sys/rpc/clnt_dg.c

   1 /*      $NetBSD: clnt_dg.c,v 1.4 2000/07/14 08:40:41 fvdl Exp $ */
   2
   3 /*
   4  * Sun RPC is a product of Sun Microsystems, Inc. and is provided for
   5  * unrestricted use provided that this legend is included on all tape
   6  * media and as a part of the software program in whole or part.  Users
   7  * may copy or modify Sun RPC without charge, but are not authorized
   8  * to license or distribute it to anyone else except as part of a product or
   9  * program developed by the user.
  10  *
  11  * SUN RPC IS PROVIDED AS IS WITH NO WARRANTIES OF ANY KIND INCLUDING THE
  12  * WARRANTIES OF DESIGN, MERCHANTIBILITY AND FITNESS FOR A PARTICULAR
  13  * PURPOSE, OR ARISING FROM A COURSE OF DEALING, USAGE OR TRADE PRACTICE.
  14  *
  15  * Sun RPC is provided with no support and without any obligation on the
  16  * part of Sun Microsystems, Inc. to assist in its use, correction,
  17  * modification or enhancement.
  18  *
  19  * SUN MICROSYSTEMS, INC. SHALL HAVE NO LIABILITY WITH RESPECT TO THE
  20  * INFRINGEMENT OF COPYRIGHTS, TRADE SECRETS OR ANY PATENTS BY SUN RPC
  21  * OR ANY PART THEREOF.
  22  *
  23  * In no event will Sun Microsystems, Inc. be liable for any lost revenue
  24  * or profits or other special, indirect and consequential damages, even if
  25  * Sun has been advised of the possibility of such damages.
  26  *
  27  * Sun Microsystems, Inc.
  28  * 2550 Garcia Avenue
  29  * Mountain View, California  94043
  30  */
  31 /*
  32  * Copyright (c) 1986-1991 by Sun Microsystems Inc.
  33  */
  34
  35 #if defined(LIBC_SCCS) && !defined(lint)
  36 #ident  "@(#)clnt_dg.c  1.23    94/04/22 SMI"
  37 static char sccsid[] = "@(#)clnt_dg.c 1.19 89/03/16 Copyr 1988 Sun Micro";
  38 #endif
  39 #include <sys/cdefs.h>
  40 __FBSDID("$FreeBSD$");
  41
  42 /*
  43  * Implements a connectionless client side RPC.
  44  */
  45
  46 #include <sys/param.h>
  47 #include <sys/systm.h>
  48 #include <sys/kernel.h>
  49 #include <sys/lock.h>
  50 #include <sys/malloc.h>
  51 #include <sys/mbuf.h>
  52 #include <sys/mutex.h>
  53 #include <sys/pcpu.h>
  54 #include <sys/proc.h>
  55 #include <sys/socket.h>
  56 #include <sys/socketvar.h>
  57 #include <sys/time.h>
  58 #include <sys/uio.h>
  59
  60 #include <net/vnet.h>
  61
  62 #include <rpc/rpc.h>
  63 #include <rpc/rpc_com.h>
  64
  65
  66 #ifdef _FREEFALL_CONFIG
  67 /*
  68  * Disable RPC exponential back-off for FreeBSD.org systems.
  69  */
  70 #define RPC_MAX_BACKOFF         1 /* second */
  71 #else
  72 #define RPC_MAX_BACKOFF         30 /* seconds */
  73 #endif
  74
  75 static bool_t time_not_ok(struct timeval *);
  76 static enum clnt_stat clnt_dg_call(CLIENT *, struct rpc_callextra *,
  77     rpcproc_t, struct mbuf *, struct mbuf **, struct timeval);
  78 static void clnt_dg_geterr(CLIENT *, struct rpc_err *);
  79 static bool_t clnt_dg_freeres(CLIENT *, xdrproc_t, void *);
  80 static void clnt_dg_abort(CLIENT *);
  81 static bool_t clnt_dg_control(CLIENT *, u_int, void *);
  82 static void clnt_dg_close(CLIENT *);
  83 static void clnt_dg_destroy(CLIENT *);
  84 static int clnt_dg_soupcall(struct socket *so, void *arg, int waitflag);
  85
  86 static struct clnt_ops clnt_dg_ops = {
  87         .cl_call =      clnt_dg_call,
  88         .cl_abort =     clnt_dg_abort,
  89         .cl_geterr =    clnt_dg_geterr,
  90         .cl_freeres =   clnt_dg_freeres,
  91         .cl_close =     clnt_dg_close,
  92         .cl_destroy =   clnt_dg_destroy,
  93         .cl_control =   clnt_dg_control
  94 };
  95
  96 static const char mem_err_clnt_dg[] = "clnt_dg_create: out of memory";
  97
  98 /*
  99  * A pending RPC request which awaits a reply. Requests which have
 100  * received their reply will have cr_xid set to zero and cr_mrep to
 101  * the mbuf chain of the reply.
 102  */
 103 struct cu_request {
 104         TAILQ_ENTRY(cu_request) cr_link;
 105         CLIENT                  *cr_client;     /* owner */
 106         uint32_t                cr_xid;         /* XID of request */
 107         struct mbuf             *cr_mrep;       /* reply received by upcall */
 108         int                     cr_error;       /* any error from upcall */
 109         char                    cr_verf[MAX_AUTH_BYTES]; /* reply verf */
 110 };
 111
 112 TAILQ_HEAD(cu_request_list, cu_request);
 113
 114 #define MCALL_MSG_SIZE 24
 115
 116 /*
 117  * This structure is pointed to by the socket buffer's sb_upcallarg
 118  * member. It is separate from the client private data to facilitate
 119  * multiple clients sharing the same socket. The cs_lock mutex is used
 120  * to protect all fields of this structure, the socket's receive
 121  * buffer SOCKBUF_LOCK is used to ensure that exactly one of these
 122  * structures is installed on the socket.
