d/freebsd-nq
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
1a330eb01d
freebsd-nq/sys/rpc/rpcclnt.c
Robert Watson 84d2b7df26 Add GIANT_REQUIRED and WITNESS sleep warnings to uprintf() and tprintf(), 
as they both interact with the tty code (!MPSAFE) and may sleep if the
tty buffer is full (per comment).

Modify all consumers of uprintf() and tprintf() to hold Giant around
calls into these functions.  In most cases, this means adding an
acquisition of Giant immediately around the function.  In some cases
(nfs_timer()), it means acquiring Giant higher up in the callout.

With these changes, UFS no longer panics on SMP when either blocks are
exhausted or inodes are exhausted under load due to races in the tty
code when running without Giant.

NB: Some reduction in calls to uprintf() in the svr4 code is probably
desirable.

NB: In the case of nfs_timer(), calling uprintf() while holding a mutex,
or even in a callout at all, is a bad idea, and will generate warnings
and potential upset.  This needs to be fixed, but was a problem before
this change.

NB: uprintf()/tprintf() sleeping is generally a bad ideas, as is having
non-MPSAFE tty code.

MFC after:	1 week
2005-09-19 16:51:43 +00:00

2097 lines
53 KiB
C
Raw Blame History

/* $FreeBSD$ */
/* $Id: rpcclnt.c,v 1.9 2003/11/05 14:59:03 rees Exp $ */
/*-
* copyright (c) 2003
* the regents of the university of michigan
* all rights reserved
*
* permission is granted to use, copy, create derivative works and redistribute
* this software and such derivative works for any purpose, so long as the name
* of the university of michigan is not used in any advertising or publicity
* pertaining to the use or distribution of this software without specific,
* written prior authorization. if the above copyright notice or any other
* identification of the university of michigan is included in any copy of any
* portion of this software, then the disclaimer below must also be included.
*
* this software is provided as is, without representation from the university
* of michigan as to its fitness for any purpose, and without warranty by the
* university of michigan of any kind, either express or implied, including
* without limitation the implied warranties of merchantability and fitness for
* a particular purpose. the regents of the university of michigan shall not be
* liable for any damages, including special, indirect, incidental, or
* consequential damages, with respect to any claim arising out of or in
* connection with the use of the software, even if it has been or is hereafter
* advised of the possibility of such damages.
*/
/*-
* Copyright (c) 1989, 1991, 1993, 1995 The Regents of the University of
* California. All rights reserved.
*
* This code is derived from software contributed to Berkeley by Rick Macklem at
* The University of Guelph.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions are
* met: 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer. 2.
* Redistributions in binary form must reproduce the above copyright notice,
* this list of conditions and the following disclaimer in the documentation
* and/or other materials provided with the distribution. 3. All advertising
* materials mentioning features or use of this software must display the
* following acknowledgement: This product includes software developed by the
* University of California, Berkeley and its contributors. 4. Neither the
* name of the University nor the names of its contributors may be used to
* endorse or promote products derived from this software without specific
* prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND ANY
* EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
* WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
* DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE FOR
* ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
* SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
* CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
* SUCH DAMAGE.
*
* @(#)nfs_socket.c 8.5 (Berkeley) 3/30/95
*/
/* XXX: kill ugly debug strings */
/* XXX: get rid of proct, as it is not even being used... (or keep it so v{2,3}
* can run, but clean it up! */
#include <sys/param.h>
#include <sys/systm.h>
#include <sys/proc.h>
#include <sys/mount.h>
#include <sys/kernel.h>
#include <sys/mbuf.h>
#include <sys/syslog.h>
#include <sys/malloc.h>
#include <sys/uio.h>
#include <sys/lock.h>
#include <sys/signalvar.h>
#include <sys/sysent.h>
#include <sys/syscall.h>
#include <sys/sysctl.h>
#include <sys/domain.h>
#include <sys/protosw.h>
#include <sys/socket.h>
#include <sys/socketvar.h>
#include <sys/mutex.h>
#include <netinet/in.h>
#include <netinet/tcp.h>
#include <nfs/rpcv2.h>
#include <rpc/rpcm_subs.h>
#include <rpc/rpcclnt.h>
/* memory management */
#ifdef __OpenBSD__
struct pool rpctask_pool;
struct pool rpcclnt_pool;
#define RPCTASKPOOL_LWM 10
#define RPCTASKPOOL_HWM 40
#else
static MALLOC_DEFINE(M_RPC, "rpcclnt", "rpc state");
#endif
#define RPC_RETURN(X) do { RPCDEBUG("returning %d", X); return X; }while(0)
/*
* Estimate rto for an nfs rpc sent via. an unreliable datagram. Use the mean
* and mean deviation of rtt for the appropriate type of rpc for the frequent
* rpcs and a default for the others. The justification for doing "other"
* this way is that these rpcs happen so infrequently that timer est. would
* probably be stale. Also, since many of these rpcs are non-idempotent, a
* conservative timeout is desired. getattr, lookup - A+2D read, write -
* A+4D other - nm_timeo
*/
#define RPC_RTO(n, t) \
((t) == 0 ? (n)->rc_timeo : \
((t) < 3 ? \
(((((n)->rc_srtt[t-1] + 3) >> 2) + (n)->rc_sdrtt[t-1] + 1) >> 1) : \
((((n)->rc_srtt[t-1] + 7) >> 3) + (n)->rc_sdrtt[t-1] + 1)))
#define RPC_SRTT(s,r) (r)->r_rpcclnt->rc_srtt[rpcclnt_proct((s),\
(r)->r_procnum) - 1]
#define RPC_SDRTT(s,r) (r)->r_rpcclnt->rc_sdrtt[rpcclnt_proct((s),\
(r)->r_procnum) - 1]
/*
* There is a congestion window for outstanding rpcs maintained per mount
* point. The cwnd size is adjusted in roughly the way that: Van Jacobson,
* Congestion avoidance and Control, In "Proceedings of SIGCOMM '88". ACM,
* August 1988. describes for TCP. The cwnd size is chopped in half on a
* retransmit timeout and incremented by 1/cwnd when each rpc reply is
* received and a full cwnd of rpcs is in progress. (The sent count and cwnd
* are scaled for integer arith.) Variants of "slow start" were tried and
* were found to be too much of a performance hit (ave. rtt 3 times larger),
* I suspect due to the large rtt that nfs rpcs have.
*/
#define RPC_CWNDSCALE 256
#define RPC_MAXCWND (RPC_CWNDSCALE * 32)
static const int rpcclnt_backoff[8] = {2, 4, 8, 16, 32, 64, 128, 256,};
/* XXX ugly debug strings */
#define RPC_ERRSTR_ACCEPTED_SIZE 6
char *rpc_errstr_accepted[RPC_ERRSTR_ACCEPTED_SIZE] = {
"", /* no good message... */
"remote server hasn't exported program.",
"remote server can't support version number.",
"program can't support procedure.",
"procedure can't decode params.",
"remote error. remote side memory allocation failure?"
};
char *rpc_errstr_denied[2] = {
"remote server doesnt support rpc version 2!",
"remote server authentication error."
