Threads in the native or blocked states don't prevent safepoints, so they may run concurrently with a safepoint operation like GC. It is not safe for handles to be allocated while the GC is visiting them, so these threads must not allocate handles. Assert only threads in the VM or generated states, which prevent safepoints until they check in, may allocate handles. (Generated code does not allocate handles, but leaf runtime entries remain in the generated state.)
Bug: https://github.com/dart-lang/sdk/issues/34883
Change-Id: I1a211778f7ef96b53a2405f0ee9dde7871b122b6
Reviewed-on: https://dart-review.googlesource.com/c/81540
Commit-Queue: Ryan Macnak <rmacnak@google.com>
Reviewed-by: Siva Annamalai <asiva@google.com>
Reviewed-by: Vyacheslav Egorov <vegorov@google.com>
In VM, 'new List' is equivalent to either 'new _GrowableList' or
'new _List' depending on the number of arguments.
This change does the transformation early (on kernel AST), in order
to have specialized representation in TFA and in bytecode.
Change-Id: I46f0db8cc19efb3a53fdbe971ac26bdd2736fbda
Reviewed-on: https://dart-review.googlesource.com/76283
Reviewed-by: Vyacheslav Egorov <vegorov@google.com>
Commit-Queue: Alexander Markov <alexmarkov@google.com>
Ensure we still start NativeSymbolResolver even if the profiler is off, since it is also used name weak handler finalizer functions and to produce stack traces for crashes.
R=bkonyi@google.com
Review-Url: https://codereview.chromium.org/2992893002 .
Note there is no option to disable the profiler. The profiler on Mac, Linux, and Android is based on SIGPROF, and there is no way to wait for all the outstanding signals as required to be able to safely free the sample buffer.
Saves ~20MB on 64-bit VMs.
R=cbernaschina@google.com, zra@google.com
Review-Url: https://codereview.chromium.org/2989093002 .
I collected statistics for the sizes and capacities of growable arrays which are promoted to old-space or survive an old-space gc when running dart2js and Fasta. For these applications, the vast majority of arrays stay empty. More than half of the total object size of promoted backing arrays is backing for empty growable arrays.
Furthermore, since the overhead for an array is 3 words (header, type parameters and length), and object sizes are rounded up to an even number of words, we waste one word for all even-sized arrays.
This CL changes the growth strategy so that empty growable arrays are created with a shared, zero-sized array as backing, avoiding the allocation of a backing array if no elements are added. When the array needs to grow, it starts out at 3 and grows to double size plus one each time: 7, 15, 31, ...
A few places in the VM code need to handle these shared, zero-sized arrays specially. In particular, the Array::MakeArray function needs to allocate a new, empty array if its result is to be returned to Dart code.
Benchmarks suggest that the change improves memory usage by a few percent overall and does not significantly affect run time.
BUG=
R=erikcorry@google.com
Review-Url: https://codereview.chromium.org/2949803002 .
Regular constructors and unnamed factory constructors:
- Were A.A, now are new A
Named constructors (including factory constructors):
- Were A.A.name, now are new A.name
Previous attempt (a84b6b7f) didn't update vm/cc tests, which broke the
build, was reverted, and also revealed an important case:
Functions within constructors:
- Were A.A.x, now are new A.x
Doing this was cleanest by switching the loop out for recursion, which
provides a nice guarantee that any anonymous function within some other
function will always have the correct prefix, and the class name (which
may or may not be prepended at the end) is only applied once, and its
existence is only checked in one place.
R=asiva@google.com
Review-Url: https://codereview.chromium.org/2823253005 .
This allows us to report the "native heap memory" value in
the Observatory for VMs running on Fuchsia.
This change also renames the flag --enable-malloc-hooks to
--profiler-native-memory to better reflect what the flag
does. (In the future we may also want to rename
--profiler to --profiler-cpu.)
R=rmacnak@google.com
Review-Url: https://codereview.chromium.org/2829833003 .
Revert "Revert "Fixed issue in MallocHooks where a MallocHookScope was accidentally removed during a merge, causing a deadlock in the hooks. Also turned off stack trace collection in a test that was timing out as a result of the stack trace collection.""
This reverts commit 1a5b555aa9.
BUG=
R=zra@google.com
Review-Url: https://codereview.chromium.org/2711353003 .
