dart-sdk/pkg/native_stack_traces
Tess Strickland 010523ee4a [vm] Fix handling of WeakSerializationReferences in v8 profile.
Instead of trying to lazily create artificial nodes when needed for
WeakSerializationReference (WSR) targets in all cases, create them
eagerly for targets in reachable WSRs, since those are guaranteed
to be needed.

Those for unreachable WSRs are still lazily created as needed, since the
WSR won't even be accessed by the clustered snapshot writer unless the
unreachable WSR is part of another unreachable object that has an
artificial node created.

This rework avoids some issues seen on upcoming CLs where the artificial
nodes for WSR targets weren't getting correctly generated.

-----

Also extend the v8 snapshot profile writer tests to check the sizes of
the text and data sections in ELF snapshots. That means the v8 snapshot
profile writer tests check up to three different measures, from most
precise to least precise, depending on the output mode:

* If writing an ELF snapshot directly: the sum of the sizes attributed
  to the text and data section symbols are checked to be exactly the
  same as the sum of the sizes of the objects in the profile.

* If writing an ELF snapshot either directly or via assembly: the sum
  of the sizes of the text and data sections are checked to be the same
  as the sum of the sizes of the objects in the profile. If using an
  assembler that merges text and data sections, then account for the
  padding between the sections using an approximate check.

* For all: Check that the size of the snapshot as a whole is
  approximately equal to the sum of the sizes of the objects in the
  profile, accounting both for possible padding and for the global
  header information and non-data/non-text sections like the dynamic
  symbol section.

TEST=vm/data{,_2}/v8_snapshot_profile_writer_test

Cq-Include-Trybots: luci.dart.try:vm-kernel-precomp-linux-debug-x64-try,vm-kernel-precomp-linux-debug-x64c-try,vm-kernel-precomp-nnbd-linux-debug-x64-try,vm-kernel-precomp-mac-release-simarm64-try
Change-Id: I66e0e7fdb5bb98045621bf516f150a4723e08147
Reviewed-on: https://dart-review.googlesource.com/c/sdk/+/198942
Commit-Queue: Tess Strickland <sstrickl@google.com>
Reviewed-by: Clement Skau <cskau@google.com>
Reviewed-by: Daco Harkes <dacoharkes@google.com>
2021-05-11 11:21:44 +00:00
..
bin Roll args and dart2js_info into the SDK. 2020-12-04 22:58:22 +00:00
lib [vm] Fix handling of WeakSerializationReferences in v8 profile. 2021-05-11 11:21:44 +00:00
.gitignore
analysis_options.yaml pkg:native_stack_trace: fix pedantic reference 2020-09-22 15:34:00 +00:00
AUTHORS
CHANGELOG.md [vm] Fix handling of WeakSerializationReferences in v8 profile. 2021-05-11 11:21:44 +00:00
LICENSE Update LICENSE 2021-04-07 10:28:38 +00:00
pubspec.yaml [vm] Fix handling of WeakSerializationReferences in v8 profile. 2021-05-11 11:21:44 +00:00
README.md

native_stack_traces

This package provides libraries and a utility for decoding non-symbolic stack traces generated by an AOT-compiled Dart application.

Converting stack traces

In some modes of AOT compilation, information on mapping execution points to source locations is no longer stored in the Dart image. Instead, this information is translated to separately stored debugging information. This debugging information can then be stripped from the application before shipping.

However, there is a drawback. Stack traces generated by such an application no longer includes file, function, and line number information (i.e., symbolic stack traces). Instead, stack trace frames simply include program counter information. Thus, to find the source information for these frames, we must use the debugging information. This means either keeping the original unstripped application, or saving the debugging information into a separate file.

Given this debugging information, the libraries in this package can turn non-symbolic stack traces back into symbolic stack traces. In addition, this package includes a command line tool decode whose output is the same as its input except that non-symbolic stack traces are translated.

Using decode

Take the following Dart code, which we put in throws.dart. The inlining pragmas are here just to ensure that bar is inlined into foo and that foo is not inlined into bar, to illustrate how inlined code is handled in the translated output.

@pragma('vm:prefer-inline')
bar() => throw null;

@pragma('vm:never-inline')
foo() => bar();

main() => foo();

Now we run the following commands:

# Make sure that we have the native_stack_traces package.
$ pub get native_stack_traces
$ pub global activate native_stack_traces

# We compile the example program, removing the source location information
# from the snapshot and saving the debugging information into throws.debug.
$ dart2native -k aot -S throws.debug -o throws.aotsnapshot throws.dart

# Run the program, saving the error output to throws.err.
$ dartaotruntime throws.aotsnapshot 2>throws.err

# Using the saved debugging information, we can translate the stack trace
# contained in throws.err to its symbolic form.
$ pub global run native_stack_traces:decode translate -d throws.debug -i throws.err

# We can also just pipe the output of running the program directly into
# the utility.
$ dartaotruntime throws.aotsnapshot |& \
    pub global run native_stack_traces:decode translate -d throws.debug

Features and bugs

Please file feature requests and bugs at the issue tracker.