[fs/lustre-release.git] / lustre / osc / osc_request.c

/* -*- mode: c; c-basic-offset: 8; indent-tabs-mode: nil; -*-
 * vim:expandtab:shiftwidth=8:tabstop=8:
 *
 *  Copyright (C) 2001-2003 Cluster File Systems, Inc.
 *   Author Peter Braam <braam@clusterfs.com>
 *
 *   This file is part of the Lustre file system, http://www.lustre.org
 *   Lustre is a trademark of Cluster File Systems, Inc.
 *
 *   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.
 *
 *   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.
 *
 *  For testing and management it is treated as an obd_device,
 *  although * it does not export a full OBD method table (the
 *  requests are coming * in over the wire, so object target modules
 *  do not have a full * method table.)
 *
 */

#ifndef EXPORT_SYMTAB
# define EXPORT_SYMTAB
#endif
#define DEBUG_SUBSYSTEM S_OSC

#ifdef __KERNEL__
# include <libcfs/libcfs.h>
#else /* __KERNEL__ */
# include <liblustre.h>
#endif

# include <lustre_dlm.h>
#include <libcfs/kp30.h>
#include <lustre_net.h>
#include <lustre/lustre_user.h>
#include <obd_ost.h>
#include <obd_lov.h>

#ifdef  __CYGWIN__
# include <ctype.h>
#endif

#include <lustre_ha.h>
#include <lprocfs_status.h>
#include <lustre_log.h>
#include <lustre_debug.h>
#include <lustre_param.h>
#include "osc_internal.h"

static quota_interface_t *quota_interface = NULL;
extern quota_interface_t osc_quota_interface;

static void osc_release_ppga(struct brw_page **ppga, obd_count count);

static quota_interface_t *quota_interface;
extern quota_interface_t osc_quota_interface;

/* Pack OSC object metadata for disk storage (LE byte order). */
static int osc_packmd(struct obd_export *exp, struct lov_mds_md **lmmp,
                      struct lov_stripe_md *lsm)
{
        int lmm_size;
        ENTRY;

        lmm_size = sizeof(**lmmp);
        if (!lmmp)
                RETURN(lmm_size);

        if (*lmmp && !lsm) {
                OBD_FREE(*lmmp, lmm_size);
                *lmmp = NULL;
                RETURN(0);
        }

        if (!*lmmp) {
                OBD_ALLOC(*lmmp, lmm_size);
                if (!*lmmp)
                        RETURN(-ENOMEM);
        }

        if (lsm) {
                LASSERT(lsm->lsm_object_id);
                (*lmmp)->lmm_object_id = cpu_to_le64(lsm->lsm_object_id);
        }

        RETURN(lmm_size);
}

/* Unpack OSC object metadata from disk storage (LE byte order). */
static int osc_unpackmd(struct obd_export *exp, struct lov_stripe_md **lsmp,
                        struct lov_mds_md *lmm, int lmm_bytes)
{
        int lsm_size;
        ENTRY;

        if (lmm != NULL) {
                if (lmm_bytes < sizeof (*lmm)) {
                        CERROR("lov_mds_md too small: %d, need %d\n",
                               lmm_bytes, (int)sizeof(*lmm));
                        RETURN(-EINVAL);
                }
                /* XXX LOV_MAGIC etc check? */

                if (lmm->lmm_object_id == 0) {
                        CERROR("lov_mds_md: zero lmm_object_id\n");
                        RETURN(-EINVAL);
                }
        }

        lsm_size = lov_stripe_md_size(1);
        if (lsmp == NULL)
                RETURN(lsm_size);

        if (*lsmp != NULL && lmm == NULL) {
                OBD_FREE(*lsmp, lsm_size);
                *lsmp = NULL;
                RETURN(0);
        }

        if (*lsmp == NULL) {
                OBD_ALLOC(*lsmp, lsm_size);
                if (*lsmp == NULL)
                        RETURN(-ENOMEM);
                loi_init((*lsmp)->lsm_oinfo);
        }

        if (lmm != NULL) {
                /* XXX zero *lsmp? */
                (*lsmp)->lsm_object_id = le64_to_cpu (lmm->lmm_object_id);
                LASSERT((*lsmp)->lsm_object_id);
        }

        (*lsmp)->lsm_maxbytes = LUSTRE_STRIPE_MAXBYTES;

        RETURN(lsm_size);
}

static int osc_getattr_interpret(struct ptlrpc_request *req,
                                 struct osc_async_args *aa, int rc)
{
        struct ost_body *body;
        ENTRY;

        if (rc != 0)
                GOTO(out, rc);

        body = lustre_swab_repbuf(req, REPLY_REC_OFF, sizeof(*body),
                                  lustre_swab_ost_body);
        if (body) {
                CDEBUG(D_INODE, "mode: %o\n", body->oa.o_mode);
                memcpy(aa->aa_oi->oi_oa, &body->oa, sizeof(*aa->aa_oi->oi_oa));

                /* This should really be sent by the OST */
                aa->aa_oi->oi_oa->o_blksize = PTLRPC_MAX_BRW_SIZE;
                aa->aa_oi->oi_oa->o_valid |= OBD_MD_FLBLKSZ;
        } else {
                CERROR("can't unpack ost_body\n");
                rc = -EPROTO;
                aa->aa_oi->oi_oa->o_valid = 0;
        }
out:
        rc = aa->aa_oi->oi_cb_up(aa->aa_oi, rc);
        RETURN(rc);
}

static int osc_getattr_async(struct obd_export *exp, struct obd_info *oinfo,
                             struct ptlrpc_request_set *set)
{
        struct ptlrpc_request *req;
        struct ost_body *body;
        int size[2] = { sizeof(struct ptlrpc_body), sizeof(*body) };
        struct osc_async_args *aa;
        ENTRY;

        req = ptlrpc_prep_req(class_exp2cliimp(exp), LUSTRE_OST_VERSION,
                              OST_GETATTR, 2, size,NULL);
        if (!req)
                RETURN(-ENOMEM);

        body = lustre_msg_buf(req->rq_reqmsg, REQ_REC_OFF, sizeof(*body));
        memcpy(&body->oa, oinfo->oi_oa, sizeof(*oinfo->oi_oa));

        ptlrpc_req_set_repsize(req, 2, size);
        req->rq_interpret_reply = osc_getattr_interpret;

        CLASSERT(sizeof(*aa) <= sizeof(req->rq_async_args));
        aa = (struct osc_async_args *)&req->rq_async_args;
        aa->aa_oi = oinfo;

        ptlrpc_set_add_req(set, req);
        RETURN (0);
}

static int osc_getattr(struct obd_export *exp, struct obd_info *oinfo)
{
        struct ptlrpc_request *req;
        struct ost_body *body;
        int rc, size[2] = { sizeof(struct ptlrpc_body), sizeof(*body) };
        ENTRY;

        req = ptlrpc_prep_req(class_exp2cliimp(exp), LUSTRE_OST_VERSION,
                              OST_GETATTR, 2, size, NULL);
        if (!req)
                RETURN(-ENOMEM);

        body = lustre_msg_buf(req->rq_reqmsg, REQ_REC_OFF, sizeof(*body));
        memcpy(&body->oa, oinfo->oi_oa, sizeof(*oinfo->oi_oa));

        ptlrpc_req_set_repsize(req, 2, size);

        rc = ptlrpc_queue_wait(req);
        if (rc) {
                CERROR("%s failed: rc = %d\n", __FUNCTION__, rc);
                GOTO(out, rc);
        }

        body = lustre_swab_repbuf(req, REPLY_REC_OFF, sizeof(*body),
                                  lustre_swab_ost_body);
        if (body == NULL) {
                CERROR ("can't unpack ost_body\n");
                GOTO (out, rc = -EPROTO);
        }

        CDEBUG(D_INODE, "mode: %o\n", body->oa.o_mode);
        memcpy(oinfo->oi_oa, &body->oa, sizeof(*oinfo->oi_oa));

        /* This should really be sent by the OST */
        oinfo->oi_oa->o_blksize = PTLRPC_MAX_BRW_SIZE;
        oinfo->oi_oa->o_valid |= OBD_MD_FLBLKSZ;

        EXIT;
 out:
        ptlrpc_req_finished(req);
        return rc;
}

static int osc_setattr(struct obd_export *exp, struct obd_info *oinfo,
                       struct obd_trans_info *oti)
{
        struct ptlrpc_request *req;
        struct ost_body *body;
        int rc, size[2] = { sizeof(struct ptlrpc_body), sizeof(*body) };
        ENTRY;

        req = ptlrpc_prep_req(class_exp2cliimp(exp), LUSTRE_OST_VERSION,
                              OST_SETATTR, 2, size, NULL);
        if (!req)
                RETURN(-ENOMEM);

        body = lustre_msg_buf(req->rq_reqmsg, REQ_REC_OFF, sizeof(*body));
        memcpy(&body->oa, oinfo->oi_oa, sizeof(*oinfo->oi_oa));

        ptlrpc_req_set_repsize(req, 2, size);

        rc = ptlrpc_queue_wait(req);
        if (rc)
                GOTO(out, rc);

        body = lustre_swab_repbuf(req, REPLY_REC_OFF, sizeof(*body),
                                  lustre_swab_ost_body);
        if (body == NULL)
                GOTO(out, rc = -EPROTO);

        memcpy(oinfo->oi_oa, &body->oa, sizeof(*oinfo->oi_oa));

        EXIT;
out:
        ptlrpc_req_finished(req);
        RETURN(rc);
}

static int osc_setattr_interpret(struct ptlrpc_request *req,
                                 struct osc_async_args *aa, int rc)
{
        struct ost_body *body;
        ENTRY;

        if (rc != 0)
                GOTO(out, rc);

        body = lustre_swab_repbuf(req, REPLY_REC_OFF, sizeof(*body),
                                  lustre_swab_ost_body);
        if (body == NULL) {
                CERROR("can't unpack ost_body\n");
                GOTO(out, rc = -EPROTO);
        }

        memcpy(aa->aa_oi->oi_oa, &body->oa, sizeof(*aa->aa_oi->oi_oa));
out:
        rc = aa->aa_oi->oi_cb_up(aa->aa_oi, rc);
        RETURN(rc);
}

static int osc_setattr_async(struct obd_export *exp, struct obd_info *oinfo,
                             struct obd_trans_info *oti,
                             struct ptlrpc_request_set *rqset)
{
        struct ptlrpc_request *req;
        struct ost_body *body;
        int size[2] = { sizeof(struct ptlrpc_body), sizeof(*body) };
        struct osc_async_args *aa;
        ENTRY;

        req = ptlrpc_prep_req(class_exp2cliimp(exp), LUSTRE_OST_VERSION,
                              OST_SETATTR, 2, size, NULL);
        if (!req)
                RETURN(-ENOMEM);

        body = lustre_msg_buf(req->rq_reqmsg, REQ_REC_OFF, sizeof(*body));

        if (oinfo->oi_oa->o_valid & OBD_MD_FLCOOKIE) {
                LASSERT(oti);
                memcpy(obdo_logcookie(oinfo->oi_oa), oti->oti_logcookies,
                       sizeof(*oti->oti_logcookies));
        }

        memcpy(&body->oa, oinfo->oi_oa, sizeof(*oinfo->oi_oa));
        ptlrpc_req_set_repsize(req, 2, size);
        /* do mds to ost setattr asynchronouly */
        if (!rqset) {
                /* Do not wait for response. */
                ptlrpcd_add_req(req);
        } else {
                req->rq_interpret_reply = osc_setattr_interpret;

                CLASSERT(sizeof(*aa) <= sizeof(req->rq_async_args));
                aa = (struct osc_async_args *)&req->rq_async_args;
                aa->aa_oi = oinfo;

                ptlrpc_set_add_req(rqset, req);
        }

        RETURN(0);
}

int osc_real_create(struct obd_export *exp, struct obdo *oa,
                    struct lov_stripe_md **ea, struct obd_trans_info *oti)
{
        struct ptlrpc_request *req;
        struct ost_body *body;
        struct lov_stripe_md *lsm;
        int rc, size[2] = { sizeof(struct ptlrpc_body), sizeof(*body) };
        ENTRY;

        LASSERT(oa);
        LASSERT(ea);

        lsm = *ea;
        if (!lsm) {
                rc = obd_alloc_memmd(exp, &lsm);
                if (rc < 0)
                        RETURN(rc);
        }

        req = ptlrpc_prep_req(class_exp2cliimp(exp), LUSTRE_OST_VERSION,
                              OST_CREATE, 2, size, NULL);
        if (!req)
                GOTO(out, rc = -ENOMEM);

        body = lustre_msg_buf(req->rq_reqmsg, REQ_REC_OFF, sizeof(*body));
        memcpy(&body->oa, oa, sizeof(body->oa));

        ptlrpc_req_set_repsize(req, 2, size);
        if (oa->o_valid & OBD_MD_FLINLINE) {
                LASSERT((oa->o_valid & OBD_MD_FLFLAGS) &&
                        oa->o_flags == OBD_FL_DELORPHAN);
                DEBUG_REQ(D_HA, req,
                          "delorphan from OST integration");
                /* Don't resend the delorphan req */
                req->rq_no_resend = req->rq_no_delay = 1;
        }

        rc = ptlrpc_queue_wait(req);
        if (rc)
                GOTO(out_req, rc);

        body = lustre_swab_repbuf(req, REPLY_REC_OFF, sizeof(*body),
                                  lustre_swab_ost_body);
        if (body == NULL) {
                CERROR ("can't unpack ost_body\n");
                GOTO (out_req, rc = -EPROTO);
        }

        memcpy(oa, &body->oa, sizeof(*oa));

        /* This should really be sent by the OST */
        oa->o_blksize = PTLRPC_MAX_BRW_SIZE;
        oa->o_valid |= OBD_MD_FLBLKSZ;

        /* XXX LOV STACKING: the lsm that is passed to us from LOV does not
         * have valid lsm_oinfo data structs, so don't go touching that.
         * This needs to be fixed in a big way.
         */
        lsm->lsm_object_id = oa->o_id;
        *ea = lsm;

        if (oti != NULL) {
                oti->oti_transno = lustre_msg_get_transno(req->rq_repmsg);

                if (oa->o_valid & OBD_MD_FLCOOKIE) {
                        if (!oti->oti_logcookies)
                                oti_alloc_cookies(oti, 1);
                        memcpy(oti->oti_logcookies, obdo_logcookie(oa),
                               sizeof(oti->oti_onecookie));
                }
        }

        CDEBUG(D_HA, "transno: "LPD64"\n",
               lustre_msg_get_transno(req->rq_repmsg));
        EXIT;
out_req:
        ptlrpc_req_finished(req);
out:
        if (rc && !*ea)
                obd_free_memmd(exp, &lsm);
        return rc;
}

static int osc_punch_interpret(struct ptlrpc_request *req,
                               struct osc_async_args *aa, int rc)
{
        struct ost_body *body;
        ENTRY;

        if (rc != 0)
                GOTO(out, rc);

        body = lustre_swab_repbuf(req, REPLY_REC_OFF, sizeof (*body),
                                  lustre_swab_ost_body);
        if (body == NULL) {
                CERROR ("can't unpack ost_body\n");
                GOTO(out, rc = -EPROTO);
        }

        memcpy(aa->aa_oi->oi_oa, &body->oa, sizeof(*aa->aa_oi->oi_oa));
out:
        rc = aa->aa_oi->oi_cb_up(aa->aa_oi, rc);
        RETURN(rc);
}

static int osc_punch(struct obd_export *exp, struct obd_info *oinfo,
                     struct obd_trans_info *oti,
                     struct ptlrpc_request_set *rqset)
{
        struct ptlrpc_request *req;
        struct osc_async_args *aa;
        struct ost_body *body;
        int size[2] = { sizeof(struct ptlrpc_body), sizeof(*body) };
        ENTRY;

        if (!oinfo->oi_oa) {
                CERROR("oa NULL\n");
                RETURN(-EINVAL);
        }

        req = ptlrpc_prep_req(class_exp2cliimp(exp), LUSTRE_OST_VERSION,
                              OST_PUNCH, 2, size, NULL);
        if (!req)
                RETURN(-ENOMEM);

        /* FIXME bug 249. Also see bug 7198 */
        if (class_exp2cliimp(exp)->imp_connect_data.ocd_connect_flags &
            OBD_CONNECT_REQPORTAL)
                req->rq_request_portal = OST_IO_PORTAL;

        body = lustre_msg_buf(req->rq_reqmsg, REQ_REC_OFF, sizeof(*body));
        memcpy(&body->oa, oinfo->oi_oa, sizeof(*oinfo->oi_oa));

        /* overload the size and blocks fields in the oa with start/end */
        body->oa.o_size = oinfo->oi_policy.l_extent.start;
        body->oa.o_blocks = oinfo->oi_policy.l_extent.end;
        body->oa.o_valid |= (OBD_MD_FLSIZE | OBD_MD_FLBLOCKS);

        ptlrpc_req_set_repsize(req, 2, size);

        req->rq_interpret_reply = osc_punch_interpret;
        CLASSERT(sizeof(*aa) <= sizeof(req->rq_async_args));
        aa = (struct osc_async_args *)&req->rq_async_args;
        aa->aa_oi = oinfo;
        ptlrpc_set_add_req(rqset, req);

