1/* $NetBSD: svc_vc.c,v 1.7 2000/08/03 00:01:53 fvdl Exp $ */ 2 3/*- 4 * SPDX-License-Identifier: BSD-3-Clause 5 * 6 * Copyright (c) 2009, Sun Microsystems, Inc. 7 * All rights reserved. 8 * 9 * Redistribution and use in source and binary forms, with or without 10 * modification, are permitted provided that the following conditions are met: 11 * - Redistributions of source code must retain the above copyright notice, 12 * this list of conditions and the following disclaimer. 13 * - Redistributions in binary form must reproduce the above copyright notice, 14 * this list of conditions and the following disclaimer in the documentation 15 * and/or other materials provided with the distribution. 16 * - Neither the name of Sun Microsystems, Inc. nor the names of its 17 * contributors may be used to endorse or promote products derived 18 * from this software without specific prior written permission. 19 * 20 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" 21 * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 22 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 23 * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE 24 * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR 25 * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF 26 * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS 27 * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN 28 * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) 29 * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE 30 * POSSIBILITY OF SUCH DAMAGE. 31 */ 32 33#if defined(LIBC_SCCS) && !defined(lint) 34static char *sccsid2 = "@(#)svc_tcp.c 1.21 87/08/11 Copyr 1984 Sun Micro"; 35static char *sccsid = "@(#)svc_tcp.c 2.2 88/08/01 4.0 RPCSRC"; 36#endif 37#include <sys/cdefs.h> 38__FBSDID("$FreeBSD$"); 39 40/* 41 * svc_vc.c, Server side for Connection Oriented based RPC. 42 * 43 * Actually implements two flavors of transporter - 44 * a tcp rendezvouser (a listner and connection establisher) 45 * and a record/tcp stream. 46 */ 47 48#include "opt_kern_tls.h" 49 50#include <sys/param.h> 51#include <sys/limits.h> 52#include <sys/lock.h> 53#include <sys/kernel.h> 54#include <sys/ktls.h> 55#include <sys/malloc.h> 56#include <sys/mbuf.h> 57#include <sys/mutex.h> 58#include <sys/proc.h> 59#include <sys/protosw.h> 60#include <sys/queue.h> 61#include <sys/socket.h> 62#include <sys/socketvar.h> 63#include <sys/sx.h> 64#include <sys/systm.h> 65#include <sys/uio.h> 66 67#include <net/vnet.h> 68 69#include <netinet/tcp.h> 70 71#include <rpc/rpc.h> 72#include <rpc/rpcsec_tls.h> 73 74#include <rpc/krpc.h> 75#include <rpc/rpc_com.h> 76 77#include <security/mac/mac_framework.h> 78 79static bool_t svc_vc_rendezvous_recv(SVCXPRT *, struct rpc_msg *, 80 struct sockaddr **, struct mbuf **); 81static enum xprt_stat svc_vc_rendezvous_stat(SVCXPRT *); 82static void svc_vc_rendezvous_destroy(SVCXPRT *); 83static bool_t svc_vc_null(void); 84static void svc_vc_destroy(SVCXPRT *); 85static enum xprt_stat svc_vc_stat(SVCXPRT *); 86static bool_t svc_vc_ack(SVCXPRT *, uint32_t *); 87static bool_t svc_vc_recv(SVCXPRT *, struct rpc_msg *, 88 struct sockaddr **, struct mbuf **); 89static bool_t svc_vc_reply(SVCXPRT *, struct rpc_msg *, 90 struct sockaddr *, struct mbuf *, uint32_t *seq); 91static bool_t svc_vc_control(SVCXPRT *xprt, const u_int rq, void *in); 92static bool_t svc_vc_rendezvous_control (SVCXPRT *xprt, const u_int rq, 93 void *in); 94static void svc_vc_backchannel_destroy(SVCXPRT *); 95static enum xprt_stat svc_vc_backchannel_stat(SVCXPRT *); 96static bool_t svc_vc_backchannel_recv(SVCXPRT *, struct rpc_msg *, 97 struct sockaddr **, struct mbuf **); 98static bool_t svc_vc_backchannel_reply(SVCXPRT *, struct rpc_msg *, 99 struct sockaddr *, struct