This is a reland of commit 4cd9c9c666
This will be relanded without changes because we discussed the .proto
definition package name clashing problem with the Dart in google3 team,
and we have decided that the best solution is to ignore the checked in
.proto files in google3 builds.
TEST=Recorded traces with the Perfetto file recorder and explored them
in Perfetto's trace viewer, CI
Original change's description:
> Reland "[VM] Begin supporting Perfetto file recorder"
>
> This is a reland of commit 7424295ce9
>
> The differences between this reland and the original CL are: now the
> Perfetto file recorder does not get built in PRODUCT builds, and it does
> not get built unless the SUPPORT_PERFETTO macro is defined.
>
> TEST=Recorded traces with the Perfetto file recorder and explored them
> in Perfetto's trace viewer, CI
>
> Original change's description:
> > [VM] Begin supporting Perfetto file recorder
> >
> > This CL adds the `TimelineEventPerfettoFileRecorder` class, which is a
> > timeline recorder that writes a trace to file in Perfetto's proto
> > format. This CL supports the recording of all types of timeline events
> > except flow events. Support for flow events will be added in a future
> > CL.
> >
> > TEST=Recorded traces with the Perfetto file recorder and explored them
> > in Perfetto's trace viewer, CI
> >
> > Change-Id: Iaa2051e536589a473c5e15f9de9bb9c251f13d38
> > Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/278942
> > Reviewed-by: Ben Konyi <bkonyi@google.com>
> > Commit-Queue: Derek Xu <derekx@google.com>
>
> Change-Id: I8713f704b5fbeed5f1231012bce8a32aaf566ae4
> Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/286020
> Reviewed-by: Ben Konyi <bkonyi@google.com>
> Commit-Queue: Derek Xu <derekx@google.com>
Change-Id: Ia97ef1f0fe73e76c7022623b9ae7e21a4ea73a46
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/295804
Commit-Queue: Derek Xu <derekx@google.com>
Reviewed-by: Ben Konyi <bkonyi@google.com>
This reverts commit 4cd9c9c666.
Reason for revert: reusing the same proto package for copied files breaks rolling it to G3.
Original change's description:
> Reland "[VM] Begin supporting Perfetto file recorder"
>
> This is a reland of commit 7424295ce9
>
> The differences between this reland and the original CL are: now the
> Perfetto file recorder does not get built in PRODUCT builds, and it does
> not get built unless the SUPPORT_PERFETTO macro is defined.
>
> TEST=Recorded traces with the Perfetto file recorder and explored them
> in Perfetto's trace viewer, CI
>
> Original change's description:
> > [VM] Begin supporting Perfetto file recorder
> >
> > This CL adds the `TimelineEventPerfettoFileRecorder` class, which is a
> > timeline recorder that writes a trace to file in Perfetto's proto
> > format. This CL supports the recording of all types of timeline events
> > except flow events. Support for flow events will be added in a future
> > CL.
> >
> > TEST=Recorded traces with the Perfetto file recorder and explored them
> > in Perfetto's trace viewer, CI
> >
> > Change-Id: Iaa2051e536589a473c5e15f9de9bb9c251f13d38
> > Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/278942
> > Reviewed-by: Ben Konyi <bkonyi@google.com>
> > Commit-Queue: Derek Xu <derekx@google.com>
>
> Change-Id: I8713f704b5fbeed5f1231012bce8a32aaf566ae4
> Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/286020
> Reviewed-by: Ben Konyi <bkonyi@google.com>
> Commit-Queue: Derek Xu <derekx@google.com>
Change-Id: I2da67ffe5bff01d568f32f682455bb2c789499b0
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/295821
Commit-Queue: Ilya Yanok <yanok@google.com>
Bot-Commit: Rubber Stamper <rubber-stamper@appspot.gserviceaccount.com>
Reviewed-by: Alexander Thomas <athom@google.com>
Reviewed-by: Oleh Prypin <oprypin@google.com>
Reviewed-by: Ivan Inozemtsev <iinozemtsev@google.com>
This is a reland of commit 7424295ce9
The differences between this reland and the original CL are: now the
Perfetto file recorder does not get built in PRODUCT builds, and it does
not get built unless the SUPPORT_PERFETTO macro is defined.
