summaryrefslogtreecommitdiff
path: root/test/-ext-/thread
AgeCommit message (Collapse)Author
2022-07-13GVL Instrumentation: remove the EXITED count assertionJean Boussier
It's very flaky for some unknown reason. Something we have an extra EXITED event. I suspect some other test is causing this. Notes: Merged: https://github.com/ruby/ruby/pull/6133
2022-07-13thread/test_instrumentation_api: cleanup all existing threads in setupJean Boussier
We saw the following failure: ``` TestThreadInstrumentation#test_thread_instrumentation [/tmp/ruby/v3/src/trunk-random3/test/-ext-/thread/test_instrumentation_api.rb:25]: Expected 0..3 to include 4. ``` Which shouldn't happen unless somehow there was a leaked thread. Notes: Merged: https://github.com/ruby/ruby/pull/6128
2022-07-12Refactor tests for ThreadInstrumentation countersNobuyoshi Nakada
* Extracted some assertions. * Assert counter values should be positive. Notes: Merged: https://github.com/ruby/ruby/pull/6111
2022-07-12Use `IO.popen` to fork and exit the child process without cleanupNobuyoshi Nakada
Notes: Merged: https://github.com/ruby/ruby/pull/6111
2022-07-08Relax assertion condition for thread local countersNobuyoshi Nakada
Recently `TestThreadInstrumentation#test_join_counters` often fails as ``` <[1, 1, 1]> expected but was <[2, 2, 2]>. ``` Probably it seems that the thread is suspended more than once. There may be no guarantee that the subject thread never be suspended more than once.
2022-07-07thread_pthread.c: call SUSPENDED event when entering native_sleepJean Boussier
[Bug #18900] Thread#join and a few other codepaths are using native sleep as a way to suspend the current thread. So we should call the relevant hook when this happen, otherwise some thread may transition directly from `RESUMED` to `READY`. Notes: Merged: https://github.com/ruby/ruby/pull/6101
2022-06-20test_instrumentation_api.rb: Allow one less exitJean Boussier
I suspect that sometimes on CI the last thread is prempted before eaching the exit hook causing the test to flake. I can't find a good way to force it to run.
2022-06-19Scale the time to wait native threads to run hookNobuyoshi Nakada
2022-06-17Debug TestThreadInstrumentationJean Boussier
It previously failed with: ``` 1) Failure: TestThreadInstrumentation#test_thread_instrumentation_fork_safe [/home/runner/work/ruby/ruby/src/test/-ext-/thread/test_instrumentation_api.rb:50]: <5> expected but was <4>. ``` Suggesting one `EXIT` event wasn't fired or processed. Adding an assetion on `Thead#status` may help figure out what is wrong. Notes: Merged: https://github.com/ruby/ruby/pull/6032
2022-06-17GVL Instrumentation API: add STARTED and EXITED eventsJean Boussier
[Feature #18339] After experimenting with the initial version of the API I figured there is a need for an exit event to cleanup instrumentation data. e.g. if you record data in a {thread_id -> data} table, you need to free associated data when a thread goes away. Notes: Merged: https://github.com/ruby/ruby/pull/6029
2022-06-07Refactor TestThreadInstrumentation to investigate CI flakinessJean Boussier
`test_thread_instrumentation_fork_safe` has been failing occasionaly on Ubuntu and Arch. At this stage we're not sure why, all we know is that the child exit with status 1. I suspect that something entirely unrelated cause the forked children to fail on exit, so by using `exit!(0)` and doing assertions in the parent I hope to be resilient to that. Notes: Merged: https://github.com/ruby/ruby/pull/5983
2022-06-06Use `sleep 0.5` for tests of GVL instrumentation APIYusuke Endoh
The tests fail randomly on some platforms. http://rubyci.s3.amazonaws.com/ubuntu/ruby-master/log/20220605T213004Z.fail.html.gz http://rubyci.s3.amazonaws.com/arch/ruby-master/log/20220605T210003Z.fail.html.gz ``` [15737/21701] TestThreadInstrumentation#test_thread_instrumentation_fork_safe/home/chkbuild/chkbuild/tmp/build/20220605T213004Z/ruby/tool/lib/test/unit/assertions.rb:109:in `assert': Expected 0 to be nonzero?. (Test::Unit::AssertionFailedError) ``` The failures seem to depend on context switches. I suspect `sleep 0.05` is too short, so this change tries to extend the wait time.
2022-06-03[Feature #18339] GVL Instrumentation APIJean Boussier
Ref: https://bugs.ruby-lang.org/issues/18339 Design: - This tries to minimize the overhead when no hook is registered. It should only incur an extra unsynchronized boolean check. - The hook list is protected with a read-write lock as to cause contention when some hooks are registered. - The hooks MUST be thread safe, and MUST NOT call into Ruby as they are executed outside the GVL. - It's simply a noop on Windows. API: ``` rb_internal_thread_event_hook_t * rb_internal_thread_add_event_hook(rb_internal_thread_event_callback callback, rb_event_flag_t internal_event, void *user_data); bool rb_internal_thread_remove_event_hook(rb_internal_thread_event_hook_t * hook); ``` You can subscribe to 3 events: - READY: called right before attempting to acquire the GVL - RESUMED: called right after successfully acquiring the GVL - SUSPENDED: called right after releasing the GVL. The hooks MUST be threadsafe, as they are executed outside of the GVL, they also MUST NOT call any Ruby API. Notes: Merged: https://github.com/ruby/ruby/pull/5500