Whamcloud - gitweb
Branch HEAD
[fs/lustre-release.git] / lustre / obdecho / echo.c
index eb31dac..76288a1 100644 (file)
 /* -*- mode: c; c-basic-offset: 8; indent-tabs-mode: nil; -*-
  * vim:expandtab:shiftwidth=8:tabstop=8:
  *
- *  linux/fs/ext2_obd/ext2_obd.c
+ *  Copyright (c) 2001-2003 Cluster File Systems, Inc.
+ *   Author: Peter Braam <braam@clusterfs.com>
+ *   Author: Andreas Dilger <adilger@clusterfs.com>
  *
- * Copyright (C) 2001  Cluster File Systems, Inc.
+ *   This file is part of the Lustre file system, http://www.lustre.org
+ *   Lustre is a trademark of Cluster File Systems, Inc.
  *
- * This code is issued under the GNU General Public License.
- * See the file COPYING in this distribution
+ *   You may have signed or agreed to another license before downloading
+ *   this software.  If so, you are bound by the terms and conditions
+ *   of that agreement, and the following does not apply to you.  See the
+ *   LICENSE file included with this distribution for more information.
  *
- * by Peter Braam <braam@clusterfs.com>
+ *   If you did not agree to a different license, then this copy of Lustre
+ *   is open source software; you can redistribute it and/or modify it
+ *   under the terms of version 2 of the GNU General Public License as
+ *   published by the Free Software Foundation.
+ *
+ *   In either case, Lustre is distributed in the hope that it will be
+ *   useful, but WITHOUT ANY WARRANTY; without even the implied warranty
+ *   of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ *   license text for more details.
  */
 
-static char rcsid[] __attribute ((unused)) = "$Id: echo.c,v 1.24 2002/08/19 23:45:00 adilger Exp $";
-#define OBDECHO_VERSION "$Revision: 1.24 $"
-
-#define EXPORT_SYMTAB
-
-#include <linux/version.h>
-#include <linux/module.h>
-#include <linux/fs.h>
-#include <linux/stat.h>
-#include <linux/locks.h>
-#include <linux/ext2_fs.h>
-#include <linux/quotaops.h>
-#include <linux/proc_fs.h>
-#include <linux/init.h>
-#include <asm/unistd.h>
+#ifndef EXPORT_SYMTAB
+# define EXPORT_SYMTAB
+#endif
 
 #define DEBUG_SUBSYSTEM S_ECHO
 
-#include <linux/obd_support.h>
-#include <linux/obd_class.h>
-#include <linux/obd_echo.h>
-#include <linux/lustre_debug.h>
-#include <linux/lustre_dlm.h>
+#include <obd_support.h>
+#include <obd_class.h>
+#include <obd_echo.h>
+#include <lustre_debug.h>
+#include <lustre_dlm.h>
+#include <lprocfs_status.h>
 
-extern struct obd_device obd_dev[MAX_OBD_DEVICES];
-static struct obdo OA;
-static obd_count GEN;
-static long echo_pages = 0;
+#define ECHO_INIT_OBJID      0x1000000000000000ULL
+#define ECHO_HANDLE_MAGIC    0xabcd0123fedc9876ULL
 
-static atomic_t echo_page_rws;
-static atomic_t echo_getattrs;
+#define ECHO_PERSISTENT_PAGES (ECHO_PERSISTENT_SIZE >> CFS_PAGE_SHIFT)
+static cfs_page_t *echo_persistent_pages[ECHO_PERSISTENT_PAGES];
 
-#define ECHO_PROC_STAT "sys/obdecho"
+enum {
+        LPROC_ECHO_READ_BYTES = 1,
+        LPROC_ECHO_WRITE_BYTES = 2,
+        LPROC_ECHO_LAST = LPROC_ECHO_WRITE_BYTES +1
+};
 