TEST=Recorded traces with the Perfetto file recorder and explored them
in Perfetto's trace viewer, CI
Original change's description:
> [VM] Begin supporting Perfetto file recorder
>
> This CL adds the `TimelineEventPerfettoFileRecorder` class, which is a
> timeline recorder that writes a trace to file in Perfetto's proto
> format. This CL supports the recording of all types of timeline events
> except flow events. Support for flow events will be added in a future
> CL.
>
> TEST=Recorded traces with the Perfetto file recorder and explored them
> in Perfetto's trace viewer, CI
>
> Change-Id: Iaa2051e536589a473c5e15f9de9bb9c251f13d38
> Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/278942
> Reviewed-by: Ben Konyi <bkonyi@google.com>
> Commit-Queue: Derek Xu <derekx@google.com>
Change-Id: I8713f704b5fbeed5f1231012bce8a32aaf566ae4
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/286020
Reviewed-by: Ben Konyi <bkonyi@google.com>
Commit-Queue: Derek Xu <derekx@google.com>
This reverts commit 7424295ce9.
Reason for revert: Causes VM crash in internal tests - b/271081079
Original change's description:
> [VM] Begin supporting Perfetto file recorder
>
> This CL adds the `TimelineEventPerfettoFileRecorder` class, which is a
> timeline recorder that writes a trace to file in Perfetto's proto
> format. This CL supports the recording of all types of timeline events
> except flow events. Support for flow events will be added in a future
> CL.
>
> TEST=Recorded traces with the Perfetto file recorder and explored them
> in Perfetto's trace viewer, CI
>
> Change-Id: Iaa2051e536589a473c5e15f9de9bb9c251f13d38
> Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/278942
> Reviewed-by: Ben Konyi <bkonyi@google.com>
> Commit-Queue: Derek Xu <derekx@google.com>
Change-Id: I1c4b0b78be3748b525b82ea66079dd95d412d82e
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/285906
Auto-Submit: Oleh Prypin <oprypin@google.com>
Bot-Commit: Rubber Stamper <rubber-stamper@appspot.gserviceaccount.com>
Reviewed-by: Alexander Thomas <athom@google.com>
Commit-Queue: Alexander Thomas <athom@google.com>
This CL adds the `TimelineEventPerfettoFileRecorder` class, which is a
timeline recorder that writes a trace to file in Perfetto's proto
format. This CL supports the recording of all types of timeline events
except flow events. Support for flow events will be added in a future
CL.
TEST=Recorded traces with the Perfetto file recorder and explored them
in Perfetto's trace viewer, CI
Change-Id: Iaa2051e536589a473c5e15f9de9bb9c251f13d38
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/278942
Reviewed-by: Ben Konyi <bkonyi@google.com>
Commit-Queue: Derek Xu <derekx@google.com>
Allows embedders to programmatically request heap snapshot from a
running isolate group instead of relying on a vm-service or hidden
internal Dart APIs
Additionally we allow to include snapshot writing functionality into
PRODUCT builds using DART_ENABLE_HEAP_SNAPSHOT_WRITER define.
TEST=vm/cc/DartAPI_WriteHeapSnapshot
Bug: b/259115846
Change-Id: Ic3ef76e5fb9adcf8f23a1959f5238742b64ecde2
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/273181
Reviewed-by: Daco Harkes <dacoharkes@google.com>
Commit-Queue: Slava Egorov <vegorov@google.com>
Previously, the VM used a linear array to cache previous instantiations
of a type arguments object. Now once the cache hits a certain number
of occupied entries, the VM changes to using a hash-based approach.