mbuf *, uint32_t *); 100static bool_t svc_vc_backchannel_control(SVCXPRT *xprt, const u_int rq, 101 void *in); 102static SVCXPRT *svc_vc_create_conn(SVCPOOL *pool, struct socket *so, 103 struct sockaddr *raddr); 104static int svc_vc_accept(struct socket *head, struct socket **sop); 105static int svc_vc_soupcall(struct socket *so, void *arg, int waitflag); 106static int svc_vc_rendezvous_soupcall(struct socket *, void *, int); 107 108static struct xp_ops svc_vc_rendezvous_ops = { 109 .xp_recv = svc_vc_rendezvous_recv, 110 .xp_stat = svc_vc_rendezvous_stat, 111 .xp_reply = (bool_t (*)(SVCXPRT *, struct rpc_msg *, 112 struct sockaddr *, struct mbuf *, uint32_t *))svc_vc_null, 113 .xp_destroy = svc_vc_rendezvous_destroy, 114 .xp_control = svc_vc_rendezvous_control 115}; 116 117static struct xp_ops svc_vc_ops = { 118 .xp_recv = svc_vc_recv, 119 .xp_stat = svc_vc_stat, 120 .xp_ack = svc_vc_ack, 121 .xp_reply = svc_vc_reply, 122 .xp_destroy = svc_vc_destroy, 123 .xp_control = svc_vc_control 124}; 125 126static struct xp_ops svc_vc_backchannel_ops = { 127 .xp_recv = svc_vc_backchannel_recv, 128 .xp_stat = svc_vc_backchannel_stat, 129 .xp_reply = svc_vc_backchannel_reply, 130 .xp_destroy = svc_vc_backchannel_destroy, 131 .xp_control = svc_vc_backchannel_control 132}; 133 134/* 135 * Usage: 136 * xprt = svc_vc_create(sock, send_buf_size, recv_buf_size); 137 * 138 * Creates, registers, and returns a (rpc) tcp based transporter. 139 * Once *xprt is initialized, it is registered as a transporter 140 * see (svc.h, xprt_register). This routine returns 141 * a NULL if a problem occurred. 142 * 143 * The filedescriptor passed in is expected to refer to a bound, but 144 * not yet connected socket. 145 * 146 * Since streams do buffered io similar to stdio, the caller can specify 147 * how big the send and receive buffers are via the second and third parms; 148 * 0 => use the system default. 149 */ 150SVCXPRT * 151svc_vc_create(SVCPOOL *pool, struct socket *so, size_t sendsize, 152 size_t recvsize) 153{ 154 SVCXPRT *xprt; 155 struct sockaddr* sa; 156 int error; 157 158 SOCK_LOCK(so); 159 if (so->so_state & (SS_ISCONNECTED|SS_ISDISCONNECTED)) { 160 SOCK_UNLOCK(so); 161 CURVNET_SET(so->so_vnet); 162 error = so->so_proto->pr_usrreqs->pru_peeraddr(so, &sa); 163 CURVNET_RESTORE(); 164 if (error) 165 return (NULL); 166 xprt = svc_vc_create_conn(pool, so, sa); 167 free(sa, M_SONAME); 168 return (xprt); 169 } 170 SOCK_UNLOCK(so); 171 172 xprt = svc_xprt_alloc(); 173 sx_init(&xprt->xp_lock, "xprt->xp_lock"); 174 xprt->xp_pool = pool; 175 xprt->xp_socket = so; 176 xprt->xp_p1 = NULL; 177 xprt->xp_p2 = NULL; 178 xprt->xp_ops = &svc_vc_rendezvous_ops; 179 180 CURVNET_SET(so->so_vnet); 181 error = so->so_proto->pr_usrreqs->pru_sockaddr(so, &sa); 182 CURVNET_RESTORE(); 183 if (error) { 184 goto cleanup_svc_vc_create; 185 } 186 187 memcpy(&xprt->xp_ltaddr, sa, sa->sa_len); 188 free(sa, M_SONAME); 189 190 xprt_register(xprt); 191 192 solisten(so, -1, curthread); 193 194 SOLISTEN_LOCK(so); 195 xprt->xp_upcallset = 1; 196 solisten_upcall_set(so, svc_vc_rendezvous_soupcall, xprt); 197 SOLISTEN_UNLOCK(so); 198 199 return (xprt); 200 201cleanup_svc_vc_create: 202 sx_destroy(&xprt->xp_lock); 203 svc_xprt_free(xprt); 204 205 return (NULL); 206} 207 208/* 209 * Create a new transport for a socket optained via soaccept(). 