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         int error;
 197
 198         if (svcaddr == NULL) {
 199                 rpc_createerr.cf_stat = RPC_UNKNOWNADDR;
 200                 return (NULL);
 201         }
 202
 203         CURVNET_SET(so->so_vnet);
 204         if (!__rpc_socket2sockinfo(so, &si)) {
 205                 rpc_createerr.cf_stat = RPC_TLIERROR;
 206                 rpc_createerr.cf_error.re_errno = 0;
 207                 CURVNET_RESTORE();
 208                 return (NULL);
 209         }
 210         CURVNET_RESTORE();
 211
 212         /*
 213          * Find the receive and the send size
 214          */
 215         sendsz = __rpc_get_t_size(si.si_af, si.si_proto, (int)sendsz);
 216         recvsz = __rpc_get_t_size(si.si_af, si.si_proto, (int)recvsz);
 217         if ((sendsz == 0) || (recvsz == 0)) {
 218                 rpc_createerr.cf_stat = RPC_TLIERROR; /* XXX */
 219                 rpc_createerr.cf_error.re_errno = 0;
 220                 return (NULL);
 221         }
 222
 223         cl = mem_alloc(sizeof (CLIENT));
 224
 225         /*
 226          * Should be multiple of 4 for XDR.
 227          */
 228         sendsz = ((sendsz + 3) / 4) * 4;
 229         recvsz = ((recvsz + 3) / 4) * 4;
 230         cu = mem_alloc(sizeof (*cu));
 231         cu->cu_threads = 0;
 232         cu->cu_closing = FALSE;
 233         cu->cu_closed = FALSE;
 234         (void) memcpy(&cu->cu_raddr, svcaddr, (size_t)svcaddr->sa_len);
 235         cu->cu_rlen = svcaddr->sa_len;
 236         /* Other values can also be set through clnt_control() */
 237         cu->cu_wait.tv_sec = 3; /* heuristically chosen */
 238         cu->cu_wait.tv_usec = 0;
 239         cu->cu_total.tv_sec = -1;
 240         cu->cu_total.tv_usec = -1;
 241         cu->cu_sendsz = sendsz;
 242         cu->cu_recvsz = recvsz;
 243         cu->cu_async = FALSE;
 244         cu->cu_connect = FALSE;
 245         cu->cu_connected = FALSE;
 246         cu->cu_waitchan = "rpcrecv";
 247         cu->cu_waitflag = 0;
 248         cu->cu_cwnd = MAXCWND / 2;
 249         cu->cu_sent = 0;
 250         cu->cu_cwnd_wait = FALSE;
 251         (void) getmicrotime(&now);
 252         cu->cu_xid = __RPC_GETXID(&now);
 253         call_msg.rm_xid = cu->cu_xid;
 254         call_msg.rm_call.cb_prog = program;
 255         call_msg.rm_call.cb_vers = version;
 256         xdrmem_create(&xdrs, cu->cu_mcallc, MCALL_MSG_SIZE, XDR_ENCODE);
 257         if (! xdr_callhdr(&xdrs, &call_msg)) {
 258                 rpc_createerr.cf_stat = RPC_CANTENCODEARGS;  /* XXX */
 259                 rpc_createerr.cf_error.re_errno = 0;
 260                 goto err2;
 261         }
 262         cu->cu_mcalllen = XDR_GETPOS(&xdrs);
 263
 264         /*
 265          * By default, closeit is always FALSE. It is users responsibility
 266          * to do a close on it, else the user may use clnt_control
 267          * to let clnt_destroy do it for him/her.
 268          */
 269         cu->cu_closeit = FALSE;
 270         cu->cu_socket = so;
 271         error = soreserve(so, (u_long)sendsz, (u_long)recvsz);
 272         if (error != 0) {
 273                 rpc_createerr.cf_stat = RPC_FAILED;
 274                 rpc_createerr.cf_error.re_errno = error;
 275                 goto err2;
 276         }
 277
 278         sb = &so->so_rcv;
 279         SOCKBUF_LOCK(&so->so_rcv);
 280 recheck_socket:
 281         if (sb->sb_upcall) {
 282                 if (sb->sb_upcall != clnt_dg_soupcall) {
 283                         SOCKBUF_UNLOCK(&so->so_rcv);
 284                         printf("clnt_dg_create(): socket already has an incompatible upcall\n");
 285                         goto err2;
 286                 }
 287                 cs = (struct cu_socket *) sb->sb_upcallarg;
 288                 mtx_lock(&cs->cs_lock);
 289                 cs->cs_refs++;
 290                 mtx_unlock(&cs->cs_lock);
 291         } else {
 292                 /*
 293                  * We are the first on this socket - allocate the
 294                  * structure and install it in the socket.
