John Sully
|
465b8a9ea7
|
write fastlock_unlock in ASM... because its faster
Former-commit-id: bad73faf28f879d32c4064389b69c83e9474115a
|
2019-03-02 16:47:27 -05:00 |
|
John Sully
|
7a4227915a
|
Rewrite our spinlock in assembly, pretty big performance improvement
Former-commit-id: 40d7a701feefd36e9e3fdb6d516228c4a70fcf3d
|
2019-02-25 18:21:27 -05:00 |
|
John Sully
|
1761aabab4
|
Lock use after free
|
2019-02-22 21:00:14 -05:00 |
|
John Sully
|
b6735ac883
|
Remove unnecessary mfences
|
2019-02-22 15:49:22 -05:00 |
|
John Sully
|
d016b967c5
|
Implement try_lock
|
2019-02-22 01:23:31 -05:00 |
|
John Sully
|
942510f1a5
|
Change fastlock implementation from naive spinlock to a ticket lock for better fairness and lower peak latency
|
2019-02-21 00:16:47 -05:00 |
|
John Sully
|
02e7fe400c
|
Cleanup lock contention, and ensure clients are written to in an unsafe way when the global lock is released
|
2019-02-20 23:30:21 -05:00 |
|
John Sully
|
29c1105132
|
Multithreading works!
|
2019-02-20 01:20:26 -05:00 |
|
John Sully
|
2526d51d1a
|
Thread safety fixes
|
2019-02-18 22:25:35 -05:00 |
|
John Sully
|
2c220889c6
|
Fix recursive fastlock...
|
2019-02-16 14:25:14 -05:00 |
|
John Sully
|
f5caec488d
|
Make the lock recursive, this is because processEventsWhileBlocked may cause us to lock multiple times
|
2019-02-15 14:11:05 -05:00 |
|
John Sully
|
988ea40101
|
Fastlock fixes
|
2019-02-10 22:00:19 -05:00 |
|
John Sully
|
c4dbc557eb
|
move ae to C++
|
2019-02-10 20:24:11 -05:00 |
|