-int
-echo_proc_read (char *page, char **start, off_t off, int count, int *eof, void *data)
+static int echo_connect(const struct lu_env *env,
+                        struct lustre_handle *conn, struct obd_device *obd,
+                        struct obd_uuid *cluuid, struct obd_connect_data *data,
+                        void *localdata)
 {
-       int                len;
-        int                attrs = atomic_read (&echo_getattrs);
-        int                pages = atomic_read (&echo_page_rws);
+        data->ocd_connect_flags &= ECHO_CONNECT_SUPPORTED;
+        return class_connect(conn, obd, cluuid);
+}
 
-       *eof = 1;
-       if (off != 0)
-               return (0);
+static int echo_disconnect(struct obd_export *exp)
+{
+        LASSERT (exp != NULL);
+
+        ldlm_cancel_locks_for_export(exp);
+
+        /* complete all outstanding replies */
+        spin_lock(&exp->exp_lock);
+        while (!list_empty(&exp->exp_outstanding_replies)) {
+                struct ptlrpc_reply_state *rs =
+                        list_entry(exp->exp_outstanding_replies.next,
+                                   struct ptlrpc_reply_state, rs_exp_list);
+                struct ptlrpc_service *svc = rs->rs_service;
+
+                spin_lock(&svc->srv_lock);
+                list_del_init(&rs->rs_exp_list);
+                ptlrpc_schedule_difficult_reply(rs);
+                spin_unlock(&svc->srv_lock);
+        }
+        spin_unlock(&exp->exp_lock);
 
-       len = sprintf (page, "%d %d\n", pages, attrs);
+        return class_disconnect(exp);
+}
 
-       *start = page;
-       return (len);
+static int echo_destroy_export(struct obd_export *exp)
+{
+        ENTRY;
+
+        target_destroy_export(exp);
+
+        RETURN(0);
 }
 
-int
-echo_proc_write (struct file *file, const char *ubuffer, unsigned long count, void *data)
+ static __u64 echo_next_id(struct obd_device *obddev)
 {
-       /* Ignore what we've been asked to write, and just zero the stats counters */
-        atomic_set (&echo_page_rws, 0);
-        atomic_set (&echo_getattrs, 0);
+        obd_id id;
+
+        spin_lock(&obddev->u.echo.eo_lock);
+        id = ++obddev->u.echo.eo_lastino;
+        spin_unlock(&obddev->u.echo.eo_lock);
 
-       return (count);
+        return id;
 }
 
-void
-echo_proc_init(void)
+int echo_create(struct obd_export *exp, struct obdo *oa,
+                struct lov_stripe_md **ea, struct obd_trans_info *oti)
 {
-        struct proc_dir_entry *entry;
+        struct obd_device *obd = class_exp2obd(exp);
+
+        if (!obd) {
+                CERROR("invalid client cookie "LPX64"\n",
+                       exp->exp_handle.h_cookie);
+                return -EINVAL;
+        }
 
-        entry = create_proc_entry(ECHO_PROC_STAT, S_IFREG|S_IRUGO|S_IWUSR,NULL);
+        if (!(oa->o_mode && S_IFMT)) {
+                CERROR("echo obd: no type!\n");
+                return -ENOENT;
+        }
 
-        if (entry == NULL) {
-                CERROR("couldn't create proc entry %s\n", ECHO_PROC_STAT);
-                return;
+        if (!(oa->o_valid & OBD_MD_FLTYPE)) {
+                CERROR("invalid o_valid "LPX64"\n", oa->o_valid);
+                return -EINVAL;
         }
 
-        entry->data = NULL;
-        entry->read_proc = echo_proc_read;
-       entry->write_proc = echo_proc_write;
+        oa->o_id = echo_next_id(obd);
+        oa->o_valid = OBD_MD_FLID;
+
+        return 0;
 }
 
