Ignore the result of pthread_kill in ubf_wakeup_thread
After an upgrade to Ruby 3.3.0, I experienced reproducible production crashes
of the form:
[BUG] pthread_kill: No such process (ESRCH)
This is the only pthread_kill call in Ruby. The result of pthread_kill was
previously ignored in Ruby 3.2 and below. Checking the result was added in
be1bbd5b7d (MaNy).
I have not yet been able to create a minimal self-contained example,
but it should be safe to remove the checks.
`hrrel / RB_HRTIME_PER_MSEC` floor the timeout value and it can
return wrong value by `Mutex#sleep` (return Integer even if
it should return nil (timeout'ed)).
This patch ceil the value and the issue was solved.
* Allows macOS users to use M:N threads (and technically FreeBSD, though it has not been verified on FreeBSD)
* Include sys/event.h header check for macros, and include sys/event.h when present
* Rename epoll_fd to more generic kq_fd (Kernel event Queue) for use by both epoll and kqueue
* MAP_STACK is not available on macOS so conditionall apply it to mmap flags
* Set fd to close on exec
* Log debug messages specific to kqueue and epoll on creation
* close_invalidate raises an error for the kqueue fd on child process fork. It's unclear rn if that's a bug, or if it's kqueue specific behavior
Use kq with rb_thread_wait_for_single_fd
* Only platforms with `USE_POLL` (linux) had changes applied to take advantage of kernel event queues. It needed to be applied to the `select` so that kqueue could be properly applied
* Clean up kqueue specific code and make sure only flags that were actually set are removed (or an error is raised)
* Also handle kevent specific errnos, since most don't apply from epoll to kqueue
* Use the more platform standard close-on-exec approach of `fcntl` and `FD_CLOEXEC`. The io-event gem uses `ioctl`, but fcntl seems to be the recommended choice. It is also what Go, Bun, and Libuv use
* We're making changes in this file anyways - may as well fix a couple spelling mistakes while here
Make sure FD_CLOEXEC carries over in dup
* Otherwise the kqueue descriptor should have FD_CLOEXEC, but doesn't and fails in assert_close_on_exec
`sched->lock_owner` can be non-NULL at fork because the timer thread
can acquire the lock while forking. `lock_owner` information is for
debugging, so we only need to clear it at fork.
I hope this patch fixes the following assertion failure:
```
thread_pthread.c:354:thread_sched_lock_:sched->lock_owner == NULL
```
[Bug #20019]
This fixes GVL instrumentation in three locations it was missing:
- Suspending when blocking on a Ractor
- Suspending when doing a coroutine transfer from an M:N thread
- Resuming after an M:N thread starts
Co-authored-by: Matthew Draper <matthew@trebex.net>
This entirely changes how it is tested. Rather than to use counters
we now record the timeline of events with associated threads which
makes it much easier to assert that certains events are only preceded
by a specific event, and makes it much easier to debug unexpected
timelines.
Co-Authored-By: Étienne Barrié <etienne.barrie@gmail.com>
Co-Authored-By: JP Camara <jp@jpcamara.com>
Co-Authored-By: John Hawthorn <john@hawthorn.email>
Context: https://github.com/ivoanjo/gvl-tracing/pull/4
Some hooks may want to collect data on a per thread basis.
Right now the only way to identify the concerned thread is to
use `rb_nativethread_self()` or similar, but even then because
of the thread cache or MaNy, two distinct Ruby threads may report
the same native thread id.
By passing `thread->self`, hooks can use it as a key to store
the metadata.
NB: Most hooks are executed outside the GVL, so such data collection
need to use a thread-safe data-structure, and shouldn't use the
reference in other ways from inside the hook.
They must also either pin that value or handle compaction.
If `RUBY_MN_THREADS=1` is given, this patch shows `+MN` in
`RUBY_DESCRIPTION` like:
```
$ RUBY_MN_THREADS=1 ./miniruby --yjit -v
ruby 3.3.0dev (2023-10-17T04:10:14Z master 908f8fffa2) +YJIT +MN [x86_64-linux]
```
Before this patch, a warning is displayed if `$VERBOSE` is given.
However it can make troubles with tests (with `$VERBOSE`), do not
show any warning with a MN threads configuration.
* on `__EMSCRIPTEN__` provides epoll* declarations, but no implementations.
* on `NON_SCALAR_THREAD_ID`, now we can not debug issues on x390s/Ubuntu so skip it.
x390s/RHEL works fine, so I think we can remove second limitation but
I could not login to it so it seems hard to debug now.
With M:N thread scheduler, the native thread (NT) related resources
should be freed when the NT is no longer needed. So the calling
`native_thread_destroy()` at the end of `is will be freed when
`thread_cleanup_func()` (at the end of Ruby thread) is not correct
timing. Call it when the corresponding Ruby thread is collected.
