Whamcloud - gitweb
smash the HEAD with the contents of b_cmd. HEAD_PRE_CMD_SMASH and
[fs/lustre-release.git] / lustre / mds / mds_open.c
index 8b8bae0..1a95d04 100644 (file)
@@ -369,11 +369,12 @@ static int mds_create_objects(struct ptlrpc_request *req, int offset,
                 GOTO(out_ids, rc = -ENOMEM);
         oa->o_mode = S_IFREG | 0600;
         oa->o_id = inode->i_ino;
+        oa->o_gr = FILTER_GROUP_FIRST_MDS + mds->mds_num;
         oa->o_generation = inode->i_generation;
         oa->o_uid = 0; /* must have 0 uid / gid on OST */
         oa->o_gid = 0;
         oa->o_valid = OBD_MD_FLID | OBD_MD_FLGENER | OBD_MD_FLTYPE |
-                OBD_MD_FLMODE | OBD_MD_FLUID | OBD_MD_FLGID;
+                OBD_MD_FLMODE | OBD_MD_FLUID | OBD_MD_FLGID | OBD_MD_FLGROUP;
         oa->o_size = 0;
 
         obdo_from_inode(oa, inode, OBD_MD_FLTYPE|OBD_MD_FLATIME|OBD_MD_FLMTIME|
@@ -388,6 +389,7 @@ static int mds_create_objects(struct ptlrpc_request *req, int offset,
                         if (rc)
                                 GOTO(out_oa, rc);
                 }
+                LASSERT(oa->o_gr >= FILTER_GROUP_FIRST_MDS);
                 rc = obd_create(mds->mds_osc_exp, oa, &lsm, &oti);
                 if (rc) {
                         int level = D_ERROR;
@@ -410,6 +412,7 @@ static int mds_create_objects(struct ptlrpc_request *req, int offset,
                         GOTO(out_oa, rc);
                 }
                 lsm->lsm_object_id = oa->o_id;
+                lsm->lsm_object_gr = oa->o_gr;
         }
         if (inode->i_size) {
                 oa->o_size = inode->i_size;
@@ -509,8 +512,8 @@ static void reconstruct_open(struct mds_update_record *rec, int offset,
 
         /* get lock (write for O_CREAT, read otherwise) */
 
-        mds_pack_inode2fid(&body->fid1, child->d_inode);
-        mds_pack_inode2body(body, child->d_inode);
+        mds_pack_inode2fid(obd, &body->fid1, child->d_inode);
+        mds_pack_inode2body(obd, body, child->d_inode);
         if (S_ISREG(child->d_inode->i_mode)) {
                 rc = mds_pack_md(obd, req->rq_repmsg, 2, body,
                                  child->d_inode, 1);
@@ -604,13 +607,15 @@ static int mds_finish_open(struct ptlrpc_request *req, struct dentry *dchild,
         struct obd_device *obd = req->rq_export->exp_obd;
         struct mds_file_data *mfd = NULL;
         obd_id *ids = NULL; /* object IDs created */
+        unsigned mode;
         int rc = 0;
         ENTRY;
 
         /* atomically create objects if necessary */
         down(&dchild->d_inode->i_sem);
-        if (S_ISREG(dchild->d_inode->i_mode) &&
-            !(body->valid & OBD_MD_FLEASIZE)) {
+        mode = dchild->d_inode->i_mode;
+        if ((S_ISREG(mode) || S_ISDIR(mode)) &&
+                        !(body->valid & OBD_MD_FLEASIZE)) {
                 rc = mds_pack_md(obd, req->rq_repmsg, 2, body,
                                  dchild->d_inode, 0);
                 if (rc) {
@@ -677,8 +682,8 @@ static int mds_open_by_fid(struct ptlrpc_request *req, struct ll_fid *fid,
         if (dchild->d_inode != NULL) {
                 up(&pending_dir->i_sem);
                 mds_inode_set_orphan(dchild->d_inode);
-                mds_pack_inode2fid(&body->fid1, dchild->d_inode);
-                mds_pack_inode2body(body, dchild->d_inode);
+                mds_pack_inode2fid(req2obd(req), &body->fid1, dchild->d_inode);
+                mds_pack_inode2body(req2obd(req), body, dchild->d_inode);
                 intent_set_disposition(rep, DISP_LOOKUP_EXECD);
                 intent_set_disposition(rep, DISP_LOOKUP_POS);
                 CWARN("Orphan %s found and opened in PENDING directory\n",
@@ -694,8 +699,8 @@ static int mds_open_by_fid(struct ptlrpc_request *req, struct ll_fid *fid,
         if (IS_ERR(dchild))
                 RETURN(PTR_ERR(dchild));
 
-        mds_pack_inode2fid(&body->fid1, dchild->d_inode);
-        mds_pack_inode2body(body, dchild->d_inode);
+        mds_pack_inode2fid(req2obd(req), &body->fid1, dchild->d_inode);
+        mds_pack_inode2body(req2obd(req), body, dchild->d_inode);
         intent_set_disposition(rep, DISP_LOOKUP_EXECD);
         intent_set_disposition(rep, DISP_LOOKUP_POS);
 
