On Sun, May 11, 2008 at 12:35:23AM -0700, Matt Thomas wrote: > > On May 10, 2008, at 9:40 PM, Bill Stouder-Studenmund wrote: > > >As part of reviving SA, I'm re-adding all of the kernel > >infrastructure we > >ripped out. > > > >In doing this, I looked at re-adding the preempt(int more) code we > >had. > >However, I have serious questions about it. Like why do it? [snip] > You went from LSONPROC to LSRUN and that's shouldn't cause an BLOCKED > upcall. > I think a BLOCKED upcall should be sent iff the state changed to > LSSLEEP. > I also think if a change to LSSTOP or LSSUSPENDED happens an upcall > should > not be sent since an external party want the lwp/proc to stop > execution and > switching to a new one to continue execution just seems wrong. Ok. Sounds good. So I wasn't totally misunderstanding it. :-) Take care, Bill
Attachment:
pgp7gzWOpiwA0.pgp
Description: PGP signature