lustre/ptlrpc/niobuf.c

   1 /* -*- mode: c; c-basic-offset: 8; indent-tabs-mode: nil; -*-
   2  * vim:expandtab:shiftwidth=8:tabstop=8:
   3  *
   4  *  Copyright (C) 2002 Cluster File Systems, Inc.
   5  *
   6  *   This file is part of Lustre, http://www.lustre.org.
   7  *
   8  *   Lustre is free software; you can redistribute it and/or
   9  *   modify it under the terms of version 2 of the GNU General Public
  10  *   License as published by the Free Software Foundation.
  11  *
  12  *   Lustre is distributed in the hope that it will be useful,
  13  *   but WITHOUT ANY WARRANTY; without even the implied warranty of
  14  *   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  15  *   GNU General Public License for more details.
  16  *
  17  *   You should have received a copy of the GNU General Public License
  18  *   along with Lustre; if not, write to the Free Software
  19  *   Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
  20  *
  21  */
  22
  23 #define DEBUG_SUBSYSTEM S_RPC
  24
  25 #include <linux/obd_support.h>
  26 #include <linux/lustre_net.h>
  27 #include <linux/lustre_lib.h>
  28 #include <linux/obd.h>
  29
  30 extern ptl_handle_eq_t request_out_eq, reply_in_eq, reply_out_eq,
  31         bulk_source_eq, bulk_sink_eq;
  32
  33 static int ptl_send_buf(struct ptlrpc_request *request,
  34                         struct ptlrpc_connection *conn, int portal)
  35 {
  36         int rc;
  37         ptl_process_id_t remote_id;
  38         ptl_handle_md_t md_h;
  39
  40         request->rq_req_md.user_ptr = request;
  41
  42         switch (request->rq_type) {
  43         case PTL_RPC_MSG_REQUEST:
  44                 request->rq_reqmsg->type = HTON__u32(request->rq_type);
  45                 request->rq_req_md.start = request->rq_reqmsg;
  46                 request->rq_req_md.length = request->rq_reqlen;
  47                 request->rq_req_md.eventq = request_out_eq;
  48                 break;
  49         case PTL_RPC_MSG_REPLY:
  50                 request->rq_repmsg->type = HTON__u32(request->rq_type);
  51                 request->rq_req_md.start = request->rq_repmsg;
  52                 request->rq_req_md.length = request->rq_replen;
  53                 request->rq_req_md.eventq = reply_out_eq;
  54                 break;
  55         default:
  56                 LBUG();
  57                 return -1; /* notreached */
  58         }
  59         request->rq_req_md.threshold = 1;
  60         request->rq_req_md.options = PTL_MD_OP_PUT;
  61         request->rq_req_md.user_ptr = request;
  62
  63         rc = PtlMDBind(conn->c_peer.peer_ni, request->rq_req_md, &md_h);
  64         if (rc != 0) {
  65                 CERROR("PtlMDBind failed: %d\n", rc);
  66                 LBUG();
  67                 return rc;
  68         }
  69
  70         remote_id.nid = conn->c_peer.peer_nid;
  71         remote_id.pid = 0;
  72
  73         CDEBUG(D_NET, "Sending %d bytes to portal %d, xid "LPD64"\n",
  74                request->rq_req_md.length, portal, request->rq_xid);
  75
  76         if (!portal)
  77                 LBUG();
  78         rc = PtlPut(md_h, PTL_NOACK_REQ, remote_id, portal, 0, request->rq_xid,
  79                     0, 0);
  80         if (rc != PTL_OK) {
  81                 CERROR("PtlPut("LPU64", %d, "LPD64") failed: %d\n", remote_id.nid,
  82                        portal, request->rq_xid, rc);
  83                 PtlMDUnlink(md_h);
  84         }
  85
  86         return rc;
  87 }
  88
  89 static inline struct iovec *
  90 ptlrpc_get_bulk_iov (struct ptlrpc_bulk_desc *desc)
  91 {
  92         struct iovec *iov;
  93
  94         if (desc->bd_page_count <= sizeof (desc->bd_iov)/sizeof (struct iovec))
  95                 return (desc->bd_iov);
  96
  97         OBD_ALLOC (iov, desc->bd_page_count * sizeof (struct iovec));
  98         if (iov == NULL)
  99                 LBUG();
 100
