net/sunrpc/xprtsock.c

   1 /*
   2  * linux/net/sunrpc/xprtsock.c
   3  *
   4  * Client-side transport implementation for sockets.
   5  *
   6  * TCP callback races fixes (C) 1998 Red Hat
   7  * TCP send fixes (C) 1998 Red Hat
   8  * TCP NFS related read + write fixes
   9  *  (C) 1999 Dave Airlie, University of Limerick, Ireland <airlied@linux.ie>
  10  *
  11  * Rewrite of larges part of the code in order to stabilize TCP stuff.
  12  * Fix behaviour when socket buffer is full.
  13  *  (C) 1999 Trond Myklebust <trond.myklebust@fys.uio.no>
  14  *
  15  * IP socket transport implementation, (C) 2005 Chuck Lever <cel@netapp.com>
  16  *
  17  * IPv6 support contributed by Gilles Quillard, Bull Open Source, 2005.
  18  *   <gilles.quillard@bull.net>
  19  */
  20
  21 #include <linux/types.h>
  22 #include <linux/string.h>
  23 #include <linux/slab.h>
  24 #include <linux/module.h>
  25 #include <linux/capability.h>
  26 #include <linux/pagemap.h>
  27 #include <linux/errno.h>
  28 #include <linux/socket.h>
  29 #include <linux/in.h>
  30 #include <linux/net.h>
  31 #include <linux/mm.h>
  32 #include <linux/un.h>
  33 #include <linux/udp.h>
  34 #include <linux/tcp.h>
  35 #include <linux/sunrpc/clnt.h>
  36 #include <linux/sunrpc/addr.h>
  37 #include <linux/sunrpc/sched.h>
  38 #include <linux/sunrpc/svcsock.h>
  39 #include <linux/sunrpc/xprtsock.h>
  40 #include <linux/file.h>
  41 #ifdef CONFIG_SUNRPC_BACKCHANNEL
  42 #include <linux/sunrpc/bc_xprt.h>
  43 #endif
  44
  45 #include <net/sock.h>
  46 #include <net/checksum.h>
  47 #include <net/udp.h>
  48 #include <net/tcp.h>
  49
  50 #include <trace/events/sunrpc.h>
  51
  52 #include "sunrpc.h"
  53
  54 static void xs_close(struct rpc_xprt *xprt);
  55
  56 /*
  57  * xprtsock tunables
  58  */
  59 static unsigned int xprt_udp_slot_table_entries = RPC_DEF_SLOT_TABLE;
  60 static unsigned int xprt_tcp_slot_table_entries = RPC_MIN_SLOT_TABLE;
  61 static unsigned int xprt_max_tcp_slot_table_entries = RPC_MAX_SLOT_TABLE;
  62
  63 static unsigned int xprt_min_resvport = RPC_DEF_MIN_RESVPORT;
  64 static unsigned int xprt_max_resvport = RPC_DEF_MAX_RESVPORT;
  65
  66 #define XS_TCP_LINGER_TO        (15U * HZ)
  67 static unsigned int xs_tcp_fin_timeout __read_mostly = XS_TCP_LINGER_TO;
  68
  69 /*
  70  * We can register our own files under /proc/sys/sunrpc by
  71  * calling register_sysctl_table() again.  The files in that
  72  * directory become the union of all files registered there.
  73  *
  74  * We simply need to make sure that we don't collide with
  75  * someone else's file names!
  76  */
  77
  78 #ifdef RPC_DEBUG
  79
  80 static unsigned int min_slot_table_size = RPC_MIN_SLOT_TABLE;
  81 static unsigned int max_slot_table_size = RPC_MAX_SLOT_TABLE;
  82 static unsigned int max_tcp_slot_table_limit = RPC_MAX_SLOT_TABLE_LIMIT;
  83 static unsigned int xprt_min_resvport_limit = RPC_MIN_RESVPORT;
  84 static unsigned int xprt_max_resvport_limit = RPC_MAX_RESVPORT;
  85
  86 static struct ctl_table_header *sunrpc_table_header;
  87
  88 /*
  89  * FIXME: changing the UDP slot table size should also resize the UDP
  90  *        socket buffers for existing UDP transports
  91  */
  92 static struct ctl_table xs_tunables_table[] = {
  93         {
  94                 .procname       = "udp_slot_table_entries",
  95                 .data           = &xprt_udp_slot_table_entries,
  96                 .maxlen         = sizeof(unsigned int),
  97                 .mode           = 0644,
  98                 .proc_handler   = proc_dointvec_minmax,
  99                 .extra1         = &min_slot_table_size,
 100                 .extra2         = &max_slot_table_size
 101         },
 102         {
 103                 .procname       = "tcp_slot_table_entries",
 104                 .data           = &xprt_tcp_slot_table_entries,
 105                 .maxlen         = sizeof(unsigned int),
 106                 .mode           = 0644,
 107                 .proc_handler   = proc_dointvec_minmax,
 108                 .extra1         = &min_slot_table_size,
 109                 .extra2         = &max_slot_table_size
 110         },
 111         {
 112                 .procname       = "tcp_max_slot_table_entries",
 113                 .data           = &xprt_max_tcp_slot_table_entries,
 114                 .maxlen         = sizeof(unsigned int),
 115                 .mode           = 0644,
 116                 .proc_handler   = proc_dointvec_minmax,
 117                 .extra1         = &min_slot_table_size,
 118                 .extra2         = &max_tcp_slot_table_limit
 119         },
 120         {
 121                 .procname       = "min_resvport",
 122                 .data           = &xprt_min_resvport,
 123                 .maxlen         = sizeof(unsigned int),
 124                 .mode           = 0644,
 125                 .proc_handler   = proc_dointvec_minmax,
 126                 .extra1         = &xprt_min_resvport_limit,
 127                 .extra2         = &xprt_max_resvport_limit
 128         },
 129         {
 130                 .procname       = "max_resvport",
 131                 .data           = &xprt_max_resvport,
 132                 .maxlen         = sizeof(unsigned int),
 133                 .mode           = 0644,
 134                 .proc_handler   = proc_dointvec_minmax,
 135                 .extra1         = &xprt_min_resvport_limit,
 136                 .extra2         = &xprt_max_resvport_limit
 137         },
 138         {
 139                 .procname       = "tcp_fin_timeout",
 140                 .data           = &xs_tcp_fin_timeout,
 141                 .maxlen         = sizeof(xs_tcp_fin_timeout),
 142                 .mode           = 0644,
 143                 .proc_handler   = proc_dointvec_jiffies,
 144         },
 145         { },
 146 };
 147
 148 static struct ctl_table sunrpc_table[] = {
 149         {
 150                 .procname       = "sunrpc",
 151                 .mode           = 0555,
 152                 .child          = xs_tunables_table
 153         },
 154         { },
 155 };
 156
 157 #endif
 158
 159 /*
 160  * Wait duration for a reply from the RPC portmapper.
 161  */
 162 #define XS_BIND_TO              (60U * HZ)
 163
 164 /*
 165  * Delay if a UDP socket connect error occurs.  This is most likely some
 166  * kind of resource problem on the local host.
 167  */
 168 #define XS_UDP_REEST_TO         (2U * HZ)
 169
 170 /*
 171  * The reestablish timeout allows clients to delay for a bit before attempting
 172  * to reconnect to a server that just dropped our connection.
 173  *
 174  * We implement an exponential backoff when trying to reestablish a TCP
 175  * transport connection with the server.  Some servers like to drop a TCP
 176  * connection when they are overworked, so we start with a short timeout and
 177  * increase over time if the server is down or not responding.
 178  */
 179 #define XS_TCP_INIT_REEST_TO    (3U * HZ)
 180 #define XS_TCP_MAX_REEST_TO     (5U * 60 * HZ)
 181
 182 /*
 183  * TCP idle timeout; client drops the transport socket if it is idle
 184  * for this long.  Note that we also timeout UDP sockets to prevent
 185  * holding port numbers when there is no RPC traffic.