 295                  */
 296                 SOCKBUF_UNLOCK(&so->so_rcv);
 297                 cs = mem_alloc(sizeof(*cs));
 298                 SOCKBUF_LOCK(&so->so_rcv);
 299                 if (sb->sb_upcall) {
 300                         /*
 301                          * We have lost a race with some other client.
 302                          */
 303                         mem_free(cs, sizeof(*cs));
 304                         goto recheck_socket;
 305                 }
 306                 mtx_init(&cs->cs_lock, "cs->cs_lock", NULL, MTX_DEF);
 307                 cs->cs_refs = 1;
 308                 cs->cs_upcallrefs = 0;
 309                 TAILQ_INIT(&cs->cs_pending);
 310                 soupcall_set(so, SO_RCV, clnt_dg_soupcall, cs);
 311         }
 312         SOCKBUF_UNLOCK(&so->so_rcv);
 313
 314         cl->cl_refs = 1;
 315         cl->cl_ops = &clnt_dg_ops;
 316         cl->cl_private = (caddr_t)(void *)cu;
 317         cl->cl_auth = authnone_create();
 318         cl->cl_tp = NULL;
 319         cl->cl_netid = NULL;
 320         return (cl);
 321 err2:
 322         if (cl) {
 323                 mem_free(cl, sizeof (CLIENT));
 324                 if (cu)
 325                         mem_free(cu, sizeof (*cu));
 326         }
 327         return (NULL);
 328 }
 329
 330 static enum clnt_stat
 331 clnt_dg_call(
 332         CLIENT          *cl,            /* client handle */
 333         struct rpc_callextra *ext,      /* call metadata */
 334         rpcproc_t       proc,           /* procedure number */
 335         struct mbuf     *args,          /* pointer to args */
 336         struct mbuf     **resultsp,     /* pointer to results */
 337         struct timeval  utimeout)       /* seconds to wait before giving up */
 338 {
 339         struct cu_data *cu = (struct cu_data *)cl->cl_private;
 340         struct cu_socket *cs;
 341         struct rpc_timers *rt;
 342         AUTH *auth;
 343         struct rpc_err *errp;
 344         enum clnt_stat stat;
 345         XDR xdrs;
 346         struct rpc_msg reply_msg;
 347         bool_t ok;
 348         int retrans;                    /* number of re-transmits so far */
 349         int nrefreshes = 2;             /* number of times to refresh cred */
 350         struct timeval *tvp;
 351         int timeout;
 352         int retransmit_time;
 353         int next_sendtime, starttime, rtt, time_waited, tv = 0;
 354         struct sockaddr *sa;
 355         socklen_t salen;
 356         uint32_t xid = 0;
 357         struct mbuf *mreq = NULL, *results;
 358         struct cu_request *cr;
 359         int error;
 360
 361         cs = cu->cu_socket->so_rcv.sb_upcallarg;
 362         cr = malloc(sizeof(struct cu_request), M_RPC, M_WAITOK);
 363
 364         mtx_lock(&cs->cs_lock);
 365
 366         if (cu->cu_closing || cu->cu_closed) {
 367                 mtx_unlock(&cs->cs_lock);
 368                 free(cr, M_RPC);
 369                 return (RPC_CANTSEND);
 370         }
 371         cu->cu_threads++;
 372
 373         if (ext) {
 374                 auth = ext->rc_auth;
 375                 errp = &ext->rc_err;
 376         } else {
 377                 auth = cl->cl_auth;
 378                 errp = &cu->cu_error;
 379         }
 380
 381         cr->cr_client = cl;
 382         cr->cr_mrep = NULL;
 383         cr->cr_error = 0;
 384
 385         if (cu->cu_total.tv_usec == -1) {
 386                 tvp = &utimeout; /* use supplied timeout */
 387         } else {
 388                 tvp = &cu->cu_total; /* use default timeout */
 389         }
 390         if (tvp->tv_sec || tvp->tv_usec)
 391                 timeout = tvtohz(tvp);
 392         else
 393                 timeout = 0;
 394
 395         if (cu->cu_connect && !cu->cu_connected) {
 396                 mtx_unlock(&cs->cs_lock);
 397                 error = soconnect(cu->cu_socket,
 398                     (struct sockaddr *)&cu->cu_raddr, curthread);
 399                 mtx_lock(&cs->cs_lock);
 400                 if (error) {
 401                         errp->re_errno = error;
 402                         errp->re_status = stat = RPC_CANTSEND;
 403                         goto out;
 404                 }
 405                 cu->cu_connected = 1;
 406         }
 407         if (cu->cu_connected) {
 408                 sa = NULL;
 409                 salen = 0;
 410         } else {
 411                 sa = (struct sockaddr *)&cu->cu_raddr;
 412                 salen = cu->cu_rlen;
 413         }
 414         time_waited = 0;
 415         retrans = 0;
 416         if (ext && ext->rc_timers) {
 417                 rt = ext->rc_timers;
 418                 if (!rt->rt_rtxcur)
 419                         rt->rt_rtxcur = tvtohz(&cu->cu_wait);
 420                 retransmit_time = next_sendtime = rt->rt_rtxcur;
 421         } else {
 422                 rt = NULL;
 423                 retransmit_time = next_sendtime = tvtohz(&cu->cu_wait);
 424         }
 425
 426         starttime = ticks;
 427
 428 call_again:
 429         mtx_assert(&cs->cs_lock, MA_OWNED);
 430
 431         cu->cu_xid++;
 432         xid = cu->cu_xid;
 433
 434 send_again:
 435         mtx_unlock(&cs->cs_lock);
 436
 437         MGETHDR(mreq, M_WAIT, MT_DATA);
 438         KASSERT(cu->cu_mcalllen <= MHLEN, ("RPC header too big"));
 439         bcopy(cu->cu_mcallc, mreq->m_data, cu->cu_mcalllen);
 440         mreq->m_len = cu->cu_mcalllen;
 441
 442         /*
 443          * The XID is the first thing in the request.
