Right each `Pointer.fromFunction()` invocation will lead to creation of
a new ffi trampoline function & it's following JITed code. In AOT we
have exactly one ffi trampoline per target/native-signature/exceptional-return
combination.
=> This CL ensures we have only one such function.
Furthermore each `Pointer.fromFunction()` will currently perform 2
runtime calls in JIT: One to create a `Function` object, the other to
JIT that function & register callback metadata.
=> This CL ensures we won't do a runtime call to get a function, instead
do it at compile-time (as in AOT)
Furthermore we eagerly assign a callback-id to the unique/deduped ffi
trampoline callbacks. Only when the application requests a pointer, do
we populate metadata on the `Thread` object.
This CL doesn't (yet) change the fact that in JIT mode we have
isolate-specific jit trampolines (that will call now shared ffi trampoline
functions).
We also avoid baking in C++ runtime function pointers in generated
code. As a result we can now preserve ffi trampolines across AppJIT
serialization.
As a nice side-effect, we remove 100 lines of code.
TEST=ffi{,_2}/ffi_callback_unique_test
Issue https://github.com/dart-lang/sdk/issues/50611
Change-Id: I458831a47b041a088086f28f825de2a3849f6adc
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/273420
Reviewed-by: Daco Harkes <dacoharkes@google.com>
Commit-Queue: Martin Kustermann <kustermann@google.com>
This CL introduces new embedding APIs for supporting heap sample
profiling. A registered sampling callback is invoked approximately every
N bytes based on an exponential distribution, providing information
about the isolate group the allocation occurred in, the user visible
name of the allocated object type, a weak persistent handle to the
allocated object, and the size of the allocation.
Sampling is triggered using artificial TLAB boundaries to cause
allocations to be sampled to take the allocation slow path where the
registered callback can be invoked with the allocation information.
Only new space allocations are currently traced, with old space
allocation support to be added in a future CL.
TEST=Dart_HeapSampling
Change-Id: I22bcdeec6e823bc1ab44898d4c596fbed7169fa1
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/264520
Commit-Queue: Ben Konyi <bkonyi@google.com>
Reviewed-by: Siva Annamalai <asiva@google.com>
Reviewed-by: Ryan Macnak <rmacnak@google.com>
Enable concurrent marking for IA32.
This removes the last write barrier relying on different alignment offsets for old and new space.
TEST=ci
Change-Id: Ib1c13124002392cf1c3ec264643325ec471a6918
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/262280
Commit-Queue: Ryan Macnak <rmacnak@google.com>
Reviewed-by: Alexander Markov <alexmarkov@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>
`MournFinalized` runs during marking and can add objects to the store
buffer. These objects are stored in the threads' store buffer block.
This block needs to be released to the central store buffer in order for
the objects' addresses to be updated during compacting.
TEST=runtime/vm/object_test.cc
TEST=tools/test.py vm/cc/Finalizer_Regress_48843
Closes: https://github.com/dart-lang/sdk/issues/48843
Change-Id: Ib2424929c86fee730d3f09fbd2f9f6c97f31abfd
Cq-Include-Trybots: luci.dart.try:vm-canary-linux-debug-try,vm-kernel-linux-debug-x64-try,vm-kernel-linux-debug-x64c-try,vm-kernel-mac-debug-x64-try,vm-kernel-win-debug-x64-try,vm-kernel-linux-release-x64-try,vm-kernel-linux-product-x64-try,vm-kernel-linux-debug-ia32-try,vm-kernel-linux-release-ia32-try
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/243262
Reviewed-by: Slava Egorov <vegorov@google.com>
Commit-Queue: Daco Harkes <dacoharkes@google.com>
Reviewed-by: Ryan Macnak <rmacnak@google.com>
If the popular store buffer targets are copied instead of promoted, the store buffer won't shrink and a second scavenge will occur that does promote them.
TEST=ci
Change-Id: I09fd7f263a2a9e99179af7ee58bc93335c4f5d05
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/241203
Reviewed-by: Siva Annamalai <asiva@google.com>
Commit-Queue: Ryan Macnak <rmacnak@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>
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>
Add a GCReason for store buffer overflow instead of using a separate print.
This is a pure refactoring that doesn't change GC policy.
TEST=ci
Change-Id: Idcd95b0701a36f45ebaa02db2c2fac9a9081d7c6
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/215102
Reviewed-by: Liam Appelbe <liama@google.com>
Commit-Queue: Ryan Macnak <rmacnak@google.com>
1) double.truncate() now simply calls toInt(), omitting
truncateToDouble() call.
2) double.floor() and ceil() are now intrinsified on arm64 and AOT/x64
and generated using DoubleToInteger instruction.
3) DoubleToInteger instruction is extended to support floor and ceil.
