Move CallBack Server thread creation, initial processing and destruction to RPC Cleanup some RPC code. Remove extraneous fields from nfs41_cb_info and clean up the code. Change KM_SLEEP in mir_nfs41_callback_thread to KM_NOSLEEP. Fix lint warnings Incorporate code review comments.
1 /* 2 * CDDL HEADER START 3 * 4 * The contents of this file are subject to the terms of the 5 * Common Development and Distribution License (the "License"). 6 * You may not use this file except in compliance with the License. 7 * 8 * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE 9 * or http://www.opensolaris.org/os/licensing. 10 * See the License for the specific language governing permissions 11 * and limitations under the License. 12 * 13 * When distributing Covered Code, include this CDDL HEADER in each 14 * file and include the License file at usr/src/OPENSOLARIS.LICENSE. 15 * If applicable, add the following below this CDDL HEADER, with the 16 * fields enclosed by brackets "[]" replaced with your own identifying 17 * information: Portions Copyright [yyyy] [name of copyright owner] 18 * 19 * CDDL HEADER END 20 */ 21 /* 22 * Copyright 2008 Sun Microsystems, Inc. All rights reserved. 23 * Use is subject to license terms. 24 */ 25 /* Copyright (c) 1990 Mentat Inc. */ 26 27 /* Copyright (c) 1984, 1986, 1987, 1988, 1989 AT&T */ 28 /* All Rights Reserved */ 29 30 /* 31 * Kernel RPC filtering module 32 */ 33 34 #include <sys/param.h> 35 #include <sys/types.h> 36 #include <sys/stream.h> 37 #include <sys/stropts.h> 38 #include <sys/tihdr.h> 39 #include <sys/timod.h> 40 #include <sys/tiuser.h> 41 #include <sys/debug.h> 42 #include <sys/signal.h> 43 #include <sys/pcb.h> 44 #include <sys/user.h> 45 #include <sys/errno.h> 46 #include <sys/cred.h> 47 #include <sys/policy.h> 48 #include <sys/inline.h> 49 #include <sys/cmn_err.h> 50 #include <sys/kmem.h> 51 #include <sys/file.h> 52 #include <sys/sysmacros.h> 53 #include <sys/systm.h> 54 #include <sys/t_lock.h> 55 #include <sys/ddi.h> 56 #include <sys/vtrace.h> 57 #include <sys/callb.h> 58 #include <sys/strsun.h> 59 60 #include <sys/strlog.h> 61 #include <rpc/rpc_com.h> 62 #include <inet/common.h> 63 #include <rpc/types.h> 64 #include <sys/time.h> 65 #include <rpc/xdr.h> 66 #include <rpc/auth.h> 67 #include <rpc/clnt.h> 68 #include <rpc/rpc_msg.h> 69 #include <rpc/clnt.h> 70 #include <rpc/svc.h> 71 #include <rpc/rpcsys.h> 72 #include <rpc/rpc_rdma.h> 73 #include <sys/sdt.h> 74 75 /* 76 * This is the loadable module wrapper. 77 */ 78 #include <sys/conf.h> 79 #include <sys/modctl.h> 80 #include <sys/syscall.h> 81 82 extern struct streamtab rpcinfo; 83 84 static struct fmodsw fsw = { 85 "rpcmod", 86 &rpcinfo, 87 D_NEW|D_MP, 88 }; 89 90 /* 91 * Module linkage information for the kernel. 92 */ 93 94 static struct modlstrmod modlstrmod = { 95 &mod_strmodops, "rpc interface str mod", &fsw 96 }; 97 98 /* 99 * For the RPC system call. 100 */ 101 static struct sysent rpcsysent = { 102 2, 103 SE_32RVAL1 | SE_ARGC | SE_NOUNLOAD, 104 rpcsys 105 }; 106 107 static struct modlsys modlsys = { 108 &mod_syscallops, 109 "RPC syscall", 110 &rpcsysent 111 }; 112 113 #ifdef _SYSCALL32_IMPL 114 static struct modlsys modlsys32 = { 115 &mod_syscallops32, 116 "32-bit RPC syscall", 117 &rpcsysent 118 }; 119 #endif /* _SYSCALL32_IMPL */ 120 121 static struct modlinkage modlinkage = { 122 MODREV_1, 123 { 124 &modlsys, 125 #ifdef _SYSCALL32_IMPL 126 &modlsys32, 127 #endif 128 &modlstrmod, 129 NULL 130 } 131 }; 132 133 int 134 _init(void) 135 { 136 int error = 0; 137 callb_id_t cid; 138 int status; 139 140 svc_init(); 141 clnt_init(); 142 cid = callb_add(connmgr_cpr_reset, 0, CB_CL_CPR_RPC, "rpc"); 143 144 if (error = mod_install(&modlinkage)) { 145 /* 146 * Could not install module, cleanup previous 147 * initialization work. 148 */ 149 clnt_fini(); 150 if (cid != NULL) 151 (void) callb_delete(cid); 152 153 return (error); 154 } 155 156 /* 157 * Load up the RDMA plugins and initialize the stats. Even if the 158 * plugins loadup fails, but rpcmod was successfully installed the 159 * counters still get initialized. 160 */ 161 rw_init(&rdma_lock, NULL, RW_DEFAULT, NULL); 162 mutex_init(&rdma_modload_lock, NULL, MUTEX_DEFAULT, NULL); 163 mt_kstat_init(); 164 165 /* 166 * Get our identification into ldi. This is used for loading 167 * other modules, e.g. rpcib. 168 */ 169 status = ldi_ident_from_mod(&modlinkage, &rpcmod_li); 170 if (status != 0) { 171 cmn_err(CE_WARN, "ldi_ident_from_mod fails with %d", status); 172 rpcmod_li = NULL; 173 } 174 175 return (error); 176 } 177 178 /* 179 * The unload entry point fails, because we advertise entry points into 180 * rpcmod from the rest of kRPC: rpcmod_release(). 181 */ 182 int 183 _fini(void) 184 { 185 return (EBUSY); 186 } 187 188 int 189 _info(struct modinfo *modinfop) 190 { 191 return (mod_info(&modlinkage, modinfop)); 192 } 193 194 extern int nulldev(); 195 196 #define RPCMOD_ID 2049 197 198 int rmm_open(), rmm_close(); 199 200 /* 201 * To save instructions, since STREAMS ignores the return value 202 * from these functions, they are defined as void here. Kind of icky, but... 203 */ 204 void rmm_rput(queue_t *, mblk_t *); 205 void rmm_wput(queue_t *, mblk_t *); 206 void rmm_rsrv(queue_t *); 207 void rmm_wsrv(queue_t *); 208 209 int rpcmodopen(), rpcmodclose(); 210 void rpcmodrput(), rpcmodwput(); 211 void rpcmodrsrv(), rpcmodwsrv(); 212 213 static void rpcmodwput_other(queue_t *, mblk_t *); 214 static int mir_close(queue_t *q); 215 static int mir_open(queue_t *q, dev_t *devp, int flag, int sflag, 216 cred_t *credp); 217 static void mir_rput(queue_t *q, mblk_t *mp); 218 static void mir_rsrv(queue_t *q); 219 static void mir_wput(queue_t *q, mblk_t *mp); 220 static void mir_wsrv(queue_t *q); 221 222 static struct module_info rpcmod_info = 223 {RPCMOD_ID, "rpcmod", 0, INFPSZ, 256*1024, 1024}; 224 225 /* 226 * Read side has no service procedure. 227 */ 228 static struct qinit rpcmodrinit = { 229 (int (*)())rmm_rput, 230 (int (*)())rmm_rsrv, 231 rmm_open, 232 rmm_close, 233 nulldev, 234 &rpcmod_info, 235 NULL 236 }; 237 238 /* 239 * The write put procedure is simply putnext to conserve stack space. 240 * The write service procedure is not used to queue data, but instead to 241 * synchronize with flow control. 242 */ 243 static struct qinit rpcmodwinit = { 244 (int (*)())rmm_wput, 245 (int (*)())rmm_wsrv, 246 rmm_open, 247 rmm_close, 248 nulldev, 249 &rpcmod_info, 250 NULL 251 }; 252 struct streamtab rpcinfo = { &rpcmodrinit, &rpcmodwinit, NULL, NULL }; 253 254 struct xprt_style_ops { 255 int (*xo_open)(); 256 int (*xo_close)(); 257 void (*xo_wput)(); 258 void (*xo_wsrv)(); 259 void (*xo_rput)(); 260 void (*xo_rsrv)(); 261 }; 262 263 static struct xprt_style_ops xprt_clts_ops = { 264 rpcmodopen, 265 rpcmodclose, 266 rpcmodwput, 267 rpcmodwsrv, 268 rpcmodrput, 269 NULL 270 }; 271 272 static struct xprt_style_ops xprt_cots_ops = { 273 mir_open, 274 mir_close, 275 mir_wput, 276 mir_wsrv, 277 mir_rput, 278 mir_rsrv 279 }; 280 281 /* 282 * Per rpcmod "slot" data structure. q->q_ptr points to one of these. 283 */ 284 struct rpcm { 285 void *rm_krpc_cell; /* Reserved for use by KRPC */ 286 struct xprt_style_ops *rm_ops; 287 int rm_type; /* Client or server side stream */ 288 #define RM_CLOSING 0x1 /* somebody is trying to close slot */ 289 uint_t rm_state; /* state of the slot. see above */ 290 uint_t rm_ref; /* cnt of external references to slot */ 291 kmutex_t rm_lock; /* mutex protecting above fields */ 292 kcondvar_t rm_cwait; /* condition for closing */ 293 zoneid_t rm_zoneid; /* zone which pushed rpcmod */ 294 }; 295 296 struct temp_slot { 297 void *cell; 298 struct xprt_style_ops *ops; 299 int type; 300 mblk_t *info_ack; 301 kmutex_t lock; 302 kcondvar_t wait; 303 }; 304 305 typedef struct mir_s { 306 void *mir_krpc_cell; /* Reserved for KRPC use. This field */ 307 /* must be first in the structure. */ 308 struct xprt_style_ops *rm_ops; 309 int mir_type; /* Client or server side stream */ 310 311 mblk_t *mir_head_mp; /* RPC msg in progress */ 312 /* 313 * mir_head_mp points the first mblk being collected in 314 * the current RPC message. Record headers are removed 315 * before data is linked into mir_head_mp. 316 */ 317 mblk_t *mir_tail_mp; /* Last mblk in mir_head_mp */ 318 /* 319 * mir_tail_mp points to the last mblk in the message 320 * chain starting at mir_head_mp. It is only valid 321 * if mir_head_mp is non-NULL and is used to add new 322 * data blocks to the end of chain quickly. 323 */ 324 325 int32_t mir_frag_len; /* Bytes seen in the current frag */ 326 /* 327 * mir_frag_len starts at -4 for beginning of each fragment. 328 * When this length is negative, it indicates the number of 329 * bytes that rpcmod needs to complete the record marker 330 * header. When it is positive or zero, it holds the number 331 * of bytes that have arrived for the current fragment and 332 * are held in mir_header_mp. 333 */ 334 335 int32_t mir_frag_header; 336 /* 337 * Fragment header as collected for the current fragment. 338 * It holds the last-fragment indicator and the number 339 * of bytes in the fragment. 340 */ 341 342 unsigned int 343 mir_ordrel_pending : 1, /* Sent T_ORDREL_REQ */ 344 mir_hold_inbound : 1, /* Hold inbound messages on server */ 345 /* side until outbound flow control */ 346 /* is relieved. */ 347 mir_closing : 1, /* The stream is being closed */ 348 mir_inrservice : 1, /* data queued or rd srv proc running */ 349 mir_inwservice : 1, /* data queued or wr srv proc running */ 350 mir_inwflushdata : 1, /* flush M_DATAs when srv runs */ 351 /* 352 * On client streams, mir_clntreq is 0 or 1; it is set 353 * to 1 whenever a new request is sent out (mir_wput) 354 * and cleared when the timer fires (mir_timer). If 355 * the timer fires with this value equal to 0, then the 356 * stream is considered idle and KRPC is notified. 357 */ 358 mir_clntreq : 1, 359 /* 360 * On server streams, stop accepting messages 361 */ 362 mir_svc_no_more_msgs : 1, 363 mir_listen_stream : 1, /* listen end point */ 364 mir_unused : 1, /* no longer used */ 365 mir_timer_call : 1, 366 mir_junk_fill_thru_bit_31 : 21; 367 368 int mir_setup_complete; /* server has initialized everything */ 369 timeout_id_t mir_timer_id; /* Timer for idle checks */ 370 clock_t mir_idle_timeout; /* Allowed idle time before shutdown */ 371 /* 372 * This value is copied from clnt_idle_timeout or 373 * svc_idle_timeout during the appropriate ioctl. 374 * Kept in milliseconds 375 */ 376 clock_t mir_use_timestamp; /* updated on client with each use */ 377 /* 378 * This value is set to lbolt 379 * every time a client stream sends or receives data. 380 * Even if the timer message arrives, we don't shutdown 381 * client unless: 382 * lbolt >= MSEC_TO_TICK(mir_idle_timeout)+mir_use_timestamp. 383 * This value is kept in HZ. 384 */ 385 386 uint_t *mir_max_msg_sizep; /* Reference to sanity check size */ 387 /* 388 * This pointer is set to &clnt_max_msg_size or 389 * &svc_max_msg_size during the appropriate ioctl. 