 444          */
 445         *mtod(mreq, uint32_t *) = htonl(xid);
 446
 447         xdrmbuf_create(&xdrs, mreq, XDR_ENCODE);
 448
 449         if (cu->cu_async == TRUE && args == NULL)
 450                 goto get_reply;
 451
 452         if ((! XDR_PUTINT32(&xdrs, &proc)) ||
 453             (! AUTH_MARSHALL(auth, xid, &xdrs,
 454                 m_copym(args, 0, M_COPYALL, M_WAITOK)))) {
 455                 errp->re_status = stat = RPC_CANTENCODEARGS;
 456                 mtx_lock(&cs->cs_lock);
 457                 goto out;
 458         }
 459         mreq->m_pkthdr.len = m_length(mreq, NULL);
 460
 461         cr->cr_xid = xid;
 462         mtx_lock(&cs->cs_lock);
 463
 464         /*
 465          * Try to get a place in the congestion window.
 466          */
 467         while (cu->cu_sent >= cu->cu_cwnd) {
 468                 cu->cu_cwnd_wait = TRUE;
 469                 error = msleep(&cu->cu_cwnd_wait, &cs->cs_lock,
 470                     cu->cu_waitflag, "rpccwnd", 0);
 471                 if (error) {
 472                         errp->re_errno = error;
 473                         errp->re_status = stat = RPC_CANTSEND;
 474                         goto out;
 475                 }
 476         }
 477         cu->cu_sent += CWNDSCALE;
 478
 479         TAILQ_INSERT_TAIL(&cs->cs_pending, cr, cr_link);
 480         mtx_unlock(&cs->cs_lock);
 481
 482         /*
 483          * sosend consumes mreq.
 484          */
 485         error = sosend(cu->cu_socket, sa, NULL, mreq, NULL, 0, curthread);
 486         mreq = NULL;
 487
 488         /*
 489          * sub-optimal code appears here because we have
 490          * some clock time to spare while the packets are in flight.
 491          * (We assume that this is actually only executed once.)
 492          */
 493         reply_msg.acpted_rply.ar_verf.oa_flavor = AUTH_NULL;
 494         reply_msg.acpted_rply.ar_verf.oa_base = cr->cr_verf;
 495         reply_msg.acpted_rply.ar_verf.oa_length = 0;
 496         reply_msg.acpted_rply.ar_results.where = NULL;
 497         reply_msg.acpted_rply.ar_results.proc = (xdrproc_t)xdr_void;
 498
 499         mtx_lock(&cs->cs_lock);
 500         if (error) {
 501                 TAILQ_REMOVE(&cs->cs_pending, cr, cr_link);
 502                 errp->re_errno = error;
 503                 errp->re_status = stat = RPC_CANTSEND;
 504                 cu->cu_sent -= CWNDSCALE;
 505                 if (cu->cu_cwnd_wait) {
 506                         cu->cu_cwnd_wait = FALSE;
 507                         wakeup(&cu->cu_cwnd_wait);
 508                 }
 509                 goto out;
 510         }
 511
 512         /*
 513          * Check to see if we got an upcall while waiting for the
 514          * lock.