 123  */
 124 struct cu_socket {
 125         struct mtx              cs_lock;
 126         int                     cs_refs;        /* Count of clients */
 127         struct cu_request_list  cs_pending;     /* Requests awaiting replies */
 128         int                     cs_upcallrefs;  /* Refcnt of upcalls in prog.*/
 129 };
 130
 131 static void clnt_dg_upcallsdone(struct socket *, struct cu_socket *);
 132
 133 /*
 134  * Private data kept per client handle
 135  */
 136 struct cu_data {
 137         int                     cu_threads;     /* # threads in clnt_vc_call */
 138         bool_t                  cu_closing;     /* TRUE if we are closing */
 139         bool_t                  cu_closed;      /* TRUE if we are closed */
 140         struct socket           *cu_socket;     /* connection socket */
 141         bool_t                  cu_closeit;     /* opened by library */
 142         struct sockaddr_storage cu_raddr;       /* remote address */
 143         int                     cu_rlen;
 144         struct timeval          cu_wait;        /* retransmit interval */
 145         struct timeval          cu_total;       /* total time for the call */
 146         struct rpc_err          cu_error;
 147         uint32_t                cu_xid;
 148         char                    cu_mcallc[MCALL_MSG_SIZE]; /* marshalled callmsg */
 149         size_t                  cu_mcalllen;
 150         size_t                  cu_sendsz;      /* send size */
 151         size_t                  cu_recvsz;      /* recv size */
 152         int                     cu_async;
 153         int                     cu_connect;     /* Use connect(). */
 154         int                     cu_connected;   /* Have done connect(). */
 155         const char              *cu_waitchan;
 156         int                     cu_waitflag;
 157         int                     cu_cwnd;        /* congestion window */
 158         int                     cu_sent;        /* number of in-flight RPCs */
 159         bool_t                  cu_cwnd_wait;
 160 };
 161
 162 #define CWNDSCALE       256
 163 #define MAXCWND         (32 * CWNDSCALE)
 164
 165 /*
 166  * Connection less client creation returns with client handle parameters.
 167  * Default options are set, which the user can change using clnt_control().
 168  * fd should be open and bound.
 169  * NB: The rpch->cl_auth is initialized to null authentication.
 170  *      Caller may wish to set this something more useful.
 171  *
 172  * sendsz and recvsz are the maximum allowable packet sizes that can be
 173  * sent and received. Normally they are the same, but they can be
 174  * changed to improve the program efficiency and buffer allocation.
 175  * If they are 0, use the transport default.
 176  *
 177  * If svcaddr is NULL, returns NULL.
 178  */
 179 CLIENT *
 180 clnt_dg_create(
 181         struct socket *so,
 182         struct sockaddr *svcaddr,       /* servers address */
 183         rpcprog_t program,              /* program number */
 184         rpcvers_t version,              /* version number */
 185         size_t sendsz,                  /* buffer recv size */
 186         size_t recvsz)                  /* buffer send size */
 187 {
 188         CLIENT *cl = NULL;              /* client handle */
 189         struct cu_data *cu = NULL;      /* private data */
 190         struct cu_socket *cs = NULL;
 191         struct sockbuf *sb;
 192         struct timeval now;
 193         struct rpc_msg call_msg;
 194         struct __rpc_sockinfo si;
 195         XDR xdrs;
 196
 197         if (svcaddr == NULL) {
 198                 rpc_createerr.cf_stat = RPC_UNKNOWNADDR;
 199                 return (NULL);
 200         }
 201
 202         CURVNET_SET(so->so_vnet);
 203         if (!__rpc_socket2sockinfo(so, &si)) {
 204                 rpc_createerr.cf_stat = RPC_TLIERROR;
 205                 rpc_createerr.cf_error.re_errno = 0;
 206                 CURVNET_RESTORE();
 207                 return (NULL);
 208         }
 209         CURVNET_RESTORE();
 210
 211         /*
 212          * Find the receive and the send size
 213          */
 214         sendsz = __rpc_get_t_size(si.si_af, si.si_proto, (int)sendsz);
 215         recvsz = __rpc_get_t_size(si.si_af, si.si_proto, (int)recvsz);
 216         if ((sendsz == 0) || (recvsz == 0)) {
 217                 rpc_createerr.cf_stat = RPC_TLIERROR; /* XXX */
 218                 rpc_createerr.cf_error.re_errno = 0;
 219                 return (NULL);
 220         }
 221
 222         cl = mem_alloc(sizeof (CLIENT));
 223
 224         /*
 225          * Should be multiple of 4 for XDR.
 226          */
 227         sendsz = ((sendsz + 3) / 4) * 4;
 228         recvsz = ((recvsz + 3) / 4) * 4;
 229         cu = mem_alloc(sizeof (*cu));
 230         cu->cu_threads = 0;
 231         cu->cu_closing = FALSE;
 232         cu->cu_closed = FALSE;
 233         (void) memcpy(&cu->cu_raddr, svcaddr, (size_t)svcaddr->sa_len);
 234         cu->cu_rlen = svcaddr->sa_len;
 235         /* Other values can also be set through clnt_control() */
 236         cu->cu_wait.tv_sec = 3; /* heuristically chosen */
 237         cu->cu_wait.tv_usec = 0;
 238         cu->cu_total.tv_sec = -1;
 239         cu->cu_total.tv_usec = -1;
 240         cu->cu_sendsz = sendsz;
 241         cu->cu_recvsz = recvsz;
 242         cu->cu_async = FALSE;
 243         cu->cu_connect = FALSE;
 244         cu->cu_connected = FALSE;
 245         cu->cu_waitchan = "rpcrecv";
 246         cu->cu_waitflag = 0;
 247         cu->cu_cwnd = MAXCWND / 2;
 248         cu->cu_sent = 0;
 249         cu->cu_cwnd_wait = FALSE;
 250         (void) getmicrotime(&now);
 251         cu->cu_xid = __RPC_GETXID(&now);
 252         call_msg.rm_xid = cu->cu_xid;
 253         call_msg.rm_call.cb_prog = program;
 254         call_msg.rm_call.cb_vers = version;
 255         xdrmem_create(&xdrs, cu->cu_mcallc, MCALL_MSG_SIZE, XDR_ENCODE);
 256         if (! xdr_callhdr(&xdrs, &call_msg)) {
 257                 rpc_createerr.cf_stat = RPC_CANTENCODEARGS;  /* XXX */
 258                 rpc_createerr.cf_error.re_errno = 0;
 259                 goto err2;
 260         }
 261         cu->cu_mcalllen = XDR_GETPOS(&xdrs);
 262
 263         /*
 264          * By default, closeit is always FALSE. It is users responsibility
 265          * to do a close on it, else the user may use clnt_control
 266          * to let clnt_destroy do it for him/her.