390 */ 391 zoneid_t mir_zoneid; /* zone which pushed rpcmod */ 392 /* Server-side fields. */ 393 int mir_ref_cnt; /* Reference count: server side only */ 394 /* counts the number of references */ 395 /* that a kernel RPC server thread */ 396 /* (see svc_run()) has on this rpcmod */ 397 /* slot. Effectively, it is the */ 398 /* number * of unprocessed messages */ 399 /* that have been passed up to the */ 400 /* KRPC layer */ 401 402 mblk_t *mir_svc_pend_mp; /* Pending T_ORDREL_IND or */ 403 /* T_DISCON_IND */ 404 405 /* 406 * these fields are for both client and server, but for debugging, 407 * it is easier to have these last in the structure. 408 */ 409 kmutex_t mir_mutex; /* Mutex and condvar for close */ 410 kcondvar_t mir_condvar; /* synchronization. */ 411 kcondvar_t mir_timer_cv; /* Timer routine sync. */ 412 void *mir_cb; /* For callbacks */ 413 } mir_t; 414 415 void tmp_rput(queue_t *q, mblk_t *mp); 416 417 struct xprt_style_ops tmpops = { 418 NULL, 419 NULL, 420 putnext, 421 NULL, 422 tmp_rput, 423 NULL 424 }; 425 426 void 427 tmp_rput(queue_t *q, mblk_t *mp) 428 { 429 struct temp_slot *t = (struct temp_slot *)(q->q_ptr); 430 struct T_info_ack *pptr; 431 432 switch (mp->b_datap->db_type) { 433 case M_PCPROTO: 434 pptr = (struct T_info_ack *)mp->b_rptr; 435 switch (pptr->PRIM_type) { 436 case T_INFO_ACK: 437 mutex_enter(&t->lock); 438 t->info_ack = mp; 439 cv_signal(&t->wait); 440 mutex_exit(&t->lock); 441 return; 442 default: 443 break; 444 } 445 default: 446 break; 447 } 448 449 /* 450 * Not an info-ack, so free it. This is ok because we should 451 * not be receiving data until the open finishes: rpcmod 452 * is pushed well before the end-point is bound to an address. 453 */ 454 freemsg(mp); 455 } 456 457 int 458 rmm_open(queue_t *q, dev_t *devp, int flag, int sflag, cred_t *crp) 459 { 460 mblk_t *bp; 461 struct temp_slot ts, *t; 462 struct T_info_ack *pptr; 463 int error = 0; 464 465 ASSERT(q != NULL); 466 /* 467 * Check for re-opens. 468 */ 469 if (q->q_ptr) { 470 TRACE_1(TR_FAC_KRPC, TR_RPCMODOPEN_END, 471 "rpcmodopen_end:(%s)", "q->qptr"); 472 return (0); 473 } 474 475 t = &ts; 476 bzero(t, sizeof (*t)); 477 q->q_ptr = (void *)t; 478 WR(q)->q_ptr = (void *)t; 479 480 /* 481 * Allocate the required messages upfront. 482 */ 483 if ((bp = allocb(sizeof (struct T_info_req) + 484 sizeof (struct T_info_ack), BPRI_LO)) == (mblk_t *)NULL) { 485 return (ENOBUFS); 486 } 487 488 mutex_init(&t->lock, NULL, MUTEX_DEFAULT, NULL); 489 cv_init(&t->wait, NULL, CV_DEFAULT, NULL); 490 491 t->ops = &tmpops; 492 493 qprocson(q); 494 bp->b_datap->db_type = M_PCPROTO; 495 *(int32_t *)bp->b_wptr = (int32_t)T_INFO_REQ; 496 bp->b_wptr += sizeof (struct T_info_req); 497 putnext(WR(q), bp); 498 499 mutex_enter(&t->lock); 500 while (t->info_ack == NULL) { 501 if (cv_wait_sig(&t->wait, &t->lock) == 0) { 502 error = EINTR; 503 break; 504 } 505 } 506 mutex_exit(&t->lock); 507 508 if (error) 509 goto out; 510 511 pptr = (struct T_info_ack *)t->info_ack->b_rptr; 512 513 if (pptr->SERV_type == T_CLTS) { 514 if ((error = rpcmodopen(q, devp, flag, sflag, crp)) == 0) 515 ((struct rpcm *)q->q_ptr)->rm_ops = &xprt_clts_ops; 516 } else { 517 if ((error = mir_open(q, devp, flag, sflag, crp)) == 0) 518 ((mir_t *)q->q_ptr)->rm_ops = &xprt_cots_ops; 519 } 520 521 out: 522 if (error) 523 qprocsoff(q); 524 525 freemsg(t->info_ack); 526 mutex_destroy(&t->lock); 527 cv_destroy(&t->wait); 528 529 return (error); 530 } 531 532 void 533 rmm_rput(queue_t *q, mblk_t *mp) 534 { 535 (*((struct temp_slot *)q->q_ptr)->ops->xo_rput)(q, mp); 536 } 537 538 void 539 rmm_rsrv(queue_t *q) 540 { 541 (*((struct temp_slot *)q->q_ptr)->ops->xo_rsrv)(q); 542 } 543 544 void 545 rmm_wput(queue_t *q, mblk_t *mp) 546 { 547 (*((struct temp_slot *)q->q_ptr)->ops->xo_wput)(q, mp); 548 } 549 550 void 551 rmm_wsrv(queue_t *q) 552 { 553 (*((struct temp_slot *)q->q_ptr)->ops->xo_wsrv)(q); 554 } 555 556 int 557 rmm_close(queue_t *q, int flag, cred_t *crp) 558 { 559 return ((*((struct temp_slot *)q->q_ptr)->ops->xo_close)(q, flag, crp)); 560 } 561 562 /* 563 * rpcmodopen - open routine gets called when the module gets pushed 564 * onto the stream. 565 */ 566 /*ARGSUSED*/ 567 int 568 rpcmodopen(queue_t *q, dev_t *devp, int flag, int sflag, cred_t *crp) 569 { 570 struct rpcm *rmp; 571 572 extern void (*rpc_rele)(queue_t *, mblk_t *); 573 static void rpcmod_release(queue_t *, mblk_t *); 574 575 TRACE_0(TR_FAC_KRPC, TR_RPCMODOPEN_START, "rpcmodopen_start:"); 576 577 /* 578 * Initialize entry points to release a rpcmod slot (and an input 579 * message if supplied) and to send an output message to the module 580 * below rpcmod. 581 */ 582 if (rpc_rele == NULL) 583 rpc_rele = rpcmod_release; 584 585 /* 586 * Only sufficiently privileged users can use this module, and it 587 * is assumed that they will use this module properly, and NOT send 588 * bulk data from downstream. 589 */ 590 if (secpolicy_rpcmod_open(crp) != 0) 591 return (EPERM); 592 593 /* 594 * Allocate slot data structure. 595 */ 596 rmp = kmem_zalloc(sizeof (*rmp), KM_SLEEP); 597 598 mutex_init(&rmp->rm_lock, NULL, MUTEX_DEFAULT, NULL); 599 cv_init(&rmp->rm_cwait, NULL, CV_DEFAULT, NULL); 600 rmp->rm_zoneid = rpc_zoneid(); 601 /* 602 * slot type will be set by kRPC client and server ioctl's 603 */ 604 rmp->rm_type = 0; 605 606 q->q_ptr = (void *)rmp; 607 WR(q)->q_ptr = (void *)rmp; 608 609 TRACE_1(TR_FAC_KRPC, TR_RPCMODOPEN_END, "rpcmodopen_end:(%s)", "end"); 610 return (0); 611 } 612 613 /* 614 * rpcmodclose - This routine gets called when the module gets popped 615 * off of the stream. 616 */ 617 /*ARGSUSED*/ 618 int 619 rpcmodclose(queue_t *q, int flag, cred_t *crp) 620 { 621 struct rpcm *rmp; 622 623 ASSERT(q != NULL); 624 rmp = (struct rpcm *)q->q_ptr; 625 626 /* 627 * Mark our state as closing. 628 */ 629 mutex_enter(&rmp->rm_lock); 630 rmp->rm_state |= RM_CLOSING; 631 632 /* 633 * Check and see if there are any messages on the queue. If so, send 634 * the messages, regardless whether the downstream module is ready to 635 * accept data. 636 */ 637 if (rmp->rm_type == RPC_SERVER) { 638 flushq(q, FLUSHDATA); 639 640 qenable(WR(q)); 641 642 if (rmp->rm_ref) { 643 mutex_exit(&rmp->rm_lock); 644 /* 645 * call into SVC to clean the queue 646 */ 647 svc_queueclean(q); 648 mutex_enter(&rmp->rm_lock); 649 650 /* 651 * Block while there are kRPC threads with a reference 652 * to this message. 653 */ 654 while (rmp->rm_ref) 655 cv_wait(&rmp->rm_cwait, &rmp->rm_lock); 656 } 657 658 mutex_exit(&rmp->rm_lock); 659 660 /* 661 * It is now safe to remove this queue from the stream. No kRPC 662 * threads have a reference to the stream, and none ever will, 663 * because RM_CLOSING is set. 664 */ 665 qprocsoff(q); 666 667 /* Notify kRPC that this stream is going away. */ 668 svc_queueclose(q); 669 } else { 670 mutex_exit(&rmp->rm_lock); 671 qprocsoff(q); 672 } 673 674 q->q_ptr = NULL; 675 WR(q)->q_ptr = NULL; 676 mutex_destroy(&rmp->rm_lock); 677 cv_destroy(&rmp->rm_cwait); 678 kmem_free(rmp, sizeof (*rmp)); 679 return (0); 680 } 681 682 #ifdef DEBUG 683 int rpcmod_send_msg_up = 0; 684 int rpcmod_send_uderr = 0; 685 int rpcmod_send_dup = 0; 686 int rpcmod_send_dup_cnt = 0; 687 #endif 688 689 /* 690 * rpcmodrput - Module read put procedure. This is called from 691 * the module, driver, or stream head downstream. 692 */ 693 void 694 rpcmodrput(queue_t *q, mblk_t *mp) 695 { 696 struct rpcm *rmp; 697 union T_primitives *pptr; 698 int hdrsz; 699 700 TRACE_0(TR_FAC_KRPC, TR_RPCMODRPUT_START, "rpcmodrput_start:"); 701 702 ASSERT(q != NULL); 703 rmp = (struct rpcm *)q->q_ptr; 704 705 if (rmp->rm_type == 0) { 706 freemsg(mp); 707 return; 708 } 709 710 #ifdef DEBUG 711 if (rpcmod_send_msg_up > 0) { 712 mblk_t *nmp = copymsg(mp); 713 if (nmp) { 714 putnext(q, nmp); 715 rpcmod_send_msg_up--; 716 } 717 } 718 if ((rpcmod_send_uderr > 0) && mp->b_datap->db_type == M_PROTO) { 719 mblk_t *nmp; 720 struct T_unitdata_ind *data; 721 struct T_uderror_ind *ud; 722 int d; 723 data = (struct T_unitdata_ind *)mp->b_rptr; 724 if (data->PRIM_type == T_UNITDATA_IND) { 725 d = sizeof (*ud) - sizeof (*data); 726 nmp = allocb(mp->b_wptr - mp->b_rptr + d, BPRI_HI); 727 if (nmp) { 728 ud = (struct T_uderror_ind *)nmp->b_rptr; 729 ud->PRIM_type = T_UDERROR_IND; 730 ud->DEST_length = data->SRC_length; 731 ud->DEST_offset = data->SRC_offset + d; 732 ud->OPT_length = data->OPT_length; 733 ud->OPT_offset = data->OPT_offset + d; 734 ud->ERROR_type = ENETDOWN; 735 if (data->SRC_length) { 736 bcopy(mp->b_rptr + 737 data->SRC_offset, 738 nmp->b_rptr + 739 ud->DEST_offset, 740 data->SRC_length); 741 } 742 if (data->OPT_length) { 743 bcopy(mp->b_rptr + 744 data->OPT_offset, 745 nmp->b_rptr + 746 ud->OPT_offset, 747 data->OPT_length); 748 } 749 nmp->b_wptr += d; 750 nmp->b_wptr += (mp->b_wptr - mp->b_rptr); 751 nmp->b_datap->db_type = M_PROTO; 752 putnext(q, nmp); 753 rpcmod_send_uderr--; 754 } 755 } 756 } 757 #endif 758 switch (mp->b_datap->db_type) { 759 default: 760 putnext(q, mp); 761 break; 762 763 case M_PROTO: 764 case M_PCPROTO: 765 ASSERT((mp->b_wptr - mp->b_rptr) >= sizeof (int32_t)); 766 pptr = (union T_primitives *)mp->b_rptr; 767 768 /* 769 * Forward this message to krpc if it is data. 770 */ 771 if (pptr->type == T_UNITDATA_IND) { 772 mblk_t *nmp; 773 774 /* 775 * Check if the module is being popped. 776 */ 777 mutex_enter(&rmp->rm_lock); 778 if (rmp->rm_state & RM_CLOSING) { 779 mutex_exit(&rmp->rm_lock); 780 putnext(q, mp); 781 break; 782 } 783 784 switch (rmp->rm_type) { 785 case RPC_CLIENT: 786 mutex_exit(&rmp->rm_lock); 787 hdrsz = mp->b_wptr - mp->b_rptr; 788 789 /* 790 * Make sure the header is sane. 791 */ 792 if (hdrsz < TUNITDATAINDSZ || 793 hdrsz < (pptr->unitdata_ind.OPT_length + 794 pptr->unitdata_ind.OPT_offset) || 795 hdrsz < (pptr->unitdata_ind.SRC_length + 796 pptr->unitdata_ind.SRC_offset)) { 797 freemsg(mp); 798 return; 799 } 800 801 /* 802 * Call clnt_clts_dispatch_notify, so that it 803 * can pass the message to the proper caller. 804 * Don't discard the header just yet since the 805 * client may need the sender's address. 806 */ 807 clnt_clts_dispatch_notify(mp, hdrsz, 808 rmp->rm_zoneid); 809 return; 810 case RPC_SERVER: 811 /* 812 * rm_krpc_cell is exclusively used by the kRPC 813 * CLTS server 814 */ 815 if (rmp->rm_krpc_cell) { 816 #ifdef DEBUG 817 /* 818 * Test duplicate request cache and 819 * rm_ref count handling by sending a 820 * duplicate every so often, if 821 * desired. 822 */ 823 if (rpcmod_send_dup && 824 rpcmod_send_dup_cnt++ % 825 rpcmod_send_dup) 826 nmp = copymsg(mp); 827 else 828 nmp = NULL; 829 #endif 830 /* 831 * Raise the reference count on this 832 * module to prevent it from being 833 * popped before krpc generates the 834 * reply. 