[Sis-SCPS-INTEREST] New SCPS RI Release on Open Channel Software...

Otto Solares solca at guug.org
Thu Mar 16 16:54:42 EST 2006


On Wed, Mar 15, 2006 at 03:07:52PM -0500, Feighery, Patrick D. wrote:
> When I used the term 'poll' I used that term loosely.  Before this new
> release, the SCPS RI did use select(). However it returned immediately
> and did not have any timeout specified.  This code was originally
> written over 10 years ago.  Now the code has a select that will wait
> for 1 msec before it times out.  After doing this I realized that the
> rate control token bucket was affected, so I made that change as well.
> After this I realized that there were a couple of other tricks I needed
> to do.

Hmmm, the problem I see is that the process never block, I am not
familiar with the source, but if you think we could make the process
block by using blocking I/O I'll take a look and probably generate
a patch.  Is there any reason why SPCP_RI doesn't use blocking I/O?
My biggest gripe will be the rate control using blocking I/O.

> As for Eventpoll (linux) and kqueue/kevent (BSDs)  I really am not
> familiar with either.  If you or someone else wanted to create this
> functionality and have it conditionally compiled, I would be happy to
> look at it.

I am familiar with epoll in linux, but basically is just a fast and
efficient select replacement.  It use will not improve the high CPU
utilization if the process never block.

-otto



More information about the Sis-SCPS-INTEREST mailing list