lwkt_token progress

Matthew Dillon dillon at apollo.backplane.com
Sun Feb 22 00:00:35 PST 2004

:Matthew Dillon <dillon at xxxxxxxxxxxxxxxxxxxx> wrote in
:news:200402212030.i1LKU9px020972 at xxxxxxxxxxxxxxxxxxxx: 
:>                          ----
:>                     Serializing Tokens
:Would this tokens writeup be useful to add to one of the Goals pages on
:the website? 

    Yes, it would be an excellent addition..  actually, if you could,
    rip out the section in Goals/threads.cgi near the end that explains
    tokens and replace it with the above documentation.

    The new token methodology is nicely solidified now.  I have successfully
    booted a UP kernel with the initial changeset.  Tomorrow I will get
    SMP working, then I have to fix all the vnode scanning interlock code
    using the new token abstraction.  (David Rhodus keeps finding panic
    bugs in the inode hash and vnode reclaimation code and it's just 
    impossible to fix using the old token model).

    I should have a patch set ready for testing tomorrow sometime.  The
    SMP interactions going on behind the scenes are a little tricky.

					Matthew Dillon 
					<dillon at xxxxxxxxxxxxx>

More information about the Kernel mailing list