 186  */
 187 #define XS_IDLE_DISC_TO         (5U * 60 * HZ)
 188
 189 #ifdef RPC_DEBUG
 190 # undef  RPC_DEBUG_DATA
 191 # define RPCDBG_FACILITY        RPCDBG_TRANS
 192 #endif
 193
 194 #ifdef RPC_DEBUG_DATA
 195 static void xs_pktdump(char *msg, u32 *packet, unsigned int count)
 196 {
 197         u8 *buf = (u8 *) packet;
 198         int j;
 199
 200         dprintk("RPC:       %s\n", msg);
 201         for (j = 0; j < count && j < 128; j += 4) {
 202                 if (!(j & 31)) {
 203                         if (j)
 204                                 dprintk("\n");
 205                         dprintk("0x%04x ", j);
 206                 }
 207                 dprintk("%02x%02x%02x%02x ",
 208                         buf[j], buf[j+1], buf[j+2], buf[j+3]);
 209         }
 210         dprintk("\n");
 211 }
 212 #else
 213 static inline void xs_pktdump(char *msg, u32 *packet, unsigned int count)
 214 {
 215         /* NOP */
 216 }
 217 #endif
 218
 219 struct sock_xprt {
 220         struct rpc_xprt         xprt;
 221
 222         /*
 223          * Network layer
 224          */
 225         struct socket *         sock;
 226         struct sock *           inet;
 227
 228         /*
 229          * State of TCP reply receive
 230          */
 231         __be32                  tcp_fraghdr,
 232                                 tcp_xid,
 233                                 tcp_calldir;
 234
 235         u32                     tcp_offset,
 236                                 tcp_reclen;
 237
 238         unsigned long           tcp_copied,
 239                                 tcp_flags;
 240
 241         /*
 242          * Connection of transports
 243          */
 244         struct delayed_work     connect_worker;
 245         struct sockaddr_storage srcaddr;
 246         unsigned short          srcport;
 247
 248         /*
 249          * UDP socket buffer size parameters
 250          */
 251         size_t                  rcvsize,
 252                                 sndsize;
 253
 254         /*
 255          * Saved socket callback addresses
 256          */
 257         void                    (*old_data_ready)(struct sock *);
 258         void                    (*old_state_change)(struct sock *);
 259         void                    (*old_write_space)(struct sock *);
 260         void                    (*old_error_report)(struct sock *);
 261 };
 262
 263 /*
 264  * TCP receive state flags
 265  */
 266 #define TCP_RCV_LAST_FRAG       (1UL << 0)
 267 #define TCP_RCV_COPY_FRAGHDR    (1UL << 1)
 268 #define TCP_RCV_COPY_XID        (1UL << 2)
 269 #define TCP_RCV_COPY_DATA       (1UL << 3)
 270 #define TCP_RCV_READ_CALLDIR    (1UL << 4)
 271 #define TCP_RCV_COPY_CALLDIR    (1UL << 5)
 272
 273 /*
 274  * TCP RPC flags
 275  */
 276 #define TCP_RPC_REPLY           (1UL << 6)
 277
 278 static inline struct rpc_xprt *xprt_from_sock(struct sock *sk)
 279 {
 280         return (struct rpc_xprt *) sk->sk_user_data;
 281 }
 282
 283 static inline struct sockaddr *xs_addr(struct rpc_xprt *xprt)
 284 {
 285         return (struct sockaddr *) &xprt->addr;
 286 }
 287
 288 static inline struct sockaddr_un *xs_addr_un(struct rpc_xprt *xprt)
 289 {
 290         return (struct sockaddr_un *) &xprt->addr;
 291 }
 292
 293 static inline struct sockaddr_in *xs_addr_in(struct rpc_xprt *xprt)
 294 {
 295         return (struct sockaddr_in *) &xprt->addr;
 296 }
 297
 298 static inline struct sockaddr_in6 *xs_addr_in6(struct rpc_xprt *xprt)
 299 {
 300         return (struct sockaddr_in6 *) &xprt->addr;
 301 }
 302
 303 static void xs_format_common_peer_addresses(struct rpc_xprt *xprt)
 304 {
 305         struct sockaddr *sap = xs_addr(xprt);
 306         struct sockaddr_in6 *sin6;
 307         struct sockaddr_in *sin;
 308         struct sockaddr_un *sun;
 309         char buf[128];
 310
 311         switch (sap->sa_family) {
 312         case AF_LOCAL:
 313                 sun = xs_addr_un(xprt);
 314                 strlcpy(buf, sun->sun_path, sizeof(buf));
 315                 xprt->address_strings[RPC_DISPLAY_ADDR] =
 316                                                 kstrdup(buf, GFP_KERNEL);
 317                 break;
 318         case AF_INET:
 319                 (void)rpc_ntop(sap, buf, sizeof(buf));
 320                 xprt->address_strings[RPC_DISPLAY_ADDR] =
 321                                                 kstrdup(buf, GFP_KERNEL);
 322                 sin = xs_addr_in(xprt);
 323                 snprintf(buf, sizeof(buf), "%08x", ntohl(sin->sin_addr.s_addr));
 324                 break;
 325         case AF_INET6:
 326                 (void)rpc_ntop(sap, buf, sizeof(buf));
 327                 xprt->address_strings[RPC_DISPLAY_ADDR] =
 328                                                 kstrdup(buf, GFP_KERNEL);
 329                 sin6 = xs_addr_in6(xprt);
 330                 snprintf(buf, sizeof(buf), "%pi6", &sin6->sin6_addr);
 331                 break;
 332         default:
 333                 BUG();
 334         }
 335
 336         xprt->address_strings[RPC_DISPLAY_HEX_ADDR] = kstrdup(buf, GFP_KERNEL);
 337 }
 338
 339 static void xs_format_common_peer_ports(struct rpc_xprt *xprt)
 340 {
 341         struct sockaddr *sap = xs_addr(xprt);
 342         char buf[128];
 343
 344         snprintf(buf, sizeof(buf), "%u", rpc_get_port(sap));
 345         xprt->address_strings[RPC_DISPLAY_PORT] = kstrdup(buf, GFP_KERNEL);
 346
 347         snprintf(buf, sizeof(buf), "%4hx", rpc_get_port(sap));
 348         xprt->address_strings[RPC_DISPLAY_HEX_PORT] = kstrdup(buf, GFP_KERNEL);
 349 }
 350
 351 static void xs_format_peer_addresses(struct rpc_xprt *xprt,
 352                                      const char *protocol,
 353                                      const char *netid)
 354 {
 355         xprt->address_strings[RPC_DISPLAY_PROTO] = protocol;
 356         xprt->address_strings[RPC_DISPLAY_NETID] = netid;
 357         xs_format_common_peer_addresses(xprt);
 358         xs_format_common_peer_ports(xprt);
 359 }
 360
 361 static void xs_update_peer_port(struct rpc_xprt *xprt)
 362 {
 363         kfree(xprt->address_strings[RPC_DISPLAY_HEX_PORT]);
 364         kfree(xprt->address_strings[RPC_DISPLAY_PORT]);
 365
 366         xs_format_common_peer_ports(xprt);
 367 }
 368
 369 static void xs_free_peer_addresses(struct rpc_xprt *xprt)
 370 {
 371         unsigned int i;
 372
 373         for (i = 0; i < RPC_DISPLAY_MAX; i++)
 374                 switch (i) {
 375                 case RPC_DISPLAY_PROTO:
 376                 case RPC_DISPLAY_NETID:
 377                         continue;
 378                 default:
 379                         kfree(xprt->address_strings[i]);
 380                 }
 381 }
 382
 383 #define XS_SENDMSG_FLAGS        (MSG_DONTWAIT | MSG_NOSIGNAL)
 384
 385 static int xs_send_kvec(struct socket *sock, struct sockaddr *addr, int addrlen, struct kvec *vec, unsigned int base, int more)
 386 {
 387         struct msghdr msg = {
 388                 .msg_name       = addr,
 389                 .msg_namelen    = addrlen,
 390                 .msg_flags      = XS_SENDMSG_FLAGS | (more ? MSG_MORE : 0),
 391         };
 392         struct kvec iov = {
 393                 .iov_base       = vec->iov_base + base,
 394                 .iov_len        = vec->iov_len - base,
 395         };
 396
 397         if (iov.iov_len != 0)
 398                 return kernel_sendmsg(sock, &msg, &iov, 1, iov.iov_len);
 399         return kernel_sendmsg(sock, &msg, NULL, 0, 0);
 400 }
 401
 402 static int xs_send_pagedata(struct socket *sock, struct xdr_buf *xdr, unsigned int base, int more, bool zerocopy)
 403 {
 404         ssize_t (*do_sendpage)(struct socket *sock, struct page *page,
 405                         int offset, size_t size, int flags);
 406         struct page **ppage;
 407         unsigned int remainder;
 408         int err, sent = 0;
 409
 410         remainder = xdr->page_len - base;
 411         base += xdr->page_base;
 412         ppage = xdr->pages + (base >> PAGE_SHIFT);
 413         base &= ~PAGE_MASK;
 414         do_sendpage = sock->ops->sendpage;
 415         if (!zerocopy)
 416                 do_sendpage = sock_no_sendpage;
 417         for(;;) {
 418                 unsigned int len = min_t(unsigned int, PAGE_SIZE - base, remainder);
 419                 int flags = XS_SENDMSG_FLAGS;
 420
 421                 remainder -= len;
 422                 if (remainder != 0 || more)
 423                         flags |= MSG_MORE;
 424                 err = do_sendpage(sock, *ppage, base, len, flags);
 425                 if (remainder == 0 || err != len)
 426                         break;
 427                 sent += err;
 428                 ppage++;
 429                 base = 0;
 430         }
 431         if (sent == 0)
 432                 return err;
 433         if (err > 0)
 434                 sent += err;
 435         return sent;
 436 }
 437
 438 /**
 439  * xs_sendpages - write pages directly to a socket
 440  * @sock: socket to send on
 441  * @addr: UDP only -- address of destination
 442  * @addrlen: UDP only -- length of destination address
 443  * @xdr: buffer containing this request
 444  * @base: starting position in the buffer
 445  * @zerocopy: true if it is safe to use sendpage()
 446  *
 447  */
 448 static int xs_sendpages(struct socket *sock, struct sockaddr *addr, int addrlen, struct xdr_buf *xdr, unsigned int base, bool zerocopy)
 449 {
 450         unsigned int remainder = xdr->len - base;
 451         int err, sent = 0;
 452
 453         if (unlikely(!sock))
 454                 return -ENOTSOCK;
 455
 456         clear_bit(SOCK_ASYNC_NOSPACE, &sock->flags);
 457         if (base != 0) {
 458                 addr = NULL;
 459                 addrlen = 0;