-void echo_proc_fini(void)
+int echo_destroy(struct obd_export *exp, struct obdo *oa,
+                 struct lov_stripe_md *ea, struct obd_trans_info *oti,
+                 struct obd_export *md_exp)
 {
-        remove_proc_entry(ECHO_PROC_STAT, 0);
+        struct obd_device *obd = class_exp2obd(exp);
+
+        ENTRY;
+        if (!obd) {
+                CERROR("invalid client cookie "LPX64"\n",
+                       exp->exp_handle.h_cookie);
+                RETURN(-EINVAL);
+        }
+
+        if (!(oa->o_valid & OBD_MD_FLID)) {
+                CERROR("obdo missing FLID valid flag: "LPX64"\n", oa->o_valid);
+                RETURN(-EINVAL);
+        }
+
+        if (oa->o_id > obd->u.echo.eo_lastino || oa->o_id < ECHO_INIT_OBJID) {
+                CERROR("bad destroy objid: "LPX64"\n", oa->o_id);
+                RETURN(-EINVAL);
+        }
+
+        RETURN(0);
 }
 
-static int echo_connect(struct lustre_handle *conn, struct obd_device *obd,
-                        char *cluuid)
+static int echo_getattr(struct obd_export *exp, struct obd_info *oinfo)
 {
-        int rc;
+        struct obd_device *obd = class_exp2obd(exp);
+        obd_id id = oinfo->oi_oa->o_id;
 
-        MOD_INC_USE_COUNT;
-        rc = class_connect(conn, obd, NULL);
+        ENTRY;
+        if (!obd) {
+                CERROR("invalid client cookie "LPX64"\n",
+                       exp->exp_handle.h_cookie);
+                RETURN(-EINVAL);
+        }
 
-        if (rc)
-                MOD_DEC_USE_COUNT;
+        if (!(oinfo->oi_oa->o_valid & OBD_MD_FLID)) {
+                CERROR("obdo missing FLID valid flag: "LPX64"\n",
+                       oinfo->oi_oa->o_valid);
+                RETURN(-EINVAL);
+        }
 
-        return rc;
+        obdo_cpy_md(oinfo->oi_oa, &obd->u.echo.eo_oa, oinfo->oi_oa->o_valid);
+        oinfo->oi_oa->o_id = id;
+
+        RETURN(0);
 }
 
-static int echo_disconnect(struct lustre_handle *conn)
+static int echo_setattr(struct obd_export *exp, struct obd_info *oinfo,
+                        struct obd_trans_info *oti)
 {
-        int rc;
+        struct obd_device *obd = class_exp2obd(exp);
 
-        rc = class_disconnect(conn);
-        if (!rc)
-                MOD_DEC_USE_COUNT;
+        ENTRY;
+        if (!obd) {
+                CERROR("invalid client cookie "LPX64"\n",
+                       exp->exp_handle.h_cookie);
+                RETURN(-EINVAL);
+        }
 
-        return rc;
+        if (!(oinfo->oi_oa->o_valid & OBD_MD_FLID)) {
+                CERROR("obdo missing FLID valid flag: "LPX64"\n",
+                       oinfo->oi_oa->o_valid);
+                RETURN(-EINVAL);
+        }
+
+        memcpy(&obd->u.echo.eo_oa, oinfo->oi_oa, sizeof(*oinfo->oi_oa));
+
+        if (oinfo->oi_oa->o_id & 4) {
+                /* Save lock to force ACKed reply */
+                ldlm_lock_addref (&obd->u.echo.eo_nl_lock, LCK_NL);
+                oti->oti_ack_locks[0].mode = LCK_NL;
+                oti->oti_ack_locks[0].lock = obd->u.echo.eo_nl_lock;
+        }
+
+        RETURN(0);
 }
 