 515          */
 516         if (cr->cr_error) {
 517                 TAILQ_REMOVE(&cs->cs_pending, cr, cr_link);
 518                 errp->re_errno = cr->cr_error;
 519                 errp->re_status = stat = RPC_CANTRECV;
 520                 cu->cu_sent -= CWNDSCALE;
 521                 if (cu->cu_cwnd_wait) {
 522                         cu->cu_cwnd_wait = FALSE;
 523                         wakeup(&cu->cu_cwnd_wait);
 524                 }
 525                 goto out;
 526         }
 527         if (cr->cr_mrep) {
 528                 TAILQ_REMOVE(&cs->cs_pending, cr, cr_link);
 529                 cu->cu_sent -= CWNDSCALE;
 530                 if (cu->cu_cwnd_wait) {
 531                         cu->cu_cwnd_wait = FALSE;
 532                         wakeup(&cu->cu_cwnd_wait);
 533                 }
 534                 goto got_reply;
 535         }
 536
 537         /*
 538          * Hack to provide rpc-based message passing
 539          */
 540         if (timeout == 0) {
 541                 TAILQ_REMOVE(&cs->cs_pending, cr, cr_link);
 542                 errp->re_status = stat = RPC_TIMEDOUT;
 543                 cu->cu_sent -= CWNDSCALE;
 544                 if (cu->cu_cwnd_wait) {
 545                         cu->cu_cwnd_wait = FALSE;
 546                         wakeup(&cu->cu_cwnd_wait);
 547                 }
 548                 goto out;
 549         }
 550
 551 get_reply:
 552         for (;;) {
 553                 /* Decide how long to wait. */
 554                 if (next_sendtime < timeout)
 555                         tv = next_sendtime;
 556                 else
 557                         tv = timeout;
 558                 tv -= time_waited;
 559
 560                 if (tv > 0) {
 561                         if (cu->cu_closing || cu->cu_closed) {
 562                                 error = 0;
 563                                 cr->cr_error = ESHUTDOWN;
 564                         } else {
 565                                 error = msleep(cr, &cs->cs_lock,
 566                                     cu->cu_waitflag, cu->cu_waitchan, tv);
 567                         }
 568                 } else {
 569                         error = EWOULDBLOCK;
 570                 }
 571
 572                 TAILQ_REMOVE(&cs->cs_pending, cr, cr_link);
 573                 cu->cu_sent -= CWNDSCALE;
 574                 if (cu->cu_cwnd_wait) {
 575                         cu->cu_cwnd_wait = FALSE;
 576                         wakeup(&cu->cu_cwnd_wait);
 577                 }
 578
 579                 if (!error) {
 580                         /*
 581                          * We were woken up by the upcall.  If the
 582                          * upcall had a receive error, report that,
 583                          * otherwise we have a reply.
 584                          */
 585                         if (cr->cr_error) {
 586                                 errp->re_errno = cr->cr_error;
 587                                 errp->re_status = stat = RPC_CANTRECV;
 588                                 goto out;
 589                         }
 590
 591                         cu->cu_cwnd += (CWNDSCALE * CWNDSCALE
 592                             + cu->cu_cwnd / 2) / cu->cu_cwnd;
 593                         if (cu->cu_cwnd > MAXCWND)
 594                                 cu->cu_cwnd = MAXCWND;
 595
 596                         if (rt) {
 597                                 /*
 598                                  * Add one to the time since a tick
 599                                  * count of N means that the actual
 600                                  * time taken was somewhere between N
 601                                  * and N+1.
 602                                  */
 603                                 rtt = ticks - starttime + 1;
 604
 605                                 /*
 606                                  * Update our estimate of the round
 607                                  * trip time using roughly the
 608                                  * algorithm described in RFC
 609                                  * 2988. Given an RTT sample R:
 610                                  *
 611                                  * RTTVAR = (1-beta) * RTTVAR + beta * |SRTT-R|
 612                                  * SRTT = (1-alpha) * SRTT + alpha * R
 613                                  *
 614                                  * where alpha = 0.125 and beta = 0.25.
 615                                  *
 616                                  * The initial retransmit timeout is
 617                                  * SRTT + 4*RTTVAR and doubles on each
 618                                  * retransmision.
 619                                  */
 620                                 if (rt->rt_srtt == 0) {
 621                                         rt->rt_srtt = rtt;
 622                                         rt->rt_deviate = rtt / 2;
 623                                 } else {
 624                                         int32_t error = rtt - rt->rt_srtt;
 625                                         rt->rt_srtt += error / 8;
 626                                         error = abs(error) - rt->rt_deviate;
 627                                         rt->rt_deviate += error / 4;
 628                                 }
 629                                 rt->rt_rtxcur = rt->rt_srtt + 4*rt->rt_deviate;
 630                         }
 631
 632                         break;
 633                 }
 634
 635                 /*
 636                  * The sleep returned an error so our request is still
 637                  * on the list. If we got EWOULDBLOCK, we may want to
 638                  * re-send the request.