 267          */
 268         cu->cu_closeit = FALSE;
 269         cu->cu_socket = so;
 270         soreserve(so, 256*1024, 256*1024);
 271
 272         sb = &so->so_rcv;
 273         SOCKBUF_LOCK(&so->so_rcv);
 274 recheck_socket:
 275         if (sb->sb_upcall) {
 276                 if (sb->sb_upcall != clnt_dg_soupcall) {
 277                         SOCKBUF_UNLOCK(&so->so_rcv);
 278                         printf("clnt_dg_create(): socket already has an incompatible upcall\n");
 279                         goto err2;
 280                 }
 281                 cs = (struct cu_socket *) sb->sb_upcallarg;
 282                 mtx_lock(&cs->cs_lock);
 283                 cs->cs_refs++;
 284                 mtx_unlock(&cs->cs_lock);
 285         } else {
 286                 /*
 287                  * We are the first on this socket - allocate the
 288                  * structure and install it in the socket.
 289                  */
 290                 SOCKBUF_UNLOCK(&so->so_rcv);
 291                 cs = mem_alloc(sizeof(*cs));
 292                 SOCKBUF_LOCK(&so->so_rcv);
 293                 if (sb->sb_upcall) {
 294                         /*
 295                          * We have lost a race with some other client.
 296                          */
 297                         mem_free(cs, sizeof(*cs));
 298                         goto recheck_socket;
 299                 }
 300                 mtx_init(&cs->cs_lock, "cs->cs_lock", NULL, MTX_DEF);
 301                 cs->cs_refs = 1;
 302                 cs->cs_upcallrefs = 0;
 303                 TAILQ_INIT(&cs->cs_pending);
 304                 soupcall_set(so, SO_RCV, clnt_dg_soupcall, cs);
 305         }
 306         SOCKBUF_UNLOCK(&so->so_rcv);
 307
 308         cl->cl_refs = 1;
 309         cl->cl_ops = &clnt_dg_ops;
 310         cl->cl_private = (caddr_t)(void *)cu;
 311         cl->cl_auth = authnone_create();
 312         cl->cl_tp = NULL;
 313         cl->cl_netid = NULL;
 314         return (cl);
 315 err2:
 316         if (cl) {
 317                 mem_free(cl, sizeof (CLIENT));
 318                 if (cu)
 319                         mem_free(cu, sizeof (*cu));
 320         }
 321         return (NULL);
 322 }
 323
 324 static enum clnt_stat
 325 clnt_dg_call(
 326         CLIENT          *cl,            /* client handle */
 327         struct rpc_callextra *ext,      /* call metadata */
 328         rpcproc_t       proc,           /* procedure number */
 329         struct mbuf     *args,          /* pointer to args */
 330         struct mbuf     **resultsp,     /* pointer to results */
 331         struct timeval  utimeout)       /* seconds to wait before giving up */
 332 {
 333         struct cu_data *cu = (struct cu_data *)cl->cl_private;
 334         struct cu_socket *cs;
 335         struct rpc_timers *rt;
 336         AUTH *auth;
 337         struct rpc_err *errp;
 338         enum clnt_stat stat;
 339         XDR xdrs;
 340         struct rpc_msg reply_msg;
 341         bool_t ok;
 342         int retrans;                    /* number of re-transmits so far */
 343         int nrefreshes = 2;             /* number of times to refresh cred */
 344         struct timeval *tvp;
 345         int timeout;
 346         int retransmit_time;
 347         int next_sendtime, starttime, rtt, time_waited, tv = 0;
 348         struct sockaddr *sa;
 349         socklen_t salen;
 350         uint32_t xid = 0;
 351         struct mbuf *mreq = NULL, *results;
 352         struct cu_request *cr;
 353         int error;
 354
 355         cs = cu->cu_socket->so_rcv.sb_upcallarg;
 356         cr = malloc(sizeof(struct cu_request), M_RPC, M_WAITOK);
 357
 358         mtx_lock(&cs->cs_lock);
 359
 360         if (cu->cu_closing || cu->cu_closed) {
 361                 mtx_unlock(&cs->cs_lock);
 362                 free(cr, M_RPC);
 363                 return (RPC_CANTSEND);
 364         }
 365         cu->cu_threads++;
 366
 367         if (ext) {
 368                 auth = ext->rc_auth;
 369                 errp = &ext->rc_err;
 370         } else {
 371                 auth = cl->cl_auth;
 372                 errp = &cu->cu_error;
 373         }
 374
 375         cr->cr_client = cl;
 376         cr->cr_mrep = NULL;
 377         cr->cr_error = 0;
 378
 379         if (cu->cu_total.tv_usec == -1) {
 380                 tvp = &utimeout; /* use supplied timeout */
 381         } else {
 382                 tvp = &cu->cu_total; /* use default timeout */
 383         }
 384         if (tvp->tv_sec || tvp->tv_usec)
 385                 timeout = tvtohz(tvp);
 386         else
 387                 timeout = 0;
 388
 389         if (cu->cu_connect && !cu->cu_connected) {
 390                 mtx_unlock(&cs->cs_lock);
 391                 error = soconnect(cu->cu_socket,
 392                     (struct sockaddr *)&cu->cu_raddr, curthread);
 393                 mtx_lock(&cs->cs_lock);
 394                 if (error) {
 395                         errp->re_errno = error;
 396                         errp->re_status = stat = RPC_CANTSEND;
 397                         goto out;
 398                 }
 399                 cu->cu_connected = 1;
 400         }
 401         if (cu->cu_connected) {
 402                 sa = NULL;
 403                 salen = 0;
 404         } else {
 405                 sa = (struct sockaddr *)&cu->cu_raddr;
 406                 salen = cu->cu_rlen;
 407         }
 408         time_waited = 0;
 409         retrans = 0;
 410         if (ext && ext->rc_timers) {
 411                 rt = ext->rc_timers;
 412                 if (!rt->rt_rtxcur)
 413                         rt->rt_rtxcur = tvtohz(&cu->cu_wait);
 414                 retransmit_time = next_sendtime = rt->rt_rtxcur;
 415         } else {
 416                 rt = NULL;
 417                 retransmit_time = next_sendtime = tvtohz(&cu->cu_wait);
 418         }
 419
 420         starttime = ticks;
 421
 422 call_again:
 423         mtx_assert(&cs->cs_lock, MA_OWNED);
 424
 425         cu->cu_xid++;
 426         xid = cu->cu_xid;
 427
 428 send_again:
 429         mtx_unlock(&cs->cs_lock);
 430
 431         MGETHDR(mreq, M_WAIT, MT_DATA);
 432         KASSERT(cu->cu_mcalllen <= MHLEN, ("RPC header too big"));
 433         bcopy(cu->cu_mcallc, mreq->m_data, cu->cu_mcalllen);
 434         mreq->m_len = cu->cu_mcalllen;
 435
 436         /*
 437          * The XID is the first thing in the request.
