clnt_dg.c revision 213756
1/*	$NetBSD: clnt_dg.c,v 1.4 2000/07/14 08:40:41 fvdl Exp $	*/
2
3/*
4 * Sun RPC is a product of Sun Microsystems, Inc. and is provided for
5 * unrestricted use provided that this legend is included on all tape
6 * media and as a part of the software program in whole or part.  Users
7 * may copy or modify Sun RPC without charge, but are not authorized
8 * to license or distribute it to anyone else except as part of a product or
9 * program developed by the user.
10 *
11 * SUN RPC IS PROVIDED AS IS WITH NO WARRANTIES OF ANY KIND INCLUDING THE
12 * WARRANTIES OF DESIGN, MERCHANTIBILITY AND FITNESS FOR A PARTICULAR
13 * PURPOSE, OR ARISING FROM A COURSE OF DEALING, USAGE OR TRADE PRACTICE.
14 *
15 * Sun RPC is provided with no support and without any obligation on the
16 * part of Sun Microsystems, Inc. to assist in its use, correction,
17 * modification or enhancement.
18 *
19 * SUN MICROSYSTEMS, INC. SHALL HAVE NO LIABILITY WITH RESPECT TO THE
20 * INFRINGEMENT OF COPYRIGHTS, TRADE SECRETS OR ANY PATENTS BY SUN RPC
21 * OR ANY PART THEREOF.
22 *
23 * In no event will Sun Microsystems, Inc. be liable for any lost revenue
24 * or profits or other special, indirect and consequential damages, even if
25 * Sun has been advised of the possibility of such damages.
26 *
27 * Sun Microsystems, Inc.
28 * 2550 Garcia Avenue
29 * Mountain View, California  94043
30 */
31/*
32 * Copyright (c) 1986-1991 by Sun Microsystems Inc.
33 */
34
35#if defined(LIBC_SCCS) && !defined(lint)
36#ident	"@(#)clnt_dg.c	1.23	94/04/22 SMI"
37static char sccsid[] = "@(#)clnt_dg.c 1.19 89/03/16 Copyr 1988 Sun Micro";
38#endif
39#include <sys/cdefs.h>
40__FBSDID("$FreeBSD: head/sys/rpc/clnt_dg.c 213756 2010-10-13 00:57:14Z rmacklem $");
41
42/*
43 * Implements a connectionless client side RPC.
44 */
45
46#include <sys/param.h>
47#include <sys/systm.h>
48#include <sys/kernel.h>
49#include <sys/lock.h>
50#include <sys/malloc.h>
51#include <sys/mbuf.h>
52#include <sys/mutex.h>
53#include <sys/pcpu.h>
54#include <sys/proc.h>
55#include <sys/socket.h>
56#include <sys/socketvar.h>
57#include <sys/time.h>
58#include <sys/uio.h>
59
60#include <net/vnet.h>
61
62#include <rpc/rpc.h>
63#include <rpc/rpc_com.h>
64
65
66#ifdef _FREEFALL_CONFIG
67/*
68 * Disable RPC exponential back-off for FreeBSD.org systems.
69 */
70#define	RPC_MAX_BACKOFF		1 /* second */
71#else
72#define	RPC_MAX_BACKOFF		30 /* seconds */
73#endif
74
75static bool_t time_not_ok(struct timeval *);
76static enum clnt_stat clnt_dg_call(CLIENT *, struct rpc_callextra *,
77    rpcproc_t, struct mbuf *, struct mbuf **, struct timeval);
78static void clnt_dg_geterr(CLIENT *, struct rpc_err *);
79static bool_t clnt_dg_freeres(CLIENT *, xdrproc_t, void *);
80static void clnt_dg_abort(CLIENT *);
81static bool_t clnt_dg_control(CLIENT *, u_int, void *);
82static void clnt_dg_close(CLIENT *);
83static void clnt_dg_destroy(CLIENT *);
84static int clnt_dg_soupcall(struct socket *so, void *arg, int waitflag);
85
86static struct clnt_ops clnt_dg_ops = {
87	.cl_call =	clnt_dg_call,
88	.cl_abort =	clnt_dg_abort,
89	.cl_geterr =	clnt_dg_geterr,
90	.cl_freeres =	clnt_dg_freeres,
91	.cl_close =	clnt_dg_close,
92	.cl_destroy =	clnt_dg_destroy,
93	.cl_control =	clnt_dg_control
94};
95
96static const char mem_err_clnt_dg[] = "clnt_dg_create: out of memory";
97
98/*
99 * A pending RPC request which awaits a reply. Requests which have
100 * received their reply will have cr_xid set to zero and cr_mrep to
101 * the mbuf chain of the reply.
102 */
103struct cu_request {
104	TAILQ_ENTRY(cu_request) cr_link;
105	CLIENT			*cr_client;	/* owner */
106	uint32_t		cr_xid;		/* XID of request */
107	struct mbuf		*cr_mrep;	/* reply received by upcall */
108	int			cr_error;	/* any error from upcall */
109	char			cr_verf[MAX_AUTH_BYTES]; /* reply verf */
110};
111
112TAILQ_HEAD(cu_request_list, cu_request);
113
114#define MCALL_MSG_SIZE 24
115
116/*
117 * This structure is pointed to by the socket buffer's sb_upcallarg
118 * member. It is separate from the client private data to facilitate
119 * multiple clients sharing the same socket. The cs_lock mutex is used
120 * to protect all fields of this structure, the socket's receive
121 * buffer SOCKBUF_LOCK is used to ensure that exactly one of these
122 * structures is installed on the socket.