The InstantiateTypeArguments stubs have not yet been updated to
traverse the hash-based cache, so once the cache has grown too large,
all attempts at instantiations, even those that are in the cache, go to
the runtime. Thus, until the stubs are updated, this is only an
improvement if the cost of traversing the linear cache dominates the
cost of making a runtime call. Our benchmarks see a ~40% performance
regression for hash-based caches of size 100 but a ~400% performance
improvement for hash-based caches of size 1000. Thus, we currently
split the difference and set the maximum size of linear caches to 500.
TEST=vm/cc/TypeArguments_Cache_ManyInstantiations
Bug: https://github.com/dart-lang/sdk/issues/48344
Change-Id: I7f1376943523bb5bcd8b175cfb1936779ea73d60
Cq-Include-Trybots: luci.dart.try:vm-kernel-precomp-dwarf-linux-product-x64-try,vm-kernel-precomp-linux-product-x64-try,vm-kernel-precomp-linux-release-x64-try,vm-kernel-precomp-nnbd-mac-release-arm64-try,vm-kernel-precomp-nnbd-linux-release-simarm_x64-try,vm-kernel-precomp-linux-release-simarm-try,vm-kernel-precomp-nnbd-linux-release-x64-try,vm-kernel-precomp-nnbd-linux-release-simarm64-try,vm-kernel-precomp-nnbd-linux-debug-simriscv64-try,vm-kernel-precomp-tsan-linux-release-x64-try,vm-kernel-tsan-linux-release-x64-try
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/265325
Reviewed-by: Martin Kustermann <kustermann@google.com>
Reviewed-by: Ryan Macnak <rmacnak@google.com>
Commit-Queue: Tess Strickland <sstrickl@google.com>
To create a Dart standalone executable on MacOS, we modify the
dartaotruntime executable to add the snapshot contents, and the
VM looks into the executable on disk to find the snapshot to load.
Previously, we did this by adding a new 64-bit segment load command
with a single section, where the section's file offset and size
describes the inserted snapshot. This meant the Mach-O header size increased by 152 bytes.
Originally, this wasn't an issue as there was plenty of padding, but
later clang updates removed most of this padding, and so writing the
new header actually overwrote the initial contents of the first section
in the file, which happens to be the __text section. In addition, since
the first section's offset was now declared to be within the header,
utilities that strictly validated the Mach-O format, like codesign,
would report errors.
This CL changes it so that we actually reserve space in the
dartaotruntime header using the -add_empty_section flag to the linker.
In addition, we change from using a segment load command to using a
(40 byte) note load command. This is because a segment load command
specifies that the contents should be loaded in memory, but we don't
use that loaded version. Instead, the VM reloads it from the executable
on disk so it can appropriately mmap the different parts of the
snapshot. A note section instead just declares a section of the
executable as arbitrary data that the owner can read from the file
and use as desired, which is semantically closer to our current usage.
This CL also adds a test to pkg/dartdev/test/commands/compile_test to
ensure that corrupting a random part of the snapshot in the executable
causes signature verification to fail.
This CL also reverts CL 256208, thus relanding the clang changes
starting from June that originally raised awareness of the issue by
greatly reduced the amount of padding after the load commands.
TEST=pkg/dartdev/test/commands/compile_test
Bug: https://github.com/dart-lang/sdk/issues/49783
Change-Id: Iee554d87b0eabaecd7a534ca4e4facfefbce6385
Cq-Include-Trybots: luci.dart.try:analyzer-mac-release-try,dart-sdk-mac-arm64-try,dart-sdk-mac-try,pkg-mac-release-arm64-try,pkg-mac-release-try,vm-kernel-precomp-mac-product-x64-try,vm-kernel-precomp-nnbd-mac-release-arm64-try
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/260108
Reviewed-by: Ryan Macnak <rmacnak@google.com>
Reviewed-by: Daco Harkes <dacoharkes@google.com>
Commit-Queue: Tess Strickland <sstrickl@google.com>
Instead apply the same approach as we do in AOT: unbox based on the
static type information. There are no TFA results available in JIT,
but we could still unbox fields when running in sound null-safety.
