Re: Threading on Solaris

[Alexandre Julliard]

Gregg Mattinson  writes:
> Why does the SYSLEVEL structure have a CRITICAL_SECTION inside it?  Shouldn't a 
> mutex be used instead?

No, a mutex is a lot slower, and there is no way for a SYSLEVEL to
report that it was abandoned. If a thread gets killed while holding a
critical section or syslevel, all bets are off anyway, since you don't
know anything about the state of the resource that was protected by
the critical section. There is simply no way to recover from that.