123 */
124struct cu_socket {
125	struct mtx		cs_lock;
126	int			cs_refs;	/* Count of clients */
127	struct cu_request_list	cs_pending;	/* Requests awaiting replies */
128	int			cs_upcallrefs;	/* Refcnt of upcalls in prog.*/
129};
130
131static void clnt_dg_upcallsdone(struct socket *, struct cu_socket *);
132
133/*
134 * Private data kept per client handle
135 */
136struct cu_data {
137	int			cu_threads;	/* # threads in clnt_vc_call */
138	bool_t			cu_closing;	/* TRUE if we are closing */
139	bool_t			cu_closed;	/* TRUE if we are closed */
140	struct socket		*cu_socket;	/* connection socket */
141	bool_t			cu_closeit;	/* opened by library */
142	struct sockaddr_storage	cu_raddr;	/* remote address */
143	int			cu_rlen;
144	struct timeval		cu_wait;	/* retransmit interval */
145	struct timeval		cu_total;	/* total time for the call */
146	struct rpc_err		cu_error;
147	uint32_t		cu_xid;
148	char			cu_mcallc[MCALL_MSG_SIZE]; /* marshalled callmsg */
149	size_t			cu_mcalllen;
150	size_t			cu_sendsz;	/* send size */
151	size_t			cu_recvsz;	/* recv size */
152	int			cu_async;
153	int			cu_connect;	/* Use connect(). */
154	int			cu_connected;	/* Have done connect(). */
155	const char		*cu_waitchan;
156	int			cu_waitflag;
157	int			cu_cwnd;	/* congestion window */
158	int			cu_sent;	/* number of in-flight RPCs */
159	bool_t			cu_cwnd_wait;
160};
161
162#define CWNDSCALE	256
163#define MAXCWND		(32 * CWNDSCALE)
164
165/*
166 * Connection less client creation returns with client handle parameters.
167 * Default options are set, which the user can change using clnt_control().
168 * fd should be open and bound.
169 * NB: The rpch->cl_auth is initialized to null authentication.
170 * 	Caller may wish to set this something more useful.
171 *
172 * sendsz and recvsz are the maximum allowable packet sizes that can be
173 * sent and received. Normally they are the same, but they can be
174 * changed to improve the program efficiency and buffer allocation.
175 * If they are 0, use the transport default.
176 *
177 * If svcaddr is NULL, returns NULL.
178 */
179CLIENT *
180clnt_dg_create(
181	struct socket *so,
182	struct sockaddr *svcaddr,	/* servers address */
183	rpcprog_t program,		/* program number */
184	rpcvers_t version,		/* version number */
185	size_t sendsz,			/* buffer recv size */
186	size_t recvsz)			/* buffer send size */
187{
188	CLIENT *cl = NULL;		/* client handle */
189	struct cu_data *cu = NULL;	/* private data */
190	struct cu_socket *cs = NULL;
191	struct sockbuf *sb;
192	struct timeval now;
193	struct rpc_msg call_msg;
194	struct __rpc_sockinfo si;
195	XDR xdrs;
196	int error;
197
198	if (svcaddr == NULL) {
199		rpc_createerr.cf_stat = RPC_UNKNOWNADDR;
200		return (NULL);
201	}
202
203	CURVNET_SET(so->so_vnet);
204	if (!__rpc_socket2sockinfo(so, &si)) {
205		rpc_createerr.cf_stat = RPC_TLIERROR;
206		rpc_createerr.cf_error.re_errno = 0;
207		CURVNET_RESTORE();
208		return (NULL);
209	}
210	CURVNET_RESTORE();
211
212	/*
213	 * Find the receive and the send size
214	 */
215	sendsz = __rpc_get_t_size(si.si_af, si.si_proto, (int)sendsz);
216	recvsz = __rpc_get_t_size(si.si_af, si.si_proto, (int)recvsz);
217	if ((sendsz == 0) || (recvsz == 0)) {
218		rpc_createerr.cf_stat = RPC_TLIERROR; /* XXX */
219		rpc_createerr.cf_error.re_errno = 0;
220		return (NULL);
221	}
222
223	cl = mem_alloc(sizeof (CLIENT));
224
225	/*
226	 * Should be multiple of 4 for XDR.
227	 */
228	sendsz = ((sendsz + 3) / 4) * 4;
229	recvsz = ((recvsz + 3) / 4) * 4;
230	cu = mem_alloc(sizeof (*cu));
231	cu->cu_threads = 0;
232	cu->cu_closing = FALSE;
233	cu->cu_closed = FALSE;
234	(void) memcpy(&cu->cu_raddr, svcaddr, (size_t)svcaddr->sa_len);
235	cu->cu_rlen = svcaddr->sa_len;
236	/* Other values can also be set through clnt_control() */
237	cu->cu_wait.tv_sec = 3;	/* heuristically chosen */
238	cu->cu_wait.tv_usec = 0;
239	cu->cu_total.tv_sec = -1;
240	cu->cu_total.tv_usec = -1;
241	cu->cu_sendsz = sendsz;
242	cu->cu_recvsz = recvsz;
243	cu->cu_async = FALSE;
244	cu->cu_connect = FALSE;
245	cu->cu_connected = FALSE;
246	cu->cu_waitchan = "rpcrecv";
247	cu->cu_waitflag = 0;
248	cu->cu_cwnd = MAXCWND / 2;
249	cu->cu_sent = 0;
250	cu->cu_cwnd_wait = FALSE;
251	(void) getmicrotime(&now);
252	cu->cu_xid = __RPC_GETXID(&now);
253	call_msg.rm_xid = cu->cu_xid;
254	call_msg.rm_call.cb_prog = program;
255	call_msg.rm_call.cb_vers = version;
256	xdrmem_create(&xdrs, cu->cu_mcallc, MCALL_MSG_SIZE, XDR_ENCODE);
257	if (! xdr_callhdr(&xdrs, &call_msg)) {
258		rpc_createerr.cf_stat = RPC_CANTENCODEARGS;  /* XXX */
259		rpc_createerr.cf_error.re_errno = 0;
260		goto err2;
261	}
262	cu->cu_mcalllen = XDR_GETPOS(&xdrs);
263
264	/*
265	 * By default, closeit is always FALSE. It is users responsibility
266	 * to do a close on it, else the user may use clnt_control
267	 * to let clnt_destroy do it for him/her.