210 */ 211SVCXPRT * 212svc_vc_create_conn(SVCPOOL *pool, struct socket *so, struct sockaddr *raddr) 213{ 214 SVCXPRT *xprt; 215 struct cf_conn *cd; 216 struct sockaddr* sa = NULL; 217 struct sockopt opt; 218 int one = 1; 219 int error; 220 221 bzero(&opt, sizeof(struct sockopt)); 222 opt.sopt_dir = SOPT_SET; 223 opt.sopt_level = SOL_SOCKET; 224 opt.sopt_name = SO_KEEPALIVE; 225 opt.sopt_val = &one; 226 opt.sopt_valsize = sizeof(one); 227 error = sosetopt(so, &opt); 228 if (error) { 229 return (NULL); 230 } 231 232 if (so->so_proto->pr_protocol == IPPROTO_TCP) { 233 bzero(&opt, sizeof(struct sockopt)); 234 opt.sopt_dir = SOPT_SET; 235 opt.sopt_level = IPPROTO_TCP; 236 opt.sopt_name = TCP_NODELAY; 237 opt.sopt_val = &one; 238 opt.sopt_valsize = sizeof(one); 239 error = sosetopt(so, &opt); 240 if (error) { 241 return (NULL); 242 } 243 } 244 245 cd = mem_alloc(sizeof(*cd)); 246 cd->strm_stat = XPRT_IDLE; 247 248 xprt = svc_xprt_alloc(); 249 sx_init(&xprt->xp_lock, "xprt->xp_lock"); 250 xprt->xp_pool = pool; 251 xprt->xp_socket = so; 252 xprt->xp_p1 = cd; 253 xprt->xp_p2 = NULL; 254 xprt->xp_ops = &svc_vc_ops; 255 256 /* 257 * See http://www.connectathon.org/talks96/nfstcp.pdf - client 258 * has a 5 minute timer, server has a 6 minute timer. 259 */ 260 xprt->xp_idletimeout = 6 * 60; 261 262 memcpy(&xprt->xp_rtaddr, raddr, raddr->sa_len); 263 264 CURVNET_SET(so->so_vnet); 265 error = so->so_proto->pr_usrreqs->pru_sockaddr(so, &sa); 266 CURVNET_RESTORE(); 267 if (error) 268 goto cleanup_svc_vc_create; 269 270 memcpy(&xprt->xp_ltaddr, sa, sa->sa_len); 271 free(sa, M_SONAME); 272 273 xprt_register(xprt); 274 275 SOCKBUF_LOCK(&so->so_rcv); 276 xprt->xp_upcallset = 1; 277 soupcall_set(so, SO_RCV, svc_vc_soupcall, xprt); 278 SOCKBUF_UNLOCK(&so->so_rcv); 279 280 /* 281 * Throw the transport into the active list in case it already 282 * has some data buffered. 283 */ 284 sx_xlock(&xprt->xp_lock); 285 xprt_active(xprt); 286 sx_xunlock(&xprt->xp_lock); 287 288 return (xprt); 289cleanup_svc_vc_create: 290 sx_destroy(&xprt->xp_lock); 291 svc_xprt_free(xprt); 292 mem_free(cd, sizeof(*cd)); 293 294 return (NULL); 295} 296 297/* 298 * Create a new transport for a backchannel on a clnt_vc socket. 299 */ 300SVCXPRT * 301svc_vc_create_backchannel(SVCPOOL *pool) 302{ 303 SVCXPRT *xprt = NULL; 304 struct cf_conn *cd = NULL; 305 306 cd = mem_alloc(sizeof(*cd)); 307 cd->strm_stat = XPRT_IDLE; 308 309 xprt = svc_xprt_alloc(); 310 sx_init(&xprt->xp_lock, "xprt->xp_lock"); 311 xprt->xp_pool = pool; 312 xprt->xp_socket = NULL; 313 xprt->xp_p1 = cd; 314 xprt->xp_p2 = NULL; 315 xprt->xp_ops = &svc_vc_backchannel_ops; 316 return (xprt); 317} 318 319/* 320 * This does all of the accept except the final call to soaccept. The 321 * caller will call soaccept after dropping its locks (soaccept may 322 * call malloc). 323 */ 324int 325svc_vc_accept(struct socket *head, struct socket **sop) 326{ 327 struct socket *so; 328 int error = 0; 329 short nbio; 330 331 /* XXXGL: shouldn't that be an assertion? */ 332 if (!SOLISTENING(head)) { 333 error = EINVAL; 334 goto done; 335 } 336#ifdef MAC 337 error = mac_socket_check_accept(curthread->td_ucred, head); 338 if (error != 0) 339 goto done; 340#endif 341 /* 342 * XXXGL: we want non-blocking semantics. The socket could be a 343 * socket created by kernel as well as socket shared with userland, 344 * so we can't be sure about presense of SS_NBIO. We also shall not 345 * toggle it on the socket, since that may surprise userland. So we 346 * set SS_NBIO only temporarily. 347 */ 348 SOLISTEN_LOCK(head); 349 nbio = head->so_state & SS_NBIO; 350 head->so_state |= SS_NBIO; 351 error = solisten_dequeue(head, &so, 0); 352 head->so_state &= (nbio & ~SS_NBIO); 353 if (error) 354 goto done; 355 356 so->so_state |= nbio; 357 *sop = so; 358 359 /* connection has been removed from the listen queue */ 360 KNOTE_UNLOCKED(&head->so_rdsel.si_note, 0); 361done: 362 return (error); 363} 364 365/*ARGSUSED*/ 366static bool_t 367svc_vc_rendezvous_recv(SVCXPRT *xprt, struct rpc_msg *msg, 368 struct sockaddr **addrp, struct mbuf **mp) 369{ 370 struct socket *so = NULL; 371 struct sockaddr *sa = NULL; 372 int error; 373 SVCXPRT *new_xprt; 374 375 /* 376 * The socket upcall calls xprt_active() which will eventually 377 * cause the server to call us here. We attempt to accept a 378 * connection from the socket and turn it into a new 379 * transport. If the accept fails, we have drained all pending 380 * connections so we call xprt_inactive(). 