TEST=ci
Cq-Include-Trybots: luci.dart.try:vm-kernel-reload-linux-release-x64-try,vm-kernel-reload-linux-debug-x64-try,vm-kernel-reload-rollback-linux-debug-x64-try,vm-kernel-reload-rollback-linux-release-x64-try,vm-kernel-precomp-linux-debug-x64-try,vm-kernel-precomp-linux-product-x64-try,vm-kernel-precomp-linux-release-x64-try,vm-kernel-precomp-nnbd-linux-release-simarm64-try,vm-kernel-linux-debug-simriscv64-try,vm-kernel-precomp-linux-debug-simriscv64-try,vm-kernel-nnbd-linux-release-ia32-try,vm-kernel-nnbd-linux-debug-x64-try,vm-kernel-nnbd-linux-debug-ia32-try,vm-kernel-nnbd-linux-release-simarm-try,vm-kernel-nnbd-linux-release-simarm64-try
Change-Id: Ide2e78c6659261ef8d245a4586cf699ea0fbb459
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/256211
Reviewed-by: Ryan Macnak <rmacnak@google.com>
Commit-Queue: Slava Egorov <vegorov@google.com>
Reviewed-by: Martin Kustermann <kustermann@google.com>
Reviewed-by: Alexander Markov <alexmarkov@google.com>
* Merge ClassTable and SharedClassTable back together;
* Simplify handling of multiple arrays growing in sync;
* Refactor how reload deals with ClassTable.
The last change is the most important because it makes it
much easier to reason about the code. We move away from
copying bits and pieces of the class table and shared
class table into reload contexts.
Having two class table fields in the isolate group makes
it easier to reason about. One field contains program
class table (one modified by kernel loader and accessed
by various program structure cid lookups) and heap
walk class table (used by GC visitors). Normally these
two fields point to the same class table, but during
hot reload we temporary split them apart: original
class table is kept as a heap walk class table, while
program class table is replaced by a clone and updated
by reload.
If reload succeeds we drop original class table and
set program class table as heap walk one.
If reload fails we drop the program class table and
restore original one from heap walk table.
TEST=ci
Cq-Include-Trybots: luci.dart.try:vm-kernel-reload-linux-release-x64-try,vm-kernel-reload-linux-debug-x64-try,vm-kernel-reload-rollback-linux-debug-x64-try,vm-kernel-reload-rollback-linux-release-x64-try,vm-kernel-linux-debug-x64-try,vm-kernel-precomp-tsan-linux-release-x64-try,vm-kernel-tsan-linux-release-x64-try,vm-kernel-precomp-asan-linux-release-x64-try,vm-kernel-asan-linux-release-x64-try
Change-Id: I8b66259fcc474dea7dd2af063e4772df99be06c4
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/258361
Commit-Queue: Slava Egorov <vegorov@google.com>
Reviewed-by: Ryan Macnak <rmacnak@google.com>
Instead of using an integer literal use constexpr popcount
function instead.
The misalignment causes an incorrect generation of a call
to a shared write barrier stub for the last allocatable register
on ARM64.
Reported by Lin Zuojian <zuojian.lzj@alibaba-inc.com>
TEST=ci
Change-Id: I69ce32da958573be0ec8967e462f8a378a778a28
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/254401
Commit-Queue: Slava Egorov <vegorov@google.com>
Reviewed-by: Ryan Macnak <rmacnak@google.com>
Examples of the new line added to non-symbolic stack traces:
os: linux arch: x64 comp: yes sim: no
(Running on linux-x64c)
os: macos arch: arm64 comp: no sim: yes
(Running on mac-simarm64)
This CL also abstracts out the separate hardcoded strings across
the codebase for host and target OS and architecture into
definitions in platform/globals.h to ensure that they stay
in sync across different uses.