-static int echo_getattr(struct lustre_handle *conn, struct obdo *oa,
-                        struct lov_stripe_md *md)
+static void
+echo_page_debug_setup(cfs_page_t *page, int rw, obd_id id,
+                      __u64 offset, int len)
 {
-        memcpy(oa, &OA, sizeof(*oa));
-        oa->o_mode = ++GEN;
+        int   page_offset = offset & ~CFS_PAGE_MASK;
+        char *addr        = ((char *)cfs_kmap(page)) + page_offset;
+
+        if (len % OBD_ECHO_BLOCK_SIZE != 0)
+                CERROR("Unexpected block size %d\n", len);
+
+        while (len > 0) {
+                if (rw & OBD_BRW_READ)
+                        block_debug_setup(addr, OBD_ECHO_BLOCK_SIZE,
+                                          offset, id);
+                else
+                        block_debug_setup(addr, OBD_ECHO_BLOCK_SIZE,
+                                          0xecc0ecc0ecc0ecc0ULL,
+                                          0xecc0ecc0ecc0ecc0ULL);
+
+                addr   += OBD_ECHO_BLOCK_SIZE;
+                offset += OBD_ECHO_BLOCK_SIZE;
+                len    -= OBD_ECHO_BLOCK_SIZE;
+        }
 
-        atomic_inc (&echo_getattrs);
+        cfs_kunmap(page);
+}
 
-        return 0;
+static int
+echo_page_debug_check(cfs_page_t *page, obd_id id,
+                      __u64 offset, int len)
+{
+        int   page_offset = offset & ~CFS_PAGE_MASK;
+        char *addr        = ((char *)cfs_kmap(page)) + page_offset;
+        int   rc          = 0;
+        int   rc2;
+
+        if (len % OBD_ECHO_BLOCK_SIZE != 0)
+                CERROR("Unexpected block size %d\n", len);
+
+        while (len > 0) {
+                rc2 = block_debug_check("echo", addr, OBD_ECHO_BLOCK_SIZE,
+                                        offset, id);
+
+                if (rc2 != 0 && rc == 0)
+                        rc = rc2;
+
+                addr   += OBD_ECHO_BLOCK_SIZE;
+                offset += OBD_ECHO_BLOCK_SIZE;
+                len    -= OBD_ECHO_BLOCK_SIZE;
+        }
+
+        cfs_kunmap(page);
+
+        return (rc);
 }
 
+/* This allows us to verify that desc_private is passed unmolested */
 #define DESC_PRIV 0x10293847
 
-int echo_preprw(int cmd, struct lustre_handle *conn, int objcount,
-                struct obd_ioobj *obj, int niocount, struct niobuf_remote *nb,
-                struct niobuf_local *res, void **desc_private)
+int echo_preprw(int cmd, struct obd_export *export, struct obdo *oa,
+                int objcount, struct obd_ioobj *obj, int niocount,
+                struct niobuf_remote *nb, struct niobuf_local *res,
+                struct obd_trans_info *oti, struct lustre_capa *unused)
 {
+        struct obd_device *obd;
         struct niobuf_local *r = res;
+        int tot_bytes = 0;
         int rc = 0;
         int i;
-
         ENTRY;
 
+        obd = export->exp_obd;
+        if (obd == NULL)
+                RETURN(-EINVAL);
+
+        /* Temp fix to stop falling foul of osc_announce_cached() */
+        oa->o_valid &= ~(OBD_MD_FLBLOCKS | OBD_MD_FLGRANT);
+
         memset(res, 0, sizeof(*res) * niocount);
 
         CDEBUG(D_PAGE, "%s %d obdos with %d IOs\n",
                cmd == OBD_BRW_READ ? "reading" : "writing", objcount, niocount);
 
-        *desc_private = (void *)DESC_PRIV;
+        if (oti)
+                oti->oti_handle = (void *)DESC_PRIV;
 
         for (i = 0; i < objcount; i++, obj++) {
+                int gfp_mask = (obj->ioo_id & 1) ? CFS_ALLOC_HIGHUSER : CFS_ALLOC_STD;
+                int ispersistent = obj->ioo_id == ECHO_PERSISTENT_OBJID;
+                int debug_setup = (!ispersistent &&
+                                   (oa->o_valid & OBD_MD_FLFLAGS) != 0 &&
+                                   (oa->o_flags & OBD_FL_DEBUG_CHECK) != 0);
                 int j;
 