381 */ 382 sx_xlock(&xprt->xp_lock); 383 384 error = svc_vc_accept(xprt->xp_socket, &so); 385 386 if (error == EWOULDBLOCK) { 387 /* 388 * We must re-test for new connections after taking 389 * the lock to protect us in the case where a new 390 * connection arrives after our call to accept fails 391 * with EWOULDBLOCK. 392 */ 393 SOLISTEN_LOCK(xprt->xp_socket); 394 if (TAILQ_EMPTY(&xprt->xp_socket->sol_comp)) 395 xprt_inactive_self(xprt); 396 SOLISTEN_UNLOCK(xprt->xp_socket); 397 sx_xunlock(&xprt->xp_lock); 398 return (FALSE); 399 } 400 401 if (error) { 402 SOLISTEN_LOCK(xprt->xp_socket); 403 if (xprt->xp_upcallset) { 404 xprt->xp_upcallset = 0; 405 soupcall_clear(xprt->xp_socket, SO_RCV); 406 } 407 SOLISTEN_UNLOCK(xprt->xp_socket); 408 xprt_inactive_self(xprt); 409 sx_xunlock(&xprt->xp_lock); 410 return (FALSE); 411 } 412 413 sx_xunlock(&xprt->xp_lock); 414 415 sa = NULL; 416 error = soaccept(so, &sa); 417 418 if (error) { 419 /* 420 * XXX not sure if I need to call sofree or soclose here. 421 */ 422 if (sa) 423 free(sa, M_SONAME); 424 return (FALSE); 425 } 426 427 /* 428 * svc_vc_create_conn will call xprt_register - we don't need 429 * to do anything with the new connection except derefence it. 430 */ 431 new_xprt = svc_vc_create_conn(xprt->xp_pool, so, sa); 432 if (!new_xprt) { 433 soclose(so); 434 } else { 435 SVC_RELEASE(new_xprt); 436 } 437 438 free(sa, M_SONAME); 439 440 return (FALSE); /* there is never an rpc msg to be processed */ 441} 442 443/*ARGSUSED*/ 444static enum xprt_stat 445svc_vc_rendezvous_stat(SVCXPRT *xprt) 446{ 447 448 return (XPRT_IDLE); 449} 450 451static void 452svc_vc_destroy_common(SVCXPRT *xprt) 453{ 454 enum clnt_stat stat; 455 uint32_t reterr; 456 457 if (xprt->xp_socket) { 458 if ((xprt->xp_tls & (RPCTLS_FLAGS_HANDSHAKE | 459 RPCTLS_FLAGS_HANDSHFAIL)) != 0) { 460 if ((xprt->xp_tls & RPCTLS_FLAGS_HANDSHAKE) != 0) { 461 /* 462 * If the upcall fails, the socket has 463 * probably been closed via the rpctlssd 464 * daemon having crashed or been 465 * restarted, so just ignore returned stat. 466 */ 467 stat = rpctls_srv_disconnect(xprt->xp_sslsec, 468 xprt->xp_sslusec, xprt->xp_sslrefno, 469 &reterr); 470 } 471 /* Must sorele() to get rid of reference. */ 472 CURVNET_SET(xprt->xp_socket->so_vnet); 473 SOCK_LOCK(xprt->xp_socket); 474 sorele(xprt->xp_socket); 475 CURVNET_RESTORE(); 476 } else 477 (void)soclose(xprt->xp_socket); 478 } 479 480 if (xprt->xp_netid) 481 (void) mem_free(xprt->xp_netid, strlen(xprt->xp_netid) + 1); 482 svc_xprt_free(xprt); 483} 484 485static void 486svc_vc_rendezvous_destroy(SVCXPRT *xprt) 487{ 488 489 SOLISTEN_LOCK(xprt->xp_socket); 490 if (xprt->xp_upcallset) { 491 xprt->xp_upcallset = 0; 492 solisten_upcall_set(xprt->xp_socket, NULL, NULL); 493 } 494 SOLISTEN_UNLOCK(xprt->xp_socket); 495 496 svc_vc_destroy_common(xprt); 497} 498 499static void 500svc_vc_destroy(SVCXPRT *xprt) 501{ 502 struct cf_conn *cd = (struct cf_conn *)xprt->xp_p1; 503 504 SOCKBUF_LOCK(&xprt->xp_socket->so_rcv); 505 if (xprt->xp_upcallset) { 506 xprt->xp_upcallset = 0; 507 if (xprt->xp_socket->so_rcv.sb_upcall != NULL) 508 soupcall_clear(xprt->xp_socket, SO_RCV); 509 } 510 SOCKBUF_UNLOCK(&xprt->xp_socket->so_rcv); 511 512 svc_vc_destroy_common(xprt); 513 514 if (cd->mreq) 515 m_freem(cd->mreq); 516 if (cd->mpending) 517 m_freem(cd->mpending); 518 mem_free(cd, sizeof(*cd)); 519} 520 521static void 