 639                  */
 640                 if (error != EWOULDBLOCK) {
 641                         errp->re_errno = error;
 642                         if (error == EINTR)
 643                                 errp->re_status = stat = RPC_INTR;
 644                         else
 645                                 errp->re_status = stat = RPC_CANTRECV;
 646                         goto out;
 647                 }
 648
 649                 time_waited = ticks - starttime;
 650
 651                 /* Check for timeout. */
 652                 if (time_waited > timeout) {
 653                         errp->re_errno = EWOULDBLOCK;
 654                         errp->re_status = stat = RPC_TIMEDOUT;
 655                         goto out;
 656                 }
 657
 658                 /* Retransmit if necessary. */
 659                 if (time_waited >= next_sendtime) {
 660                         cu->cu_cwnd /= 2;
 661                         if (cu->cu_cwnd < CWNDSCALE)
 662                                 cu->cu_cwnd = CWNDSCALE;
 663                         if (ext && ext->rc_feedback) {
 664                                 mtx_unlock(&cs->cs_lock);
 665                                 if (retrans == 0)
 666                                         ext->rc_feedback(FEEDBACK_REXMIT1,
 667                                             proc, ext->rc_feedback_arg);
 668                                 else
 669                                         ext->rc_feedback(FEEDBACK_REXMIT2,
 670                                             proc, ext->rc_feedback_arg);
 671                                 mtx_lock(&cs->cs_lock);
 672                         }
 673                         if (cu->cu_closing || cu->cu_closed) {
 674                                 errp->re_errno = ESHUTDOWN;
 675                                 errp->re_status = stat = RPC_CANTRECV;
 676                                 goto out;
 677                         }
 678                         retrans++;
 679                         /* update retransmit_time */
 680                         if (retransmit_time < RPC_MAX_BACKOFF * hz)
 681                                 retransmit_time = 2 * retransmit_time;
 682                         next_sendtime += retransmit_time;
 683                         goto send_again;
 684                 }
 685                 TAILQ_INSERT_TAIL(&cs->cs_pending, cr, cr_link);
 686         }
 687
 688 got_reply:
 689         /*
 690          * Now decode and validate the response. We need to drop the
 691          * lock since xdr_replymsg may end up sleeping in malloc.
 692          */
 693         mtx_unlock(&cs->cs_lock);
 694
 695         if (ext && ext->rc_feedback)
 696                 ext->rc_feedback(FEEDBACK_OK, proc, ext->rc_feedback_arg);
 697
 698         xdrmbuf_create(&xdrs, cr->cr_mrep, XDR_DECODE);
 699         ok = xdr_replymsg(&xdrs, &reply_msg);
 700         cr->cr_mrep = NULL;
 701
 702         if (ok) {
 703                 if ((reply_msg.rm_reply.rp_stat == MSG_ACCEPTED) &&
 704                     (reply_msg.acpted_rply.ar_stat == SUCCESS))
 705                         errp->re_status = stat = RPC_SUCCESS;
 706                 else
 707                         stat = _seterr_reply(&reply_msg, &(cu->cu_error));
 708
 709                 if (errp->re_status == RPC_SUCCESS) {
 710                         results = xdrmbuf_getall(&xdrs);
 711                         if (! AUTH_VALIDATE(auth, xid,
 712                                 &reply_msg.acpted_rply.ar_verf,
 713                                 &results)) {
 714                                 errp->re_status = stat = RPC_AUTHERROR;
 715                                 errp->re_why = AUTH_INVALIDRESP;
 716                                 if (retrans &&
 717                                     auth->ah_cred.oa_flavor == RPCSEC_GSS) {
 718                                         /*
 719                                          * If we retransmitted, its
 720                                          * possible that we will
 721                                          * receive a reply for one of
 722                                          * the earlier transmissions
 723                                          * (which will use an older
 724                                          * RPCSEC_GSS sequence
 725                                          * number). In this case, just
 726                                          * go back and listen for a
 727                                          * new reply. We could keep a
 728                                          * record of all the seq
 729                                          * numbers we have transmitted
 730                                          * so far so that we could
 731                                          * accept a reply for any of
 732                                          * them here.