        RETURN(0);
}

static int osc_sync(struct obd_export *exp, struct obdo *oa,
                    struct lov_stripe_md *md, obd_size start, obd_size end)
{
        struct ptlrpc_request *req;
        struct ost_body *body;
        int rc, size[2] = { sizeof(struct ptlrpc_body), sizeof(*body) };
        ENTRY;

        if (!oa) {
                CERROR("oa NULL\n");
                RETURN(-EINVAL);
        }

        req = ptlrpc_prep_req(class_exp2cliimp(exp), LUSTRE_OST_VERSION,
                              OST_SYNC, 2, size, NULL);
        if (!req)
                RETURN(-ENOMEM);

        body = lustre_msg_buf(req->rq_reqmsg, REQ_REC_OFF, sizeof(*body));
        memcpy(&body->oa, oa, sizeof(*oa));

        /* overload the size and blocks fields in the oa with start/end */
        body->oa.o_size = start;
        body->oa.o_blocks = end;
        body->oa.o_valid |= (OBD_MD_FLSIZE | OBD_MD_FLBLOCKS);

        ptlrpc_req_set_repsize(req, 2, size);

        rc = ptlrpc_queue_wait(req);
        if (rc)
                GOTO(out, rc);

        body = lustre_swab_repbuf(req, REPLY_REC_OFF, sizeof(*body),
                                  lustre_swab_ost_body);
        if (body == NULL) {
                CERROR ("can't unpack ost_body\n");
                GOTO (out, rc = -EPROTO);
        }

        memcpy(oa, &body->oa, sizeof(*oa));

        EXIT;
 out:
        ptlrpc_req_finished(req);
        return rc;
}

/* Destroy requests can be async always on the client, and we don't even really
 * care about the return code since the client cannot do anything at all about
 * a destroy failure.
 * When the MDS is unlinking a filename, it saves the file objects into a
 * recovery llog, and these object records are cancelled when the OST reports
 * they were destroyed and sync'd to disk (i.e. transaction committed).
 * If the client dies, or the OST is down when the object should be destroyed,
 * the records are not cancelled, and when the OST reconnects to the MDS next,
 * it will retrieve the llog unlink logs and then sends the log cancellation
 * cookies to the MDS after committing destroy transactions. */
static int osc_destroy(struct obd_export *exp, struct obdo *oa,
                       struct lov_stripe_md *ea, struct obd_trans_info *oti,
                       struct obd_export *md_export)
{
        struct ptlrpc_request *req;
        struct ost_body *body;
        int size[2] = { sizeof(struct ptlrpc_body), sizeof(*body) };
        ENTRY;

        if (!oa) {
                CERROR("oa NULL\n");
                RETURN(-EINVAL);
        }

        req = ptlrpc_prep_req(class_exp2cliimp(exp), LUSTRE_OST_VERSION,
                              OST_DESTROY, 2, size, NULL);
        if (!req)
                RETURN(-ENOMEM);

        /* FIXME bug 249. Also see bug 7198 */
        if (class_exp2cliimp(exp)->imp_connect_data.ocd_connect_flags &
            OBD_CONNECT_REQPORTAL)
                req->rq_request_portal = OST_IO_PORTAL;

        body = lustre_msg_buf(req->rq_reqmsg, REQ_REC_OFF, sizeof(*body));

        if (oti != NULL && oa->o_valid & OBD_MD_FLCOOKIE) {
                memcpy(obdo_logcookie(oa), oti->oti_logcookies,
                       sizeof(*oti->oti_logcookies));
        }

        memcpy(&body->oa, oa, sizeof(*oa));
        ptlrpc_req_set_repsize(req, 2, size);

        ptlrpcd_add_req(req);
        RETURN(0);
}

static void osc_announce_cached(struct client_obd *cli, struct obdo *oa,
                                long writing_bytes)
{
        obd_flag bits = OBD_MD_FLBLOCKS|OBD_MD_FLGRANT;

        LASSERT(!(oa->o_valid & bits));

        oa->o_valid |= bits;
        client_obd_list_lock(&cli->cl_loi_list_lock);
        oa->o_dirty = cli->cl_dirty;
        if (cli->cl_dirty > cli->cl_dirty_max) {
                CERROR("dirty %lu > dirty_max %lu\n",
                       cli->cl_dirty, cli->cl_dirty_max);
                oa->o_undirty = 0;
        } else if (atomic_read(&obd_dirty_pages) > obd_max_dirty_pages) {
                CERROR("dirty %d > system dirty_max %d\n",
                       atomic_read(&obd_dirty_pages), obd_max_dirty_pages);
                oa->o_undirty = 0;
        } else if (cli->cl_dirty_max - cli->cl_dirty > 0x7fffffff) {
                CERROR("dirty %lu - dirty_max %lu too big???\n",
                       cli->cl_dirty, cli->cl_dirty_max);
                oa->o_undirty = 0;
        } else {
                long max_in_flight = (cli->cl_max_pages_per_rpc << CFS_PAGE_SHIFT)*
                                (cli->cl_max_rpcs_in_flight + 1);
                oa->o_undirty = max(cli->cl_dirty_max, max_in_flight);
        }
        oa->o_grant = cli->cl_avail_grant;
        oa->o_dropped = cli->cl_lost_grant;
        cli->cl_lost_grant = 0;
        client_obd_list_unlock(&cli->cl_loi_list_lock);
        CDEBUG(D_CACHE,"dirty: "LPU64" undirty: %u dropped %u grant: "LPU64"\n",
               oa->o_dirty, oa->o_undirty, oa->o_dropped, oa->o_grant);
}

/* caller must hold loi_list_lock */
static void osc_consume_write_grant(struct client_obd *cli,struct brw_page *pga)
{
        atomic_inc(&obd_dirty_pages);
        cli->cl_dirty += CFS_PAGE_SIZE;
        cli->cl_avail_grant -= CFS_PAGE_SIZE;
        pga->flag |= OBD_BRW_FROM_GRANT;
        CDEBUG(D_CACHE, "using %lu grant credits for brw %p page %p\n",
               CFS_PAGE_SIZE, pga, pga->pg);
        LASSERT(cli->cl_avail_grant >= 0);
}

/* the companion to osc_consume_write_grant, called when a brw has completed.
 * must be called with the loi lock held. */
static void osc_release_write_grant(struct client_obd *cli,
                                    struct brw_page *pga, int sent)
{
        int blocksize = cli->cl_import->imp_obd->obd_osfs.os_bsize ? : 4096;
        ENTRY;

        if (!(pga->flag & OBD_BRW_FROM_GRANT)) {
                EXIT;
                return;
        }

        pga->flag &= ~OBD_BRW_FROM_GRANT;
        atomic_dec(&obd_dirty_pages);
        cli->cl_dirty -= CFS_PAGE_SIZE;
        if (!sent) {
                cli->cl_lost_grant += CFS_PAGE_SIZE;
                CDEBUG(D_CACHE, "lost grant: %lu avail grant: %lu dirty: %lu\n",
                       cli->cl_lost_grant, cli->cl_avail_grant, cli->cl_dirty);
        } else if (CFS_PAGE_SIZE != blocksize && pga->count != CFS_PAGE_SIZE) {
                /* For short writes we shouldn't count parts of pages that
                 * span a whole block on the OST side, or our accounting goes
                 * wrong.  Should match the code in filter_grant_check. */
                int offset = pga->off & ~CFS_PAGE_MASK;
                int count = pga->count + (offset & (blocksize - 1));
                int end = (offset + pga->count) & (blocksize - 1);
                if (end)
                        count += blocksize - end;

                cli->cl_lost_grant += CFS_PAGE_SIZE - count;
                CDEBUG(D_CACHE, "lost %lu grant: %lu avail: %lu dirty: %lu\n",
                       CFS_PAGE_SIZE - count, cli->cl_lost_grant,
                       cli->cl_avail_grant, cli->cl_dirty);
        }

        EXIT;
}

static unsigned long rpcs_in_flight(struct client_obd *cli)
{
        return cli->cl_r_in_flight + cli->cl_w_in_flight;
}

/* caller must hold loi_list_lock */
void osc_wake_cache_waiters(struct client_obd *cli)
{
        struct list_head *l, *tmp;
        struct osc_cache_waiter *ocw;

        ENTRY;
        list_for_each_safe(l, tmp, &cli->cl_cache_waiters) {
                /* if we can't dirty more, we must wait until some is written */
                if ((cli->cl_dirty + CFS_PAGE_SIZE > cli->cl_dirty_max) ||
                   ((atomic_read(&obd_dirty_pages)+1)>(obd_max_dirty_pages))) {
                        CDEBUG(D_CACHE, "no dirty room: dirty: %ld "
                               "osc max %ld, sys max %d\n", cli->cl_dirty,
                               cli->cl_dirty_max, obd_max_dirty_pages);
                        return;
                }

                /* if still dirty cache but no grant wait for pending RPCs that
                 * may yet return us some grant before doing sync writes */
                if (cli->cl_w_in_flight && cli->cl_avail_grant < CFS_PAGE_SIZE) {
                        CDEBUG(D_CACHE, "%u BRW writes in flight, no grant\n",
                               cli->cl_w_in_flight);
                        return;
                }

                ocw = list_entry(l, struct osc_cache_waiter, ocw_entry);
                list_del_init(&ocw->ocw_entry);
                if (cli->cl_avail_grant < CFS_PAGE_SIZE) {
                        /* no more RPCs in flight to return grant, do sync IO */
                        ocw->ocw_rc = -EDQUOT;
                        CDEBUG(D_INODE, "wake oap %p for sync\n", ocw->ocw_oap);
                } else {
                        osc_consume_write_grant(cli,
                                                &ocw->ocw_oap->oap_brw_page);
                }

                cfs_waitq_signal(&ocw->ocw_waitq);
        }

        EXIT;
}

static void osc_init_grant(struct client_obd *cli, struct obd_connect_data *ocd)
{
        client_obd_list_lock(&cli->cl_loi_list_lock);
        cli->cl_avail_grant = ocd->ocd_grant;
        client_obd_list_unlock(&cli->cl_loi_list_lock);

        CDEBUG(D_CACHE, "setting cl_avail_grant: %ld cl_lost_grant: %ld\n",
               cli->cl_avail_grant, cli->cl_lost_grant);
        LASSERT(cli->cl_avail_grant >= 0);
}

static void osc_update_grant(struct client_obd *cli, struct ost_body *body)
{
        client_obd_list_lock(&cli->cl_loi_list_lock);
        CDEBUG(D_CACHE, "got "LPU64" extra grant\n", body->oa.o_grant);
        cli->cl_avail_grant += body->oa.o_grant;
        /* waiters are woken in brw_interpret_oap */
        client_obd_list_unlock(&cli->cl_loi_list_lock);
}

/* We assume that the reason this OSC got a short read is because it read
 * beyond the end of a stripe file; i.e. lustre is reading a sparse file
 * via the LOV, and it _knows_ it's reading inside the file, it's just that
 * this stripe never got written at or beyond this stripe offset yet. */
static void handle_short_read(int nob_read, obd_count page_count,
                              struct brw_page **pga)
{
        char *ptr;
        int i = 0;

        /* skip bytes read OK */
        while (nob_read > 0) {
                LASSERT (page_count > 0);

                if (pga[i]->count > nob_read) {
                        /* EOF inside this page */
                        ptr = cfs_kmap(pga[i]->pg) + 
                                (pga[i]->off & ~CFS_PAGE_MASK);
                        memset(ptr + nob_read, 0, pga[i]->count - nob_read);
                        cfs_kunmap(pga[i]->pg);
                        page_count--;
                        i++;
                        break;
                }

                nob_read -= pga[i]->count;
                page_count--;
                i++;
        }

        /* zero remaining pages */
        while (page_count-- > 0) {
                ptr = cfs_kmap(pga[i]->pg) + (pga[i]->off & ~CFS_PAGE_MASK);
                memset(ptr, 0, pga[i]->count);
                cfs_kunmap(pga[i]->pg);
                i++;
        }
}

static int check_write_rcs(struct ptlrpc_request *req,
                           int requested_nob, int niocount,
                           obd_count page_count, struct brw_page **pga)
{
        int    *remote_rcs, i;

        /* return error if any niobuf was in error */
        remote_rcs = lustre_swab_repbuf(req, REQ_REC_OFF + 1,
                                        sizeof(*remote_rcs) * niocount, NULL);
        if (remote_rcs == NULL) {
                CERROR("Missing/short RC vector on BRW_WRITE reply\n");
                return(-EPROTO);
        }
        if (lustre_msg_swabbed(req->rq_repmsg))
                for (i = 0; i < niocount; i++)
                        __swab32s(&remote_rcs[i]);

        for (i = 0; i < niocount; i++) {
                if (remote_rcs[i] < 0)
                        return(remote_rcs[i]);

                if (remote_rcs[i] != 0) {
                        CERROR("rc[%d] invalid (%d) req %p\n",
                                i, remote_rcs[i], req);
                        return(-EPROTO);
                }
        }

        if (req->rq_bulk->bd_nob_transferred != requested_nob) {
                CERROR("Unexpected # bytes transferred: %d (requested %d)\n",
                       requested_nob, req->rq_bulk->bd_nob_transferred);
                return(-EPROTO);
        }

        return (0);
}

static inline int can_merge_pages(struct brw_page *p1, struct brw_page *p2)
{
        if (p1->flag != p2->flag) {
                unsigned mask = ~OBD_BRW_FROM_GRANT;

                /* warn if we try to combine flags that we don't know to be
                 * safe to combine */
                if ((p1->flag & mask) != (p2->flag & mask))
                        CERROR("is it ok to have flags 0x%x and 0x%x in the "
                               "same brw?\n", p1->flag, p2->flag);
                return 0;
        }

        return (p1->off + p1->count == p2->off);
}

static obd_count osc_checksum_bulk(int nob, obd_count pg_count,
                                   struct brw_page **pga)
{
        __u32 cksum = ~0;
        int i = 0;

        LASSERT (pg_count > 0);
        while (nob > 0 && pg_count > 0) {
                char *ptr = cfs_kmap(pga[i]->pg);
                int off = pga[i]->off & ~CFS_PAGE_MASK;
                int count = pga[i]->count > nob ? nob : pga[i]->count;

                /* corrupt the data before we compute the checksum, to
                 * simulate an OST->client data error */
                if (i == 0 &&OBD_FAIL_CHECK_ONCE(OBD_FAIL_OSC_CHECKSUM_RECEIVE))
                        memcpy(ptr + off, "bad1", min(4, nob));
                cksum = crc32_le(cksum, ptr + off, count);
                cfs_kunmap(pga[i]->pg);
                LL_CDEBUG_PAGE(D_PAGE, pga[i]->pg, "off %d checksum %x\n",
                               off, cksum);

                nob -= pga[i]->count;
                pg_count--;
                i++;
        }
        /* For sending we only compute the wrong checksum instead
         * of corrupting the data so it is still correct on a redo */
        if (OBD_FAIL_CHECK_ONCE(OBD_FAIL_OSC_CHECKSUM_SEND))
                cksum++;

        return cksum;
}

static int osc_brw_prep_request(int cmd, struct client_obd *cli,struct obdo *oa,
                                struct lov_stripe_md *lsm, obd_count page_count,
                                struct brw_page **pga,
                                struct ptlrpc_request **reqp)
{
        struct ptlrpc_request   *req;
        struct ptlrpc_bulk_desc *desc;
        struct ost_body         *body;
        struct obd_ioobj        *ioobj;
        struct niobuf_remote    *niobuf;
        int size[4] = { sizeof(struct ptlrpc_body), sizeof(*body) };
        int niocount, i, requested_nob, opc, rc;
        struct ptlrpc_request_pool *pool;
        struct osc_brw_async_args *aa;

        ENTRY;
        opc = ((cmd & OBD_BRW_WRITE) != 0) ? OST_WRITE : OST_READ;
        pool = ((cmd & OBD_BRW_WRITE) != 0) ? cli->cl_import->imp_rq_pool :NULL;

        for (niocount = i = 1; i < page_count; i++) {
                if (!can_merge_pages(pga[i - 1], pga[i]))
                        niocount++;
        }

        size[REQ_REC_OFF + 1] = sizeof(*ioobj);
        size[REQ_REC_OFF + 2] = niocount * sizeof(*niobuf);

        OBD_FAIL_RETURN(OBD_FAIL_OSC_BRW_PREP_REQ, -ENOMEM);
        req = ptlrpc_prep_req_pool(cli->cl_import, LUSTRE_OST_VERSION, opc, 4, size,
                                   NULL, pool);
        if (req == NULL)
                RETURN (-ENOMEM);

        /* FIXME bug 249. Also see bug 7198 */
        if (cli->cl_import->imp_connect_data.ocd_connect_flags &
            OBD_CONNECT_REQPORTAL)
                req->rq_request_portal = OST_IO_PORTAL;

        if (opc == OST_WRITE)
                desc = ptlrpc_prep_bulk_imp (req, page_count,
                                             BULK_GET_SOURCE, OST_BULK_PORTAL);
        else
                desc = ptlrpc_prep_bulk_imp (req, page_count,
                                             BULK_PUT_SINK, OST_BULK_PORTAL);
        if (desc == NULL)
                GOTO(out, rc = -ENOMEM);
        /* NB request now owns desc and will free it when it gets freed */

        body = lustre_msg_buf(req->rq_reqmsg, REQ_REC_OFF, sizeof(*body));
        ioobj = lustre_msg_buf(req->rq_reqmsg, REQ_REC_OFF + 1, sizeof(*ioobj));
        niobuf = lustre_msg_buf(req->rq_reqmsg, REQ_REC_OFF + 2,
                                niocount * sizeof(*niobuf));

        memcpy(&body->oa, oa, sizeof(*oa));

        obdo_to_ioobj(oa, ioobj);
        ioobj->ioo_bufcnt = niocount;