268	 */
269	cu->cu_closeit = FALSE;
270	cu->cu_socket = so;
271	error = soreserve(so, (u_long)sendsz, (u_long)recvsz);
272	if (error != 0) {
273		rpc_createerr.cf_stat = RPC_FAILED;
274		rpc_createerr.cf_error.re_errno = error;
275		goto err2;
276	}
277
278	sb = &so->so_rcv;
279	SOCKBUF_LOCK(&so->so_rcv);
280recheck_socket:
281	if (sb->sb_upcall) {
282		if (sb->sb_upcall != clnt_dg_soupcall) {
283			SOCKBUF_UNLOCK(&so->so_rcv);
284			printf("clnt_dg_create(): socket already has an incompatible upcall\n");
285			goto err2;
286		}
287		cs = (struct cu_socket *) sb->sb_upcallarg;
288		mtx_lock(&cs->cs_lock);
289		cs->cs_refs++;
290		mtx_unlock(&cs->cs_lock);
291	} else {
292		/*
293		 * We are the first on this socket - allocate the
294		 * structure and install it in the socket.
295		 */
296		SOCKBUF_UNLOCK(&so->so_rcv);
297		cs = mem_alloc(sizeof(*cs));
298		SOCKBUF_LOCK(&so->so_rcv);
299		if (sb->sb_upcall) {
300			/*
301			 * We have lost a race with some other client.
302			 */
303			mem_free(cs, sizeof(*cs));
304			goto recheck_socket;
305		}
306		mtx_init(&cs->cs_lock, "cs->cs_lock", NULL, MTX_DEF);
307		cs->cs_refs = 1;
308		cs->cs_upcallrefs = 0;
309		TAILQ_INIT(&cs->cs_pending);
310		soupcall_set(so, SO_RCV, clnt_dg_soupcall, cs);
311	}
312	SOCKBUF_UNLOCK(&so->so_rcv);
313
314	cl->cl_refs = 1;
315	cl->cl_ops = &clnt_dg_ops;
316	cl->cl_private = (caddr_t)(void *)cu;
317	cl->cl_auth = authnone_create();
318	cl->cl_tp = NULL;
319	cl->cl_netid = NULL;
320	return (cl);
321err2:
322	if (cl) {
323		mem_free(cl, sizeof (CLIENT));
324		if (cu)
325			mem_free(cu, sizeof (*cu));
326	}
327	return (NULL);
328}
329
330static enum clnt_stat
331clnt_dg_call(
332	CLIENT		*cl,		/* client handle */
333	struct rpc_callextra *ext,	/* call metadata */
334	rpcproc_t	proc,		/* procedure number */
335	struct mbuf	*args,		/* pointer to args */
336	struct mbuf	**resultsp,	/* pointer to results */
337	struct timeval	utimeout)	/* seconds to wait before giving up */
338{
339	struct cu_data *cu = (struct cu_data *)cl->cl_private;
340	struct cu_socket *cs;
341	struct rpc_timers *rt;
342	AUTH *auth;
343	struct rpc_err *errp;
344	enum clnt_stat stat;
345	XDR xdrs;
346	struct rpc_msg reply_msg;
347	bool_t ok;
348	int retrans;			/* number of re-transmits so far */
349	int nrefreshes = 2;		/* number of times to refresh cred */
350	struct timeval *tvp;
351	int timeout;
352	int retransmit_time;
353	int next_sendtime, starttime, rtt, time_waited, tv = 0;
354	struct sockaddr *sa;
355	socklen_t salen;
356	uint32_t xid = 0;
357	struct mbuf *mreq = NULL, *results;
358	struct cu_request *cr;
359	int error;
360
361	cs = cu->cu_socket->so_rcv.sb_upcallarg;
362	cr = malloc(sizeof(struct cu_request), M_RPC, M_WAITOK);
363
364	mtx_lock(&cs->cs_lock);
365
366	if (cu->cu_closing || cu->cu_closed) {
367		mtx_unlock(&cs->cs_lock);
368		free(cr, M_RPC);
369		return (RPC_CANTSEND);
370	}
371	cu->cu_threads++;
372
373	if (ext) {
374		auth = ext->rc_auth;
375		errp = &ext->rc_err;
376	} else {
377		auth = cl->cl_auth;
378		errp = &cu->cu_error;
379	}
380
381	cr->cr_client = cl;
382	cr->cr_mrep = NULL;
383	cr->cr_error = 0;
384
385	if (cu->cu_total.tv_usec == -1) {
386		tvp = &utimeout; /* use supplied timeout */
387	} else {
388		tvp = &cu->cu_total; /* use default timeout */
389	}
390	if (tvp->tv_sec || tvp->tv_usec)