 101         return (iov);
 102 }
 103
 104 static inline void
 105 ptlrpc_put_bulk_iov (struct ptlrpc_bulk_desc *desc, struct iovec *iov)
 106 {
 107         if (desc->bd_page_count <= sizeof (desc->bd_iov)/sizeof (struct iovec))
 108                 return;
 109
 110         OBD_FREE (iov, desc->bd_page_count * sizeof (struct iovec));
 111 }
 112
 113 int ptlrpc_send_bulk(struct ptlrpc_bulk_desc *desc)
 114 {
 115         int rc;
 116         struct list_head *tmp, *next;
 117         ptl_process_id_t remote_id;
 118         __u32 xid = 0;
 119         struct iovec *iov;
 120         ENTRY;
 121
 122         iov = ptlrpc_get_bulk_iov (desc);
 123         if (iov == NULL)
 124                 RETURN (-ENOMEM);
 125
 126         desc->bd_md.start = iov;
 127         desc->bd_md.niov = 0;
 128         desc->bd_md.length = 0;
 129         desc->bd_md.eventq = bulk_source_eq;
 130         desc->bd_md.threshold = 2; /* SENT and ACK */
 131         desc->bd_md.options = PTL_MD_OP_PUT | PTL_MD_IOV;
 132         desc->bd_md.user_ptr = desc;
 133
 134         atomic_set (&desc->bd_source_callback_count, 2);
 135
 136         list_for_each_safe(tmp, next, &desc->bd_page_list) {
 137                 struct ptlrpc_bulk_page *bulk;
 138                 bulk = list_entry(tmp, struct ptlrpc_bulk_page, bp_link);
 139
 140                 LASSERT (desc->bd_md.niov < desc->bd_page_count);
 141
 142                 if (desc->bd_md.niov == 0)
 143                         xid = bulk->bp_xid;
 144                 LASSERT (xid == bulk->bp_xid);   /* should all be the same */
 145
 146                 iov[desc->bd_md.niov].iov_base = bulk->bp_buf;
 147                 iov[desc->bd_md.niov].iov_len = bulk->bp_buflen;
 148                 desc->bd_md.niov++;
 149                 desc->bd_md.length += bulk->bp_buflen;
 150         }
 151
 152         LASSERT (desc->bd_md.niov == desc->bd_page_count);
 153         LASSERT (desc->bd_md.niov != 0);
 154
 155         rc = PtlMDBind(desc->bd_connection->c_peer.peer_ni, desc->bd_md,
 156                        &desc->bd_md_h);
 157
 158         ptlrpc_put_bulk_iov (desc, iov);        /* move down to reduce latency to send */
 159
 160         if (rc != PTL_OK) {
 161                 CERROR("PtlMDBind failed: %d\n", rc);
 162                 LBUG();
 163                 RETURN(rc);
 164         }
 165
 166         remote_id.nid = desc->bd_connection->c_peer.peer_nid;
 167         remote_id.pid = 0;
 168
 169         CDEBUG(D_NET, "Sending %u pages %u bytes to portal %d nid "LPX64" pid %d xid %d\n",
 170                desc->bd_md.niov, desc->bd_md.length,
 171                desc->bd_portal, remote_id.nid, remote_id.pid, xid);
 172
 173         rc = PtlPut(desc->bd_md_h, PTL_ACK_REQ, remote_id,
 174                     desc->bd_portal, 0, xid, 0, 0);
 175         if (rc != PTL_OK) {
 176                 CERROR("PtlPut("LPU64", %d, %d) failed: %d\n",
 177                        remote_id.nid, desc->bd_portal, xid, rc);
 178                 PtlMDUnlink(desc->bd_md_h);
 179                 LBUG();
 180                 RETURN(rc);
 181         }
 182
 183         RETURN(0);
 184 }
 185
 186 int ptlrpc_register_bulk(struct ptlrpc_bulk_desc *desc)
 187 {
 188         struct list_head *tmp, *next;
 189         int rc;
 190         __u32 xid = 0;
 191         struct iovec *iov;
 192         ptl_process_id_t source_id;
 193         ENTRY;
 194
 195         iov = ptlrpc_get_bulk_iov (desc);
 196         if (iov == NULL)
 197                 return (-ENOMEM);
 198
 199         desc->bd_md.start = iov;
 200         desc->bd_md.niov = 0;
 201         desc->bd_md.length = 0;
 202         desc->bd_md.threshold = 1;
 203         desc->bd_md.options = PTL_MD_OP_PUT | PTL_MD_IOV;
 204         desc->bd_md.user_ptr = desc;
 205         desc->bd_md.eventq = bulk_sink_eq;
 206
 207         list_for_each_safe(tmp, next, &desc->bd_page_list) {
 208                 struct ptlrpc_bulk_page *bulk;
