| Age | Commit message (Collapse) | Author |
|---|
|
|
|
I noticed this was missing while adding spec comments a bit ago.
It's small and easy enough to implement, might as well make us
more POSIX compliant.
|
|
|
|
All the way back in commit 1670ee5aba09, the default clock for
condition variables was set to `CLOCK_MONOTONIC`, because there was no
other clock available.
However, if a condition variable is initialized without any additional
attributes by an application, they sometimes assume that the absolute
time that is passed to e.g. `pthread_cond_timedwait` is actually based
on a realtime clock, as can be seen here in SDL2:
https://github.com/SerenityPorts/SDL/blob/6f419bdf5f56be236c070a9d364e0d238b868565/src/thread/pthread/SDL_syscond.c#L99
Additionally, the glibc implementation defaults to a realtime clock:
https://github.com/bminor/glibc/blob/aac54dcd378209bbdddbcec749561b1d8f167d11/nptl/pthread_cond_init.c#L42
...so we probably should do so as well :^)
|
|
We only ever use private futexes, so it doesn't make sense to carry
around all the complexity required for global (cross-process) futexes.
|
|
Performance go brrrrr
|
|
|
|
It would be enough to use relaxed ordering here if it weren't for
the mutex, which we also need to store and retrieve. To ensure the
pthread_cond_broadcast() call sees the store, use release and acquire
as appropriate. Thankfully, both of these are on the slow paths.
|
|
This implementation features a fast path for pthread_cond_signal() and
pthread_cond_broadcast() for the case there's no thread waiting, and
does not exhibit the "thundering herd" issue in
pthread_cond_broadcast().
Fixes https://github.com/SerenityOS/serenity/issues/8432
|