};
#define RPC_ERRSTR_AUTH_SIZE 6
char *rpc_errstr_auth[RPC_ERRSTR_AUTH_SIZE] = {
"",
"auth error: bad credential (seal broken).",
"auth error: client must begin new session.",
"auth error: bad verifier (seal broken).",
"auth error: verifier expired or replayed.",
"auth error: rejected for security reasons.",
};
/*
* Static data, mostly RPC constants in XDR form
*/
static u_int32_t rpc_reply, rpc_call, rpc_vers;
/*
* rpc_msgdenied, rpc_mismatch, rpc_auth_unix, rpc_msgaccepted,
* rpc_autherr, rpc_auth_kerb;
*/
static u_int32_t rpcclnt_xid = 0;
static u_int32_t rpcclnt_xid_touched = 0;
struct rpcstats rpcstats;
int rpcclnt_ticks;
SYSCTL_NODE(_kern, OID_AUTO, rpc, CTLFLAG_RD, 0, "RPC Subsystem");
SYSCTL_UINT(_kern_rpc, OID_AUTO, retries, CTLFLAG_RD, &rpcstats.rpcretries, 0, "retries");
SYSCTL_UINT(_kern_rpc, OID_AUTO, request, CTLFLAG_RD, &rpcstats.rpcrequests, 0, "request");
SYSCTL_UINT(_kern_rpc, OID_AUTO, timeouts, CTLFLAG_RD, &rpcstats.rpctimeouts, 0, "timeouts");
SYSCTL_UINT(_kern_rpc, OID_AUTO, unexpected, CTLFLAG_RD, &rpcstats.rpcunexpected, 0, "unexpected");
SYSCTL_UINT(_kern_rpc, OID_AUTO, invalid, CTLFLAG_RD, &rpcstats.rpcinvalid, 0, "invalid");
#ifdef RPCCLNT_DEBUG
int rpcdebugon = 0;
SYSCTL_UINT(_kern_rpc, OID_AUTO, debug_on, CTLFLAG_RW, &rpcdebugon, 0, "RPC Debug messages");
#endif
/*
* Queue head for rpctask's
*/
static
TAILQ_HEAD(, rpctask) rpctask_q;
struct callout rpcclnt_callout;
#ifdef __OpenBSD__
static int rpcclnt_send(struct socket *, struct mbuf *, struct mbuf *, struct rpctask *);
static int rpcclnt_receive(struct rpctask *, struct mbuf **, struct mbuf **, RPC_EXEC_CTX);
#else
static int rpcclnt_send(struct socket *, struct sockaddr *, struct mbuf *, struct rpctask *);
static int rpcclnt_receive(struct rpctask *, struct sockaddr **, struct mbuf **, RPC_EXEC_CTX);
#endif
static int rpcclnt_msg(RPC_EXEC_CTX, const char *, char *);
static int rpcclnt_reply(struct rpctask *, RPC_EXEC_CTX);
static void rpcclnt_timer(void *);
static int rpcclnt_sndlock(int *, struct rpctask *);
static void rpcclnt_sndunlock(int *);
static int rpcclnt_rcvlock(struct rpctask *);
static void rpcclnt_rcvunlock(int *);
#if 0
void rpcclnt_realign(struct mbuf *, int);
#else
static void rpcclnt_realign(struct mbuf **, int);
#endif
static struct mbuf *rpcclnt_buildheader(struct rpcclnt *, int, struct mbuf *, u_int32_t, int *, struct mbuf **, struct ucred *);
static int rpcm_disct(struct mbuf **, caddr_t *, int, int, caddr_t *);
static u_int32_t rpcclnt_proct(struct rpcclnt *, u_int32_t);
static int rpc_adv(struct mbuf **, caddr_t *, int, int);
static void rpcclnt_softterm(struct rpctask * task);
static int rpcauth_buildheader(struct rpc_auth * auth, struct ucred *, struct mbuf **, caddr_t *);
void
rpcclnt_init(void)
{
#ifdef __OpenBSD__
static struct timeout rpcclnt_timer_to;
#endif
rpcclnt_ticks = (hz * RPC_TICKINTVL + 500) / 1000;
if (rpcclnt_ticks < 1)
rpcclnt_ticks = 1;
rpcstats.rpcretries = 0;
rpcstats.rpcrequests = 0;
rpcstats.rpctimeouts = 0;
rpcstats.rpcunexpected = 0;
rpcstats.rpcinvalid = 0;
/*
* rpc constants how about actually using more than one of these!
*/
rpc_reply = txdr_unsigned(RPC_REPLY);
rpc_vers = txdr_unsigned(RPC_VER2);
rpc_call = txdr_unsigned(RPC_CALL);
#if 0
rpc_msgdenied = txdr_unsigned(RPC_MSGDENIED);
rpc_msgaccepted = txdr_unsigned(RPC_MSGACCEPTED);
rpc_mismatch = txdr_unsigned(RPC_MISMATCH);
rpc_autherr = txdr_unsigned(RPC_AUTHERR);
rpc_auth_unix = txdr_unsigned(RPCAUTH_UNIX);
rpc_auth_kerb = txdr_unsigned(RPCAUTH_KERB4);
#endif
/* initialize rpctask queue */
TAILQ_INIT(&rpctask_q);
#ifdef __OpenBSD__
/* initialize pools */
pool_init(&rpctask_pool, sizeof(struct rpctask), 0, 0, RPCTASKPOOL_LWM,
"rpctask_p", NULL);
pool_setlowat(&rpctask_pool, RPCTASKPOOL_LWM);
pool_sethiwat(&rpctask_pool, RPCTASKPOOL_HWM);
pool_init(&rpcclnt_pool, sizeof(struct rpcclnt), 0, 0, 1, "rpcclnt_p", NULL);
/* initialize timers */
timeout_set(&rpcclnt_timer_to, rpcclnt_timer, &rpcclnt_timer_to);
rpcclnt_timer(&rpcclnt_timer_to);
#else /* !__OpenBSD__ */
callout_init(&rpcclnt_callout, 0);
#endif /* !__OpenBSD__ */
RPCDEBUG("rpc initialed");
return;
}
void
rpcclnt_uninit(void)
{
RPCDEBUG("uninit");
/* XXX delete sysctl variables? */
callout_stop(&rpcclnt_callout);
}
int
rpcclnt_setup(clnt, program, addr, sotype, soproto, auth, max_read_size, max_write_size, flags)
struct rpcclnt * clnt;
struct rpc_program * program;
struct sockaddr * addr;
int sotype;
int soproto;
struct rpc_auth * auth;
int max_read_size;
int max_write_size;
int flags;
{
if (clnt == NULL || program == NULL || addr == NULL || auth == NULL)
RPC_RETURN (EFAULT);
if (program->prog_name == NULL)
RPC_RETURN (EFAULT);
clnt->rc_prog = program;
clnt->rc_name = addr;
clnt->rc_sotype = sotype;
clnt->rc_soproto = soproto;
clnt->rc_auth = auth;
clnt->rc_rsize = max_read_size;
clnt->rc_wsize = max_write_size;
clnt->rc_flag = flags;
clnt->rc_proctlen = 0;
clnt->rc_proct = NULL;
RPC_RETURN (0);
}
/*
* Initialize sockets and congestion for a new RPC connection. We do not free
* the sockaddr if error.
*/
int
rpcclnt_connect(rpc, td)
struct rpcclnt *rpc;
RPC_EXEC_CTX td;
{
struct socket *so;
int s, error, rcvreserve, sndreserve;
struct sockaddr *saddr;
#ifdef __OpenBSD__
struct sockaddr_in *sin;
struct mbuf *m;
#else
struct sockaddr_in sin;
int soarg;
struct sockopt opt;
#endif
if (rpc == NULL) {
RPCDEBUG("no rpcclnt struct!\n");
RPC_RETURN(EFAULT);
}
GIANT_REQUIRED; /* XXX until socket locking done */
/* create the socket */
rpc->rc_so = NULL;
saddr = rpc->rc_name;
error = socreate(saddr->sa_family, &rpc->rc_so, rpc->rc_sotype,
rpc->rc_soproto, td->td_ucred, td);
if (error) {
RPCDEBUG("error %d in socreate()", error);
RPC_RETURN(error);
}
so = rpc->rc_so;
rpc->rc_soflags = so->so_proto->pr_flags;
/*
* Some servers require that the client port be a reserved port
* number. We always allocate a reserved port, as this prevents
* filehandle disclosure through UDP port capture.
*/
if (saddr->sa_family == AF_INET) {
#ifdef __OpenBSD__
struct mbuf *mopt;
int *ip;
#endif
#ifdef __OpenBSD__
MGET(mopt, M_TRYWAIT, MT_SOOPTS);
mopt->m_len = sizeof(int);
ip = mtod(mopt, int *);
*ip = IP_PORTRANGE_LOW;
error = sosetopt(so, IPPROTO_IP, IP_PORTRANGE, mopt);
#else
soarg = IP_PORTRANGE_LOW;
bzero(&opt, sizeof(struct sockopt));
opt.sopt_dir = SOPT_SET;
opt.sopt_level = IPPROTO_IP;
opt.sopt_name = IP_PORTRANGE;
opt.sopt_val = &soarg;
opt.sopt_valsize = sizeof(soarg);
error = sosetopt(so, &opt);
#endif
if (error)
goto bad;
#if __OpenBSD__
MGET(m, M_TRYWAIT, MT_SONAME);
sin = mtod(m, struct sockaddr_in *);
sin->sin_len = m->m_len = sizeof(struct sockaddr_in);
sin->sin_family = AF_INET;
sin->sin_addr.s_addr = INADDR_ANY;
sin->sin_port = htons(0);
error = sobind(so, m);
m_freem(m);
#else
sin.sin_len = sizeof(struct sockaddr_in);
sin.sin_family = AF_INET;
sin.sin_addr.s_addr = INADDR_ANY;
sin.sin_port = htons(0);
error = sobind(so, (struct sockaddr *) & sin, td);
#endif
if (error)
goto bad;
#if __OpenBSD__
MGET(mopt, M_TRYWAIT, MT_SOOPTS);
mopt->m_len = sizeof(int);
ip = mtod(mopt, int *);
*ip = IP_PORTRANGE_DEFAULT;
error = sosetopt(so, IPPROTO_IP, IP_PORTRANGE, mopt);
#else
soarg = IP_PORTRANGE_DEFAULT;
bzero(&opt, sizeof(struct sockopt));
opt.sopt_dir = SOPT_SET;
opt.sopt_level = IPPROTO_IP;
opt.sopt_name = IP_PORTRANGE;
opt.sopt_val = &soarg;
opt.sopt_valsize = sizeof(soarg);
error = sosetopt(so, &opt);
#endif
if (error)
goto bad;
}
/*
* Protocols that do not require connections may be optionally left
* unconnected for servers that reply from a port other than
* NFS_PORT.
