[ovs-dev] [PATCH v3 4/8] Support decoding of NTR selection method

Simon Horman simon.horman at netronome.com
Wed Mar 18 23:33:28 UTC 2015


On Tue, Mar 17, 2015 at 03:10:18PM -0700, Ben Pfaff wrote:
> On Mon, Mar 09, 2015 at 10:10:59AM +0900, Simon Horman wrote:
> > This is in preparation for supporting group mod and desc reply
> > messages with an NTR selection method group experimenter property.
> > 
> > Currently decoding always fails as it only allows properties for known
> > selection methods and no selection methods are known yet. A subsequent
> > patch will propose a hash selection method.
> > 
> > NTR selection method
> > Signed-off-by: Simon Horman <simon.horman at netronome.com>
> > 
> > ---
> > 
> > v3
> > * Add check to only permit known selection methods: currently none
> > * Use fixed array for fields_array rather than constructing a list
> > * Use NTR instead of NMX as Netronome extension prefix
> > 
> > v2
> > * Use list of struct field_array of TLVs rather than OF1.1 match
> >   for fields field of NTR selection method property
> 
> In oxm_pull_field_array(), I think that it would be appropriate to
> pass NULL to nx_pull_match_entry() for the "mask" parameter, to
> disallow having a mask at all.
> 
> nx_pull_match_entry() checks the value as a value, whereas you want to
> check it as a mask, so probably it's best to bypass
> nx_pull_match_entry() entirely and call nx_pull_entry__() directly so
> that you can check it as a mask yourself.

Thanks. I think you may have mentioned this once before but fixing
it slipped my mind. In any case I have now done so.

> It seems like "struct ofgroup" could be simpler if you just included a
> struct ofputil_group_props in there, instead of each of the members of
> that struct individually.

Thanks, that is a nice simplification.



More information about the dev mailing list