TEST=vm/dart{,_2}/use_dwarf_stack_traces_flag
Issue: https://github.com/flutter/flutter/pull/101586
Change-Id: Ifdfea5138dd1003f561da0174e89aebc165bf9b0
Cq-Include-Trybots: luci.dart.try:vm-kernel-precomp-dwarf-linux-product-x64-try,vm-kernel-precomp-linux-release-simarm-try,vm-kernel-precomp-linux-release-simarm64-try,vm-kernel-precomp-linux-release-simarm_x64-try,vm-kernel-precomp-linux-release-x64-try,vm-kernel-precomp-mac-product-x64-try,vm-kernel-precomp-nnbd-linux-release-x64-try,vm-kernel-precomp-nnbd-linux-release-simarm_x64-try,vm-kernel-precomp-win-release-x64-try,vm-kernel-precomp-nnbd-mac-release-arm64-try,vm-kernel-precomp-mac-release-simarm64-try,vm-ffi-android-release-arm-try,vm-ffi-android-release-arm64c-try
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/253283
Reviewed-by: Ryan Macnak <rmacnak@google.com>
Commit-Queue: Tess Strickland <sstrickl@google.com>
Reviewed-by: Daco Harkes <dacoharkes@google.com>
Even if the main program disables executable stacks, dlopen'ing a library that doesn't itself disable executable stack will switch the stack to executable. (Presumably in the name of compatability with GNU nested functions.)
Dart does not need executable stacks, and executable stacks are an additional attack surface.
TEST=readelf
Bug: b/229648756
Change-Id: Ia8c234ebc6178a26528d37b1359a10dd42039a9b
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/241540
Reviewed-by: Tess Strickland <sstrickl@google.com>
Commit-Queue: Ryan Macnak <rmacnak@google.com>
* Use CompilerPass::GenerateCode in AOT to ensure we print
the graph if the code generation crashes.
* Enable Dart_DumpNativeStackTrace in gen_snapshot even in
PRODUCT builds.
Additionally enable disassembler in gen_snapshot even in
PRODUCT builds.
TEST=tested manually
Change-Id: I951d62ca07c7882fa1d8e765d4b5ffd26edd9456
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/233880
Reviewed-by: Alexander Markov <alexmarkov@google.com>
Auto-Submit: Slava Egorov <vegorov@google.com>
Commit-Queue: Slava Egorov <vegorov@google.com>
This CL implements a slightly altered compilation process for `dart
compile` on MacOS. Instead of the traditional approach of
concatenating a dart snapshot with the dart runtime, this new
implementation uses a new MachO segment to store the snapshot. Doing
so ensure users can properly sign the resulting binaries.
The dart runtime has also been updated to look for the snapshots in
this new segment.
There are new unit tests ensuring the dart runtime can correctly
identify MachO-formatted executables and an end-to-end test ensuring
`dart compile` produces code that has the new segment and produces
code that can be signed.
TEST=pkg/dartdev/test/commands/compile_test
Change-Id: Ibdce80455e02bc2b6fd345d80be8d75f989a9c28
Cq-Include-Trybots: luci.dart.try:analyzer-mac-release-try,dart-sdk-mac-arm64-try,dart-sdk-mac-try,pkg-mac-release-arm64-try,pkg-mac-release-try,vm-kernel-mac-debug-x64-try,vm-kernel-mac-product-x64-try,vm-kernel-mac-release-arm64-try,vm-kernel-mac-release-x64-try
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/228080
Reviewed-by: Martin Kustermann <kustermann@google.com>
Reviewed-by: Tess Strickland <sstrickl@google.com>
Commit-Queue: Tess Strickland <sstrickl@google.com>
Previously all arrays were excluded from WB elimination pass
to avoid invariant restoration code creating excessive work
for the GC and to avoid dealing with card marking in the
invariant restoration code.
It seems reasonable to enable this for small arrays of up to 8
elements. The cut of point of 8 elements was chosen based on the
cut of point for literal<N> specialisations provided by the
core library for creating small literal arrays.
