Whamcloud - gitweb
LU-6142 osd: Fix style issues for osd_object.c 57/54257/3
authorArshad Hussain <arshad.hussain@aeoncomputing.com>
Sun, 3 Mar 2024 16:55:29 +0000 (22:25 +0530)
committerOleg Drokin <green@whamcloud.com>
Tue, 30 Apr 2024 06:50:28 +0000 (06:50 +0000)
This patch fixes issues reported by checkpatch
for file lustre/osd-zfs/osd_object.c

Test-Parameters: trivial
Signed-off-by: Arshad Hussain <arshad.hussain@aeoncomputing.com>
Change-Id: I32a91583b37752a722cf558dfa14f191163090b3
Reviewed-on: https://review.whamcloud.com/c/fs/lustre-release/+/54257
Tested-by: jenkins <devops@whamcloud.com>
Tested-by: Maloo <maloo@whamcloud.com>
Reviewed-by: Timothy Day <timday@amazon.com>
Reviewed-by: Andreas Dilger <adilger@whamcloud.com>
Reviewed-by: Oleg Drokin <green@whamcloud.com>
lustre/osd-zfs/osd_object.c

index 6ef58d7..eb91671 100644 (file)
@@ -193,8 +193,8 @@ static int __osd_object_attr_get(const struct lu_env *env, struct osd_device *o,
        struct lu_buf buf;
        int cnt = 0;
        int              rc;
-       ENTRY;
 
+       ENTRY;
        LASSERT(obj->oo_dn != NULL);
 
        la->la_valid |= LA_ATIME | LA_MTIME | LA_CTIME | LA_BTIME | LA_MODE |
@@ -353,8 +353,8 @@ static void osd_obj_set_blksize(const struct lu_env *env,
        dnode_t *dn = obj->oo_dn;
        uint32_t blksz;
        int rc = 0;
-       ENTRY;
 
+       ENTRY;
        LASSERT(!osd_oti_get(env)->oti_in_trans);
 
        tx = dmu_tx_create(osd->od_os);
@@ -368,8 +368,8 @@ static void osd_obj_set_blksize(const struct lu_env *env,
        rc = -dmu_tx_assign(tx, TXG_WAIT);
        if (rc) {
                dmu_tx_abort(tx);
-               CERROR("%s: fail to assign tx to set blksize for "DFID
-                      ": rc = %d\n", osd->od_svname, PFID(fid), rc);
+               CERROR("%s: fail to assign tx to set blksize for "DFID": rc = %d\n",
+                      osd->od_svname, PFID(fid), rc);
                RETURN_EXIT;
        }
 
@@ -413,8 +413,8 @@ static int osd_object_init0(const struct lu_env *env, struct osd_object *obj)
        struct osd_device       *osd = osd_obj2dev(obj);
        const struct lu_fid     *fid = lu_object_fid(&obj->oo_dt.do_lu);
        int                      rc = 0;
-       ENTRY;
 
+       ENTRY;
        LASSERT(obj->oo_dn);
 
        rc = osd_object_sa_init(obj, osd);
@@ -457,8 +457,8 @@ static int osd_check_lma(const struct lu_env *env, struct osd_object *obj)
        int                     rc;
        struct lustre_mdt_attrs *lma;
        const struct lu_fid *rfid = lu_object_fid(&obj->oo_dt.do_lu);
-       ENTRY;
 
+       ENTRY;
        BUILD_BUG_ON(sizeof(info->oti_buf) < sizeof(*lma));
        lma = (struct lustre_mdt_attrs *)info->oti_buf;
        buf.lb_buf = lma;
@@ -470,17 +470,16 @@ static int osd_check_lma(const struct lu_env *env, struct osd_object *obj)
                lustre_lma_swab(lma);
                if (unlikely((lma->lma_incompat & ~LMA_INCOMPAT_SUPP) ||
                             CFS_FAIL_CHECK(OBD_FAIL_OSD_LMA_INCOMPAT))) {
-                       CWARN("%s: unsupported incompat LMA feature(s) %#x for "
-                             "fid = "DFID"\n", osd_obj2dev(obj)->od_svname,
-                             lma->lma_incompat & ~LMA_INCOMPAT_SUPP,
-                             PFID(rfid));
                        rc = -EOPNOTSUPP;
-               } else if (unlikely(!lu_fid_eq(rfid, &lma->lma_self_fid))) {
-                       CERROR("%s: FID-in-LMA "DFID" does not match the "
-                             "object self-fid "DFID"\n",
+                       CWARN("%s: unsupported incompat LMA feature(s) %#x for fid = "DFID": rc = %d\n",
                              osd_obj2dev(obj)->od_svname,
-                             PFID(&lma->lma_self_fid), PFID(rfid));
+                             lma->lma_incompat & ~LMA_INCOMPAT_SUPP,
+                             PFID(rfid), rc);
+               } else if (unlikely(!lu_fid_eq(rfid, &lma->lma_self_fid))) {
                        rc = -EREMCHG;
+                       CERROR("%s: FID-in-LMA "DFID" does not match the object self-fid "DFID": rc = %d\n",
+                             osd_obj2dev(obj)->od_svname,
+                             PFID(&lma->lma_self_fid), PFID(rfid), rc);
                } else {
                        struct osd_device *osd = osd_obj2dev(obj);
 