@@ -768,13 +773,18 @@ int mds_open(struct mds_update_record *rec, int offset,
         struct mds_body *body = NULL;
         struct dentry *dchild = NULL, *dparent = NULL;
         struct mds_export_data *med;
-        struct lustre_handle parent_lockh;
+        struct lustre_handle parent_lockh[2];
         int rc = 0, cleanup_phase = 0, acc_mode, created = 0;
         int parent_mode = LCK_PR;
         void *handle = NULL;
         struct dentry_params dp;
+        struct mea *mea = NULL;
+        int mea_size;
         ENTRY;
 
+        parent_lockh[0].cookie = 0;
+        parent_lockh[1].cookie = 0;
+
         if (offset == 2) { /* intent */
                 rep = lustre_msg_buf(req->rq_repmsg, 0, sizeof (*rep));
                 body = lustre_msg_buf(req->rq_repmsg, 1, sizeof (*body));
@@ -818,11 +828,32 @@ int mds_open(struct mds_update_record *rec, int offset,
         acc_mode = accmode(rec->ur_flags);
 
         /* Step 1: Find and lock the parent */
-        if (rec->ur_flags & O_CREAT)
+        if (rec->ur_flags & O_CREAT) {
+                /* XXX Well, in fact we only need this lock mode change if
+                   in addition to O_CREAT, the file does not exist.
+                   But we do not know if it exists or not yet */
                 parent_mode = LCK_PW;
+        }
+        
+        if (rec->ur_namelen == 1) {
+                /* client (LMV) wants to open the file by fid */
+                CDEBUG(D_OTHER, "OPEN by fid %u/%u/%u\n",
+                                (unsigned) rec->ur_fid1->mds,
+                                (unsigned) rec->ur_fid1->id,
+                                (unsigned) rec->ur_fid1->generation);
+                dchild = mds_fid2dentry(mds, rec->ur_fid1, NULL);
+                if (IS_ERR(dchild)) {
+                        rc = PTR_ERR(dparent);
+                        CERROR("child lookup by a fid error %d\n", rc);
+                        GOTO(cleanup, rc);
+                }
+                goto got_child;
+        }
+        
         dparent = mds_fid2locked_dentry(obd, rec->ur_fid1, NULL, parent_mode,
-                                        &parent_lockh, rec->ur_name,
-                                        rec->ur_namelen - 1);
+                                        parent_lockh, rec->ur_name,
+                                        rec->ur_namelen - 1,
+                                        MDS_INODELOCK_UPDATE);
         if (IS_ERR(dparent)) {
                 rc = PTR_ERR(dparent);
                 CERROR("parent lookup error %d\n", rc);
@@ -832,6 +863,21 @@ int mds_open(struct mds_update_record *rec, int offset,
 
         cleanup_phase = 1; /* parent dentry and lock */
 
+        /* try to retrieve MEA data for this dir */
+        rc = mds_get_lmv_attr(obd, dparent->d_inode, &mea, &mea_size);
+       
+        if (mea != NULL) {
+                /* dir is already splitted, check is requested filename
+                 * should live at this MDS or at another one */
+                int i;
+                i = mea_name2idx(mea, rec->ur_name, rec->ur_namelen - 1);
+                if (mea->mea_master != i) {
+                        CERROR("inapropriate MDS(%d) for %s. should be %d\n",
+                                mea->mea_master, rec->ur_name, i);
+                        GOTO(cleanup, rc = -ESTALE);
+                }
+        }
+
         /* Step 2: Lookup the child */
         dchild = ll_lookup_one_len(rec->ur_name, dparent, rec->ur_namelen - 1);
         if (IS_ERR(dchild)) {
@@ -840,9 +886,45 @@ int mds_open(struct mds_update_record *rec, int offset,
                 GOTO(cleanup, rc);
         }
 
+got_child:
         cleanup_phase = 2; /* child dentry */
 
         intent_set_disposition(rep, DISP_LOOKUP_EXECD);
+
+        if (dchild->d_flags & DCACHE_CROSS_REF) {
+                struct ldlm_res_id res_id = { . name = {0} };
+                ldlm_policy_data_t policy;
+                int flags = 0;
+                CDEBUG(D_OTHER, "cross reference: %lu/%lu/%lu\n",
+                       (unsigned long) dchild->d_mdsnum,
+                       (unsigned long) dchild->d_inum,
+                       (unsigned long) dchild->d_generation);
+                body->valid |= OBD_MD_FLID | OBD_MD_MDS;
+                body->fid1.id = dchild->d_inum;
+                body->fid1.mds = dchild->d_mdsnum;
+                body->fid1.generation = dchild->d_generation;
+                intent_set_disposition(rep, DISP_LOOKUP_POS);
+                res_id.name[0] = dchild->d_inum;
+                res_id.name[1] = dchild->d_generation;
+                policy.l_inodebits.bits = MDS_INODELOCK_LOOKUP;
+                rc = ldlm_cli_enqueue(NULL, NULL, obd->obd_namespace,
+                                      res_id, LDLM_IBITS, &policy,
+                                      LCK_PR, &flags,
+                                      mds_blocking_ast,
+                                      ldlm_completion_ast, NULL, NULL,
+                                      NULL, 0, NULL, child_lockh);
+#ifdef S_PDIROPS
+                if (parent_lockh[1].cookie != 0)
+                        ldlm_lock_decref(parent_lockh + 1, LCK_CW);
+#endif
+                ldlm_lock_decref(parent_lockh, parent_mode);
+                if (mea)
+                        OBD_FREE(mea, mea_size);
+                l_dput(dchild);
+                l_dput(dparent);
+                RETURN(rc);
+        }
+        
         if (dchild->d_inode)
                 intent_set_disposition(rep, DISP_LOOKUP_POS);
         else
@@ -854,6 +936,16 @@ int mds_open(struct mds_update_record *rec, int offset,
                 struct iattr iattr;
                 struct inode *inode;
 
