[prev in list] [next in list] [prev in thread] [next in thread] 

List:       dragonfly-kernel
Subject:    Re: Soft token staleness question
From:       Matthew Dillon <dillon () apollo ! backplane ! com>
Date:       2010-05-27 19:40:43
Message-ID: 201005271940.o4RJehqH062432 () apollo ! backplane ! com
[Download RAW message or body]


:Hi,
:
:Why is lwkt_token_is_stale if 0-ed out?
:
:If I understand the motivation behind tokens, higher level code can take it
:an call lower level code which can block. But without the ability to check
:staleness, how you know if lower level code blocked and allowed your
:assumptions made under the token earlier to be invalidated?
:
:Thanks,
:-- vs

    It's hard to say how useful a staleness indicator is because using
    such an indicator means having more complex, harder-to-debug code,
    verses simply coding things to assume the lock was temporarily
    broken whenever something might block.

    I think there are a few cases where it would be useful, particularly
    in the VM subsystem when it eventually gets converted over to MPSAFE
    code.  It would not be useful for most of the existing use cases.

					-Matt
					Matthew Dillon 
					<dillon@backplane.com>
[prev in list] [next in list] [prev in thread] [next in thread]