On Mon, 2010-10-04 at 11:56 -0700, Joe Eykholt wrote:
> On 10/4/10 11:23 AM, Bhanu Gollapudi wrote:
> > When number of NPIV ports created are greater than the xids
> > allocated per pool -- for eg., creating 255 NPIV ports on a
> > system with nr_cpu_ids of 32, with each pool containing 128
> > xids -- and then generating a link event - for eg.,
> > shutdown/no shutdown -- on the switch port causes the hang
> > with the following stack trace.
> >
> > Call Trace:
> > schedule_timeout+0x19d/0x230
> > wait_for_common+0xc0/0x170
> > __cancel_work_timer+0xcf/0x1b0
> > fc_disc_stop+0x16/0x30 [libfc]
> > fc_lport_reset_locked+0x47/0x90 [libfc]
> > fc_lport_enter_reset+0x67/0xe0 [libfc]
> > fc_lport_disc_callback+0xbc/0xe0 [libfc]
> > fc_disc_done+0xa8/0xf0 [libfc]
> > fc_disc_timeout+0x29/0x40 [libfc]
> > run_workqueue+0xb8/0x140
> > worker_thread+0x96/0x110
> > kthread+0x96/0xa0
> > child_rip+0xa/0x20
> >
> > Fix is to not cancel the disc_work if discovery is already
> > stopped.
> >
> > Signed-off-by: Bhanu Prakash Gollapudi <[email protected]>
> > ---
> > drivers/scsi/libfc/fc_disc.c | 8 ++++++--
> > include/scsi/libfc.h | 4 +++-
> > 2 files changed, 9 insertions(+), 3 deletions(-)
> >
> > diff --git a/drivers/scsi/libfc/fc_disc.c b/drivers/scsi/libfc/fc_disc.c
> > index 32f67c4..cd1dfc9 100644
> > --- a/drivers/scsi/libfc/fc_disc.c
> > +++ b/drivers/scsi/libfc/fc_disc.c
> > @@ -223,6 +223,7 @@ static void fc_disc_restart(struct fc_disc *disc)
> >
> > FC_DISC_DBG(disc, "Restarting discovery\n");
> >
> > + disc->disc_stopped = 0;
> > disc->requested = 1;
> > if (disc->pending)
> > return;
> > @@ -333,8 +334,10 @@ static void fc_disc_error(struct fc_disc *disc, struct
> > fc_frame *fp)
> > }
> > disc->retry_count++;
> > schedule_delayed_work(&disc->disc_work, delay);
> > - } else
> > + } else {
> > + disc->disc_stopped = 1;
>
> I think you can use disc_done == 0 instead of disc_stopped.
> Also, this is the only place disc_stopped is set, and maybe
> other calls to disc_done should set it, too, so setting it
> in disc_done would be better.
Thanks Joe, I think you were referring to disc->pending.
>
> > fc_disc_done(disc, DISC_EV_FAILED);
> > + }
> > }
> > }
> >
> > @@ -685,7 +688,8 @@ void fc_disc_stop(struct fc_lport *lport)
> > struct fc_disc *disc = &lport->disc;
> >
> > if (disc) {
>
> BTW, note there's no need to test disc for non-NULL here,
> since it contained in the lport. Minor. We can fix it later.
I'll take care of this as well, and re-submit the patch after testing.
Thanks again for your review.
>
> > - cancel_delayed_work_sync(&disc->disc_work);
> > + if (!disc->disc_stopped)
> Change to:
> if (disc->disc_pending)
> > + cancel_delayed_work_sync(&disc->disc_work);
> > fc_disc_stop_rports(disc);
> > }
> > }
> > diff --git a/include/scsi/libfc.h b/include/scsi/libfc.h
> > index 74e0985..36edda1 100644
> > --- a/include/scsi/libfc.h
> > +++ b/include/scsi/libfc.h
> > @@ -736,7 +736,8 @@ struct libfc_function_template {
> > * struct fc_disc - Discovery context
> > * @retry_count: Number of retries
> > * @pending: 1 if discovery is pending, 0 if not
> > - * @requesting: 1 if discovery has been requested, 0 if not
> > + * @requested: 1 if discovery has been requested, 0 if not
> > + * @disc_stopped: 1 if discovery is stopped, 0 if not
> > * @seq_count: Number of sequences used for discovery
> > * @buf_len: Length of the discovery buffer
> > * @disc_id: Discovery ID
> > @@ -752,6 +753,7 @@ struct fc_disc {
> > unsigned char retry_count;
> > unsigned char pending;
> > unsigned char requested;
> > + unsigned char disc_stopped;
> > unsigned short seq_count;
> > unsigned char buf_len;
> > u16 disc_id;
>
_______________________________________________
devel mailing list
[email protected]
http://www.open-fcoe.org/mailman/listinfo/devel
