Andy Zhou <az...@ovn.org> wrote on 22/01/2016 04:09:47 AM:
> On Sat, Jan 16, 2016 at 12:16 AM, Liran Schour <lir...@il.ibm.com>
wrote:
> That will be used to by jsonrpc-server for monitor_cond_change
> method See ovsdb-server (1) man page.
>
> Signed-off-by: Liran Schour <lir...@il.ibm.com>
> ---
> ovsdb/monitor.c | 52
+++++++++++++++++++++++++++++++++++++++++++++++++++-
> ovsdb/monitor.h | 7 +++++++
> 2 files changed, 58 insertions(+), 1 deletion(-)
>
> diff --git a/ovsdb/monitor.c b/ovsdb/monitor.c
> index 5e3117a..cf87b98 100644
> --- a/ovsdb/monitor.c
> +++ b/ovsdb/monitor.c
> @@ -591,6 +591,50 @@ ovsdb_monitor_table_condition_add(
> return NULL;
> }
>
> +struct ovsdb_error *
> +ovsdb_monitor_table_condition_change(
> + struct ovsdb_monitor_session_condition
> *condition,
> + const struct ovsdb_table *table,
> + const struct json *added,
> + const struct json *removed)
> +{
> + struct ovsdb_monitor_table_condition *mtc =
> + shash_find_data(&condition->tables, table->schema->name);
> + struct ovsdb_condition added_cnd = OVSDB_CONDITION_INITIALIZER;
> + struct ovsdb_condition removed_cnd = OVSDB_CONDITION_INITIALIZER;
> + struct ovsdb_error *error;
> +
> + if (added) {
> + error = ovsdb_condition_from_json(table->schema, added,
> + NULL, &added_cnd);
> + if (error) {
> + return error;
> + }
> + }
> + if (removed) {
> + error = ovsdb_condition_from_json(table->schema, removed,
> + NULL, &removed_cnd);
> + if (error) {
> + return error;
> + }
> + }
> + if (!ovsdb_conditions_includes(&mtc->new_condition,
> + &added_cnd, &removed_cnd)) {
> + return ovsdb_syntax_error(
> + removed, NULL,
> + "conditions does not include removed");
> + }
> +
> + ovsdb_condition_add(&mtc->new_condition, &added_cnd);
> + ovsdb_condition_remove(&mtc->new_condition, &removed_cnd);
> +
> + if (!ovsdb_condition_empty(&mtc->new_condition)) {
> + condition->can_cache = false;
> + }
> If we do:
> condition->can_cache = ovsdb_condition_empty(&mtc->new_condition);
> then we don't have worry about how can_cache is initialized. Slightly
better
> for maintenance. I think.
>
Not exactly. Can_cache is true only if all tables has an empty condition.
Maybe we should hold a counter of the number of tables that has a none
empty condition. Then a counter==0 means that can_cache is true. Will
change the code to include that. (By that we will add the possibility that
a session is being backed to a state of can_cache == true)
> +
> + return NULL;
> +}
> +
> /* Set session's condition in this ovsdb_monitor */
> static void
> ovsdb_monitor_set_condition(
> @@ -764,7 +808,7 @@ ovsdb_monitor_compose_row_update(
> * for 'row' within * 'mt', or NULL if no row update should be sent.
> *
> * The caller should specify 'initial' as true if the returned JSON is
> - * going to be used as part of the initial reply to a "monitor2"
request,
> + * going to be used as part of the initial reply to a "monitor_cond"
request,
> * false if it is going to be used as part of an "update2"
notification.
> *
> * 'changed' must be a scratch buffer for internal use that is at least
> @@ -948,6 +992,12 @@ ovsdb_monitor_get_update(
>
> ovsdb_monitor_table_untrack_changes(mt, prev_txn);
> ovsdb_monitor_table_track_changes(mt, next_txn);
> +
> + if (dbmon->condition && ovsdb_condition_cmp(mt->old_condition,
> + mt->new_condition))
{
> + ovsdb_condition_destroy(mt->old_condition);
> + ovsdb_condition_clone(mt->old_condition,
mt->new_condition);
> + }
> }
> *unflushed = next_txn;
> ovsdb_monitor_unset_condition(dbmon);
> diff --git a/ovsdb/monitor.h b/ovsdb/monitor.h
> index d4063f7..0f6c360 100644
> --- a/ovsdb/monitor.h
> +++ b/ovsdb/monitor.h
> @@ -106,4 +106,11 @@ void ovsdb_monitor_session_condition_bind(
> const struct ovsdb_monitor_session_condition
*,
> const struct ovsdb_monitor *);
>
> +struct ovsdb_error *
> +ovsdb_monitor_table_condition_change(
> + struct ovsdb_monitor_session_condition
*condition,
> + const struct ovsdb_table *table,
> + const struct json *added,
> + const struct json *removed);
> +
> #endif
> --
> 2.1.4
>
>
> _______________________________________________
> dev mailing list
> dev@openvswitch.org
> http://openvswitch.org/mailman/listinfo/dev
_______________________________________________
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