*/
if (rpc->rc_flag & RPCCLNT_NOCONN) {
if (rpc->rc_soflags & PR_CONNREQUIRED) {
error = ENOTCONN;
goto bad;
}
} else {
error = soconnect(so, saddr, td);
if (error)
goto bad;
/*
* Wait for the connection to complete. Cribbed from the
* connect system call but with the wait timing out so that
* interruptible mounts don't hang here for a long time.
*/
#ifdef __OpenBSD__
s = splsoftnet();
#else
s = splnet();
#endif
while ((so->so_state & SS_ISCONNECTING) && so->so_error == 0) {
(void)tsleep((caddr_t) & so->so_timeo, PSOCK,
"rpc", 2 * hz);
/*
* XXX needs to catch interrupt signals. something
* like this: if ((so->so_state & SS_ISCONNECTING) &&
* so->so_error == 0 && rep && (error =
* nfs_sigintr(nmp, rep, rep->r_td)) != 0) {
* so->so_state &= ~SS_ISCONNECTING; splx(s); goto
* bad; }
*/
}
if (so->so_error) {
error = so->so_error;
so->so_error = 0;
splx(s);
goto bad;
}
splx(s);
}
if (rpc->rc_flag & (RPCCLNT_SOFT | RPCCLNT_INT)) {
so->so_rcv.sb_timeo = (5 * hz);
so->so_snd.sb_timeo = (5 * hz);
} else {
so->so_rcv.sb_timeo = 0;
so->so_snd.sb_timeo = 0;
}
if (rpc->rc_sotype == SOCK_DGRAM) {
sndreserve = rpc->rc_wsize + RPC_MAXPKTHDR;
rcvreserve = rpc->rc_rsize + RPC_MAXPKTHDR;
} else if (rpc->rc_sotype == SOCK_SEQPACKET) {
sndreserve = (rpc->rc_wsize + RPC_MAXPKTHDR) * 2;
rcvreserve = (rpc->rc_rsize + RPC_MAXPKTHDR) * 2;
} else {
if (rpc->rc_sotype != SOCK_STREAM)
panic("rpcclnt_connect() bad sotype");
if (so->so_proto->pr_flags & PR_CONNREQUIRED) {
#ifdef __OpenBSD__
MGET(m, M_TRYWAIT, MT_SOOPTS);
*mtod(m, int32_t *) = 1;
m->m_len = sizeof(int32_t);
sosetopt(so, SOL_SOCKET, SO_KEEPALIVE, m);
#else
soarg = 1;
bzero(&opt, sizeof(struct sockopt));
opt.sopt_dir = SOPT_SET;
opt.sopt_level = SOL_SOCKET;
opt.sopt_name = SO_KEEPALIVE;
opt.sopt_val = &soarg;
opt.sopt_valsize = sizeof(soarg);
sosetopt(so, &opt);
#endif
}
if (so->so_proto->pr_protocol == IPPROTO_TCP) {
#ifdef __OpenBSD__
MGET(m, M_TRYWAIT, MT_SOOPTS);
*mtod(m, int32_t *) = 1;
m->m_len = sizeof(int32_t);
sosetopt(so, IPPROTO_TCP, TCP_NODELAY, m);
#else
soarg = 1;
bzero(&opt, sizeof(struct sockopt));
opt.sopt_dir = SOPT_SET;
opt.sopt_level = IPPROTO_TCP;
opt.sopt_name = TCP_NODELAY;
opt.sopt_val = &soarg;
opt.sopt_valsize = sizeof(soarg);
sosetopt(so, &opt);
#endif
}
sndreserve = (rpc->rc_wsize + RPC_MAXPKTHDR +
sizeof(u_int32_t)) * 2;
rcvreserve = (rpc->rc_rsize + RPC_MAXPKTHDR +
sizeof(u_int32_t)) * 2;
}
error = soreserve(so, sndreserve, rcvreserve);
if (error)
goto bad;
so->so_rcv.sb_flags |= SB_NOINTR;
so->so_snd.sb_flags |= SB_NOINTR;
/* Initialize other non-zero congestion variables */
rpc->rc_srtt[0] = rpc->rc_srtt[1] = rpc->rc_srtt[2] =
rpc->rc_srtt[3] = (RPC_TIMEO << 3);
rpc->rc_sdrtt[0] = rpc->rc_sdrtt[1] = rpc->rc_sdrtt[2] =
rpc->rc_sdrtt[3] = 0;
rpc->rc_cwnd = RPC_MAXCWND / 2; /* Initial send window */
rpc->rc_sent = 0;
rpc->rc_timeouts = 0;
RPC_RETURN(0);
bad:
rpcclnt_disconnect(rpc);
RPC_RETURN(error);
}
/*
* Reconnect routine: Called when a connection is broken on a reliable
* protocol. - clean up the old socket - nfs_connect() again - set
* R_MUSTRESEND for all outstanding requests on mount point If this fails the
* mount point is DEAD! nb: Must be called with the nfs_sndlock() set on the
* mount point.
*/
int
rpcclnt_reconnect(rep, td)
struct rpctask *rep;
RPC_EXEC_CTX td;
{
struct rpctask *rp;
struct rpcclnt *rpc = rep->r_rpcclnt;
int error;
rpcclnt_disconnect(rpc);
while ((error = rpcclnt_connect(rpc, td)) != 0) {
if (error == EINTR || error == ERESTART)
RPC_RETURN(EINTR);
tsleep(&lbolt, PSOCK, "rpccon", 0);
}
/*
* Loop through outstanding request list and fix up all requests on
* old socket.
*/
for (rp = TAILQ_FIRST(&rpctask_q); rp != NULL;
rp = TAILQ_NEXT(rp, r_chain)) {
if (rp->r_rpcclnt == rpc)
rp->r_flags |= R_MUSTRESEND;
}
RPC_RETURN(0);
}
/*
* NFS disconnect. Clean up and unlink.
*/
void
rpcclnt_disconnect(rpc)
struct rpcclnt *rpc;
{
struct socket *so;
GIANT_REQUIRED; /* XXX until socket locking done */
if (rpc->rc_so) {
so = rpc->rc_so;
rpc->rc_so = NULL;
soshutdown(so, 2);
soclose(so);
}
}
void
rpcclnt_safedisconnect(struct rpcclnt * rpc)
{
struct rpctask dummytask;
bzero(&dummytask, sizeof(dummytask));
dummytask.r_rpcclnt = rpc;
rpcclnt_rcvlock(&dummytask);
rpcclnt_disconnect(rpc);
rpcclnt_rcvunlock(&rpc->rc_flag);
}
/*
* This is the nfs send routine. For connection based socket types, it must
* be called with an nfs_sndlock() on the socket. "rep == NULL" indicates
* that it has been called from a server. For the client side: - return EINTR
* if the RPC is terminated, 0 otherwise - set R_MUSTRESEND if the send fails
* for any reason - do any cleanup required by recoverable socket errors
* (???) For the server side: - return EINTR or ERESTART if interrupted by a
* signal - return EPIPE if a connection is lost for connection based sockets
* (TCP...) - do any cleanup required by recoverable socket errors (???)
*/
static int
rpcclnt_send(so, nam, top, rep)
struct socket *so;
#ifdef __OpenBSD__
struct mbuf *nam;
#else
struct sockaddr *nam;
#endif
struct mbuf *top;
struct rpctask *rep;
{
#ifdef __OpenBSD__
struct mbuf *sendnam;
#else
struct sockaddr *sendnam;
struct thread *td = curthread;
#endif
int error, soflags, flags;
GIANT_REQUIRED; /* XXX until socket locking done */
if (rep) {
if (rep->r_flags & R_SOFTTERM) {
m_freem(top);
RPC_RETURN(EINTR);
}
if ((so = rep->r_rpcclnt->rc_so) == NULL) {
rep->r_flags |= R_MUSTRESEND;
m_freem(top);
RPC_RETURN(0);
}
rep->r_flags &= ~R_MUSTRESEND;
soflags = rep->r_rpcclnt->rc_soflags;
} else
soflags = so->so_proto->pr_flags;
if ((soflags & PR_CONNREQUIRED) || (so->so_state & SS_ISCONNECTED))
sendnam = NULL;
else
sendnam = nam;
if (so->so_type == SOCK_SEQPACKET)
flags = MSG_EOR;
else
flags = 0;
error = sosend(so, sendnam, NULL, top, NULL, flags, td);
if (error) {
if (rep) {
log(LOG_INFO, "rpc send error %d for service %s\n", error,
rep->r_rpcclnt->rc_prog->prog_name);
/*
* Deal with errors for the client side.