835 */ 836 rmp->rm_ref++; 837 mutex_exit(&rmp->rm_lock); 838 839 /* 840 * Submit the message to krpc. 841 */ 842 svc_queuereq(q, mp); 843 #ifdef DEBUG 844 /* 845 * Send duplicate if we created one. 846 */ 847 if (nmp) { 848 mutex_enter(&rmp->rm_lock); 849 rmp->rm_ref++; 850 mutex_exit(&rmp->rm_lock); 851 svc_queuereq(q, nmp); 852 } 853 #endif 854 } else { 855 mutex_exit(&rmp->rm_lock); 856 freemsg(mp); 857 } 858 return; 859 default: 860 mutex_exit(&rmp->rm_lock); 861 freemsg(mp); 862 return; 863 } /* end switch(rmp->rm_type) */ 864 } else if (pptr->type == T_UDERROR_IND) { 865 mutex_enter(&rmp->rm_lock); 866 hdrsz = mp->b_wptr - mp->b_rptr; 867 868 /* 869 * Make sure the header is sane 870 */ 871 if (hdrsz < TUDERRORINDSZ || 872 hdrsz < (pptr->uderror_ind.OPT_length + 873 pptr->uderror_ind.OPT_offset) || 874 hdrsz < (pptr->uderror_ind.DEST_length + 875 pptr->uderror_ind.DEST_offset)) { 876 mutex_exit(&rmp->rm_lock); 877 freemsg(mp); 878 return; 879 } 880 881 /* 882 * In the case where a unit data error has been 883 * received, all we need to do is clear the message from 884 * the queue. 885 */ 886 mutex_exit(&rmp->rm_lock); 887 freemsg(mp); 888 RPCLOG(32, "rpcmodrput: unitdata error received at " 889 "%ld\n", gethrestime_sec()); 890 return; 891 } /* end else if (pptr->type == T_UDERROR_IND) */ 892 893 putnext(q, mp); 894 break; 895 } /* end switch (mp->b_datap->db_type) */ 896 897 TRACE_0(TR_FAC_KRPC, TR_RPCMODRPUT_END, 898 "rpcmodrput_end:"); 899 /* 900 * Return codes are not looked at by the STREAMS framework. 901 */ 902 } 903 904 /* 905 * write put procedure 906 */ 907 void 908 rpcmodwput(queue_t *q, mblk_t *mp) 909 { 910 struct rpcm *rmp; 911 912 ASSERT(q != NULL); 913 914 switch (mp->b_datap->db_type) { 915 case M_PROTO: 916 case M_PCPROTO: 917 break; 918 default: 919 rpcmodwput_other(q, mp); 920 return; 921 } 922 923 /* 924 * Check to see if we can send the message downstream. 925 */ 926 if (canputnext(q)) { 927 putnext(q, mp); 928 return; 929 } 930 931 rmp = (struct rpcm *)q->q_ptr; 932 ASSERT(rmp != NULL); 933 934 /* 935 * The first canputnext failed. Try again except this time with the 936 * lock held, so that we can check the state of the stream to see if 937 * it is closing. If either of these conditions evaluate to true 938 * then send the meesage. 939 */ 940 mutex_enter(&rmp->rm_lock); 941 if (canputnext(q) || (rmp->rm_state & RM_CLOSING)) { 942 mutex_exit(&rmp->rm_lock); 943 putnext(q, mp); 944 } else { 945 /* 946 * canputnext failed again and the stream is not closing. 947 * Place the message on the queue and let the service 948 * procedure handle the message. 949 */ 950 mutex_exit(&rmp->rm_lock); 951 (void) putq(q, mp); 952 } 953 } 954 955 static void 956 rpcmodwput_other(queue_t *q, mblk_t *mp) 957 { 958 struct rpcm *rmp; 959 struct iocblk *iocp; 960 961 rmp = (struct rpcm *)q->q_ptr; 962 ASSERT(rmp != NULL); 963 964 switch (mp->b_datap->db_type) { 965 case M_IOCTL: 966 iocp = (struct iocblk *)mp->b_rptr; 967 ASSERT(iocp != NULL); 968 switch (iocp->ioc_cmd) { 969 case RPC_CLIENT: 970 case RPC_SERVER: 971 mutex_enter(&rmp->rm_lock); 972 rmp->rm_type = iocp->ioc_cmd; 973 mutex_exit(&rmp->rm_lock); 974 mp->b_datap->db_type = M_IOCACK; 975 qreply(q, mp); 976 return; 977 default: 978 /* 979 * pass the ioctl downstream and hope someone 980 * down there knows how to handle it. 981 */ 982 putnext(q, mp); 983 return; 984 } 985 default: 986 break; 987 } 988 /* 989 * This is something we definitely do not know how to handle, just 990 * pass the message downstream 991 */ 992 putnext(q, mp); 993 } 994 995 /* 996 * Module write service procedure. This is called by downstream modules 997 * for back enabling during flow control. 998 */ 999 void 1000 rpcmodwsrv(queue_t *q) 1001 { 1002 struct rpcm *rmp; 1003 mblk_t *mp = NULL; 1004 1005 rmp = (struct rpcm *)q->q_ptr; 1006 ASSERT(rmp != NULL); 1007 1008 /* 1009 * Get messages that may be queued and send them down stream 1010 */ 1011 while ((mp = getq(q)) != NULL) { 1012 /* 1013 * Optimize the service procedure for the server-side, by 1014 * avoiding a call to canputnext(). 1015 */ 1016 if (rmp->rm_type == RPC_SERVER || canputnext(q)) { 1017 putnext(q, mp); 1018 continue; 1019 } 1020 (void) putbq(q, mp); 1021 return; 1022 } 1023 } 1024 1025 static void 1026 rpcmod_release(queue_t *q, mblk_t *bp) 1027 { 1028 struct rpcm *rmp; 1029 1030 /* 1031 * For now, just free the message. 1032 */ 1033 if (bp) 1034 freemsg(bp); 1035 rmp = (struct rpcm *)q->q_ptr; 1036 1037 mutex_enter(&rmp->rm_lock); 1038 rmp->rm_ref--; 1039 1040 if (rmp->rm_ref == 0 && (rmp->rm_state & RM_CLOSING)) { 1041 cv_broadcast(&rmp->rm_cwait); 1042 } 1043 1044 mutex_exit(&rmp->rm_lock); 1045 } 1046 1047 /* 1048 * This part of rpcmod is pushed on a connection-oriented transport for use 1049 * by RPC. It serves to bypass the Stream head, implements 1050 * the record marking protocol, and dispatches incoming RPC messages. 1051 */ 1052 1053 /* Default idle timer values */ 1054 #define MIR_CLNT_IDLE_TIMEOUT (5 * (60 * 1000L)) /* 5 minutes */ 1055 #define MIR_SVC_IDLE_TIMEOUT (6 * (60 * 1000L)) /* 6 minutes */ 1056 #define MIR_SVC_ORDREL_TIMEOUT (10 * (60 * 1000L)) /* 10 minutes */ 1057 #define MIR_LASTFRAG 0x80000000 /* Record marker */ 1058 1059 #define DLEN(mp) (mp->b_cont ? msgdsize(mp) : (mp->b_wptr - mp->b_rptr)) 1060 1061 #define MIR_SVC_QUIESCED(mir) \ 1062 (mir->mir_ref_cnt == 0 && mir->mir_inrservice == 0) 1063 1064 #define MIR_CLEAR_INRSRV(mir_ptr) { \ 1065 (mir_ptr)->mir_inrservice = 0; \ 1066 if ((mir_ptr)->mir_type == RPC_SERVER && \ 1067 (mir_ptr)->mir_closing) \ 1068 cv_signal(&(mir_ptr)->mir_condvar); \ 1069 } 1070 1071 /* 1072 * Don't block service procedure (and mir_close) if 1073 * we are in the process of closing. 1074 */ 1075 #define MIR_WCANPUTNEXT(mir_ptr, write_q) \ 1076 (canputnext(write_q) || ((mir_ptr)->mir_svc_no_more_msgs == 1)) 1077 1078 static int mir_clnt_dup_request(queue_t *q, mblk_t *mp); 1079 static void mir_rput_proto(queue_t *q, mblk_t *mp); 1080 static int mir_svc_policy_notify(queue_t *q, int event); 1081 static void mir_svc_release(queue_t *wq, mblk_t *mp); 1082 static void mir_svc_start(queue_t *wq); 1083 static void mir_svc_idle_start(queue_t *, mir_t *); 1084 static void mir_svc_idle_stop(queue_t *, mir_t *); 1085 static void mir_svc_start_close(queue_t *, mir_t *); 1086 static void mir_clnt_idle_do_stop(queue_t *); 1087 static void mir_clnt_idle_stop(queue_t *, mir_t *); 1088 static void mir_clnt_idle_start(queue_t *, mir_t *); 1089 static void mir_wput(queue_t *q, mblk_t *mp); 1090 static void mir_wput_other(queue_t *q, mblk_t *mp); 1091 static void mir_wsrv(queue_t *q); 1092 static void mir_disconnect(queue_t *, mir_t *ir); 1093 static int mir_check_len(queue_t *, int32_t, mblk_t *); 1094 static void mir_timer(void *); 1095 1096 extern void (*mir_rele)(queue_t *, mblk_t *); 1097 extern void (*mir_start)(queue_t *); 1098 extern void (*clnt_stop_idle)(queue_t *); 1099 1100 clock_t clnt_idle_timeout = MIR_CLNT_IDLE_TIMEOUT; 1101 clock_t svc_idle_timeout = MIR_SVC_IDLE_TIMEOUT; 1102 1103 /* 1104 * Timeout for subsequent notifications of idle connection. This is 1105 * typically used to clean up after a wedged orderly release. 1106 */ 1107 clock_t svc_ordrel_timeout = MIR_SVC_ORDREL_TIMEOUT; /* milliseconds */ 1108 1109 extern uint_t *clnt_max_msg_sizep; 1110 extern uint_t *svc_max_msg_sizep; 1111 uint_t clnt_max_msg_size = RPC_MAXDATASIZE; 1112 uint_t svc_max_msg_size = RPC_MAXDATASIZE; 1113 uint_t mir_krpc_cell_null; 1114 1115 uint32_t cb_live = 0; 1116 1117 /* 1118 * XXXsessions 1119 * This global is used to control which path we take in mir_set_cbinfo 1120 * when we detect the race between nfs41_callback_thread() and 1121 * nfs4_sequence_heartbeat_thread(). 1122 * If it's non-zero, we simply print a message. 1123 * If it's zero, we bump cb_live and wait for the condition to clear. 1124 * Unfortunately, we tend to hang during data server recovery. 1125 * The sessions team is working on a fix for this. 1126 */ 1127 int mir_set_cbinfo_hack = 1; 1128 1129 void 1130 mir_set_cbinfo(queue_t *wq, void *info) 1131 { 1132 mir_t *mir = (mir_t *)wq->q_ptr; 1133 struct __svccb *scb = mir->mir_cb; 1134 1135 if (scb != NULL) { 1136 if (mir_set_cbinfo_hack) { 1137 /* 1138 * This is a hack to prevent the hang we get 1139 * w/DS Recovery 1140 */ 1141 cmn_err(CE_WARN, "mir_set_cbinfo: scb != NULL"); 1142 } else { 1143 /* 1144 * XXX 1145 * Race condition between nfs41_callback_thread() 1146 * and nfs4_sequence_heartbeat_thread() is causing 1147 * us to hit ASSERT, since heartbeat thread reaches 1148 * this code before r_flags is marked as SVCCB_DEAD. 1149 * 1150 * Need to revisit this; NFS layer shouldn't be 1151 * cv_wait()'ing the RPC layer. 1152 */ 1153 mutex_enter(&scb->r_lock); 1154 while (!(scb->r_flags & SVCCB_DEAD)) { 1155 cb_live++; 1156 cv_wait(&scb->r_cbwait, &scb->r_lock); 1157 } 1158 mutex_exit(&scb->r_lock); 1159 kmem_free(scb, sizeof (*scb)); 1160 mir->mir_cb = NULL; 1161 } 1162 } 1163 mir->mir_cb = (void *)info; 1164 } 1165 1166 void 1167 mir_clear_cbinfo(queue_t *wq) 1168 { 1169 mir_t *mir = (mir_t *)wq->q_ptr; 1170 struct __svccb *scb = mir->mir_cb; 1171 1172 if (scb == NULL) 1173 return; 1174 1175 if (scb->r_flags & SVCCB_DEAD) { 1176 kmem_free(scb, sizeof (*scb)); 1177 } 1178 mir->mir_cb = NULL; 1179 } 1180 1181 void 1182 mir_check_cb(void *handlecb, queue_t *wq) 1183 { 1184 ASSERT(handlecb == ((mir_t *)wq->q_ptr)->mir_cb); 1185 } 1186 1187 1188 SVCCB * 1189 mir_get_svccb(queue_t *wq) 1190 { 1191 mir_t *mir; 1192 mir = (mir_t *)wq->q_ptr; 1193 return ((SVCCB *)mir->mir_cb); 1194 } 1195 1196 static void 1197 mir_timer_stop(mir_t *mir) 1198 { 1199 timeout_id_t tid; 1200 1201 ASSERT(MUTEX_HELD(&mir->mir_mutex)); 1202 1203 /* 1204 * Since the mir_mutex lock needs to be released to call 1205 * untimeout(), we need to make sure that no other thread 1206 * can start/stop the timer (changing mir_timer_id) during 1207 * that time. The mir_timer_call bit and the mir_timer_cv 1208 * condition variable are used to synchronize this. Setting 1209 * mir_timer_call also tells mir_timer() (refer to the comments 1210 * in mir_timer()) that it does not need to do anything. 1211 */ 1212 while (mir->mir_timer_call) 1213 cv_wait(&mir->mir_timer_cv, &mir->mir_mutex); 1214 mir->mir_timer_call = B_TRUE; 1215 1216 if ((tid = mir->mir_timer_id) != 0) { 1217 mir->mir_timer_id = 0; 1218 mutex_exit(&mir->mir_mutex); 1219 (void) untimeout(tid); 1220 mutex_enter(&mir->mir_mutex); 1221 } 1222 mir->mir_timer_call = B_FALSE; 1223 cv_broadcast(&mir->mir_timer_cv); 1224 } 1225 1226 static void 1227 mir_timer_start(queue_t *q, mir_t *mir, clock_t intrvl) 1228 { 1229 timeout_id_t tid; 1230 1231 ASSERT(MUTEX_HELD(&mir->mir_mutex)); 1232 1233 while (mir->mir_timer_call) 1234 cv_wait(&mir->mir_timer_cv, &mir->mir_mutex); 1235 mir->mir_timer_call = B_TRUE; 1236 1237 if ((tid = mir->mir_timer_id) != 0) { 1238 mutex_exit(&mir->mir_mutex); 1239 (void) untimeout(tid); 1240 mutex_enter(&mir->mir_mutex); 1241 } 1242 /* Only start the timer when it is not closing. */ 1243 if (!mir->mir_closing) { 1244 mir->mir_timer_id = timeout(mir_timer, q, 1245 MSEC_TO_TICK(intrvl)); 1246 } 1247 mir->mir_timer_call = B_FALSE; 1248 cv_broadcast(&mir->mir_timer_cv); 1249 } 1250 1251 static int 1252 mir_clnt_dup_request(queue_t *q, mblk_t *mp) 1253 { 1254 mblk_t *mp1; 1255 uint32_t new_xid; 1256 uint32_t old_xid; 1257 1258 ASSERT(MUTEX_HELD(&((mir_t *)q->q_ptr)->mir_mutex)); 1259 new_xid = BE32_TO_U32(&mp->b_rptr[4]); 1260 /* 1261 * This loop is a bit tacky -- it walks the STREAMS list of 1262 * flow-controlled messages. 