-nnbd-linux-release-x64-try
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/151239
Commit-Queue: Daco Harkes <dacoharkes@google.com>
Reviewed-by: Martin Kustermann <kustermann@google.com>
Reviewed-by: Alexander Markov <alexmarkov@google.com>
Since CodeSourceMaps are not serialized when non-symbolic stack traces
are enabled, use a different CSM encoding in this mode. Here, we write
out the line number as before and then the column as a separately
written integer.
Cq-Include-Trybots: luci.dart.try:vm-kernel-precomp-nnbd-linux-release-x64-try,vm-kernel-precomp-linux-release-x64-try,vm-kernel-precomp-linux-product-x64-try,vm-kernel-precomp-linux-debug-x64-try,vm-kernel-precomp-linux-release-simarm_x64-try
Change-Id: If9aec2f52551d04eb45626744957e625bcc17d3d
Bug: https://github.com/dart-lang/sdk/issues/42397
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/151824
Commit-Queue: Tess Strickland <sstrickl@google.com>
Reviewed-by: Daco Harkes <dacoharkes@google.com>
Changes:
Doing this always in precompiled mode meant increased data segment sizes
when CodeSourceMaps are stored, since encoded line/column information is
larger in the LEB-like encoding used. Now we only store column
information when we produce non-symbolic stacks, since the increased
space needed to store the columns is instead in DWARF sections and can
be stripped or elided.
Original description:
Previously, we passed line number information to the stack trace printer
and to DWARF by changing the non-special positions in the CodeSourceMap
to line numbers in precompiled mode. However, doing this lost column
information.
We get the column information back in the majority of cases by encoding
the line number and column information when neither is too large to pack
together into 30 bits. (Here, 20 bits for line and 10 bits for column.)
Otherwise, we just store the line information as before, though due to
using a bit to encode whether column info exists, it's reduced to 30
bits. If the line info is too big for that, we just return kNoSourcePos.
Cq-Include-Trybots: luci.dart.try:vm-kernel-precomp-nnbd-linux-release-x64-try,vm-kernel-precomp-linux-release-x64-try,vm-kernel-precomp-linux-product-x64-try,vm-kernel-precomp-linux-debug-x64-try,vm-kernel-precomp-linux-release-simarm_x64-try
Change-Id: Ia8baee71468da6100a170fa305d03059ffd17f78
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/151822
Commit-Queue: Tess Strickland <sstrickl@google.com>
Reviewed-by: Martin Kustermann <kustermann@google.com>
This reverts commit 7c34a80c11.
Reason for revert: This change seems to cause >2% code size regression
in flutter-release mode.
Original change's description:
> [vm/aot] Keep column information when possible for precompiled mode.
>
> Previously, we passed line number information to the stack trace printer
> and to DWARF by changing the non-special positions in the CodeSourceMap
> to line numbers in precompiled mode. However, doing this lost column
> information.
>
> We get the column information back in the majority of cases by encoding
> the line number and column information when neither is too large to pack
> together into 30 bits. (Here, 20 bits for line and 10 bits for column.)
> Otherwise, we just store the line information as before, though due to
> using a bit to encode whether column info exists, it's reduced to 30
> bits. If the line info is too big for that, we just return kNoSourcePos.
>
> Cq-Include-Trybots: luci.dart.try:vm-kernel-precomp-nnbd-linux-release-x64-try,vm-kernel-precomp-linux-release-x64-try,vm-kernel-precomp-linux-product-x64-try,vm-kernel-precomp-linux-debug-x64-try,vm-kernel-precomp-linux-release-simarm_x64-try
> Change-Id: Id1c826f10871e2f304fa40a59d8b704404d3a2c9
> Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/151507
> Commit-Queue: Tess Strickland <sstrickl@google.com>
> Reviewed-by: Ryan Macnak <rmacnak@google.com>
TBR=rmacnak@google.com,alexmarkov@google.com,sstrickl@google.com
Change-Id: I5cf97543d1ac2731bb27bdb58ae97af6f22f2cfb
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Cq-Include-Trybots: luci.dart.try:vm-kernel-precomp-nnbd-linux-release-x64-try, vm-kernel-precomp-linux-release-x64-try, vm-kernel-precomp-linux-product-x64-try, vm-kernel-precomp-linux-debug-x64-try, vm-kernel-precomp-linux-release-simarm_x64-try
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/151820
Reviewed-by: Martin Kustermann <kustermann@google.com>
Commit-Queue: Martin Kustermann <kustermann@google.com>
Previously, we passed line number information to the stack trace printer
and to DWARF by changing the non-special positions in the CodeSourceMap
to line numbers in precompiled mode. However, doing this lost column
information.