+                if ((rc = mds_try_to_split_dir(obd, dparent, &mea, 0))) {
+                        if (rc > 0) {
+                                /* dir got splitted */
+                                GOTO(cleanup, rc = -ESTALE);
+                        } else {
+                                /* error happened during spitting */
+                                GOTO(cleanup, rc);
+                        }
+                }
+
                 if (!(rec->ur_flags & MDS_OPEN_CREAT)) {
                         /* It's negative and we weren't supposed to create it */
                         GOTO(cleanup, rc = -ENOENT);
@@ -917,8 +1009,8 @@ int mds_open(struct mds_update_record *rec, int offset,
 
         LASSERT(!mds_inode_is_orphan(dchild->d_inode));
 
-        mds_pack_inode2fid(&body->fid1, dchild->d_inode);
-        mds_pack_inode2body(body, dchild->d_inode);
+        mds_pack_inode2fid(obd, &body->fid1, dchild->d_inode);
+        mds_pack_inode2body(obd, body, dchild->d_inode);
 
         if (S_ISREG(dchild->d_inode->i_mode)) {
                 /* Check permissions etc */
@@ -963,6 +1055,24 @@ int mds_open(struct mds_update_record *rec, int offset,
                 }
         }
 
+        if (rc == 0) {
+                struct ldlm_res_id res_id = { . name = {0} };
+                ldlm_policy_data_t policy;
+                int flags = 0;
+                res_id.name[0] = dchild->d_inode->i_ino;
+                res_id.name[1] = dchild->d_inode->i_generation;
+                policy.l_inodebits.bits = MDS_INODELOCK_LOOKUP |
+                                                MDS_INODELOCK_UPDATE;
+                rc = ldlm_cli_enqueue(NULL, NULL, obd->obd_namespace,
+                                      res_id, LDLM_IBITS, &policy,
+                                      LCK_PR, &flags,
+                                      mds_blocking_ast,
+                                      ldlm_completion_ast, NULL, NULL,
+                                      NULL, 0, NULL, child_lockh);
+                if (rc == 0)
+                        cleanup_phase = 3;
+        }
+
         /* Step 5: mds_open it */
         rc = mds_finish_open(req, dchild, body, rec->ur_flags, &handle, rec,
                              rep);
@@ -972,7 +1082,11 @@ int mds_open(struct mds_update_record *rec, int offset,
         rc = mds_finish_transno(mds, dchild ? dchild->d_inode : NULL, handle,
                                 req, rc, rep ? rep->lock_policy_res1 : 0);
 
+
         switch (cleanup_phase) {
+        case 3:
+                if (rc)
+                        ldlm_lock_decref(child_lockh, LCK_PR);
         case 2:
                 if (rc && created) {
                         int err = vfs_unlink(dparent->d_inode, dchild);
@@ -990,13 +1104,21 @@ int mds_open(struct mds_update_record *rec, int offset,
                         break;
 
                 l_dput(dparent);
+#ifdef S_PDIROPS
+                if (parent_lockh[1].cookie != 0)
+                        ldlm_lock_decref(parent_lockh + 1, LCK_CW);
+#endif
                 if (rc)
-                        ldlm_lock_decref(&parent_lockh, parent_mode);
+                        ldlm_lock_decref(parent_lockh, parent_mode);
                 else
-                        ptlrpc_save_lock (req, &parent_lockh, parent_mode);
+                        ptlrpc_save_lock (req, parent_lockh, parent_mode);
         }
         if (rc == 0)
                 atomic_inc(&mds->mds_open_count);
+        if (mea)
+                OBD_FREE(mea, mea_size);
+        if ((cleanup_phase != 3) && !rc)
+                rc = ENOLCK;
         RETURN(rc);
 }
 
@@ -1189,8 +1311,8 @@ int mds_close(struct ptlrpc_request *req)
                 body = lustre_msg_buf(req->rq_repmsg, 0, sizeof (*body));
                 LASSERT(body != NULL);
 
-                mds_pack_inode2fid(&body->fid1, inode);
-                mds_pack_inode2body(body, inode);
+                mds_pack_inode2fid(obd, &body->fid1, inode);
+                mds_pack_inode2body(obd, body, inode);
                 mds_pack_md(obd, req->rq_repmsg, 1, body, inode, 1);
         }
         spin_lock(&med->med_open_lock);