                 for (j = 0 ; j < obj->ioo_bufcnt ; j++, nb++, r++) {
-                        r->page = alloc_pages(GFP_KERNEL, 0);
-                        if (!r->page) {
-                                CERROR("can't get page %d/%d for id "LPU64"\n",
-                                       j, obj->ioo_bufcnt, obj->ioo_id);
-                                GOTO(preprw_cleanup, rc = -ENOMEM);
+
+                        if (ispersistent &&
+                            (nb->offset >> CFS_PAGE_SHIFT) < ECHO_PERSISTENT_PAGES) {
+                                r->page = echo_persistent_pages[nb->offset >>
+                                                                CFS_PAGE_SHIFT];
+                                /* Take extra ref so __free_pages() can be called OK */
+                                cfs_get_page (r->page);
+                        } else {
+                                OBD_PAGE_ALLOC(r->page, gfp_mask);
+                                if (r->page == NULL) {
+                                        CERROR("can't get page %u/%u for id "
+                                               LPU64"\n",
+                                               j, obj->ioo_bufcnt, obj->ioo_id);
+                                        GOTO(preprw_cleanup, rc = -ENOMEM);
+                                }
                         }
-                        echo_pages++;
+
+                        tot_bytes += nb->len;
+
+                        atomic_inc(&obd->u.echo.eo_prep);
 
                         r->offset = nb->offset;
-                        r->addr = kmap(r->page);
                         r->len = nb->len;
+                        LASSERT((r->offset & ~CFS_PAGE_MASK) + r->len <= CFS_PAGE_SIZE);
+
+                        CDEBUG(D_PAGE, "$$$$ get page %p @ "LPU64" for %d\n",
+                               r->page, r->offset, r->len);
 
-                        CDEBUG(D_PAGE, "$$$$ get page %p, addr %p@"LPU64"\n",
-                               r->page, r->addr, r->offset);
                         if (cmd & OBD_BRW_READ)
-                                page_debug_setup(r->addr, r->len, r->offset,
-                                                 obj->ioo_id);
+                                r->rc = r->len;
+
+                        if (debug_setup)
+                                echo_page_debug_setup(r->page, cmd, obj->ioo_id,
+                                                      r->offset, r->len);
                 }
         }
-        CDEBUG(D_PAGE, "%ld pages allocated after prep\n", echo_pages);
+        if (cmd & OBD_BRW_READ)
+                lprocfs_counter_add(obd->obd_stats, LPROC_ECHO_READ_BYTES,
+                                    tot_bytes);
+        else
+                lprocfs_counter_add(obd->obd_stats, LPROC_ECHO_WRITE_BYTES,
+                                    tot_bytes);
+
+        CDEBUG(D_PAGE, "%d pages allocated after prep\n",
+               atomic_read(&obd->u.echo.eo_prep));
 
         RETURN(0);
 
@@ -184,131 +360,263 @@ preprw_cleanup:
          */
         CERROR("cleaning up %ld pages (%d obdos)\n", (long)(r - res), objcount);
         while (r-- > res) {
-                kunmap(r->page);
-                __free_pages(r->page, 0);
-                echo_pages--;
+                cfs_kunmap(r->page);
+                /* NB if this is a persistent page, __free_pages will just
+                 * lose the extra ref gained above */
+                OBD_PAGE_FREE(r->page);
+                atomic_dec(&obd->u.echo.eo_prep);
         }
         memset(res, 0, sizeof(*res) * niocount);
 
         return rc;
 }
 
-int echo_commitrw(int cmd, struct lustre_handle *conn, int objcount,
-                  struct obd_ioobj *obj, int niocount, struct niobuf_local *res,
-                  void *desc_private)
+int echo_commitrw(int cmd, struct obd_export *export, struct obdo *oa,
+                  int objcount, struct obd_ioobj *obj, int niocount,
+                  struct niobuf_local *res, struct obd_trans_info *oti, int rc)
 {
+        struct obd_device *obd;
         struct niobuf_local *r = res;
-        int rc = 0;
-        int i;
+        int i, vrc = 0;
         ENTRY;
 
