On Wed, Mar 05 2014 at 3:59am -0500,
Hannes Reinecke <hare@xxxxxxx> wrote:
> On 03/03/2014 12:12 AM, Junichi Nomura wrote:
> > On 02/28/14 23:33, Hannes Reinecke wrote:
> >> @@ -1214,9 +1189,12 @@ static void pg_init_done(void *data, int errors)
> >>
> >> if (!m->pg_init_required)
> >> m->queue_io = 0;
> >> -
> >> - m->pg_init_delay_retry = delay_retry;
> >> - queue_work(kmultipathd, &m->process_queued_ios);
> >> + else if (m->current_pg) {
> >> + m->pg_init_delay_retry = delay_retry;
> >> + if (__pg_init_all_paths(m))
> >> + goto out;
> >> + m->queue_io = 0;
> >> + }
> >
> > Hi Hannes,
> >
> > If m->pg_init_required && !m->current_pg,
> > doesn't m->queue_io remain 1 while nobody kicks pg_init?
> >
> > That's why I suggested this:
> >
> > if (m->pg_init_required) {
> > m->pg_init_delay_retry = delay_retry;
> > if (__pg_init_all_paths(m))
> > goto out;
> > }
> > m->queue_io = 0;
> >
> Yeah, sorry. You are right.
>
> Mike, can you fix it up or do you need an additional round of patches?
I'll take care of it. I won't be looking at these patches until next
week though.
Mike
--
dm-devel mailing list
dm-devel@xxxxxxxxxx
https://www.redhat.com/mailman/listinfo/dm-devel