TEST=vm/cc/IRTest_WriteBarrierElimination_Arrays
Cq-Include-Trybots: luci.dart.try:vm-kernel-precomp-dwarf-linux-product-x64-try,vm-kernel-precomp-linux-debug-simarm_x64-try,vm-kernel-precomp-linux-debug-x64-try,vm-kernel-precomp-linux-debug-x64c-try,vm-kernel-precomp-linux-product-x64-try,vm-kernel-precomp-linux-release-simarm-try,vm-kernel-precomp-linux-release-simarm64-try,vm-kernel-precomp-linux-release-simarm_x64-try,vm-kernel-precomp-linux-release-x64-try
Change-Id: I2b3169865f07c3ff95820c1bc6718943e96bd33b
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/229903
Reviewed-by: Alexander Markov <alexmarkov@google.com>
Reviewed-by: Ryan Macnak <rmacnak@google.com>
Commit-Queue: Slava Egorov <vegorov@google.com>
Implements a backend targeting RV32GC and RV64GC, based on Linux standardizing around GC. The assembler is written to make it easy to disable usage of C, but because the sizes of some instruction sequences are compile-time constants, an additional build configuration would need to be defined to make use of it.
The assembler and disassembler cover every RV32/64GC instruction. The simulator covers all instructions except accessing CSRs and the floating point state accessible through such, include accrued exceptions and dynamic rounding mode.
Quirks:
- RISC-V is a compare-and-branch architecture, but some existing "architecture-independent" parts of the Dart compiler assume a condition code architecture. To avoid rewriting these parts, we use a peephole in the assembler to map to compare-and-branch. See Assembler::BranchIf. Luckily nothing depended on taking multiple branches on the same condition code set.
- There are no hardware overflow checks, so we must use Hacker's Delight style software checks. Often these are very cheap: if the sign of one operand is known, a single branch is needed.
- The ranges of RISC-V branches and jumps are such that we use 3 levels of generation for forward branches, instead of the 2 levels of near and far branches used on ARM[64]. Nearly all code is handled by the first two levels with 20-bits of range, with enormous regex matchers triggering the third level that uses aupic+jalr to get 32-bits of range.
- For PC-relative calls in AOT, we always generate auipc+jalr pairs with 32-bits of range, so we never generate trampolines.
- Only a subset of registers are available in some compressed instructions, so we assign the most popular uses to these registers. In particular, THR, TMP[2], CODE and PP. This has the effect of assigning CODE and PP to volatile registers in the C calling convention, whereas they are assigned preserved registers on the other architectures. As on ARM64, PP is untagged; this is so short indices can be accessed with a compressed instruction.
- There are no push or pop instructions, so combining pushes and pops is preferred so we can update SP once.
- The C calling convention has a strongly aligned stack, but unlike on ARM64 we don't need to use an alternate stack pointer. The author ensured language was added to the RISC-V psABI making the OS responsible for realigning the stack pointer for signal handlers, allowing Dart to leave the stack pointer misaligned from the C calling convention's point of view until a foreign call.
- We don't bother with the link register tracking done on ARM[64]. Instead we make use of an alternate link register to avoid inline spilling in the write barrier.
Unimplemented:
- non-trivial FFI cases
- Compressed pointers - No intention to implement.
- Unboxed SIMD - We might make use of the V extension registers when the V extension is ratified.
- BigInt intrinsics
TEST=existing tests for IL level, new tests for assembler/disassembler/simulator
Bug: https://github.com/dart-lang/sdk/issues/38587
Bug: https://github.com/dart-lang/sdk/issues/48164
Change-Id: I991d1df4be5bf55efec5371b767b332d37dfa3e0
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/217289
Reviewed-by: Alexander Markov <alexmarkov@google.com>
Reviewed-by: Daco Harkes <dacoharkes@google.com>
Reviewed-by: Slava Egorov <vegorov@google.com>
Commit-Queue: Ryan Macnak <rmacnak@google.com>
Having uint32_t field in UntaggedCompressedStackMaps::Payload
implies 4 byte alignment of the whole structure and the field
itself. We would prefer, however, to avoid this requirement because
we are packing these structures tight in AOT snapshots without any
padding in between.
