Whamcloud - gitweb
landing b_cmobd_merge on HEAD
[fs/lustre-release.git] / lustre / mds / mds_open.c
index 2546cc6..47471f2 100644 (file)
@@ -304,13 +304,15 @@ static void mds_objids_from_lmm(obd_id *ids, struct lov_mds_md *lmm,
 static int mds_create_objects(struct ptlrpc_request *req, int offset,
                               struct mds_update_record *rec,
                               struct mds_obd *mds, struct obd_device *obd,
-                              struct inode *inode, void **handle, obd_id **ids)
+                              struct dentry *dchild, void **handle, 
+                              obd_id **ids)
 {
         struct obdo *oa;
         struct obd_trans_info oti = { 0 };
         struct mds_body *body;
         struct lov_stripe_md *lsm = NULL;
         struct lov_mds_md *lmm = NULL;
+        struct inode *inode = dchild->d_inode;
         void *lmm_buf;
         int rc, lmm_bufsize, lmm_size;
         ENTRY;
@@ -360,7 +362,6 @@ static int mds_create_objects(struct ptlrpc_request *req, int offset,
                 RETURN(0);
         }
 
-        
         if (OBD_FAIL_CHECK_ONCE(OBD_FAIL_MDS_ALLOC_OBDO))
                 GOTO(out_ids, rc = -ENOMEM);
 
@@ -381,9 +382,9 @@ static int mds_create_objects(struct ptlrpc_request *req, int offset,
                         OBD_MD_FLCTIME);
 
         if (!(rec->ur_flags & MDS_OPEN_HAS_OBJS)) {
-                /* check if things like lstripe/lfs stripe are sending us the ea */
+                /* check if things like lfs setstripe are sending us the ea */
                 if (rec->ur_flags & MDS_OPEN_HAS_EA) {
-                        rc = obd_iocontrol(OBD_IOC_LOV_SETSTRIPE, 
+                        rc = obd_iocontrol(OBD_IOC_LOV_SETSTRIPE,
                                            mds->mds_osc_exp,
                                            0, &lsm, rec->ur_eadata);
                         if (rc)
@@ -561,7 +562,7 @@ static void reconstruct_open(struct mds_update_record *rec, int offset,
                 mfd = NULL;
         }
 
-#warning "XXX fixme"
+        /* #warning "XXX fixme" bug 2991 */
         /* Here it used to LASSERT(mfd) if exp_outstanding_reply != NULL.
          * Now that exp_outstanding_reply is a list, it's just using mfd != NULL
          * to detect a re-open */
@@ -575,6 +576,10 @@ static void reconstruct_open(struct mds_update_record *rec, int offset,
                         GOTO(out_dput, req->rq_status = -ENOMEM);
                 }
                 put_child = 0;
+        } else {
+                body->handle.cookie = mfd->mfd_handle.h_cookie;
+                CDEBUG(D_INODE, "resend mfd %p, cookie "LPX64"\n", mfd,
+                       mfd->mfd_handle.h_cookie);
         }
 
  out_dput:
@@ -614,8 +619,8 @@ static int mds_finish_open(struct ptlrpc_request *req, struct dentry *dchild,
         /* atomically create objects if necessary */
         down(&dchild->d_inode->i_sem);
         mode = dchild->d_inode->i_mode;
-        if ((S_ISREG(mode) || S_ISDIR(mode)) &&
-                        !(body->valid & OBD_MD_FLEASIZE)) {
+        if ((S_ISREG(mode) && (body->valid & OBD_MD_FLEASIZE)) || 
+            (S_ISDIR(mode) && (body->valid & OBD_MD_FLDIREA))) {
                 rc = mds_pack_md(obd, req->rq_repmsg, 2, body,
                                  dchild->d_inode, 0);
                 if (rc) {
@@ -626,7 +631,7 @@ static int mds_finish_open(struct ptlrpc_request *req, struct dentry *dchild,
         if (rec != NULL) {
                 /* no EA: create objects */
                 rc = mds_create_objects(req, 2, rec, mds, obd,
-                                        dchild->d_inode, handle, &ids);
+                                        dchild, handle, &ids);
                 if (rc) {
                         CERROR("mds_create_objects: rc = %d\n", rc);
                         up(&dchild->d_inode->i_sem);
@@ -718,7 +723,7 @@ int mds_pin(struct ptlrpc_request *req)
 {
         struct obd_device *obd = req->rq_export->exp_obd;
         struct mds_body *request_body, *reply_body;
-        struct obd_run_ctxt saved;
+        struct lvfs_run_ctxt saved;
         int rc, size = sizeof(*reply_body);
         ENTRY;
 