 209                 bulk = list_entry(tmp, struct ptlrpc_bulk_page, bp_link);
 210
 211                 LASSERT (desc->bd_md.niov < desc->bd_page_count);
 212
 213                 if (desc->bd_md.niov == 0)
 214                         xid = bulk->bp_xid;
 215                 LASSERT (xid == bulk->bp_xid);   /* should all be the same */
 216
 217                 iov[desc->bd_md.niov].iov_base = bulk->bp_buf;
 218                 iov[desc->bd_md.niov].iov_len = bulk->bp_buflen;
 219                 desc->bd_md.niov++;
 220                 desc->bd_md.length += bulk->bp_buflen;
 221         }
 222
 223         LASSERT (desc->bd_md.niov == desc->bd_page_count);
 224         LASSERT (desc->bd_md.niov != 0);
 225
 226         source_id.nid = desc->bd_connection->c_peer.peer_nid;
 227         source_id.pid = PTL_PID_ANY;
 228
 229         rc = PtlMEAttach(desc->bd_connection->c_peer.peer_ni,
 230                          desc->bd_portal, source_id, xid, 0,
 231                          PTL_UNLINK, PTL_INS_AFTER, &desc->bd_me_h);
 232
 233         ptlrpc_put_bulk_iov (desc, iov);
 234
 235         if (rc != PTL_OK) {
 236                 CERROR("PtlMEAttach failed: %d\n", rc);
 237                 LBUG();
 238                 GOTO(cleanup, rc);
 239         }
 240
 241         rc = PtlMDAttach(desc->bd_me_h, desc->bd_md, PTL_UNLINK,
 242                          &desc->bd_md_h);
 243         if (rc != PTL_OK) {
 244                 CERROR("PtlMDAttach failed: %d\n", rc);
 245                 LBUG();
 246                 GOTO(cleanup, rc);
 247         }
 248
 249         CDEBUG(D_NET, "Setup bulk sink buffers: %u pages %u bytes, xid %u, "
 250                "portal %u\n", desc->bd_md.niov, desc->bd_md.length,
 251                xid, desc->bd_portal);
 252
 253         RETURN(0);
 254
 255  cleanup:
 256         ptlrpc_abort_bulk(desc);
 257
 258         return rc;
 259 }
 260
 261 int ptlrpc_abort_bulk(struct ptlrpc_bulk_desc *desc)
 262 {
 263         /* This should be safe: these handles are initialized to be
 264          * invalid in ptlrpc_prep_bulk() */
 265         PtlMDUnlink(desc->bd_md_h);
 266         PtlMEUnlink(desc->bd_me_h);
 267
 268         return 0;
 269 }
 270
 271 int ptlrpc_reply(struct ptlrpc_service *svc, struct ptlrpc_request *req)
 272 {
 273         if (req->rq_repmsg == NULL) {
 274                 CERROR("bad: someone called ptlrpc_reply when they meant "
 275                        "ptlrpc_error\n");
 276                 return -EINVAL;
 277         }
 278
 279         /* FIXME: we need to increment the count of handled events */
 280         if (req->rq_type != PTL_RPC_MSG_ERR)
 281                 req->rq_type = PTL_RPC_MSG_REPLY;
 282         //req->rq_repmsg->conn = req->rq_connection->c_remote_conn;
 283         //req->rq_repmsg->token = req->rq_connection->c_remote_token;
 284         req->rq_repmsg->status = HTON__u32(req->rq_status);
 285         return ptl_send_buf(req, req->rq_connection, svc->srv_rep_portal);
 286 }
 287
 288 int ptlrpc_error(struct ptlrpc_service *svc, struct ptlrpc_request *req)
 289 {
 290         int rc;
 291         ENTRY;
 292
 293         if (req->rq_repmsg) {
 294                 CERROR("req already has repmsg\n");
 295                 LBUG();
 296         }
 297
 298         rc = lustre_pack_msg(0, NULL, NULL, &req->rq_replen, &req->rq_repmsg);
 299         if (rc)
 300                 RETURN(rc);
 301
 302         req->rq_repmsg->type = HTON__u32(PTL_RPC_MSG_ERR);
 303
 304         rc = ptlrpc_reply(svc, req);
 305         RETURN(rc);
 306 }
 307
 308 int ptl_send_rpc(struct ptlrpc_request *request)
 309 {
 310         int rc;
 311         char *repbuf;
 312         ptl_process_id_t source_id;
 313
 314         ENTRY;
 315
 316         if (request->rq_type != PTL_RPC_MSG_REQUEST) {
 317                 CERROR("wrong packet type sent %d\n",
 318                        NTOH__u32(request->rq_reqmsg->type));