        LASSERT (page_count > 0);
        for (requested_nob = i = 0; i < page_count; i++, niobuf++) {
                struct brw_page *pg = pga[i];
                struct brw_page *pg_prev = pga[i - 1];

                LASSERT(pg->count > 0);
                LASSERTF((pg->off & ~CFS_PAGE_MASK) + pg->count <= CFS_PAGE_SIZE,
                         "i: %d pg: %p off: "LPU64", count: %u\n", i, pg,
                         pg->off, pg->count);
#ifdef __LINUX__
                LASSERTF(i == 0 || pg->off > pg_prev->off,
                         "i %d p_c %u pg %p [pri %lu ind %lu] off "LPU64
                         " prev_pg %p [pri %lu ind %lu] off "LPU64"\n",
                         i, page_count,
                         pg->pg, page_private(pg->pg), pg->pg->index, pg->off,
                         pg_prev->pg, page_private(pg_prev->pg),
                         pg_prev->pg->index, pg_prev->off);
#else
                LASSERTF(i == 0 || pg->off > pg_prev->off,
                         "i %d p_c %u\n", i, page_count);
#endif
                LASSERT((pga[0]->flag & OBD_BRW_SRVLOCK) ==
                        (pg->flag & OBD_BRW_SRVLOCK));

                ptlrpc_prep_bulk_page(desc, pg->pg, pg->off & ~CFS_PAGE_MASK,
                                      pg->count);
                requested_nob += pg->count;

                if (i > 0 && can_merge_pages(pg_prev, pg)) {
                        niobuf--;
                        niobuf->len += pg->count;
                } else {
                        niobuf->offset = pg->off;
                        niobuf->len    = pg->count;
                        niobuf->flags  = pg->flag;
                }
        }

        LASSERT((void *)(niobuf - niocount) ==
                lustre_msg_buf(req->rq_reqmsg, REQ_REC_OFF + 2,
                               niocount * sizeof(*niobuf)));
        osc_announce_cached(cli, &body->oa, opc == OST_WRITE ? requested_nob:0);

        /* size[REQ_REC_OFF] still sizeof (*body) */
        if (opc == OST_WRITE) {
                if (unlikely(cli->cl_checksum)) {
                        body->oa.o_valid |= OBD_MD_FLCKSUM;
                        body->oa.o_cksum = osc_checksum_bulk(requested_nob,
                                                             page_count, pga);
                        CDEBUG(D_PAGE, "checksum at write origin: %x\n",
                               body->oa.o_cksum);
                        /* save this in 'oa', too, for later checking */
                        oa->o_valid |= OBD_MD_FLCKSUM;
                } else {
                        /* clear out the checksum flag, in case this is a
                         * resend but cl_checksum is no longer set. b=11238 */
                        oa->o_valid &= ~OBD_MD_FLCKSUM;
                }
                oa->o_cksum = body->oa.o_cksum;
                /* 1 RC per niobuf */
                size[REPLY_REC_OFF + 1] = sizeof(__u32) * niocount;
                ptlrpc_req_set_repsize(req, 3, size);
        } else {
                if (unlikely(cli->cl_checksum))
                        body->oa.o_valid |= OBD_MD_FLCKSUM;
                /* 1 RC for the whole I/O */
                ptlrpc_req_set_repsize(req, 2, size);
        }

        CLASSERT(sizeof(*aa) <= sizeof(req->rq_async_args));
        aa = (struct osc_brw_async_args *)&req->rq_async_args;
        aa->aa_oa = oa;
        aa->aa_requested_nob = requested_nob;
        aa->aa_nio_count = niocount;
        aa->aa_page_count = page_count;
        aa->aa_retries = 5;     /*retry for checksum errors; lprocfs? */
        aa->aa_ppga = pga;
        aa->aa_cli = cli;
        INIT_LIST_HEAD(&aa->aa_oaps);

        *reqp = req;
        RETURN (0);

 out:
        ptlrpc_req_finished (req);
        RETURN (rc);
}

static int check_write_checksum(struct obdo *oa, const lnet_process_id_t *peer,
                                 __u32 client_cksum, __u32 server_cksum, int nob,
                                 obd_count page_count, struct brw_page **pga)
{
        __u32 new_cksum;
        char *msg;

        if (server_cksum == client_cksum) {
                CDEBUG(D_PAGE, "checksum %x confirmed\n", client_cksum);
                return 0;
        }

        new_cksum = osc_checksum_bulk(nob, page_count, pga);

        if (new_cksum == server_cksum)
                msg = "changed on the client after we checksummed it";
        else if (new_cksum == client_cksum)
                msg = "changed in transit before arrival at OST";
        else
                msg = "changed in transit AND doesn't match the original";

        LCONSOLE_ERROR("BAD WRITE CHECKSUM: %s: from %s inum "LPU64"/"LPU64
                       " object "LPU64"/"LPU64" extent ["LPU64"-"LPU64"]\n",
                       msg, libcfs_nid2str(peer->nid),
                       oa->o_valid & OBD_MD_FLFID ? oa->o_fid : (__u64)0,
                       oa->o_valid & OBD_MD_FLFID ? oa->o_generation : (__u64)0,
                       oa->o_id,
                       oa->o_valid & OBD_MD_FLGROUP ? oa->o_gr : (__u64)0,
                       pga[0]->off,
                       pga[page_count-1]->off + pga[page_count-1]->count - 1);
        CERROR("original client csum %x, server csum %x, client csum now %x\n",
               client_cksum, server_cksum, new_cksum);

        return 1;
}

/* Note rc enters this function as number of bytes transferred */
static int osc_brw_fini_request(struct ptlrpc_request *req, int rc)
{
        struct osc_brw_async_args *aa = (void *)&req->rq_async_args;
        const lnet_process_id_t *peer =
                        &req->rq_import->imp_connection->c_peer;
        struct client_obd *cli = aa->aa_cli;
        struct ost_body *body;
        __u32 client_cksum = 0;
        ENTRY;

        if (rc < 0 && rc != -EDQUOT)
                RETURN(rc);

        LASSERTF(req->rq_repmsg != NULL, "rc = %d\n", rc);
        body = lustre_swab_repbuf(req, REPLY_REC_OFF, sizeof(*body),
                                  lustre_swab_ost_body);
        if (body == NULL) {
                CERROR ("Can't unpack body\n");
                RETURN(-EPROTO);
        }

        /* set/clear over quota flag for a uid/gid */
        if (lustre_msg_get_opc(req->rq_reqmsg) == OST_WRITE &&
            body->oa.o_valid & (OBD_MD_FLUSRQUOTA | OBD_MD_FLGRPQUOTA))
                lquota_setdq(quota_interface, cli, body->oa.o_uid,
                             body->oa.o_gid, body->oa.o_valid,
                             body->oa.o_flags);

        if (rc < 0)
                RETURN(rc);

        if (unlikely(aa->aa_oa->o_valid & OBD_MD_FLCKSUM))
                client_cksum = aa->aa_oa->o_cksum; /* save for later */

        osc_update_grant(cli, body);

        if (lustre_msg_get_opc(req->rq_reqmsg) == OST_WRITE) {
                if (rc > 0) {
                        CERROR ("Unexpected +ve rc %d\n", rc);
                        RETURN(-EPROTO);
                }
                LASSERT(req->rq_bulk->bd_nob == aa->aa_requested_nob);

                if (unlikely((aa->aa_oa->o_valid & OBD_MD_FLCKSUM) &&
                             client_cksum &&
                             check_write_checksum(&body->oa, peer, client_cksum,
                                                 body->oa.o_cksum,
                                                 aa->aa_requested_nob,
                                                 aa->aa_page_count,
                                                 aa->aa_ppga)))
                        RETURN(-EAGAIN);

                rc = check_write_rcs(req, aa->aa_requested_nob,aa->aa_nio_count,
                                     aa->aa_page_count, aa->aa_ppga);
                GOTO(out, rc);
        }

        /* The rest of this function executes only for OST_READs */
        if (rc > aa->aa_requested_nob) {
                CERROR("Unexpected rc %d (%d requested)\n", rc,
                       aa->aa_requested_nob);
                RETURN(-EPROTO);
        }

        if (rc != req->rq_bulk->bd_nob_transferred) {
                CERROR ("Unexpected rc %d (%d transferred)\n",
                        rc, req->rq_bulk->bd_nob_transferred);
                return (-EPROTO);
        }

        if (rc < aa->aa_requested_nob)
                handle_short_read(rc, aa->aa_page_count, aa->aa_ppga);

        if (unlikely(body->oa.o_valid & OBD_MD_FLCKSUM)) {
                static int cksum_counter;
                __u32 server_cksum = body->oa.o_cksum;
                client_cksum = osc_checksum_bulk(rc, aa->aa_page_count,
                                                 aa->aa_ppga);

                if (server_cksum == ~0 && rc > 0) {
                        CERROR("Protocol error: server %s set the 'checksum' "
                               "bit, but didn't send a checksum.  Not fatal, "
                               "but please tell CFS.\n",
                               libcfs_nid2str(peer->nid));
                } else if (server_cksum != client_cksum) {
                        LCONSOLE_ERROR("%s: BAD READ CHECKSUM: from %s inum "
                                       LPU64"/"LPU64" object "LPU64"/"LPU64
                                       " extent ["LPU64"-"LPU64"]\n",
                                       req->rq_import->imp_obd->obd_name,
                                       libcfs_nid2str(peer->nid),
                                       body->oa.o_valid & OBD_MD_FLFID ?
                                                body->oa.o_fid : (__u64)0,
                                       body->oa.o_valid & OBD_MD_FLFID ?
                                                body->oa.o_generation :(__u64)0,
                                       body->oa.o_id,
                                       body->oa.o_valid & OBD_MD_FLGROUP ?
                                                body->oa.o_gr : (__u64)0,
                                       aa->aa_ppga[0]->off,
                                       aa->aa_ppga[aa->aa_page_count-1]->off +
                                       aa->aa_ppga[aa->aa_page_count-1]->count -
                                                                        1);
                        CERROR("client %x, server %x\n",
                               client_cksum, server_cksum);
                        cksum_counter = 0;
                        aa->aa_oa->o_cksum = client_cksum;
                        rc = -EAGAIN;
                } else {
                        cksum_counter++;
                        CDEBUG(D_PAGE, "checksum %x confirmed\n", client_cksum);
                        rc = 0;
                }
        } else if (unlikely(client_cksum)) {
                static int cksum_missed;

                cksum_missed++;
                if ((cksum_missed & (-cksum_missed)) == cksum_missed)
                        CERROR("Checksum %u requested from %s but not sent\n",
                               cksum_missed, libcfs_nid2str(peer->nid));
        } else {
                rc = 0;
        }
out:
        if (rc >= 0)
                memcpy(aa->aa_oa, &body->oa, sizeof(*aa->aa_oa));

        RETURN(rc);
}

static int osc_brw_internal(int cmd, struct obd_export *exp,struct obdo *oa,
                            struct lov_stripe_md *lsm,
                            obd_count page_count, struct brw_page **pga)
{
        struct ptlrpc_request *request;
        int                    rc, retries = 5; /* lprocfs? */
        ENTRY;

restart_bulk:
        rc = osc_brw_prep_request(cmd, &exp->exp_obd->u.cli, oa, lsm,
                                  page_count, pga, &request);
        if (rc != 0)
                return (rc);

        rc = ptlrpc_queue_wait(request);

        if (rc == -ETIMEDOUT && request->rq_resend) {
                DEBUG_REQ(D_HA, request,  "BULK TIMEOUT");
                ptlrpc_req_finished(request);
                goto restart_bulk;
        }

        rc = osc_brw_fini_request(request, rc);

        ptlrpc_req_finished(request);
        if (rc == -EAGAIN) {
                if (retries-- > 0)
                        goto restart_bulk;
                rc = -EIO;
        }
        RETURN(rc);
}

int osc_brw_redo_request(struct ptlrpc_request *request,
                         struct osc_brw_async_args *aa)
{
        struct ptlrpc_request *new_req;
        struct ptlrpc_request_set *set = request->rq_set;
        struct osc_brw_async_args *new_aa;
        struct osc_async_page *oap;
        int rc = 0;
        ENTRY;

        if (aa->aa_retries-- <= 0) {
                CERROR("too many checksum retries, returning error\n");
                RETURN(-EIO);
        }

        DEBUG_REQ(D_ERROR, request, "redo for checksum error");
        list_for_each_entry(oap, &aa->aa_oaps, oap_rpc_item) {
                if (oap->oap_request != NULL) {
                        LASSERTF(request == oap->oap_request,
                                 "request %p != oap_request %p\n",
                                 request, oap->oap_request);
                        if (oap->oap_interrupted) {
                                ptlrpc_mark_interrupted(oap->oap_request);
                                rc = -EINTR;
                                break;
                        }
                }
        }
        if (rc)
                RETURN(rc);

        rc = osc_brw_prep_request(lustre_msg_get_opc(request->rq_reqmsg) ==
                                        OST_WRITE ? OBD_BRW_WRITE :OBD_BRW_READ,
                                  aa->aa_cli, aa->aa_oa,
                                  NULL /* lsm unused by osc currently */,
                                  aa->aa_page_count, aa->aa_ppga, &new_req);
        if (rc)
                RETURN(rc);

        /* New request takes over pga and oaps from old request.
         * Note that copying a list_head doesn't work, need to move it... */
        new_req->rq_interpret_reply = request->rq_interpret_reply;
        new_req->rq_async_args = request->rq_async_args;
        new_aa = (struct osc_brw_async_args *)&new_req->rq_async_args;
        INIT_LIST_HEAD(&new_aa->aa_oaps);
        list_splice(&aa->aa_oaps, &new_aa->aa_oaps);
        INIT_LIST_HEAD(&aa->aa_oaps);

        list_for_each_entry(oap, &new_aa->aa_oaps, oap_rpc_item) {
                if (oap->oap_request) {
                        ptlrpc_req_finished(oap->oap_request);
                        oap->oap_request = ptlrpc_request_addref(new_req);
                }
        }

        ptlrpc_set_add_req(set, new_req);

        RETURN(0);
}

static int brw_interpret(struct ptlrpc_request *request, void *data, int rc)
{
        struct osc_brw_async_args *aa = data;
        int                        i;
        ENTRY;

        rc = osc_brw_fini_request(request, rc);
        if (rc == -EAGAIN) {
                rc = osc_brw_redo_request(request, aa);
                if (rc == 0)
                        RETURN(0);
        }

        spin_lock(&aa->aa_cli->cl_loi_list_lock);
        for (i = 0; i < aa->aa_page_count; i++)
                osc_release_write_grant(aa->aa_cli, aa->aa_ppga[i], 1);
        spin_unlock(&aa->aa_cli->cl_loi_list_lock);

        osc_release_ppga(aa->aa_ppga, aa->aa_page_count);

        RETURN(rc);
}

static int async_internal(int cmd, struct obd_export *exp, struct obdo *oa,
                          struct lov_stripe_md *lsm, obd_count page_count,
                          struct brw_page **pga, struct ptlrpc_request_set *set)
{
        struct ptlrpc_request     *request;
        struct client_obd         *cli = &exp->exp_obd->u.cli;
        int                        rc, i;
        ENTRY;

        /* Consume write credits even if doing a sync write -
         * otherwise we may run out of space on OST due to grant. */
        spin_lock(&cli->cl_loi_list_lock);
        for (i = 0; i < page_count; i++) {
                if (cli->cl_avail_grant >= CFS_PAGE_SIZE)
                        osc_consume_write_grant(cli, pga[i]);
        }
        spin_unlock(&cli->cl_loi_list_lock);

        rc = osc_brw_prep_request(cmd, &exp->exp_obd->u.cli, oa, lsm,
                                  page_count, pga, &request);

        if (rc == 0) {
                request->rq_interpret_reply = brw_interpret;
                ptlrpc_set_add_req(set, request);
        } else {
                spin_lock(&cli->cl_loi_list_lock);
                for (i = 0; i < page_count; i++)
                        osc_release_write_grant(cli, pga[i], 0);
                spin_unlock(&cli->cl_loi_list_lock);
        }

        RETURN (rc);
}

/*
 * ugh, we want disk allocation on the target to happen in offset order.  we'll
 * follow sedgewicks advice and stick to the dead simple shellsort -- it'll do
 * fine for our small page arrays and doesn't require allocation.  its an
 * insertion sort that swaps elements that are strides apart, shrinking the
 * stride down until its '1' and the array is sorted.
 */
static void sort_brw_pages(struct brw_page **array, int num)
{
        int stride, i, j;
        struct brw_page *tmp;

        if (num == 1)
                return;
        for (stride = 1; stride < num ; stride = (stride * 3) + 1)
                ;

        do {
                stride /= 3;
                for (i = stride ; i < num ; i++) {
                        tmp = array[i];
                        j = i;
                        while (j >= stride && array[j-stride]->off > tmp->off) {
                                array[j] = array[j - stride];
                                j -= stride;
                        }
                        array[j] = tmp;
                }
        } while (stride > 1);
}

static obd_count max_unfragmented_pages(struct brw_page **pg, obd_count pages)
{
        int count = 1;
        int offset;
        int i = 0;