@@ -729,10 +734,10 @@ int mds_pin(struct ptlrpc_request *req)
                 RETURN(rc);
         reply_body = lustre_msg_buf(req->rq_repmsg, 0, sizeof(*reply_body));
 
-        push_ctxt(&saved, &obd->obd_ctxt, NULL);
+        push_ctxt(&saved, &obd->obd_lvfs_ctxt, NULL);
         rc = mds_open_by_fid(req, &request_body->fid1, reply_body,
                              request_body->flags, NULL, NULL);
-        pop_ctxt(&saved, &obd->obd_ctxt, NULL);
+        pop_ctxt(&saved, &obd->obd_lvfs_ctxt, NULL);
 
         RETURN(rc);
 }
@@ -865,6 +870,8 @@ int mds_open(struct mds_update_record *rec, int offset,
 
         /* try to retrieve MEA data for this dir */
         rc = mds_get_lmv_attr(obd, dparent->d_inode, &mea, &mea_size);
+        if (rc)
+                GOTO(cleanup, rc);
        
         if (mea != NULL) {
                 /* dir is already splitted, check is requested filename
@@ -981,15 +988,15 @@ got_child:
                 }
 
                 created = 1;
-                LTIME_S(iattr.ia_atime) = rec->ur_time;
-                LTIME_S(iattr.ia_ctime) = rec->ur_time;
-                LTIME_S(iattr.ia_mtime) = rec->ur_time;
+                LTIME_S(iattr.ia_atime) = LTIME_S(rec->ur_time);
+                LTIME_S(iattr.ia_ctime) = LTIME_S(rec->ur_time);
+                LTIME_S(iattr.ia_mtime) = LTIME_S(rec->ur_time);
 
-                iattr.ia_uid = rec->ur_fsuid;
+                iattr.ia_uid = rec->_ur_fsuid;
                 if (dparent->d_inode->i_mode & S_ISGID)
                         iattr.ia_gid = dparent->d_inode->i_gid;
                 else
-                        iattr.ia_gid = rec->ur_fsgid;
+                        iattr.ia_gid = rec->_ur_fsgid;
 
                 iattr.ia_valid = ATTR_UID | ATTR_GID | ATTR_ATIME |
                         ATTR_MTIME | ATTR_CTIME;
@@ -1004,7 +1011,7 @@ got_child:
                 if (rc)
                         CERROR("error on parent setattr: rc = %d\n", rc);
 
-                acc_mode = 0;                  /* Don't check for permissions */
+                acc_mode = 0;           /* Don't check for permissions */
         }
 
         LASSERT(!mds_inode_is_orphan(dchild->d_inode));
@@ -1044,9 +1051,10 @@ got_child:
         if ((rec->ur_flags & MDS_OPEN_DIRECTORY) &&
             !S_ISDIR(dchild->d_inode->i_mode))
                 GOTO(cleanup, rc = -ENOTDIR);
-        if (S_ISDIR(dchild->d_inode->i_mode)) { 
-                if (rec->ur_flags & MDS_OPEN_CREAT || rec->ur_flags & FMODE_WRITE) {
+
+        if (S_ISDIR(dchild->d_inode->i_mode)) {
+                if (rec->ur_flags & MDS_OPEN_CREAT ||
+                    rec->ur_flags & FMODE_WRITE) {
                         /*we are tryying to create or write a exist dir*/
                         GOTO(cleanup, rc = -EISDIR);
                 }
@@ -1082,7 +1090,6 @@ got_child:
         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)
@@ -1143,15 +1150,15 @@ int mds_mfd_close(struct ptlrpc_request *req, struct obd_device *obd,
         void *handle = NULL;
         struct mds_body *request_body = NULL, *reply_body = NULL;
         struct dentry_params dp;
-        struct lov_mds_md *lmm;
+        struct iattr iattr = { 0 };
         ENTRY;
 
