| commit | f56345bcd10e20c4d0735486368cbc157b4e05ca | [log] [tgz] |
|---|---|---|
| author | Daniele Di Proietto <ddiproietto@google.com> | Wed Nov 02 10:16:11 2022 +0000 |
| committer | Lalit Maganti <lalitm@google.com> | Wed Mar 29 15:03:56 2023 +0100 |
| tree | 71384e2f54d7f3e53cbb6bb4736c60cbbf923675 | |
| parent | ad418a3f70f33f6b50cd008b565360ef95446063 [diff] |
heapprofd: Fix crash when tracing service disconnects
When the tracing service (traced) disconnects (because it crashes or
other reasons), if heapprofd has an active client, it will crash,
because:
* HeapprofdProducer::Restart() is executed.
* HeapprofdProducer destructor destroys its unwinding_workers_
* UnwindingWorker destroys its members. In order:
* thread_task_runner_ is destroyed, so that the thread is stopped.
* (If there's one connected client) client_data_ is destroyed.
* ClientData includes a UnixSocket, which calls PostTask on
thread_task_runner_.
* thread_task_runner_ was destroyed earlier.
This patch makes fixes the problem by destroying client_data_ (on
thread_task_runner_) before destroying the thread_task_runner_ itself.
(this is a cherry-pick of just the test part of ag/20347035 since the
actual fix landed through a merge from tm-dev)
Bug: 233338564
Bug: 155759761
Change-Id: Id1874abbff8f7a7b702542a943a8de3f3a52277a
Merged-In: Id1874abbff8f7a7b702542a943a8de3f3a52277a
Perfetto is a production-grade open-source stack for performance instrumentation and trace analysis. It offers services and libraries and for recording system-level and app-level traces, native + java heap profiling, a library for analyzing traces using SQL and a web-based UI to visualize and explore multi-GB traces.
See https://perfetto.dev/docs or the /docs/ directory for documentation.