 438          */
 439         *mtod(mreq, uint32_t *) = htonl(xid);
 440
 441         xdrmbuf_create(&xdrs, mreq, XDR_ENCODE);
 442
 443         if (cu->cu_async == TRUE && args == NULL)
 444                 goto get_reply;
 445
 446         if ((! XDR_PUTINT32(&xdrs, &proc)) ||
 447             (! AUTH_MARSHALL(auth, xid, &xdrs,
 448                 m_copym(args, 0, M_COPYALL, M_WAITOK)))) {
 449                 errp->re_status = stat = RPC_CANTENCODEARGS;
 450                 mtx_lock(&cs->cs_lock);
 451                 goto out;
 452         }
 453         mreq->m_pkthdr.len = m_length(mreq, NULL);
 454
 455         cr->cr_xid = xid;
 456         mtx_lock(&cs->cs_lock);
 457
 458         /*
 459          * Try to get a place in the congestion window.
 460          */
 461         while (cu->cu_sent >= cu->cu_cwnd) {
 462                 cu->cu_cwnd_wait = TRUE;
 463                 error = msleep(&cu->cu_cwnd_wait, &cs->cs_lock,
 464                     cu->cu_waitflag, "rpccwnd", 0);
 465                 if (error) {
 466                         errp->re_errno = error;
 467                         errp->re_status = stat = RPC_CANTSEND;
 468                         goto out;
 469                 }
 470         }
 471         cu->cu_sent += CWNDSCALE;
 472
 473         TAILQ_INSERT_TAIL(&cs->cs_pending, cr, cr_link);
 474         mtx_unlock(&cs->cs_lock);
 475
 476         /*
 477          * sosend consumes mreq.
 478          */
 479         error = sosend(cu->cu_socket, sa, NULL, mreq, NULL, 0, curthread);
 480         mreq = NULL;
 481
 482         /*
 483          * sub-optimal code appears here because we have
 484          * some clock time to spare while the packets are in flight.
 485          * (We assume that this is actually only executed once.)
 486          */
 487         reply_msg.acpted_rply.ar_verf.oa_flavor = AUTH_NULL;
 488         reply_msg.acpted_rply.ar_verf.oa_base = cr->cr_verf;
 489         reply_msg.acpted_rply.ar_verf.oa_length = 0;
 490         reply_msg.acpted_rply.ar_results.where = NULL;
 491         reply_msg.acpted_rply.ar_results.proc = (xdrproc_t)xdr_void;
 492
 493         mtx_lock(&cs->cs_lock);
 494         if (error) {
 495                 TAILQ_REMOVE(&cs->cs_pending, cr, cr_link);
 496                 errp->re_errno = error;
 497                 errp->re_status = stat = RPC_CANTSEND;
 498                 cu->cu_sent -= CWNDSCALE;
 499                 if (cu->cu_cwnd_wait) {
 500                         cu->cu_cwnd_wait = FALSE;
 501                         wakeup(&cu->cu_cwnd_wait);
 502                 }
 503                 goto out;
 504         }
 505
 506         /*
 507          * Check to see if we got an upcall while waiting for the
 508          * lock.