391		timeout = tvtohz(tvp);
392	else
393		timeout = 0;
394
395	if (cu->cu_connect && !cu->cu_connected) {
396		mtx_unlock(&cs->cs_lock);
397		error = soconnect(cu->cu_socket,
398		    (struct sockaddr *)&cu->cu_raddr, curthread);
399		mtx_lock(&cs->cs_lock);
400		if (error) {
401			errp->re_errno = error;
402			errp->re_status = stat = RPC_CANTSEND;
403			goto out;
404		}
405		cu->cu_connected = 1;
406	}
407	if (cu->cu_connected) {
408		sa = NULL;
409		salen = 0;
410	} else {
411		sa = (struct sockaddr *)&cu->cu_raddr;
412		salen = cu->cu_rlen;
413	}
414	time_waited = 0;
415	retrans = 0;
416	if (ext && ext->rc_timers) {
417		rt = ext->rc_timers;
418		if (!rt->rt_rtxcur)
419			rt->rt_rtxcur = tvtohz(&cu->cu_wait);
420		retransmit_time = next_sendtime = rt->rt_rtxcur;
421	} else {
422		rt = NULL;
423		retransmit_time = next_sendtime = tvtohz(&cu->cu_wait);
424	}
425
426	starttime = ticks;
427
428call_again:
429	mtx_assert(&cs->cs_lock, MA_OWNED);
430
431	cu->cu_xid++;
432	xid = cu->cu_xid;
433
434send_again:
435	mtx_unlock(&cs->cs_lock);
436
437	MGETHDR(mreq, M_WAIT, MT_DATA);
438	KASSERT(cu->cu_mcalllen <= MHLEN, ("RPC header too big"));
439	bcopy(cu->cu_mcallc, mreq->m_data, cu->cu_mcalllen);
440	mreq->m_len = cu->cu_mcalllen;
441
442	/*
443	 * The XID is the first thing in the request.
444	 */
445	*mtod(mreq, uint32_t *) = htonl(xid);
446
447	xdrmbuf_create(&xdrs, mreq, XDR_ENCODE);
448
449	if (cu->cu_async == TRUE && args == NULL)
450		goto get_reply;
451
452	if ((! XDR_PUTINT32(&xdrs, &proc)) ||
453	    (! AUTH_MARSHALL(auth, xid, &xdrs,
454		m_copym(args, 0, M_COPYALL, M_WAITOK)))) {
455		errp->re_status = stat = RPC_CANTENCODEARGS;
456		mtx_lock(&cs->cs_lock);
457		goto out;
458	}
459	mreq->m_pkthdr.len = m_length(mreq, NULL);
460
461	cr->cr_xid = xid;
462	mtx_lock(&cs->cs_lock);
463
464	/*
465	 * Try to get a place in the congestion window.
466	 */
467	while (cu->cu_sent >= cu->cu_cwnd) {
468		cu->cu_cwnd_wait = TRUE;
469		error = msleep(&cu->cu_cwnd_wait, &cs->cs_lock,
470		    cu->cu_waitflag, "rpccwnd", 0);
471		if (error) {
472			errp->re_errno = error;
473			errp->re_status = stat = RPC_CANTSEND;
474			goto out;
475		}
476	}
477	cu->cu_sent += CWNDSCALE;
478
479	TAILQ_INSERT_TAIL(&cs->cs_pending, cr, cr_link);
480	mtx_unlock(&cs->cs_lock);
481
482	/*
483	 * sosend consumes mreq.
484	 */
485	error = sosend(cu->cu_socket, sa, NULL, mreq, NULL, 0, curthread);
486	mreq = NULL;
487
488	/*
489	 * sub-optimal code appears here because we have
490	 * some clock time to spare while the packets are in flight.
491	 * (We assume that this is actually only executed once.)
492	 */
493	reply_msg.acpted_rply.ar_verf.oa_flavor = AUTH_NULL;
494	reply_msg.acpted_rply.ar_verf.oa_base = cr->cr_verf;
495	reply_msg.acpted_rply.ar_verf.oa_length = 0;
496	reply_msg.acpted_rply.ar_results.where = NULL;
497	reply_msg.acpted_rply.ar_results.proc = (xdrproc_t)xdr_void;
498
499	mtx_lock(&cs->cs_lock);
500	if (error) {
501		TAILQ_REMOVE(&cs->cs_pending, cr, cr_link);
502		errp->re_errno = error;
503		errp->re_status = stat = RPC_CANTSEND;
504		cu->cu_sent -= CWNDSCALE;
505		if (cu->cu_cwnd_wait) {
506			cu->cu_cwnd_wait = FALSE;
507			wakeup(&cu->cu_cwnd_wait);
508		}
509		goto out;
510	}
511
512	/*
513	 * Check to see if we got an upcall while waiting for the
514	 * lock.