522svc_vc_backchannel_destroy(SVCXPRT *xprt) 523{ 524 struct cf_conn *cd = (struct cf_conn *)xprt->xp_p1; 525 struct mbuf *m, *m2; 526 527 svc_xprt_free(xprt); 528 m = cd->mreq; 529 while (m != NULL) { 530 m2 = m; 531 m = m->m_nextpkt; 532 m_freem(m2); 533 } 534 mem_free(cd, sizeof(*cd)); 535} 536 537/*ARGSUSED*/ 538static bool_t 539svc_vc_control(SVCXPRT *xprt, const u_int rq, void *in) 540{ 541 return (FALSE); 542} 543 544static bool_t 545svc_vc_rendezvous_control(SVCXPRT *xprt, const u_int rq, void *in) 546{ 547 548 return (FALSE); 549} 550 551static bool_t 552svc_vc_backchannel_control(SVCXPRT *xprt, const u_int rq, void *in) 553{ 554 555 return (FALSE); 556} 557 558static enum xprt_stat 559svc_vc_stat(SVCXPRT *xprt) 560{ 561 struct cf_conn *cd; 562 563 cd = (struct cf_conn *)(xprt->xp_p1); 564 565 if (cd->strm_stat == XPRT_DIED) 566 return (XPRT_DIED); 567 568 if (cd->mreq != NULL && cd->resid == 0 && cd->eor) 569 return (XPRT_MOREREQS); 570 571 if (soreadable(xprt->xp_socket)) 572 return (XPRT_MOREREQS); 573 574 return (XPRT_IDLE); 575} 576 577static bool_t 578svc_vc_ack(SVCXPRT *xprt, uint32_t *ack) 579{ 580 581 *ack = atomic_load_acq_32(&xprt->xp_snt_cnt); 582 *ack -= sbused(&xprt->xp_socket->so_snd); 583 return (TRUE); 584} 585 586static enum xprt_stat 587svc_vc_backchannel_stat(SVCXPRT *xprt) 588{ 589 struct cf_conn *cd; 590 591 cd = (struct cf_conn *)(xprt->xp_p1); 592 593 if (cd->mreq != NULL) 594 return (XPRT_MOREREQS); 595 596 return (XPRT_IDLE); 597} 598 599/* 600 * If we have an mbuf chain in cd->mpending, try to parse a record from it, 601 * leaving the result in cd->mreq. If we don't have a complete record, leave 602 * the partial result in cd->mreq and try to read more from the socket. 603 */ 604static int 605svc_vc_process_pending(SVCXPRT *xprt) 606{ 607 struct cf_conn *cd = (struct cf_conn *) xprt->xp_p1; 608 struct socket *so = xprt->xp_socket; 609 struct mbuf *m; 610 611 /* 612 * If cd->resid is non-zero, we have part of the 613 * record already, otherwise we are expecting a record 614 * marker. 615 */ 616 if (!cd->resid && cd->mpending) { 617 /* 618 * See if there is enough data buffered to 619 * make up a record marker. Make sure we can 620 * handle the case where the record marker is 621 * split across more than one mbuf. 622 */ 623 size_t n = 0; 624 uint32_t header; 625 626 m = cd->mpending; 627 while (n < sizeof(uint32_t) && m) { 628 n += m->m_len; 629 m = m->m_next; 630 } 631 if (n < sizeof(uint32_t)) { 632 so->so_rcv.sb_lowat = sizeof(uint32_t) - n; 633 return (FALSE); 634 } 635 m_copydata(cd->mpending, 0, sizeof(header), 636 (char *)&header); 637 header = ntohl(header); 638 cd->eor = (header & 0x80000000) != 0; 639 cd->resid = header & 0x7fffffff; 640 m_adj(cd->mpending, sizeof(uint32_t)); 641 } 642 643 /* 644 * Start pulling off mbufs from cd->mpending 645 * until we either have a complete record or 646 * we run out of data. We use m_split to pull 647 * data - it will pull as much as possible and 648 * split the last mbuf if necessary. 649 */ 650 while (cd->mpending && cd->resid) { 651 m = cd->mpending; 652 if (cd->mpending->m_next 653 || cd->mpending->m_len > cd->resid) 654 cd->mpending = m_split(cd->mpending, 655 cd->resid, M_WAITOK); 656 else 657 cd->mpending = NULL; 658 if (cd->mreq) 659 m_last(cd->mreq)->m_next = m; 660 else 661 cd->mreq = m; 662 while (m) { 663 cd->resid -= m->m_len; 664 m = m->m_next; 665 } 666 } 667 668 /* 669 * Block receive upcalls if we have more data pending, 670 * otherwise report our need. 671 */ 672 if (cd->mpending) 673 so->so_rcv.sb_lowat = INT_MAX; 674 else 675 so->so_rcv.sb_lowat = 676 imax(1, imin(cd->resid, so->so_rcv.sb_hiwat / 2)); 677 return (TRUE); 678} 679 680static bool_t 681svc_vc_recv(SVCXPRT *xprt, struct rpc_msg *msg, 682 struct sockaddr **addrp, struct mbuf **mp) 683{ 684 struct cf_conn *cd = (struct cf_conn *) xprt->xp_p1; 685 struct uio uio; 686 struct mbuf *m, *ctrl; 687 struct socket* so = xprt->xp_socket; 688 XDR xdrs; 689 int error, rcvflag; 690 uint32_t reterr, xid_plus_direction[2]; 691 struct cmsghdr *cmsg; 692 struct tls_get_record tgr; 693 enum clnt_stat ret; 694 695 /* 696 * Serialise access to the socket and our own record parsing 697 * state. 