*/
if (rep->r_flags & R_SOFTTERM)
error = EINTR;
else
rep->r_flags |= R_MUSTRESEND;
} else
log(LOG_INFO, "rpc service send error %d\n", error);
/*
* Handle any recoverable (soft) socket errors here.
*/
if (error != EINTR && error != ERESTART &&
error != EWOULDBLOCK && error != EPIPE)
error = 0;
}
RPC_RETURN(error);
}
/*
* Receive a Sun RPC Request/Reply. For SOCK_DGRAM, the work is all done by
* soreceive(), but for SOCK_STREAM we must deal with the Record Mark and
* consolidate the data into a new mbuf list. nb: Sometimes TCP passes the
* data up to soreceive() in long lists of small mbufs. For SOCK_STREAM we
* must be very careful to read an entire record once we have read any of it,
* even if the system call has been interrupted.
*/
static int
rpcclnt_receive(rep, aname, mp, td)
struct rpctask *rep;
#ifdef __OpenBSD__
struct mbuf **aname;
#else
struct sockaddr **aname;
#endif
struct mbuf **mp;
RPC_EXEC_CTX td;
{
struct socket *so;
struct uio auio;
struct iovec aio;
struct mbuf *m;
struct mbuf *control;
u_int32_t len;
#ifdef __OpenBSD__
struct mbuf **getnam;
#else
struct sockaddr **getnam;
#endif
int error, sotype, rcvflg;
GIANT_REQUIRED; /* XXX until socket locking done */
/*
* Set up arguments for soreceive()
*/
*mp = NULL;
*aname = NULL;
sotype = rep->r_rpcclnt->rc_sotype;
/*
* For reliable protocols, lock against other senders/receivers in
* case a reconnect is necessary. For SOCK_STREAM, first get the
* Record Mark to find out how much more there is to get. We must
* lock the socket against other receivers until we have an entire
* rpc request/reply.
*/
if (sotype != SOCK_DGRAM) {
error = rpcclnt_sndlock(&rep->r_rpcclnt->rc_flag, rep);
if (error)
RPC_RETURN(error);
tryagain:
/*
* Check for fatal errors and resending request.
*/
/*
* Ugh: If a reconnect attempt just happened, rc_so would
* have changed. NULL indicates a failed attempt that has
* essentially shut down this mount point.
*/
if (rep->r_mrep || (rep->r_flags & R_SOFTTERM)) {
rpcclnt_sndunlock(&rep->r_rpcclnt->rc_flag);
RPC_RETURN(EINTR);
}
so = rep->r_rpcclnt->rc_so;
if (!so) {
error = rpcclnt_reconnect(rep, td);
if (error) {
rpcclnt_sndunlock(&rep->r_rpcclnt->rc_flag);
RPC_RETURN(error);
}
goto tryagain;
}
while (rep->r_flags & R_MUSTRESEND) {
m = m_copym(rep->r_mreq, 0, M_COPYALL, M_TRYWAIT);
rpcstats.rpcretries++;
error = rpcclnt_send(so, rep->r_rpcclnt->rc_name, m, rep);
if (error) {
if (error == EINTR || error == ERESTART ||
(error = rpcclnt_reconnect(rep, td)) != 0) {
rpcclnt_sndunlock(&rep->r_rpcclnt->rc_flag);
RPC_RETURN(error);
}
goto tryagain;
}
}
rpcclnt_sndunlock(&rep->r_rpcclnt->rc_flag);
if (sotype == SOCK_STREAM) {
aio.iov_base = (caddr_t) & len;
aio.iov_len = sizeof(u_int32_t);
auio.uio_iov = &aio;
auio.uio_iovcnt = 1;
auio.uio_segflg = UIO_SYSSPACE;
auio.uio_rw = UIO_READ;
auio.uio_offset = 0;
auio.uio_resid = sizeof(u_int32_t);
#ifdef __OpenBSD__
auio.uio_procp = td;
#else
auio.uio_td = td;
#endif
do {
rcvflg = MSG_WAITALL;
error = soreceive(so, NULL, &auio, NULL, NULL, &rcvflg);
if (error == EWOULDBLOCK && rep) {
if (rep->r_flags & R_SOFTTERM)
RPC_RETURN(EINTR);
}
} while (error == EWOULDBLOCK);
if (!error && auio.uio_resid > 0) {
log(LOG_INFO,
"short receive (%zu/%zu) from rpc server %s\n",
sizeof(u_int32_t) - auio.uio_resid,
sizeof(u_int32_t),
rep->r_rpcclnt->rc_prog->prog_name);
error = EPIPE;
}
if (error)
goto errout;
len = ntohl(len) & ~0x80000000;
/*
* This is SERIOUS! We are out of sync with the
* sender and forcing a disconnect/reconnect is all I
* can do.
*/
if (len > RPC_MAXPACKET) {
log(LOG_ERR, "%s (%d) from rpc server %s\n",
"impossible packet length",
len,
rep->r_rpcclnt->rc_prog->prog_name);
error = EFBIG;
goto errout;
}
auio.uio_resid = len;
do {
rcvflg = MSG_WAITALL;
error = soreceive(so, NULL, &auio, mp, NULL, &rcvflg);
} while (error == EWOULDBLOCK || error == EINTR ||
error == ERESTART);
if (!error && auio.uio_resid > 0) {
log(LOG_INFO,
"short receive (%d/%d) from rpc server %s\n",
len - auio.uio_resid, len,
rep->r_rpcclnt->rc_prog->prog_name);
error = EPIPE;
}
} else {
/*
* NB: Since uio_resid is big, MSG_WAITALL is ignored
* and soreceive() will return when it has either a
* control msg or a data msg. We have no use for
* control msg., but must grab them and then throw
* them away so we know what is going on.
*/
auio.uio_resid = len = 100000000; /* Anything Big */
#ifdef __OpenBSD__
auio.uio_procp = td;
#else
auio.uio_td = td;
#endif
do {
rcvflg = 0;
error = soreceive(so, NULL, &auio, mp, &control, &rcvflg);
if (control)
m_freem(control);
if (error == EWOULDBLOCK && rep) {
if (rep->r_flags & R_SOFTTERM)
RPC_RETURN(EINTR);
}
} while (error == EWOULDBLOCK ||
(!error && *mp == NULL && control));
if ((rcvflg & MSG_EOR) == 0)
printf("Egad!!\n");
if (!error && *mp == NULL)
error = EPIPE;
len -= auio.uio_resid;
}
errout:
if (error && error != EINTR && error != ERESTART) {
m_freem(*mp);
*mp = (struct mbuf *) 0;
if (error != EPIPE)
log(LOG_INFO,
"receive error %d from rpc server %s\n",
error,
rep->r_rpcclnt->rc_prog->prog_name);
error = rpcclnt_sndlock(&rep->r_rpcclnt->rc_flag, rep);
if (!error)
error = rpcclnt_reconnect(rep, td);
if (!error)
goto tryagain;
}
} else {
if ((so = rep->r_rpcclnt->rc_so) == NULL)
RPC_RETURN(EACCES);
if (so->so_state & SS_ISCONNECTED)
getnam = NULL;
else
getnam = aname;
auio.uio_resid = len = 1000000;
#ifdef __OpenBSD__
auio.uio_procp = td;
#else
auio.uio_td = td;
#endif
do {
rcvflg = 0;
error = soreceive(so, getnam, &auio, mp, NULL, &rcvflg);
RPCDEBUG("soreceivce returns %d", error);
if (error == EWOULDBLOCK && (rep->r_flags & R_SOFTTERM)) {
RPCDEBUG("wouldblock && softerm -> EINTR");
RPC_RETURN(EINTR);
}
} while (error == EWOULDBLOCK);
len -= auio.uio_resid;
}
if (error) {
m_freem(*mp);
*mp = NULL;
} else {
/*
* Search for any mbufs that are not a multiple of 4 bytes
* long or with m_data not longword aligned. These could
* cause pointer alignment problems, so copy them to well
* aligned mbufs.
*/
rpcclnt_realign(mp, 5 * RPCX_UNSIGNED);
}
RPC_RETURN(error);
}
/*
* Implement receipt of reply on a socket. We must search through the list of
* received datagrams matching them with outstanding requests using the xid,
* until ours is found.