 509          */
 510         if (cr->cr_error) {
 511                 TAILQ_REMOVE(&cs->cs_pending, cr, cr_link);
 512                 errp->re_errno = cr->cr_error;
 513                 errp->re_status = stat = RPC_CANTRECV;
 514                 cu->cu_sent -= CWNDSCALE;
 515                 if (cu->cu_cwnd_wait) {
 516                         cu->cu_cwnd_wait = FALSE;
 517                         wakeup(&cu->cu_cwnd_wait);
 518                 }
 519                 goto out;
 520         }
 521         if (cr->cr_mrep) {
 522                 TAILQ_REMOVE(&cs->cs_pending, cr, cr_link);
 523                 cu->cu_sent -= CWNDSCALE;
 524                 if (cu->cu_cwnd_wait) {
 525                         cu->cu_cwnd_wait = FALSE;
 526                         wakeup(&cu->cu_cwnd_wait);
 527                 }
 528                 goto got_reply;
 529         }
 530
 531         /*
 532          * Hack to provide rpc-based message passing
 533          */
 534         if (timeout == 0) {
 535                 TAILQ_REMOVE(&cs->cs_pending, cr, cr_link);
 536                 errp->re_status = stat = RPC_TIMEDOUT;
 537                 cu->cu_sent -= CWNDSCALE;
 538                 if (cu->cu_cwnd_wait) {
 539                         cu->cu_cwnd_wait = FALSE;
 540                         wakeup(&cu->cu_cwnd_wait);
 541                 }
 542                 goto out;
 543         }
 544
 545 get_reply:
 546         for (;;) {
 547                 /* Decide how long to wait. */
 548                 if (next_sendtime < timeout)
 549                         tv = next_sendtime;
 550                 else
 551                         tv = timeout;
 552                 tv -= time_waited;
 553
 554                 if (tv > 0) {
 555                         if (cu->cu_closing || cu->cu_closed) {
 556                                 error = 0;
 557                                 cr->cr_error = ESHUTDOWN;
 558                         } else {
 559                                 error = msleep(cr, &cs->cs_lock,
 560                                     cu->cu_waitflag, cu->cu_waitchan, tv);
 561                         }
 562                 } else {
 563                         error = EWOULDBLOCK;
 564                 }
 565
 566                 TAILQ_REMOVE(&cs->cs_pending, cr, cr_link);
 567                 cu->cu_sent -= CWNDSCALE;
 568                 if (cu->cu_cwnd_wait) {
 569                         cu->cu_cwnd_wait = FALSE;
 570                         wakeup(&cu->cu_cwnd_wait);
 571                 }
 572
 573                 if (!error) {
 574                         /*
 575                          * We were woken up by the upcall.  If the
 576                          * upcall had a receive error, report that,
 577                          * otherwise we have a reply.
 578                          */
 579                         if (cr->cr_error) {
 580                                 errp->re_errno = cr->cr_error;
 581                                 errp->re_status = stat = RPC_CANTRECV;
 582                                 goto out;
 583                         }
 584
 585                         cu->cu_cwnd += (CWNDSCALE * CWNDSCALE
 586                             + cu->cu_cwnd / 2) / cu->cu_cwnd;
 587                         if (cu->cu_cwnd > MAXCWND)
 588                                 cu->cu_cwnd = MAXCWND;
 589
 590                         if (rt) {
 591                                 /*
 592                                  * Add one to the time since a tick
 593                                  * count of N means that the actual
 594                                  * time taken was somewhere between N
 595                                  * and N+1.
 596                                  */
 597                                 rtt = ticks - starttime + 1;
 598
 599                                 /*
 600                                  * Update our estimate of the round
 601                                  * trip time using roughly the
 602                                  * algorithm described in RFC
 603                                  * 2988. Given an RTT sample R:
 604                                  *
 605                                  * RTTVAR = (1-beta) * RTTVAR + beta * |SRTT-R|
 606                                  * SRTT = (1-alpha) * SRTT + alpha * R
 607                                  *
 608                                  * where alpha = 0.125 and beta = 0.25.
 609                                  *
 610                                  * The initial retransmit timeout is
 611                                  * SRTT + 4*RTTVAR and doubles on each
 612                                  * retransmision.
 613                                  */
 614                                 if (rt->rt_srtt == 0) {
 615                                         rt->rt_srtt = rtt;
 616                                         rt->rt_deviate = rtt / 2;
 617                                 } else {
 618                                         int32_t error = rtt - rt->rt_srtt;
 619                                         rt->rt_srtt += error / 8;
 620                                         error = abs(error) - rt->rt_deviate;
 621                                         rt->rt_deviate += error / 4;
 622                                 }
 623                                 rt->rt_rtxcur = rt->rt_srtt + 4*rt->rt_deviate;
 624                         }
 625
 626                         break;
 627                 }
 628
 629                 /*
 630                  * The sleep returned an error so our request is still
 631                  * on the list. If we got EWOULDBLOCK, we may want to
 632                  * re-send the request.
 633                  */
 634                 if (error != EWOULDBLOCK) {
 635                         errp->re_errno = error;
 636                         if (error == EINTR)
 637                                 errp->re_status = stat = RPC_INTR;
 638                         else
 639                                 errp->re_status = stat = RPC_CANTRECV;
 640                         goto out;
 641                 }
 642
 643                 time_waited = ticks - starttime;
 644
 645                 /* Check for timeout. */
 646                 if (time_waited > timeout) {
 647                         errp->re_errno = EWOULDBLOCK;
 648                         errp->re_status = stat = RPC_TIMEDOUT;
 649                         goto out;
 650                 }
 651
 652                 /* Retransmit if necessary. */
 653                 if (time_waited >= next_sendtime) {
 654                         cu->cu_cwnd /= 2;
 655                         if (cu->cu_cwnd < CWNDSCALE)
 656                                 cu->cu_cwnd = CWNDSCALE;
 657                         if (ext && ext->rc_feedback) {
 658                                 mtx_unlock(&cs->cs_lock);
 659                                 if (retrans == 0)
 660                                         ext->rc_feedback(FEEDBACK_REXMIT1,
 661                                             proc, ext->rc_feedback_arg);
 662                                 else
 663                                         ext->rc_feedback(FEEDBACK_REXMIT2,
 664                                             proc, ext->rc_feedback_arg);
 665                                 mtx_lock(&cs->cs_lock);
 666                         }
 667                         if (cu->cu_closing || cu->cu_closed) {
 668                                 errp->re_errno = ESHUTDOWN;
 669                                 errp->re_status = stat = RPC_CANTRECV;
 670                                 goto out;
 671                         }
 672                         retrans++;
 673                         /* update retransmit_time */
 674                         if (retransmit_time < RPC_MAX_BACKOFF * hz)
 675                                 retransmit_time = 2 * retransmit_time;
 676                         next_sendtime += retransmit_time;
 677                         goto send_again;
 678                 }
 679                 TAILQ_INSERT_TAIL(&cs->cs_pending, cr, cr_link);
 680         }
 681
 682 got_reply:
 683         /*
 684          * Now decode and validate the response. We need to drop the
 685          * lock since xdr_replymsg may end up sleeping in malloc.