698 */ 699 sx_xlock(&xprt->xp_lock); 700 701 for (;;) { 702 /* If we have no request ready, check pending queue. */ 703 while (cd->mpending && 704 (cd->mreq == NULL || cd->resid != 0 || !cd->eor)) { 705 if (!svc_vc_process_pending(xprt)) 706 break; 707 } 708 709 /* Process and return complete request in cd->mreq. */ 710 if (cd->mreq != NULL && cd->resid == 0 && cd->eor) { 711 712 /* 713 * Now, check for a backchannel reply. 714 * The XID is in the first uint32_t of the reply 715 * and the message direction is the second one. 716 */ 717 if ((cd->mreq->m_len >= sizeof(xid_plus_direction) || 718 m_length(cd->mreq, NULL) >= 719 sizeof(xid_plus_direction)) && 720 xprt->xp_p2 != NULL) { 721 m_copydata(cd->mreq, 0, 722 sizeof(xid_plus_direction), 723 (char *)xid_plus_direction); 724 xid_plus_direction[0] = 725 ntohl(xid_plus_direction[0]); 726 xid_plus_direction[1] = 727 ntohl(xid_plus_direction[1]); 728 /* Check message direction. */ 729 if (xid_plus_direction[1] == REPLY) { 730 clnt_bck_svccall(xprt->xp_p2, 731 cd->mreq, 732 xid_plus_direction[0]); 733 cd->mreq = NULL; 734 continue; 735 } 736 } 737 738 xdrmbuf_create(&xdrs, cd->mreq, XDR_DECODE); 739 cd->mreq = NULL; 740 741 /* Check for next request in a pending queue. */ 742 svc_vc_process_pending(xprt); 743 if (cd->mreq == NULL || cd->resid != 0) { 744 SOCKBUF_LOCK(&so->so_rcv); 745 if (!soreadable(so)) 746 xprt_inactive_self(xprt); 747 SOCKBUF_UNLOCK(&so->so_rcv); 748 } 749 750 sx_xunlock(&xprt->xp_lock); 751 752 if (! xdr_callmsg(&xdrs, msg)) { 753 XDR_DESTROY(&xdrs); 754 return (FALSE); 755 } 756 757 *addrp = NULL; 758 *mp = xdrmbuf_getall(&xdrs); 759 XDR_DESTROY(&xdrs); 760 761 return (TRUE); 762 } 763 764 /* 765 * If receiving is disabled so that a TLS handshake can be 766 * done by the rpctlssd daemon, return FALSE here. 767 */ 768 rcvflag = MSG_DONTWAIT; 769 if ((xprt->xp_tls & RPCTLS_FLAGS_HANDSHAKE) != 0) 770 rcvflag |= MSG_TLSAPPDATA; 771tryagain: 772 if (xprt->xp_dontrcv) { 773 sx_xunlock(&xprt->xp_lock); 774 return (FALSE); 775 } 776 777 /* 778 * The socket upcall calls xprt_active() which will eventually 779 * cause the server to call us here. We attempt to 780 * read as much as possible from the socket and put 781 * the result in cd->mpending. If the read fails, 782 * we have drained both cd->mpending and the socket so 783 * we can call xprt_inactive(). 784 */ 785 uio.uio_resid = 1000000000; 786 uio.uio_td = curthread; 787 ctrl = m = NULL; 788 error = soreceive(so, NULL, &uio, &m, &ctrl, &rcvflag); 789 790 if (error == EWOULDBLOCK) { 791 /* 792 * We must re-test for readability after 793 * taking the lock to protect us in the case 794 * where a new packet arrives on the socket 795 * after our call to soreceive fails with 796 * EWOULDBLOCK. 797 */ 798 SOCKBUF_LOCK(&so->so_rcv); 799 if (!soreadable(so)) 800 xprt_inactive_self(xprt); 801 SOCKBUF_UNLOCK(&so->so_rcv); 802 sx_xunlock(&xprt->xp_lock); 803 return (FALSE); 804 } 805 806 /* 807 * A return of ENXIO indicates that there is a 808 * non-application data record at the head of the 809 * socket's receive queue, for TLS connections. 810 * This record needs to be handled in userland 811 * via an SSL_read() call, so do an upcall to the daemon. 