1263 */ 1264 if ((mp1 = q->q_first) != NULL) { 1265 do { 1266 old_xid = BE32_TO_U32(&mp1->b_rptr[4]); 1267 if (new_xid == old_xid) 1268 return (1); 1269 } while ((mp1 = mp1->b_next) != NULL); 1270 } 1271 return (0); 1272 } 1273 1274 static int 1275 mir_close(queue_t *q) 1276 { 1277 mir_t *mir = q->q_ptr; 1278 mblk_t *mp; 1279 bool_t queue_cleaned = FALSE; 1280 1281 RPCLOG(32, "rpcmod: mir_close of q 0x%p\n", (void *)q); 1282 ASSERT(MUTEX_NOT_HELD(&mir->mir_mutex)); 1283 mutex_enter(&mir->mir_mutex); 1284 if ((mp = mir->mir_head_mp) != NULL) { 1285 mir->mir_head_mp = NULL; 1286 mir->mir_tail_mp = NULL; 1287 freemsg(mp); 1288 } 1289 /* 1290 * Set mir_closing so we get notified when MIR_SVC_QUIESCED() 1291 * is TRUE. And mir_timer_start() won't start the timer again. 1292 */ 1293 mir->mir_closing = B_TRUE; 1294 mir_timer_stop(mir); 1295 1296 if (mir->mir_type == RPC_SERVER) { 1297 flushq(q, FLUSHDATA); /* Ditch anything waiting on read q */ 1298 1299 /* 1300 * This will prevent more requests from arriving and 1301 * will force rpcmod to ignore flow control. 1302 */ 1303 mir_svc_start_close(WR(q), mir); 1304 1305 while ((!MIR_SVC_QUIESCED(mir)) || mir->mir_inwservice == 1) { 1306 1307 if (mir->mir_ref_cnt && !mir->mir_inrservice && 1308 (queue_cleaned == FALSE)) { 1309 /* 1310 * call into SVC to clean the queue 1311 */ 1312 mutex_exit(&mir->mir_mutex); 1313 svc_queueclean(q); 1314 queue_cleaned = TRUE; 1315 mutex_enter(&mir->mir_mutex); 1316 continue; 1317 } 1318 1319 /* 1320 * Bugid 1253810 - Force the write service 1321 * procedure to send its messages, regardless 1322 * whether the downstream module is ready 1323 * to accept data. 1324 */ 1325 if (mir->mir_inwservice == 1) 1326 qenable(WR(q)); 1327 1328 cv_wait(&mir->mir_condvar, &mir->mir_mutex); 1329 } 1330 1331 mutex_exit(&mir->mir_mutex); 1332 /* 1333 * Destroy the cm_entry 1334 */ 1335 connmgr_cb_destroy(WR(q)); 1336 qprocsoff(q); 1337 1338 /* Notify KRPC that this stream is going away. */ 1339 svc_queueclose(q); 1340 } else { 1341 mutex_exit(&mir->mir_mutex); 1342 qprocsoff(q); 1343 } 1344 1345 mutex_destroy(&mir->mir_mutex); 1346 cv_destroy(&mir->mir_condvar); 1347 cv_destroy(&mir->mir_timer_cv); 1348 kmem_free(mir, sizeof (mir_t)); 1349 return (0); 1350 } 1351 1352 /* 1353 * This is server side only (RPC_SERVER). 1354 * 1355 * Exit idle mode. 1356 */ 1357 static void 1358 mir_svc_idle_stop(queue_t *q, mir_t *mir) 1359 { 1360 ASSERT(MUTEX_HELD(&mir->mir_mutex)); 1361 ASSERT((q->q_flag & QREADR) == 0); 1362 ASSERT(mir->mir_type == RPC_SERVER); 1363 RPCLOG(16, "rpcmod: mir_svc_idle_stop of q 0x%p\n", (void *)q); 1364 1365 mir_timer_stop(mir); 1366 } 1367 1368 /* 1369 * This is server side only (RPC_SERVER). 1370 * 1371 * Start idle processing, which will include setting idle timer if the 1372 * stream is not being closed. 1373 */ 1374 static void 1375 mir_svc_idle_start(queue_t *q, mir_t *mir) 1376 { 1377 ASSERT(MUTEX_HELD(&mir->mir_mutex)); 1378 ASSERT((q->q_flag & QREADR) == 0); 1379 ASSERT(mir->mir_type == RPC_SERVER); 1380 RPCLOG(16, "rpcmod: mir_svc_idle_start q 0x%p\n", (void *)q); 1381 1382 /* 1383 * Don't re-start idle timer if we are closing queues. 1384 */ 1385 if (mir->mir_closing) { 1386 RPCLOG(16, "mir_svc_idle_start - closing: 0x%p\n", 1387 (void *)q); 1388 1389 /* 1390 * We will call mir_svc_idle_start() whenever MIR_SVC_QUIESCED() 1391 * is true. When it is true, and we are in the process of 1392 * closing the stream, signal any thread waiting in 1393 * mir_close(). 1394 */ 1395 if (mir->mir_inwservice == 0) 1396 cv_signal(&mir->mir_condvar); 1397 1398 } else { 1399 RPCLOG(16, "mir_svc_idle_start - reset %s timer\n", 1400 mir->mir_ordrel_pending ? "ordrel" : "normal"); 1401 /* 1402 * Normal condition, start the idle timer. If an orderly 1403 * release has been sent, set the timeout to wait for the 1404 * client to close its side of the connection. Otherwise, 1405 * use the normal idle timeout. 1406 */ 1407 mir_timer_start(q, mir, mir->mir_ordrel_pending ? 1408 svc_ordrel_timeout : mir->mir_idle_timeout); 1409 } 1410 } 1411 1412 /* ARGSUSED */ 1413 static int 1414 mir_open(queue_t *q, dev_t *devp, int flag, int sflag, cred_t *credp) 1415 { 1416 mir_t *mir; 1417 1418 RPCLOG(32, "rpcmod: mir_open of q 0x%p\n", (void *)q); 1419 /* Set variables used directly by KRPC. */ 1420 if (!mir_rele) 1421 mir_rele = mir_svc_release; 1422 if (!mir_start) 1423 mir_start = mir_svc_start; 1424 if (!clnt_stop_idle) 1425 clnt_stop_idle = mir_clnt_idle_do_stop; 1426 if (!clnt_max_msg_sizep) 1427 clnt_max_msg_sizep = &clnt_max_msg_size; 1428 if (!svc_max_msg_sizep) 1429 svc_max_msg_sizep = &svc_max_msg_size; 1430 1431 /* Allocate a zero'ed out mir structure for this stream. */ 1432 mir = kmem_zalloc(sizeof (mir_t), KM_SLEEP); 1433 1434 /* 1435 * We set hold inbound here so that incoming messages will 1436 * be held on the read-side queue until the stream is completely 1437 * initialized with a RPC_CLIENT or RPC_SERVER ioctl. During 1438 * the ioctl processing, the flag is cleared and any messages that 1439 * arrived between the open and the ioctl are delivered to KRPC. 1440 * 1441 * Early data should never arrive on a client stream since 1442 * servers only respond to our requests and we do not send any. 1443 * until after the stream is initialized. Early data is 1444 * very common on a server stream where the client will start 1445 * sending data as soon as the connection is made (and this 1446 * is especially true with TCP where the protocol accepts the 1447 * connection before nfsd or KRPC is notified about it). 1448 */ 1449 1450 mir->mir_hold_inbound = 1; 1451 1452 /* 1453 * Start the record marker looking for a 4-byte header. When 1454 * this length is negative, it indicates that rpcmod is looking 1455 * for bytes to consume for the record marker header. When it 1456 * is positive, it holds the number of bytes that have arrived 1457 * for the current fragment and are being held in mir_header_mp. 1458 */ 1459 1460 mir->mir_frag_len = -(int32_t)sizeof (uint32_t); 1461 1462 mir->mir_zoneid = rpc_zoneid(); 1463 mutex_init(&mir->mir_mutex, NULL, MUTEX_DEFAULT, NULL); 1464 cv_init(&mir->mir_condvar, NULL, CV_DRIVER, NULL); 1465 cv_init(&mir->mir_timer_cv, NULL, CV_DRIVER, NULL); 1466 1467 q->q_ptr = (char *)mir; 1468 WR(q)->q_ptr = (char *)mir; 1469 1470 /* 1471 * We noenable the read-side queue because we don't want it 1472 * automatically enabled by putq. We enable it explicitly 1473 * in mir_wsrv when appropriate. (See additional comments on 1474 * flow control at the beginning of mir_rsrv.) 1475 */ 1476 noenable(q); 1477 1478 qprocson(q); 1479 return (0); 1480 } 1481 void 1482 mir_queue_rele(queue_t *q) 1483 { 1484 mir_t *mir; 1485 1486 ASSERT(q != NULL); 1487 mir = (mir_t *)q->q_ptr; 1488 ASSERT(mir != NULL); 1489 1490 mutex_enter(&mir->mir_mutex); 1491 mir->mir_ref_cnt--; 1492 mutex_exit(&mir->mir_mutex); 1493 } 1494 1495 void 1496 mir_queue_hold(queue_t *q) 1497 { 1498 mir_t *mir; 1499 1500 ASSERT(q != NULL); 1501 mir = (mir_t *)q->q_ptr; 1502 ASSERT(mir != NULL); 1503 1504 mutex_enter(&mir->mir_mutex); 1505 mir->mir_ref_cnt++; 1506 mutex_exit(&mir->mir_mutex); 1507 } 1508 1509 /* 1510 * Copy out the RPC transaction id and RPC Direction 1511 * from the mblk chain. Leave the mblk intact. 1512 */ 1513 bool_t 1514 mir_dir_xid(mblk_t *mp, uint32_t *dir, uint32_t *xid) 1515 { 1516 unsigned char *p; 1517 unsigned char *rptr; 1518 mblk_t *tmp; 1519 int i, get_rpcdir; 1520 uint32_t d_tmp = 0; 1521 1522 /* 1523 * If we can just grab the XID and RPC direction flag great. 1524 */ 1525 if ((IS_P2ALIGNED(mp->b_rptr, (sizeof (uint64_t)))) && 1526 (mp->b_wptr - mp->b_rptr) >= (sizeof (uint64_t))) { 1527 *xid = *((uint32_t *)mp->b_rptr); 1528 *dir = ntohl(*((uint32_t *)(mp->b_rptr + sizeof (uint32_t)))); 1529 return (TRUE); 1530 } 1531 1532 /* 1533 * Otherwise we need to copy byte-by-byte 1534 */ 1535 DTRACE_PROBE(krpc__i__bytecopy); 1536 1537 i = get_rpcdir = 0; 1538 p = (unsigned char *)xid; 1539 tmp = mp; 1540 1541 /* 1542 * While we have not exhausted the entire mblk chain: 1543 * copy the first sizeof uint32_t value into xid, and 1544 * then the second sizeof uint32_t value into a temporary 1545 * so that we can convert from network byte order. 1546 * 1547 * Should we exhaust the entire mblk chain in attempting 1548 * to do this, return FALSE. 1549 */ 1550 while (tmp) { 1551 rptr = tmp->b_rptr; 1552 while (rptr < tmp->b_wptr) { 1553 *p++ = *rptr++; 1554 /* 1555 * Have we collected enough bytes for 1556 * a uint32_t ? 1557 */ 1558 if (++i == sizeof (uint32_t)) { 1559 /* 1560 * If yes, do we need to switch to 1561 * RPC Direction or are we all done ? 1562 */ 1563 if (get_rpcdir) { 1564 /* Got it all */ 1565 *dir = ntohl(d_tmp); 1566 return (TRUE); 1567 } 1568 /* start to collect RPC Direction */ 1569 get_rpcdir++; 1570 i = 0; 1571 p = (unsigned char *)&d_tmp; 1572 } 1573 } 1574 tmp = tmp->b_cont; 1575 } 1576 1577 /* We didn't get both of them.. */ 1578 DTRACE_PROBE(krpc__e__mblk_exhausted); 1579 return (FALSE); 1580 } 1581 1582 /* 1583 * Read-side put routine for both the client and server side. Does the 1584 * record marking for incoming RPC messages, and when complete, dispatches 1585 * the message to either the client or server. 1586 */ 1587 static void 1588 mir_rput(queue_t *q, mblk_t *mp) 1589 { 1590 int excess; 1591 int32_t frag_len, frag_header; 1592 mblk_t *cont_mp, *head_mp, *tail_mp, *mp1; 1593 mir_t *mir = q->q_ptr; 1594 boolean_t stop_timer = B_FALSE; 1595 uint32_t xid; 1596 uint32_t dir; 1597 1598 ASSERT(mir != NULL); 1599 1600 /* 1601 * If the stream has not been set up as a RPC_CLIENT or RPC_SERVER 1602 * with the corresponding ioctl, then don't accept 1603 * any inbound data. This should never happen for streams 1604 * created by nfsd or client-side KRPC because they are careful 1605 * to set the mode of the stream before doing anything else. 1606 */ 1607 if (mir->mir_type == 0) { 1608 freemsg(mp); 1609 return; 1610 } 1611 1612 ASSERT(MUTEX_NOT_HELD(&mir->mir_mutex)); 1613 1614 switch (mp->b_datap->db_type) { 1615 case M_DATA: 1616 break; 1617 case M_PROTO: 1618 case M_PCPROTO: 1619 if (MBLKL(mp) < sizeof (t_scalar_t)) { 1620 RPCLOG(1, "mir_rput: runt TPI message (%d bytes)\n", 1621 (int)MBLKL(mp)); 1622 freemsg(mp); 1623 return; 1624 } 1625 if (((union T_primitives *)mp->b_rptr)->type != T_DATA_IND) { 1626 mir_rput_proto(q, mp); 1627 return; 1628 } 1629 1630 /* Throw away the T_DATA_IND block and continue with data. */ 1631 mp1 = mp; 1632 mp = mp->b_cont; 1633 freeb(mp1); 1634 break; 1635 case M_SETOPTS: 1636 /* 1637 * If a module on the stream is trying set the Stream head's 1638 * high water mark, then set our hiwater to the requested 1639 * value. We are the "stream head" for all inbound 1640 * data messages since messages are passed directly to KRPC. 1641 */ 1642 if (MBLKL(mp) >= sizeof (struct stroptions)) { 1643 struct stroptions *stropts; 1644 1645 stropts = (struct stroptions *)mp->b_rptr; 1646 if ((stropts->so_flags & SO_HIWAT) && 1647 !