 460         }
 461
 462         if (base < xdr->head[0].iov_len || addr != NULL) {
 463                 unsigned int len = xdr->head[0].iov_len - base;
 464                 remainder -= len;
 465                 err = xs_send_kvec(sock, addr, addrlen, &xdr->head[0], base, remainder != 0);
 466                 if (remainder == 0 || err != len)
 467                         goto out;
 468                 sent += err;
 469                 base = 0;
 470         } else
 471                 base -= xdr->head[0].iov_len;
 472
 473         if (base < xdr->page_len) {
 474                 unsigned int len = xdr->page_len - base;
 475                 remainder -= len;
 476                 err = xs_send_pagedata(sock, xdr, base, remainder != 0, zerocopy);
 477                 if (remainder == 0 || err != len)
 478                         goto out;
 479                 sent += err;
 480                 base = 0;
 481         } else
 482                 base -= xdr->page_len;
 483
 484         if (base >= xdr->tail[0].iov_len)
 485                 return sent;
 486         err = xs_send_kvec(sock, NULL, 0, &xdr->tail[0], base, 0);
 487 out:
 488         if (sent == 0)
 489                 return err;
 490         if (err > 0)
 491                 sent += err;
 492         return sent;
 493 }
 494
 495 static void xs_nospace_callback(struct rpc_task *task)
 496 {
 497         struct sock_xprt *transport = container_of(task->tk_rqstp->rq_xprt, struct sock_xprt, xprt);
 498
 499         transport->inet->sk_write_pending--;
 500         clear_bit(SOCK_ASYNC_NOSPACE, &transport->sock->flags);
 501 }
 502
 503 /**
 504  * xs_nospace - place task on wait queue if transmit was incomplete
 505  * @task: task to put to sleep
 506  *
 507  */
 508 static int xs_nospace(struct rpc_task *task)
 509 {
 510         struct rpc_rqst *req = task->tk_rqstp;
 511         struct rpc_xprt *xprt = req->rq_xprt;
 512         struct sock_xprt *transport = container_of(xprt, struct sock_xprt, xprt);
 513         struct sock *sk = transport->inet;
 514         int ret = -EAGAIN;
 515
 516         dprintk("RPC: %5u xmit incomplete (%u left of %u)\n",
 517                         task->tk_pid, req->rq_slen - req->rq_bytes_sent,
 518                         req->rq_slen);
 519
 520         /* Protect against races with write_space */
 521         spin_lock_bh(&xprt->transport_lock);
 522
 523         /* Don't race with disconnect */
 524         if (xprt_connected(xprt)) {
 525                 if (test_bit(SOCK_ASYNC_NOSPACE, &transport->sock->flags)) {
 526                         /*
 527                          * Notify TCP that we're limited by the application
 528                          * window size
 529                          */
 530                         set_bit(SOCK_NOSPACE, &transport->sock->flags);
 531                         sk->sk_write_pending++;
 532                         /* ...and wait for more buffer space */
 533                         xprt_wait_for_buffer_space(task, xs_nospace_callback);
 534                 }
 535         } else {
 536                 clear_bit(SOCK_ASYNC_NOSPACE, &transport->sock->flags);
 537                 ret = -ENOTCONN;
 538         }
 539
 540         spin_unlock_bh(&xprt->transport_lock);
 541
 542         /* Race breaker in case memory is freed before above code is called */
 543         sk->sk_write_space(sk);
 544         return ret;
 545 }
 546
 547 /*
 548  * Construct a stream transport record marker in @buf.
 549  */
 550 static inline void xs_encode_stream_record_marker(struct xdr_buf *buf)
 551 {
 552         u32 reclen = buf->len - sizeof(rpc_fraghdr);
 553         rpc_fraghdr *base = buf->head[0].iov_base;
 554         *base = cpu_to_be32(RPC_LAST_STREAM_FRAGMENT | reclen);
 555 }
 556
 557 /**
 558  * xs_local_send_request - write an RPC request to an AF_LOCAL socket
 559  * @task: RPC task that manages the state of an RPC request
 560  *
 561  * Return values:
 562  *        0:    The request has been sent
 563  *   EAGAIN:    The socket was blocked, please call again later to
 564  *              complete the request
 565  * ENOTCONN:    Caller needs to invoke connect logic then call again
 566  *    other:    Some other error occured, the request was not sent
 567  */
 568 static int xs_local_send_request(struct rpc_task *task)
 569 {
 570         struct rpc_rqst *req = task->tk_rqstp;
 571         struct rpc_xprt *xprt = req->rq_xprt;
 572         struct sock_xprt *transport =
 573                                 container_of(xprt, struct sock_xprt, xprt);
 574         struct xdr_buf *xdr = &req->rq_snd_buf;
 575         int status;
 576
 577         xs_encode_stream_record_marker(&req->rq_snd_buf);
 578
 579         xs_pktdump("packet data:",
 580                         req->rq_svec->iov_base, req->rq_svec->iov_len);
 581
 582         status = xs_sendpages(transport->sock, NULL, 0,
 583                                                 xdr, req->rq_bytes_sent, true);
 584         dprintk("RPC:       %s(%u) = %d\n",
 585                         __func__, xdr->len - req->rq_bytes_sent, status);
 586         if (likely(status >= 0)) {
 587                 req->rq_bytes_sent += status;
 588                 req->rq_xmit_bytes_sent += status;
 589                 if (likely(req->rq_bytes_sent >= req->rq_slen)) {
 590                         req->rq_bytes_sent = 0;
 591                         return 0;
 592                 }
 593                 status = -EAGAIN;
 594         }
 595
 596         switch (status) {
 597         case -ENOBUFS:
 598         case -EAGAIN:
 599                 status = xs_nospace(task);
 600                 break;
 601         default:
 602                 dprintk("RPC:       sendmsg returned unrecognized error %d\n",
 603                         -status);
 604         case -EPIPE:
 605                 xs_close(xprt);
 606                 status = -ENOTCONN;
 607         }
 608
 609         return status;
 610 }
 611
 612 /**
 613  * xs_udp_send_request - write an RPC request to a UDP socket
 614  * @task: address of RPC task that manages the state of an RPC request
 615  *
 616  * Return values:
 617  *        0:    The request has been sent
 618  *   EAGAIN:    The socket was blocked, please call again later to
 619  *              complete the request
 620  * ENOTCONN:    Caller needs to invoke connect logic then call again
 621  *    other:    Some other error occurred, the request was not sent
 622  */
 623 static int xs_udp_send_request(struct rpc_task *task)
 624 {
 625         struct rpc_rqst *req = task->tk_rqstp;
 626         struct rpc_xprt *xprt = req->rq_xprt;
 627         struct sock_xprt *transport = container_of(xprt, struct sock_xprt, xprt);
 628         struct xdr_buf *xdr = &req->rq_snd_buf;
 629         int status;
 630
 631         xs_pktdump("packet data:",
 632                                 req->rq_svec->iov_base,
 633                                 req->rq_svec->iov_len);
 634
 635         if (!xprt_bound(xprt))
 636                 return -ENOTCONN;
 637         status = xs_sendpages(transport->sock,
 638                               xs_addr(xprt),
 639                               xprt->addrlen, xdr,
 640                               req->rq_bytes_sent, true);
 641
 642         dprintk("RPC:       xs_udp_send_request(%u) = %d\n",
 643                         xdr->len - req->rq_bytes_sent, status);
 644
 645         if (status >= 0) {
 646                 req->rq_xmit_bytes_sent += status;
 647                 if (status >= req->rq_slen)
 648                         return 0;
 649                 /* Still some bytes left; set up for a retry later. */
 650                 status = -EAGAIN;
 651         }
 652
 653         switch (status) {
 654         case -ENOTSOCK:
 655                 status = -ENOTCONN;
 656                 /* Should we call xs_close() here? */
 657                 break;
 658         case -EAGAIN:
 659                 status = xs_nospace(task);
 660                 break;
 661         default:
 662                 dprintk("RPC:       sendmsg returned unrecognized error %d\n",
 663                         -status);
 664         case -ENETUNREACH:
 665         case -ENOBUFS:
 666         case -EPIPE:
 667         case -ECONNREFUSED:
 668                 /* When the server has died, an ICMP port unreachable message
 669                  * prompts ECONNREFUSED. */
 670                 clear_bit(SOCK_ASYNC_NOSPACE, &transport->sock->flags);
 671         }
 672
 673         return status;
 674 }
 675
 676 /**
 677  * xs_tcp_shutdown - gracefully shut down a TCP socket
 678  * @xprt: transport
 679  *
 680  * Initiates a graceful shutdown of the TCP socket by calling the
 681  * equivalent of shutdown(SHUT_WR);
 682  */
 683 static void xs_tcp_shutdown(struct rpc_xprt *xprt)
 684 {
 685         struct sock_xprt *transport = container_of(xprt, struct sock_xprt, xprt);
 686         struct socket *sock = transport->sock;
 687
 688         if (sock != NULL) {
 689                 kernel_sock_shutdown(sock, SHUT_WR);
 690                 trace_rpc_socket_shutdown(xprt, sock);
 691         }
 692 }
 693
 694 /**
 695  * xs_tcp_send_request - write an RPC request to a TCP socket
 696  * @task: address of RPC task that manages the state of an RPC request
 697  *
 698  * Return values:
 699  *        0:    The request has been sent
 700  *   EAGAIN:    The socket was blocked, please call again later to
 701  *              complete the request
 702  * ENOTCONN:    Caller needs to invoke connect logic then call again
 703  *    other:    Some other error occurred, the request was not sent
 704  *
 705  * XXX: In the case of soft timeouts, should we eventually give up
 706  *      if sendmsg is not able to make progress?