        LASSERT (pages > 0);
        offset = pg[i]->off & (~CFS_PAGE_MASK);

        for (;;) {
                pages--;
                if (pages == 0)         /* that's all */
                        return count;

                if (offset + pg[i]->count < CFS_PAGE_SIZE)
                        return count;   /* doesn't end on page boundary */

                i++;
                offset = pg[i]->off & (~CFS_PAGE_MASK);
                if (offset != 0)        /* doesn't start on page boundary */
                        return count;

                count++;
        }
}

static struct brw_page **osc_build_ppga(struct brw_page *pga, obd_count count)
{
        struct brw_page **ppga;
        int i;

        OBD_ALLOC(ppga, sizeof(*ppga) * count);
        if (ppga == NULL)
                return NULL;

        for (i = 0; i < count; i++)
                ppga[i] = pga + i;
        return ppga;
}

static void osc_release_ppga(struct brw_page **ppga, obd_count count)
{
        LASSERT(ppga != NULL);
        OBD_FREE(ppga, sizeof(*ppga) * count);
}

static int osc_brw(int cmd, struct obd_export *exp, struct obd_info *oinfo,
                   obd_count page_count, struct brw_page *pga,
                   struct obd_trans_info *oti)
{
        struct obdo *saved_oa = NULL;
        struct brw_page **ppga, **orig;
        struct obd_import *imp = class_exp2cliimp(exp);
        struct client_obd *cli = &imp->imp_obd->u.cli;
        int rc, page_count_orig;
        ENTRY;

        if (cmd & OBD_BRW_CHECK) {
                /* The caller just wants to know if there's a chance that this
                 * I/O can succeed */

                if (imp == NULL || imp->imp_invalid)
                        RETURN(-EIO);
                RETURN(0);
        }

        /* test_brw with a failed create can trip this, maybe others. */
        LASSERT(cli->cl_max_pages_per_rpc);

        rc = 0;

        orig = ppga = osc_build_ppga(pga, page_count);
        if (ppga == NULL)
                RETURN(-ENOMEM);
        page_count_orig = page_count;

        sort_brw_pages(ppga, page_count);
        while (page_count) {
                obd_count pages_per_brw;

                if (page_count > cli->cl_max_pages_per_rpc)
                        pages_per_brw = cli->cl_max_pages_per_rpc;
                else
                        pages_per_brw = page_count;

                pages_per_brw = max_unfragmented_pages(ppga, pages_per_brw);

                if (saved_oa != NULL) {
                        /* restore previously saved oa */
                        *oinfo->oi_oa = *saved_oa;
                } else if (page_count > pages_per_brw) {
                        /* save a copy of oa (brw will clobber it) */
                        saved_oa = obdo_alloc();
                        if (saved_oa == NULL)
                                GOTO(out, rc = -ENOMEM);
                        *saved_oa = *oinfo->oi_oa;
                }

                rc = osc_brw_internal(cmd, exp, oinfo->oi_oa, oinfo->oi_md,
                                      pages_per_brw, ppga);

                if (rc != 0)
                        break;

                page_count -= pages_per_brw;
                ppga += pages_per_brw;
        }

out:
        osc_release_ppga(orig, page_count_orig);

        if (saved_oa != NULL)
                obdo_free(saved_oa);

        RETURN(rc);
}

static int osc_brw_async(int cmd, struct obd_export *exp,
                         struct obd_info *oinfo, obd_count page_count,
                         struct brw_page *pga, struct obd_trans_info *oti,
                         struct ptlrpc_request_set *set)
{
        struct brw_page **ppga, **orig;
        int page_count_orig;
        int rc = 0;
        ENTRY;

        if (cmd & OBD_BRW_CHECK) {
                /* The caller just wants to know if there's a chance that this
                 * I/O can succeed */
                struct obd_import *imp = class_exp2cliimp(exp);

                if (imp == NULL || imp->imp_invalid)
                        RETURN(-EIO);
                RETURN(0);
        }

        orig = ppga = osc_build_ppga(pga, page_count);
        if (ppga == NULL)
                RETURN(-ENOMEM);
        page_count_orig = page_count;

        sort_brw_pages(ppga, page_count);
        while (page_count) {
                struct brw_page **copy;
                obd_count pages_per_brw;

                pages_per_brw = min_t(obd_count, page_count,
                    class_exp2cliimp(exp)->imp_obd->u.cli.cl_max_pages_per_rpc);

                pages_per_brw = max_unfragmented_pages(ppga, pages_per_brw);

                /* use ppga only if single RPC is going to fly */
                if (pages_per_brw != page_count_orig || ppga != orig) {
                        OBD_ALLOC(copy, pages_per_brw * sizeof(*copy));
                        if (copy == NULL)
                                GOTO(out, rc = -ENOMEM);
                        memcpy(copy, ppga, pages_per_brw * sizeof(*copy));
                } else
                        copy = ppga;

                rc = async_internal(cmd, exp, oinfo->oi_oa, oinfo->oi_md,
                                    pages_per_brw, copy, set);

                if (rc != 0) {
                        if (copy != ppga)
                                OBD_FREE(copy, pages_per_brw * sizeof(*copy));
                        break;
                }

                if (copy == orig) {
                        /* we passed it to async_internal() which is
                         * now responsible for releasing memory */
                        orig = NULL;
                }

                page_count -= pages_per_brw;
                ppga += pages_per_brw;
        }
out:
        if (orig)
                osc_release_ppga(orig, page_count_orig);
        RETURN(rc);
}

static void osc_check_rpcs(struct client_obd *cli);

/* The companion to osc_enter_cache(), called when @oap is no longer part of
 * the dirty accounting.  Writeback completes or truncate happens before
 * writing starts.  Must be called with the loi lock held. */
static void osc_exit_cache(struct client_obd *cli, struct osc_async_page *oap,
                           int sent)
{
        osc_release_write_grant(cli, &oap->oap_brw_page, sent);
}

/* This maintains the lists of pending pages to read/write for a given object
 * (lop).  This is used by osc_check_rpcs->osc_next_loi() and loi_list_maint()
 * to quickly find objects that are ready to send an RPC. */
static int lop_makes_rpc(struct client_obd *cli, struct loi_oap_pages *lop,
                         int cmd)
{
        int optimal;
        ENTRY;

        if (lop->lop_num_pending == 0)
                RETURN(0);

        /* if we have an invalid import we want to drain the queued pages
         * by forcing them through rpcs that immediately fail and complete
         * the pages.  recovery relies on this to empty the queued pages
         * before canceling the locks and evicting down the llite pages */
        if (cli->cl_import == NULL || cli->cl_import->imp_invalid)
                RETURN(1);

        /* stream rpcs in queue order as long as as there is an urgent page
         * queued.  this is our cheap solution for good batching in the case
         * where writepage marks some random page in the middle of the file
         * as urgent because of, say, memory pressure */
        if (!list_empty(&lop->lop_urgent)) {
                CDEBUG(D_CACHE, "urgent request forcing RPC\n");
                RETURN(1);
        }

        /* fire off rpcs when we have 'optimal' rpcs as tuned for the wire. */
        optimal = cli->cl_max_pages_per_rpc;
        if (cmd & OBD_BRW_WRITE) {
                /* trigger a write rpc stream as long as there are dirtiers
                 * waiting for space.  as they're waiting, they're not going to
                 * create more pages to coallesce with what's waiting.. */
                if (!list_empty(&cli->cl_cache_waiters)) {
                        CDEBUG(D_CACHE, "cache waiters forcing RPC\n");
                        RETURN(1);
                }

                /* +16 to avoid triggering rpcs that would want to include pages
                 * that are being queued but which can't be made ready until
                 * the queuer finishes with the page. this is a wart for
                 * llite::commit_write() */
                optimal += 16;
        }
        if (lop->lop_num_pending >= optimal)
                RETURN(1);

        RETURN(0);
}

static void on_list(struct list_head *item, struct list_head *list,
                    int should_be_on)
{
        if (list_empty(item) && should_be_on)
                list_add_tail(item, list);
        else if (!list_empty(item) && !should_be_on)
                list_del_init(item);
}

/* maintain the loi's cli list membership invariants so that osc_send_oap_rpc
 * can find pages to build into rpcs quickly */
static void loi_list_maint(struct client_obd *cli, struct lov_oinfo *loi)
{
        on_list(&loi->loi_cli_item, &cli->cl_loi_ready_list,
                lop_makes_rpc(cli, &loi->loi_write_lop, OBD_BRW_WRITE) ||
                lop_makes_rpc(cli, &loi->loi_read_lop, OBD_BRW_READ));

        on_list(&loi->loi_write_item, &cli->cl_loi_write_list,
                loi->loi_write_lop.lop_num_pending);

        on_list(&loi->loi_read_item, &cli->cl_loi_read_list,
                loi->loi_read_lop.lop_num_pending);
}

static void lop_update_pending(struct client_obd *cli,
                               struct loi_oap_pages *lop, int cmd, int delta)
{
        lop->lop_num_pending += delta;
        if (cmd & OBD_BRW_WRITE)
                cli->cl_pending_w_pages += delta;
        else
                cli->cl_pending_r_pages += delta;
}

/* this is called when a sync waiter receives an interruption.  Its job is to
 * get the caller woken as soon as possible.  If its page hasn't been put in an
 * rpc yet it can dequeue immediately.  Otherwise it has to mark the rpc as
 * desiring interruption which will forcefully complete the rpc once the rpc
 * has timed out */
static void osc_occ_interrupted(struct oig_callback_context *occ)
{
        struct osc_async_page *oap;
        struct loi_oap_pages *lop;
        struct lov_oinfo *loi;
        ENTRY;

        /* XXX member_of() */
        oap = list_entry(occ, struct osc_async_page, oap_occ);

        client_obd_list_lock(&oap->oap_cli->cl_loi_list_lock);

        oap->oap_interrupted = 1;

        /* ok, it's been put in an rpc. only one oap gets a request reference */
        if (oap->oap_request != NULL) {
                ptlrpc_mark_interrupted(oap->oap_request);
                ptlrpcd_wake(oap->oap_request);
                GOTO(unlock, 0);
        }

        /* we don't get interruption callbacks until osc_trigger_group_io()
         * has been called and put the sync oaps in the pending/urgent lists.*/
        if (!list_empty(&oap->oap_pending_item)) {
                list_del_init(&oap->oap_pending_item);
                list_del_init(&oap->oap_urgent_item);

                loi = oap->oap_loi;
                lop = (oap->oap_cmd & OBD_BRW_WRITE) ?
                        &loi->loi_write_lop : &loi->loi_read_lop;
                lop_update_pending(oap->oap_cli, lop, oap->oap_cmd, -1);
                loi_list_maint(oap->oap_cli, oap->oap_loi);

                oig_complete_one(oap->oap_oig, &oap->oap_occ, -EINTR);
                oap->oap_oig = NULL;
        }

unlock:
        client_obd_list_unlock(&oap->oap_cli->cl_loi_list_lock);
}

/* this is trying to propogate async writeback errors back up to the
 * application.  As an async write fails we record the error code for later if
 * the app does an fsync.  As long as errors persist we force future rpcs to be
 * sync so that the app can get a sync error and break the cycle of queueing
 * pages for which writeback will fail. */
static void osc_process_ar(struct osc_async_rc *ar, struct ptlrpc_request *req,
                           int rc)
{
        if (rc) {
                if (!ar->ar_rc)
                        ar->ar_rc = rc;

                ar->ar_force_sync = 1;
                ar->ar_min_xid = ptlrpc_sample_next_xid();
                return;

        }

        if (ar->ar_force_sync && req && (ptlrpc_req_xid(req) >= ar->ar_min_xid))
                ar->ar_force_sync = 0;
}

static void osc_oap_to_pending(struct osc_async_page *oap)
{
        struct loi_oap_pages *lop;

        if (oap->oap_cmd & OBD_BRW_WRITE)
                lop = &oap->oap_loi->loi_write_lop;
        else
                lop = &oap->oap_loi->loi_read_lop;

        if (oap->oap_async_flags & ASYNC_URGENT)
                list_add(&oap->oap_urgent_item, &lop->lop_urgent);
        list_add_tail(&oap->oap_pending_item, &lop->lop_pending);
        lop_update_pending(oap->oap_cli, lop, oap->oap_cmd, 1);
}

/* this must be called holding the loi list lock to give coverage to exit_cache,
 * async_flag maintenance, and oap_request */
static void osc_ap_completion(struct client_obd *cli, struct obdo *oa,
                              struct osc_async_page *oap, int sent, int rc)
{
        ENTRY;
        oap->oap_async_flags = 0;
        oap->oap_interrupted = 0;

        if (oap->oap_cmd & OBD_BRW_WRITE) {
                osc_process_ar(&cli->cl_ar, oap->oap_request, rc);
                osc_process_ar(&oap->oap_loi->loi_ar, oap->oap_request, rc);
        }

        if (oap->oap_request != NULL) {
                ptlrpc_req_finished(oap->oap_request);
                oap->oap_request = NULL;
        }

        if (rc == 0 && oa != NULL) {
                if (oa->o_valid & OBD_MD_FLBLOCKS)
                        oap->oap_loi->loi_lvb.lvb_blocks = oa->o_blocks;
                if (oa->o_valid & OBD_MD_FLMTIME)
                        oap->oap_loi->loi_lvb.lvb_mtime = oa->o_mtime;
                if (oa->o_valid & OBD_MD_FLATIME)
                        oap->oap_loi->loi_lvb.lvb_atime = oa->o_atime;
                if (oa->o_valid & OBD_MD_FLCTIME)
                        oap->oap_loi->loi_lvb.lvb_ctime = oa->o_ctime;
        }

        if (oap->oap_oig) {
                osc_exit_cache(cli, oap, sent);
                oig_complete_one(oap->oap_oig, &oap->oap_occ, rc);
                oap->oap_oig = NULL;
                EXIT;
                return;
        }

        rc = oap->oap_caller_ops->ap_completion(oap->oap_caller_data,
                                                oap->oap_cmd, oa, rc);

        /* ll_ap_completion (from llite) drops PG_locked. so, a new
         * I/O on the page could start, but OSC calls it under lock
         * and thus we can add oap back to pending safely */
        if (rc)
                /* upper layer wants to leave the page on pending queue */
                osc_oap_to_pending(oap);
        else
                osc_exit_cache(cli, oap, sent);
        EXIT;
}

static int brw_interpret_oap(struct ptlrpc_request *request, void *data, int rc)
{
        struct osc_brw_async_args *aa = data;
        struct osc_async_page *oap, *tmp;
        struct client_obd *cli;
        ENTRY;

        rc = osc_brw_fini_request(request, rc);
        CDEBUG(D_INODE, "request %p aa %p rc %d\n", request, aa, rc);
        if (rc == -EAGAIN) {
                rc = osc_brw_redo_request(request, aa);
                if (rc == 0)
                        RETURN(0);
                GOTO(out, rc);
        }

        cli = aa->aa_cli;

        client_obd_list_lock(&cli->cl_loi_list_lock);

        /* We need to decrement before osc_ap_completion->osc_wake_cache_waiters
         * is called so we know whether to go to sync BRWs or wait for more
         * RPCs to complete */
        if (lustre_msg_get_opc(request->rq_reqmsg) == OST_WRITE)
                cli->cl_w_in_flight--;
        else
                cli->cl_r_in_flight--;

        /* the caller may re-use the oap after the completion call so
         * we need to clean it up a little */
        list_for_each_entry_safe(oap, tmp, &aa->aa_oaps, oap_rpc_item) {
                list_del_init(&oap->oap_rpc_item);
                osc_ap_completion(cli, aa->aa_oa, oap, 1, rc);
        }

        osc_wake_cache_waiters(cli);
        osc_check_rpcs(cli);

        client_obd_list_unlock(&cli->cl_loi_list_lock);

        obdo_free(aa->aa_oa);

        rc = 0;
out:
        osc_release_ppga(aa->aa_ppga, aa->aa_page_count);
        RETURN(rc);
}

static struct ptlrpc_request *osc_build_req(struct client_obd *cli,
                                            struct list_head *rpc_list,
                                            int page_count, int cmd)
{
        struct ptlrpc_request *req;
        struct brw_page **pga = NULL;
        struct osc_brw_async_args *aa;
        struct obdo *oa = NULL;
        struct obd_async_page_ops *ops = NULL;
        void *caller_data = NULL;
        struct osc_async_page *oap;
        int i, rc;

        ENTRY;
        LASSERT(!list_empty(rpc_list));

        OBD_ALLOC(pga, sizeof(*pga) * page_count);
        if (pga == NULL)
                RETURN(ERR_PTR(-ENOMEM));

        oa = obdo_alloc();
        if (oa == NULL)
                GOTO(out, req = ERR_PTR(-ENOMEM));

        i = 0;
        list_for_each_entry(oap, rpc_list, oap_rpc_item) {
                if (ops == NULL) {
                        ops = oap->oap_caller_ops;
                        caller_data = oap->oap_caller_data;
                }
                pga[i] = &oap->oap_brw_page;
                pga[i]->off = oap->oap_obj_off + oap->oap_page_off;
                CDEBUG(0, "put page %p index %lu oap %p flg %x to pga\n",
                       pga[i]->pg, cfs_page_index(oap->oap_page), oap, pga[i]->flag);
                i++;
        }