We get the column information back in the majority of cases by encoding
the line number and column information when neither is too large to pack
together into 30 bits. (Here, 20 bits for line and 10 bits for column.)
Otherwise, we just store the line information as before, though due to
using a bit to encode whether column info exists, it's reduced to 30
bits. If the line info is too big for that, we just return kNoSourcePos.
Cq-Include-Trybots: luci.dart.try:vm-kernel-precomp-nnbd-linux-release-x64-try,vm-kernel-precomp-linux-release-x64-try,vm-kernel-precomp-linux-product-x64-try,vm-kernel-precomp-linux-debug-x64-try,vm-kernel-precomp-linux-release-simarm_x64-try
Change-Id: Id1c826f10871e2f304fa40a59d8b704404d3a2c9
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/151507
Commit-Queue: Tess Strickland <sstrickl@google.com>
Reviewed-by: Ryan Macnak <rmacnak@google.com>
This is a reland of 6a98e2719e
Original change's description:
> [VM/nnbd] Canonicalize TypeParameter objects in the VM.
>
> Prior to this CL, type parameters were all assumed canonical, even duplicate ones.
>
> Per a new convention introduced in this CL, the type parameter array in generic classes and generic functions contains canonical type parameters. As these type parameters get cloned with a different nullability, they are inserted in a new hash table of canonical type parameters.
>
> This fixes performance issue https://github.com/dart-lang/sdk/issues/41421
>
> Change-Id: I9086158fa6b6261e9997bb50edec6d7c54abbfa1
> Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/148223
> Commit-Queue: Régis Crelier <regis@google.com>
> Reviewed-by: Ryan Macnak <rmacnak@google.com>
Change-Id: I4a42d0f9d53c2244b1a638432ed2cd6a055fa72a
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/150266
Commit-Queue: Régis Crelier <regis@google.com>
Reviewed-by: Alexander Markov <alexmarkov@google.com>
Reviewed-by: Ryan Macnak <rmacnak@google.com>
Prior to this CL, type parameters were all assumed canonical, even duplicate ones.
Per a new convention introduced in this CL, the type parameter array in generic classes and generic functions contains canonical type parameters. As these type parameters get cloned with a different nullability, they are inserted in a new hash table of canonical type parameters.
This fixes performance issue https://github.com/dart-lang/sdk/issues/41421
Change-Id: I9086158fa6b6261e9997bb50edec6d7c54abbfa1
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/148223
Commit-Queue: Régis Crelier <regis@google.com>
Reviewed-by: Ryan Macnak <rmacnak@google.com>
Clean up signatures of runtime functions in object.{cc,h} using the trail to avoid cycles. Move the trail to the last parameter of these functions.
Introduce a trail in IsSubtypeOf() to avoid cycles introduced by bounds of F-bounded types. Issue uncovered by the new normalization tests.
Change-Id: I3241c7e4023a09c122e1594b7aff90b5b103f4f4
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/150180
Commit-Queue: Régis Crelier <regis@google.com>
Reviewed-by: Alexander Markov <alexmarkov@google.com>
Initializer functions are created lazily at compile time when generating
LoadField which calls initializer. This may happen in the background
compiler thread, where Field objects are cloned.
EnsureInitializerFunction should be called from original field, and it
should be properly synchronized to make sure there are no races with
mutator thread.
Change-Id: Icc339afa2ee410385019c72c717737719767d367
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/150166
Commit-Queue: Alexander Markov <alexmarkov@google.com>
Reviewed-by: Martin Kustermann <kustermann@google.com>
To avoid stopping all mutators when updating subtype test caches, we
ensure that any updates to the subtype test cache will be visible in a
consistent way to mutators by adding a store-release barrier when
updating the backing array of the STC.
Since STCs are always grown by allocating a new array with size+1, there
is no need to actually add barriers when accessing the individual
entries of the backing store.
To prevent multiple writers creating STCs and patching the pool or
updating a STC we add a mutex.
Issue https://github.com/dart-lang/sdk/issues/36097
Change-Id: Ic705619fe16772565ac11438fd15d3977cbb49c0
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/149164
Reviewed-by: Ryan Macnak <rmacnak@google.com>
Reviewed-by: Alexander Aprelev <aam@google.com>
Commit-Queue: Martin Kustermann <kustermann@google.com>
Regis Crelier