-        CDEBUG(D_PAGE, "%s %d obdos with %d IOs\n",
-               cmd == OBD_BRW_READ ? "reading" : "writing", objcount, niocount);
+        obd = export->exp_obd;
+        if (obd == NULL)
+                RETURN(-EINVAL);
+
+        if (rc)
+                GOTO(commitrw_cleanup, rc);
+
+        if ((cmd & OBD_BRW_RWMASK) == OBD_BRW_READ) {
+                CDEBUG(D_PAGE, "reading %d obdos with %d IOs\n",
+                       objcount, niocount);
+        } else {
+                CDEBUG(D_PAGE, "writing %d obdos with %d IOs\n",
+                       objcount, niocount);
+        }
 
         if (niocount && !r) {
                 CERROR("NULL res niobuf with niocount %d\n", niocount);
                 RETURN(-EINVAL);
         }
 
-        LASSERT(desc_private == (void *)DESC_PRIV);
+        LASSERT(oti == NULL || oti->oti_handle == (void *)DESC_PRIV);
 
         for (i = 0; i < objcount; i++, obj++) {
+                int verify = (rc == 0 &&
+                              obj->ioo_id != ECHO_PERSISTENT_OBJID &&
+                              (oa->o_valid & OBD_MD_FLFLAGS) != 0 &&
+                              (oa->o_flags & OBD_FL_DEBUG_CHECK) != 0);
                 int j;
 
                 for (j = 0 ; j < obj->ioo_bufcnt ; j++, r++) {
-                        struct page *page = r->page;
+                        cfs_page_t *page = r->page;
                         void *addr;
 
-                        if (!page || !(addr = page_address(page)) ||
-                            !kern_addr_valid(addr)) {
-
-                                CERROR("bad page "LPU64":%p, buf %d/%d\n",
+                        if (page == NULL) {
+                                CERROR("null page objid "LPU64":%p, buf %d/%d\n",
                                        obj->ioo_id, page, j, obj->ioo_bufcnt);
                                 GOTO(commitrw_cleanup, rc = -EFAULT);
                         }
 
-                        atomic_inc (&echo_page_rws);
+                        addr = cfs_kmap(page);
 
                         CDEBUG(D_PAGE, "$$$$ use page %p, addr %p@"LPU64"\n",
                                r->page, addr, r->offset);
-                        if (cmd & OBD_BRW_WRITE)
-                                page_debug_check("echo", addr, r->len,
-                                                 r->offset, obj->ioo_id);
 
-                        kunmap(page);
-                        __free_pages(page, 0);
-                        echo_pages--;
+                        if (verify) {
+                                vrc = echo_page_debug_check(page, obj->ioo_id,
+                                                            r->offset, r->len);
+                                /* check all the pages always */
+                                if (vrc != 0 && rc == 0)
+                                        rc = vrc;
+                        }
+
+                        cfs_kunmap(page);
+                        /* NB see comment above regarding persistent pages */
+                        OBD_PAGE_FREE(page);
+                        atomic_dec(&obd->u.echo.eo_prep);
                 }
         }
-        CDEBUG(D_PAGE, "%ld pages remain after commit\n", echo_pages);
-        RETURN(0);
+        CDEBUG(D_PAGE, "%d pages remain after commit\n",
+               atomic_read(&obd->u.echo.eo_prep));
+        RETURN(rc);
 
 commitrw_cleanup:
         CERROR("cleaning up %ld pages (%d obdos)\n",
                niocount - (long)(r - res) - 1, objcount);
         while (++r < res + niocount) {
-                struct page *page = r->page;
+                cfs_page_t *page = r->page;
 
-                kunmap(page);
-                __free_pages(page, 0);
-                echo_pages--;
+                /* NB see comment above regarding persistent pages */
+                OBD_PAGE_FREE(page);
+                atomic_dec(&obd->u.echo.eo_prep);
         }
         return rc;
 }
 
