In the source:
https://en.wikipedia.org/wiki/Spinlock
I have question about the sentence:
On later implementations of the x86 architecture, spin_unlock can safely use an unlocked MOV instead of the slower locker XCHG. This is due to subtle memory ordering rules which support this, even though MOV is not a full memory barrier.
What is sutble memory ordering rules?
Why MOV is not a full memory barrier and what is instruction(s) that has a full memory barrier?
And what the meaning of the sentence? And how it can express in other simple words?
Thanks for answering...
https://en.wikipedia.org/wiki/Spinlock
I have question about the sentence:
On later implementations of the x86 architecture, spin_unlock can safely use an unlocked MOV instead of the slower locker XCHG. This is due to subtle memory ordering rules which support this, even though MOV is not a full memory barrier.
What is sutble memory ordering rules?
Why MOV is not a full memory barrier and what is instruction(s) that has a full memory barrier?
And what the meaning of the sentence? And how it can express in other simple words?
Thanks for answering...