        /* always get the data for the obdo for the rpc */
        LASSERT(ops != NULL);
        ops->ap_fill_obdo(caller_data, cmd, oa);

        sort_brw_pages(pga, page_count);
        rc = osc_brw_prep_request(cmd, cli, oa, NULL, page_count, pga, &req);
        if (rc != 0) {
                CERROR("prep_req failed: %d\n", rc);
                GOTO(out, req = ERR_PTR(rc));
        }

        /* Need to update the timestamps after the request is built in case
         * we race with setattr (locally or in queue at OST).  If OST gets
         * later setattr before earlier BRW (as determined by the request xid),
         * the OST will not use BRW timestamps.  Sadly, there is no obvious
         * way to do this in a single call.  bug 10150 */
        ops->ap_update_obdo(caller_data, cmd, oa,
                            OBD_MD_FLMTIME | OBD_MD_FLCTIME | OBD_MD_FLATIME);

        CLASSERT(sizeof(*aa) <= sizeof(req->rq_async_args));
        aa = (struct osc_brw_async_args *)&req->rq_async_args;
        INIT_LIST_HEAD(&aa->aa_oaps);
        list_splice(rpc_list, &aa->aa_oaps);
        INIT_LIST_HEAD(rpc_list);

out:
        if (IS_ERR(req)) {
                if (oa)
                        obdo_free(oa);
                if (pga)
                        OBD_FREE(pga, sizeof(*pga) * page_count);
        }
        RETURN(req);
}

/* the loi lock is held across this function but it's allowed to release
 * and reacquire it during its work */
static int osc_send_oap_rpc(struct client_obd *cli, struct lov_oinfo *loi,
                            int cmd, struct loi_oap_pages *lop)
{
        struct ptlrpc_request *req;
        obd_count page_count = 0;
        struct osc_async_page *oap = NULL, *tmp;
        struct osc_brw_async_args *aa;
        struct obd_async_page_ops *ops;
        CFS_LIST_HEAD(rpc_list);
        unsigned int ending_offset;
        unsigned  starting_offset = 0;
        ENTRY;

        /* first we find the pages we're allowed to work with */
        list_for_each_entry_safe(oap, tmp, &lop->lop_pending, oap_pending_item){
                ops = oap->oap_caller_ops;

                LASSERT(oap->oap_magic == OAP_MAGIC);

                /* in llite being 'ready' equates to the page being locked
                 * until completion unlocks it.  commit_write submits a page
                 * as not ready because its unlock will happen unconditionally
                 * as the call returns.  if we race with commit_write giving
                 * us that page we dont' want to create a hole in the page
                 * stream, so we stop and leave the rpc to be fired by
                 * another dirtier or kupdated interval (the not ready page
                 * will still be on the dirty list).  we could call in
                 * at the end of ll_file_write to process the queue again. */
                if (!(oap->oap_async_flags & ASYNC_READY)) {
                        int rc = ops->ap_make_ready(oap->oap_caller_data, cmd);
                        if (rc < 0)
                                CDEBUG(D_INODE, "oap %p page %p returned %d "
                                                "instead of ready\n", oap,
                                                oap->oap_page, rc);
                        switch (rc) {
                        case -EAGAIN:
                                /* llite is telling us that the page is still
                                 * in commit_write and that we should try
                                 * and put it in an rpc again later.  we
                                 * break out of the loop so we don't create
                                 * a hole in the sequence of pages in the rpc
                                 * stream.*/
                                oap = NULL;
                                break;
                        case -EINTR:
                                /* the io isn't needed.. tell the checks
                                 * below to complete the rpc with EINTR */
                                oap->oap_async_flags |= ASYNC_COUNT_STABLE;
                                oap->oap_count = -EINTR;
                                break;
                        case 0:
                                oap->oap_async_flags |= ASYNC_READY;
                                break;
                        default:
                                LASSERTF(0, "oap %p page %p returned %d "
                                            "from make_ready\n", oap,
                                            oap->oap_page, rc);
                                break;
                        }
                }
                if (oap == NULL)
                        break;
                /*
                 * Page submitted for IO has to be locked. Either by
                 * ->ap_make_ready() or by higher layers.
                 *
                 * XXX nikita: this assertion should be adjusted when lustre
                 * starts using PG_writeback for pages being written out.
                 */
#if defined(__KERNEL__) && defined(__LINUX__)
                LASSERT(PageLocked(oap->oap_page));
#endif
                /* If there is a gap at the start of this page, it can't merge
                 * with any previous page, so we'll hand the network a
                 * "fragmented" page array that it can't transfer in 1 RDMA */
                if (page_count != 0 && oap->oap_page_off != 0)
                        break;

                /* take the page out of our book-keeping */
                list_del_init(&oap->oap_pending_item);
                lop_update_pending(cli, lop, cmd, -1);
                list_del_init(&oap->oap_urgent_item);

                if (page_count == 0)
                        starting_offset = (oap->oap_obj_off+oap->oap_page_off) &
                                          (PTLRPC_MAX_BRW_SIZE - 1);

                /* ask the caller for the size of the io as the rpc leaves. */
                if (!(oap->oap_async_flags & ASYNC_COUNT_STABLE))
                        oap->oap_count =
                                ops->ap_refresh_count(oap->oap_caller_data,cmd);
                if (oap->oap_count <= 0) {
                        CDEBUG(D_CACHE, "oap %p count %d, completing\n", oap,
                               oap->oap_count);
                        osc_ap_completion(cli, NULL, oap, 0, oap->oap_count);
                        continue;
                }

                /* now put the page back in our accounting */
                list_add_tail(&oap->oap_rpc_item, &rpc_list);
                if (++page_count >= cli->cl_max_pages_per_rpc)
                        break;

                /* End on a PTLRPC_MAX_BRW_SIZE boundary.  We want full-sized
                 * RPCs aligned on PTLRPC_MAX_BRW_SIZE boundaries to help reads
                 * have the same alignment as the initial writes that allocated
                 * extents on the server. */
                ending_offset = (oap->oap_obj_off + oap->oap_page_off +
                                 oap->oap_count) & (PTLRPC_MAX_BRW_SIZE - 1);
                if (ending_offset == 0)
                        break;

                /* If there is a gap at the end of this page, it can't merge
                 * with any subsequent pages, so we'll hand the network a
                 * "fragmented" page array that it can't transfer in 1 RDMA */
                if (oap->oap_page_off + oap->oap_count < CFS_PAGE_SIZE)
                        break;
        }

        osc_wake_cache_waiters(cli);

        if (page_count == 0)
                RETURN(0);

        loi_list_maint(cli, loi);

        client_obd_list_unlock(&cli->cl_loi_list_lock);

        req = osc_build_req(cli, &rpc_list, page_count, cmd);
        if (IS_ERR(req)) {
                /* this should happen rarely and is pretty bad, it makes the
                 * pending list not follow the dirty order */
                client_obd_list_lock(&cli->cl_loi_list_lock);
                list_for_each_entry_safe(oap, tmp, &rpc_list, oap_rpc_item) {
                        list_del_init(&oap->oap_rpc_item);

                        /* queued sync pages can be torn down while the pages
                         * were between the pending list and the rpc */
                        if (oap->oap_interrupted) {
                                CDEBUG(D_INODE, "oap %p interrupted\n", oap);
                                osc_ap_completion(cli, NULL, oap, 0,
                                                  oap->oap_count);
                                continue;
                        }
                        osc_ap_completion(cli, NULL, oap, 0, PTR_ERR(req));
                }
                loi_list_maint(cli, loi);
                RETURN(PTR_ERR(req));
        }

        aa = (struct osc_brw_async_args *)&req->rq_async_args;
        if (cmd == OBD_BRW_READ) {
                lprocfs_oh_tally_log2(&cli->cl_read_page_hist, page_count);
                lprocfs_oh_tally(&cli->cl_read_rpc_hist, cli->cl_r_in_flight);
                lprocfs_oh_tally_log2(&cli->cl_read_offset_hist,
                                      (starting_offset >> CFS_PAGE_SHIFT) + 1);
                ptlrpc_lprocfs_brw(req, OST_READ, aa->aa_requested_nob);
        } else {
                lprocfs_oh_tally_log2(&cli->cl_write_page_hist, page_count);
                lprocfs_oh_tally(&cli->cl_write_rpc_hist,
                                 cli->cl_w_in_flight);
                lprocfs_oh_tally_log2(&cli->cl_write_offset_hist,
                                      (starting_offset >> CFS_PAGE_SHIFT) + 1);
                ptlrpc_lprocfs_brw(req, OST_WRITE, aa->aa_requested_nob);
        }

        client_obd_list_lock(&cli->cl_loi_list_lock);

        if (cmd == OBD_BRW_READ)
                cli->cl_r_in_flight++;
        else
                cli->cl_w_in_flight++;

        /* queued sync pages can be torn down while the pages
         * were between the pending list and the rpc */
        tmp = NULL;
        list_for_each_entry(oap, &aa->aa_oaps, oap_rpc_item) {
                /* only one oap gets a request reference */
                if (tmp == NULL)
                        tmp = oap;
                if (oap->oap_interrupted && !req->rq_intr) {
                        CDEBUG(D_INODE, "oap %p in req %p interrupted\n",
                               oap, req);
                        ptlrpc_mark_interrupted(req);
                }
        }
        if (tmp != NULL)
                tmp->oap_request = ptlrpc_request_addref(req);

        DEBUG_REQ(D_INODE, req, "%d pages, aa %p. now %dr/%dw in flight",
                  page_count, aa, cli->cl_r_in_flight, cli->cl_w_in_flight);

        req->rq_interpret_reply = brw_interpret_oap;
        ptlrpcd_add_req(req);
        RETURN(1);
}

#define LOI_DEBUG(LOI, STR, args...)                                     \
        CDEBUG(D_INODE, "loi ready %d wr %d:%d rd %d:%d " STR,           \
               !list_empty(&(LOI)->loi_cli_item),                        \
               (LOI)->loi_write_lop.lop_num_pending,                     \
               !list_empty(&(LOI)->loi_write_lop.lop_urgent),            \
               (LOI)->loi_read_lop.lop_num_pending,                      \
               !list_empty(&(LOI)->loi_read_lop.lop_urgent),             \
               args)                                                     \

/* This is called by osc_check_rpcs() to find which objects have pages that
 * we could be sending.  These lists are maintained by lop_makes_rpc(). */
struct lov_oinfo *osc_next_loi(struct client_obd *cli)
{
        ENTRY;
        /* first return all objects which we already know to have
         * pages ready to be stuffed into rpcs */
        if (!list_empty(&cli->cl_loi_ready_list))
                RETURN(list_entry(cli->cl_loi_ready_list.next,
                                  struct lov_oinfo, loi_cli_item));

        /* then if we have cache waiters, return all objects with queued
         * writes.  This is especially important when many small files
         * have filled up the cache and not been fired into rpcs because
         * they don't pass the nr_pending/object threshhold */
        if (!list_empty(&cli->cl_cache_waiters) &&
            !list_empty(&cli->cl_loi_write_list))
                RETURN(list_entry(cli->cl_loi_write_list.next,
                                  struct lov_oinfo, loi_write_item));

        /* then return all queued objects when we have an invalid import
         * so that they get flushed */
        if (cli->cl_import == NULL || cli->cl_import->imp_invalid) {
                if (!list_empty(&cli->cl_loi_write_list))
                        RETURN(list_entry(cli->cl_loi_write_list.next,
                                          struct lov_oinfo, loi_write_item));
                if (!list_empty(&cli->cl_loi_read_list))
                        RETURN(list_entry(cli->cl_loi_read_list.next,
                                          struct lov_oinfo, loi_read_item));
        }
        RETURN(NULL);
}

/* called with the loi list lock held */
static void osc_check_rpcs(struct client_obd *cli)
{
        struct lov_oinfo *loi;
        int rc = 0, race_counter = 0;
        ENTRY;

        while ((loi = osc_next_loi(cli)) != NULL) {
                LOI_DEBUG(loi, "%lu in flight\n", rpcs_in_flight(cli));

                if (rpcs_in_flight(cli) >= cli->cl_max_rpcs_in_flight)
                        break;

                /* attempt some read/write balancing by alternating between
                 * reads and writes in an object.  The makes_rpc checks here
                 * would be redundant if we were getting read/write work items
                 * instead of objects.  we don't want send_oap_rpc to drain a
                 * partial read pending queue when we're given this object to
                 * do io on writes while there are cache waiters */
                if (lop_makes_rpc(cli, &loi->loi_write_lop, OBD_BRW_WRITE)) {
                        rc = osc_send_oap_rpc(cli, loi, OBD_BRW_WRITE,
                                              &loi->loi_write_lop);
                        if (rc < 0)
                                break;
                        if (rc > 0)
                                race_counter = 0;
                        else
                                race_counter++;
                }
                if (lop_makes_rpc(cli, &loi->loi_read_lop, OBD_BRW_READ)) {
                        rc = osc_send_oap_rpc(cli, loi, OBD_BRW_READ,
                                              &loi->loi_read_lop);
                        if (rc < 0)
                                break;
                        if (rc > 0)
                                race_counter = 0;
                        else
                                race_counter++;
                }

                /* attempt some inter-object balancing by issueing rpcs
                 * for each object in turn */
                if (!list_empty(&loi->loi_cli_item))
                        list_del_init(&loi->loi_cli_item);
                if (!list_empty(&loi->loi_write_item))
                        list_del_init(&loi->loi_write_item);
                if (!list_empty(&loi->loi_read_item))
                        list_del_init(&loi->loi_read_item);

                loi_list_maint(cli, loi);

                /* send_oap_rpc fails with 0 when make_ready tells it to
                 * back off.  llite's make_ready does this when it tries
                 * to lock a page queued for write that is already locked.
                 * we want to try sending rpcs from many objects, but we
                 * don't want to spin failing with 0.  */
                if (race_counter == 10)
                        break;
        }
        EXIT;
}

/* we're trying to queue a page in the osc so we're subject to the
 * 'cl_dirty_max' limit on the number of pages that can be queued in the osc.
 * If the osc's queued pages are already at that limit, then we want to sleep
 * until there is space in the osc's queue for us.  We also may be waiting for
 * write credits from the OST if there are RPCs in flight that may return some
 * before we fall back to sync writes.
 *
 * We need this know our allocation was granted in the presence of signals */
static int ocw_granted(struct client_obd *cli, struct osc_cache_waiter *ocw)
{
        int rc;
        ENTRY;
        client_obd_list_lock(&cli->cl_loi_list_lock);
        rc = list_empty(&ocw->ocw_entry) || rpcs_in_flight(cli) == 0;
        client_obd_list_unlock(&cli->cl_loi_list_lock);
        RETURN(rc);
};

/* Caller must hold loi_list_lock - we drop/regain it if we need to wait for
 * grant or cache space. */
static int osc_enter_cache(struct client_obd *cli, struct lov_oinfo *loi,
                           struct osc_async_page *oap)
{
        struct osc_cache_waiter ocw;
        struct l_wait_info lwi = { 0 };
        ENTRY;

        CDEBUG(D_CACHE, "dirty: %ld/%d dirty_max: %ld/%d dropped: %lu "
               "grant: %lu\n", cli->cl_dirty, atomic_read(&obd_dirty_pages),
               cli->cl_dirty_max, obd_max_dirty_pages,
               cli->cl_lost_grant, cli->cl_avail_grant);

        /* force the caller to try sync io.  this can jump the list
         * of queued writes and create a discontiguous rpc stream */
        if (cli->cl_dirty_max < CFS_PAGE_SIZE || cli->cl_ar.ar_force_sync ||
            loi->loi_ar.ar_force_sync)
                RETURN(-EDQUOT);

        /* Hopefully normal case - cache space and write credits available */
        if ((cli->cl_dirty + CFS_PAGE_SIZE <= cli->cl_dirty_max) &&
            (atomic_read(&obd_dirty_pages) + 1 <= obd_max_dirty_pages) &&
            (cli->cl_avail_grant >= CFS_PAGE_SIZE)) {
                /* account for ourselves */
                osc_consume_write_grant(cli, &oap->oap_brw_page);
                RETURN(0);
        }

        /* Make sure that there are write rpcs in flight to wait for.  This
         * is a little silly as this object may not have any pending but
         * other objects sure might. */
        if (cli->cl_w_in_flight) {
                list_add_tail(&ocw.ocw_entry, &cli->cl_cache_waiters);
                cfs_waitq_init(&ocw.ocw_waitq);
                ocw.ocw_oap = oap;
                ocw.ocw_rc = 0;

                loi_list_maint(cli, loi);
                osc_check_rpcs(cli);
                client_obd_list_unlock(&cli->cl_loi_list_lock);

                CDEBUG(D_CACHE, "sleeping for cache space\n");
                l_wait_event(ocw.ocw_waitq, ocw_granted(cli, &ocw), &lwi);

                client_obd_list_lock(&cli->cl_loi_list_lock);
                if (!list_empty(&ocw.ocw_entry)) {
                        list_del(&ocw.ocw_entry);
                        RETURN(-EINTR);
                }
                RETURN(ocw.ocw_rc);
        }