*/
/* ARGSUSED */
static int
rpcclnt_reply(myrep, td)
struct rpctask *myrep;
RPC_EXEC_CTX td;
{
struct rpctask *rep;
struct rpcclnt *rpc = myrep->r_rpcclnt;
int32_t t1;
struct mbuf *mrep, *md;
#ifdef __OpenBSD__
struct mbuf *nam;
#else
struct sockaddr *nam;
#endif
u_int32_t rxid, *tl;
caddr_t dpos, cp2;
int error;
/*
* Loop around until we get our own reply
*/
for (;;) {
/*
* Lock against other receivers so that I don't get stuck in
* sbwait() after someone else has received my reply for me.
* Also necessary for connection based protocols to avoid
* race conditions during a reconnect.
*/
error = rpcclnt_rcvlock(myrep);
if (error)
RPC_RETURN(error);
/* Already received, bye bye */
if (myrep->r_mrep != NULL) {
rpcclnt_rcvunlock(&rpc->rc_flag);
RPC_RETURN(0);
}
/*
* Get the next Rpc reply off the socket
*/
error = rpcclnt_receive(myrep, &nam, &mrep, td);
rpcclnt_rcvunlock(&rpc->rc_flag);
if (error) {
/*
* Ignore routing errors on connectionless
* protocols??
*/
if (RPCIGNORE_SOERROR(rpc->rc_soflags, error)) {
rpc->rc_so->so_error = 0;
if (myrep->r_flags & R_GETONEREP)
RPC_RETURN(0);
RPCDEBUG("ingoring routing error on connectionless protocol.");
continue;
}
RPC_RETURN(error);
}
#ifdef __OpenBSD__
if (nam)
m_freem(nam);
#else
if (nam)
FREE(nam, M_SONAME);
#endif
/*
* Get the xid and check that it is an rpc reply
*/
md = mrep;
dpos = mtod(md, caddr_t);
rpcm_dissect(tl, u_int32_t *, 2 * RPCX_UNSIGNED);
rxid = *tl++;
if (*tl != rpc_reply) {
rpcstats.rpcinvalid++;
m_freem(mrep);
rpcmout:
if (myrep->r_flags & R_GETONEREP)
RPC_RETURN(0);
continue;
}
/*
* Loop through the request list to match up the reply Iff no
* match, just drop the datagram
*/
TAILQ_FOREACH(rep, &rpctask_q, r_chain) {
if (rep->r_mrep == NULL && rxid == rep->r_xid) {
/* Found it.. */
rep->r_mrep = mrep;
rep->r_md = md;
rep->r_dpos = dpos;
/*
* Update congestion window. Do the additive
* increase of one rpc/rtt.
*/
if (rpc->rc_cwnd <= rpc->rc_sent) {
rpc->rc_cwnd +=
(RPC_CWNDSCALE * RPC_CWNDSCALE +
(rpc->rc_cwnd >> 1)) / rpc->rc_cwnd;
if (rpc->rc_cwnd > RPC_MAXCWND)
rpc->rc_cwnd = RPC_MAXCWND;
}
rep->r_flags &= ~R_SENT;
rpc->rc_sent -= RPC_CWNDSCALE;
/*
* Update rtt using a gain of 0.125 on the
* mean and a gain of 0.25 on the deviation.
*/
if (rep->r_flags & R_TIMING) {
/*
* Since the timer resolution of
* NFS_HZ is so course, it can often
* result in r_rtt == 0. Since r_rtt
* == N means that the actual rtt is
* between N+dt and N+2-dt ticks, add
* 1.
*/
t1 = rep->r_rtt + 1;
t1 -= (RPC_SRTT(rpc, rep) >> 3);
RPC_SRTT(rpc, rep) += t1;
if (t1 < 0)
t1 = -t1;
t1 -= (RPC_SDRTT(rpc, rep) >> 2);
RPC_SDRTT(rpc, rep) += t1;
}
rpc->rc_timeouts = 0;
break;
}
}
/*
* If not matched to a request, drop it. If it's mine, get
* out.
*/
if (rep == 0) {
rpcstats.rpcunexpected++;
RPCDEBUG("rpc reply not matched\n");
m_freem(mrep);
} else if (rep == myrep) {
if (rep->r_mrep == NULL)
panic("rpcreply nil");
RPC_RETURN(0);
}
if (myrep->r_flags & R_GETONEREP)
RPC_RETURN(0);
}
}
/* XXX: ignores tryagain! */
/*
* code from nfs_request - goes something like this - fill in task struct -
* links task into list - calls nfs_send() for first transmit - calls
* nfs_receive() to get reply - fills in reply (which should be initialized
* prior to calling), which is valid when 0 is returned and is NEVER freed in
* this function
*
* always frees the request header, but NEVER frees 'mrest'
*
*/
/*
* ruthtype
* pcclnt_setauth() should be used before calling this. EAUTH is returned if
* authentication fails.
*/
/*
* note that reply->result_* are invalid unless reply->type ==
* RPC_MSGACCEPTED and reply->status == RPC_SUCCESS and that reply->verf_*
* are invalid unless reply->type == RPC_MSGACCEPTED
*/
int
rpcclnt_request(rpc, mrest, procnum, td, cred, reply)
struct rpcclnt *rpc;
struct mbuf *mrest;
int procnum;
RPC_EXEC_CTX td;
struct ucred *cred;
struct rpc_reply *reply;
{
struct mbuf *m, *mrep;
struct rpctask *task;
u_int32_t *tl;
struct mbuf *md, *mheadend;
caddr_t dpos, cp2;
int t1, s, error = 0, mrest_len;
u_int32_t xid;
#ifdef __OpenBSD__
task = pool_get(&rpctask_pool, PR_WAITOK);
#else
MALLOC(task, struct rpctask *, sizeof(struct rpctask), M_RPC, (M_WAITOK | M_ZERO));
#endif
task->r_rpcclnt = rpc;
task->r_procnum = procnum;
task->r_td = td;
mrest_len = m_length(mrest, NULL);
m = rpcclnt_buildheader(rpc, procnum, mrest, mrest_len, &xid, &mheadend,
cred);
/*
* For stream protocols, insert a Sun RPC Record Mark.
*/
if (rpc->rc_sotype == SOCK_STREAM) {
M_PREPEND(m, RPCX_UNSIGNED, M_TRYWAIT);
*mtod(m, u_int32_t *) = htonl(0x80000000 |
(m->m_pkthdr.len - RPCX_UNSIGNED));
}
task->r_mreq = m;
task->r_xid = xid;
if (rpc->rc_flag & RPCCLNT_SOFT)
task->r_retry = rpc->rc_retry;
else
task->r_retry = RPC_MAXREXMIT + 1; /* past clip limit */
task->r_rtt = task->r_rexmit = 0;
if (rpcclnt_proct(rpc, procnum) > 0)
task->r_flags = R_TIMING;
else
task->r_flags = 0;
task->r_mrep = NULL;
/*
* Do the client side RPC.
*/
rpcstats.rpcrequests++;
/*
* Chain request into list of outstanding requests. Be sure to put it
* LAST so timer finds oldest requests first.
*/
s = splsoftclock();
if (TAILQ_EMPTY(&rpctask_q))
callout_reset(&rpcclnt_callout, rpcclnt_ticks, rpcclnt_timer,
NULL);
TAILQ_INSERT_TAIL(&rpctask_q, task, r_chain);
/*
* If backing off another request or avoiding congestion, don't send
* this one now but let timer do it. If not timing a request, do it
* now.
*/
if (rpc->rc_so && (rpc->rc_sotype != SOCK_DGRAM ||
(rpc->rc_flag & RPCCLNT_DUMBTIMR) ||
rpc->rc_sent < rpc->rc_cwnd)) {
splx(s);
if (rpc->rc_soflags & PR_CONNREQUIRED)
error = rpcclnt_sndlock(&rpc->rc_flag, task);
if (!error) {
error = rpcclnt_send(rpc->rc_so, rpc->rc_name,
m_copym(m, 0, M_COPYALL, M_TRYWAIT),
task);
if (rpc->rc_soflags & PR_CONNREQUIRED)
rpcclnt_sndunlock(&rpc->rc_flag);
}
if (!error && (task->r_flags & R_MUSTRESEND) == 0) {
rpc->rc_sent += RPC_CWNDSCALE;
task->r_flags |= R_SENT;
}
} else {
splx(s);
task->r_rtt = -1;
}
/*
* Wait for the reply from our send or the timer's.
*/
if (!error || error == EPIPE)
error = rpcclnt_reply(task, td);
/*
* RPC done, unlink the request.