(stropts->so_flags & SO_BAND)) { 1648 (void) strqset(q, QHIWAT, 0, stropts->so_hiwat); 1649 } 1650 } 1651 putnext(q, mp); 1652 return; 1653 case M_FLUSH: 1654 RPCLOG(32, "mir_rput: ignoring M_FLUSH %x ", *mp->b_rptr); 1655 RPCLOG(32, "on q 0x%p\n", (void *)q); 1656 putnext(q, mp); 1657 return; 1658 default: 1659 putnext(q, mp); 1660 return; 1661 } 1662 1663 mutex_enter(&mir->mir_mutex); 1664 1665 /* 1666 * If this connection is closing, don't accept any new messages. 1667 */ 1668 if (mir->mir_svc_no_more_msgs) { 1669 ASSERT(mir->mir_type == RPC_SERVER); 1670 mutex_exit(&mir->mir_mutex); 1671 freemsg(mp); 1672 return; 1673 } 1674 1675 /* Get local copies for quicker access. */ 1676 frag_len = mir->mir_frag_len; 1677 frag_header = mir->mir_frag_header; 1678 head_mp = mir->mir_head_mp; 1679 tail_mp = mir->mir_tail_mp; 1680 1681 /* Loop, processing each message block in the mp chain separately. */ 1682 do { 1683 cont_mp = mp->b_cont; 1684 mp->b_cont = NULL; 1685 1686 /* 1687 * Drop zero-length mblks to prevent unbounded kernel memory 1688 * consumption. 1689 */ 1690 if (MBLKL(mp) == 0) { 1691 freeb(mp); 1692 continue; 1693 } 1694 1695 /* 1696 * If frag_len is negative, we're still in the process of 1697 * building frag_header -- try to complete it with this mblk. 1698 */ 1699 while (frag_len < 0 && mp->b_rptr < mp->b_wptr) { 1700 frag_len++; 1701 frag_header <<= 8; 1702 frag_header += *mp->b_rptr++; 1703 } 1704 1705 if (MBLKL(mp) == 0 && frag_len < 0) { 1706 /* 1707 * We consumed this mblk while trying to complete the 1708 * fragment header. Free it and move on. 1709 */ 1710 freeb(mp); 1711 continue; 1712 } 1713 1714 ASSERT(frag_len >= 0); 1715 1716 /* 1717 * Now frag_header has the number of bytes in this fragment 1718 * and we're just waiting to collect them all. Chain our 1719 * latest mblk onto the list and see if we now have enough 1720 * bytes to complete the fragment. 1721 */ 1722 if (head_mp == NULL) { 1723 ASSERT(tail_mp == NULL); 1724 head_mp = tail_mp = mp; 1725 } else { 1726 tail_mp->b_cont = mp; 1727 tail_mp = mp; 1728 } 1729 1730 frag_len += MBLKL(mp); 1731 excess = frag_len - (frag_header & ~MIR_LASTFRAG); 1732 if (excess < 0) { 1733 /* 1734 * We still haven't received enough data to complete 1735 * the fragment, so continue on to the next mblk. 1736 */ 1737 continue; 1738 } 1739 1740 /* 1741 * We've got a complete fragment. If there are excess bytes, 1742 * then they're part of the next fragment's header (of either 1743 * this RPC message or the next RPC message). Split that part 1744 * into its own mblk so that we can safely freeb() it when 1745 * building frag_header above. 1746 */ 1747 if (excess > 0) { 1748 if ((mp1 = dupb(mp)) == NULL && 1749 (mp1 = copyb(mp)) == NULL) { 1750 freemsg(head_mp); 1751 freemsg(cont_mp); 1752 RPCLOG0(1, "mir_rput: dupb/copyb failed\n"); 1753 mir->mir_frag_header = 0; 1754 mir->mir_frag_len = -(int32_t)sizeof (uint32_t); 1755 mir->mir_head_mp = NULL; 1756 mir->mir_tail_mp = NULL; 1757 mir_disconnect(q, mir); /* drops mir_mutex */ 1758 return; 1759 } 1760 1761 /* 1762 * Relink the message chain so that the next mblk is 1763 * the next fragment header, followed by the rest of 1764 * the message chain. 1765 */ 1766 mp1->b_cont = cont_mp; 1767 cont_mp = mp1; 1768 1769 /* 1770 * Data in the new mblk begins at the next fragment, 1771 * and data in the old mblk ends at the next fragment. 1772 */ 1773 mp1->b_rptr = mp1->b_wptr - excess; 1774 mp->b_wptr -= excess; 1775 } 1776 1777 /* 1778 * Reset frag_len and frag_header for the next fragment. 1779 */ 1780 frag_len = -(int32_t)sizeof (uint32_t); 1781 if (!(frag_header & MIR_LASTFRAG)) { 1782 /* 1783 * The current fragment is complete, but more 1784 * fragments need to be processed before we can 1785 * pass along the RPC message headed at head_mp. 1786 */ 1787 frag_header = 0; 1788 continue; 1789 } 1790 frag_header = 0; 1791 1792 /* 1793 * Get msg direction and handle to the appropriate ctxt 1794 */ 1795 if (!mir_dir_xid(head_mp, &dir, &xid)) { 1796 /* XXX - if we can't get the dir, we're hosed */ 1797 mutex_exit(&mir->mir_mutex); 1798 freemsg(head_mp); 1799 return; 1800 } 1801 1802 /* 1803 * We've got a complete RPC message; pass it to the 1804 * appropriate consumer. 1805 */ 1806 switch (mir->mir_type) { 1807 case RPC_CLIENT: 1808 switch (dir) { 1809 case REPLY: 1810 if (clnt_dispatch_notify(head_mp, 1811 mir->mir_zoneid, xid)) { 1812 /* 1813 * Mark this stream as active. This marker 1814 * is used in mir_timer(). 1815 */ 1816 mir->mir_clntreq = 1; 1817 mir->mir_use_timestamp = lbolt; 1818 } else 1819 freemsg(head_mp); 1820 break; 1821 1822 case CALL: 1823 default: 1824 { 1825 SVCCB *svccb; 1826 ASSERT(dir == CALL); 1827 1828 svccb = (SVCCB *)mir->mir_cb; 1829 if (svccb->r_flags & SVCCB_DEAD) { 1830 zcmn_err(getzoneid(), CE_NOTE, 1831 "Callback On Dead Session" 1832 "%p %p", (void *)mir, 1833 (void *)svccb); 1834 } else { 1835 svccb->r_mp = head_mp; 1836 cv_signal(&svccb->r_cbwait); 1837 break; 1838 } 1839 } 1840 1841 } 1842 break; 1843 1844 case RPC_SERVER: 1845 switch (dir) { 1846 case REPLY: 1847 /* 1848 * RPC Server initiated a Callback RPC and 1849 * is receiving a reply from the RPC Client. 1850 */ 1851 if (clnt_dispatch_notify(head_mp, 1852 global_zone->zone_id, xid)) { 1853 /* 1854 * Mark this stream as active. 1855 * This marker is used in mir_timer(). 1856 */ 1857 mir->mir_clntreq = 0; 1858 } else 1859 freemsg(head_mp); 1860 break; 1861 1862 case CALL: 1863 default: 1864 /* 1865 * Check for flow control before 1866 * passing the message to KRPC. 1867 */ 1868 if (!mir->mir_hold_inbound) { 1869 if (mir->mir_krpc_cell) { 1870 /* 1871 * If the reference count is 0 1872 * (not including this request), 1873 * then the stream is transitioning 1874 * from idle to non-idle. In this case, 1875 * we cancel the idle timer. 1876 */ 1877 if (mir->mir_ref_cnt++ == 0) 1878 stop_timer = B_TRUE; 1879 if (mir_check_len(q, 1880 (int32_t)msgdsize(mp), mp)) 1881 return; 1882 svc_queuereq(q, head_mp); /* to KRPC */ 1883 } else { 1884 /* 1885 * Count # of times this happens. 1886 * Should be never, but experience 1887 * shows otherwise. 1888 */ 1889 mir_krpc_cell_null++; 1890 freemsg(head_mp); 1891 } 1892 1893 } else { 1894 /* 1895 * If the outbound side of the stream 1896 * is flow controlled, then hold this 1897 * message until client catches up. 1898 * mir_hold_inbound is set in mir_wput 1899 * and cleared in mir_wsrv. 1900 */ 1901 (void) putq(q, head_mp); 1902 mir->mir_inrservice = B_TRUE; 1903 } 1904 break; 1905 } 1906 break; /* RPC_SERVER */ 1907 1908 default: 1909 RPCLOG(1, "mir_rput: unknown mir_type %d\n", 1910 mir->mir_type); 1911 freemsg(head_mp); 1912 break; 1913 } 1914 1915 /* 1916 * Reset the chain since we're starting on a new RPC message. 1917 */ 1918 head_mp = tail_mp = NULL; 1919 } while ((mp = cont_mp) != NULL); 1920 1921 /* 1922 * Sanity check the message length; if it's too large mir_check_len() 1923 * will shutdown the connection, drop mir_mutex, and return non-zero. 1924 */ 1925 if (head_mp != NULL && mir->mir_setup_complete && 1926 mir_check_len(q, frag_len, head_mp)) 1927 return; 1928 1929 /* Save our local copies back in the mir structure. */ 1930 mir->mir_frag_header = frag_header; 1931 mir->mir_frag_len = frag_len; 1932 mir->mir_head_mp = head_mp; 1933 mir->mir_tail_mp = tail_mp; 1934 1935 /* 1936 * The timer is stopped after the whole message chain is processed. 1937 * The reason is that stopping the timer releases the mir_mutex 1938 * lock temporarily. This means that the request can be serviced 1939 * while we are still processing the message chain. This is not 1940 * good. So we stop the timer here instead. 1941 * 1942 * Note that if the timer fires before we stop it, it will not 1943 * do any harm as MIR_SVC_QUIESCED() is false and mir_timer() 1944 * will just return. 1945 */ 1946 if (stop_timer) { 1947 RPCLOG(16, "mir_rput: stopping idle timer on 0x%p because " 1948 "ref cnt going to non zero\n", (void *)WR(q)); 1949 mir_svc_idle_stop(WR(q), mir); 1950 } 1951 mutex_exit(&mir->mir_mutex); 1952 } 1953 1954 static void 1955 mir_rput_proto(queue_t *q, mblk_t *mp) 1956 { 1957 mir_t *mir = (mir_t *)q->q_ptr; 1958 uint32_t type; 1959 uint32_t reason = 0; 1960 1961 ASSERT(MUTEX_NOT_HELD(&mir->mir_mutex)); 1962 1963 type = ((union T_primitives *)mp->b_rptr)->type; 1964 switch (mir->mir_type) { 1965 case RPC_CLIENT: 1966 switch (type) { 1967 case T_DISCON_IND: 1968 reason = ((struct T_discon_ind *) 1969 (mp->b_rptr))->DISCON_reason; 1970 /*FALLTHROUGH*/ 1971 case T_ORDREL_IND: 1972 mutex_enter(&mir->mir_mutex); 1973 if (mir->mir_head_mp) { 1974 freemsg(mir->mir_head_mp); 1975 mir->mir_head_mp = (mblk_t *)0; 1976 mir->mir_tail_mp = (mblk_t *)0; 1977 } 1978 /* 1979 * We are disconnecting, but not necessarily 1980 * closing. By not closing, we will fail to 1981 * pick up a possibly changed global timeout value, 1982 * unless we store it now. 1983 */ 1984 mir->mir_idle_timeout = clnt_idle_timeout; 1985 mir_clnt_idle_stop(WR(q), mir); 1986 1987 /* 1988 * Even though we are unconnected, we still 1989 * leave the idle timer going on the client. The 1990 * reason for is that if we've disconnected due 1991 * to a server-side disconnect, reset, or connection 1992 * timeout, there is a possibility the client may 1993 * retry the RPC request. This retry needs to done on 1994 * the same bound address for the server to interpret 1995 * it as such. However, we don't want 1996 * to wait forever for that possibility. If the 1997 * end-point stays unconnected for mir_idle_timeout 1998 * units of time, then that is a signal to the 1999 * connection manager to give up waiting for the 2000 * application (eg. NFS) to send a retry. 2001 */ 2002 mir_clnt_idle_start(WR(q), mir); 2003 mutex_exit(&mir->mir_mutex); 2004 clnt_dispatch_notifyall(WR(q), type, reason); 2005 freemsg(mp); 2006 return; 2007 case T_ERROR_ACK: 2008 { 2009 struct T_error_ack *terror; 2010 2011 terror = (struct T_error_ack *)mp->b_rptr; 2012 RPCLOG(1, "mir_rput_proto T_ERROR_ACK for queue 0x%p", 2013 (void *)q); 2014 RPCLOG(1, " ERROR_prim: %s,", 2015 rpc_tpiprim2name(terror->ERROR_prim)); 2016 RPCLOG(1, " TLI_error: %s,", 2017 rpc_tpierr2name(terror->TLI_error)); 2018 RPCLOG(1, " UNIX_error: %d\n", terror->UNIX_error); 2019 if (terror->ERROR_prim == T_DISCON_REQ) { 2020 clnt_dispatch_notifyall(WR(q), type, reason); 2021 freemsg(mp); 2022 return; 2023 } else { 2024 if (clnt_dispatch_notifyconn(WR(q), mp)) 2025 return; 2026 } 2027 break; 2028 } 2029 case T_OK_ACK: 2030 { 2031 struct T_ok_ack *tok = (struct T_ok_ack *)mp->b_rptr; 2032 2033 if (tok->CORRECT_prim == T_DISCON_REQ) { 2034 clnt_dispatch_notifyall(WR(q), type, reason); 2035 freemsg(mp); 2036 return; 2037 } else { 2038 if (clnt_dispatch_notifyconn(WR(q), mp)) 2039 return; 2040 } 2041 break; 2042 } 2043 case T_CONN_CON: 2044 case T_INFO_ACK: 2045 case T_OPTMGMT_ACK: 2046 if (clnt_dispatch_notifyconn(WR(q), mp)) 2047 return; 2048 break; 2049 case T_BIND_ACK: 2050 break; 2051 default: 2052 RPCLOG(1, "mir_rput: unexpected message %d " 2053 "for KRPC client\n", 2054 ((union T_primitives *)mp->b_rptr)->type); 2055 break; 2056 } 2057 break; 2058 2059 case RPC_SERVER: 2060 switch (type) { 2061 case T_BIND_ACK: 2062 { 2063 struct T_bind_ack *tbind; 2064 2065 /* 2066 * If this is a listening stream, then shut 2067 * off the idle timer. 