On Tue, Dec 12, 2023 at 7:45 AM Dumitru Ceara <[email protected]> wrote:
>
> On 12/12/23 05:04, Mike Pattrick wrote:
> > In the case of a weak reference, clearing all change_seqno's can delete
> > useful information. Instead of clearing all seqno's when removing
> > track_node, only clear those values in cases including row insertion,
> > and row deleting if no dst_arcs remain.
> >
> > Fixes: 95689f166818 ("ovsdb-idl: Preserve references for deleted rows.")
> > Reported-at: https://issues.redhat.com/browse/FDP-193
> > Signed-off-by: Mike Pattrick <[email protected]>
> > ---
>
> Thanks for the fix, Mike!  Overall I think the idea of the fix is OK, I
> only have one comment below.
>
> >  lib/ovsdb-idl.c | 13 +++++++++++--
> >  1 file changed, 11 insertions(+), 2 deletions(-)
> >
> > diff --git a/lib/ovsdb-idl.c b/lib/ovsdb-idl.c
> > index 634fbb56d..77f1a6db4 100644
> > --- a/lib/ovsdb-idl.c
> > +++ b/lib/ovsdb-idl.c
> > @@ -177,6 +177,7 @@ static void 
> > ovsdb_idl_row_mark_backrefs_for_reparsing(struct ovsdb_idl_row *);
> >  static void ovsdb_idl_row_track_change(struct ovsdb_idl_row *,
> >                                         enum ovsdb_idl_change);
> >  static void ovsdb_idl_row_untrack_change(struct ovsdb_idl_row *);
> > +static void ovsdb_idl_row_clear_changeseqno(struct ovsdb_idl_row *);
> >
> >  static void ovsdb_idl_txn_abort_all(struct ovsdb_idl *);
> >  static bool ovsdb_idl_txn_extract_mutations(struct ovsdb_idl_row *,
> > @@ -1374,6 +1375,7 @@ ovsdb_idl_track_clear__(struct ovsdb_idl *idl, bool 
> > flush_all)
> >                      row->updated = NULL;
> >                  }
> >                  ovsdb_idl_row_untrack_change(row);
> > +                ovsdb_idl_row_clear_changeseqno(row);
> >
> >                  if (ovsdb_idl_row_is_orphan(row)) {
> >                      ovsdb_idl_row_unparse(row);
> > @@ -1554,6 +1556,8 @@ ovsdb_idl_reparse_deleted(struct ovsdb_idl *db)
> >          if (ovs_list_is_empty(&row->dst_arcs)
> >                  || ovsdb_idl_track_is_set(row->table)) {
> >              ovsdb_idl_row_track_change(row, OVSDB_IDL_CHANGE_DELETE);
> > +        } else {
> > +            ovsdb_idl_row_clear_changeseqno(row);
>
> I'm not sure we need this call to ovsdb_idl_row_clear_changeseqno()
> here.  If I comment it out all tests still pass.  What case is it taking
> care of?
>
> If a deleted row is "re-inserted" we already cleared the change seqno in
> ovsdb_idl_process_update().
>
> If a deleted row is not referenced anymore (no dst arcs) then it's fine
> to keep its deletion change seqno.
>
> Or am I missing something?

No, that's a good point. I can resubmit without that line.

-M

>
> FWIW, I did run the OVN test suite with this OVS change applied and
> those tests passed too:
>
> https://github.com/dceara/ovn/actions/runs/7180690942
>
> Thanks,
> Dumitru
>
> >          }
> >      }
> >  }
> > @@ -1632,6 +1636,7 @@ ovsdb_idl_process_update(struct ovsdb_idl_table 
> > *table,
> >                                   ru->columns);
> >          } else if (ovsdb_idl_row_is_orphan(row)) {
> >              ovsdb_idl_row_untrack_change(row);
> > +            ovsdb_idl_row_clear_changeseqno(row);
> >              ovsdb_idl_insert_row(row, ru->columns);
> >          } else {
> >              VLOG_ERR_RL(&semantic_rl, "cannot add existing row "UUID_FMT" 
> > to "
> > @@ -2283,11 +2288,15 @@ ovsdb_idl_row_untrack_change(struct ovsdb_idl_row 
> > *row)
> >          return;
> >      }
> >
> > +    ovs_list_remove(&row->track_node);
> > +    ovs_list_init(&row->track_node);
> > +}
> > +
> > +static void ovsdb_idl_row_clear_changeseqno(struct ovsdb_idl_row *row)
> > +{
> >      row->change_seqno[OVSDB_IDL_CHANGE_INSERT] =
> >          row->change_seqno[OVSDB_IDL_CHANGE_MODIFY] =
> >          row->change_seqno[OVSDB_IDL_CHANGE_DELETE] = 0;
> > -    ovs_list_remove(&row->track_node);
> > -    ovs_list_init(&row->track_node);
> >  }
> >
> >  static struct ovsdb_idl_row *
>

_______________________________________________
dev mailing list
[email protected]
https://mail.openvswitch.org/mailman/listinfo/ovs-dev

Reply via email to