515	 */
516	if (cr->cr_error) {
517		TAILQ_REMOVE(&cs->cs_pending, cr, cr_link);
518		errp->re_errno = cr->cr_error;
519		errp->re_status = stat = RPC_CANTRECV;
520		cu->cu_sent -= CWNDSCALE;
521		if (cu->cu_cwnd_wait) {
522			cu->cu_cwnd_wait = FALSE;
523			wakeup(&cu->cu_cwnd_wait);
524		}
525		goto out;
526	}
527	if (cr->cr_mrep) {
528		TAILQ_REMOVE(&cs->cs_pending, cr, cr_link);
529		cu->cu_sent -= CWNDSCALE;
530		if (cu->cu_cwnd_wait) {
531			cu->cu_cwnd_wait = FALSE;
532			wakeup(&cu->cu_cwnd_wait);
533		}
534		goto got_reply;
535	}
536
537	/*
538	 * Hack to provide rpc-based message passing
539	 */
540	if (timeout == 0) {
541		TAILQ_REMOVE(&cs->cs_pending, cr, cr_link);
542		errp->re_status = stat = RPC_TIMEDOUT;
543		cu->cu_sent -= CWNDSCALE;
544		if (cu->cu_cwnd_wait) {
545			cu->cu_cwnd_wait = FALSE;
546			wakeup(&cu->cu_cwnd_wait);
547		}
548		goto out;
549	}
550
551get_reply:
552	for (;;) {
553		/* Decide how long to wait. */
554		if (next_sendtime < timeout)
555			tv = next_sendtime;
556		else
557			tv = timeout;
558		tv -= time_waited;
559
560		if (tv > 0) {
561			if (cu->cu_closing || cu->cu_closed) {
562				error = 0;
563				cr->cr_error = ESHUTDOWN;
564			} else {
565				error = msleep(cr, &cs->cs_lock,
566				    cu->cu_waitflag, cu->cu_waitchan, tv);
567			}
568		} else {
569			error = EWOULDBLOCK;
570		}
571
572		TAILQ_REMOVE(&cs->cs_pending, cr, cr_link);
573		cu->cu_sent -= CWNDSCALE;
574		if (cu->cu_cwnd_wait) {
575			cu->cu_cwnd_wait = FALSE;
576			wakeup(&cu->cu_cwnd_wait);
577		}
578
579		if (!error) {
580			/*
581			 * We were woken up by the upcall.  If the
582			 * upcall had a receive error, report that,
583			 * otherwise we have a reply.
584			 */
585			if (cr->cr_error) {
586				errp->re_errno = cr->cr_error;
587				errp->re_status = stat = RPC_CANTRECV;
588				goto out;
589			}
590
591			cu->cu_cwnd += (CWNDSCALE * CWNDSCALE
592			    + cu->cu_cwnd / 2) / cu->cu_cwnd;
593			if (cu->cu_cwnd > MAXCWND)
594				cu->cu_cwnd = MAXCWND;
595
596			if (rt) {
597				/*
598				 * Add one to the time since a tick
599				 * count of N means that the actual
600				 * time taken was somewhere between N
601				 * and N+1.
602				 */
603				rtt = ticks - starttime + 1;
604
605				/*
606				 * Update our estimate of the round
607				 * trip time using roughly the
608				 * algorithm described in RFC
609				 * 2988. Given an RTT sample R:
610				 *
611				 * RTTVAR = (1-beta) * RTTVAR + beta * |SRTT-R|
612				 * SRTT = (1-alpha) * SRTT + alpha * R
613				 *
614				 * where alpha = 0.125 and beta = 0.25.
615				 *
616				 * The initial retransmit timeout is
617				 * SRTT + 4*RTTVAR and doubles on each
618				 * retransmision.
619				 */
620				if (rt->rt_srtt == 0) {
621					rt->rt_srtt = rtt;
622					rt->rt_deviate = rtt / 2;
623				} else {
624					int32_t error = rtt - rt->rt_srtt;
625					rt->rt_srtt += error / 8;
626					error = abs(error) - rt->rt_deviate;
627					rt->rt_deviate += error / 4;
628				}
629				rt->rt_rtxcur = rt->rt_srtt + 4*rt->rt_deviate;
630			}
631
632			break;
633		}
634
635		/*
636		 * The sleep returned an error so our request is still
637		 * on the list. If we got EWOULDBLOCK, we may want to
638		 * re-send the request.
639		 */
640		if (error != EWOULDBLOCK) {
641			errp->re_errno = error;
642			if (error == EINTR)
643				errp->re_status = stat = RPC_INTR;
644			else
645				errp->re_status = stat = RPC_CANTRECV;
646			goto out;
647		}
648
649		time_waited = ticks - starttime;
650
651		/* Check for timeout. */
652		if (time_waited > timeout) {
653			errp->re_errno = EWOULDBLOCK;
654			errp->re_status = stat = RPC_TIMEDOUT;
655			goto out;
656		}
657
658		/* Retransmit if necessary. */
659		if (time_waited >= next_sendtime) {
660			cu->cu_cwnd /= 2;
661			if (cu->cu_cwnd < CWNDSCALE)
662				cu->cu_cwnd = CWNDSCALE;
663			if (ext && ext->rc_feedback) {
664				mtx_unlock(&cs->cs_lock);
665				if (retrans == 0)
666					ext->rc_feedback(FEEDBACK_REXMIT1,
667					    proc, ext->rc_feedback_arg);
668				else
669					ext->rc_feedback(FEEDBACK_REXMIT2,
670					    proc, ext->rc_feedback_arg);
671				mtx_lock(&cs->cs_lock);
672			}
673			if (cu->cu_closing || cu->cu_closed) {
674				errp->re_errno = ESHUTDOWN;
675				errp->re_status = stat = RPC_CANTRECV;
676				goto out;
677			}
678			retrans++;
679			/* update retransmit_time */
680			if (retransmit_time < RPC_MAX_BACKOFF * hz)
681				retransmit_time = 2 * retransmit_time;
682			next_sendtime += retransmit_time;
683			goto send_again;
684		}
685		TAILQ_INSERT_TAIL(&cs->cs_pending, cr, cr_link);
686	}
687
688got_reply:
689	/*
690	 * Now decode and validate the response. We need to drop the
691	 * lock since xdr_replymsg may end up sleeping in malloc.