2068 */ 2069 tbind = (struct T_bind_ack *)mp->b_rptr; 2070 if (tbind->CONIND_number > 0) { 2071 mutex_enter(&mir->mir_mutex); 2072 mir_svc_idle_stop(WR(q), mir); 2073 2074 /* 2075 * mark this as a listen endpoint 2076 * for special handling. 2077 */ 2078 2079 mir->mir_listen_stream = 1; 2080 mutex_exit(&mir->mir_mutex); 2081 } 2082 break; 2083 } 2084 case T_DISCON_IND: 2085 case T_ORDREL_IND: 2086 RPCLOG(16, "mir_rput_proto: got %s indication\n", 2087 type == T_DISCON_IND ? "disconnect" 2088 : "orderly release"); 2089 2090 /* 2091 * For listen endpoint just pass 2092 * on the message. 2093 */ 2094 2095 if (mir->mir_listen_stream) 2096 break; 2097 2098 2099 mutex_enter(&mir->mir_mutex); 2100 2101 /* 2102 * If client wants to break off connection, record 2103 * that fact. 2104 */ 2105 mir_svc_start_close(WR(q), mir); 2106 2107 /* 2108 * If we are idle, then send the orderly release 2109 * or disconnect indication to nfsd. 2110 */ 2111 if (MIR_SVC_QUIESCED(mir)) { 2112 mutex_exit(&mir->mir_mutex); 2113 break; 2114 } 2115 2116 RPCLOG(16, "mir_rput_proto: not idle, so " 2117 "disconnect/ord rel indication not passed " 2118 "upstream on 0x%p\n", (void *)q); 2119 2120 /* 2121 * Hold the indication until we get idle 2122 * If there already is an indication stored, 2123 * replace it if the new one is a disconnect. The 2124 * reasoning is that disconnection takes less time 2125 * to process, and once a client decides to 2126 * disconnect, we should do that. 2127 */ 2128 if (mir->mir_svc_pend_mp) { 2129 if (type == T_DISCON_IND) { 2130 RPCLOG(16, "mir_rput_proto: replacing" 2131 " held disconnect/ord rel" 2132 " indication with disconnect on" 2133 " 0x%p\n", (void *)q); 2134 2135 freemsg(mir->mir_svc_pend_mp); 2136 mir->mir_svc_pend_mp = mp; 2137 } else { 2138 RPCLOG(16, "mir_rput_proto: already " 2139 "held a disconnect/ord rel " 2140 "indication. freeing ord rel " 2141 "ind on 0x%p\n", (void *)q); 2142 freemsg(mp); 2143 } 2144 } else 2145 mir->mir_svc_pend_mp = mp; 2146 2147 mutex_exit(&mir->mir_mutex); 2148 return; 2149 2150 default: 2151 /* nfsd handles server-side non-data messages. */ 2152 break; 2153 } 2154 break; 2155 2156 default: 2157 break; 2158 } 2159 2160 putnext(q, mp); 2161 } 2162 2163 /* 2164 * The server-side read queues are used to hold inbound messages while 2165 * outbound flow control is exerted. When outbound flow control is 2166 * relieved, mir_wsrv qenables the read-side queue. Read-side queues 2167 * are not enabled by STREAMS and are explicitly noenable'ed in mir_open. 2168 * 2169 * For the server side, we have two types of messages queued. The first type 2170 * are messages that are ready to be XDR decoded and and then sent to the 2171 * RPC program's dispatch routine. The second type are "raw" messages that 2172 * haven't been processed, i.e. assembled from rpc record fragements into 2173 * full requests. The only time we will see the second type of message 2174 * queued is if we have a memory allocation failure while processing a 2175 * a raw message. The field mir_first_non_processed_mblk will mark the 2176 * first such raw message. So the flow for server side is: 2177 * 2178 * - send processed queued messages to kRPC until we run out or find 2179 * one that needs additional processing because we were short on memory 2180 * earlier 2181 * - process a message that was deferred because of lack of 2182 * memory 2183 * - continue processing messages until the queue empties or we 2184 * have to stop because of lack of memory 2185 * - during each of the above phase, if the queue is empty and 2186 * there are no pending messages that were passed to the RPC 2187 * layer, send upstream the pending disconnect/ordrel indication if 2188 * there is one 2189 * 2190 * The read-side queue is also enabled by a bufcall callback if dupmsg 2191 * fails in mir_rput. 2192 */ 2193 static void 2194 mir_rsrv(queue_t *q) 2195 { 2196 mir_t *mir; 2197 mblk_t *mp; 2198 mblk_t *cmp = NULL; 2199 boolean_t stop_timer = B_FALSE; 2200 2201 mir = (mir_t *)q->q_ptr; 2202 mutex_enter(&mir->mir_mutex); 2203 2204 mp = NULL; 2205 switch (mir->mir_type) { 2206 case RPC_SERVER: 2207 if (mir->mir_ref_cnt == 0) 2208 mir->mir_hold_inbound = 0; 2209 if (mir->mir_hold_inbound) { 2210 2211 ASSERT(cmp == NULL); 2212 if (q->q_first == NULL) { 2213 2214 MIR_CLEAR_INRSRV(mir); 2215 2216 if (MIR_SVC_QUIESCED(mir)) { 2217 cmp = mir->mir_svc_pend_mp; 2218 mir->mir_svc_pend_mp = NULL; 2219 } 2220 } 2221 2222 mutex_exit(&mir->mir_mutex); 2223 2224 if (cmp != NULL) { 2225 RPCLOG(16, "mir_rsrv: line %d: sending a held " 2226 "disconnect/ord rel indication upstream\n", 2227 __LINE__); 2228 putnext(q, cmp); 2229 } 2230 2231 return; 2232 } 2233 while (mp = getq(q)) { 2234 if (mir->mir_krpc_cell && 2235 (mir->mir_svc_no_more_msgs == 0)) { 2236 /* 2237 * If we were idle, turn off idle timer since 2238 * we aren't idle any more. 2239 */ 2240 if (mir->mir_ref_cnt++ == 0) 2241 stop_timer = B_TRUE; 2242 if (mir_check_len(q, 2243 (int32_t)msgdsize(mp), mp)) 2244 return; 2245 svc_queuereq(q, mp); 2246 } else { 2247 /* 2248 * Count # of times this happens. Should be 2249 * never, but experience shows otherwise. 2250 */ 2251 if (mir->mir_krpc_cell == NULL) 2252 mir_krpc_cell_null++; 2253 freemsg(mp); 2254 } 2255 } 2256 break; 2257 case RPC_CLIENT: 2258 break; 2259 default: 2260 RPCLOG(1, "mir_rsrv: unexpected mir_type %d\n", mir->mir_type); 2261 2262 if (q->q_first == NULL) 2263 MIR_CLEAR_INRSRV(mir); 2264 2265 mutex_exit(&mir->mir_mutex); 2266 2267 return; 2268 } 2269 2270 /* 2271 * The timer is stopped after all the messages are processed. 2272 * The reason is that stopping the timer releases the mir_mutex 2273 * lock temporarily. This means that the request can be serviced 2274 * while we are still processing the message queue. This is not 2275 * good. So we stop the timer here instead. 2276 */ 2277 if (stop_timer) { 2278 RPCLOG(16, "mir_rsrv stopping idle timer on 0x%p because ref " 2279 "cnt going to non zero\n", (void *)WR(q)); 2280 mir_svc_idle_stop(WR(q), mir); 2281 } 2282 2283 if (q->q_first == NULL) { 2284 2285 MIR_CLEAR_INRSRV(mir); 2286 2287 ASSERT(cmp == NULL); 2288 if (mir->mir_type == RPC_SERVER && MIR_SVC_QUIESCED(mir)) { 2289 cmp = mir->mir_svc_pend_mp; 2290 mir->mir_svc_pend_mp = NULL; 2291 } 2292 2293 mutex_exit(&mir->mir_mutex); 2294 2295 if (cmp != NULL) { 2296 RPCLOG(16, "mir_rsrv: line %d: sending a held " 2297 "disconnect/ord rel indication upstream\n", 2298 __LINE__); 2299 putnext(q, cmp); 2300 } 2301 2302 return; 2303 } 2304 mutex_exit(&mir->mir_mutex); 2305 } 2306 2307 static int mir_svc_policy_fails; 2308 2309 /* 2310 * Called to send an event code to nfsd/lockd so that it initiates 2311 * connection close. 2312 */ 2313 static int 2314 mir_svc_policy_notify(queue_t *q, int event) 2315 { 2316 mblk_t *mp; 2317 #ifdef DEBUG 2318 mir_t *mir = (mir_t *)q->q_ptr; 2319 ASSERT(MUTEX_NOT_HELD(&mir->mir_mutex)); 2320 #endif 2321 ASSERT(q->q_flag & QREADR); 2322 2323 /* 2324 * Create an M_DATA message with the event code and pass it to the 2325 * Stream head (nfsd or whoever created the stream will consume it). 2326 */ 2327 mp = allocb(sizeof (int), BPRI_HI); 2328 2329 if (!mp) { 2330 2331 mir_svc_policy_fails++; 2332 RPCLOG(16, "mir_svc_policy_notify: could not allocate event " 2333 "%d\n", event); 2334 return (ENOMEM); 2335 } 2336 2337 U32_TO_BE32(event, mp->b_rptr); 2338 mp->b_wptr = mp->b_rptr + sizeof (int); 2339 putnext(q, mp); 2340 return (0); 2341 } 2342 2343 /* 2344 * Server side: start the close phase. We want to get this rpcmod slot in an 2345 * idle state before mir_close() is called. 2346 */ 2347 static void 2348 mir_svc_start_close(queue_t *wq, mir_t *mir) 2349 { 2350 ASSERT(MUTEX_HELD(&mir->mir_mutex)); 2351 ASSERT((wq->q_flag & QREADR) == 0); 2352 ASSERT(mir->mir_type == RPC_SERVER); 2353 2354 2355 /* 2356 * Do not accept any more messages. 2357 */ 2358 mir->mir_svc_no_more_msgs = 1; 2359 2360 /* 2361 * Next two statements will make the read service procedure invoke 2362 * svc_queuereq() on everything stuck in the streams read queue. 2363 * It's not necessary because enabling the write queue will 2364 * have the same effect, but why not speed the process along? 2365 */ 2366 mir->mir_hold_inbound = 0; 2367 qenable(RD(wq)); 2368 2369 /* 2370 * Meanwhile force the write service procedure to send the 2371 * responses downstream, regardless of flow control. 2372 */ 2373 qenable(wq); 2374 } 2375 2376 /* 2377 * This routine is called directly by KRPC after a request is completed, 2378 * whether a reply was sent or the request was dropped. 2379 */ 2380 static void 2381 mir_svc_release(queue_t *wq, mblk_t *mp) 2382 { 2383 mir_t *mir = (mir_t *)wq->q_ptr; 2384 mblk_t *cmp = NULL; 2385 2386 ASSERT((wq->q_flag & QREADR) == 0); 2387 if (mp) 2388 freemsg(mp); 2389 2390 mutex_enter(&mir->mir_mutex); 2391 2392 /* 2393 * Start idle processing if this is the last reference. 2394 */ 2395 if ((mir->mir_ref_cnt == 1) && (mir->mir_inrservice == 0)) { 2396 cmp = mir->mir_svc_pend_mp; 2397 mir->mir_svc_pend_mp = NULL; 2398 } 2399 2400 if (cmp) { 2401 RPCLOG(16, "mir_svc_release: sending a held " 2402 "disconnect/ord rel indication upstream on queue 0x%p\n", 2403 (void *)RD(wq)); 2404 2405 mutex_exit(&mir->mir_mutex); 2406 2407 putnext(RD(wq), cmp); 2408 2409 mutex_enter(&mir->mir_mutex); 2410 } 2411 2412 /* 2413 * Start idle processing if this is the last reference. 2414 */ 2415 if (mir->mir_ref_cnt == 1 && mir->mir_inrservice == 0) { 2416 2417 RPCLOG(16, "mir_svc_release starting idle timer on 0x%p " 2418 "because ref cnt is zero\n", (void *) wq); 2419 2420 mir_svc_idle_start(wq, mir); 2421 } 2422 2423 mir->mir_ref_cnt--; 2424 ASSERT(mir->mir_ref_cnt >= 0); 2425 2426 /* 2427 * Wake up the thread waiting to close. 2428 */ 2429 2430 if ((mir->mir_ref_cnt == 0) && mir->mir_closing) 2431 cv_signal(&mir->mir_condvar); 2432 2433 mutex_exit(&mir->mir_mutex); 2434 } 2435 2436 /* 2437 * This routine is called by server-side KRPC when it is ready to 2438 * handle inbound messages on the stream. 