@@ -547,8 +546,8 @@ static int osd_object_init(const struct lu_env *env, struct lu_object *l,
        int rc = 0;
        int rc1;
        bool remote = false;
-       ENTRY;
 
+       ENTRY;
        LASSERT(osd_invariant(obj));
 
        if (fid_is_otable_it(&l->lo_header->loh_fid)) {
@@ -633,7 +632,8 @@ trigger:
         *
         * During the OI scrub, if we cannot find the OI mapping, we may still
         * have change to map the FID to local OID via lookup the dir
-        * /REMOTE_PARENT_DIR. */
+        * /REMOTE_PARENT_DIR.
+        */
        if (!remote && !fid_is_on_ost(env, osd, fid)) {
                osd_fid2str(name, fid, sizeof(info->oti_str));
                rc = osd_zap_lookup(osd, osd->od_remote_parent_dir,
@@ -717,7 +717,8 @@ osd_object_unlinked_add(struct osd_object *obj, struct osd_thandle *oh)
 
        /* the object is supposed to be exclusively locked by
         * the caller (osd_destroy()), while the transaction
-        * (oh) is per-thread and not shared */
+        * (oh) is per-thread and not shared
+        */
        if (likely(list_empty(&obj->oo_unlinked_linkage))) {
                list_add(&obj->oo_unlinked_linkage, &oh->ot_unlinked_list);
                rc = 0;
@@ -759,8 +760,8 @@ static int osd_declare_destroy(const struct lu_env *env, struct dt_object *dt,
        dnode_t *dn;
        int                      rc;
        uint64_t                 zapid;
-       ENTRY;
 
+       ENTRY;
        LASSERT(th != NULL);
        if (unlikely(obj->oo_dn == NULL))
                RETURN(-ENOENT);
@@ -803,8 +804,7 @@ static int osd_declare_destroy(const struct lu_env *env, struct dt_object *dt,
                osd_tx_hold_zap(oh->ot_tx, osd->od_remote_parent_dir,
                                NULL, FALSE, NULL);
 
-       /* will help to find FID->ino when this object is being
-        * added to PENDING/ */
+       /* will help to find FID->ino when this obj is being added to PENDING */
        osd_idc_find_and_init(env, osd, obj);
 
        RETURN(0);
@@ -822,8 +822,8 @@ static int osd_destroy(const struct lu_env *env, struct dt_object *dt,
        int                      rc;
        uint64_t                 oid, zapid;
        dnode_t *zdn;
-       ENTRY;
 
+       ENTRY;
        down_write(&obj->oo_guard);
 
        if (unlikely(!dt_object_exists(dt) || obj->oo_destroyed))
@@ -856,7 +856,8 @@ static int osd_destroy(const struct lu_env *env, struct dt_object *dt,
                /* this may happen if the destroy wasn't declared
                 * e.g. when the object is created and then destroyed
                 * in the same transaction - we don't need additional
-                * space for destroy specifically */
+                * space for destroy specifically
+                */
                LASSERT(obj->oo_attr.la_size <= osd_sync_destroy_max_size);
                rc = -dmu_object_free(osd->od_os, oid, oh->ot_tx);
                if (rc)
@@ -883,7 +884,8 @@ static int osd_destroy(const struct lu_env *env, struct dt_object *dt,
        }
 