 686          */
 687         mtx_unlock(&cs->cs_lock);
 688
 689         if (ext && ext->rc_feedback)
 690                 ext->rc_feedback(FEEDBACK_OK, proc, ext->rc_feedback_arg);
 691
 692         xdrmbuf_create(&xdrs, cr->cr_mrep, XDR_DECODE);
 693         ok = xdr_replymsg(&xdrs, &reply_msg);
 694         cr->cr_mrep = NULL;
 695
 696         if (ok) {
 697                 if ((reply_msg.rm_reply.rp_stat == MSG_ACCEPTED) &&
 698                     (reply_msg.acpted_rply.ar_stat == SUCCESS))
 699                         errp->re_status = stat = RPC_SUCCESS;
 700                 else
 701                         stat = _seterr_reply(&reply_msg, &(cu->cu_error));
 702
 703                 if (errp->re_status == RPC_SUCCESS) {
 704                         results = xdrmbuf_getall(&xdrs);
 705                         if (! AUTH_VALIDATE(auth, xid,
 706                                 &reply_msg.acpted_rply.ar_verf,
 707                                 &results)) {
 708                                 errp->re_status = stat = RPC_AUTHERROR;
 709                                 errp->re_why = AUTH_INVALIDRESP;
 710                                 if (retrans &&
 711                                     auth->ah_cred.oa_flavor == RPCSEC_GSS) {
 712                                         /*
 713                                          * If we retransmitted, its
 714                                          * possible that we will
 715                                          * receive a reply for one of
 716                                          * the earlier transmissions
 717                                          * (which will use an older
 718                                          * RPCSEC_GSS sequence
 719                                          * number). In this case, just
 720                                          * go back and listen for a
 721                                          * new reply. We could keep a
 722                                          * record of all the seq
 723                                          * numbers we have transmitted
 724                                          * so far so that we could
 725                                          * accept a reply for any of
 726                                          * them here.
 727                                          */
 728                                         XDR_DESTROY(&xdrs);
 729                                         mtx_lock(&cs->cs_lock);
 730                                         TAILQ_INSERT_TAIL(&cs->cs_pending,
 731                                             cr, cr_link);
 732                                         cr->cr_mrep = NULL;
 733                                         goto get_reply;
 734                                 }
 735                         } else {
 736                                 *resultsp = results;
 737                         }
 738                 }               /* end successful completion */
 739                 /*
 740                  * If unsuccesful AND error is an authentication error
 741                  * then refresh credentials and try again, else break
 742                  */
 743                 else if (stat == RPC_AUTHERROR)
 744                         /* maybe our credentials need to be refreshed ... */
 745                         if (nrefreshes > 0 &&
 746                             AUTH_REFRESH(auth, &reply_msg)) {
 747                                 nrefreshes--;
 748                                 XDR_DESTROY(&xdrs);
 749                                 mtx_lock(&cs->cs_lock);
 750                                 goto call_again;
 751                         }
 752                 /* end of unsuccessful completion */
 753         }       /* end of valid reply message */
 754         else {
 755                 errp->re_status = stat = RPC_CANTDECODERES;
 756
 757         }
 758         XDR_DESTROY(&xdrs);
 759         mtx_lock(&cs->cs_lock);
 760 out:
 761         mtx_assert(&cs->cs_lock, MA_OWNED);
 762
 763         if (mreq)
 764                 m_freem(mreq);
 765         if (cr->cr_mrep)
 766                 m_freem(cr->cr_mrep);
 767
 768         cu->cu_threads--;
 769         if (cu->cu_closing)
 770                 wakeup(cu);
 771
 772         mtx_unlock(&cs->cs_lock);
 773
 774         if (auth && stat != RPC_SUCCESS)
 775                 AUTH_VALIDATE(auth, xid, NULL, NULL);
 776
 777         free(cr, M_RPC);
 778
 779         return (stat);
 780 }
 781
 782 static void
 783 clnt_dg_geterr(CLIENT *cl, struct rpc_err *errp)
 784 {
 785         struct cu_data *cu = (struct cu_data *)cl->cl_private;
 786
 787         *errp = cu->cu_error;
 788 }
 789
 790 static bool_t
 791 clnt_dg_freeres(CLIENT *cl, xdrproc_t xdr_res, void *res_ptr)
 792 {
 793         XDR xdrs;
 794         bool_t dummy;
 795
 796         xdrs.x_op = XDR_FREE;
 797         dummy = (*xdr_res)(&xdrs, res_ptr);
 798
 799         return (dummy);
 800 }
 801
 802 /*ARGSUSED*/
 803 static void
 804 clnt_dg_abort(CLIENT *h)
 805 {
 806 }
 807
 808 static bool_t
 809 clnt_dg_control(CLIENT *cl, u_int request, void *info)
 810 {
 811         struct cu_data *cu = (struct cu_data *)cl->cl_private;
 812         struct cu_socket *cs;
 813         struct sockaddr *addr;
 814
 815         cs = cu->cu_socket->so_rcv.sb_upcallarg;
 816         mtx_lock(&cs->cs_lock);
 817
 818         switch (request) {
 819         case CLSET_FD_CLOSE:
 820                 cu->cu_closeit = TRUE;
 821                 mtx_unlock(&cs->cs_lock);
 822                 return (TRUE);
 823         case CLSET_FD_NCLOSE:
 824                 cu->cu_closeit = FALSE;
 825                 mtx_unlock(&cs->cs_lock);
 826                 return (TRUE);
 827         }
 828
 829         /* for other requests which use info */
 830         if (info == NULL) {
 831                 mtx_unlock(&cs->cs_lock);
 832                 return (FALSE);
 833         }
 834         switch (request) {
 835         case CLSET_TIMEOUT:
 836                 if (time_not_ok((struct timeval *)info)) {
 837                         mtx_unlock(&cs->cs_lock);
 838                         return (FALSE);
 839                 }
 840                 cu->cu_total = *(struct timeval *)info;
 841                 break;
 842         case CLGET_TIMEOUT:
 843                 *(struct timeval *)info = cu->cu_total;
 844                 break;
 845         case CLSET_RETRY_TIMEOUT:
 846                 if (time_not_ok((struct timeval *)info)) {
 847                         mtx_unlock(&cs->cs_lock);
 848                         return (FALSE);
 849                 }
 850                 cu->cu_wait = *(struct timeval *)info;
 851                 break;
 852         case CLGET_RETRY_TIMEOUT:
 853                 *(struct timeval *)info = cu->cu_wait;
 854                 break;
 855         case CLGET_SVC_ADDR:
 856                 /*
 857                  * Slightly different semantics to userland - we use
 858                  * sockaddr instead of netbuf.