692	 */
693	mtx_unlock(&cs->cs_lock);
694
695	if (ext && ext->rc_feedback)
696		ext->rc_feedback(FEEDBACK_OK, proc, ext->rc_feedback_arg);
697
698	xdrmbuf_create(&xdrs, cr->cr_mrep, XDR_DECODE);
699	ok = xdr_replymsg(&xdrs, &reply_msg);
700	cr->cr_mrep = NULL;
701
702	if (ok) {
703		if ((reply_msg.rm_reply.rp_stat == MSG_ACCEPTED) &&
704		    (reply_msg.acpted_rply.ar_stat == SUCCESS))
705			errp->re_status = stat = RPC_SUCCESS;
706		else
707			stat = _seterr_reply(&reply_msg, &(cu->cu_error));
708
709		if (errp->re_status == RPC_SUCCESS) {
710			results = xdrmbuf_getall(&xdrs);
711			if (! AUTH_VALIDATE(auth, xid,
712				&reply_msg.acpted_rply.ar_verf,
713				&results)) {
714				errp->re_status = stat = RPC_AUTHERROR;
715				errp->re_why = AUTH_INVALIDRESP;
716				if (retrans &&
717				    auth->ah_cred.oa_flavor == RPCSEC_GSS) {
718					/*
719					 * If we retransmitted, its
720					 * possible that we will
721					 * receive a reply for one of
722					 * the earlier transmissions
723					 * (which will use an older
724					 * RPCSEC_GSS sequence
725					 * number). In this case, just
726					 * go back and listen for a
727					 * new reply. We could keep a
728					 * record of all the seq
729					 * numbers we have transmitted
730					 * so far so that we could
731					 * accept a reply for any of
732					 * them here.
733					 */
734					XDR_DESTROY(&xdrs);
735					mtx_lock(&cs->cs_lock);
736					TAILQ_INSERT_TAIL(&cs->cs_pending,
737					    cr, cr_link);
738					cr->cr_mrep = NULL;
739					goto get_reply;
740				}
741			} else {
742				*resultsp = results;
743			}
744		}		/* end successful completion */
745		/*
746		 * If unsuccesful AND error is an authentication error
747		 * then refresh credentials and try again, else break
748		 */
749		else if (stat == RPC_AUTHERROR)
750			/* maybe our credentials need to be refreshed ... */
751			if (nrefreshes > 0 &&
752			    AUTH_REFRESH(auth, &reply_msg)) {
753				nrefreshes--;
754				XDR_DESTROY(&xdrs);
755				mtx_lock(&cs->cs_lock);
756				goto call_again;
757			}
758		/* end of unsuccessful completion */
759	}	/* end of valid reply message */
760	else {
761		errp->re_status = stat = RPC_CANTDECODERES;
762
763	}
764	XDR_DESTROY(&xdrs);
765	mtx_lock(&cs->cs_lock);
766out:
767	mtx_assert(&cs->cs_lock, MA_OWNED);
768
769	if (mreq)
770		m_freem(mreq);
771	if (cr->cr_mrep)
772		m_freem(cr->cr_mrep);
773
774	cu->cu_threads--;
775	if (cu->cu_closing)
776		wakeup(cu);
777
778	mtx_unlock(&cs->cs_lock);
779
780	if (auth && stat != RPC_SUCCESS)
781		AUTH_VALIDATE(auth, xid, NULL, NULL);
782
783	free(cr, M_RPC);
784
785	return (stat);
786}
787
788static void
789clnt_dg_geterr(CLIENT *cl, struct rpc_err *errp)
790{
791	struct cu_data *cu = (struct cu_data *)cl->cl_private;
792
793	*errp = cu->cu_error;
794}
795
796static bool_t
797clnt_dg_freeres(CLIENT *cl, xdrproc_t xdr_res, void *res_ptr)
798{
799	XDR xdrs;
800	bool_t dummy;
801
802	xdrs.x_op = XDR_FREE;
803	dummy = (*xdr_res)(&xdrs, res_ptr);
804
805	return (dummy);
806}
807
808/*ARGSUSED*/
809static void
810clnt_dg_abort(CLIENT *h)
811{
812}
813
814static bool_t
815clnt_dg_control(CLIENT *cl, u_int request, void *info)
816{
817	struct cu_data *cu = (struct cu_data *)cl->cl_private;
818	struct cu_socket *cs;
819	struct sockaddr *addr;
820
821	cs = cu->cu_socket->so_rcv.sb_upcallarg;
822	mtx_lock(&cs->cs_lock);
823
824	switch (request) {
825	case CLSET_FD_CLOSE:
826		cu->cu_closeit = TRUE;
827		mtx_unlock(&cs->cs_lock);
828		return (TRUE);
829	case CLSET_FD_NCLOSE:
830		cu->cu_closeit = FALSE;
831		mtx_unlock(&cs->cs_lock);
832		return (TRUE);
833	}
834
835	/* for other requests which use info */
836	if (info == NULL) {
837		mtx_unlock(&cs->cs_lock);
838		return (FALSE);
839	}
840	switch (request) {
841	case CLSET_TIMEOUT:
842		if (time_not_ok((struct timeval *)info)) {
843			mtx_unlock(&cs->cs_lock);
844			return (FALSE);
845		}
846		cu->cu_total = *(struct timeval *)info;
847		break;
848	case CLGET_TIMEOUT:
849		*(struct timeval *)info = cu->cu_total;
850		break;
851	case CLSET_RETRY_TIMEOUT:
852		if (time_not_ok((struct timeval *)info)) {
853			mtx_unlock(&cs->cs_lock);
854			return (FALSE);
855		}
856		cu->cu_wait = *(struct timeval *)info;
857		break;
858	case CLGET_RETRY_TIMEOUT:
859		*(struct timeval *)info = cu->cu_wait;
860		break;
861	case CLGET_SVC_ADDR:
862		/*
863		 * Slightly different semantics to userland - we use
864		 * sockaddr instead of netbuf.