 707  */
 708 static int xs_tcp_send_request(struct rpc_task *task)
 709 {
 710         struct rpc_rqst *req = task->tk_rqstp;
 711         struct rpc_xprt *xprt = req->rq_xprt;
 712         struct sock_xprt *transport = container_of(xprt, struct sock_xprt, xprt);
 713         struct xdr_buf *xdr = &req->rq_snd_buf;
 714         bool zerocopy = true;
 715         int status;
 716
 717         xs_encode_stream_record_marker(&req->rq_snd_buf);
 718
 719         xs_pktdump("packet data:",
 720                                 req->rq_svec->iov_base,
 721                                 req->rq_svec->iov_len);
 722         /* Don't use zero copy if this is a resend. If the RPC call
 723          * completes while the socket holds a reference to the pages,
 724          * then we may end up resending corrupted data.
 725          */
 726         if (task->tk_flags & RPC_TASK_SENT)
 727                 zerocopy = false;
 728
 729         /* Continue transmitting the packet/record. We must be careful
 730          * to cope with writespace callbacks arriving _after_ we have
 731          * called sendmsg(). */
 732         while (1) {
 733                 status = xs_sendpages(transport->sock,
 734                                         NULL, 0, xdr, req->rq_bytes_sent,
 735                                         zerocopy);
 736
 737                 dprintk("RPC:       xs_tcp_send_request(%u) = %d\n",
 738                                 xdr->len - req->rq_bytes_sent, status);
 739
 740                 if (unlikely(status < 0))
 741                         break;
 742
 743                 /* If we've sent the entire packet, immediately
 744                  * reset the count of bytes sent. */
 745                 req->rq_bytes_sent += status;
 746                 req->rq_xmit_bytes_sent += status;
 747                 if (likely(req->rq_bytes_sent >= req->rq_slen)) {
 748                         req->rq_bytes_sent = 0;
 749                         return 0;
 750                 }
 751
 752                 if (status != 0)
 753                         continue;
 754                 status = -EAGAIN;
 755                 break;
 756         }
 757
 758         switch (status) {
 759         case -ENOTSOCK:
 760                 status = -ENOTCONN;
 761                 /* Should we call xs_close() here? */
 762                 break;
 763         case -ENOBUFS:
 764         case -EAGAIN:
 765                 status = xs_nospace(task);
 766                 break;
 767         default:
 768                 dprintk("RPC:       sendmsg returned unrecognized error %d\n",
 769                         -status);
 770         case -ECONNRESET:
 771                 xs_tcp_shutdown(xprt);
 772         case -ECONNREFUSED:
 773         case -ENOTCONN:
 774         case -EPIPE:
 775                 clear_bit(SOCK_ASYNC_NOSPACE, &transport->sock->flags);
 776         }
 777
 778         return status;
 779 }
 780
 781 /**
 782  * xs_tcp_release_xprt - clean up after a tcp transmission
 783  * @xprt: transport
 784  * @task: rpc task
 785  *
 786  * This cleans up if an error causes us to abort the transmission of a request.
 787  * In this case, the socket may need to be reset in order to avoid confusing
 788  * the server.
 789  */
 790 static void xs_tcp_release_xprt(struct rpc_xprt *xprt, struct rpc_task *task)
 791 {
 792         struct rpc_rqst *req;
 793
 794         if (task != xprt->snd_task)
 795                 return;
 796         if (task == NULL)
 797                 goto out_release;
 798         req = task->tk_rqstp;
 799         if (req == NULL)
 800                 goto out_release;
 801         if (req->rq_bytes_sent == 0)
 802                 goto out_release;
 803         if (req->rq_bytes_sent == req->rq_snd_buf.len)
 804                 goto out_release;
 805         set_bit(XPRT_CLOSE_WAIT, &xprt->state);
 806 out_release:
 807         xprt_release_xprt(xprt, task);
 808 }
 809
 810 static void xs_save_old_callbacks(struct sock_xprt *transport, struct sock *sk)
 811 {
 812         transport->old_data_ready = sk->sk_data_ready;
 813         transport->old_state_change = sk->sk_state_change;
 814         transport->old_write_space = sk->sk_write_space;
 815         transport->old_error_report = sk->sk_error_report;
 816 }
 817
 818 static void xs_restore_old_callbacks(struct sock_xprt *transport, struct sock *sk)
 819 {
 820         sk->sk_data_ready = transport->old_data_ready;
 821         sk->sk_state_change = transport->old_state_change;
 822         sk->sk_write_space = transport->old_write_space;
 823         sk->sk_error_report = transport->old_error_report;
 824 }
 825
 826 /**
 827  * xs_error_report - callback to handle TCP socket state errors
 828  * @sk: socket
 829  *
 830  * Note: we don't call sock_error() since there may be a rpc_task
 831  * using the socket, and so we don't want to clear sk->sk_err.
 832  */
 833 static void xs_error_report(struct sock *sk)
 834 {
 835         struct rpc_xprt *xprt;
 836         int err;
 837
 838         read_lock_bh(&sk->sk_callback_lock);
 839         if (!(xprt = xprt_from_sock(sk)))
 840                 goto out;
 841
 842         err = -sk->sk_err;
 843         if (err == 0)
 844                 goto out;
 845         dprintk("RPC:       xs_error_report client %p, error=%d...\n",
 846                         xprt, -err);
 847         trace_rpc_socket_error(xprt, sk->sk_socket, err);
 848         if (test_bit(XPRT_CONNECTION_REUSE, &xprt->state))
 849                 goto out;
 850         xprt_wake_pending_tasks(xprt, err);
 851  out:
 852         read_unlock_bh(&sk->sk_callback_lock);
 853 }
 854
 855 static void xs_reset_transport(struct sock_xprt *transport)
 856 {
 857         struct socket *sock = transport->sock;
 858         struct sock *sk = transport->inet;
 859
 860         if (sk == NULL)
 861                 return;
 862
 863         transport->srcport = 0;
 864
 865         write_lock_bh(&sk->sk_callback_lock);
 866         transport->inet = NULL;
 867         transport->sock = NULL;
 868
 869         sk->sk_user_data = NULL;
 870
 871         xs_restore_old_callbacks(transport, sk);
 872         write_unlock_bh(&sk->sk_callback_lock);
 873
 874         trace_rpc_socket_close(&transport->xprt, sock);
 875         sock_release(sock);
 876 }
 877
 878 /**
 879  * xs_close - close a socket
 880  * @xprt: transport
 881  *
 882  * This is used when all requests are complete; ie, no DRC state remains
 883  * on the server we want to save.
 884  *
 885  * The caller _must_ be holding XPRT_LOCKED in order to avoid issues with
 886  * xs_reset_transport() zeroing the socket from underneath a writer.
 887  */
 888 static void xs_close(struct rpc_xprt *xprt)
 889 {
 890         struct sock_xprt *transport = container_of(xprt, struct sock_xprt, xprt);
 891
 892         dprintk("RPC:       xs_close xprt %p\n", xprt);
 893
 894         cancel_delayed_work_sync(&transport->connect_worker);
 895
 896         xs_reset_transport(transport);
 897         xprt->reestablish_timeout = 0;
 898
 899         smp_mb__before_atomic();
 900         clear_bit(XPRT_CONNECTION_ABORT, &xprt->state);
 901         clear_bit(XPRT_CLOSE_WAIT, &xprt->state);
 902         clear_bit(XPRT_CLOSING, &xprt->state);
 903         smp_mb__after_atomic();
 904         xprt_disconnect_done(xprt);
 905 }
 906
 907 static void xs_tcp_close(struct rpc_xprt *xprt)
 908 {
 909         if (test_and_clear_bit(XPRT_CONNECTION_CLOSE, &xprt->state))
 910                 xs_close(xprt);
 911         else
 912                 xs_tcp_shutdown(xprt);
 913 }
 914
 915 static void xs_xprt_free(struct rpc_xprt *xprt)
 916 {
 917         xs_free_peer_addresses(xprt);
 918         xprt_free(xprt);
 919 }
 920
 921 /**
 922  * xs_destroy - prepare to shutdown a transport
 923  * @xprt: doomed transport
 924  *
 925  */
 926 static void xs_destroy(struct rpc_xprt *xprt)
 927 {
 928         dprintk("RPC:       xs_destroy xprt %p\n", xprt);
 929
 930         xs_close(xprt);
 931         xs_xprt_free(xprt);
 932         module_put(THIS_MODULE);
 933 }
 934
 935 static int xs_local_copy_to_xdr(struct xdr_buf *xdr, struct sk_buff *skb)
 936 {
 937         struct xdr_skb_reader desc = {
 938                 .skb            = skb,
 939                 .offset         = sizeof(rpc_fraghdr),
 940                 .count          = skb->len - sizeof(rpc_fraghdr),
 941         };
 942
 943         if (xdr_partial_copy_from_skb(xdr, 0, &desc, xdr_skb_read_bits) < 0)
 944                 return -1;
 945         if (desc.count)
 946                 return -1;
 947         return 0;
 948 }
 949
 950 /**
 951  * xs_local_data_ready - "data ready" callback for AF_LOCAL sockets
 952  * @sk: socket with data to read
 953  * @len: how much data to read
 954  *
 955  * Currently this assumes we can read the whole reply in a single gulp.
