[ovs-dev] [tests+nxm-ofctl 26/42] ovs-ofctl: Reimplement dumping particular tables.
Ben Pfaff
blp at nicira.com
Fri Dec 3 18:33:52 UTC 2010
That's right, we only have a single table, number 0, so that's why it
worked.
I looked more carefully through the history and saw that this got broken
only on Nov. 8 with commit 8050b31d6 "ofp-parse: Refactor flow
parsing". So it hasn't been broken for long.
On Thu, Dec 02, 2010 at 06:38:30PM -0800, Justin Pettit wrote:
> Why did this work before? Is it because we only had one table (with
> an id of 0), so it was only dumping the contents of that one? If so,
> the problem seems a little more serious than the description
> describes...but not much since we do only define a single table.
>
> Looks good.
>
> --Justin
>
>
> On Nov 23, 2010, at 2:43 PM, Ben Pfaff wrote:
>
> > "dump-flows" and "dump-aggregate" are documented to accept a "table"
> > value to dump only a particular OpenFlow table, but somewhere along the
> > line this got broken. This commit should fix it (though I haven't tested
> > it).
> > ---
> > utilities/ovs-ofctl.c | 2 ++
> > 1 files changed, 2 insertions(+), 0 deletions(-)
> >
> > diff --git a/utilities/ovs-ofctl.c b/utilities/ovs-ofctl.c
> > index 80e3eb3..cc520cc 100644
> > --- a/utilities/ovs-ofctl.c
> > +++ b/utilities/ovs-ofctl.c
> > @@ -478,6 +478,7 @@ do_dump_flows(int argc, char *argv[])
> > parse_ofp_str(&pf, NULL, argc > 2 ? argv[2] : "");
> > ofputil_cls_rule_to_match(&pf.rule, NXFF_OPENFLOW10, &req->match);
> > memset(&req->pad, 0, sizeof req->pad);
> > + req->table_id = pf.table_idx;
> > req->out_port = htons(pf.out_port);
> >
> > dump_stats_transaction(argv[1], request);
> > @@ -494,6 +495,7 @@ do_dump_aggregate(int argc, char *argv[])
> > parse_ofp_str(&pf, NULL, argc > 2 ? argv[2] : "");
> > ofputil_cls_rule_to_match(&pf.rule, NXFF_OPENFLOW10, &req->match);
> > memset(&req->pad, 0, sizeof req->pad);
> > + req->table_id = pf.table_idx;
> > req->out_port = htons(pf.out_port);
> >
> > dump_stats_transaction(argv[1], request);
> > --
> > 1.7.1
> >
> >
> > _______________________________________________
> > dev mailing list
> > dev at openvswitch.org
> > http://openvswitch.org/mailman/listinfo/dev_openvswitch.org
>
More information about the dev
mailing list