        RETURN(-EDQUOT);
}

int osc_prep_async_page(struct obd_export *exp, struct lov_stripe_md *lsm,
                        struct lov_oinfo *loi, cfs_page_t *page,
                        obd_off offset, struct obd_async_page_ops *ops,
                        void *data, void **res)
{
        struct osc_async_page *oap;
        ENTRY;

        if (!page)
                return size_round(sizeof(*oap));

        oap = *res;
        oap->oap_magic = OAP_MAGIC;
        oap->oap_cli = &exp->exp_obd->u.cli;
        oap->oap_loi = loi;

        oap->oap_caller_ops = ops;
        oap->oap_caller_data = data;

        oap->oap_page = page;
        oap->oap_obj_off = offset;

        CFS_INIT_LIST_HEAD(&oap->oap_pending_item);
        CFS_INIT_LIST_HEAD(&oap->oap_urgent_item);
        CFS_INIT_LIST_HEAD(&oap->oap_rpc_item);

        oap->oap_occ.occ_interrupted = osc_occ_interrupted;

        CDEBUG(D_CACHE, "oap %p page %p obj off "LPU64"\n", oap, page, offset);
        RETURN(0);
}

struct osc_async_page *oap_from_cookie(void *cookie)
{
        struct osc_async_page *oap = cookie;
        if (oap->oap_magic != OAP_MAGIC)
                return ERR_PTR(-EINVAL);
        return oap;
};

static int osc_queue_async_io(struct obd_export *exp, struct lov_stripe_md *lsm,
                              struct lov_oinfo *loi, void *cookie,
                              int cmd, obd_off off, int count,
                              obd_flag brw_flags, enum async_flags async_flags)
{
        struct client_obd *cli = &exp->exp_obd->u.cli;
        struct osc_async_page *oap;
        int rc = 0;
        ENTRY;

        oap = oap_from_cookie(cookie);
        if (IS_ERR(oap))
                RETURN(PTR_ERR(oap));

        if (cli->cl_import == NULL || cli->cl_import->imp_invalid)
                RETURN(-EIO);

        if (!list_empty(&oap->oap_pending_item) ||
            !list_empty(&oap->oap_urgent_item) ||
            !list_empty(&oap->oap_rpc_item))
                RETURN(-EBUSY);

        /* check if the file's owner/group is over quota */
#ifdef HAVE_QUOTA_SUPPORT
        if ((cmd & OBD_BRW_WRITE) && !(cmd & OBD_BRW_NOQUOTA)){
                struct obd_async_page_ops *ops;
                struct obdo *oa;

                oa = obdo_alloc();
                if (oa == NULL)
                        RETURN(-ENOMEM);

                ops = oap->oap_caller_ops;
                ops->ap_fill_obdo(oap->oap_caller_data, cmd, oa);
                if (lquota_chkdq(quota_interface, cli, oa->o_uid, oa->o_gid) ==
                    NO_QUOTA)
                        rc = -EDQUOT;

                obdo_free(oa);
                if (rc)
                        RETURN(rc);
        }
#endif

        if (loi == NULL)
                loi = &lsm->lsm_oinfo[0];

        client_obd_list_lock(&cli->cl_loi_list_lock);

        oap->oap_cmd = cmd;
        oap->oap_page_off = off;
        oap->oap_count = count;
        oap->oap_brw_flags = brw_flags;
        oap->oap_async_flags = async_flags;

        if (cmd & OBD_BRW_WRITE) {
                rc = osc_enter_cache(cli, loi, oap);
                if (rc) {
                        client_obd_list_unlock(&cli->cl_loi_list_lock);
                        RETURN(rc);
                }
        }

        osc_oap_to_pending(oap);
        loi_list_maint(cli, loi);

        LOI_DEBUG(loi, "oap %p page %p added for cmd %d\n", oap, oap->oap_page,
                  cmd);

        osc_check_rpcs(cli);
        client_obd_list_unlock(&cli->cl_loi_list_lock);

        RETURN(0);
}

/* aka (~was & now & flag), but this is more clear :) */
#define SETTING(was, now, flag) (!(was & flag) && (now & flag))

static int osc_set_async_flags(struct obd_export *exp,
                               struct lov_stripe_md *lsm,
                               struct lov_oinfo *loi, void *cookie,
                               obd_flag async_flags)
{
        struct client_obd *cli = &exp->exp_obd->u.cli;
        struct loi_oap_pages *lop;
        struct osc_async_page *oap;
        int rc = 0;
        ENTRY;

        oap = oap_from_cookie(cookie);
        if (IS_ERR(oap))
                RETURN(PTR_ERR(oap));

        /*
         * bug 7311: OST-side locking is only supported for liblustre for now
         * (and liblustre never calls obd_set_async_flags(). I hope.), generic
         * implementation has to handle case where OST-locked page was picked
         * up by, e.g., ->writepage().
         */
        LASSERT(!(oap->oap_brw_flags & OBD_BRW_SRVLOCK));
        LASSERT(!LIBLUSTRE_CLIENT); /* check that liblustre angels do fear to
                                     * tread here. */

        if (cli->cl_import == NULL || cli->cl_import->imp_invalid)
                RETURN(-EIO);

        if (loi == NULL)
                loi = &lsm->lsm_oinfo[0];

        if (oap->oap_cmd & OBD_BRW_WRITE) {
                lop = &loi->loi_write_lop;
        } else {
                lop = &loi->loi_read_lop;
        }

        client_obd_list_lock(&cli->cl_loi_list_lock);

        if (list_empty(&oap->oap_pending_item))
                GOTO(out, rc = -EINVAL);

        if ((oap->oap_async_flags & async_flags) == async_flags)
                GOTO(out, rc = 0);

        if (SETTING(oap->oap_async_flags, async_flags, ASYNC_READY))
                oap->oap_async_flags |= ASYNC_READY;

        if (SETTING(oap->oap_async_flags, async_flags, ASYNC_URGENT)) {
                if (list_empty(&oap->oap_rpc_item)) {
                        list_add(&oap->oap_urgent_item, &lop->lop_urgent);
                        loi_list_maint(cli, loi);
                }
        }

        LOI_DEBUG(loi, "oap %p page %p has flags %x\n", oap, oap->oap_page,
                        oap->oap_async_flags);
out:
        osc_check_rpcs(cli);
        client_obd_list_unlock(&cli->cl_loi_list_lock);
        RETURN(rc);
}

static int osc_queue_group_io(struct obd_export *exp, struct lov_stripe_md *lsm,
                             struct lov_oinfo *loi,
                             struct obd_io_group *oig, void *cookie,
                             int cmd, obd_off off, int count,
                             obd_flag brw_flags,
                             obd_flag async_flags)
{
        struct client_obd *cli = &exp->exp_obd->u.cli;
        struct osc_async_page *oap;
        struct loi_oap_pages *lop;
        int rc = 0;
        ENTRY;

        oap = oap_from_cookie(cookie);
        if (IS_ERR(oap))
                RETURN(PTR_ERR(oap));

        if (cli->cl_import == NULL || cli->cl_import->imp_invalid)
                RETURN(-EIO);

        if (!list_empty(&oap->oap_pending_item) ||
            !list_empty(&oap->oap_urgent_item) ||
            !list_empty(&oap->oap_rpc_item))
                RETURN(-EBUSY);

        if (loi == NULL)
                loi = &lsm->lsm_oinfo[0];

        client_obd_list_lock(&cli->cl_loi_list_lock);

        oap->oap_cmd = cmd;
        oap->oap_page_off = off;
        oap->oap_count = count;
        oap->oap_brw_flags = brw_flags;
        oap->oap_async_flags = async_flags;

        if (cmd & OBD_BRW_WRITE)
                lop = &loi->loi_write_lop;
        else
                lop = &loi->loi_read_lop;

        list_add_tail(&oap->oap_pending_item, &lop->lop_pending_group);
        if (oap->oap_async_flags & ASYNC_GROUP_SYNC) {
                oap->oap_oig = oig;
                rc = oig_add_one(oig, &oap->oap_occ);
        }

        LOI_DEBUG(loi, "oap %p page %p on group pending: rc %d\n",
                  oap, oap->oap_page, rc);

        client_obd_list_unlock(&cli->cl_loi_list_lock);

        RETURN(rc);
}

static void osc_group_to_pending(struct client_obd *cli, struct lov_oinfo *loi,
                                 struct loi_oap_pages *lop, int cmd)
{
        struct list_head *pos, *tmp;
        struct osc_async_page *oap;

        list_for_each_safe(pos, tmp, &lop->lop_pending_group) {
                oap = list_entry(pos, struct osc_async_page, oap_pending_item);
                list_del(&oap->oap_pending_item);
                osc_oap_to_pending(oap);
        }
        loi_list_maint(cli, loi);
}

static int osc_trigger_group_io(struct obd_export *exp,
                                struct lov_stripe_md *lsm,
                                struct lov_oinfo *loi,
                                struct obd_io_group *oig)
{
        struct client_obd *cli = &exp->exp_obd->u.cli;
        ENTRY;

        if (loi == NULL)
                loi = &lsm->lsm_oinfo[0];

        client_obd_list_lock(&cli->cl_loi_list_lock);

        osc_group_to_pending(cli, loi, &loi->loi_write_lop, OBD_BRW_WRITE);
        osc_group_to_pending(cli, loi, &loi->loi_read_lop, OBD_BRW_READ);

        osc_check_rpcs(cli);
        client_obd_list_unlock(&cli->cl_loi_list_lock);

        RETURN(0);
}

static int osc_teardown_async_page(struct obd_export *exp,
                                   struct lov_stripe_md *lsm,
                                   struct lov_oinfo *loi, void *cookie)
{
        struct client_obd *cli = &exp->exp_obd->u.cli;
        struct loi_oap_pages *lop;
        struct osc_async_page *oap;
        int rc = 0;
        ENTRY;

        oap = oap_from_cookie(cookie);
        if (IS_ERR(oap))
                RETURN(PTR_ERR(oap));

        if (loi == NULL)
                loi = &lsm->lsm_oinfo[0];

        if (oap->oap_cmd & OBD_BRW_WRITE) {
                lop = &loi->loi_write_lop;
        } else {
                lop = &loi->loi_read_lop;
        }

        client_obd_list_lock(&cli->cl_loi_list_lock);

        if (!list_empty(&oap->oap_rpc_item))
                GOTO(out, rc = -EBUSY);

        osc_exit_cache(cli, oap, 0);
        osc_wake_cache_waiters(cli);

        if (!list_empty(&oap->oap_urgent_item)) {
                list_del_init(&oap->oap_urgent_item);
                oap->oap_async_flags &= ~ASYNC_URGENT;
        }
        if (!list_empty(&oap->oap_pending_item)) {
                list_del_init(&oap->oap_pending_item);
                lop_update_pending(cli, lop, oap->oap_cmd, -1);
        }
        loi_list_maint(cli, loi);

        LOI_DEBUG(loi, "oap %p page %p torn down\n", oap, oap->oap_page);
out:
        client_obd_list_unlock(&cli->cl_loi_list_lock);
        RETURN(rc);
}

static void osc_set_data_with_check(struct lustre_handle *lockh, void *data,
                                    int flags)
{
        struct ldlm_lock *lock = ldlm_handle2lock(lockh);

        if (lock == NULL) {
                CERROR("lockh %p, data %p - client evicted?\n", lockh, data);
                return;
        }
        lock_res_and_lock(lock);
#ifdef __KERNEL__
#ifdef __LINUX__
        /* Liang XXX: Darwin and Winnt checking should be added */
        if (lock->l_ast_data && lock->l_ast_data != data) {
                struct inode *new_inode = data;
                struct inode *old_inode = lock->l_ast_data;
                if (!(old_inode->i_state & I_FREEING))
                        LDLM_ERROR(lock, "inconsistent l_ast_data found");
                LASSERTF(old_inode->i_state & I_FREEING,
                         "Found existing inode %p/%lu/%u state %lu in lock: "
                         "setting data to %p/%lu/%u\n", old_inode,
                         old_inode->i_ino, old_inode->i_generation,
                         old_inode->i_state,
                         new_inode, new_inode->i_ino, new_inode->i_generation);
        }
#endif
#endif
        lock->l_ast_data = data;
        lock->l_flags |= (flags & LDLM_FL_NO_LRU);
        unlock_res_and_lock(lock);
        LDLM_LOCK_PUT(lock);
}

static int osc_change_cbdata(struct obd_export *exp, struct lov_stripe_md *lsm,
                             ldlm_iterator_t replace, void *data)
{
        struct ldlm_res_id res_id = { .name = {lsm->lsm_object_id} };
        struct obd_device *obd = class_exp2obd(exp);

        ldlm_resource_iterate(obd->obd_namespace, &res_id, replace, data);
        return 0;
}

static int osc_enqueue_fini(struct ptlrpc_request *req, struct obd_info *oinfo,
                            int intent, int rc)
{
        ENTRY;

        if (intent) {
                /* The request was created before ldlm_cli_enqueue call. */
                if (rc == ELDLM_LOCK_ABORTED) {
                        struct ldlm_reply *rep;

                        /* swabbed by ldlm_cli_enqueue() */
                        LASSERT_REPSWABBED(req, DLM_LOCKREPLY_OFF);
                        rep = lustre_msg_buf(req->rq_repmsg, DLM_LOCKREPLY_OFF,
                                             sizeof(*rep));
                        LASSERT(rep != NULL);
                        if (rep->lock_policy_res1)
                                rc = rep->lock_policy_res1;
                }
        }

        if ((intent && rc == ELDLM_LOCK_ABORTED) || !rc) {
                CDEBUG(D_INODE,"got kms "LPU64" blocks "LPU64" mtime "LPU64"\n",
                       oinfo->oi_md->lsm_oinfo->loi_lvb.lvb_size,
                       oinfo->oi_md->lsm_oinfo->loi_lvb.lvb_blocks,
                       oinfo->oi_md->lsm_oinfo->loi_lvb.lvb_mtime);
        }

        /* Call the update callback. */
        rc = oinfo->oi_cb_up(oinfo, rc);
        RETURN(rc);
}

static int osc_enqueue_interpret(struct ptlrpc_request *req,
                                 struct osc_enqueue_args *aa, int rc)
{
        int intent = aa->oa_ei->ei_flags & LDLM_FL_HAS_INTENT;
        struct lov_stripe_md *lsm = aa->oa_oi->oi_md;
        struct ldlm_lock *lock;

        /* ldlm_cli_enqueue is holding a reference on the lock, so it must
         * be valid. */
        lock = ldlm_handle2lock(aa->oa_oi->oi_lockh);

        /* Complete obtaining the lock procedure. */
        rc = ldlm_cli_enqueue_fini(aa->oa_exp, req, aa->oa_ei->ei_type, 1,
                                   aa->oa_ei->ei_mode,
                                   &aa->oa_ei->ei_flags,
                                   &lsm->lsm_oinfo->loi_lvb,
                                   sizeof(lsm->lsm_oinfo->loi_lvb),
                                   lustre_swab_ost_lvb,
                                   aa->oa_oi->oi_lockh, rc);

        /* Complete osc stuff. */
        rc = osc_enqueue_fini(req, aa->oa_oi, intent, rc);

        /* Release the lock for async request. */
        if (lustre_handle_is_used(aa->oa_oi->oi_lockh) && rc == ELDLM_OK)
                ldlm_lock_decref(aa->oa_oi->oi_lockh, aa->oa_ei->ei_mode);

        LASSERTF(lock != NULL, "lockh %p, req %p, aa %p - client evicted?\n",
                 aa->oa_oi->oi_lockh, req, aa);
        LDLM_LOCK_PUT(lock);
        return rc;
}

/* When enqueuing asynchronously, locks are not ordered, we can obtain a lock
 * from the 2nd OSC before a lock from the 1st one. This does not deadlock with
 * other synchronous requests, however keeping some locks and trying to obtain
 * others may take a considerable amount of time in a case of ost failure; and
 * when other sync requests do not get released lock from a client, the client
 * is excluded from the cluster -- such scenarious make the life difficult, so
 * release locks just after they are obtained. */
static int osc_enqueue(struct obd_export *exp, struct obd_info *oinfo,
                       struct obd_enqueue_info *einfo)
{
        struct ldlm_res_id res_id = { .name = {oinfo->oi_md->lsm_object_id} };
        struct obd_device *obd = exp->exp_obd;
        struct ldlm_reply *rep;
        struct ptlrpc_request *req = NULL;
        int intent = einfo->ei_flags & LDLM_FL_HAS_INTENT;
        int rc;
        ENTRY;

        /* Filesystem lock extents are extended to page boundaries so that
         * dealing with the page cache is a little smoother.  */
        oinfo->oi_policy.l_extent.start -=
                oinfo->oi_policy.l_extent.start & ~CFS_PAGE_MASK;
        oinfo->oi_policy.l_extent.end |= ~CFS_PAGE_MASK;

        if (oinfo->oi_md->lsm_oinfo->loi_kms_valid == 0)
                goto no_match;

        /* Next, search for already existing extent locks that will cover us */
        rc = ldlm_lock_match(obd->obd_namespace, einfo->ei_flags | LDLM_FL_LVB_READY, &res_id,
                             einfo->ei_type, &oinfo->oi_policy, einfo->ei_mode,
                             oinfo->oi_lockh);
        if (rc == 1) {
                osc_set_data_with_check(oinfo->oi_lockh, einfo->ei_cbdata,
                                        einfo->ei_flags);
                if (intent) {
                        /* I would like to be able to ASSERT here that rss <=
                         * kms, but I can't, for reasons which are explained in
                         * lov_enqueue() */
                }