-static int echo_setup(struct obd_device *obddev, obd_count len, void *buf)
+static int echo_setup(struct obd_device *obd, struct lustre_cfg *lcfg)
 {
+        struct lprocfs_static_vars lvars;
+        int                        rc;
+        int                        lock_flags = 0;
+        struct ldlm_res_id         res_id = {.name = {1}};
         ENTRY;
 
-        obddev->obd_namespace =
-                ldlm_namespace_new("echo-tgt", LDLM_NAMESPACE_SERVER);
-        if (obddev->obd_namespace == NULL) {
+        spin_lock_init(&obd->u.echo.eo_lock);
+        obd->u.echo.eo_lastino = ECHO_INIT_OBJID;
+
+        obd->obd_namespace = ldlm_namespace_new(obd, "echo-tgt",
+                                                LDLM_NAMESPACE_SERVER,
+                                                LDLM_NAMESPACE_MODEST);
+        if (obd->obd_namespace == NULL) {
                 LBUG();
                 RETURN(-ENOMEM);
         }
 
+        rc = ldlm_cli_enqueue_local(obd->obd_namespace, &res_id, LDLM_PLAIN,
+                                    NULL, LCK_NL, &lock_flags, NULL,
+                                    ldlm_completion_ast, NULL, NULL,
+                                    0, NULL, &obd->u.echo.eo_nl_lock);
+        LASSERT (rc == ELDLM_OK);
+
+        lprocfs_echo_init_vars(&lvars);
+        if (lprocfs_obd_setup(obd, lvars.obd_vars) == 0 &&
+            lprocfs_alloc_obd_stats(obd, LPROC_ECHO_LAST) == 0) {
+                lprocfs_counter_init(obd->obd_stats, LPROC_ECHO_READ_BYTES,
+                                     LPROCFS_CNTR_AVGMINMAX,
+                                     "read_bytes", "bytes");
+                lprocfs_counter_init(obd->obd_stats, LPROC_ECHO_WRITE_BYTES,
+                                     LPROCFS_CNTR_AVGMINMAX,
+                                     "write_bytes", "bytes");
+        }
+
+        ptlrpc_init_client (LDLM_CB_REQUEST_PORTAL, LDLM_CB_REPLY_PORTAL,
+                            "echo_ldlm_cb_client", &obd->obd_ldlm_client);
         RETURN(0);
 }
 
-static int echo_cleanup(struct obd_device *obddev)
+static int echo_cleanup(struct obd_device *obd)
 {
+        int leaked;
         ENTRY;
 
-        ldlm_namespace_free(obddev->obd_namespace);
+        lprocfs_obd_cleanup(obd);
+        lprocfs_free_obd_stats(obd);
+
+        ldlm_lock_decref (&obd->u.echo.eo_nl_lock, LCK_NL);
+
+        /* XXX Bug 3413; wait for a bit to ensure the BL callback has
+         * happened before calling ldlm_namespace_free() */
+        set_current_state (TASK_UNINTERRUPTIBLE);
+        cfs_schedule_timeout (CFS_TASK_UNINT, cfs_time_seconds(1));
+
+        ldlm_namespace_free(obd->obd_namespace, NULL, obd->obd_force);
+        obd->obd_namespace = NULL;
+
+        leaked = atomic_read(&obd->u.echo.eo_prep);
+        if (leaked != 0)
+                CERROR("%d prep/commitrw pages leaked\n", leaked);
 
         RETURN(0);
 }
 
-struct obd_ops echo_obd_ops = {
-        o_connect:     echo_connect,
-        o_disconnect:  echo_disconnect,
-        o_getattr:     echo_getattr,
-        o_preprw:      echo_preprw,
-        o_commitrw:    echo_commitrw,
-        o_setup:       echo_setup,
-        o_cleanup:     echo_cleanup
+static struct obd_ops echo_obd_ops = {
+        .o_owner           = THIS_MODULE,
+        .o_connect         = echo_connect,
+        .o_disconnect      = echo_disconnect,
+        .o_destroy_export  = echo_destroy_export,
+        .o_create          = echo_create,
+        .o_destroy         = echo_destroy,
+        .o_getattr         = echo_getattr,
+        .o_setattr         = echo_setattr,
+        .o_preprw          = echo_preprw,
+        .o_commitrw        = echo_commitrw,
+        .o_setup           = echo_setup,
+        .o_cleanup         = echo_cleanup
 };
 