CodeSourceMap now contains a small bytecode that describes changes to the token position and inlining stack as one moves through a function. This information is used by the disassembler to display inlining comments and by the profiler to properly blame ticks against source code.
Slightly shrinks the total size of code metadata; on dart2js compiling hello world, the final heap is
before 59.6MB total, 12.8MB List, 1.1MB CodeSourceMap
after 59.4MB total, 12.1MB List, 1.5MB CodeSourceMap
To be filtered and included in AOT code to expand inline frames in stack traces.
R=johnmccutchan@google.com
Review-Url: https://codereview.chromium.org/2670843006 .
The analyzer has a stricter patch parser than the VM. Patch files
cannot change signatures of patched members. Specifically, they cannot
change:
- the return type
- a parameter's name
- a parameter to an initializing formal
- an optional parameter's default value
BUG=
R=asiva@google.com
Review-Url: https://codereview.chromium.org/2612043002 .
Before, the VM's dart:typed_data was a complete replacement of the SDK's
dart:typed_data implementation instead of a patch. This is unlike all
the other SDK libraries. This difference requires special-casing for
dart:typed_data in tools that handle the SDK libraries (e.g., the
Analyzer's patching support, the GN build).
This change makes dart:typed_data back into a patch to the SDK's
implementation. It reintroduces a distinction between abstract
interface and concrete implementation classes, so there are more
classes.
BUG=
R=fschneider@google.com, vegorov@google.com
Committed: https://github.com/dart-lang/sdk/commit/a9b906d319c32525a0600c75c008c92753591d86
Review-Url: https://codereview.chromium.org/2571563005 .
Before, the VM's dart:typed_data was a complete replacement of the SDK's
dart:typed_data implementation instead of a patch. This is unlike all
the other SDK libraries. This difference requires special-casing for
dart:typed_data in tools that handle the SDK libraries (e.g., the
Analyzer's patching support, the GN build).
This change makes dart:typed_data back into a patch to the SDK's
implementation. It reintroduces a distinction between abstract
interface and concrete implementation classes, so there are more
classes.
BUG=
R=fschneider@google.com
Review-Url: https://codereview.chromium.org/2571563005 .
We have a racy read of the object header in the stack frame walker
called from the profiler (in this case the gc marker is running at
the same time).
It is appropriate to avoid this even though we think the access is
safe as the objects are in old space and never forwarded and only
the mark bits are manipulated while marking
BUG=
R=johnmccutchan@google.com
Review URL: https://codereview.chromium.org/2517473002 .
The call sequence is very similar to a classic IC call, except the guarded class and the target are loaded indirectly from the constant pool instead of as immediates. In the monomorphic case, we call directly to the expected target with a class check in the callee. In the unlinked, polymorphic and megamorphic cases, we call a stub; these case are now call-through instead of call-and-return.
Every code, except stubs involved in switchable calls, includes the class check sequence at the beginning. So we now distinguish between a checked and an unchecked entry point. Generated code except the switchable call continues to use the unchecked entry point.
PC offsets are calculated relative to the beginning of the instruction stream, rather than either entry point.
BUG=
R=fschneider@google.com
Review URL: https://codereview.chromium.org/2226893002 .
This is a cut of the work that Todd and I collaborated on in the reload branch.
In this CL, we've dropped the loader port hacks, in other words, on stack reloading in the standalone embedder does not work yet.
- [x] Support for hot reloading of isolate source code
- [x] Unit test harness and many tests
- [x] Service protocol and Observatory support
- [x] Product build does not include support for hot reloading.
R=rmacnak@google.com
Review URL: https://codereview.chromium.org/1965823002 .
Service:
- [x] Add "_profilerMode" to VM response. "VM" or "Dart" depending on --profile-vm flag.
- [x] Output a tree where VM tag are treated as executing "exit frame" functions when profiling Dart.
- [x] tick "exit frame" tag functions when profiling Dart.
Observatory:
- [x] don't include option for VM tags when profiling Dart.
- [x] default to no tags when profiling Dart.
- [x] indent the colored bar closer to the down arrow.
- [x] don't show percentages for user tag rows.
- [x] Update cpu profile table view to use virtual tree.
- [x] Update class allocation view to use virtual tree.
BUG=
R=rmacnak@google.com
Review URL: https://codereview.chromium.org/1851293005 .
Ryan Macnak