 733                                          */
 734                                         XDR_DESTROY(&xdrs);
 735                                         mtx_lock(&cs->cs_lock);
 736                                         TAILQ_INSERT_TAIL(&cs->cs_pending,
 737                                             cr, cr_link);
 738                                         cr->cr_mrep = NULL;
 739                                         goto get_reply;
 740                                 }
 741                         } else {
 742                                 *resultsp = results;
 743                         }
 744                 }               /* end successful completion */
 745                 /*
 746                  * If unsuccesful AND error is an authentication error
 747                  * then refresh credentials and try again, else break
 748                  */
 749                 else if (stat == RPC_AUTHERROR)
 750                         /* maybe our credentials need to be refreshed ... */
 751                         if (nrefreshes > 0 &&
 752                             AUTH_REFRESH(auth, &reply_msg)) {
 753                                 nrefreshes--;
 754                                 XDR_DESTROY(&xdrs);
 755                                 mtx_lock(&cs->cs_lock);
 756                                 goto call_again;
 757                         }
 758                 /* end of unsuccessful completion */
 759         }       /* end of valid reply message */
 760         else {
 761                 errp->re_status = stat = RPC_CANTDECODERES;
 762
 763         }
 764         XDR_DESTROY(&xdrs);
 765         mtx_lock(&cs->cs_lock);
 766 out:
 767         mtx_assert(&cs->cs_lock, MA_OWNED);
 768
 769         if (mreq)
 770                 m_freem(mreq);
 771         if (cr->cr_mrep)
 772                 m_freem(cr->cr_mrep);
 773
 774         cu->cu_threads--;
 775         if (cu->cu_closing)
 776                 wakeup(cu);
 777
 778         mtx_unlock(&cs->cs_lock);
 779
 780         if (auth && stat != RPC_SUCCESS)
 781                 AUTH_VALIDATE(auth, xid, NULL, NULL);
 782
 783         free(cr, M_RPC);
 784
 785         return (stat);
 786 }
 787
 788 static void
 789 clnt_dg_geterr(CLIENT *cl, struct rpc_err *errp)
 790 {
 791         struct cu_data *cu = (struct cu_data *)cl->cl_private;
 792
 793         *errp = cu->cu_error;
 794 }
 795
 796 static bool_t
 797 clnt_dg_freeres(CLIENT *cl, xdrproc_t xdr_res, void *res_ptr)
 798 {
 799         XDR xdrs;
 800         bool_t dummy;
 801
 802         xdrs.x_op = XDR_FREE;
 803         dummy = (*xdr_res)(&xdrs, res_ptr);
 804
 805         return (dummy);
 806 }
 807
 808 /*ARGSUSED*/
 809 static void
 810 clnt_dg_abort(CLIENT *h)
 811 {
 812 }
 813
 814 static bool_t
 815 clnt_dg_control(CLIENT *cl, u_int request, void *info)
 816 {
 817         struct cu_data *cu = (struct cu_data *)cl->cl_private;
 818         struct cu_socket *cs;
 819         struct sockaddr *addr;
 820
 821         cs = cu->cu_socket->so_rcv.sb_upcallarg;
 822         mtx_lock(&cs->cs_lock);
 823
 824         switch (request) {
 825         case CLSET_FD_CLOSE:
 826                 cu->cu_closeit = TRUE;
 827                 mtx_unlock(&cs->cs_lock);
 828                 return (TRUE);
 829         case CLSET_FD_NCLOSE:
 830                 cu->cu_closeit = FALSE;
 831                 mtx_unlock(&cs->cs_lock);
 832                 return (TRUE);
 833         }
 834
 835         /* for other requests which use info */
 836         if (info == NULL) {
 837                 mtx_unlock(&cs->cs_lock);
 838                 return (FALSE);
 839         }
 840         switch (request) {
 841         case CLSET_TIMEOUT:
 842                 if (time_not_ok((struct timeval *)info)) {
 843                         mtx_unlock(&cs->cs_lock);
 844                         return (FALSE);
 845                 }
 846                 cu->cu_total = *(struct timeval *)info;
 847                 break;
 848         case CLGET_TIMEOUT:
 849                 *(struct timeval *)info = cu->cu_total;
 850                 break;
 851         case CLSET_RETRY_TIMEOUT:
 852                 if (time_not_ok((struct timeval *)info)) {
 853                         mtx_unlock(&cs->cs_lock);
 854                         return (FALSE);
 855                 }
 856                 cu->cu_wait = *(struct timeval *)info;
 857                 break;
 858         case CLGET_RETRY_TIMEOUT:
 859                 *(struct timeval *)info = cu->cu_wait;
 860                 break;
 861         case CLGET_SVC_ADDR:
 862                 /*
 863                  * Slightly different semantics to userland - we use
 864                  * sockaddr instead of netbuf.