865		 */
866		memcpy(info, &cu->cu_raddr, cu->cu_raddr.ss_len);
867		break;
868	case CLSET_SVC_ADDR:		/* set to new address */
869		addr = (struct sockaddr *)info;
870		(void) memcpy(&cu->cu_raddr, addr, addr->sa_len);
871		break;
872	case CLGET_XID:
873		*(uint32_t *)info = cu->cu_xid;
874		break;
875
876	case CLSET_XID:
877		/* This will set the xid of the NEXT call */
878		/* decrement by 1 as clnt_dg_call() increments once */
879		cu->cu_xid = *(uint32_t *)info - 1;
880		break;
881
882	case CLGET_VERS:
883		/*
884		 * This RELIES on the information that, in the call body,
885		 * the version number field is the fifth field from the
886		 * begining of the RPC header. MUST be changed if the
887		 * call_struct is changed
888		 */
889		*(uint32_t *)info =
890		    ntohl(*(uint32_t *)(void *)(cu->cu_mcallc +
891		    4 * BYTES_PER_XDR_UNIT));
892		break;
893
894	case CLSET_VERS:
895		*(uint32_t *)(void *)(cu->cu_mcallc + 4 * BYTES_PER_XDR_UNIT)
896			= htonl(*(uint32_t *)info);
897		break;
898
899	case CLGET_PROG:
900		/*
901		 * This RELIES on the information that, in the call body,
902		 * the program number field is the fourth field from the
903		 * begining of the RPC header. MUST be changed if the
904		 * call_struct is changed
905		 */
906		*(uint32_t *)info =
907		    ntohl(*(uint32_t *)(void *)(cu->cu_mcallc +
908		    3 * BYTES_PER_XDR_UNIT));
909		break;
910
911	case CLSET_PROG:
912		*(uint32_t *)(void *)(cu->cu_mcallc + 3 * BYTES_PER_XDR_UNIT)
913			= htonl(*(uint32_t *)info);
914		break;
915	case CLSET_ASYNC:
916		cu->cu_async = *(int *)info;
917		break;
918	case CLSET_CONNECT:
919		cu->cu_connect = *(int *)info;
920		break;
921	case CLSET_WAITCHAN:
922		cu->cu_waitchan = (const char *)info;
923		break;
924	case CLGET_WAITCHAN:
925		*(const char **) info = cu->cu_waitchan;
926		break;
927	case CLSET_INTERRUPTIBLE:
928		if (*(int *) info)
929			cu->cu_waitflag = PCATCH;
930		else
931			cu->cu_waitflag = 0;
932		break;
933	case CLGET_INTERRUPTIBLE:
934		if (cu->cu_waitflag)
935			*(int *) info = TRUE;
936		else
937			*(int *) info = FALSE;
938		break;
939	default:
940		mtx_unlock(&cs->cs_lock);
941		return (FALSE);
942	}
943	mtx_unlock(&cs->cs_lock);
944	return (TRUE);
945}
946
947static void
948clnt_dg_close(CLIENT *cl)
949{
950	struct cu_data *cu = (struct cu_data *)cl->cl_private;
951	struct cu_socket *cs;
952	struct cu_request *cr;
953
954	cs = cu->cu_socket->so_rcv.sb_upcallarg;
955	mtx_lock(&cs->cs_lock);
956
957	if (cu->cu_closed) {
958		mtx_unlock(&cs->cs_lock);
959		return;
960	}
961
962	if (cu->cu_closing) {
963		while (cu->cu_closing)
964			msleep(cu, &cs->cs_lock, 0, "rpcclose", 0);
965		KASSERT(cu->cu_closed, ("client should be closed"));
966		mtx_unlock(&cs->cs_lock);
967		return;
968	}
969
970	/*
971	 * Abort any pending requests and wait until everyone
972	 * has finished with clnt_vc_call.