812 */ 813 if ((xprt->xp_tls & RPCTLS_FLAGS_HANDSHAKE) != 0 && 814 error == ENXIO) { 815 /* Disable reception. */ 816 xprt->xp_dontrcv = TRUE; 817 sx_xunlock(&xprt->xp_lock); 818 ret = rpctls_srv_handlerecord(xprt->xp_sslsec, 819 xprt->xp_sslusec, xprt->xp_sslrefno, 820 &reterr); 821 sx_xlock(&xprt->xp_lock); 822 xprt->xp_dontrcv = FALSE; 823 if (ret != RPC_SUCCESS || reterr != RPCTLSERR_OK) { 824 /* 825 * All we can do is soreceive() it and 826 * then toss it. 827 */ 828 rcvflag = MSG_DONTWAIT; 829 goto tryagain; 830 } 831 sx_xunlock(&xprt->xp_lock); 832 xprt_active(xprt); /* Harmless if already active. */ 833 return (FALSE); 834 } 835 836 if (error) { 837 SOCKBUF_LOCK(&so->so_rcv); 838 if (xprt->xp_upcallset) { 839 xprt->xp_upcallset = 0; 840 soupcall_clear(so, SO_RCV); 841 } 842 SOCKBUF_UNLOCK(&so->so_rcv); 843 xprt_inactive_self(xprt); 844 cd->strm_stat = XPRT_DIED; 845 sx_xunlock(&xprt->xp_lock); 846 return (FALSE); 847 } 848 849 if (!m) { 850 /* 851 * EOF - the other end has closed the socket. 852 */ 853 xprt_inactive_self(xprt); 854 cd->strm_stat = XPRT_DIED; 855 sx_xunlock(&xprt->xp_lock); 856 return (FALSE); 857 } 858 859 /* Process any record header(s). */ 860 if (ctrl != NULL) { 861 cmsg = mtod(ctrl, struct cmsghdr *); 862 if (cmsg->cmsg_type == TLS_GET_RECORD && 863 cmsg->cmsg_len == CMSG_LEN(sizeof(tgr))) { 864 memcpy(&tgr, CMSG_DATA(cmsg), sizeof(tgr)); 865 /* 866 * This should have been handled by 867 * the rpctls_svc_handlerecord() 868 * upcall. If not, all we can do is 869 * toss it away. 870 */ 871 if (tgr.tls_type != TLS_RLTYPE_APP) { 872 m_freem(m); 873 m_free(ctrl); 874 rcvflag = MSG_DONTWAIT | MSG_TLSAPPDATA; 875 goto tryagain; 876 } 877 } 878 m_free(ctrl); 879 } 880 881 if (cd->mpending) 882 m_last(cd->mpending)->m_next = m; 883 else 884 cd->mpending = m; 885 } 886} 887 888static bool_t 889svc_vc_backchannel_recv(SVCXPRT *xprt, struct rpc_msg *msg, 890 struct sockaddr **addrp, struct mbuf **mp) 891{ 892 struct cf_conn *cd = (struct cf_conn *) xprt->xp_p1; 893 struct ct_data *ct; 894 struct mbuf *m; 895 XDR xdrs; 896 897 sx_xlock(&xprt->xp_lock); 898 ct = (struct ct_data *)xprt->xp_p2; 899 if (ct == NULL) { 900 sx_xunlock(&xprt->xp_lock); 901 return (FALSE); 902 } 903 mtx_lock(&ct->ct_lock); 904 m = cd->mreq; 905 if (m == NULL) { 906 xprt_inactive_self(xprt); 907 mtx_unlock(&ct->ct_lock); 908 sx_xunlock(&xprt->xp_lock); 909 return (FALSE); 910 } 911 cd->mreq = m->m_nextpkt; 912 mtx_unlock(&ct->ct_lock); 913 sx_xunlock(&xprt->xp_lock); 914 915 xdrmbuf_create(&xdrs, m, XDR_DECODE); 916 if (! xdr_callmsg(&xdrs, msg)) { 917 XDR_DESTROY(&xdrs); 918 return (FALSE); 919 } 920 *addrp = NULL; 921 *mp = xdrmbuf_getall(&xdrs); 922 XDR_DESTROY(&xdrs); 923 return (TRUE); 924} 925 926static bool_t 927svc_vc_reply(SVCXPRT *xprt, struct rpc_msg *msg, 928 struct sockaddr *addr, struct mbuf *m, uint32_t *seq) 929{ 930 XDR xdrs; 931 struct mbuf *mrep; 932 bool_t stat = TRUE; 933 int error, len, maxextsiz; 934#ifdef KERN_TLS 935 u_int maxlen; 936#endif 937 938 /* 939 * Leave space for record mark. 940 */ 941 mrep = m_gethdr(M_WAITOK, MT_DATA); 942 mrep->m_data += sizeof(uint32_t); 943 944 xdrmbuf_create(&xdrs, mrep, XDR_ENCODE); 945 946 if (msg->rm_reply.rp_stat == MSG_ACCEPTED && 947 msg->rm_reply.rp_acpt.ar_stat == SUCCESS) { 948 if (!xdr_replymsg(&xdrs, msg)) 949 stat = FALSE; 950 else 951 xdrmbuf_append(&xdrs, m); 952 } else { 953 stat = xdr_replymsg(&xdrs, msg); 954 } 955 956 if (stat) { 957 m_fixhdr(mrep); 958 959 /* 960 * Prepend a record marker containing the reply length. 