-static int __init obdecho_init(void)
+extern int echo_client_init(void);
+extern void echo_client_exit(void);
+
+static void
+echo_persistent_pages_fini (void)
 {
-        printk(KERN_INFO "Echo OBD driver " OBDECHO_VERSION " info@clusterfs.com\n");
+        int     i;
+
+        for (i = 0; i < ECHO_PERSISTENT_PAGES; i++)
+                if (echo_persistent_pages[i] != NULL) {
+                        OBD_PAGE_FREE(echo_persistent_pages[i]);
+                        echo_persistent_pages[i] = NULL;
+                }
+}
+
+static int
+echo_persistent_pages_init (void)
+{
+        cfs_page_t *pg;
+        int          i;
+
+        for (i = 0; i < ECHO_PERSISTENT_PAGES; i++) {
+                int gfp_mask = (i < ECHO_PERSISTENT_PAGES/2) ?
+                        CFS_ALLOC_STD : CFS_ALLOC_HIGHUSER;
+
+                OBD_PAGE_ALLOC(pg, gfp_mask);
+                if (pg == NULL) {
+                        echo_persistent_pages_fini ();
+                        return (-ENOMEM);
+                }
 
-        echo_proc_init();
+                memset (cfs_kmap (pg), 0, CFS_PAGE_SIZE);
+                cfs_kunmap (pg);
 
-        return class_register_type(&echo_obd_ops, OBD_ECHO_DEVICENAME);
+                echo_persistent_pages[i] = pg;
+        }
+
+        return (0);
 }
 
-static void __exit obdecho_exit(void)
+static int __init obdecho_init(void)
 {
-        echo_proc_fini ();
+        struct lprocfs_static_vars lvars;
+        int rc;
+
+        ENTRY;
+        printk(KERN_INFO "Lustre: Echo OBD driver; info@clusterfs.com\n");
+
+        LASSERT(CFS_PAGE_SIZE % OBD_ECHO_BLOCK_SIZE == 0);
 
-        CERROR("%ld prep/commitrw pages leaked\n", echo_pages);
-        class_unregister_type(OBD_ECHO_DEVICENAME);
+        lprocfs_echo_init_vars(&lvars);
+
+        rc = echo_persistent_pages_init ();
+        if (rc != 0)
+                goto failed_0;
+
+        rc = class_register_type(&echo_obd_ops, NULL, lvars.module_vars,
+                                 LUSTRE_ECHO_NAME, NULL);
+        if (rc != 0)
+                goto failed_1;
+
+        rc = echo_client_init();
+        if (rc == 0)
+                RETURN (0);
+
+        class_unregister_type(LUSTRE_ECHO_NAME);
+ failed_1:
+        echo_persistent_pages_fini ();
+ failed_0:
+        RETURN(rc);
+}
+
+static void /*__exit*/ obdecho_exit(void)
+{
+        echo_client_exit();
+        class_unregister_type(LUSTRE_ECHO_NAME);
+        echo_persistent_pages_fini ();
 }
 
-MODULE_AUTHOR("Cluster Filesystems Inc. <info@clusterfs.com>");
-MODULE_DESCRIPTION("Lustre Testing Echo OBD driver " OBDECHO_VERSION);
+MODULE_AUTHOR("Cluster File Systems, Inc. <info@clusterfs.com>");
+MODULE_DESCRIPTION("Lustre Testing Echo OBD driver");
 MODULE_LICENSE("GPL");
 
-module_init(obdecho_init);
-module_exit(obdecho_exit);
+cfs_module(obdecho, "1.0.0", obdecho_init, obdecho_exit);