Re: Call for Developers! Userland threading

[Matthew Dillon <dillon@xxxxxxxxxxxxxxxxxxxx>]

:So it sounds like theres going to have to be some user backing code to 
:register a sort of handler.  Preemption is needed to allow signals to catch 
:at all if you were caught in a tight loop (unless kernel takes over like 
:KILL).  Should it be as complex as having the kernel creating a userland 
:trapframe when its doing an "upcall" so the handler can decide what to do, 
:and if it's not important, return to the interrupted thread?  Otherwise the 
:kernel could probably have different paths for sending critical vs. 
:non-critical messages to a process? It seems the first is going to have to be 
:the approach for per process signals also.  I dont fully know the POSIX spec, 
:but from my limited knowledge, it seems this would work.
:- -- 
:- -Craig

    It's actually less complex for the kernel to just do an unconditional
    upcall and let the userland program figure out what to do with it.

						-Matt