This change rewrites the implementation to use `memcpy(...)` to remove
any alignment requirements. Note that this `memcpy(...)` will
be optimized down to a single memory move on all platforms we currently
support as they support necessary unaligned load/store instructions.
This also means that the original code worked just fine, except when
running under UBSAN.
Fixes https://github.com/dart-lang/sdk/issues/47971
TEST=ubsan bots
Cq-Include-Trybots: luci.dart.try:vm-kernel-precomp-ubsan-linux-release-x64-try,vm-kernel-ubsan-linux-release-x64-try
Change-Id: I26a30123fcbc6d2c711f039f15f749913ce4d7bd
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/226100
Reviewed-by: Alexander Aprelev <aam@google.com>
Commit-Queue: Slava Egorov <vegorov@google.com>
The failure is within std::ostringstream without any obvious errors
on our side, but just appending to std::string instead should be
fine for a failure case.
TEST=vm/cc/IsolateReload_PendingUnqualifiedCall_StaticToInstance on MSAN trybot
Change-Id: I36e91c4ba4019856c709834d29ce9c4595e6a336
Cq-Include-Trybots: luci.dart.try:vm-kernel-msan-linux-release-x64-try
Bug: https://github.com/dart-lang/sdk/issues/47933
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/224200
Reviewed-by: Daco Harkes <dacoharkes@google.com>
Reviewed-by: Martin Kustermann <kustermann@google.com>
Commit-Queue: Tess Strickland <sstrickl@google.com>
- Fix disassembly of 32-bit stxr/ldxr/stlr/ldar.
- Add missing disassembly of rbit.
- Use the standard names for arithmetic/logical ops with immediates.
- Fix name of csel.
- Fix operand ordering for conditional moves.
- Add preferred disassembly 'ret lr' -> 'ret'.
- Add preferred disassembly 'ands zr, a, b' -> 'tst a, b'.
- Add preferred disassembly 'sub a, zr, b' -> 'neg a, b'.
Overall our disassembly is nearly the same as standard A64 disassembly, except we use a "w" suffix for 32-bit operations instead of using "w" versus "x" register names, and we use our own register names.
TEST=ci
Change-Id: Ifb4e587b5cc490beb4c6a09145149d95b09d412d
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/223680
Reviewed-by: Alexander Markov <alexmarkov@google.com>
Commit-Queue: Ryan Macnak <rmacnak@google.com>
TSAN instruments C++ code by adding prologue/epilogue code which
maintains a shadow stack. Using setjmp()/longjmp() is intercepted by
TSAN and correspondingly unwinds the shadow stack.
When Dart VM throws exceptions we call the JumpToFrame stub
from C++ which will directly reset the stack to the exception handler
catch entry. This leaves the TSAN shadow stack unchanged.
This means whenever an exception is thrown we leak frames in TSAN's
shadow stack. Due to using a fixed-size shadow stack, it will cause a
buffer-overflow in TSAN when too many such frame leaks happen. This can
cause arbitrary memory to be overriden, leading to awkward crashes.
This is especially an issue on the "iso-stres" builder because it
launches - in the same process - *many* small tests, more easily hitting
that limit.
This CL will workaround the TSAN issue by making runtime call save it's
state via setjmp() and make exception throughing process go via
longjmp() (which TSAN will intercept) before actually calling the
JumpToFrame stub.
=> This will ensure the TSAN shadow stack is correctly maintained.
The [jmp_buf]'s encoding of register state is non-trivial (e.g. it uses
XOR'ing of the actual saved state under certain glibc versions). So we
store any state we need to pass to the target of the `longjmp()` on the
[Thread] instead of overriding the [jmp_buf]s register state with the
arguments.
Issue https://github.com/dart-lang/sdk/issues/47472#issuecomment-948235479
TEST=vm/dart{,_2}/regress47472_test.dart
Change-Id: Ifbf6580aa15bcce54d0584cdc3cd18cc19be0a9c
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/222300
Reviewed-by: Ryan Macnak <rmacnak@google.com>
Commit-Queue: Martin Kustermann <kustermann@google.com>
Derek Xu