*/
s = splsoftclock();
TAILQ_REMOVE(&rpctask_q, task, r_chain);
if (TAILQ_EMPTY(&rpctask_q))
callout_stop(&rpcclnt_callout);
splx(s);
/*
* Decrement the outstanding request count.
*/
if (task->r_flags & R_SENT) {
task->r_flags &= ~R_SENT; /* paranoia */
rpc->rc_sent -= RPC_CWNDSCALE;
}
/*
* If there was a successful reply and a tprintf msg. tprintf a
* response.
*/
if (!error && (task->r_flags & R_TPRINTFMSG)) {
mtx_lock(&Giant);
rpcclnt_msg(task->r_td, rpc->rc_prog->prog_name,
"is alive again");
mtx_unlock(&Giant);
}
/* free request header (leaving mrest) */
mheadend->m_next = NULL;
m_freem(task->r_mreq);
/* initialize reply */
reply->mrep = task->r_mrep;
reply->verf_md = NULL;
reply->result_md = NULL;
mrep = task->r_mrep;
md = task->r_md;
dpos = task->r_dpos;
/* task structure is no longer needed */
#ifdef __OpenBSD__
pool_put(&rpctask_pool, task);
#else
FREE(task, M_RPC);
#endif
if (error)
goto rpcmout;
/*
* break down the rpc header and check if ok
*/
rpcm_dissect(tl, u_int32_t *, RPCX_UNSIGNED);
reply->stat.type = fxdr_unsigned(u_int32_t, *tl);
if (reply->stat.type == RPC_MSGDENIED) {
rpcm_dissect(tl, u_int32_t *, RPCX_UNSIGNED);
reply->stat.status = fxdr_unsigned(u_int32_t, *tl);
switch (reply->stat.status) {
case RPC_MISMATCH:
rpcm_dissect(tl, u_int32_t *, 2 * RPCX_UNSIGNED);
reply->stat.mismatch_info.low = fxdr_unsigned(u_int32_t, *tl++);
reply->stat.mismatch_info.high = fxdr_unsigned(u_int32_t, *tl);
error = EOPNOTSUPP;
break;
case RPC_AUTHERR:
rpcm_dissect(tl, u_int32_t *, RPCX_UNSIGNED);
reply->stat.autherr = fxdr_unsigned(u_int32_t, *tl);
error = EACCES;
break;
default:
error = EBADRPC;
break;
}
goto rpcmout;
} else if (reply->stat.type != RPC_MSGACCEPTED) {
error = EBADRPC;
goto rpcmout;
}
rpcm_dissect(tl, u_int32_t *, 2 * RPCX_UNSIGNED);
reply->verf_md = md;
reply->verf_dpos = dpos;
reply->verf_type = fxdr_unsigned(u_int32_t, *tl++);
reply->verf_size = fxdr_unsigned(u_int32_t, *tl);
if (reply->verf_size != 0)
rpcm_adv(rpcm_rndup(reply->verf_size));
rpcm_dissect(tl, u_int32_t *, RPCX_UNSIGNED);
reply->stat.status = fxdr_unsigned(u_int32_t, *tl);
if (reply->stat.status == RPC_SUCCESS) {
if ((uint32_t)(dpos - mtod(md, caddr_t)) >= md->m_len) {
RPCDEBUG("where is the next mbuf?");
RPCDEBUG("%d -> %d",
(int)(dpos - mtod(md, caddr_t)), md->m_len);
if (md->m_next == NULL) {
error = EBADRPC;
goto rpcmout;
} else {
reply->result_md = md->m_next;
reply->result_dpos = mtod(reply->result_md,
caddr_t);
}
} else {
reply->result_md = md;
reply->result_dpos = dpos;
}
} else if (reply->stat.status == RPC_PROGMISMATCH) {
rpcm_dissect(tl, u_int32_t *, 2 * RPCX_UNSIGNED);
reply->stat.mismatch_info.low = fxdr_unsigned(u_int32_t, *tl++);
reply->stat.mismatch_info.high = fxdr_unsigned(u_int32_t, *tl);
error = EOPNOTSUPP;
goto rpcmout;
} else {
error = EPROTONOSUPPORT;
goto rpcmout;
}
error = 0;
rpcmout:
RPC_RETURN(error);
}
/*
* Nfs timer routine Scan the nfsreq list and retranmit any requests that
* have timed out To avoid retransmission attempts on STREAM sockets (in the
* future) make sure to set the r_retry field to 0 (implies nm_retry == 0).
*/
void
rpcclnt_timer(arg)
void *arg;
{
#ifdef __OpenBSD__
struct timeout *to = (struct timeout *) arg;
#endif
struct rpctask *rep;
struct mbuf *m;
struct socket *so;
struct rpcclnt *rpc;
int timeo;
int s, error;
#ifndef __OpenBSD__
struct thread *td = curthread;
#endif
#if __OpenBSD__
s = splsoftnet();
#else
s = splnet();
#endif
mtx_lock(&Giant); /* rpc_msg -> tprintf */
TAILQ_FOREACH(rep, &rpctask_q, r_chain) {
rpc = rep->r_rpcclnt;
if (rep->r_mrep || (rep->r_flags & R_SOFTTERM))
continue;
if (rpcclnt_sigintr(rpc, rep, rep->r_td)) {
rep->r_flags |= R_SOFTTERM;
continue;
}
if (rep->r_rtt >= 0) {
rep->r_rtt++;
if (rpc->rc_flag & RPCCLNT_DUMBTIMR)
timeo = rpc->rc_timeo;
else
timeo = RPC_RTO(rpc, rpcclnt_proct(rep->r_rpcclnt,
rep->r_procnum));
if (rpc->rc_timeouts > 0)
timeo *= rpcclnt_backoff[rpc->rc_timeouts - 1];
if (rep->r_rtt <= timeo)
continue;
if (rpc->rc_timeouts < 8)
rpc->rc_timeouts++;
}
/*
* Check for server not responding
*/
if ((rep->r_flags & R_TPRINTFMSG) == 0 &&
rep->r_rexmit > rpc->rc_deadthresh) {
rpcclnt_msg(rep->r_td, rpc->rc_prog->prog_name,
"not responding");
rep->r_flags |= R_TPRINTFMSG;
}
if (rep->r_rexmit >= rep->r_retry) { /* too many */
rpcstats.rpctimeouts++;
rep->r_flags |= R_SOFTTERM;
continue;
}
if (rpc->rc_sotype != SOCK_DGRAM) {
if (++rep->r_rexmit > RPC_MAXREXMIT)
rep->r_rexmit = RPC_MAXREXMIT;
continue;
}
if ((so = rpc->rc_so) == NULL)
continue;
/*
* If there is enough space and the window allows.. Resend it
* Set r_rtt to -1 in case we fail to send it now.
*/
rep->r_rtt = -1;
if (sbspace(&so->so_snd) >= rep->r_mreq->m_pkthdr.len &&
((rpc->rc_flag & RPCCLNT_DUMBTIMR) ||
(rep->r_flags & R_SENT) ||
rpc->rc_sent < rpc->rc_cwnd) &&
(m = m_copym(rep->r_mreq, 0, M_COPYALL, M_DONTWAIT))) {
if ((rpc->rc_flag & RPCCLNT_NOCONN) == 0)
error = (*so->so_proto->pr_usrreqs->pru_send) (so, 0, m,
NULL, NULL, td);
else
error = (*so->so_proto->pr_usrreqs->pru_send)(so, 0, m, rpc->rc_name, NULL, td);
if (error) {
if (RPCIGNORE_SOERROR(rpc->rc_soflags, error))
so->so_error = 0;
} else {
/*
* Iff first send, start timing else turn
* timing off, backoff timer and divide
* congestion window by 2.
*/
if (rep->r_flags & R_SENT) {
rep->r_flags &= ~R_TIMING;
if (++rep->r_rexmit > RPC_MAXREXMIT)
rep->r_rexmit = RPC_MAXREXMIT;
rpc->rc_cwnd >>= 1;
if (rpc->rc_cwnd < RPC_CWNDSCALE)
rpc->rc_cwnd = RPC_CWNDSCALE;
rpcstats.rpcretries++;
} else {
rep->r_flags |= R_SENT;
rpc->rc_sent += RPC_CWNDSCALE;
}
rep->r_rtt = 0;
}
}
}
mtx_unlock(&Giant); /* rpc_msg -> tprintf */
splx(s);
#ifdef __OpenBSD__
timeout_add(rpcclnt_timer, to, rpcclnt_ticks);
#else
callout_reset(&rpcclnt_callout, rpcclnt_ticks, rpcclnt_timer, NULL);
#endif
}
/*
* Test for a termination condition pending on the process. This is used for
* RPCCLNT_INT mounts.