 319                 LBUG();
 320                 RETURN(EINVAL);
 321         }
 322         if (request->rq_replen == 0) {
 323                 CERROR("request->rq_replen is 0!\n");
 324                 RETURN(EINVAL);
 325         }
 326
 327         /* request->rq_repmsg is set only when the reply comes in, in
 328          * client_packet_callback() */
 329         if (request->rq_reply_md.start)
 330                 OBD_FREE(request->rq_reply_md.start, request->rq_replen);
 331
 332         OBD_ALLOC(repbuf, request->rq_replen);
 333         if (!repbuf) {
 334                 LBUG();
 335                 RETURN(ENOMEM);
 336         }
 337
 338         // down(&request->rq_client->cli_rpc_sem);
 339
 340         source_id.nid = request->rq_connection->c_peer.peer_nid;
 341         source_id.pid = PTL_PID_ANY;
 342
 343         rc = PtlMEAttach(request->rq_connection->c_peer.peer_ni,
 344                          request->rq_import->imp_client->cli_reply_portal,
 345                          source_id, request->rq_xid, 0, PTL_UNLINK,
 346                          PTL_INS_AFTER, &request->rq_reply_me_h);
 347         if (rc != PTL_OK) {
 348                 CERROR("PtlMEAttach failed: %d\n", rc);
 349                 LBUG();
 350                 GOTO(cleanup, rc);
 351         }
 352
 353         request->rq_reply_md.start = repbuf;
 354         request->rq_reply_md.length = request->rq_replen;
 355         request->rq_reply_md.threshold = 1;
 356         request->rq_reply_md.options = PTL_MD_OP_PUT;
 357         request->rq_reply_md.user_ptr = request;
 358         request->rq_reply_md.eventq = reply_in_eq;
 359
 360         rc = PtlMDAttach(request->rq_reply_me_h, request->rq_reply_md,
 361                          PTL_UNLINK, &request->rq_reply_md_h);
 362         if (rc != PTL_OK) {
 363                 CERROR("PtlMDAttach failed: %d\n", rc);
 364                 LBUG();
 365                 GOTO(cleanup2, rc);
 366         }
 367
 368         CDEBUG(D_NET, "Setup reply buffer: %u bytes, xid "LPU64", portal %u\n",
 369                request->rq_replen, request->rq_xid,
 370                request->rq_import->imp_client->cli_reply_portal);
 371
 372         rc = ptl_send_buf(request, request->rq_connection,
 373                           request->rq_import->imp_client->cli_request_portal);
 374         RETURN(rc);
 375
 376  cleanup2:
 377         PtlMEUnlink(request->rq_reply_me_h);
 378  cleanup:
 379         OBD_FREE(repbuf, request->rq_replen);
 380         // up(&request->rq_client->cli_rpc_sem);
 381
 382         return rc;
 383 }
 384
 385 void ptlrpc_link_svc_me(struct ptlrpc_request_buffer_desc *rqbd)
 386 {
 387         struct ptlrpc_service *service = rqbd->rqbd_service;
 388         static ptl_process_id_t match_id = {PTL_NID_ANY, PTL_PID_ANY};
 389         int rc;
 390         ptl_md_t dummy;
 391         ptl_handle_md_t md_h;
 392
 393         /* Attach the leading ME on which we build the ring */
 394         rc = PtlMEAttach(service->srv_self.peer_ni, service->srv_req_portal,
 395                          match_id, 0, ~0,
 396                          PTL_UNLINK, PTL_INS_AFTER, &rqbd->rqbd_me_h);
 397         if (rc != PTL_OK) {
 398                 CERROR("PtlMEAttach failed: %d\n", rc);
 399                 LBUG();
 400         }
 401
 402         dummy.start         = rqbd->rqbd_buffer;
 403         dummy.length        = service->srv_buf_size;
 404         dummy.max_offset    = service->srv_buf_size;
 405         dummy.threshold     = 1;
 406         dummy.options       = PTL_MD_OP_PUT;
 407         dummy.user_ptr      = rqbd;
 408         dummy.eventq        = service->srv_eq_h;
 409
 410         rc = PtlMDAttach(rqbd->rqbd_me_h, dummy, PTL_UNLINK, &md_h);
 411         if (rc != PTL_OK) {
 412                 /* cleanup */
 413                 CERROR("PtlMDAttach failed: %d\n", rc);
 414                 LBUG();
 415         }
 416 }