-        if (req != NULL) {
+        if (req && req->rq_reqmsg != NULL)
                 request_body = lustre_msg_buf(req->rq_reqmsg, 0,
                                               sizeof(*request_body));
+        if (req && req->rq_repmsg != NULL)
                 reply_body = lustre_msg_buf(req->rq_repmsg, 0,
                                             sizeof(*reply_body));
-        }
 
         fidlen = ll_fid2str(fidname, inode->i_ino, inode->i_generation);
 
@@ -1167,6 +1174,8 @@ int mds_mfd_close(struct ptlrpc_request *req, struct obd_device *obd,
         }
 
         if (last_orphan && unlink_orphan) {
+                struct lov_mds_md *lmm = NULL;
+                int stripe_count = 0;
                 LASSERT(rc == 0); /* mds_put_write_access must have succeeded */
 
                 CDEBUG(D_HA, "destroying orphan object %s\n", fidname);
@@ -1184,20 +1193,22 @@ int mds_mfd_close(struct ptlrpc_request *req, struct obd_device *obd,
                 LASSERT(pending_child->d_inode != NULL);
 
                 cleanup_phase = 2; /* dput(pending_child) when finished */
-                lmm = lustre_msg_buf(req->rq_repmsg, 1, 0);
-                handle = fsfilt_start_log(obd, pending_dir,
-                                          FSFILT_OP_UNLINK, NULL,
-                                          le32_to_cpu(lmm->lmm_stripe_count));
+                if (req != NULL && req->rq_repmsg != NULL) {
+                        lmm = lustre_msg_buf(req->rq_repmsg, 1, 0);
+                        stripe_count = le32_to_cpu(lmm->lmm_stripe_count);
+                }
+
+                handle = fsfilt_start_log(obd, pending_dir, FSFILT_OP_UNLINK,
+                                          NULL, stripe_count);
                 if (IS_ERR(handle)) {
                         rc = PTR_ERR(handle);
                         handle = NULL;
                         GOTO(cleanup, rc);
                 }
 
-                if (req != NULL &&
+                if (req != NULL && req->rq_repmsg != NULL &&
                     (reply_body->valid & OBD_MD_FLEASIZE) &&
-                    mds_log_op_unlink(obd, pending_child->d_inode,
-                                      lustre_msg_buf(req->rq_repmsg, 1, 0),
+                    mds_log_op_unlink(obd, pending_child->d_inode, lmm,
                                       req->rq_repmsg->buflens[1],
                                       lustre_msg_buf(req->rq_repmsg, 2, 0),
                                       req->rq_repmsg->buflens[2]) > 0) {
@@ -1213,32 +1224,65 @@ int mds_mfd_close(struct ptlrpc_request *req, struct obd_device *obd,
                         rc = vfs_unlink(pending_dir, pending_child);
                 if (rc)
                         CERROR("error unlinking orphan %s: rc %d\n",fidname,rc);
-        } else if (mfd->mfd_mode & FMODE_WRITE && rc == 0) {
+
+                goto out; /* Don't bother updating attrs on unlinked inode */
+        }
+
+        if (request_body != NULL && mfd->mfd_mode & FMODE_WRITE && rc == 0) {
                 /* Update the on-disk attributes if this was the last write
                  * close, and all information was provided (i.e., rc == 0)
                  *
                  * XXX this should probably be abstracted with mds_reint_setattr
                  */
+
 #if 0
-                struct iattr iattr;
+                if (request_body->valid & OBD_MD_FLMTIME &&
+                    LTIME_S(request_body->mtime) > LTIME_S(inode->i_mtime)) {
+                        LTIME_S(iattr.ia_mtime) = LTIME_S(request_body->mtime);
+                        iattr.ia_valid |= ATTR_MTIME;
+                }
+                if (request_body->valid & OBD_MD_FLCTIME &&
+                    LTIME_S(request_body->ctime) > LTIME_S(inode->i_ctime)) {
+                        LTIME_S(iattr.ia_ctime) = LTIME_S(request_body->ctime);
+                        iattr.ia_valid |= ATTR_CTIME;
+                }
 