*/
int
rpcclnt_sigintr(rpc, task, pr)
struct rpcclnt *rpc;
struct rpctask *task;
RPC_EXEC_CTX pr;
{
struct proc *p;
sigset_t tmpset;
if (rpc == NULL)
return EFAULT;
/* XXX deal with forced unmounts */
if (task && (task->r_flags & R_SOFTTERM))
RPC_RETURN(EINTR);
if (!(rpc->rc_flag & RPCCLNT_INT))
RPC_RETURN(0);
if (pr == NULL)
return (0);
#ifdef __OpenBSD__
p = pr;
if (p && p->p_siglist &&
(((p->p_siglist & ~p->p_sigmask) & ~p->p_sigignore) &
RPCINT_SIGMASK))
RPC_RETURN(EINTR);
#else
p = pr->td_proc;
PROC_LOCK(p);
tmpset = p->p_siglist;
SIGSETNAND(tmpset, pr->td_sigmask);
mtx_lock(&p->p_sigacts->ps_mtx);
SIGSETNAND(tmpset, p->p_sigacts->ps_sigignore);
mtx_unlock(&p->p_sigacts->ps_mtx);
if (SIGNOTEMPTY(p->p_siglist) && RPCCLNTINT_SIGMASK(tmpset)) {
PROC_UNLOCK(p);
RPC_RETURN(EINTR);
}
PROC_UNLOCK(p);
#endif
RPC_RETURN(0);
}
/*
* Lock a socket against others. Necessary for STREAM sockets to ensure you
* get an entire rpc request/reply and also to avoid race conditions between
* the processes with nfs requests in progress when a reconnect is necessary.
*/
static int
rpcclnt_sndlock(flagp, task)
int *flagp;
struct rpctask *task;
{
RPC_EXEC_CTX p;
int slpflag = 0, slptimeo = 0;
p = task->r_td;
if (task->r_rpcclnt->rc_flag & RPCCLNT_INT)
slpflag = PCATCH;
while (*flagp & RPCCLNT_SNDLOCK) {
if (rpcclnt_sigintr(task->r_rpcclnt, task, p))
RPC_RETURN(EINTR);
*flagp |= RPCCLNT_WANTSND;
(void)tsleep((caddr_t) flagp, slpflag | (PZERO - 1), "rpcsndlck",
slptimeo);
if (slpflag == PCATCH) {
slpflag = 0;
slptimeo = 2 * hz;
}
}
*flagp |= RPCCLNT_SNDLOCK;
RPC_RETURN(0);
}
/*
* Unlock the stream socket for others.
*/
static void
rpcclnt_sndunlock(flagp)
int *flagp;
{
if ((*flagp & RPCCLNT_SNDLOCK) == 0)
panic("rpc sndunlock");
*flagp &= ~RPCCLNT_SNDLOCK;
if (*flagp & RPCCLNT_WANTSND) {
*flagp &= ~RPCCLNT_WANTSND;
wakeup((caddr_t) flagp);
}
}
static int
rpcclnt_rcvlock(task)
struct rpctask *task;
{
int *flagp = &task->r_rpcclnt->rc_flag;
int slpflag, slptimeo = 0;
if (*flagp & RPCCLNT_INT)
slpflag = PCATCH;
else
slpflag = 0;
while (*flagp & RPCCLNT_RCVLOCK) {
if (rpcclnt_sigintr(task->r_rpcclnt, task, task->r_td))
RPC_RETURN(EINTR);
*flagp |= RPCCLNT_WANTRCV;
(void)tsleep((caddr_t) flagp, slpflag | (PZERO - 1), "rpcrcvlk",
slptimeo);
if (slpflag == PCATCH) {
slpflag = 0;
slptimeo = 2 * hz;
}
}
*flagp |= RPCCLNT_RCVLOCK;
RPC_RETURN(0);
}
/*
* Unlock the stream socket for others.
*/
static void
rpcclnt_rcvunlock(flagp)
int *flagp;
{
if ((*flagp & RPCCLNT_RCVLOCK) == 0)
panic("nfs rcvunlock");
*flagp &= ~RPCCLNT_RCVLOCK;
if (*flagp & RPCCLNT_WANTRCV) {
*flagp &= ~RPCCLNT_WANTRCV;
wakeup((caddr_t) flagp);
}
}
#if 0
/*
* Check for badly aligned mbuf data areas and realign data in an mbuf list
* by copying the data areas up, as required.
*/
void
rpcclnt_realign(m, hsiz)
struct mbuf *m;
int hsiz;
{
struct mbuf *m2;
int siz, mlen, olen;
caddr_t tcp, fcp;
struct mbuf *mnew;
while (m) {
/*
* This never happens for UDP, rarely happens for TCP but
* frequently happens for iso transport.
*/
if ((m->m_len & 0x3) || (mtod(m, long)&0x3)) {
olen = m->m_len;
fcp = mtod(m, caddr_t);
if ((long)fcp & 0x3) {
if (m->m_flags & M_PKTHDR)
m_tag_delete_chain(m, NULL);
m->m_flags &= ~M_PKTHDR;
if (m->m_flags & M_EXT)
m->m_data = m->m_ext.ext_buf +
((m->m_ext.ext_size - olen) & ~0x3);
else
m->m_data = m->m_dat;
}
m->m_len = 0;
tcp = mtod(m, caddr_t);
mnew = m;
m2 = m->m_next;
/*
* If possible, only put the first invariant part of
* the RPC header in the first mbuf.
*/
mlen = M_TRAILINGSPACE(m);
if (olen <= hsiz && mlen > hsiz)
mlen = hsiz;
/* Loop through the mbuf list consolidating data. */
while (m) {
while (olen > 0) {
if (mlen == 0) {
if (m2->m_flags & M_PKTHDR)
m_tag_delete_chain(m2, NULL);
m2->m_flags &= ~M_PKTHDR;
if (m2->m_flags & M_EXT)
m2->m_data = m2->m_ext.ext_buf;
else
m2->m_data = m2->m_dat;
m2->m_len = 0;
mlen = M_TRAILINGSPACE(m2);
tcp = mtod(m2, caddr_t);
mnew = m2;
m2 = m2->m_next;
}
siz = min(mlen, olen);
if (tcp != fcp)
bcopy(fcp, tcp, siz);
mnew->m_len += siz;
mlen -= siz;
olen -= siz;
tcp += siz;
fcp += siz;
}
m = m->m_next;
if (m) {
olen = m->m_len;
fcp = mtod(m, caddr_t);
}
}
/*
* Finally, set m_len == 0 for any trailing mbufs
* that have been copied out of.
*/
while (m2) {
m2->m_len = 0;
m2 = m2->m_next;
}
return;
}
m = m->m_next;
}
}
#else
static void
rpcclnt_realign(struct mbuf **pm, int hsiz)
{
struct mbuf *m;
struct mbuf *n = NULL;
int off = 0;
RPCDEBUG("in rpcclnt_realign()");
while ((m = *pm) != NULL) {
if ((m->m_len & 0x3) || (mtod(m, intptr_t) & 0x3)) {
MGET(n, M_TRYWAIT, MT_DATA);
if (m->m_len >= MINCLSIZE) {
MCLGET(n, M_TRYWAIT);
}
n->m_len = 0;
break;
}
pm = &m->m_next;
}
/*
* If n is non-NULL, loop on m copying data, then replace the
* portion of the chain that had to be realigned.
*/
if (n != NULL) {
while (m) {
m_copyback(n, off, m->m_len, mtod(m, caddr_t));
off += m->m_len;
m = m->m_next;
}
m_freem(*pm);
*pm = n;
}
RPCDEBUG("leave rpcclnt_realign()");
}
#endif
static int
rpcclnt_msg(p, server, msg)
RPC_EXEC_CTX p;
const char *server;
char *msg;
{
#ifdef __OpenBSD__
tpr_t tpr;
struct proc *pr = p;
if (p)
tpr = tprintf_open(p);
else
tpr = NULL;
tprintf(tpr, "rpc server %s: %s\n", server, msg);
tprintf_close(tpr);
RPC_RETURN(0);
#else
GIANT_REQUIRED;
tprintf(p ? p->td_proc : NULL, LOG_INFO,
"nfs server %s: %s\n", server, msg);
RPC_RETURN(0);
#endif
}
/*
* Build the RPC header and fill in the authorization info. The authorization
* string argument is only used when the credentials come from outside of the
* kernel (AUTH_KERB). (likewise, the ucred is only used when inside the
* kernel) Returns the head of the mbuf list.