 859                  */
 860                 memcpy(info, &cu->cu_raddr, cu->cu_raddr.ss_len);
 861                 break;
 862         case CLSET_SVC_ADDR:            /* set to new address */
 863                 addr = (struct sockaddr *)info;
 864                 (void) memcpy(&cu->cu_raddr, addr, addr->sa_len);
 865                 break;
 866         case CLGET_XID:
 867                 *(uint32_t *)info = cu->cu_xid;
 868                 break;
 869
 870         case CLSET_XID:
 871                 /* This will set the xid of the NEXT call */
 872                 /* decrement by 1 as clnt_dg_call() increments once */
 873                 cu->cu_xid = *(uint32_t *)info - 1;
 874                 break;
 875
 876         case CLGET_VERS:
 877                 /*
 878                  * This RELIES on the information that, in the call body,
 879                  * the version number field is the fifth field from the
 880                  * begining of the RPC header. MUST be changed if the
 881                  * call_struct is changed
 882                  */
 883                 *(uint32_t *)info =
 884                     ntohl(*(uint32_t *)(void *)(cu->cu_mcallc +
 885                     4 * BYTES_PER_XDR_UNIT));
 886                 break;
 887
 888         case CLSET_VERS:
 889                 *(uint32_t *)(void *)(cu->cu_mcallc + 4 * BYTES_PER_XDR_UNIT)
 890                         = htonl(*(uint32_t *)info);
 891                 break;
 892
 893         case CLGET_PROG:
 894                 /*
 895                  * This RELIES on the information that, in the call body,
 896                  * the program number field is the fourth field from the
 897                  * begining of the RPC header. MUST be changed if the
 898                  * call_struct is changed
 899                  */
 900                 *(uint32_t *)info =
 901                     ntohl(*(uint32_t *)(void *)(cu->cu_mcallc +
 902                     3 * BYTES_PER_XDR_UNIT));
 903                 break;
 904
 905         case CLSET_PROG:
 906                 *(uint32_t *)(void *)(cu->cu_mcallc + 3 * BYTES_PER_XDR_UNIT)
 907                         = htonl(*(uint32_t *)info);
 908                 break;
 909         case CLSET_ASYNC:
 910                 cu->cu_async = *(int *)info;
 911                 break;
 912         case CLSET_CONNECT:
 913                 cu->cu_connect = *(int *)info;
 914                 break;
 915         case CLSET_WAITCHAN:
 916                 cu->cu_waitchan = (const char *)info;
 917                 break;
 918         case CLGET_WAITCHAN:
 919                 *(const char **) info = cu->cu_waitchan;
 920                 break;
 921         case CLSET_INTERRUPTIBLE:
 922                 if (*(int *) info)
 923                         cu->cu_waitflag = PCATCH;
 924                 else
 925                         cu->cu_waitflag = 0;
 926                 break;
 927         case CLGET_INTERRUPTIBLE:
 928                 if (cu->cu_waitflag)
 929                         *(int *) info = TRUE;
 930                 else
 931                         *(int *) info = FALSE;
 932                 break;
 933         default:
 934                 mtx_unlock(&cs->cs_lock);
 935                 return (FALSE);
 936         }
 937         mtx_unlock(&cs->cs_lock);
 938         return (TRUE);
 939 }
 940
 941 static void
 942 clnt_dg_close(CLIENT *cl)
 943 {
 944         struct cu_data *cu = (struct cu_data *)cl->cl_private;
 945         struct cu_socket *cs;
 946         struct cu_request *cr;
 947
 948         cs = cu->cu_socket->so_rcv.sb_upcallarg;
 949         mtx_lock(&cs->cs_lock);
 950
 951         if (cu->cu_closed) {
 952                 mtx_unlock(&cs->cs_lock);
 953                 return;
 954         }
 955
 956         if (cu->cu_closing) {
 957                 while (cu->cu_closing)
 958                         msleep(cu, &cs->cs_lock, 0, "rpcclose", 0);
 959                 KASSERT(cu->cu_closed, ("client should be closed"));
 960                 mtx_unlock(&cs->cs_lock);
 961                 return;
 962         }
 963
 964         /*
 965          * Abort any pending requests and wait until everyone
 966          * has finished with clnt_vc_call.