2439 */ 2440 static void 2441 mir_svc_start(queue_t *wq) 2442 { 2443 mir_t *mir = (mir_t *)wq->q_ptr; 2444 2445 /* 2446 * no longer need to take the mir_mutex because the 2447 * mir_setup_complete field has been moved out of 2448 * the binary field protected by the mir_mutex. 2449 */ 2450 2451 mir->mir_setup_complete = 1; 2452 qenable(RD(wq)); 2453 } 2454 2455 /* 2456 * client side wrapper for stopping timer with normal idle timeout. 2457 */ 2458 static void 2459 mir_clnt_idle_stop(queue_t *wq, mir_t *mir) 2460 { 2461 ASSERT(MUTEX_HELD(&mir->mir_mutex)); 2462 ASSERT((wq->q_flag & QREADR) == 0); 2463 ASSERT(mir->mir_type == RPC_CLIENT); 2464 2465 mir_timer_stop(mir); 2466 } 2467 2468 /* 2469 * client side wrapper for stopping timer with normal idle timeout. 2470 */ 2471 static void 2472 mir_clnt_idle_start(queue_t *wq, mir_t *mir) 2473 { 2474 ASSERT(MUTEX_HELD(&mir->mir_mutex)); 2475 ASSERT((wq->q_flag & QREADR) == 0); 2476 ASSERT(mir->mir_type == RPC_CLIENT); 2477 2478 mir_timer_start(wq, mir, mir->mir_idle_timeout); 2479 } 2480 2481 /* 2482 * client side only. Forces rpcmod to stop sending T_ORDREL_REQs on 2483 * end-points that aren't connected. 2484 */ 2485 static void 2486 mir_clnt_idle_do_stop(queue_t *wq) 2487 { 2488 mir_t *mir = (mir_t *)wq->q_ptr; 2489 2490 RPCLOG(1, "mir_clnt_idle_do_stop: wq 0x%p\n", (void *)wq); 2491 ASSERT(MUTEX_NOT_HELD(&mir->mir_mutex)); 2492 mutex_enter(&mir->mir_mutex); 2493 mir_clnt_idle_stop(wq, mir); 2494 mutex_exit(&mir->mir_mutex); 2495 } 2496 2497 /* 2498 * Timer handler. It handles idle timeout and memory shortage problem. 2499 */ 2500 static void 2501 mir_timer(void *arg) 2502 { 2503 queue_t *wq = (queue_t *)arg; 2504 mir_t *mir = (mir_t *)wq->q_ptr; 2505 boolean_t notify; 2506 2507 mutex_enter(&mir->mir_mutex); 2508 2509 /* 2510 * mir_timer_call is set only when either mir_timer_[start|stop] 2511 * is progressing. And mir_timer() can only be run while they 2512 * are progressing if the timer is being stopped. So just 2513 * return. 2514 */ 2515 if (mir->mir_timer_call) { 2516 mutex_exit(&mir->mir_mutex); 2517 return; 2518 } 2519 mir->mir_timer_id = 0; 2520 2521 switch (mir->mir_type) { 2522 case RPC_CLIENT: 2523 2524 /* 2525 * For clients, the timer fires at clnt_idle_timeout 2526 * intervals. If the activity marker (mir_clntreq) is 2527 * zero, then the stream has been idle since the last 2528 * timer event and we notify KRPC. If mir_clntreq is 2529 * non-zero, then the stream is active and we just 2530 * restart the timer for another interval. mir_clntreq 2531 * is set to 1 in mir_wput for every request passed 2532 * downstream. 2533 * 2534 * If this was a memory shortage timer reset the idle 2535 * timeout regardless; the mir_clntreq will not be a 2536 * valid indicator. 2537 * 2538 * The timer is initially started in mir_wput during 2539 * RPC_CLIENT ioctl processing. 2540 * 2541 * The timer interval can be changed for individual 2542 * streams with the ND variable "mir_idle_timeout". 2543 */ 2544 if (mir->mir_clntreq > 0 && mir->mir_use_timestamp + 2545 MSEC_TO_TICK(mir->mir_idle_timeout) - lbolt >= 0) { 2546 clock_t tout; 2547 2548 tout = mir->mir_idle_timeout - 2549 TICK_TO_MSEC(lbolt - mir->mir_use_timestamp); 2550 if (tout < 0) 2551 tout = 1000; 2552 #if 0 2553 printf("mir_timer[%d < %d + %d]: reset client timer " 2554 "to %d (ms)\n", TICK_TO_MSEC(lbolt), 2555 TICK_TO_MSEC(mir->mir_use_timestamp), 2556 mir->mir_idle_timeout, tout); 2557 #endif 2558 mir->mir_clntreq = 0; 2559 mir_timer_start(wq, mir, tout); 2560 mutex_exit(&mir->mir_mutex); 2561 return; 2562 } 2563 #if 0 2564 printf("mir_timer[%d]: doing client timeout\n", lbolt / hz); 2565 #endif 2566 /* 2567 * We are disconnecting, but not necessarily 2568 * closing. By not closing, we will fail to 2569 * pick up a possibly changed global timeout value, 2570 * unless we store it now. 2571 */ 2572 mir->mir_idle_timeout = clnt_idle_timeout; 2573 mir_clnt_idle_start(wq, mir); 2574 2575 mutex_exit(&mir->mir_mutex); 2576 /* 2577 * We pass T_ORDREL_REQ as an integer value 2578 * to KRPC as the indication that the stream 2579 * is idle. This is not a T_ORDREL_REQ message, 2580 * it is just a convenient value since we call 2581 * the same KRPC routine for T_ORDREL_INDs and 2582 * T_DISCON_INDs. 2583 */ 2584 clnt_dispatch_notifyall(wq, T_ORDREL_REQ, 0); 2585 return; 2586 2587 case RPC_SERVER: 2588 2589 /* 2590 * For servers, the timer is only running when the stream 2591 * is really idle or memory is short. The timer is started 2592 * by mir_wput when mir_type is set to RPC_SERVER and 2593 * by mir_svc_idle_start whenever the stream goes idle 2594 * (mir_ref_cnt == 0). The timer is cancelled in 2595 * mir_rput whenever a new inbound request is passed to KRPC 2596 * and the stream was previously idle. 2597 * 2598 * The timer interval can be changed for individual 2599 * streams with the ND variable "mir_idle_timeout". 2600 * 2601 * If the stream is not idle do nothing. 2602 */ 2603 if (!MIR_SVC_QUIESCED(mir)) { 2604 mutex_exit(&mir->mir_mutex); 2605 return; 2606 } 2607 2608 notify = !mir->mir_inrservice; 2609 mutex_exit(&mir->mir_mutex); 2610 2611 /* 2612 * If there is no packet queued up in read queue, the stream 2613 * is really idle so notify nfsd to close it. 2614 */ 2615 if (notify) { 2616 RPCLOG(16, "mir_timer: telling stream head listener " 2617 "to close stream (0x%p)\n", (void *) RD(wq)); 2618 (void) mir_svc_policy_notify(RD(wq), 1); 2619 } 2620 return; 2621 default: 2622 RPCLOG(1, "mir_timer: unexpected mir_type %d\n", 2623 mir->mir_type); 2624 mutex_exit(&mir->mir_mutex); 2625 return; 2626 } 2627 } 2628 2629 /* 2630 * Called by the RPC package to send either a call or a return, or a 2631 * transport connection request. Adds the record marking header. 2632 */ 2633 static void 2634 mir_wput(queue_t *q, mblk_t *mp) 2635 { 2636 uint_t frag_header; 2637 mir_t *mir = (mir_t *)q->q_ptr; 2638 uchar_t *rptr = mp->b_rptr; 2639 uint32_t xid; 2640 uint32_t dir; 2641 2642 if (!mir) { 2643 freemsg(mp); 2644 return; 2645 } 2646 2647 if (mp->b_datap->db_type != M_DATA) { 2648 mir_wput_other(q, mp); 2649 return; 2650 } 2651 2652 if (mir->mir_ordrel_pending == 1) { 2653 freemsg(mp); 2654 RPCLOG(16, "mir_wput wq 0x%p: got data after T_ORDREL_REQ\n", 2655 (void *)q); 2656 return; 2657 } 2658 2659 frag_header = (uint_t)DLEN(mp); 2660 frag_header |= MIR_LASTFRAG; 2661 2662 /* Stick in the 4 byte record marking header. */ 2663 if ((rptr - mp->b_datap->db_base) < sizeof (uint32_t) || 2664 !IS_P2ALIGNED(mp->b_rptr, sizeof (uint32_t))) { 2665 /* 2666 * Since we know that M_DATA messages are created exclusively 2667 * by KRPC, we expect that KRPC will leave room for our header 2668 * and 4 byte align which is normal for XDR. 2669 * If KRPC (or someone else) does not cooperate, then we 2670 * just throw away the message. 2671 */ 2672 RPCLOG(1, "mir_wput: KRPC did not leave space for record " 2673 "fragment header (%d bytes left)\n", 2674 (int)(rptr - mp->b_datap->db_base)); 2675 freemsg(mp); 2676 return; 2677 } 2678 rptr -= sizeof (uint32_t); 2679 *(uint32_t *)rptr = htonl(frag_header); 2680 mp->b_rptr = rptr; 2681 2682 mutex_enter(&mir->mir_mutex); 2683 if (mir->mir_type == RPC_CLIENT) { 2684 /* 2685 * For the client, set mir_clntreq to indicate that the 2686 * connection is active. 2687 */ 2688 mir->mir_clntreq = 1; 2689 mir->mir_use_timestamp = lbolt; 2690 } 2691 2692 /* 2693 * If we haven't already queued some data and the downstream module 2694 * can accept more data, send it on, otherwise we queue the message 2695 * and take other actions depending on mir_type. 2696 */ 2697 if (!mir->mir_inwservice && MIR_WCANPUTNEXT(mir, q)) { 2698 mutex_exit(&mir->mir_mutex); 2699 2700 /* 2701 * Now we pass the RPC message downstream. 2702 */ 2703 putnext(q, mp); 2704 return; 2705 } 2706 2707 /* 2708 * Get msg direction and handle to the appropriate ctxt 2709 */ 2710 if (!mir_dir_xid(mp, &dir, &xid)) { 2711 /* XXX - if we can't get the dir, we're hosed */ 2712 mutex_exit(&mir->mir_mutex); 2713 freemsg(mp); 2714 return; 2715 } 2716 2717 switch (mir->mir_type) { 2718 case RPC_CLIENT: 2719 /* 2720 * Check for a previous duplicate request on the 2721 * queue. If there is one, then we throw away 2722 * the current message and let the previous one 2723 * go through. If we can't find a duplicate, then 2724 * send this one. This tap dance is an effort 2725 * to reduce traffic and processing requirements 2726 * under load conditions. 2727 */ 2728 if (mir_clnt_dup_request(q, mp)) { 2729 mutex_exit(&mir->mir_mutex); 2730 freemsg(mp); 2731 return; 2732 } 2733 break; 2734 2735 case RPC_SERVER: 2736 switch (dir) { 2737 case CALL: 2738 /* 2739 * RPC Server doing Callball RPC 2740 */ 2741 if (mir_clnt_dup_request(q, mp)) { 2742 mutex_exit(&mir->mir_mutex); 2743 freemsg(mp); 2744 return; 2745 } 2746 break; 2747 2748 case REPLY: 2749 default: 2750 /* 2751 * Set mir_hold_inbound so that new inbound RPC 2752 * messages will be held until the client catches 2753 * up on the earlier replies. This flag is cleared 2754 * in mir_wsrv after flow control is relieved; 2755 * the read-side queue is also enabled at that time. 2756 */ 2757 mir->mir_hold_inbound = 1; 2758 break; 2759 } 2760 break; 2761 default: 2762 RPCLOG(1, "mir_wput: unexpected mir_type %d\n", mir->mir_type); 2763 break; 2764 } 2765 mir->mir_inwservice = 1; 2766 (void) putq(q, mp); 2767 mutex_exit(&mir->mir_mutex); 2768 } 2769 2770 static void 2771 mir_wput_other(queue_t *q, mblk_t *mp) 2772 { 2773 mir_t *mir = (mir_t *)q->q_ptr; 2774 struct iocblk *iocp; 2775 uchar_t *rptr = mp->b_rptr; 2776 bool_t flush_in_svc = FALSE; 2777 2778 ASSERT(MUTEX_NOT_HELD(&mir->mir_mutex)); 2779 switch (mp->b_datap->db_type) { 2780 case M_IOCTL: 2781 iocp = (struct iocblk *)rptr; 2782 switch (iocp->ioc_cmd) { 2783 case RPC_CLIENT: 2784 mutex_enter(&mir->mir_mutex); 2785 if (mir->mir_type != 0 && 2786 mir->mir_type != iocp->ioc_cmd) { 2787 ioc_eperm: 2788 mutex_exit(&mir->mir_mutex); 2789 iocp->ioc_error = EPERM; 2790 iocp->ioc_count = 0; 2791 mp->b_datap->db_type = M_IOCACK; 2792 qreply(q, mp); 2793 return; 2794 } 2795 2796 mir->mir_type = iocp->ioc_cmd; 2797 2798 /* 2799 * Clear mir_hold_inbound which was set to 1 by 2800 * mir_open. This flag is not used on client 2801 * streams. 