*/
static struct mbuf *
rpcclnt_buildheader(rc, procid, mrest, mrest_len, xidp, mheadend, cred)
struct rpcclnt *rc;
int procid;
struct mbuf *mrest;
u_int32_t mrest_len;
int *xidp;
struct mbuf **mheadend;
struct ucred * cred;
{
/* register */ struct mbuf *mb;
register u_int32_t *tl;
/* register */ caddr_t bpos;
struct mbuf *mreq, *mb2;
int error;
MGETHDR(mb, M_TRYWAIT, MT_DATA);
if (6 * RPCX_UNSIGNED >= MINCLSIZE) {
MCLGET(mb, M_TRYWAIT);
} else if (6 * RPCX_UNSIGNED < MHLEN) {
MH_ALIGN(mb, 6 * RPCX_UNSIGNED);
} else {
RPCDEBUG("mbuf too small");
panic("cheap bailout");
}
mb->m_len = 0;
mreq = mb;
bpos = mtod(mb, caddr_t);
/*
* First the RPC header.
*/
rpcm_build(tl, u_int32_t *, 6 * RPCX_UNSIGNED);
/* Get a new (non-zero) xid */
if ((rpcclnt_xid == 0) && (rpcclnt_xid_touched == 0)) {
rpcclnt_xid = arc4random();
rpcclnt_xid_touched = 1;
} else {
while ((*xidp = arc4random() % 256) == 0);
rpcclnt_xid += *xidp;
}
/* XXX: funky... */
*tl++ = *xidp = txdr_unsigned(rpcclnt_xid);
*tl++ = rpc_call;
*tl++ = rpc_vers;
*tl++ = txdr_unsigned(rc->rc_prog->prog_id);
*tl++ = txdr_unsigned(rc->rc_prog->prog_version);
*tl++ = txdr_unsigned(procid);
if ((error = rpcauth_buildheader(rc->rc_auth, cred, &mb, &bpos))) {
RPCDEBUG("rpcauth_buildheader failed %d", error);
return NULL;
}
mb->m_next = mrest;
*mheadend = mb;
mreq->m_pkthdr.len = m_length(mreq, NULL);
mreq->m_pkthdr.rcvif = NULL;
return (mreq);
}
/*
* Help break down an mbuf chain by setting the first siz bytes contiguous
* pointed to by returned val. This is used by the macros rpcm_dissect and
* rpcm_dissecton for tough cases. (The macros use the vars. dpos and dpos2)
*/
static int
rpcm_disct(mdp, dposp, siz, left, cp2)
struct mbuf **mdp;
caddr_t *dposp;
int siz;
int left;
caddr_t *cp2;
{
struct mbuf *mp, *mp2;
int siz2, xfer;
caddr_t p;
mp = *mdp;
while (left == 0) {
*mdp = mp = mp->m_next;
if (mp == NULL)
RPC_RETURN(EBADRPC);
left = mp->m_len;
*dposp = mtod(mp, caddr_t);
}
if (left >= siz) {
*cp2 = *dposp;
*dposp += siz;
} else if (mp->m_next == NULL) {
RPC_RETURN(EBADRPC);
} else if (siz > MHLEN) {
panic("rpc S too big");
} else {
MGET(mp2, M_TRYWAIT, MT_DATA);
mp2->m_next = mp->m_next;
mp->m_next = mp2;
mp->m_len -= left;
mp = mp2;
*cp2 = p = mtod(mp, caddr_t);
bcopy(*dposp, p, left); /* Copy what was left */
siz2 = siz - left;
p += left;
mp2 = mp->m_next;
/* Loop around copying up the siz2 bytes */
while (siz2 > 0) {
if (mp2 == NULL)
RPC_RETURN(EBADRPC);
xfer = (siz2 > mp2->m_len) ? mp2->m_len : siz2;
if (xfer > 0) {
bcopy(mtod(mp2, caddr_t), p, xfer);
RPCMADV(mp2, xfer);
mp2->m_len -= xfer;
p += xfer;
siz2 -= xfer;
}
if (siz2 > 0)
mp2 = mp2->m_next;
}
mp->m_len = siz;
*mdp = mp2;
*dposp = mtod(mp2, caddr_t);
}
RPC_RETURN(0);
}
static u_int32_t
rpcclnt_proct(rpc, procid)
struct rpcclnt *rpc;
u_int32_t procid;
{
if (rpc->rc_proctlen != 0 && rpc->rc_proct != NULL &&
procid < rpc->rc_proctlen) {
return rpc->rc_proct[procid];
}
return (0);
}
static int
rpc_adv(mdp, dposp, offs, left)
struct mbuf **mdp;
caddr_t *dposp;
int offs;
int left;
{
struct mbuf *m;
int s;
m = *mdp;
s = left;
while (s < offs) {
offs -= s;
m = m->m_next;
if (m == NULL)
RPC_RETURN(EBADRPC);
s = m->m_len;
}
*mdp = m;
*dposp = mtod(m, caddr_t) + offs;
RPC_RETURN(0);
}
int
rpcclnt_cancelreqs(rpc)
struct rpcclnt *rpc;
{
struct rpctask *task;
int i, s;
s = splnet();
TAILQ_FOREACH(task, &rpctask_q, r_chain) {
if (rpc != task->r_rpcclnt || task->r_mrep != NULL ||
(task->r_flags & R_SOFTTERM))
continue;
rpcclnt_softterm(task);
}
splx(s);
for (i = 0; i < 30; i++) {
s = splnet();
TAILQ_FOREACH(task, &rpctask_q, r_chain) {
if (rpc == task->r_rpcclnt)
break;
}
splx(s);
if (task == NULL)
return (0);
tsleep(&lbolt, PSOCK, "nfscancel", 0);
}
return (EBUSY);
}
static void
rpcclnt_softterm(struct rpctask * task)
{
task->r_flags |= R_SOFTTERM;
if (task->r_flags & R_SENT) {
task->r_rpcclnt->rc_sent -= RPC_CWNDSCALE;
task->r_flags &= ~R_SENT;
}
}
#ifndef __OpenBSD__
/* called by rpcclnt_get() */
void
rpcclnt_create(struct rpcclnt ** rpc)
{
MALLOC(*rpc, struct rpcclnt *, sizeof(struct rpcclnt), M_RPC, M_WAITOK | M_ZERO);
}
/* called by rpcclnt_put() */
void
rpcclnt_destroy(struct rpcclnt * rpc)
{
if (rpc != NULL) {
FREE(rpc, M_RPC);
} else {
RPCDEBUG("attempting to free a NULL rpcclnt (not dereferenced)");
}
}
#endif /* !__OpenBSD__ */
/* XXX: add a lock around the auth structure in struct rpcclnt and make this
* call safe for calling durring a connection */
static int
rpcauth_buildheader(struct rpc_auth * auth, struct ucred * cred, struct mbuf ** mhdr, caddr_t * bp)
{
size_t authsiz, verfsiz;
uint32_t mlen, grpsiz;
register struct mbuf *mb, *mb2;
caddr_t bpos;
register u_int32_t *tl;
register int i;
if (auth == NULL || mhdr == NULL)
return EFAULT;
switch (auth->auth_type) {
case RPCAUTH_NULL:
authsiz = 0;
verfsiz = 0;
break;
case RPCAUTH_UNIX:
authsiz = (5 + cred->cr_ngroups) * RPCX_UNSIGNED;
verfsiz = 0;
break;
default:
return EPROTONOSUPPORT;
break;
};
mlen = rpcm_rndup(authsiz) + rpcm_rndup(verfsiz) + 4 * RPCX_UNSIGNED;
mb = *mhdr;
bpos = *bp;
rpcm_build(tl, u_int32_t *, mlen);
*bp = bpos;
*mhdr = mb;
*tl++ = txdr_unsigned(auth->auth_type);
*tl++ = txdr_unsigned(authsiz);
switch (auth->auth_type) {
case RPCAUTH_UNIX:
*tl++ = 0;
*tl++ = 0;
*tl++ = txdr_unsigned(cred->cr_uid);
*tl++ = txdr_unsigned(cred->cr_groups[0]);
grpsiz = cred->cr_ngroups;
*tl++ = txdr_unsigned(grpsiz);
/* XXX: groups[0] is already sent... */
for (i = 0 ; i < grpsiz ; i++) {
*tl++ = txdr_unsigned(cred->cr_groups[i]);
}
/* null verification header */
*tl++ = txdr_unsigned(RPCAUTH_NULL);
*tl++ = 0;
break;
case RPCAUTH_NULL:
/* just a null verf header */
*tl++ = txdr_unsigned(RPCAUTH_NULL);
*tl = 0;
break;
default:
panic("inconsistent rpc auth type");
break;
}
return 0;
}
Reference in New Issue View Git Blame Copy Permalink