From: Benjamin Marzinski Subject: Re: [PATCH 66/74] libmultipath: update_pathvec_from_dm: handle pp->mpp mismatch Date: Sun, 19 Jul 2020 00:12:00 -0500 Message-ID: <20200719051200.GX11089@octiron.msp.redhat.com> References: <20200709105145.9211-1-mwilck@suse.com> <20200709105145.9211-14-mwilck@suse.com> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <20200709105145.9211-14-mwilck@suse.com> List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: dm-devel-bounces@redhat.com Errors-To: dm-devel-bounces@redhat.com Content-Disposition: inline To: mwilck@suse.com Cc: dm-devel@redhat.com List-Id: dm-devel.ids On Thu, Jul 09, 2020 at 12:51:37PM +0200, mwilck@suse.com wrote: > From: Martin Wilck > > Treat this like a WWID mismatch. > > Signed-off-by: Martin Wilck > --- > libmultipath/structs_vec.c | 37 +++++++++++++++++++++++-------------- > 1 file changed, 23 insertions(+), 14 deletions(-) > > diff --git a/libmultipath/structs_vec.c b/libmultipath/structs_vec.c > index 5dd37d5..8651b98 100644 > --- a/libmultipath/structs_vec.c > +++ b/libmultipath/structs_vec.c > @@ -118,6 +118,12 @@ bool update_pathvec_from_dm(vector pathvec, struct multipath *mpp, > goto delete_pg; > > vector_foreach_slot(pgp->paths, pp, j) { > + > + if (pp->mpp && pp->mpp != mpp) { > + condlog(0, "BUG: %s: found path %s which is already in %s", > + mpp->alias, pp->dev, pp->mpp->alias); > + goto bad_path; > + } > pp->mpp = mpp; > > if (pp->udev) { > @@ -163,25 +169,28 @@ bool update_pathvec_from_dm(vector pathvec, struct multipath *mpp, > > condlog(0, "%s: path %s WWID %s doesn't match, removing from map", > mpp->wwid, pp->dev_t, pp->wwid); > - /* > - * This path exists, but in the wong map. > - * We can't reload the map from here. > - * Instead, treat this path like "missing udev", > - * which it probably is. > - * check_path() will trigger an uevent > - * and reset pp->tick. > - */ > - must_reload = true; > - pp->mpp = NULL; > - dm_fail_path(mpp->alias, pp->dev_t); > - vector_del_slot(pgp->paths, j--); > - pp->initialized = INIT_MISSING_UDEV; > - pp->tick = 1; > + goto bad_path; > } > condlog(2, "%s: adding new path %s", > mpp->alias, pp->dev); > store_path(pathvec, pp); > + > } > + continue; > + > + bad_path: > + /* > + * This path exists, but in the wrong map. > + * We can't reload the map from here. > + * Instead, treat this path like "missing udev". > + * check_path() will trigger an uevent and reset pp->tick. > + */ > + must_reload = true; > + pp->mpp = NULL; > + dm_fail_path(mpp->alias, pp->dev_t); > + vector_del_slot(pgp->paths, j--); > + pp->initialized = INIT_MISSING_UDEV; > + pp->tick = 1; Is there a reason not to call orphan_path() to clean up things like any open fd, until we figure out what to do with the path. > } > if (VECTOR_SIZE(pgp->paths) != 0) > continue; > -- > 2.26.2