[ofa-general] [PATCH] opensm/osm_multicast.[ch]: simplify flows, remove unused functions
Sasha Khapyorsky
sashak at voltaire.com
Tue Sep 16 08:47:52 PDT 2008
Hi Hal,
On 09:52 Tue 16 Sep , Hal Rosenstock wrote:
> > +void
> > +osm_mgrp_send_delete_notice(IN osm_subn_t * const p_subn,
> > + IN osm_log_t * const p_log, IN osm_mgrp_t * p_mgrp)
> > +{
> > + mgrp_send_notice(p_subn, p_log, p_mgrp, 67);
> > }
>
> Any reason not to eliminate this extra call level ?
Basically I agree, and in future it can be eliminated (I have few more
patches in the stack :)). This patch is "cosmetic" one, so I didn't change
any OpenSM internal API.
> > -/**********************************************************************
> > - **********************************************************************/
> > void
> > osm_mgrp_send_create_notice(IN osm_subn_t * const p_subn,
> > IN osm_log_t * const p_log, IN osm_mgrp_t * p_mgrp)
> > {
> > - ib_mad_notice_attr_t notice;
> > - ib_api_status_t status;
> > -
> > - OSM_LOG_ENTER(p_log);
> > -
> > - /* prepare the needed info */
> > -
> > - /* details of the notice */
> > - notice.generic_type = 0x83; /* Generic SubnMgt type */
> > - ib_notice_set_prod_type_ho(¬ice, 4); /* A Class Manager generator */
> > - notice.g_or_v.generic.trap_num = CL_HTON16(66); /* create of mcg */
> > - /* The sm_base_lid is saved in network order already. */
> > - notice.issuer_lid = p_subn->sm_base_lid;
> > - /* following o14-12.1.11 and table 120 p726 */
> > - /* we need to provide the MGID */
> > - memcpy(&(notice.data_details.ntc_64_67.gid),
> > - &(p_mgrp->mcmember_rec.mgid), sizeof(ib_gid_t));
> > -
> > - /* According to page 653 - the issuer gid in this case of trap
> > - is the SM gid, since the SM is the initiator of this trap. */
> > - notice.issuer_gid.unicast.prefix = p_subn->opt.subnet_prefix;
> > - notice.issuer_gid.unicast.interface_id = p_subn->sm_port_guid;
> > -
> > - status = osm_report_notice(p_log, p_subn, ¬ice);
> > - if (status != IB_SUCCESS) {
> > - OSM_LOG(p_log, OSM_LOG_ERROR, "ERR 7602: "
> > - "Error sending trap reports (%s)\n",
> > - ib_get_err_str(status));
> > - goto Exit;
> > - }
> > -
> > -Exit:
> > - OSM_LOG_EXIT(p_log);
> > + mgrp_send_notice(p_subn, p_log, p_mgrp, 66);
>
> Similarly, any reason not to eliminate this extra calling level ?
Ditto.
Sasha
More information about the general
mailing list