        /* Remove the OI mapping after the destroy to handle the race with
-        * OI scrub that may insert missed OI mapping during the interval. */
+        * OI scrub that may insert missed OI mapping during the interval.
+        */
        rc = osd_zap_remove(osd, zapid, zdn, buf, oh->ot_tx);
        if (unlikely(rc == -ENOENT))
                rc = 0;
@@ -899,7 +901,7 @@ out:
        if (rc == 0)
                obj->oo_destroyed = 1;
        up_write(&obj->oo_guard);
-       RETURN (0);
+       RETURN(0);
 }
 
 static void osd_object_delete(const struct lu_env *env, struct lu_object *l)
@@ -948,7 +950,7 @@ static int osd_object_print(const struct lu_env *env, void *cookie,
 }
 
 static void osd_read_lock(const struct lu_env *env, struct dt_object *dt,
-                         unsigned role)
+                         unsigned int role)
 {
        struct osd_object *obj = osd_dt_obj(dt);
 
@@ -958,7 +960,7 @@ static void osd_read_lock(const struct lu_env *env, struct dt_object *dt,
 }
 
 static void osd_write_lock(const struct lu_env *env, struct dt_object *dt,
-                          unsigned role)
+                          unsigned int role)
 {
        struct osd_object *obj = osd_dt_obj(dt);
 
@@ -1036,10 +1038,12 @@ static int osd_attr_get(const struct lu_env *env, struct dt_object *dt,
 
        /* with ZFS_DEBUG zrl_add_debug() called by DB_DNODE_ENTER()
         * from within sa_object_size() can block on a mutex, so
-        * we can't call sa_object_size() holding rwlock */
+        * we can't call sa_object_size() holding rwlock
+        */
        sa_object_size(obj->oo_sa_hdl, &blksize, &blocks);
        /* we do not control size of indices, so always calculate
-        * it from number of blocks reported by DMU */
+        * it from number of blocks reported by DMU
+        */
        if (S_ISDIR(attr->la_mode)) {
                attr->la_size = 512 * blocks;
                rc = -zap_count(osd->od_os, obj->oo_dn->dn_object,
@@ -1161,9 +1165,8 @@ static int osd_declare_attr_set(const struct lu_env *env,
        uint32_t                 blksize;
        int                      rc = 0;
        bool                     found;
-       ENTRY;
-
 
+       ENTRY;
        LASSERT(handle != NULL);
        LASSERT(osd_invariant(obj));
 
@@ -1177,7 +1180,8 @@ static int osd_declare_attr_set(const struct lu_env *env,
        LASSERT(oh->ot_tx != NULL);
        /* regular attributes are part of the bonus buffer */
        /* let's check whether this object is already part of
-        * transaction.. */
+        * transaction..
+        */
        found = false;
        for (txh = list_head(&oh->ot_tx->tx_holds); txh;
             txh = list_next(&oh->ot_tx->tx_holds, txh)) {
@@ -1186,7 +1190,8 @@ static int osd_declare_attr_set(const struct lu_env *env,
                if (txh->txh_dnode->dn_object != obj->oo_dn->dn_object)
                        continue;
                /* this object is part of the transaction already
-                * we don't need to declare bonus again */
+                * we don't need to declare bonus again
+                */
                found = true;
                break;
        }
@@ -1283,7 +1288,6 @@ static int osd_attr_set(const struct lu_env *env, struct dt_object *dt,
        int                      rc = 0;
 
        ENTRY;
-
        down_read(&obj->oo_guard);
        if (unlikely(!dt_object_exists(dt) || obj->oo_destroyed))
                GOTO(out, rc = -ENOENT);
@@ -1293,8 +1297,7 @@ static int osd_attr_set(const struct lu_env *env, struct dt_object *dt,
        LASSERT(obj->oo_sa_hdl);
 
        oh = container_of(handle, struct osd_thandle, ot_super);
-       /* Assert that the transaction has been assigned to a
-          transaction group. */
+       /* Assert that the transaction has been assigned to a transaction grp */
        LASSERT(oh->ot_tx->tx_txg != 0);
 
        if (CFS_FAIL_CHECK(OBD_FAIL_OSD_FID_MAPPING) && !osd->od_is_ost) {
@@ -1346,7 +1349,8 @@ static int osd_attr_set(const struct lu_env *env, struct dt_object *dt,
                        buf.lb_len = sizeof(info->oti_buf);
 