 956  */
 957 static void xs_local_data_ready(struct sock *sk)
 958 {
 959         struct rpc_task *task;
 960         struct rpc_xprt *xprt;
 961         struct rpc_rqst *rovr;
 962         struct sk_buff *skb;
 963         int err, repsize, copied;
 964         u32 _xid;
 965         __be32 *xp;
 966
 967         read_lock_bh(&sk->sk_callback_lock);
 968         dprintk("RPC:       %s...\n", __func__);
 969         xprt = xprt_from_sock(sk);
 970         if (xprt == NULL)
 971                 goto out;
 972
 973         skb = skb_recv_datagram(sk, 0, 1, &err);
 974         if (skb == NULL)
 975                 goto out;
 976
 977         repsize = skb->len - sizeof(rpc_fraghdr);
 978         if (repsize < 4) {
 979                 dprintk("RPC:       impossible RPC reply size %d\n", repsize);
 980                 goto dropit;
 981         }
 982
 983         /* Copy the XID from the skb... */
 984         xp = skb_header_pointer(skb, sizeof(rpc_fraghdr), sizeof(_xid), &_xid);
 985         if (xp == NULL)
 986                 goto dropit;
 987
 988         /* Look up and lock the request corresponding to the given XID */
 989         spin_lock(&xprt->transport_lock);
 990         rovr = xprt_lookup_rqst(xprt, *xp);
 991         if (!rovr)
 992                 goto out_unlock;
 993         task = rovr->rq_task;
 994
 995         copied = rovr->rq_private_buf.buflen;
 996         if (copied > repsize)
 997                 copied = repsize;
 998
 999         if (xs_local_copy_to_xdr(&rovr->rq_private_buf, skb)) {
1000                 dprintk("RPC:       sk_buff copy failed\n");
1001                 goto out_unlock;
1002         }
1003
1004         xprt_complete_rqst(task, copied);
1005
1006  out_unlock:
1007         spin_unlock(&xprt->transport_lock);
1008  dropit:
1009         skb_free_datagram(sk, skb);
1010  out:
1011         read_unlock_bh(&sk->sk_callback_lock);
1012 }
1013
1014 /**
1015  * xs_udp_data_ready - "data ready" callback for UDP sockets
1016  * @sk: socket with data to read
1017  * @len: how much data to read
1018  *
1019  */
1020 static void xs_udp_data_ready(struct sock *sk)
1021 {
1022         struct rpc_task *task;
1023         struct rpc_xprt *xprt;
1024         struct rpc_rqst *rovr;
1025         struct sk_buff *skb;
1026         int err, repsize, copied;
1027         u32 _xid;
1028         __be32 *xp;
1029
1030         read_lock_bh(&sk->sk_callback_lock);
1031         dprintk("RPC:       xs_udp_data_ready...\n");
1032         if (!(xprt = xprt_from_sock(sk)))
1033                 goto out;
1034
1035         if ((skb = skb_recv_datagram(sk, 0, 1, &err)) == NULL)
1036                 goto out;
1037
1038         repsize = skb->len - sizeof(struct udphdr);
1039         if (repsize < 4) {
1040                 dprintk("RPC:       impossible RPC reply size %d!\n", repsize);
1041                 goto dropit;
1042         }
1043
1044         /* Copy the XID from the skb... */
1045         xp = skb_header_pointer(skb, sizeof(struct udphdr),
1046                                 sizeof(_xid), &_xid);
1047         if (xp == NULL)
1048                 goto dropit;
1049
1050         /* Look up and lock the request corresponding to the given XID */
1051         spin_lock(&xprt->transport_lock);
1052         rovr = xprt_lookup_rqst(xprt, *xp);
1053         if (!rovr)
1054                 goto out_unlock;
1055         task = rovr->rq_task;
1056
1057         if ((copied = rovr->rq_private_buf.buflen) > repsize)
1058                 copied = repsize;
1059
1060         /* Suck it into the iovec, verify checksum if not done by hw. */
1061         if (csum_partial_copy_to_xdr(&rovr->rq_private_buf, skb)) {
1062                 UDPX_INC_STATS_BH(sk, UDP_MIB_INERRORS);
1063                 goto out_unlock;
1064         }
1065
1066         UDPX_INC_STATS_BH(sk, UDP_MIB_INDATAGRAMS);
1067
1068         xprt_adjust_cwnd(xprt, task, copied);
1069         xprt_complete_rqst(task, copied);
1070
1071  out_unlock:
1072         spin_unlock(&xprt->transport_lock);
1073  dropit:
1074         skb_free_datagram(sk, skb);
1075  out:
1076         read_unlock_bh(&sk->sk_callback_lock);
1077 }
1078
1079 /*
1080  * Helper function to force a TCP close if the server is sending
1081  * junk and/or it has put us in CLOSE_WAIT
1082  */
1083 static void xs_tcp_force_close(struct rpc_xprt *xprt)
1084 {
1085         set_bit(XPRT_CONNECTION_CLOSE, &xprt->state);
1086         xprt_force_disconnect(xprt);
1087 }
1088
1089 static inline void xs_tcp_read_fraghdr(struct rpc_xprt *xprt, struct xdr_skb_reader *desc)
1090 {
1091         struct sock_xprt *transport = container_of(xprt, struct sock_xprt, xprt);
1092         size_t len, used;
1093         char *p;
1094
1095         p = ((char *) &transport->tcp_fraghdr) + transport->tcp_offset;
1096         len = sizeof(transport->tcp_fraghdr) - transport->tcp_offset;
1097         used = xdr_skb_read_bits(desc, p, len);
1098         transport->tcp_offset += used;
1099         if (used != len)
1100                 return;
1101
1102         transport->tcp_reclen = ntohl(transport->tcp_fraghdr);
1103         if (transport->tcp_reclen & RPC_LAST_STREAM_FRAGMENT)
1104                 transport->tcp_flags |= TCP_RCV_LAST_FRAG;
1105         else
1106                 transport->tcp_flags &= ~TCP_RCV_LAST_FRAG;
1107         transport->tcp_reclen &= RPC_FRAGMENT_SIZE_MASK;
1108
1109         transport->tcp_flags &= ~TCP_RCV_COPY_FRAGHDR;
1110         transport->tcp_offset = 0;
1111
1112         /* Sanity check of the record length */
1113         if (unlikely(transport->tcp_reclen < 8)) {
1114                 dprintk("RPC:       invalid TCP record fragment length\n");
1115                 xs_tcp_force_close(xprt);
1116                 return;
1117         }
1118         dprintk("RPC:       reading TCP record fragment of length %d\n",
1119                         transport->tcp_reclen);
1120 }
1121
1122 static void xs_tcp_check_fraghdr(struct sock_xprt *transport)
1123 {
1124         if (transport->tcp_offset == transport->tcp_reclen) {
1125                 transport->tcp_flags |= TCP_RCV_COPY_FRAGHDR;
1126                 transport->tcp_offset = 0;
1127                 if (transport->tcp_flags & TCP_RCV_LAST_FRAG) {
1128                         transport->tcp_flags &= ~TCP_RCV_COPY_DATA;
1129                         transport->tcp_flags |= TCP_RCV_COPY_XID;
1130                         transport->tcp_copied = 0;
1131                 }
1132         }
1133 }
1134
1135 static inline void xs_tcp_read_xid(struct sock_xprt *transport, struct xdr_skb_reader *desc)
1136 {
1137         size_t len, used;
1138         char *p;
1139
1140         len = sizeof(transport->tcp_xid) - transport->tcp_offset;
1141         dprintk("RPC:       reading XID (%Zu bytes)\n", len);
1142         p = ((char *) &transport->tcp_xid) + transport->tcp_offset;
1143         used = xdr_skb_read_bits(desc, p, len);
1144         transport->tcp_offset += used;
1145         if (used != len)
1146                 return;
1147         transport->tcp_flags &= ~TCP_RCV_COPY_XID;
1148         transport->tcp_flags |= TCP_RCV_READ_CALLDIR;
1149         transport->tcp_copied = 4;
1150         dprintk("RPC:       reading %s XID %08x\n",
1151                         (transport->tcp_flags & TCP_RPC_REPLY) ? "reply for"
1152                                                               : "request with",
1153                         ntohl(transport->tcp_xid));
1154         xs_tcp_check_fraghdr(transport);
1155 }
1156
1157 static inline void xs_tcp_read_calldir(struct sock_xprt *transport,
1158                                        struct xdr_skb_reader *desc)
1159 {
1160         size_t len, used;
1161         u32 offset;
1162         char *p;
1163
1164         /*
1165          * We want transport->tcp_offset to be 8 at the end of this routine
1166          * (4 bytes for the xid and 4 bytes for the call/reply flag).
1167          * When this function is called for the first time,
1168          * transport->tcp_offset is 4 (after having already read the xid).