On arm64 DoubleToInteger is implemented using fcvtms and fcvtps
instructions. On x64 DoubleToInteger is implemented using roundsd
under a check if it is supported (with a fallback to a stub and a
runtime call).
AOT/x64:
Before: BenchFloor(RunTime): 318.82148549569655 us.
After: BenchFloor(RunTime): 133.29430189936687 us.
TEST=ci
Closes https://github.com/dart-lang/sdk/issues/46876
Closes https://github.com/dart-lang/sdk/issues/46650
Change-Id: I16ca18faf97954f8e8e25f0b72a2bbfac5bdc672
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/212381
Commit-Queue: Alexander Markov <alexmarkov@google.com>
Reviewed-by: Slava Egorov <vegorov@google.com>
In the unoptimized code unboxed double inputs are now unboxed in the
instruction prologue, and unboxed double output is boxed in the
instruction epilogue. Double values are still always boxed on the
expression stack in the unoptimized mode.
Support for unboxed doubles in unoptimized code allows us to have
one implementation of double and dart:math built-in functions,
shared across optimized and unoptimized modes.
TEST=ci
Issue: https://github.com/dart-lang/sdk/issues/46650
Change-Id: Ic1e42b72ee80cb2c12019b7abab8111240b05efd
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/211302
Commit-Queue: Alexander Markov <alexmarkov@google.com>
Reviewed-by: Slava Egorov <vegorov@google.com>
This reverts commit 1796160606.
Reason for revert: Still seeing failures.
Original change's description:
> Reland "[ package:dds ] Add support for caching CPU samples based on UserTag"
>
> This reverts commit 4160747ef6.
>
> It seems like 9397b8ff05 should have fixed the original failures we were seeing and the failing bots in question seem to be passing when running try jobs.
>
> TBR=asiva@ (discussed offline)
>
> TEST=Existing service tests + get_cached_cpu_samples_test
> Change-Id: I1c50c0e79375df819a0bd68e68ac28c3064874c8
> Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/208441
> Reviewed-by: Ben Konyi <bkonyi@google.com>
> Commit-Queue: Ben Konyi <bkonyi@google.com>
TBR=bkonyi@google.com,asiva@google.com
Change-Id: Ic522ab61b430673533c94ecd0c91bee569a0add6
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/208960
Reviewed-by: Ben Konyi <bkonyi@google.com>
Commit-Queue: Ben Konyi <bkonyi@google.com>
This reverts commit 4160747ef6.
It seems like 9397b8ff05 should have fixed the original failures we were seeing and the failing bots in question seem to be passing when running try jobs.
TBR=asiva@ (discussed offline)
TEST=Existing service tests + get_cached_cpu_samples_test
Change-Id: I1c50c0e79375df819a0bd68e68ac28c3064874c8
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/208441
Reviewed-by: Ben Konyi <bkonyi@google.com>
Commit-Queue: Ben Konyi <bkonyi@google.com>
DDS can be configured to listen for CPU sample events and cache samples
that were collected while certain UserTags are active. These cached
samples are stored in a ring buffer and are stored until the isolate
shuts down.
TEST=pkg/dds/test/get_cached_cpu_samples_test.dart
Change-Id: Ib20770f59f1672c703413486f87795b3bb23f676
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/207206
Commit-Queue: Ben Konyi <bkonyi@google.com>
Reviewed-by: Kenzie Schmoll <kenzieschmoll@google.com>
It's mainly hoisting an expression out of
IsolateGroup::{Increase,Decrease}MutatorCount to callers of it.
Right now we consider a Thread::{Exit,Enter} to be nested if there are
existing Dart frames on the stack. A future CL will have a case where
there's a nested exit&re-enter without Dart frames on the stack.
Trying to do this automatically by checking whether there's an
active api scope / zone on the thread fails due to some
asymmetry in our codebase.
Issue https://github.com/dart-lang/sdk/issues/36097
TEST=Refactoring of existing code.
Change-Id: Id70eda03efb2099487c5f0c8affd25ddff926553
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/186287
Commit-Queue: Martin Kustermann <kustermann@google.com>
Reviewed-by: Alexander Aprelev <aam@google.com>
Reviewed-by: Ryan Macnak <rmacnak@google.com>
Currently we have things called XPtr which are not what you get from ptr().
Old world:
handle->raw() returns RawObject* (tagged)
raw_obj->ptr() returns RawObject* (untagged)
After 6fe15f6df9:
handle->raw() returns ObjectPtr
obj_ptr->ptr() returns ObjectLayout*
New world:
handle->ptr() returns ObjectPtr
obj_ptr->untag() returns UntaggedObject*
TEST=ci
Change-Id: I6c7f34014cf20737607caaf84979838300d12df2
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/149367
Commit-Queue: Ryan Macnak <rmacnak@google.com>
Reviewed-by: Martin Kustermann <kustermann@google.com>
Reviewed-by: Siva Annamalai <asiva@google.com>
Reviewed-by: Vyacheslav Egorov <vegorov@google.com>
This CL adds passing structs by value in FFI trampolines.
