LU-14037 osd: track commit cb in flight

and wait for completion in osd-ldiskfs as mntput() is not synchronous
and returns quickly scheduling real umount to another thread while
osd is shutting down. as a result commit callbacks may come when
osd has been released yet.
the patch replaces lu_device_get() and lu_device_put() for transactions
with a dedicated counter which later used to wait for callback completion.

Signed-off-by: Alex Zhuravlev <bzzz@whamcloud.com>
Change-Id: Iaa6089b9158bc2a4bea8e33f4cfcce27395689ca
Reviewed-on: https://review.whamcloud.com/40258
Reviewed-by: Mike Pershin <mpershin@whamcloud.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Lai Siyao <lai.siyao@whamcloud.com>
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>

diff --git a/lustre/osd-ldiskfs/osd_handler.c b/lustre/osd-ldiskfs/osd_handler.c
index ec9febb..554a26d 100644 (file)
--- a/lustre/osd-ldiskfs/osd_handler.c
+++ b/lustre/osd-ldiskfs/osd_handler.c
@@ -1695,6 +1695,7 @@ static void osd_trans_commit_cb(struct super_block *sb,
        struct osd_thandle *oh = container_of(jcb, struct osd_thandle, ot_jcb);
        struct thandle *th = &oh->ot_super;
        struct lu_device *lud = &th->th_dev->dd_lu_dev;
        struct osd_thandle *oh = container_of(jcb, struct osd_thandle, ot_jcb);
        struct thandle *th = &oh->ot_super;
        struct lu_device *lud = &th->th_dev->dd_lu_dev;

+       struct osd_device *osd = osd_dev(lud);

        struct dt_txn_commit_cb *dcb, *tmp;
 
        LASSERT(oh->ot_handle == NULL);
        struct dt_txn_commit_cb *dcb, *tmp;
 
        LASSERT(oh->ot_handle == NULL);


@@ -1714,7 +1715,8 @@ static void osd_trans_commit_cb(struct super_block *sb,
        }
 
        lu_ref_del_at(&lud->ld_reference, &oh->ot_dev_link, "osd-tx", th);
        }
 
        lu_ref_del_at(&lud->ld_reference, &oh->ot_dev_link, "osd-tx", th);

-       lu_device_put(lud);
+       if (atomic_dec_and_test(&osd->od_commit_cb_in_flight))
+               wake_up(&osd->od_commit_cb_done);

        th->th_dev = NULL;
 
        OBD_FREE_PTR(oh);
        th->th_dev = NULL;
 
        OBD_FREE_PTR(oh);


@@ -1897,7 +1899,7 @@ static int osd_trans_start(const struct lu_env *env, struct dt_device *d,
                oh->ot_handle = jh;
                LASSERT(oti->oti_txns == 0);
 
                oh->ot_handle = jh;
                LASSERT(oti->oti_txns == 0);
 

-               lu_device_get(&d->dd_lu_dev);
+               atomic_inc(&dev->od_commit_cb_in_flight);

                lu_ref_add_at(&d->dd_lu_dev.ld_reference, &oh->ot_dev_link,
                              "osd-tx", th);
                oti->oti_txns++;
                lu_ref_add_at(&d->dd_lu_dev.ld_reference, &oh->ot_dev_link,
                              "osd-tx", th);
                oti->oti_txns++;


@@ -7579,6 +7581,8 @@ static void osd_umount(const struct lu_env *env, struct osd_device *o)
        if (o->od_mnt != NULL) {
                shrink_dcache_sb(osd_sb(o));
                osd_sync(env, &o->od_dt_dev);
        if (o->od_mnt != NULL) {
                shrink_dcache_sb(osd_sb(o));
                osd_sync(env, &o->od_dt_dev);

+               wait_event(o->od_commit_cb_done,
+                         !atomic_read(&o->od_commit_cb_in_flight));

 
                mntput(o->od_mnt);
                o->od_mnt = NULL;
 
                mntput(o->od_mnt);
                o->od_mnt = NULL;


@@ -7816,6 +7820,7 @@ static int osd_device_init0(const struct lu_env *env,
        spin_lock_init(&o->od_lock);
        o->od_index_backup_policy = LIBP_NONE;
        o->od_t10_type = 0;
        spin_lock_init(&o->od_lock);
        o->od_index_backup_policy = LIBP_NONE;
        o->od_t10_type = 0;

+       init_waitqueue_head(&o->od_commit_cb_done);

 
        o->od_read_cache = 1;
        o->od_writethrough_cache = 1;
 
        o->od_read_cache = 1;
        o->od_writethrough_cache = 1;

diff --git a/lustre/osd-ldiskfs/osd_internal.h b/lustre/osd-ldiskfs/osd_internal.h
index 2ff5711..7351b0a 100644 (file)
--- a/lustre/osd-ldiskfs/osd_internal.h
+++ b/lustre/osd-ldiskfs/osd_internal.h
@@ -343,6 +343,8 @@ struct osd_device {
        int                      od_index_backup_stop;
        /* T10PI type, zero if not supported  */
        enum osd_t10_type        od_t10_type;
        int                      od_index_backup_stop;
        /* T10PI type, zero if not supported  */
        enum osd_t10_type        od_t10_type;

+       atomic_t                 od_commit_cb_in_flight;
+       wait_queue_head_t        od_commit_cb_done;

 };
 
 static inline struct qsd_instance *osd_def_qsd(struct osd_device *osd)
 };
 
 static inline struct qsd_instance *osd_def_qsd(struct osd_device *osd)