                        /* Please do NOT call osd_xattr_get() directly, that
-                        * will cause recursive down_read() on oo_guard. */
+                        * will cause recursive down_read() on oo_guard.
+                        */
                        rc = osd_xattr_get_internal(env, obj, &buf,
                                                    XATTR_NAME_LMA, &size);
                        if (!rc && unlikely(size < sizeof(*lma))) {
@@ -1453,8 +1457,7 @@ lock:
        }
        if (valid & LA_FLAGS) {
                osa->flags = attrs_fs2zfs(la->la_flags);
-               /* many flags are not supported by zfs, so ensure a good cached
-                * copy */
+               /* many flags not supported by zfs, ensure good cached copy */
                obj->oo_attr.la_flags = attrs_zfs2fs(osa->flags);
 #ifdef ZFS_PROJINHERIT
                if (obj->oo_with_projid && osd->od_projectused_dn)
@@ -1489,7 +1492,6 @@ out:
  *
  * XXX temporary solution.
  */
-
 static void osd_ah_init(const struct lu_env *env, struct dt_allocation_hint *ah,
                        struct dt_object *parent, struct dt_object *child,
                        umode_t child_mode)
@@ -1519,19 +1521,19 @@ static int osd_declare_create(const struct lu_env *env, struct dt_object *dt,
        uint64_t                 zapid;
        dnode_t                 *dn;
        int                      rc, dnode_size;
-       ENTRY;
 
+       ENTRY;
        LASSERT(dof);
 
        switch (dof->dof_type) {
-               case DFT_REGULAR:
-               case DFT_SYM:
-               case DFT_NODE:
-                       if (obj->oo_dt.do_body_ops == NULL)
-                               obj->oo_dt.do_body_ops = &osd_body_ops;
-                       break;
-               default:
-                       break;
+       case DFT_REGULAR:
+       case DFT_SYM:
+       case DFT_NODE:
+               if (obj->oo_dt.do_body_ops == NULL)
+                       obj->oo_dt.do_body_ops = &osd_body_ops;
+               break;
+       default:
+               break;
        }
 
        LASSERT(handle != NULL);
@@ -1544,24 +1546,24 @@ static int osd_declare_create(const struct lu_env *env, struct dt_object *dt,
        dnode_size = size_roundup_power2(obj->oo_ea_in_bonus + 32);
 
        switch (dof->dof_type) {
-               case DFT_DIR:
-                       dt->do_index_ops = &osd_dir_ops;
-                       fallthrough;
-               case DFT_INDEX:
-                       /* for zap create */
-                       dmu_tx_hold_zap(oh->ot_tx, DMU_NEW_OBJECT, FALSE, NULL);
-                       dmu_tx_hold_sa_create(oh->ot_tx, dnode_size);
-                       break;
-               case DFT_REGULAR:
-               case DFT_SYM:
-               case DFT_NODE:
-                       /* first, we'll create new object */
-                       dmu_tx_hold_sa_create(oh->ot_tx, dnode_size);
-                       break;
-
-               default:
-                       LBUG();
-                       break;
+       case DFT_DIR:
+               dt->do_index_ops = &osd_dir_ops;
+               fallthrough;
+       case DFT_INDEX:
+               /* for zap create */
+               dmu_tx_hold_zap(oh->ot_tx, DMU_NEW_OBJECT, FALSE, NULL);
+               dmu_tx_hold_sa_create(oh->ot_tx, dnode_size);
+               break;
+       case DFT_REGULAR:
+       case DFT_SYM:
+       case DFT_NODE:
+               /* first, we'll create new object */
+               dmu_tx_hold_sa_create(oh->ot_tx, dnode_size);
+               break;
+
+       default:
+               LBUG();
+               break;
        }
 
        /* and we'll add it to some mapping */
@@ -1764,7 +1766,8 @@ int __osd_object_create(const struct lu_env *env, struct osd_device *osd,
        int size;
 
        /* Use DMU_OTN_UINT8_METADATA for local objects so their data blocks
-        * would get an additional ditto copy */
+        * would get an additional ditto copy
+        */
        if (unlikely(S_ISREG(la->la_mode) &&
                     fid_seq_is_local_file(fid_seq(fid))))
                type = DMU_OTN_UINT8_METADATA;
@@ -1793,15 +1796,15 @@ int __osd_object_create(const struct lu_env *env, struct osd_device *osd,
  * This is fine for directories today, because storing the FID in the dirent
  * will also require a FAT ZAP.  If there is a new type of micro ZAP created
  * then we might need to re-evaluate the use of this flag and instead do
- * a conversion from the different internal ZAP hash formats being used. */
+ * a conversion from the different internal ZAP hash formats being used.
+ */
 int __osd_zap_create(const struct lu_env *env, struct osd_device *osd,
                     dnode_t **dnp, dmu_tx_t *tx, struct lu_attr *la,
-                    unsigned dnsize, zap_flags_t flags)
+                    unsigned int dnsize, zap_flags_t flags)
 {
        uint64_t oid;
 