                /* We already have a lock, and it's referenced */
                oinfo->oi_cb_up(oinfo, ELDLM_OK);

                /* For async requests, decref the lock. */
                if (einfo->ei_rqset)
                        ldlm_lock_decref(oinfo->oi_lockh, einfo->ei_mode);

                RETURN(ELDLM_OK);
        }

        /* If we're trying to read, we also search for an existing PW lock.  The
         * VFS and page cache already protect us locally, so lots of readers/
         * writers can share a single PW lock.
         *
         * There are problems with conversion deadlocks, so instead of
         * converting a read lock to a write lock, we'll just enqueue a new
         * one.
         *
         * At some point we should cancel the read lock instead of making them
         * send us a blocking callback, but there are problems with canceling
         * locks out from other users right now, too. */

        if (einfo->ei_mode == LCK_PR) {
                rc = ldlm_lock_match(obd->obd_namespace, einfo->ei_flags | LDLM_FL_LVB_READY,
                                     &res_id, einfo->ei_type, &oinfo->oi_policy,
                                     LCK_PW, oinfo->oi_lockh);
                if (rc == 1) {
                        /* FIXME: This is not incredibly elegant, but it might
                         * be more elegant than adding another parameter to
                         * lock_match.  I want a second opinion. */
                        /* addref the lock only if not async requests. */
                        if (!einfo->ei_rqset)
                                ldlm_lock_addref(oinfo->oi_lockh, LCK_PR);
                        osc_set_data_with_check(oinfo->oi_lockh,
                                                einfo->ei_cbdata,
                                                einfo->ei_flags);
                        oinfo->oi_cb_up(oinfo, ELDLM_OK);
                        ldlm_lock_decref(oinfo->oi_lockh, LCK_PW);
                        RETURN(ELDLM_OK);
                }
        }

 no_match:
        if (intent) {
                int size[3] = {
                        [MSG_PTLRPC_BODY_OFF] = sizeof(struct ptlrpc_body),
                        [DLM_LOCKREQ_OFF]     = sizeof(struct ldlm_request) };

                req = ptlrpc_prep_req(class_exp2cliimp(exp), LUSTRE_DLM_VERSION,
                                      LDLM_ENQUEUE, 2, size, NULL);
                if (req == NULL)
                        RETURN(-ENOMEM);

                size[DLM_LOCKREPLY_OFF] = sizeof(*rep);
                size[DLM_REPLY_REC_OFF] = 
                        sizeof(oinfo->oi_md->lsm_oinfo->loi_lvb);
                ptlrpc_req_set_repsize(req, 3, size);
        }

        /* users of osc_enqueue() can pass this flag for ldlm_lock_match() */
        einfo->ei_flags &= ~LDLM_FL_BLOCK_GRANTED;

        rc = ldlm_cli_enqueue(exp, &req, res_id, einfo->ei_type,
                              &oinfo->oi_policy, einfo->ei_mode,
                              &einfo->ei_flags, einfo->ei_cb_bl,
                              einfo->ei_cb_cp, einfo->ei_cb_gl,
                              einfo->ei_cbdata,
                              &oinfo->oi_md->lsm_oinfo->loi_lvb,
                              sizeof(oinfo->oi_md->lsm_oinfo->loi_lvb),
                              lustre_swab_ost_lvb, oinfo->oi_lockh,
                              einfo->ei_rqset ? 1 : 0);
        if (einfo->ei_rqset) {
                if (!rc) {
                        struct osc_enqueue_args *aa;
                        CLASSERT(sizeof(*aa) <= sizeof(req->rq_async_args));
                        aa = (struct osc_enqueue_args *)&req->rq_async_args;
                        aa->oa_oi = oinfo;
                        aa->oa_ei = einfo;
                        aa->oa_exp = exp;

                        req->rq_interpret_reply = osc_enqueue_interpret;
                        ptlrpc_set_add_req(einfo->ei_rqset, req);
                } else if (intent) {
                        ptlrpc_req_finished(req);
                }
                RETURN(rc);
        }

        rc = osc_enqueue_fini(req, oinfo, intent, rc);
        if (intent)
                ptlrpc_req_finished(req);

        RETURN(rc);
}

static int osc_match(struct obd_export *exp, struct lov_stripe_md *lsm,
                     __u32 type, ldlm_policy_data_t *policy, __u32 mode,
                     int *flags, void *data, struct lustre_handle *lockh)
{
        struct ldlm_res_id res_id = { .name = {lsm->lsm_object_id} };
        struct obd_device *obd = exp->exp_obd;
        int rc;
        int lflags = *flags;
        ENTRY;

        OBD_FAIL_RETURN(OBD_FAIL_OSC_MATCH, -EIO);

        /* Filesystem lock extents are extended to page boundaries so that
         * dealing with the page cache is a little smoother */
        policy->l_extent.start -= policy->l_extent.start & ~CFS_PAGE_MASK;
        policy->l_extent.end |= ~CFS_PAGE_MASK;

        /* Next, search for already existing extent locks that will cover us */
        rc = ldlm_lock_match(obd->obd_namespace, lflags | LDLM_FL_LVB_READY, &res_id, type,
                             policy, mode, lockh);
        if (rc) {
                //if (!(*flags & LDLM_FL_TEST_LOCK))
                        osc_set_data_with_check(lockh, data, lflags);
                RETURN(rc);
        }
        /* If we're trying to read, we also search for an existing PW lock.  The
         * VFS and page cache already protect us locally, so lots of readers/
         * writers can share a single PW lock. */
        if (mode == LCK_PR) {
                rc = ldlm_lock_match(obd->obd_namespace, lflags | LDLM_FL_LVB_READY, 
                                     &res_id, type,
                                     policy, LCK_PW, lockh);
                if (rc == 1 && !(lflags & LDLM_FL_TEST_LOCK)) {
                        /* FIXME: This is not incredibly elegant, but it might
                         * be more elegant than adding another parameter to
                         * lock_match.  I want a second opinion. */
                        osc_set_data_with_check(lockh, data, lflags);
                        ldlm_lock_addref(lockh, LCK_PR);
                        ldlm_lock_decref(lockh, LCK_PW);
                }
        }
        RETURN(rc);
}

static int osc_cancel(struct obd_export *exp, struct lov_stripe_md *md,
                      __u32 mode, struct lustre_handle *lockh)
{
        ENTRY;

        if (unlikely(mode == LCK_GROUP))
                ldlm_lock_decref_and_cancel(lockh, mode);
        else
                ldlm_lock_decref(lockh, mode);

        RETURN(0);
}

static int osc_cancel_unused(struct obd_export *exp,
                             struct lov_stripe_md *lsm, int flags, void *opaque)
{
        struct obd_device *obd = class_exp2obd(exp);
        struct ldlm_res_id res_id = { .name = {lsm->lsm_object_id} };

        return ldlm_cli_cancel_unused(obd->obd_namespace, &res_id, flags,
                                      opaque);
}

static int osc_join_lru(struct obd_export *exp,
                        struct lov_stripe_md *lsm, int join)
{
        struct obd_device *obd = class_exp2obd(exp);
        struct ldlm_res_id res_id = { .name = {lsm->lsm_object_id} };

        return ldlm_cli_join_lru(obd->obd_namespace, &res_id, join);
}

static int osc_statfs_interpret(struct ptlrpc_request *req,
                                struct osc_async_args *aa, int rc)
{
        struct obd_statfs *msfs;
        ENTRY;

        if (rc != 0)
                GOTO(out, rc);

        msfs = lustre_swab_repbuf(req, REPLY_REC_OFF, sizeof(*msfs),
                                  lustre_swab_obd_statfs);
        if (msfs == NULL) {
                CERROR("Can't unpack obd_statfs\n");
                GOTO(out, rc = -EPROTO);
        }

        memcpy(aa->aa_oi->oi_osfs, msfs, sizeof(*msfs));
out:
        rc = aa->aa_oi->oi_cb_up(aa->aa_oi, rc);
        RETURN(rc);
}

static int osc_statfs_async(struct obd_device *obd, struct obd_info *oinfo,
                            __u64 max_age, struct ptlrpc_request_set *rqset)
{
        struct ptlrpc_request *req;
        struct osc_async_args *aa;
        int size[2] = { sizeof(struct ptlrpc_body), sizeof(*oinfo->oi_osfs) };
        ENTRY;

        /* We could possibly pass max_age in the request (as an absolute
         * timestamp or a "seconds.usec ago") so the target can avoid doing
         * extra calls into the filesystem if that isn't necessary (e.g.
         * during mount that would help a bit).  Having relative timestamps
         * is not so great if request processing is slow, while absolute
         * timestamps are not ideal because they need time synchronization. */
        req = ptlrpc_prep_req(obd->u.cli.cl_import, LUSTRE_OST_VERSION,
                              OST_STATFS, 1, NULL, NULL);
        if (!req)
                RETURN(-ENOMEM);

        ptlrpc_req_set_repsize(req, 2, size);
        req->rq_request_portal = OST_CREATE_PORTAL; //XXX FIXME bug 249

        req->rq_interpret_reply = osc_statfs_interpret;
        CLASSERT(sizeof(*aa) <= sizeof(req->rq_async_args));
        aa = (struct osc_async_args *)&req->rq_async_args;
        aa->aa_oi = oinfo;

        ptlrpc_set_add_req(rqset, req);
        RETURN(0);
}

static int osc_statfs(struct obd_device *obd, struct obd_statfs *osfs,
                      __u64 max_age)
{
        struct obd_statfs *msfs;
        struct ptlrpc_request *req;
        int rc, size[2] = { sizeof(struct ptlrpc_body), sizeof(*osfs) };
        ENTRY;

        /* We could possibly pass max_age in the request (as an absolute
         * timestamp or a "seconds.usec ago") so the target can avoid doing
         * extra calls into the filesystem if that isn't necessary (e.g.
         * during mount that would help a bit).  Having relative timestamps
         * is not so great if request processing is slow, while absolute
         * timestamps are not ideal because they need time synchronization. */
        req = ptlrpc_prep_req(obd->u.cli.cl_import, LUSTRE_OST_VERSION,
                              OST_STATFS, 1, NULL, NULL);
        if (!req)
                RETURN(-ENOMEM);

        ptlrpc_req_set_repsize(req, 2, size);
        req->rq_request_portal = OST_CREATE_PORTAL; //XXX FIXME bug 249

        rc = ptlrpc_queue_wait(req);
        if (rc)
                GOTO(out, rc);

        msfs = lustre_swab_repbuf(req, REPLY_REC_OFF, sizeof(*msfs),
                                  lustre_swab_obd_statfs);
        if (msfs == NULL) {
                CERROR("Can't unpack obd_statfs\n");
                GOTO(out, rc = -EPROTO);
        }

        memcpy(osfs, msfs, sizeof(*osfs));

        EXIT;
 out:
        ptlrpc_req_finished(req);
        return rc;
}

/* Retrieve object striping information.
 *
 * @lmmu is a pointer to an in-core struct with lmm_ost_count indicating
 * the maximum number of OST indices which will fit in the user buffer.
 * lmm_magic must be LOV_MAGIC (we only use 1 slot here).
 */
static int osc_getstripe(struct lov_stripe_md *lsm, struct lov_user_md *lump)
{
        struct lov_user_md lum, *lumk;
        int rc = 0, lum_size;
        ENTRY;

        if (!lsm)
                RETURN(-ENODATA);

        if (copy_from_user(&lum, lump, sizeof(lum)))
                RETURN(-EFAULT);

        if (lum.lmm_magic != LOV_USER_MAGIC)
                RETURN(-EINVAL);

        if (lum.lmm_stripe_count > 0) {
                lum_size = sizeof(lum) + sizeof(lum.lmm_objects[0]);
                OBD_ALLOC(lumk, lum_size);
                if (!lumk)
                        RETURN(-ENOMEM);

                lumk->lmm_objects[0].l_object_id = lsm->lsm_object_id;
        } else {
                lum_size = sizeof(lum);
                lumk = &lum;
        }

        lumk->lmm_object_id = lsm->lsm_object_id;
        lumk->lmm_stripe_count = 1;

        if (copy_to_user(lump, lumk, lum_size))
                rc = -EFAULT;

        if (lumk != &lum)
                OBD_FREE(lumk, lum_size);

        RETURN(rc);
}


static int osc_iocontrol(unsigned int cmd, struct obd_export *exp, int len,
                         void *karg, void *uarg)
{
        struct obd_device *obd = exp->exp_obd;
        struct obd_ioctl_data *data = karg;
        int err = 0;
        ENTRY;

#if (LINUX_VERSION_CODE < KERNEL_VERSION(2,5,0))
        MOD_INC_USE_COUNT;
#else
        if (!try_module_get(THIS_MODULE)) {
                CERROR("Can't get module. Is it alive?");
                return -EINVAL;
        }
#endif
        switch (cmd) {
        case OBD_IOC_LOV_GET_CONFIG: {
                char *buf;
                struct lov_desc *desc;
                struct obd_uuid uuid;

                buf = NULL;
                len = 0;
                if (obd_ioctl_getdata(&buf, &len, (void *)uarg))
                        GOTO(out, err = -EINVAL);

                data = (struct obd_ioctl_data *)buf;

                if (sizeof(*desc) > data->ioc_inllen1) {
                        obd_ioctl_freedata(buf, len);
                        GOTO(out, err = -EINVAL);
                }

                if (data->ioc_inllen2 < sizeof(uuid)) {
                        obd_ioctl_freedata(buf, len);
                        GOTO(out, err = -EINVAL);
                }

                desc = (struct lov_desc *)data->ioc_inlbuf1;
                desc->ld_tgt_count = 1;
                desc->ld_active_tgt_count = 1;
                desc->ld_default_stripe_count = 1;
                desc->ld_default_stripe_size = 0;
                desc->ld_default_stripe_offset = 0;
                desc->ld_pattern = 0;
                memcpy(&desc->ld_uuid, &obd->obd_uuid, sizeof(uuid));

                memcpy(data->ioc_inlbuf2, &obd->obd_uuid, sizeof(uuid));

                err = copy_to_user((void *)uarg, buf, len);
                if (err)
                        err = -EFAULT;
                obd_ioctl_freedata(buf, len);
                GOTO(out, err);
        }
        case LL_IOC_LOV_SETSTRIPE:
                err = obd_alloc_memmd(exp, karg);
                if (err > 0)
                        err = 0;
                GOTO(out, err);
        case LL_IOC_LOV_GETSTRIPE:
                err = osc_getstripe(karg, uarg);
                GOTO(out, err);
        case OBD_IOC_CLIENT_RECOVER:
                err = ptlrpc_recover_import(obd->u.cli.cl_import,
                                            data->ioc_inlbuf1);
                if (err > 0)
                        err = 0;
                GOTO(out, err);
        case IOC_OSC_SET_ACTIVE:
                err = ptlrpc_set_import_active(obd->u.cli.cl_import,
                                               data->ioc_offset);
                GOTO(out, err);
        case OBD_IOC_POLL_QUOTACHECK:
                err = lquota_poll_check(quota_interface, exp,
                                        (struct if_quotacheck *)karg);
                GOTO(out, err);
        default:
                CDEBUG(D_INODE, "unrecognised ioctl %#x by %s\n",
                       cmd, cfs_curproc_comm());
                GOTO(out, err = -ENOTTY);
        }
out:
#if (LINUX_VERSION_CODE < KERNEL_VERSION(2,5,0))
        MOD_DEC_USE_COUNT;
#else
        module_put(THIS_MODULE);
#endif
        return err;
}

static int osc_get_info(struct obd_export *exp, obd_count keylen,
                        void *key, __u32 *vallen, void *val)
{
        ENTRY;
        if (!vallen || !val)
                RETURN(-EFAULT);

        if (keylen > strlen("lock_to_stripe") &&
            strcmp(key, "lock_to_stripe") == 0) {
                __u32 *stripe = val;
                *vallen = sizeof(*stripe);
                *stripe = 0;
                RETURN(0);
        } else if (keylen >= strlen("last_id") && strcmp(key, "last_id") == 0) {
                struct ptlrpc_request *req;
                obd_id *reply;
                char *bufs[2] = { NULL, key };
                int rc, size[2] = { sizeof(struct ptlrpc_body), keylen };

                req = ptlrpc_prep_req(class_exp2cliimp(exp), LUSTRE_OST_VERSION,
                                      OST_GET_INFO, 2, size, bufs);
                if (req == NULL)
                        RETURN(-ENOMEM);

                size[REPLY_REC_OFF] = *vallen;
                ptlrpc_req_set_repsize(req, 2, size);
                rc = ptlrpc_queue_wait(req);
                if (rc)
                        GOTO(out, rc);

                reply = lustre_swab_repbuf(req, REPLY_REC_OFF, sizeof(*reply),
                                           lustre_swab_ost_last_id);
                if (reply == NULL) {
                        CERROR("Can't unpack OST last ID\n");
                        GOTO(out, rc = -EPROTO);
                }
                *((obd_id *)val) = *reply;
        out:
                ptlrpc_req_finished(req);
                RETURN(rc);
        }
        RETURN(-EINVAL);
}

static int osc_setinfo_mds_conn_interpret(struct ptlrpc_request *req,
                                          void *aa, int rc)
{
        struct llog_ctxt *ctxt;
        struct obd_import *imp = req->rq_import;
        ENTRY;

        if (rc != 0)
                RETURN(rc);

        ctxt = llog_get_context(imp->imp_obd, LLOG_MDS_OST_ORIG_CTXT);
        if (ctxt) {
                if (rc == 0)
                        rc = llog_initiator_connect(ctxt);
                else
                        CERROR("cannot establish connection for "
                               "ctxt %p: %d\n", ctxt, rc);
        }

        imp->imp_server_timeout = 1;
        CDEBUG(D_HA, "pinging OST %s\n", obd2cli_tgt(imp->imp_obd));
        imp->imp_pingable = 1;