 865                  */
 866                 memcpy(info, &cu->cu_raddr, cu->cu_raddr.ss_len);
 867                 break;
 868         case CLSET_SVC_ADDR:            /* set to new address */
 869                 addr = (struct sockaddr *)info;
 870                 (void) memcpy(&cu->cu_raddr, addr, addr->sa_len);
 871                 break;
 872         case CLGET_XID:
 873                 *(uint32_t *)info = cu->cu_xid;
 874                 break;
 875
 876         case CLSET_XID:
 877                 /* This will set the xid of the NEXT call */
 878                 /* decrement by 1 as clnt_dg_call() increments once */
 879                 cu->cu_xid = *(uint32_t *)info - 1;
 880                 break;
 881
 882         case CLGET_VERS:
 883                 /*
 884                  * This RELIES on the information that, in the call body,
 885                  * the version number field is the fifth field from the
 886                  * begining of the RPC header. MUST be changed if the
 887                  * call_struct is changed
 888                  */
 889                 *(uint32_t *)info =
 890                     ntohl(*(uint32_t *)(void *)(cu->cu_mcallc +
 891                     4 * BYTES_PER_XDR_UNIT));
 892                 break;
 893
 894         case CLSET_VERS:
 895                 *(uint32_t *)(void *)(cu->cu_mcallc + 4 * BYTES_PER_XDR_UNIT)
 896                         = htonl(*(uint32_t *)info);
 897                 break;
 898
 899         case CLGET_PROG:
 900                 /*
 901                  * This RELIES on the information that, in the call body,
 902                  * the program number field is the fourth field from the
 903                  * begining of the RPC header. MUST be changed if the
 904                  * call_struct is changed
 905                  */
 906                 *(uint32_t *)info =
 907                     ntohl(*(uint32_t *)(void *)(cu->cu_mcallc +
 908                     3 * BYTES_PER_XDR_UNIT));
 909                 break;
 910
 911         case CLSET_PROG:
 912                 *(uint32_t *)(void *)(cu->cu_mcallc + 3 * BYTES_PER_XDR_UNIT)
 913                         = htonl(*(uint32_t *)info);
 914                 break;
 915         case CLSET_ASYNC:
 916                 cu->cu_async = *(int *)info;
 917                 break;
 918         case CLSET_CONNECT:
 919                 cu->cu_connect = *(int *)info;
 920                 break;
 921         case CLSET_WAITCHAN:
 922                 cu->cu_waitchan = (const char *)info;
 923                 break;
 924         case CLGET_WAITCHAN:
 925                 *(const char **) info = cu->cu_waitchan;
 926                 break;
 927         case CLSET_INTERRUPTIBLE:
 928                 if (*(int *) info)
 929                         cu->cu_waitflag = PCATCH;
 930                 else
 931                         cu->cu_waitflag = 0;
 932                 break;
 933         case CLGET_INTERRUPTIBLE:
 934                 if (cu->cu_waitflag)
 935                         *(int *) info = TRUE;
 936                 else
 937                         *(int *) info = FALSE;
 938                 break;
 939         default:
 940                 mtx_unlock(&cs->cs_lock);
 941                 return (FALSE);
 942         }
 943         mtx_unlock(&cs->cs_lock);
 944         return (TRUE);
 945 }
 946
 947 static void
 948 clnt_dg_close(CLIENT *cl)
 949 {
 950         struct cu_data *cu = (struct cu_data *)cl->cl_private;
 951         struct cu_socket *cs;
 952         struct cu_request *cr;
 953
 954         cs = cu->cu_socket->so_rcv.sb_upcallarg;
 955         mtx_lock(&cs->cs_lock);
 956
 957         if (cu->cu_closed) {
 958                 mtx_unlock(&cs->cs_lock);
 959                 return;
 960         }
 961
 962         if (cu->cu_closing) {
 963                 while (cu->cu_closing)
 964                         msleep(cu, &cs->cs_lock, 0, "rpcclose", 0);
 965                 KASSERT(cu->cu_closed, ("client should be closed"));
 966                 mtx_unlock(&cs->cs_lock);
 967                 return;
 968         }
 969
 970         /*
 971          * Abort any pending requests and wait until everyone
 972          * has finished with clnt_vc_call.
 973          */
 974         cu->cu_closing = TRUE;
 975         TAILQ_FOREACH(cr, &cs->cs_pending, cr_link) {
 976                 if (cr->cr_client == cl) {
 977                         cr->cr_xid = 0;
 978                         cr->cr_error = ESHUTDOWN;
 979                         wakeup(cr);
 980                 }
 981         }
 982
 983         while (cu->cu_threads)
 984                 msleep(cu, &cs->cs_lock, 0, "rpcclose", 0);
 985
 986         cu->cu_closing = FALSE;
 987         cu->cu_closed = TRUE;
 988
 989         mtx_unlock(&cs->cs_lock);
 990         wakeup(cu);
 991 }
 992
 993 static void
 994 clnt_dg_destroy(CLIENT *cl)
 995 {
 996         struct cu_data *cu = (struct cu_data *)cl->cl_private;
 997         struct cu_socket *cs;
 998         struct socket *so = NULL;
 999         bool_t lastsocketref;
1000
1001         cs = cu->cu_socket->so_rcv.sb_upcallarg;
1002         clnt_dg_close(cl);
1003
1004         mtx_lock(&cs->cs_lock);
1005
1006         cs->cs_refs--;
1007         if (cs->cs_refs == 0) {
1008                 mtx_unlock(&cs->cs_lock);
1009                 SOCKBUF_LOCK(&cu->cu_socket->so_rcv);
1010                 soupcall_clear(cu->cu_socket, SO_RCV);
1011                 clnt_dg_upcallsdone(cu->cu_socket, cs);
1012                 SOCKBUF_UNLOCK(&cu->cu_socket->so_rcv);
1013                 mtx_destroy(&cs->cs_lock);
1014                 mem_free(cs, sizeof(*cs));
1015                 lastsocketref = TRUE;
1016         } else {
1017                 mtx_unlock(&cs->cs_lock);
1018                 lastsocketref = FALSE;
1019         }
1020
1021         if (cu->cu_closeit && lastsocketref) {
1022                 so = cu->cu_socket;
1023                 cu->cu_socket = NULL;
1024         }
1025
1026         if (so)
1027                 soclose(so);
1028
1029         if (cl->cl_netid && cl->cl_netid[0])
1030                 mem_free(cl->cl_netid, strlen(cl->cl_netid) +1);
1031         if (cl->cl_tp && cl->cl_tp[0])
1032                 mem_free(cl->cl_tp, strlen(cl->cl_tp) +1);
1033         mem_free(cu, sizeof (*cu));
1034         mem_free(cl, sizeof (CLIENT));
1035 }
1036
1037 /*
1038  * Make sure that the time is not garbage.  -1 value is allowed.