-       /* Assert that the transaction has been assigned to a
-          transaction group. */
+       /* Assert that the transaction has been assigned to a transaction grp */
        LASSERT(tx->tx_txg != 0);
        *dnp = NULL;
 
@@ -1827,7 +1830,8 @@ static dnode_t *osd_mkidx(const struct lu_env *env, struct osd_object *obj,
        /* Index file should be created as regular file in order not to confuse
         * ZPL which could interpret them as directory.
         * We set ZAP_FLAG_UINT64_KEY to let ZFS know than we are going to use
-        * binary keys */
+        * binary keys
+        */
        LASSERT(S_ISREG(la->la_mode));
        rc = __osd_zap_create(env, osd, &dn, oh->ot_tx, la,
                osd_find_dnsize(osd, obj->oo_ea_in_bonus), ZAP_FLAG_UINT64_KEY);
@@ -1868,7 +1872,8 @@ static dnode_t *osd_mkreg(const struct lu_env *env, struct osd_object *obj,
                /* The minimum block size must be at least page size otherwise
                 * it will break the assumption in tgt_thread_big_cache where
                 * the array size is PTLRPC_MAX_BRW_PAGES. It will also affect
-                * RDMA due to subpage transfer size */
+                * RDMA due to subpage transfer size
+                */
                rc = -dmu_object_set_blocksize(osd->od_os, dn->dn_object,
                                               PAGE_SIZE, 0, oh->ot_tx);
                if (unlikely(rc)) {
@@ -1878,7 +1883,8 @@ static dnode_t *osd_mkreg(const struct lu_env *env, struct osd_object *obj,
                }
        } else if ((fid_is_llog(fid))) {
                rc = -dmu_object_set_blocksize(osd->od_os, dn->dn_object,
-                                              LLOG_MIN_CHUNK_SIZE, 0, oh->ot_tx);
+                                              LLOG_MIN_CHUNK_SIZE, 0,
+                                              oh->ot_tx);
                if (unlikely(rc)) {
                        CERROR("%s: can't change blocksize: %d\n",
                               osd->od_svname, rc);
@@ -1974,12 +1980,12 @@ static int osd_create(const struct lu_env *env, struct dt_object *dt,
        __u32 compat = 0;
 
        ENTRY;
-
        LASSERT(!fid_is_acct(fid));
 
        /* concurrent create declarations should not see
         * the object inconsistent (db, attr, etc).
-        * in regular cases acquisition should be cheap */
+        * in regular cases acquisition should be cheap
+        */
        down_write(&obj->oo_guard);
 
        if (unlikely(dt_object_exists(dt)))
@@ -1993,8 +1999,9 @@ static int osd_create(const struct lu_env *env, struct dt_object *dt,
 
        LASSERT(obj->oo_dn == NULL);
 
-       /* to follow ZFS on-disk format we need
-        * to initialize parent dnode properly */
+       /* to follow ZFS on-disk format we need to initialize parent dnode
+        * properly
+        */
        if (hint != NULL && hint->dah_parent != NULL &&
            !dt_object_remote(hint->dah_parent))
                parent = osd_dt_obj(hint->dah_parent)->oo_dn->dn_object;
@@ -2107,7 +2114,6 @@ static int osd_ref_add(const struct lu_env *env, struct dt_object *dt,
        int rc;
 
        ENTRY;
-
        down_read(&obj->oo_guard);
        if (unlikely(!dt_object_exists(dt) || obj->oo_destroyed))
                GOTO(out, rc = -ENOENT);
@@ -2148,7 +2154,6 @@ static int osd_ref_del(const struct lu_env *env, struct dt_object *dt,
        int                      rc;
 
        ENTRY;
-
        down_read(&obj->oo_guard);
 
        if (unlikely(!dt_object_exists(dt) || obj->oo_destroyed))
@@ -2176,8 +2181,8 @@ static int osd_object_sync(const struct lu_env *env, struct dt_object *dt,
 {
        struct osd_device *osd = osd_obj2dev(osd_dt_obj(dt));
        uint64_t txg = 0;
-       ENTRY;
 
+       ENTRY;
        if (osd->od_dt_dev.dd_rdonly)
                RETURN(0);