1169          */
1170         offset = transport->tcp_offset - sizeof(transport->tcp_xid);
1171         len = sizeof(transport->tcp_calldir) - offset;
1172         dprintk("RPC:       reading CALL/REPLY flag (%Zu bytes)\n", len);
1173         p = ((char *) &transport->tcp_calldir) + offset;
1174         used = xdr_skb_read_bits(desc, p, len);
1175         transport->tcp_offset += used;
1176         if (used != len)
1177                 return;
1178         transport->tcp_flags &= ~TCP_RCV_READ_CALLDIR;
1179         /*
1180          * We don't yet have the XDR buffer, so we will write the calldir
1181          * out after we get the buffer from the 'struct rpc_rqst'
1182          */
1183         switch (ntohl(transport->tcp_calldir)) {
1184         case RPC_REPLY:
1185                 transport->tcp_flags |= TCP_RCV_COPY_CALLDIR;
1186                 transport->tcp_flags |= TCP_RCV_COPY_DATA;
1187                 transport->tcp_flags |= TCP_RPC_REPLY;
1188                 break;
1189         case RPC_CALL:
1190                 transport->tcp_flags |= TCP_RCV_COPY_CALLDIR;
1191                 transport->tcp_flags |= TCP_RCV_COPY_DATA;
1192                 transport->tcp_flags &= ~TCP_RPC_REPLY;
1193                 break;
1194         default:
1195                 dprintk("RPC:       invalid request message type\n");
1196                 xs_tcp_force_close(&transport->xprt);
1197         }
1198         xs_tcp_check_fraghdr(transport);
1199 }
1200
1201 static inline void xs_tcp_read_common(struct rpc_xprt *xprt,
1202                                      struct xdr_skb_reader *desc,
1203                                      struct rpc_rqst *req)
1204 {
1205         struct sock_xprt *transport =
1206                                 container_of(xprt, struct sock_xprt, xprt);
1207         struct xdr_buf *rcvbuf;
1208         size_t len;
1209         ssize_t r;
1210
1211         rcvbuf = &req->rq_private_buf;
1212
1213         if (transport->tcp_flags & TCP_RCV_COPY_CALLDIR) {
1214                 /*
1215                  * Save the RPC direction in the XDR buffer
1216                  */
1217                 memcpy(rcvbuf->head[0].iov_base + transport->tcp_copied,
1218                         &transport->tcp_calldir,
1219                         sizeof(transport->tcp_calldir));
1220                 transport->tcp_copied += sizeof(transport->tcp_calldir);
1221                 transport->tcp_flags &= ~TCP_RCV_COPY_CALLDIR;
1222         }
1223
1224         len = desc->count;
1225         if (len > transport->tcp_reclen - transport->tcp_offset) {
1226                 struct xdr_skb_reader my_desc;
1227
1228                 len = transport->tcp_reclen - transport->tcp_offset;
1229                 memcpy(&my_desc, desc, sizeof(my_desc));
1230                 my_desc.count = len;
1231                 r = xdr_partial_copy_from_skb(rcvbuf, transport->tcp_copied,
1232                                           &my_desc, xdr_skb_read_bits);
1233                 desc->count -= r;
1234                 desc->offset += r;
1235         } else
1236                 r = xdr_partial_copy_from_skb(rcvbuf, transport->tcp_copied,
1237                                           desc, xdr_skb_read_bits);
1238
1239         if (r > 0) {
1240                 transport->tcp_copied += r;
1241                 transport->tcp_offset += r;
1242         }
1243         if (r != len) {
1244                 /* Error when copying to the receive buffer,
1245                  * usually because we weren't able to allocate
1246                  * additional buffer pages. All we can do now
1247                  * is turn off TCP_RCV_COPY_DATA, so the request
1248                  * will not receive any additional updates,
1249                  * and time out.
1250                  * Any remaining data from this record will
1251                  * be discarded.
1252                  */
1253                 transport->tcp_flags &= ~TCP_RCV_COPY_DATA;
1254                 dprintk("RPC:       XID %08x truncated request\n",
1255                                 ntohl(transport->tcp_xid));
1256                 dprintk("RPC:       xprt = %p, tcp_copied = %lu, "
1257                                 "tcp_offset = %u, tcp_reclen = %u\n",
1258                                 xprt, transport->tcp_copied,
1259                                 transport->tcp_offset, transport->tcp_reclen);
1260                 return;
1261         }
1262
1263         dprintk("RPC:       XID %08x read %Zd bytes\n",
1264                         ntohl(transport->tcp_xid), r);
1265         dprintk("RPC:       xprt = %p, tcp_copied = %lu, tcp_offset = %u, "
1266                         "tcp_reclen = %u\n", xprt, transport->tcp_copied,
1267                         transport->tcp_offset, transport->tcp_reclen);
1268
1269         if (transport->tcp_copied == req->rq_private_buf.buflen)
1270                 transport->tcp_flags &= ~TCP_RCV_COPY_DATA;
1271         else if (transport->tcp_offset == transport->tcp_reclen) {
1272                 if (transport->tcp_flags & TCP_RCV_LAST_FRAG)
1273                         transport->tcp_flags &= ~TCP_RCV_COPY_DATA;
1274         }
1275 }
1276
1277 /*
1278  * Finds the request corresponding to the RPC xid and invokes the common
1279  * tcp read code to read the data.
1280  */
1281 static inline int xs_tcp_read_reply(struct rpc_xprt *xprt,
1282                                     struct xdr_skb_reader *desc)
1283 {
1284         struct sock_xprt *transport =
1285                                 container_of(xprt, struct sock_xprt, xprt);
1286         struct rpc_rqst *req;
1287
1288         dprintk("RPC:       read reply XID %08x\n", ntohl(transport->tcp_xid));
1289
1290         /* Find and lock the request corresponding to this xid */
1291         spin_lock(&xprt->transport_lock);
1292         req = xprt_lookup_rqst(xprt, transport->tcp_xid);
1293         if (!req) {
1294                 dprintk("RPC:       XID %08x request not found!\n",
1295                                 ntohl(transport->tcp_xid));
1296                 spin_unlock(&xprt->transport_lock);
1297                 return -1;
1298         }
1299
1300         xs_tcp_read_common(xprt, desc, req);
1301
1302         if (!(transport->tcp_flags & TCP_RCV_COPY_DATA))
1303                 xprt_complete_rqst(req->rq_task, transport->tcp_copied);
1304
1305         spin_unlock(&xprt->transport_lock);
1306         return 0;
1307 }
1308
1309 #if defined(CONFIG_SUNRPC_BACKCHANNEL)
1310 /*
1311  * Obtains an rpc_rqst previously allocated and invokes the common
1312  * tcp read code to read the data.  The result is placed in the callback
1313  * queue.
1314  * If we're unable to obtain the rpc_rqst we schedule the closing of the
1315  * connection and return -1.
1316  */
1317 static int xs_tcp_read_callback(struct rpc_xprt *xprt,
1318                                        struct xdr_skb_reader *desc)
1319 {
1320         struct sock_xprt *transport =
1321                                 container_of(xprt, struct sock_xprt, xprt);
1322         struct rpc_rqst *req;
1323
1324         /* Look up and lock the request corresponding to the given XID */
1325         spin_lock(&xprt->transport_lock);
1326         req = xprt_lookup_bc_request(xprt, transport->tcp_xid);
1327         if (req == NULL) {
1328                 spin_unlock(&xprt->transport_lock);
1329                 printk(KERN_WARNING "Callback slot table overflowed\n");
1330                 xprt_force_disconnect(xprt);
1331                 return -1;
1332         }
1333
1334         dprintk("RPC:       read callback  XID %08x\n", ntohl(req->rq_xid));
1335         xs_tcp_read_common(xprt, desc, req);
1336
1337         if (!(transport->tcp_flags & TCP_RCV_COPY_DATA))
1338                 xprt_complete_bc_request(req, transport->tcp_copied);
1339         spin_unlock(&xprt->transport_lock);
1340
1341         return 0;
1342 }
1343
1344 static inline int _xs_tcp_read_data(struct rpc_xprt *xprt,
1345                                         struct xdr_skb_reader *desc)
1346 {
1347         struct sock_xprt *transport =
1348                                 container_of(xprt, struct sock_xprt, xprt);
1349
1350         return (transport->tcp_flags & TCP_RPC_REPLY) ?
1351                 xs_tcp_read_reply(xprt, desc) :
1352                 xs_tcp_read_callback(xprt, desc);
1353 }
1354 #else
1355 static inline int _xs_tcp_read_data(struct rpc_xprt *xprt,
1356                                         struct xdr_skb_reader *desc)
1357 {
1358         return xs_tcp_read_reply(xprt, desc);
1359 }
1360 #endif /* CONFIG_SUNRPC_BACKCHANNEL */
1361
1362 /*
1363  * Read data off the transport.  This can be either an RPC_CALL or an
1364  * RPC_REPLY.  Relay the processing to helper functions.
1365  */
1366 static void xs_tcp_read_data(struct rpc_xprt *xprt,
1367                                     struct xdr_skb_reader *desc)
1368 {
1369         struct sock_xprt *transport =
1370                                 container_of(xprt, struct sock_xprt, xprt);
1371
1372         if (_xs_tcp_read_data(xprt, desc) == 0)
1373                 xs_tcp_check_fraghdr(transport);
1374         else {
1375                 /*
1376                  * The transport_lock protects the request handling.
1377                  * There's no need to hold it to update the tcp_flags.