1039  */
1040 static bool_t
1041 time_not_ok(struct timeval *t)
1042 {
1043         return (t->tv_sec < -1 || t->tv_sec > 100000000 ||
1044                 t->tv_usec < -1 || t->tv_usec > 1000000);
1045 }
1046
1047 int
1048 clnt_dg_soupcall(struct socket *so, void *arg, int waitflag)
1049 {
1050         struct cu_socket *cs = (struct cu_socket *) arg;
1051         struct uio uio;
1052         struct mbuf *m;
1053         struct mbuf *control;
1054         struct cu_request *cr;
1055         int error, rcvflag, foundreq;
1056         uint32_t xid;
1057
1058         cs->cs_upcallrefs++;
1059         uio.uio_resid = 1000000000;
1060         uio.uio_td = curthread;
1061         do {
1062                 SOCKBUF_UNLOCK(&so->so_rcv);
1063                 m = NULL;
1064                 control = NULL;
1065                 rcvflag = MSG_DONTWAIT;
1066                 error = soreceive(so, NULL, &uio, &m, &control, &rcvflag);
1067                 if (control)
1068                         m_freem(control);
1069                 SOCKBUF_LOCK(&so->so_rcv);
1070
1071                 if (error == EWOULDBLOCK)
1072                         break;
1073
1074                 /*
1075                  * If there was an error, wake up all pending
1076                  * requests.
1077                  */
1078                 if (error) {
1079                         mtx_lock(&cs->cs_lock);
1080                         TAILQ_FOREACH(cr, &cs->cs_pending, cr_link) {
1081                                 cr->cr_xid = 0;
1082                                 cr->cr_error = error;
1083                                 wakeup(cr);
1084                         }
1085                         mtx_unlock(&cs->cs_lock);
1086                         break;
1087                 }
1088
1089                 /*
1090                  * The XID is in the first uint32_t of the reply.
1091                  */
1092                 if (m->m_len < sizeof(xid) && m_length(m, NULL) < sizeof(xid))
1093                         /*
1094                          * Should never happen.
1095                          */
1096                         continue;
1097
1098                 m_copydata(m, 0, sizeof(xid), (char *)&xid);
1099                 xid = ntohl(xid);
1100
1101                 /*
1102                  * Attempt to match this reply with a pending request.
1103                  */
1104                 mtx_lock(&cs->cs_lock);
1105                 foundreq = 0;
1106                 TAILQ_FOREACH(cr, &cs->cs_pending, cr_link) {
1107                         if (cr->cr_xid == xid) {
1108                                 /*
1109                                  * This one matches. We leave the
1110                                  * reply mbuf in cr->cr_mrep. Set the
1111                                  * XID to zero so that we will ignore
1112                                  * any duplicated replies that arrive
1113                                  * before clnt_dg_call removes it from
1114                                  * the queue.
1115                                  */
1116                                 cr->cr_xid = 0;
1117                                 cr->cr_mrep = m;
1118                                 cr->cr_error = 0;
1119                                 foundreq = 1;
1120                                 wakeup(cr);
1121                                 break;
1122                         }
1123                 }
1124                 mtx_unlock(&cs->cs_lock);
1125
1126                 /*
1127                  * If we didn't find the matching request, just drop
1128                  * it - its probably a repeated reply.
1129                  */
1130                 if (!foundreq)
1131                         m_freem(m);
1132         } while (m);
1133         cs->cs_upcallrefs--;
1134         if (cs->cs_upcallrefs < 0)
1135                 panic("rpcdg upcall refcnt");
1136         if (cs->cs_upcallrefs == 0)
1137                 wakeup(&cs->cs_upcallrefs);
1138         return (SU_OK);
1139 }
1140
1141 /*
1142  * Wait for all upcalls in progress to complete.
1143  */
1144 static void
1145 clnt_dg_upcallsdone(struct socket *so, struct cu_socket *cs)
1146 {
1147
1148         SOCKBUF_LOCK_ASSERT(&so->so_rcv);
1149
1150         while (cs->cs_upcallrefs > 0)
1151                 (void) msleep(&cs->cs_upcallrefs, SOCKBUF_MTX(&so->so_rcv), 0,
1152                     "rpcdgup", 0);
1153 }