973	 */
974	cu->cu_closing = TRUE;
975	TAILQ_FOREACH(cr, &cs->cs_pending, cr_link) {
976		if (cr->cr_client == cl) {
977			cr->cr_xid = 0;
978			cr->cr_error = ESHUTDOWN;
979			wakeup(cr);
980		}
981	}
982
983	while (cu->cu_threads)
984		msleep(cu, &cs->cs_lock, 0, "rpcclose", 0);
985
986	cu->cu_closing = FALSE;
987	cu->cu_closed = TRUE;
988
989	mtx_unlock(&cs->cs_lock);
990	wakeup(cu);
991}
992
993static void
994clnt_dg_destroy(CLIENT *cl)
995{
996	struct cu_data *cu = (struct cu_data *)cl->cl_private;
997	struct cu_socket *cs;
998	struct socket *so = NULL;
999	bool_t lastsocketref;
1000
1001	cs = cu->cu_socket->so_rcv.sb_upcallarg;
1002	clnt_dg_close(cl);
1003
1004	mtx_lock(&cs->cs_lock);
1005
1006	cs->cs_refs--;
1007	if (cs->cs_refs == 0) {
1008		mtx_unlock(&cs->cs_lock);
1009		SOCKBUF_LOCK(&cu->cu_socket->so_rcv);
1010		soupcall_clear(cu->cu_socket, SO_RCV);
1011		clnt_dg_upcallsdone(cu->cu_socket, cs);
1012		SOCKBUF_UNLOCK(&cu->cu_socket->so_rcv);
1013		mtx_destroy(&cs->cs_lock);
1014		mem_free(cs, sizeof(*cs));
1015		lastsocketref = TRUE;
1016	} else {
1017		mtx_unlock(&cs->cs_lock);
1018		lastsocketref = FALSE;
1019	}
1020
1021	if (cu->cu_closeit && lastsocketref) {
1022		so = cu->cu_socket;
1023		cu->cu_socket = NULL;
1024	}
1025
1026	if (so)
1027		soclose(so);
1028
1029	if (cl->cl_netid && cl->cl_netid[0])
1030		mem_free(cl->cl_netid, strlen(cl->cl_netid) +1);
1031	if (cl->cl_tp && cl->cl_tp[0])
1032		mem_free(cl->cl_tp, strlen(cl->cl_tp) +1);
1033	mem_free(cu, sizeof (*cu));
1034	mem_free(cl, sizeof (CLIENT));
1035}
1036
1037/*
1038 * Make sure that the time is not garbage.  -1 value is allowed.
1039 */
1040static bool_t
1041time_not_ok(struct timeval *t)
1042{
1043	return (t->tv_sec < -1 || t->tv_sec > 100000000 ||
1044		t->tv_usec < -1 || t->tv_usec > 1000000);
1045}
1046
1047int
1048clnt_dg_soupcall(struct socket *so, void *arg, int waitflag)
1049{
1050	struct cu_socket *cs = (struct cu_socket *) arg;
1051	struct uio uio;
1052	struct mbuf *m;
1053	struct mbuf *control;
1054	struct cu_request *cr;
1055	int error, rcvflag, foundreq;
1056	uint32_t xid;
1057
1058	cs->cs_upcallrefs++;
1059	uio.uio_resid = 1000000000;
1060	uio.uio_td = curthread;
1061	do {
1062		SOCKBUF_UNLOCK(&so->so_rcv);
1063		m = NULL;
1064		control = NULL;
1065		rcvflag = MSG_DONTWAIT;
1066		error = soreceive(so, NULL, &uio, &m, &control, &rcvflag);
1067		if (control)
1068			m_freem(control);
1069		SOCKBUF_LOCK(&so->so_rcv);
1070
1071		if (error == EWOULDBLOCK)
1072			break;
1073
1074		/*
1075		 * If there was an error, wake up all pending
1076		 * requests.
1077		 */
1078		if (error) {
1079			mtx_lock(&cs->cs_lock);
1080			TAILQ_FOREACH(cr, &cs->cs_pending, cr_link) {
1081				cr->cr_xid = 0;
1082				cr->cr_error = error;
1083				wakeup(cr);
1084			}
1085			mtx_unlock(&cs->cs_lock);
1086			break;
1087		}
1088
1089		/*
1090		 * The XID is in the first uint32_t of the reply.
1091		 */
1092		if (m->m_len < sizeof(xid))
1093			m = m_pullup(m, sizeof(xid));
1094		if (!m)
1095			/*
1096			 * Should never happen.
1097			 */
1098			continue;
1099
1100		xid = ntohl(*mtod(m, uint32_t *));
1101
1102		/*
1103		 * Attempt to match this reply with a pending request.
1104		 */
1105		mtx_lock(&cs->cs_lock);
1106		foundreq = 0;
1107		TAILQ_FOREACH(cr, &cs->cs_pending, cr_link) {
1108			if (cr->cr_xid == xid) {
1109				/*
1110				 * This one matches. We leave the
1111				 * reply mbuf in cr->cr_mrep. Set the
1112				 * XID to zero so that we will ignore
1113				 * any duplicated replies that arrive
1114				 * before clnt_dg_call removes it from
1115				 * the queue.
1116				 */
1117				cr->cr_xid = 0;
1118				cr->cr_mrep = m;
1119				cr->cr_error = 0;
1120				foundreq = 1;
1121				wakeup(cr);
1122				break;
1123			}
1124		}
1125		mtx_unlock(&cs->cs_lock);
1126
1127		/*
1128		 * If we didn't find the matching request, just drop
1129		 * it - its probably a repeated reply.
1130		 */
1131		if (!foundreq)
1132			m_freem(m);
1133	} while (m);
1134	cs->cs_upcallrefs--;
1135	if (cs->cs_upcallrefs < 0)
1136		panic("rpcdg upcall refcnt");
1137	if (cs->cs_upcallrefs == 0)
1138		wakeup(&cs->cs_upcallrefs);
1139	return (SU_OK);
1140}
1141
1142/*
1143 * Wait for all upcalls in progress to complete.
1144 */
1145static void
1146clnt_dg_upcallsdone(struct socket *so, struct cu_socket *cs)
1147{
1148
1149	SOCKBUF_LOCK_ASSERT(&so->so_rcv);
1150
1151	while (cs->cs_upcallrefs > 0)
1152		(void) msleep(&cs->cs_upcallrefs, SOCKBUF_MTX(&so->so_rcv), 0,
1153		    "rpcdgup", 0);
1154}
1155