2802 */ 2803 mir->mir_hold_inbound = 0; 2804 mir->mir_max_msg_sizep = &clnt_max_msg_size; 2805 2806 /* 2807 * Start the idle timer. See mir_timer() for more 2808 * information on how client timers work. 2809 */ 2810 mir->mir_idle_timeout = clnt_idle_timeout; 2811 mir_clnt_idle_start(q, mir); 2812 mutex_exit(&mir->mir_mutex); 2813 2814 mp->b_datap->db_type = M_IOCACK; 2815 qreply(q, mp); 2816 return; 2817 case RPC_SERVER: 2818 mutex_enter(&mir->mir_mutex); 2819 if (mir->mir_type != 0 && 2820 mir->mir_type != iocp->ioc_cmd) 2821 goto ioc_eperm; 2822 2823 /* 2824 * We don't clear mir_hold_inbound here because 2825 * mir_hold_inbound is used in the flow control 2826 * model. If we cleared it here, then we'd commit 2827 * a small violation to the model where the transport 2828 * might immediately block downstream flow. 2829 */ 2830 2831 mir->mir_type = iocp->ioc_cmd; 2832 mir->mir_max_msg_sizep = &svc_max_msg_size; 2833 2834 /* 2835 * Start the idle timer. See mir_timer() for more 2836 * information on how server timers work. 2837 * 2838 * Note that it is important to start the idle timer 2839 * here so that connections time out even if we 2840 * never receive any data on them. 2841 */ 2842 mir->mir_idle_timeout = svc_idle_timeout; 2843 RPCLOG(16, "mir_wput_other starting idle timer on 0x%p " 2844 "because we got RPC_SERVER ioctl\n", (void *)q); 2845 mir_svc_idle_start(q, mir); 2846 mutex_exit(&mir->mir_mutex); 2847 2848 mp->b_datap->db_type = M_IOCACK; 2849 qreply(q, mp); 2850 return; 2851 default: 2852 break; 2853 } 2854 break; 2855 2856 case M_PROTO: 2857 if (mir->mir_type == RPC_CLIENT) { 2858 /* 2859 * We are likely being called from the context of a 2860 * service procedure. So we need to enqueue. However 2861 * enqueing may put our message behind data messages. 2862 * So flush the data first. 2863 */ 2864 flush_in_svc = TRUE; 2865 } 2866 if ((mp->b_wptr - rptr) < sizeof (uint32_t) || 2867 !IS_P2ALIGNED(rptr, sizeof (uint32_t))) 2868 break; 2869 2870 switch (((union T_primitives *)rptr)->type) { 2871 case T_DATA_REQ: 2872 /* Don't pass T_DATA_REQ messages downstream. */ 2873 freemsg(mp); 2874 return; 2875 case T_ORDREL_REQ: 2876 RPCLOG(8, "mir_wput_other wq 0x%p: got T_ORDREL_REQ\n", 2877 (void *)q); 2878 mutex_enter(&mir->mir_mutex); 2879 if (mir->mir_type != RPC_SERVER) { 2880 /* 2881 * We are likely being called from 2882 * clnt_dispatch_notifyall(). Sending 2883 * a T_ORDREL_REQ will result in 2884 * a some kind of _IND message being sent, 2885 * will be another call to 2886 * clnt_dispatch_notifyall(). To keep the stack 2887 * lean, queue this message. 2888 */ 2889 mir->mir_inwservice = 1; 2890 (void) putq(q, mp); 2891 mutex_exit(&mir->mir_mutex); 2892 return; 2893 } 2894 2895 /* 2896 * Mark the structure such that we don't accept any 2897 * more requests from client. We could defer this 2898 * until we actually send the orderly release 2899 * request downstream, but all that does is delay 2900 * the closing of this stream. 2901 */ 2902 RPCLOG(16, "mir_wput_other wq 0x%p: got T_ORDREL_REQ " 2903 " so calling mir_svc_start_close\n", (void *)q); 2904 2905 mir_svc_start_close(q, mir); 2906 2907 /* 2908 * If we have sent down a T_ORDREL_REQ, don't send 2909 * any more. 2910 */ 2911 if (mir->mir_ordrel_pending) { 2912 freemsg(mp); 2913 mutex_exit(&mir->mir_mutex); 2914 return; 2915 } 2916 2917 /* 2918 * If the stream is not idle, then we hold the 2919 * orderly release until it becomes idle. This 2920 * ensures that KRPC will be able to reply to 2921 * all requests that we have passed to it. 2922 * 2923 * We also queue the request if there is data already 2924 * queued, because we cannot allow the T_ORDREL_REQ 2925 * to go before data. When we had a separate reply 2926 * count, this was not a problem, because the 2927 * reply count was reconciled when mir_wsrv() 2928 * completed. 2929 */ 2930 if (!MIR_SVC_QUIESCED(mir) || 2931 mir->mir_inwservice == 1) { 2932 mir->mir_inwservice = 1; 2933 (void) putq(q, mp); 2934 2935 RPCLOG(16, "mir_wput_other: queuing " 2936 "T_ORDREL_REQ on 0x%p\n", (void *)q); 2937 2938 mutex_exit(&mir->mir_mutex); 2939 return; 2940 } 2941 2942 /* 2943 * Mark the structure so that we know we sent 2944 * an orderly release request, and reset the idle timer. 2945 */ 2946 mir->mir_ordrel_pending = 1; 2947 2948 RPCLOG(16, "mir_wput_other: calling mir_svc_idle_start" 2949 " on 0x%p because we got T_ORDREL_REQ\n", 2950 (void *)q); 2951 2952 mir_svc_idle_start(q, mir); 2953 mutex_exit(&mir->mir_mutex); 2954 2955 /* 2956 * When we break, we will putnext the T_ORDREL_REQ. 2957 */ 2958 break; 2959 2960 case T_CONN_REQ: 2961 mutex_enter(&mir->mir_mutex); 2962 if (mir->mir_head_mp != NULL) { 2963 freemsg(mir->mir_head_mp); 2964 mir->mir_head_mp = NULL; 2965 mir->mir_tail_mp = NULL; 2966 } 2967 mir->mir_frag_len = -(int32_t)sizeof (uint32_t); 2968 /* 2969 * Restart timer in case mir_clnt_idle_do_stop() was 2970 * called. 2971 */ 2972 mir->mir_idle_timeout = clnt_idle_timeout; 2973 mir_clnt_idle_stop(q, mir); 2974 mir_clnt_idle_start(q, mir); 2975 mutex_exit(&mir->mir_mutex); 2976 break; 2977 2978 default: 2979 /* 2980 * T_DISCON_REQ is one of the interesting default 2981 * cases here. Ideally, an M_FLUSH is done before 2982 * T_DISCON_REQ is done. However, that is somewhat 2983 * cumbersome for clnt_cots.c to do. So we queue 2984 * T_DISCON_REQ, and let the service procedure 2985 * flush all M_DATA. 2986 */ 2987 break; 2988 } 2989 /* fallthru */; 2990 default: 2991 if (mp->b_datap->db_type >= QPCTL) { 2992 if (mp->b_datap->db_type == M_FLUSH) { 2993 if (mir->mir_type == RPC_CLIENT && 2994 *mp->b_rptr & FLUSHW) { 2995 RPCLOG(32, "mir_wput_other: flushing " 2996 "wq 0x%p\n", (void *)q); 2997 if (*mp->b_rptr & FLUSHBAND) { 2998 flushband(q, *(mp->b_rptr + 1), 2999 FLUSHDATA); 3000 } else { 3001 flushq(q, FLUSHDATA); 3002 } 3003 } else { 3004 RPCLOG(32, "mir_wput_other: ignoring " 3005 "M_FLUSH on wq 0x%p\n", (void *)q); 3006 } 3007 } 3008 break; 3009 } 3010 3011 mutex_enter(&mir->mir_mutex); 3012 if (mir->mir_inwservice == 0 && MIR_WCANPUTNEXT(mir, q)) { 3013 mutex_exit(&mir->mir_mutex); 3014 break; 3015 } 3016 mir->mir_inwservice = 1; 3017 mir->mir_inwflushdata = flush_in_svc; 3018 (void) putq(q, mp); 3019 mutex_exit(&mir->mir_mutex); 3020 qenable(q); 3021 3022 return; 3023 } 3024 putnext(q, mp); 3025 } 3026 3027 static void 3028 mir_wsrv(queue_t *q) 3029 { 3030 mblk_t *mp; 3031 mir_t *mir; 3032 bool_t flushdata; 3033 3034 mir = (mir_t *)q->q_ptr; 3035 mutex_enter(&mir->mir_mutex); 3036 3037 flushdata = mir->mir_inwflushdata; 3038 mir->mir_inwflushdata = 0; 3039 3040 while (mp = getq(q)) { 3041 if (mp->b_datap->db_type == M_DATA) { 3042 /* 3043 * Do not send any more data if we have sent 3044 * a T_ORDREL_REQ. 3045 */ 3046 if (flushdata || mir->mir_ordrel_pending == 1) { 3047 freemsg(mp); 3048 continue; 3049 } 3050 3051 /* 3052 * Make sure that the stream can really handle more 3053 * data. 3054 */ 3055 if (!MIR_WCANPUTNEXT(mir, q)) { 3056 (void) putbq(q, mp); 3057 mutex_exit(&mir->mir_mutex); 3058 return; 3059 } 3060 3061 /* 3062 * Now we pass the RPC message downstream. 3063 */ 3064 mutex_exit(&mir->mir_mutex); 3065 putnext(q, mp); 3066 mutex_enter(&mir->mir_mutex); 3067 continue; 3068 } 3069 3070 /* 3071 * This is not an RPC message, pass it downstream 3072 * (ignoring flow control) if the server side is not sending a 3073 * T_ORDREL_REQ downstream. 3074 */ 3075 if (mir->mir_type != RPC_SERVER || 3076 ((union T_primitives *)mp->b_rptr)->type != 3077 T_ORDREL_REQ) { 3078 mutex_exit(&mir->mir_mutex); 3079 putnext(q, mp); 3080 mutex_enter(&mir->mir_mutex); 3081 continue; 3082 } 3083 3084 if (mir->mir_ordrel_pending == 1) { 3085 /* 3086 * Don't send two T_ORDRELs 3087 */ 3088 freemsg(mp); 3089 continue; 3090 } 3091 3092 /* 3093 * Mark the structure so that we know we sent an orderly 3094 * release request. We will check to see slot is idle at the 3095 * end of this routine, and if so, reset the idle timer to 3096 * handle orderly release timeouts. 3097 */ 3098 mir->mir_ordrel_pending = 1; 3099 RPCLOG(16, "mir_wsrv: sending ordrel req on q 0x%p\n", 3100 (void *)q); 3101 /* 3102 * Send the orderly release downstream. If there are other 3103 * pending replies we won't be able to send them. However, 3104 * the only reason we should send the orderly release is if 3105 * we were idle, or if an unusual event occurred. 3106 */ 3107 mutex_exit(&mir->mir_mutex); 3108 putnext(q, mp); 3109 mutex_enter(&mir->mir_mutex); 3110 } 3111 3112 if (q->q_first == NULL) 3113 /* 3114 * If we call mir_svc_idle_start() below, then 3115 * clearing mir_inwservice here will also result in 3116 * any thread waiting in mir_close() to be signaled. 3117 */ 3118 mir->mir_inwservice = 0; 3119 3120 if (mir->mir_type != RPC_SERVER) { 3121 mutex_exit(&mir->mir_mutex); 3122 return; 3123 } 3124 3125 /* 3126 * If idle we call mir_svc_idle_start to start the timer (or wakeup 3127 * a close). Also make sure not to start the idle timer on the 3128 * listener stream. This can cause nfsd to send an orderly release 3129 * command on the listener stream. 3130 */ 3131 if (MIR_SVC_QUIESCED(mir) && !(mir->mir_listen_stream)) { 3132 RPCLOG(16, "mir_wsrv: calling mir_svc_idle_start on 0x%p " 3133 "because mir slot is idle\n", (void *)q); 3134 mir_svc_idle_start(q, mir); 3135 } 3136 3137 /* 3138 * If outbound flow control has been relieved, then allow new 3139 * inbound requests to be processed. 3140 */ 3141 if (mir->mir_hold_inbound) { 3142 mir->mir_hold_inbound = 0; 3143 qenable(RD(q)); 3144 } 3145 mutex_exit(&mir->mir_mutex); 3146 } 3147 3148 static void 3149 mir_disconnect(queue_t *q, mir_t *mir) 3150 { 3151 ASSERT(MUTEX_HELD(&mir->mir_mutex)); 3152 3153 switch (mir->mir_type) { 3154 case RPC_CLIENT: 3155 /* 3156 * We are disconnecting, but not necessarily 3157 * closing. By not closing, we will fail to 3158 * pick up a possibly changed global timeout value, 3159 * unless we store it now. 3160 */ 3161 mir->mir_idle_timeout = clnt_idle_timeout; 3162 mir_clnt_idle_start(WR(q), mir); 3163 mutex_exit(&mir->mir_mutex); 3164 3165 /* 3166 * T_DISCON_REQ is passed to KRPC as an integer value 3167 * (this is not a TPI message). It is used as a 3168 * convenient value to indicate a sanity check 3169 * failure -- the same KRPC routine is also called 3170 * for T_DISCON_INDs and T_ORDREL_INDs. 3171 */ 3172 clnt_dispatch_notifyall(WR(q), T_DISCON_REQ, 0); 3173 break; 3174 3175 case RPC_SERVER: 3176 mir->mir_svc_no_more_msgs = 1; 3177 mir_svc_idle_stop(WR(q), mir); 3178 mutex_exit(&mir->mir_mutex); 3179 RPCLOG(16, "mir_disconnect: telling " 3180 "stream head listener to disconnect stream " 3181 "(0x%p)\n", (void *) q); 3182 (void) mir_svc_policy_notify(q, 2); 3183 break; 3184 3185 default: 3186 mutex_exit(&mir->mir_mutex); 3187 break; 3188 } 3189 } 3190 3191 /* 3192 * Sanity check the message length, and if it's too large, shutdown the 3193 * connection. Returns 1 if the connection is shutdown; 0 otherwise. 3194 */ 3195 static int 3196 mir_check_len(queue_t *q, int32_t frag_len, mblk_t *head_mp) 3197 { 3198 mir_t *mir = q->q_ptr; 3199 uint_t maxsize = 0; 3200 3201 if (mir->mir_max_msg_sizep != NULL) 3202 maxsize = *mir->mir_max_msg_sizep; 3203 3204 if (maxsize == 0 || frag_len <= (int)maxsize) 3205 return (0); 3206 3207 freemsg(head_mp); 3208 mir->mir_head_mp = NULL; 3209 mir->mir_tail_mp = NULL; 3210 mir->mir_frag_header = 0; 3211 mir->mir_frag_len = -(int32_t)sizeof (uint32_t); 3212 if (mir->mir_type != RPC_SERVER || mir->mir_setup_complete) { 3213 cmn_err(CE_NOTE, 3214 "KRPC: record fragment from %s of size(%d) exceeds " 3215 "maximum (%u). Disconnecting", 3216 (mir->mir_type == RPC_CLIENT) ? "server" : 3217 (mir->mir_type == RPC_SERVER) ? "client" : 3218 "test tool", frag_len, maxsize); 3219 } 3220 3221 mir_disconnect(q, mir); 3222 return (1); 3223 } --- EOF ---