        RETURN(rc);
}

static int osc_set_info_async(struct obd_export *exp, obd_count keylen,
                              void *key, obd_count vallen, void *val,
                              struct ptlrpc_request_set *set)
{
        struct ptlrpc_request *req;
        struct obd_device  *obd = exp->exp_obd;
        struct obd_import *imp = class_exp2cliimp(exp);
        int size[3] = { sizeof(struct ptlrpc_body), keylen, vallen };
        char *bufs[3] = { NULL, key, val };
        ENTRY;

        OBD_FAIL_TIMEOUT(OBD_FAIL_OSC_SHUTDOWN, 10);

        if (KEY_IS(KEY_NEXT_ID)) {
                if (vallen != sizeof(obd_id))
                        RETURN(-EINVAL);
                obd->u.cli.cl_oscc.oscc_next_id = *((obd_id*)val) + 1;
                CDEBUG(D_HA, "%s: set oscc_next_id = "LPU64"\n",
                       exp->exp_obd->obd_name,
                       obd->u.cli.cl_oscc.oscc_next_id);

                RETURN(0);
        }

        if (KEY_IS("unlinked")) {
                struct osc_creator *oscc = &obd->u.cli.cl_oscc;
                spin_lock(&oscc->oscc_lock);
                oscc->oscc_flags &= ~OSCC_FLAG_NOSPC;
                spin_unlock(&oscc->oscc_lock);
                RETURN(0);
        }

        if (KEY_IS(KEY_INIT_RECOV)) {
                if (vallen != sizeof(int))
                        RETURN(-EINVAL);
                imp->imp_initial_recov = *(int *)val;
                CDEBUG(D_HA, "%s: set imp_initial_recov = %d\n",
                       exp->exp_obd->obd_name,
                       imp->imp_initial_recov);
                RETURN(0);
        }

        if (KEY_IS("checksum")) {
                if (vallen != sizeof(int))
                        RETURN(-EINVAL);
                exp->exp_obd->u.cli.cl_checksum = (*(int *)val) ? 1 : 0;
                RETURN(0);
        }

        if (!set)
                RETURN(-EINVAL);

        /* We pass all other commands directly to OST. Since nobody calls osc
           methods directly and everybody is supposed to go through LOV, we
           assume lov checked invalid values for us.
           The only recognised values so far are evict_by_nid and mds_conn.
           Even if something bad goes through, we'd get a -EINVAL from OST
           anyway. */

        req = ptlrpc_prep_req(imp, LUSTRE_OST_VERSION, OST_SET_INFO, 3, size,
                              bufs);
        if (req == NULL)
                RETURN(-ENOMEM);

        if (KEY_IS("mds_conn"))
                req->rq_interpret_reply = osc_setinfo_mds_conn_interpret;

        ptlrpc_req_set_repsize(req, 1, NULL);
        ptlrpc_set_add_req(set, req);
        ptlrpc_check_set(set);

        RETURN(0);
}


static struct llog_operations osc_size_repl_logops = {
        lop_cancel: llog_obd_repl_cancel
};

static struct llog_operations osc_mds_ost_orig_logops;
static int osc_llog_init(struct obd_device *obd, struct obd_device *tgt,
                         int count, struct llog_catid *catid, 
                         struct obd_uuid *uuid)
{
        int rc;
        ENTRY;

        spin_lock(&obd->obd_dev_lock);
        if (osc_mds_ost_orig_logops.lop_setup != llog_obd_origin_setup) {
                osc_mds_ost_orig_logops = llog_lvfs_ops;
                osc_mds_ost_orig_logops.lop_setup = llog_obd_origin_setup;
                osc_mds_ost_orig_logops.lop_cleanup = llog_obd_origin_cleanup;
                osc_mds_ost_orig_logops.lop_add = llog_obd_origin_add;
                osc_mds_ost_orig_logops.lop_connect = llog_origin_connect;
        }
        spin_unlock(&obd->obd_dev_lock);

        rc = llog_setup(obd, LLOG_MDS_OST_ORIG_CTXT, tgt, count,
                        &catid->lci_logid, &osc_mds_ost_orig_logops);
        if (rc) {
                CERROR("failed LLOG_MDS_OST_ORIG_CTXT\n");
                GOTO (out, rc);
        }

        rc = llog_setup(obd, LLOG_SIZE_REPL_CTXT, tgt, count, NULL,
                        &osc_size_repl_logops);
        if (rc) 
                CERROR("failed LLOG_SIZE_REPL_CTXT\n");
out:
        if (rc) {
                CERROR("osc '%s' tgt '%s' cnt %d catid %p rc=%d\n", 
                       obd->obd_name, tgt->obd_name, count, catid, rc);
                CERROR("logid "LPX64":0x%x\n",
                       catid->lci_logid.lgl_oid, catid->lci_logid.lgl_ogen);
        }
        RETURN(rc);
}

static int osc_llog_finish(struct obd_device *obd, int count)
{
        struct llog_ctxt *ctxt;
        int rc = 0, rc2 = 0;
        ENTRY;

        ctxt = llog_get_context(obd, LLOG_MDS_OST_ORIG_CTXT);
        if (ctxt)
                rc = llog_cleanup(ctxt);

        ctxt = llog_get_context(obd, LLOG_SIZE_REPL_CTXT);
        if (ctxt)
                rc2 = llog_cleanup(ctxt);
        if (!rc)
                rc = rc2;

        RETURN(rc);
}

static int osc_reconnect(struct obd_export *exp, struct obd_device *obd,
                         struct obd_uuid *cluuid,
                         struct obd_connect_data *data)
{
        struct client_obd *cli = &obd->u.cli;

        if (data != NULL && (data->ocd_connect_flags & OBD_CONNECT_GRANT)) {
                long lost_grant;

                client_obd_list_lock(&cli->cl_loi_list_lock);
                data->ocd_grant = cli->cl_avail_grant ?:
                                2 * cli->cl_max_pages_per_rpc << CFS_PAGE_SHIFT;
                lost_grant = cli->cl_lost_grant;
                cli->cl_lost_grant = 0;
                client_obd_list_unlock(&cli->cl_loi_list_lock);

                CDEBUG(D_CACHE, "request ocd_grant: %d cl_avail_grant: %ld "
                       "cl_lost_grant: %ld\n", data->ocd_grant,
                       cli->cl_avail_grant, lost_grant);
                CDEBUG(D_RPCTRACE, "ocd_connect_flags: "LPX64" ocd_version: %d"
                       " ocd_grant: %d\n", data->ocd_connect_flags,
                       data->ocd_version, data->ocd_grant);
        }

        RETURN(0);
}

static int osc_disconnect(struct obd_export *exp)
{
        struct obd_device *obd = class_exp2obd(exp);
        struct llog_ctxt *ctxt = llog_get_context(obd, LLOG_SIZE_REPL_CTXT);
        int rc;

        if (obd->u.cli.cl_conn_count == 1)
                /* flush any remaining cancel messages out to the target */
                llog_sync(ctxt, exp);

        rc = client_disconnect_export(exp);
        return rc;
}

static int osc_import_event(struct obd_device *obd,
                            struct obd_import *imp,
                            enum obd_import_event event)
{
        struct client_obd *cli;
        int rc = 0;

        ENTRY;
        LASSERT(imp->imp_obd == obd);

        switch (event) {
        case IMP_EVENT_DISCON: {
                /* Only do this on the MDS OSC's */
                if (imp->imp_server_timeout) {
                        struct osc_creator *oscc = &obd->u.cli.cl_oscc;

                        spin_lock(&oscc->oscc_lock);
                        oscc->oscc_flags |= OSCC_FLAG_RECOVERING;
                        spin_unlock(&oscc->oscc_lock);
                }

                break;
        }
        case IMP_EVENT_INACTIVE: {
                rc = obd_notify_observer(obd, obd, OBD_NOTIFY_INACTIVE, NULL);
                break;
        }
        case IMP_EVENT_INVALIDATE: {
                struct ldlm_namespace *ns = obd->obd_namespace;

                /* Reset grants */
                cli = &obd->u.cli;
                client_obd_list_lock(&cli->cl_loi_list_lock);
                cli->cl_avail_grant = 0;
                cli->cl_lost_grant = 0;
                /* all pages go to failing rpcs due to the invalid import */
                osc_check_rpcs(cli);
                client_obd_list_unlock(&cli->cl_loi_list_lock);

                ldlm_namespace_cleanup(ns, LDLM_FL_LOCAL_ONLY);

                break;
        }
        case IMP_EVENT_ACTIVE: {
                /* Only do this on the MDS OSC's */
                if (imp->imp_server_timeout) {
                        struct osc_creator *oscc = &obd->u.cli.cl_oscc;

                        spin_lock(&oscc->oscc_lock);
                        oscc->oscc_flags &= ~OSCC_FLAG_NOSPC;
                        spin_unlock(&oscc->oscc_lock);
                }
                rc = obd_notify_observer(obd, obd, OBD_NOTIFY_ACTIVE, NULL);
                break;
        }
        case IMP_EVENT_OCD: {
                struct obd_connect_data *ocd = &imp->imp_connect_data;

                if (ocd->ocd_connect_flags & OBD_CONNECT_GRANT)
                        osc_init_grant(&obd->u.cli, ocd);

                /* See bug 7198 */
                if (ocd->ocd_connect_flags & OBD_CONNECT_REQPORTAL)
                        imp->imp_client->cli_request_portal =OST_REQUEST_PORTAL;

                rc = obd_notify_observer(obd, obd, OBD_NOTIFY_OCD, NULL);
                break;
        }
        default:
                CERROR("Unknown import event %d\n", event);
                LBUG();
        }
        RETURN(rc);
}

int osc_setup(struct obd_device *obd, obd_count len, void *buf)
{
        int rc;
        ENTRY;

        ENTRY;
        rc = ptlrpcd_addref();
        if (rc)
                RETURN(rc);

        rc = client_obd_setup(obd, len, buf);
        if (rc) {
                ptlrpcd_decref();
        } else {
                struct lprocfs_static_vars lvars;
                struct client_obd *cli = &obd->u.cli;

                lprocfs_init_vars(osc, &lvars);
                if (lprocfs_obd_setup(obd, lvars.obd_vars) == 0) {
                        lproc_osc_attach_seqstat(obd);
                        ptlrpc_lprocfs_register_obd(obd);
                }

                oscc_init(obd);
                /* We need to allocate a few requests more, because
                   brw_interpret_oap tries to create new requests before freeing
                   previous ones. Ideally we want to have 2x max_rpcs_in_flight
                   reserved, but I afraid that might be too much wasted RAM
                   in fact, so 2 is just my guess and still should work. */
                cli->cl_import->imp_rq_pool =
                        ptlrpc_init_rq_pool(cli->cl_max_rpcs_in_flight + 2,
                                            OST_MAXREQSIZE,
                                            ptlrpc_add_rqs_to_pool);
        }

        RETURN(rc);
}

static int osc_precleanup(struct obd_device *obd, enum obd_cleanup_stage stage)
{
        int rc = 0;
        ENTRY;

        switch (stage) {
        case OBD_CLEANUP_EARLY: {
                struct obd_import *imp;
                imp = obd->u.cli.cl_import;
                CDEBUG(D_HA, "Deactivating import %s\n", obd->obd_name);
                /* ptlrpc_abort_inflight to stop an mds_lov_synchronize */
                ptlrpc_deactivate_import(imp);
                break;
        }
        case OBD_CLEANUP_EXPORTS: {
                /* If we set up but never connected, the
                   client import will not have been cleaned. */
                if (obd->u.cli.cl_import) {
                        struct obd_import *imp;
                        imp = obd->u.cli.cl_import;
                        CDEBUG(D_CONFIG, "%s: client import never connected\n",
                               obd->obd_name);
                        ptlrpc_invalidate_import(imp);
                        ptlrpc_free_rq_pool(imp->imp_rq_pool);
                        class_destroy_import(imp);
                        obd->u.cli.cl_import = NULL;
                }
                break;
        }
        case OBD_CLEANUP_SELF_EXP:
                rc = obd_llog_finish(obd, 0);
                if (rc != 0)
                        CERROR("failed to cleanup llogging subsystems\n");
                break;
        case OBD_CLEANUP_OBD:
                break;
        }
        RETURN(rc);
}

int osc_cleanup(struct obd_device *obd)
{
        struct osc_creator *oscc = &obd->u.cli.cl_oscc;
        int rc;

        ENTRY;
        ptlrpc_lprocfs_unregister_obd(obd);
        lprocfs_obd_cleanup(obd);

        spin_lock(&oscc->oscc_lock);
        oscc->oscc_flags &= ~OSCC_FLAG_RECOVERING;
        oscc->oscc_flags |= OSCC_FLAG_EXITING;
        spin_unlock(&oscc->oscc_lock);

        /* free memory of osc quota cache */
        lquota_cleanup(quota_interface, obd);

        rc = client_obd_cleanup(obd);

        ptlrpcd_decref();
        RETURN(rc);
}

static int osc_process_config(struct obd_device *obd, obd_count len, void *buf)
{
        struct lustre_cfg *lcfg = buf;
        struct lprocfs_static_vars lvars;
        int rc = 0;

        lprocfs_init_vars(osc, &lvars);

        rc = class_process_proc_param(PARAM_OSC, lvars.obd_vars, lcfg, obd);
        return(rc);
}

struct obd_ops osc_obd_ops = {
        .o_owner                = THIS_MODULE,
        .o_setup                = osc_setup,
        .o_precleanup           = osc_precleanup,
        .o_cleanup              = osc_cleanup,
        .o_add_conn             = client_import_add_conn,
        .o_del_conn             = client_import_del_conn,
        .o_connect              = client_connect_import,
        .o_reconnect            = osc_reconnect,
        .o_disconnect           = osc_disconnect,
        .o_statfs               = osc_statfs,
        .o_statfs_async         = osc_statfs_async,
        .o_packmd               = osc_packmd,
        .o_unpackmd             = osc_unpackmd,
        .o_create               = osc_create,
        .o_destroy              = osc_destroy,
        .o_getattr              = osc_getattr,
        .o_getattr_async        = osc_getattr_async,
        .o_setattr              = osc_setattr,
        .o_setattr_async        = osc_setattr_async,
        .o_brw                  = osc_brw,
        .o_brw_async            = osc_brw_async,
        .o_prep_async_page      = osc_prep_async_page,
        .o_queue_async_io       = osc_queue_async_io,
        .o_set_async_flags      = osc_set_async_flags,
        .o_queue_group_io       = osc_queue_group_io,
        .o_trigger_group_io     = osc_trigger_group_io,
        .o_teardown_async_page  = osc_teardown_async_page,
        .o_punch                = osc_punch,
        .o_sync                 = osc_sync,
        .o_enqueue              = osc_enqueue,
        .o_match                = osc_match,
        .o_change_cbdata        = osc_change_cbdata,
        .o_cancel               = osc_cancel,
        .o_cancel_unused        = osc_cancel_unused,
        .o_join_lru             = osc_join_lru,
        .o_iocontrol            = osc_iocontrol,
        .o_get_info             = osc_get_info,
        .o_set_info_async       = osc_set_info_async,
        .o_import_event         = osc_import_event,
        .o_llog_init            = osc_llog_init,
        .o_llog_finish          = osc_llog_finish,
        .o_process_config       = osc_process_config,
};

int __init osc_init(void)
{
        struct lprocfs_static_vars lvars;
        int rc;
        ENTRY;

        lprocfs_init_vars(osc, &lvars);

        request_module("lquota");
        quota_interface = PORTAL_SYMBOL_GET(osc_quota_interface);
        lquota_init(quota_interface);
        init_obd_quota_ops(quota_interface, &osc_obd_ops);

        rc = class_register_type(&osc_obd_ops, lvars.module_vars,
                                 LUSTRE_OSC_NAME);
        if (rc) {
                if (quota_interface)
                        PORTAL_SYMBOL_PUT(osc_quota_interface);
                RETURN(rc);
        }

        RETURN(rc);
}

#ifdef __KERNEL__
static void /*__exit*/ osc_exit(void)
{
        lquota_exit(quota_interface);
        if (quota_interface)
                PORTAL_SYMBOL_PUT(osc_quota_interface);

        class_unregister_type(LUSTRE_OSC_NAME);
}

MODULE_AUTHOR("Cluster File Systems, Inc. <info@clusterfs.com>");
MODULE_DESCRIPTION("Lustre Object Storage Client (OSC)");
MODULE_LICENSE("GPL");

cfs_module(osc, LUSTRE_VERSION_STRING, osc_init, osc_exit);
#endif