Hi Zhouhui, Your patch is not a formatted one. At least, it should include Signed-off-by Thanks, Changwei On 2018/3/12 15:47, wuzhouhui wrote: > When flush works belong a specific multipath, we flush workqueue only > in necessary. > > (That's my first time to send a patch, please remind me if anything > wrong) > > diff --git a/drivers/md/dm-mpath.c b/drivers/md/dm-mpath.c > index f7810cc..0f14732 100644 > --- a/drivers/md/dm-mpath.c > +++ b/drivers/md/dm-mpath.c > @@ -1154,9 +1154,10 @@ static void flush_multipath_work(struct multipath *m) > set_bit(MPATHF_PG_INIT_DISABLED, &m->flags); > smp_mb__after_atomic(); > > - flush_workqueue(kmpath_handlerd); > + if (atomic_read(&m->pg_init_in_progress)) > + flush_workqueue(kmpath_handlerd); > multipath_wait_for_pg_init_completion(m); > - flush_workqueue(kmultipathd); > + flush_work(&m->process_queued_bios); > flush_work(&m->trigger_event); > > clear_bit(MPATHF_PG_INIT_DISABLED, &m->flags); > > -- > dm-devel mailing list > dm-devel@xxxxxxxxxx > https://www.redhat.com/mailman/listinfo/dm-devel > -- dm-devel mailing list dm-devel@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/dm-devel
