1
0
mirror of https://github.com/RPCS3/rpcs3.git synced 2024-11-25 20:22:30 +01:00
Commit Graph

21 Commits

Author SHA1 Message Date
Ivan Chikish
d34287b2cc Linux: use futex_waitv syscall for atomic waiting
In order to make this possible, some unnecessary features were removed.
2023-08-02 21:46:06 +03:00
Eladash
2ec039365f shared_mutex: Optimize busy-waiting by detecting waiters and try to steal the notifying bit
Add an unused has_waiters() method.
2022-08-07 20:23:54 +03:00
Nekotekina
4cfa9b11f3 Move busy_wait() to asm.hpp 2020-12-18 12:23:53 +03:00
Nekotekina
e055d16b2c Replace verify() with ensure() with auto src location.
Expression ensure(x) returns x.
Using comma operator removed.
2020-12-09 15:43:38 +03:00
Nekotekina
cc07e5306e Improve shared_mutex
Wait/notify only on c_sig, should reduce spurious wakeup noise.
2020-11-06 00:10:16 +03:00
Nekotekina
050c3e1d6b Rewrite cpu_thread::suspend_all
Now it's a function of higher order.
Make only one thread do the hard work of thread pausing.
2020-10-10 13:58:48 +03:00
Nekotekina
8886414013 Use atomic wait in shared_mutex and semaphore 2019-07-29 03:04:55 +03:00
Nekotekina
090c71aa7c TSX: Improve cpu_thread::suspend_all implementation
Implement low_lock and vip_lock (for shared_mutex).
Try to simplify suspend_all implementation with updated shared_mutex.
2019-07-24 17:25:52 +03:00
Nekotekina
96cabeadff Rewrite condition variables
Implement helper functions balanced_wait_until and balanced_awaken
They include new path for Windows 8.1+ (WaitOnAddress)

shared_mutex, cond_variable, cond_one, cond_x16 modified to use it
Added helper function utils::popcnt16
Replace most semaphore<> with shared_mutex
2018-11-29 01:30:05 +03:00
Nekotekina
6104685ad6 Implement cond_one sync primitive
Change futex() args to use unsigned int
2018-11-05 13:14:11 +03:00
Nekotekina
7a024f3355 Implement shared_mutex::lock_unlock
Minor fix for shared_mutex::try_lock - don't optimize for pessimistic case
2018-10-02 02:22:26 +03:00
Nekotekina
9e5b633779 shared_mutex cleanup
Rewrite to use unsigned arithmetic, shrink to u32
Ensure zero default unlocked state (will need it later)
Inline all public methods, rewrite lock_upgrade()
Remove try_lock_degrade(), lock_degrade() methods
Implement lock_downgrade() (now trivial)
Remove is_reading(), added is_free()
Added reader_lock::try_upgrade()
2018-09-27 12:16:43 +03:00
Nekotekina
445e5def42 Remove safe_reader_lock, safe_writer_lock
It worked as a simple recursive locker for shared_mutex
But its design is flawed and thus can't be fully implemented
2018-09-15 17:09:56 +03:00
Nekotekina
fb5cdf9769 atomic_t<>: extend fetch_op to support cancellation
Use std::invoke inside atomic_op/fetch_op
Remove op_fetch because it's easily replaced
Add fetch_dec_sat algorithm (conditional decrement)
2018-09-08 00:32:04 +03:00
Nekotekina
7f542a5c99 Fix safe_writer_lock (typo) 2018-03-11 19:33:01 +03:00
Nekotekina
beab5291bb mutex.cpp: minor update
New concept (incomplete)
2017-10-05 19:51:37 +03:00
Ryan Gonzalez
e447f071ce Misc compile fixes 2017-04-18 22:37:17 +03:00
Ivan
1ae334e500 Fix freezes on Linux (#2632) 2017-04-04 23:27:16 +03:00
Nekotekina
f35babad98 shared_mutex fix (linux) 2017-02-24 18:48:53 +03:00
Nekotekina
b1aa87b515 Busy waiting added 2017-02-16 00:13:55 +03:00
Nekotekina
1c14d872a8 New shared_mutex
Experimental sync utils
New semaphore<>
New cond_variable
New owned_mutex
2017-01-25 00:38:07 +03:00