1378                  */
1379                 transport->tcp_flags &= ~TCP_RCV_COPY_DATA;
1380         }
1381 }
1382
1383 static inline void xs_tcp_read_discard(struct sock_xprt *transport, struct xdr_skb_reader *desc)
1384 {
1385         size_t len;
1386
1387         len = transport->tcp_reclen - transport->tcp_offset;
1388         if (len > desc->count)
1389                 len = desc->count;
1390         desc->count -= len;
1391         desc->offset += len;
1392         transport->tcp_offset += len;
1393         dprintk("RPC:       discarded %Zu bytes\n", len);
1394         xs_tcp_check_fraghdr(transport);
1395 }
1396
1397 static int xs_tcp_data_recv(read_descriptor_t *rd_desc, struct sk_buff *skb, unsigned int offset, size_t len)
1398 {
1399         struct rpc_xprt *xprt = rd_desc->arg.data;
1400         struct sock_xprt *transport = container_of(xprt, struct sock_xprt, xprt);
1401         struct xdr_skb_reader desc = {
1402                 .skb    = skb,
1403                 .offset = offset,
1404                 .count  = len,
1405         };
1406
1407         dprintk("RPC:       xs_tcp_data_recv started\n");
1408         do {
1409                 /* Read in a new fragment marker if necessary */
1410                 /* Can we ever really expect to get completely empty fragments? */
1411                 if (transport->tcp_flags & TCP_RCV_COPY_FRAGHDR) {
1412                         xs_tcp_read_fraghdr(xprt, &desc);
1413                         continue;
1414                 }
1415                 /* Read in the xid if necessary */
1416                 if (transport->tcp_flags & TCP_RCV_COPY_XID) {
1417                         xs_tcp_read_xid(transport, &desc);
1418                         continue;
1419                 }
1420                 /* Read in the call/reply flag */
1421                 if (transport->tcp_flags & TCP_RCV_READ_CALLDIR) {
1422                         xs_tcp_read_calldir(transport, &desc);
1423                         continue;
1424                 }
1425                 /* Read in the request data */
1426                 if (transport->tcp_flags & TCP_RCV_COPY_DATA) {
1427                         xs_tcp_read_data(xprt, &desc);
1428                         continue;
1429                 }
1430                 /* Skip over any trailing bytes on short reads */
1431                 xs_tcp_read_discard(transport, &desc);
1432         } while (desc.count);
1433         dprintk("RPC:       xs_tcp_data_recv done\n");
1434         return len - desc.count;
1435 }
1436
1437 /**
1438  * xs_tcp_data_ready - "data ready" callback for TCP sockets
1439  * @sk: socket with data to read
1440  * @bytes: how much data to read
1441  *
1442  */
1443 static void xs_tcp_data_ready(struct sock *sk)
1444 {
1445         struct rpc_xprt *xprt;
1446         read_descriptor_t rd_desc;
1447         int read;
1448
1449         dprintk("RPC:       xs_tcp_data_ready...\n");
1450
1451         read_lock_bh(&sk->sk_callback_lock);
1452         if (!(xprt = xprt_from_sock(sk)))
1453                 goto out;
1454         /* Any data means we had a useful conversation, so
1455          * the we don't need to delay the next reconnect
1456          */
1457         if (xprt->reestablish_timeout)
1458                 xprt->reestablish_timeout = 0;
1459
1460         /* We use rd_desc to pass struct xprt to xs_tcp_data_recv */
1461         rd_desc.arg.data = xprt;
1462         do {
1463                 rd_desc.count = 65536;
1464                 read = tcp_read_sock(sk, &rd_desc, xs_tcp_data_recv);
1465         } while (read > 0);
1466 out:
1467         read_unlock_bh(&sk->sk_callback_lock);
1468 }
1469
1470 /*
1471  * Do the equivalent of linger/linger2 handling for dealing with
1472  * broken servers that don't close the socket in a timely
1473  * fashion
1474  */
1475 static void xs_tcp_schedule_linger_timeout(struct rpc_xprt *xprt,
1476                 unsigned long timeout)
1477 {
1478         struct sock_xprt *transport;
1479
1480         if (xprt_test_and_set_connecting(xprt))
1481                 return;
1482         set_bit(XPRT_CONNECTION_ABORT, &xprt->state);
1483         transport = container_of(xprt, struct sock_xprt, xprt);
1484         queue_delayed_work(rpciod_workqueue, &transport->connect_worker,
1485                            timeout);
1486 }
1487
1488 static void xs_tcp_cancel_linger_timeout(struct rpc_xprt *xprt)
1489 {
1490         struct sock_xprt *transport;
1491
1492         transport = container_of(xprt, struct sock_xprt, xprt);
1493
1494         if (!test_bit(XPRT_CONNECTION_ABORT, &xprt->state) ||
1495             !cancel_delayed_work(&transport->connect_worker))
1496                 return;
1497         clear_bit(XPRT_CONNECTION_ABORT, &xprt->state);
1498         xprt_clear_connecting(xprt);
1499 }
1500
1501 static void xs_sock_reset_connection_flags(struct rpc_xprt *xprt)
1502 {
1503         smp_mb__before_atomic();
1504         clear_bit(XPRT_CONNECTION_ABORT, &xprt->state);
1505         clear_bit(XPRT_CONNECTION_CLOSE, &xprt->state);
1506         clear_bit(XPRT_CLOSE_WAIT, &xprt->state);
1507         clear_bit(XPRT_CLOSING, &xprt->state);
1508         smp_mb__after_atomic();
1509 }
1510
1511 static void xs_sock_mark_closed(struct rpc_xprt *xprt)
1512 {
1513         xs_sock_reset_connection_flags(xprt);
1514         /* Mark transport as closed and wake up all pending tasks */
1515         xprt_disconnect_done(xprt);
1516 }
1517
1518 /**
1519  * xs_tcp_state_change - callback to handle TCP socket state changes
1520  * @sk: socket whose state has changed
1521  *
1522  */
1523 static void xs_tcp_state_change(struct sock *sk)
1524 {
1525         struct rpc_xprt *xprt;
1526
1527         read_lock_bh(&sk->sk_callback_lock);
1528         if (!(xprt = xprt_from_sock(sk)))
1529                 goto out;
1530         dprintk("RPC:       xs_tcp_state_change client %p...\n", xprt);
1531         dprintk("RPC:       state %x conn %d dead %d zapped %d sk_shutdown %d\n",
1532                         sk->sk_state, xprt_connected(xprt),
1533                         sock_flag(sk, SOCK_DEAD),
1534                         sock_flag(sk, SOCK_ZAPPED),
1535                         sk->sk_shutdown);
1536
1537         trace_rpc_socket_state_change(xprt, sk->sk_socket);
1538         switch (sk->sk_state) {
1539         case TCP_ESTABLISHED:
1540                 spin_lock(&xprt->transport_lock);
1541                 if (!xprt_test_and_set_connected(xprt)) {
1542                         struct sock_xprt *transport = container_of(xprt,
1543                                         struct sock_xprt, xprt);
1544
1545                         /* Reset TCP record info */
1546                         transport->tcp_offset = 0;
1547                         transport->tcp_reclen = 0;
1548                         transport->tcp_copied = 0;
1549                         transport->tcp_flags =
1550                                 TCP_RCV_COPY_FRAGHDR | TCP_RCV_COPY_XID;
1551                         xprt->connect_cookie++;
1552
1553                         xprt_wake_pending_tasks(xprt, -EAGAIN);
1554                 }
1555                 spin_unlock(&xprt->transport_lock);
1556                 break;
1557         case TCP_FIN_WAIT1:
1558                 /* The client initiated a shutdown of the socket */
1559                 xprt->connect_cookie++;
1560                 xprt->reestablish_timeout = 0;
1561                 set_bit(XPRT_CLOSING, &xprt->state);
1562                 smp_mb__before_atomic();
1563                 clear_bit(XPRT_CONNECTED, &xprt->state);
1564                 clear_bit(XPRT_CLOSE_WAIT, &xprt->state);
1565                 smp_mb__after_atomic();
1566                 xs_tcp_schedule_linger_timeout(xprt, xs_tcp_fin_timeout);
1567                 break;
1568         case TCP_CLOSE_WAIT:
1569                 /* The server initiated a shutdown of the socket */
1570                 xprt->connect_cookie++;
1571                 clear_bit(XPRT_CONNECTED, &xprt->state);
1572                 xs_tcp_force_close(xprt);
1573         case TCP_CLOSING:
1574                 /*
1575                  * If the server closed down the connection, make sure that
1576                  * we back off before reconnecting
1577                  */
1578                 if (xprt->reestablish_timeout < XS_TCP_INIT_REEST_TO)
1579                         xprt->reestablish_timeout = XS_TCP_INIT_REEST_TO;
1580                 break;
1581         case TCP_LAST_ACK:
1582                 set_bit(XPRT_CLOSING, &xprt->state);
1583                 xs_tcp_schedule_linger_timeout(xprt, xs_tcp_fin_timeout);
1584                 smp_mb__before_atomic();
1585                 clear_bit(XPRT_CONNECTED, &xprt->state);
1586                 smp_mb__after_atomic();
1587                 break;
1588         case TCP_CLOSE:
1589                 xs_tcp_cancel_linger_timeout(xprt);
1590                 xs_sock_mark_closed(xprt);
1591         }
1592  out:
1593         read_unlock_bh(&sk->sk_callback_lock);
1594 }
1595
1596 static void xs_write_space(struct sock *sk)
1597 {
1598         struct socket *sock;
1599         struct rpc_xprt *xprt;
1600
1601         if (unlikely(!(sock = sk->sk_socket)))
1602                 return;
1603         clear_bit(SOCK_NOSPACE, &sock->flags);
1604
1605         if (unlikely(!(xprt = xprt_from_sock(sk))))
1606                 return;
1607         if (test_and_clear_bit(SOCK_ASYNC_NOSPACE, &sock->flags) == 0)
1608                 return;
1609
1610         xprt_write_space(xprt);
1611 }
1612
1613 /**
1614  * xs_udp_write_space - callback invoked when socket buffer space
1615  *                             becomes available
1616  * @sk: socket whose state has changed
1617  *
1618  * Called when more output buffer space is available for this socket.
