On Tue, 2020-05-19 at 02:22 +0300, Roman Bolshakov wrote:
> On Mon, May 18, 2020 at 09:31:42PM +0300, Roman Bolshakov wrote:
> >
>
> P.S. A little bit cleaner alternative would be to avoid session
> deletion
> only if scan needed is set (that allows session deletion of initiator
> ports after fabric discovery if N_Port ID change happened). Please
> let
> me know if I should submit v2 like this:
>
> diff --git a/drivers/scsi/qla2xxx/qla_gs.c
> b/drivers/scsi/qla2xxx/qla_gs.c
> index 42c3ad27f1cb..b9955af5cffe 100644
> --- a/drivers/scsi/qla2xxx/qla_gs.c
> +++ b/drivers/scsi/qla2xxx/qla_gs.c
> @@ -3495,8 +3495,10 @@ void qla24xx_async_gnnft_done(scsi_qla_host_t
> *vha, srb_t *sp)
> if ((fcport->flags & FCF_FABRIC_DEVICE) == 0) {
> qla2x00_clear_loop_id(fcport);
> fcport->flags |= FCF_FABRIC_DEVICE;
> - } else if (fcport->d_id.b24 != rp->id.b24 ||
> - fcport->scan_needed) {
> + } else if (fcport->d_id.b24 != rp->id.b24 ||
> + (fcport->scan_needed &&
> + fcport->port_type != FCT_INITIATOR
> &&
> + fcport->port_type !=
> FCT_NVME_INITIATOR)) {
> qlt_schedule_sess_for_deletion(fcport);
> }
> fcport->d_id.b24 = rp->id.b24;
Yes, this looks like an improvement to me. To express it in my own
words (as the logic is subtle): wrt the original code, this only
changes the behavior if the NPort ID is unchanged but "scan_needed" is
set - in this case the session used to be deleted, but with this patch
it isn't any more.
Thanks,
Martin