 967          */
 968         cu->cu_closing = TRUE;
 969         TAILQ_FOREACH(cr, &cs->cs_pending, cr_link) {
 970                 if (cr->cr_client == cl) {
 971                         cr->cr_xid = 0;
 972                         cr->cr_error = ESHUTDOWN;
 973                         wakeup(cr);
 974                 }
 975         }
 976
 977         while (cu->cu_threads)
 978                 msleep(cu, &cs->cs_lock, 0, "rpcclose", 0);
 979
 980         cu->cu_closing = FALSE;
 981         cu->cu_closed = TRUE;
 982
 983         mtx_unlock(&cs->cs_lock);
 984         wakeup(cu);
 985 }
 986
 987 static void
 988 clnt_dg_destroy(CLIENT *cl)
 989 {
 990         struct cu_data *cu = (struct cu_data *)cl->cl_private;
 991         struct cu_socket *cs;
 992         struct socket *so = NULL;
 993         bool_t lastsocketref;
 994
 995         cs = cu->cu_socket->so_rcv.sb_upcallarg;
 996         clnt_dg_close(cl);
 997
 998         mtx_lock(&cs->cs_lock);
 999
1000         cs->cs_refs--;
1001         if (cs->cs_refs == 0) {
1002                 mtx_unlock(&cs->cs_lock);
1003                 SOCKBUF_LOCK(&cu->cu_socket->so_rcv);
1004                 soupcall_clear(cu->cu_socket, SO_RCV);
1005                 clnt_dg_upcallsdone(cu->cu_socket, cs);
1006                 SOCKBUF_UNLOCK(&cu->cu_socket->so_rcv);
1007                 mtx_destroy(&cs->cs_lock);
1008                 mem_free(cs, sizeof(*cs));
1009                 lastsocketref = TRUE;
1010         } else {
1011                 mtx_unlock(&cs->cs_lock);
1012                 lastsocketref = FALSE;
1013         }
1014
1015         if (cu->cu_closeit && lastsocketref) {
1016                 so = cu->cu_socket;
1017                 cu->cu_socket = NULL;
1018         }
1019
1020         if (so)
1021                 soclose(so);
1022
1023         if (cl->cl_netid && cl->cl_netid[0])
1024                 mem_free(cl->cl_netid, strlen(cl->cl_netid) +1);
1025         if (cl->cl_tp && cl->cl_tp[0])
1026                 mem_free(cl->cl_tp, strlen(cl->cl_tp) +1);
1027         mem_free(cu, sizeof (*cu));
1028         mem_free(cl, sizeof (CLIENT));
1029 }
1030
1031 /*
1032  * Make sure that the time is not garbage.  -1 value is allowed.
1033  */
1034 static bool_t
1035 time_not_ok(struct timeval *t)
1036 {
1037         return (t->tv_sec < -1 || t->tv_sec > 100000000 ||
1038                 t->tv_usec < -1 || t->tv_usec > 1000000);
1039 }
1040
1041 int
1042 clnt_dg_soupcall(struct socket *so, void *arg, int waitflag)
1043 {
1044         struct cu_socket *cs = (struct cu_socket *) arg;
1045         struct uio uio;
1046         struct mbuf *m;
1047         struct mbuf *control;
1048         struct cu_request *cr;
1049         int error, rcvflag, foundreq;
1050         uint32_t xid;
1051
1052         cs->cs_upcallrefs++;
1053         uio.uio_resid = 1000000000;
1054         uio.uio_td = curthread;
1055         do {
1056                 SOCKBUF_UNLOCK(&so->so_rcv);
1057                 m = NULL;
1058                 control = NULL;
1059                 rcvflag = MSG_DONTWAIT;
1060                 error = soreceive(so, NULL, &uio, &m, &control, &rcvflag);
1061                 if (control)
1062                         m_freem(control);
1063                 SOCKBUF_LOCK(&so->so_rcv);
1064
1065                 if (error == EWOULDBLOCK)
1066                         break;
1067
1068                 /*
1069                  * If there was an error, wake up all pending
1070                  * requests.
1071                  */
1072                 if (error) {
1073                         mtx_lock(&cs->cs_lock);
1074                         TAILQ_FOREACH(cr, &cs->cs_pending, cr_link) {
1075                                 cr->cr_xid = 0;
1076                                 cr->cr_error = error;
1077                                 wakeup(cr);
1078                         }
1079                         mtx_unlock(&cs->cs_lock);
1080                         break;
1081                 }
1082
1083                 /*
1084                  * The XID is in the first uint32_t of the reply.
1085                  */
1086                 if (m->m_len < sizeof(xid))
1087                         m = m_pullup(m, sizeof(xid));
1088                 if (!m)
1089                         /*
1090                          * Should never happen.
1091                          */
1092                         continue;
1093
1094                 xid = ntohl(*mtod(m, uint32_t *));
1095
1096                 /*
1097                  * Attempt to match this reply with a pending request.
1098                  */
1099                 mtx_lock(&cs->cs_lock);
1100                 foundreq = 0;
1101                 TAILQ_FOREACH(cr, &cs->cs_pending, cr_link) {
1102                         if (cr->cr_xid == xid) {
1103                                 /*
1104                                  * This one matches. We leave the
1105                                  * reply mbuf in cr->cr_mrep. Set the
1106                                  * XID to zero so that we will ignore
1107                                  * any duplicated replies that arrive
1108                                  * before clnt_dg_call removes it from
1109                                  * the queue.
1110                                  */
1111                                 cr->cr_xid = 0;
1112                                 cr->cr_mrep = m;
1113                                 cr->cr_error = 0;
1114                                 foundreq = 1;
1115                                 wakeup(cr);
1116                                 break;
1117                         }
1118                 }
1119                 mtx_unlock(&cs->cs_lock);
1120
1121                 /*
1122                  * If we didn't find the matching request, just drop
1123                  * it - its probably a repeated reply.
1124                  */
1125                 if (!foundreq)
1126                         m_freem(m);
1127         } while (m);
1128         cs->cs_upcallrefs--;
1129         if (cs->cs_upcallrefs < 0)
1130                 panic("rpcdg upcall refcnt");
1131         if (cs->cs_upcallrefs == 0)
1132                 wakeup(&cs->cs_upcallrefs);
1133         return (SU_OK);
1134 }
1135
1136 /*
1137  * Wait for all upcalls in progress to complete.
1138  */
1139 static void
1140 clnt_dg_upcallsdone(struct socket *so, struct cu_socket *cs)
1141 {
1142
1143         SOCKBUF_LOCK_ASSERT(&so->so_rcv);
1144
1145         while (cs->cs_upcallrefs > 0)
1146                 (void) msleep(&cs->cs_upcallrefs, SOCKBUF_MTX(&so->so_rcv), 0,
1147                     "rpcdgup", 0);
1148 }