[ovs-dev] [ext-272 0/4] Implement EXT-272 (role reply generation_ids)

Ben Pfaff blp at nicira.com
Wed Feb 13 19:18:36 UTC 2013


On Wed, Feb 13, 2013 at 07:47:38AM +0000, Rajahalme, Jarno (NSN - FI/Espoo) wrote:
> On Feb 12, 2013, at 18:52 , ext Ben Pfaff wrote:
> 
> > On Tue, Feb 12, 2013 at 03:31:16PM +0000, Rajahalme, Jarno (NSN - FI/Espoo) wrote:
> >> The patches look good, but as I haven't seen the EXT-272, I don't know
> >> if there is a possible conflict with the all-ones value returned when
> >> the generation id is undefined. Have you proposed to exclude the
> >> all-ones value from the set of valid generation ID values? Or does it
> >> matter?
> > 
> > There's nothing special in EXT-272 about excluding all-ones.  I think
> > the logic is that UINT64_MAX is unlikely to actually come up if the
> > generation number starts from zero.
> 
> The logic with not having the generation ID initially defined is to
> allow for the controllers to pick any arbitrary number
> initially. This number may be persistent across individual
> controller restarts. After that the generation ID must increase, but
> each step can be as large as 2^63-1 (less than half of the ID
> space). I agree that starting from zero and increasing by one every
> now and then would take a long time to reach UINT64_MAX.

I guess that anyone who wants to use this feature of the replies will
have to avoid using UINT64_MAX as a sequence number.

> > You ought to be able to get access to the Open Networking Foundation bug
> > tracker because Nokia Siemens Networks is an ONF member.  See
> > https://www.opennetworking.org/membership/member-resources and click on
> > "Create your ONF Jira account."
> 
> Haven't bothered so far, as we have other people who get paid to do
> that :-) I'll see...

Few good developers participate.  On the other hand, it can waste
time.



More information about the dev mailing list