[ovs-discuss] About multiple ovs-vsctl execution

[Ben Pfaff]

You don't need to resubmit it, but it would save me a small amount of
time if you did.

I have not been reviewing or applying patches for a few days because I
am working on a research paper.  The paper deadline is Wednesday
night, so you can expect me to resume more normal work the following
day.

On Mon, Jan 27, 2014 at 06:43:21AM +0000, Ken Ajiro wrote:
> 
>   Hello Ben,
> 
>   To reflect this fix to repository, should I submit patch
>   although the patch was recreated by you? (If I should,
>   I'd like to submit it to ovs-dev as a formal patch.)
> 
>  On Fri, 24 Jan 2014 04:01:04 +0000,
>  ajiro at mxw.nes.nec.co.jp wrote
>  in E-mail "Re: [ovs-discuss] About multiple ovs-vsctl execution":
> 
> >   Yes, your patch also solved my problem.
> > 
> >   Thanks,
> >   Ken
> > 
> >  On Thu, 23 Jan 2014 17:24:27 +0000,
> >  blp at nicira.com wrote
> >  in E-mail "Re: [ovs-discuss] About multiple ovs-vsctl execution":
> > 
> > > >   When two ovs-vsctl update map type column at same time,
> > > >   one ovs-vsctl's update will be discarded although all
> > > >   ovs-vsctl succeeded.
> > > 
> > > Good catch!
> > > 
> > > Does the following simpler patch also solve the problem for you?
> > > 
> > > Thanks,
> > > 
> > > Ben.
> > > 
> > > diff --git a/utilities/ovs-vsctl.c b/utilities/ovs-vsctl.c
> > > index 528b40c..d09775f 100644
> > > --- a/utilities/ovs-vsctl.c
> > > +++ b/utilities/ovs-vsctl.c
> > > @@ -3352,6 +3352,8 @@ set_column(const struct vsctl_table_class *table,
> > >  
> > >          ovsdb_datum_union(&datum, ovsdb_idl_read(row, column),
> > >                            &column->type, false);
> > > +
> > > +        ovsdb_idl_txn_verify(row, column);
> > >          ovsdb_idl_txn_write(row, column, &datum);
> > >      } else {
> > >          struct ovsdb_datum datum;
> > _______________________________________________
> > discuss mailing list
> > discuss at openvswitch.org
> > http://openvswitch.org/mailman/listinfo/discuss