961 */ 962 M_PREPEND(mrep, sizeof(uint32_t), M_WAITOK); 963 len = mrep->m_pkthdr.len; 964 *mtod(mrep, uint32_t *) = 965 htonl(0x80000000 | (len - sizeof(uint32_t))); 966 967 /* For RPC-over-TLS, copy mrep to a chain of ext_pgs. */ 968 if ((xprt->xp_tls & RPCTLS_FLAGS_HANDSHAKE) != 0) { 969 /* 970 * Copy the mbuf chain to a chain of 971 * ext_pgs mbuf(s) as required by KERN_TLS. 972 */ 973 maxextsiz = TLS_MAX_MSG_SIZE_V10_2; 974#ifdef KERN_TLS 975 if (rpctls_getinfo(&maxlen, false, false)) 976 maxextsiz = min(maxextsiz, maxlen); 977#endif 978 mrep = _rpc_copym_into_ext_pgs(mrep, maxextsiz); 979 } 980 atomic_add_32(&xprt->xp_snd_cnt, len); 981 /* 982 * sosend consumes mreq. 983 */ 984 error = sosend(xprt->xp_socket, NULL, NULL, mrep, NULL, 985 0, curthread); 986 if (!error) { 987 atomic_add_rel_32(&xprt->xp_snt_cnt, len); 988 if (seq) 989 *seq = xprt->xp_snd_cnt; 990 stat = TRUE; 991 } else 992 atomic_subtract_32(&xprt->xp_snd_cnt, len); 993 } else { 994 m_freem(mrep); 995 } 996 997 XDR_DESTROY(&xdrs); 998 999 return (stat); 1000} 1001 1002static bool_t 1003svc_vc_backchannel_reply(SVCXPRT *xprt, struct rpc_msg *msg, 1004 struct sockaddr *addr, struct mbuf *m, uint32_t *seq) 1005{ 1006 struct ct_data *ct; 1007 XDR xdrs; 1008 struct mbuf *mrep; 1009 bool_t stat = TRUE; 1010 int error, maxextsiz; 1011#ifdef KERN_TLS 1012 u_int maxlen; 1013#endif 1014 1015 /* 1016 * Leave space for record mark. 1017 */ 1018 mrep = m_gethdr(M_WAITOK, MT_DATA); 1019 mrep->m_data += sizeof(uint32_t); 1020 1021 xdrmbuf_create(&xdrs, mrep, XDR_ENCODE); 1022 1023 if (msg->rm_reply.rp_stat == MSG_ACCEPTED && 1024 msg->rm_reply.rp_acpt.ar_stat == SUCCESS) { 1025 if (!xdr_replymsg(&xdrs, msg)) 1026 stat = FALSE; 1027 else 1028 xdrmbuf_append(&xdrs, m); 1029 } else { 1030 stat = xdr_replymsg(&xdrs, msg); 1031 } 1032 1033 if (stat) { 1034 m_fixhdr(mrep); 1035 1036 /* 1037 * Prepend a record marker containing the reply length. 1038 */ 1039 M_PREPEND(mrep, sizeof(uint32_t), M_WAITOK); 1040 *mtod(mrep, uint32_t *) = 1041 htonl(0x80000000 | (mrep->m_pkthdr.len 1042 - sizeof(uint32_t))); 1043 1044 /* For RPC-over-TLS, copy mrep to a chain of ext_pgs. */ 1045 if ((xprt->xp_tls & RPCTLS_FLAGS_HANDSHAKE) != 0) { 1046 /* 1047 * Copy the mbuf chain to a chain of 1048 * ext_pgs mbuf(s) as required by KERN_TLS. 1049 */ 1050 maxextsiz = TLS_MAX_MSG_SIZE_V10_2; 1051#ifdef KERN_TLS 1052 if (rpctls_getinfo(&maxlen, false, false)) 1053 maxextsiz = min(maxextsiz, maxlen); 1054#endif 1055 mrep = _rpc_copym_into_ext_pgs(mrep, maxextsiz); 1056 } 1057 sx_xlock(&xprt->xp_lock); 1058 ct = (struct ct_data *)xprt->xp_p2; 1059 if (ct != NULL) 1060 error = sosend(ct->ct_socket, NULL, NULL, mrep, NULL, 1061 0, curthread); 1062 else 1063 error = EPIPE; 1064 sx_xunlock(&xprt->xp_lock); 1065 if (!error) { 1066 stat = TRUE; 1067 } 1068 } else { 1069 m_freem(mrep); 1070 } 1071 1072 XDR_DESTROY(&xdrs); 1073 1074 return (stat); 1075} 1076 1077static bool_t 1078svc_vc_null() 1079{ 1080 1081 return (FALSE); 1082} 1083 1084static int 1085svc_vc_soupcall(struct socket *so, void *arg, int waitflag) 1086{ 1087 SVCXPRT *xprt = (SVCXPRT *) arg; 1088 1089 if (soreadable(xprt->xp_socket)) 1090 xprt_active(xprt); 1091 return (SU_OK); 1092} 1093 1094static int 1095svc_vc_rendezvous_soupcall(struct socket *head, void *arg, int waitflag) 1096{ 1097 SVCXPRT *xprt = (SVCXPRT *) arg; 1098 1099 if (!TAILQ_EMPTY(&head->sol_comp)) 1100 xprt_active(xprt); 1101 return (SU_OK); 1102} 1103 1104#if 0 1105/* 1106 * Get the effective UID of the sending process. Used by rpcbind, keyserv 1107 * and rpc.yppasswdd on AF_LOCAL. 1108 */ 1109int 1110__rpc_get_local_uid(SVCXPRT *transp, uid_t *uid) { 1111 int sock, ret; 1112 gid_t egid; 1113 uid_t euid; 1114 struct sockaddr *sa; 1115 1116 sock = transp->xp_fd; 1117 sa = (struct sockaddr *)transp->xp_rtaddr; 1118 if (sa->sa_family == AF_LOCAL) { 1119 ret = getpeereid(sock, &euid, &egid); 1120 if (ret == 0) 1121 *uid = euid; 1122 return (ret); 1123 } else 1124 return (-1); 1125} 1126#endif 1127