Nested structs and inline arrays are future work.
C defines passing empty structs as undefined behavior, so that is not
supported in this CL.
Suggested review order:
1) commit message
2) ffi/marshaller (decisions for what is done in IL and what in MC)
3) frontend/kernel_to_il (IL construction)
4) backend/il (MC generation from IL)
5) rest in VM
Overall architecture is that structs are split up into word-size chunks
in IL when this is possible: 1 definition in IL per chunk, 1 Location in
IL per chunk, and 1 NativeLocation for the backend per chunk.
In some cases it is not possible or less convenient to split into
chunks. In these cases TypedDataBase objects are stored into and loaded
from directly in machine code.
The various cases:
- FFI call arguments which are not passed as pointers: pass individual
chunks to FFI call which already have the right location.
- FFI call arguments which are passed as pointers: Pass in TypedDataBase
to FFI call, allocate space on the stack, and make a copy on the stack
and pass the copies' address to the callee.
- FFI call return value: pass in TypedData to FFI call, and copy result
in machine code.
- FFI callback arguments which are not passed as pointers: IL definition
for each chunk, and populate a new TypedData with those chunks.
- FFI callback arguments which are passed as pointer: IL definition for
the pointer, and copying of contents in IL.
- FFI return value when location is pointer: Copy data to callee result
location in IL.
- FFI return value when location is not a pointer: Copy data in machine
code to the right registers.
Some other notes about the implementation:
- Due to Store/LoadIndexed loading doubles from float arrays, we use
a int32 instead and use the BitCastInstr.
- Linux ia32 uses `ret 4` when returning structs by value. This requires
special casing in the FFI callback trampolines to either use `ret` or
`ret 4` when returning.
- The 1 IL definition, 1 Location, and 1 NativeLocation approach does
not remove the need for special casing PairLocations in the machine
code generation because they are 1 Location belonging to 1 definition.
Because of the amount of corner cases in the calling conventions that
need to be covered, the tests are generated, rather than hand-written.
ABIs tested on CQ: x64 (Linux, MacOS, Windows), ia32 (Linux, Windows),
arm (Android softFP, Linux hardFP), arm64 Android.
ABIs tested locally through Flutter: ia32 Android (emulator), x64 iOS
(simulator), arm64 iOS.
ABIs not tested: arm iOS.
TEST=runtime/bin/ffi_test/ffi_test_functions_generated.cc
TEST=runtime/bin/ffi_test/ffi_test_functions.cc
TEST=tests/{ffi,ffi_2}/function_structs_by_value_generated_test.dart
TEST=tests/{ffi,ffi_2}/function_callbacks_structs_by_value_generated_tes
TEST=tests/{ffi,ffi_2}/function_callbacks_structs_by_value_test.dart
TEST=tests/{ffi,ffi_2}/vmspecific_static_checks_test.dart
Closes https://github.com/dart-lang/sdk/issues/36730.
Change-Id: I474d3a4ee1faadbe767ddadd1b696e24d8dc364c
Cq-Include-Trybots: luci.dart.try:dart-sdk-linux-try,dart-sdk-mac-try,dart-sdk-win-try,vm-ffi-android-debug-arm-try,vm-ffi-android-debug-arm64-try,vm-kernel-asan-linux-release-x64-try,vm-kernel-mac-debug-x64-try,vm-kernel-linux-debug-ia32-try,vm-kernel-linux-debug-x64-try,vm-kernel-nnbd-linux-debug-x64-try,vm-kernel-nnbd-linux-debug-ia32-try,vm-kernel-nnbd-mac-release-x64-try,vm-kernel-nnbd-win-debug-x64-try,vm-kernel-precomp-linux-debug-x64-try,vm-kernel-precomp-linux-debug-simarm_x64-try,vm-kernel-precomp-nnbd-linux-debug-x64-try,vm-kernel-precomp-win-release-x64-try,vm-kernel-reload-linux-debug-x64-try,vm-kernel-reload-rollback-linux-debug-x64-try,vm-kernel-win-debug-x64-try,vm-kernel-win-debug-ia32-try,vm-precomp-ffi-qemu-linux-release-arm-try,vm-kernel-precomp-obfuscate-linux-release-x64-try,vm-kernel-msan-linux-release-x64-try,vm-kernel-precomp-msan-linux-release-x64-try,vm-kernel-precomp-android-release-arm_x64-try,analyzer-analysis-server-linux-try
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/140290
Commit-Queue: Daco Harkes <dacoharkes@google.com>
Reviewed-by: Martin Kustermann <kustermann@google.com>
Ryan Macnak