This patch introduce M:N thread scheduler for Ractor system.
In general, M:N thread scheduler employs N native threads (OS threads)
to manage M user-level threads (Ruby threads in this case).
On the Ruby interpreter, 1 native thread is provided for 1 Ractor
and all Ruby threads are managed by the native thread.
From Ruby 1.9, the interpreter uses 1:1 thread scheduler which means
1 Ruby thread has 1 native thread. M:N scheduler change this strategy.
Because of compatibility issue (and stableness issue of the implementation)
main Ractor doesn't use M:N scheduler on default. On the other words,
threads on the main Ractor will be managed with 1:1 thread scheduler.
There are additional settings by environment variables:
`RUBY_MN_THREADS=1` enables M:N thread scheduler on the main ractor.
Note that non-main ractors use the M:N scheduler without this
configuration. With this configuration, single ractor applications
run threads on M:1 thread scheduler (green threads, user-level threads).
`RUBY_MAX_CPU=n` specifies maximum number of native threads for
M:N scheduler (default: 8).
This patch will be reverted soon if non-easy issues are found.
[Bug #19842]
This patch fixes a potential busy-loop in the thread scheduler. If there
are two threads, the main thread (where Ruby signal handlers must run)
and a sleeping thread, it is possible for the following sequence of
events to occur:
* The sleeping thread is in native_sleep -> sigwait_sleep A signal
* arives, kicking this thread out of rb_sigwait_sleep The sleeping
* thread calls THREAD_BLOCKING_END and eventually
thread_sched_to_running_common
* the sleeping thread writes into the sigwait_fd pipe by calling
rb_thread_wakeup_timer_thread
* the sleeping thread re-loops around in native_sleep() because
the desired sleep time has not actually yet expired
* that calls rb_sigwait_sleep again the ppoll() in rb_sigwait_sleep
* immediately returns because
of the byte written into the sigwait_fd by
rb_thread_wakeup_timer_thread
* that wakes the thread up again and kicks the whole cycle off again.
Such a loop can only be broken by the main thread waking up and handling
the signal, such that ubf_threads_empty() below becomes true again;
however this loop can actually keep things so busy (and cause so much
contention on the main thread's interrupt_lock) that the main thread
doesn't deal with the signal for many seconds. This seems particuarly
likely on FreeBSD 13.
(the cycle can also be broken by the sleeping thread finally elapsing
its desired sleep time).
The fix for _this_ loop is to only wakeup the timer thrad in
thread_sched_to_running_common if the current thread is not itself the
sigwait thread.
An almost identical loop also happens in the same circumstances because
the call to check_signals_nogvl (through sigwait_timeout) in
rb_sigwait_sleep returns true if there is any pending signal for the
main thread to handle. That then causes rb_sigwait_sleep to skip over
sleeping entirely.
This is unnescessary and counterproductive, I believe; if the main
thread needs to be woken up that is done inline in check_signals_nogvl
anyway.
See https://bugs.ruby-lang.org/issues/19680
Show native thread's serial on `RUBY_DEBUG_LOG`.
`nt->serial` is also stored into `ruby_nt_serial` if the compiler
supports `RB_THREAD_LOCAL_SPECIFIER`.
[Feature #19443]
Until recently most libc would cache `getpid()` so this was a cheap check to make.
However as of glibc version 2.25 the PID cache is removed and calls to getpid() always
invoke the actual system call which significantly degrades the performance of existing applications.
The reason glibc removed the cache is that some libraries were bypassing fork(2)
by issuing system calls themselves, causing stale cache issues.
That isn't a concern for Ruby as bypassing MRI's primitive for forking would
render the VM unusable, so we can safely cache the PID.
* Revert "Remove special handling of `SIGCHLD`. (#7482)"
This reverts commit 44a0711eab.
* Revert "Remove prototypes for functions that are no longer used. (#7497)"
This reverts commit 4dce12bead.
* Revert "Remove SIGCHLD `waidpid`. (#7476)"
This reverts commit 1658e7d966.
* Fix change to rjit variable name.
```
1) Failure:
TestThreadInstrumentation#test_thread_instrumentation [/tmp/ruby/src/trunk-repeat20-asserts/test/-ext-/thread/test_instrumentation_api.rb:33]:
Call counters[4]: [3, 4, 4, 4, 0].
Expected 0 to be > 0.
```
We fire the EXIT hook after the call to `thread_sched_to_dead` which
mean another thread might be running before the `EXIT` hook have been
executed.
[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`.
[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.
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.
Takashi Kokubun
Daisuke Fujimura (fd0)
Koichi Sasada
JP Camara
John Hawthorn
Jean Boussier
Sergey Fedorov
Kazuhiro NISHIYAMA
Nobuyoshi Nakada
KJ Tsanaktsidis
Matt Valentine-House
Samuel Williams
Jean Boussier