                 /* XXX can't set block count with fsfilt_setattr (!) */
-                iattr.ia_valid = ATTR_CTIME | ATTR_ATIME |
-                        ATTR_MTIME | ATTR_SIZE;
-                iattr.ia_atime = request_body->atime;
-                iattr.ia_ctime = request_body->ctime;
-                iattr.ia_mtime = request_body->mtime;
-                iattr.ia_size = request_body->size;
-                /* iattr.ia_blocks = request_body->blocks */
+                if (request_body->valid & OBD_MD_FLSIZE) {
+                        iattr.ia_valid |= ATTR_SIZE;
+                        iattr.ia_size = request_body->size;
+                }
+                /* if (request_body->valid & OBD_MD_FLBLOCKS) {
+                        iattr.ia_valid |= ATTR_BLOCKS;
+                        iattr.ia_blocks = request_body->blocks
+                } */
+
+#endif
+        }
+
+        if (request_body != NULL && request_body->valid & OBD_MD_FLATIME) {
+                /* Only start a transaction to write out only the atime if
+                 * it is more out-of-date than the specified limit.  If we
+                 * are already going to write out the atime then do it anyway.
+                 * */
+                if ((LTIME_S(request_body->atime) >
+                     LTIME_S(inode->i_atime) + MAX_ATIME_DIFF) ||
+                    (iattr.ia_valid != 0 &&
+                     LTIME_S(request_body->atime) > LTIME_S(inode->i_atime))) {
+                        LTIME_S(iattr.ia_atime) = LTIME_S(request_body->atime);
+                        iattr.ia_valid |= ATTR_ATIME;
+                }
+        }
 
+        if (iattr.ia_valid != 0) {
                 handle = fsfilt_start(obd, inode, FSFILT_OP_SETATTR, NULL);
                 if (IS_ERR(handle))
                         GOTO(cleanup, rc = PTR_ERR(handle));
                 rc = fsfilt_setattr(obd, mfd->mfd_dentry, handle, &iattr, 0);
                 if (rc)
                         CERROR("error in setattr(%s): rc %d\n", fidname, rc);
-#endif
         }
+out:
         /* If other clients have this file open for write, rc will be > 0 */
         if (rc > 0)
                 rc = 0;
@@ -1247,7 +1291,7 @@ int mds_mfd_close(struct ptlrpc_request *req, struct obd_device *obd,
 
  cleanup:
         atomic_dec(&mds->mds_open_count);
-        if (req) {
+        if (req != NULL && reply_body != NULL) {
                 rc = mds_finish_transno(mds, pending_dir, handle, req, rc, 0);
         } else if (handle) {
                 int err = fsfilt_commit(obd, pending_dir, handle, 0);
@@ -1273,7 +1317,7 @@ int mds_close(struct ptlrpc_request *req)
         struct obd_device *obd = req->rq_export->exp_obd;
         struct mds_body *body;
         struct mds_file_data *mfd;
-        struct obd_run_ctxt saved;
+        struct lvfs_run_ctxt saved;
         struct inode *inode;
         int rc, repsize[3] = {sizeof(struct mds_body),
                               obd->u.mds.mds_max_mdsize,
@@ -1319,9 +1363,9 @@ int mds_close(struct ptlrpc_request *req)
         list_del(&mfd->mfd_list);
         spin_unlock(&med->med_open_lock);
 
-        push_ctxt(&saved, &obd->obd_ctxt, NULL);
-        req->rq_status = mds_mfd_close(rc ? NULL : req, obd, mfd, 1);
-        pop_ctxt(&saved, &obd->obd_ctxt, NULL);
+        push_ctxt(&saved, &obd->obd_lvfs_ctxt, NULL);
+        req->rq_status = mds_mfd_close(req, obd, mfd, 1);
+        pop_ctxt(&saved, &obd->obd_lvfs_ctxt, NULL);
 
         if (OBD_FAIL_CHECK(OBD_FAIL_MDS_CLOSE_PACK)) {
                 CERROR("test case OBD_FAIL_MDS_CLOSE_PACK\n");