1619  * We try not to wake our writers until they can make "significant"
1620  * progress, otherwise we'll waste resources thrashing kernel_sendmsg
1621  * with a bunch of small requests.
1622  */
1623 static void xs_udp_write_space(struct sock *sk)
1624 {
1625         read_lock_bh(&sk->sk_callback_lock);
1626
1627         /* from net/core/sock.c:sock_def_write_space */
1628         if (sock_writeable(sk))
1629                 xs_write_space(sk);
1630
1631         read_unlock_bh(&sk->sk_callback_lock);
1632 }
1633
1634 /**
1635  * xs_tcp_write_space - callback invoked when socket buffer space
1636  *                             becomes available
1637  * @sk: socket whose state has changed
1638  *
1639  * Called when more output buffer space is available for this socket.
1640  * We try not to wake our writers until they can make "significant"
1641  * progress, otherwise we'll waste resources thrashing kernel_sendmsg
1642  * with a bunch of small requests.
1643  */
1644 static void xs_tcp_write_space(struct sock *sk)
1645 {
1646         read_lock_bh(&sk->sk_callback_lock);
1647
1648         /* from net/core/stream.c:sk_stream_write_space */
1649         if (sk_stream_is_writeable(sk))
1650                 xs_write_space(sk);
1651
1652         read_unlock_bh(&sk->sk_callback_lock);
1653 }
1654
1655 static void xs_udp_do_set_buffer_size(struct rpc_xprt *xprt)
1656 {
1657         struct sock_xprt *transport = container_of(xprt, struct sock_xprt, xprt);
1658         struct sock *sk = transport->inet;
1659
1660         if (transport->rcvsize) {
1661                 sk->sk_userlocks |= SOCK_RCVBUF_LOCK;
1662                 sk->sk_rcvbuf = transport->rcvsize * xprt->max_reqs * 2;
1663         }
1664         if (transport->sndsize) {
1665                 sk->sk_userlocks |= SOCK_SNDBUF_LOCK;
1666                 sk->sk_sndbuf = transport->sndsize * xprt->max_reqs * 2;
1667                 sk->sk_write_space(sk);
1668         }
1669 }
1670
1671 /**
1672  * xs_udp_set_buffer_size - set send and receive limits
1673  * @xprt: generic transport
1674  * @sndsize: requested size of send buffer, in bytes
1675  * @rcvsize: requested size of receive buffer, in bytes
1676  *
1677  * Set socket send and receive buffer size limits.
1678  */
1679 static void xs_udp_set_buffer_size(struct rpc_xprt *xprt, size_t sndsize, size_t rcvsize)
1680 {
1681         struct sock_xprt *transport = container_of(xprt, struct sock_xprt, xprt);
1682
1683         transport->sndsize = 0;
1684         if (sndsize)
1685                 transport->sndsize = sndsize + 1024;
1686         transport->rcvsize = 0;
1687         if (rcvsize)
1688                 transport->rcvsize = rcvsize + 1024;
1689
1690         xs_udp_do_set_buffer_size(xprt);
1691 }
1692
1693 /**
1694  * xs_udp_timer - called when a retransmit timeout occurs on a UDP transport
1695  * @task: task that timed out
1696  *
1697  * Adjust the congestion window after a retransmit timeout has occurred.
1698  */
1699 static void xs_udp_timer(struct rpc_xprt *xprt, struct rpc_task *task)
1700 {
1701         xprt_adjust_cwnd(xprt, task, -ETIMEDOUT);
1702 }
1703
1704 static unsigned short xs_get_random_port(void)
1705 {
1706         unsigned short range = xprt_max_resvport - xprt_min_resvport;
1707         unsigned short rand = (unsigned short) prandom_u32() % range;
1708         return rand + xprt_min_resvport;
1709 }
1710
1711 /**
1712  * xs_set_port - reset the port number in the remote endpoint address
1713  * @xprt: generic transport
1714  * @port: new port number
1715  *
1716  */
1717 static void xs_set_port(struct rpc_xprt *xprt, unsigned short port)
1718 {
1719         dprintk("RPC:       setting port for xprt %p to %u\n", xprt, port);
1720
1721         rpc_set_port(xs_addr(xprt), port);
1722         xs_update_peer_port(xprt);
1723 }
1724
1725 static unsigned short xs_get_srcport(struct sock_xprt *transport)
1726 {
1727         unsigned short port = transport->srcport;
1728
1729         if (port == 0 && transport->xprt.resvport)
1730                 port = xs_get_random_port();
1731         return port;
1732 }
1733
1734 static unsigned short xs_next_srcport(struct sock_xprt *transport, unsigned short port)
1735 {
1736         if (transport->srcport != 0)
1737                 transport->srcport = 0;
1738         if (!transport->xprt.resvport)
1739                 return 0;
1740         if (port <= xprt_min_resvport || port > xprt_max_resvport)
1741                 return xprt_max_resvport;
1742         return --port;
1743 }
1744 static int xs_bind(struct sock_xprt *transport, struct socket *sock)
1745 {
1746         struct sockaddr_storage myaddr;
1747         int err, nloop = 0;
1748         unsigned short port = xs_get_srcport(transport);
1749         unsigned short last;
1750
1751         /*
1752          * If we are asking for any ephemeral port (i.e. port == 0 &&
1753          * transport->xprt.resvport == 0), don't bind.  Let the local
1754          * port selection happen implicitly when the socket is used
1755          * (for example at connect time).
1756          *
1757          * This ensures that we can continue to establish TCP
1758          * connections even when all local ephemeral ports are already
1759          * a part of some TCP connection.  This makes no difference
1760          * for UDP sockets, but also doens't harm them.
1761          *
1762          * If we're asking for any reserved port (i.e. port == 0 &&
1763          * transport->xprt.resvport == 1) xs_get_srcport above will
1764          * ensure that port is non-zero and we will bind as needed.
1765          */
1766         if (port == 0)
1767                 return 0;
1768
1769         memcpy(&myaddr, &transport->srcaddr, transport->xprt.addrlen);
1770         do {
1771                 rpc_set_port((struct sockaddr *)&myaddr, port);
1772                 err = kernel_bind(sock, (struct sockaddr *)&myaddr,
1773                                 transport->xprt.addrlen);
1774                 if (err == 0) {
1775                         transport->srcport = port;
1776                         break;
1777                 }
1778                 last = port;
1779                 port = xs_next_srcport(transport, port);
1780                 if (port > last)
1781                         nloop++;
1782         } while (err == -EADDRINUSE && nloop != 2);
1783
1784         if (myaddr.ss_family == AF_INET)
1785                 dprintk("RPC:       %s %pI4:%u: %s (%d)\n", __func__,
1786                                 &((struct sockaddr_in *)&myaddr)->sin_addr,
1787                                 port, err ? "failed" : "ok", err);
1788         else
1789                 dprintk("RPC:       %s %pI6:%u: %s (%d)\n", __func__,
1790                                 &((struct sockaddr_in6 *)&myaddr)->sin6_addr,
1791                                 port, err ? "failed" : "ok", err);
1792         return err;
1793 }
1794
1795 /*
1796  * We don't support autobind on AF_LOCAL sockets
1797  */
1798 static void xs_local_rpcbind(struct rpc_task *task)
1799 {
1800         rcu_read_lock();
1801         xprt_set_bound(rcu_dereference(task->tk_client->cl_xprt));
1802         rcu_read_unlock();
1803 }
1804
1805 static void xs_local_set_port(struct rpc_xprt *xprt, unsigned short port)
1806 {
1807 }
1808
1809 #ifdef CONFIG_DEBUG_LOCK_ALLOC
1810 static struct lock_class_key xs_key[2];
1811 static struct lock_class_key xs_slock_key[2];
1812
1813 static inline void xs_reclassify_socketu(struct socket *sock)
1814 {
1815         struct sock *sk = sock->sk;
1816
1817         sock_lock_init_class_and_name(sk, "slock-AF_LOCAL-RPC",
1818                 &xs_slock_key[1], "sk_lock-AF_LOCAL-RPC", &xs_key[1]);
1819 }
1820
1821 static inline void xs_reclassify_socket4(struct socket *sock)
1822 {
1823         struct sock *sk = sock->sk;
1824
1825         sock_lock_init_class_and_name(sk, "slock-AF_INET-RPC",
1826                 &xs_slock_key[0], "sk_lock-AF_INET-RPC", &xs_key[0]);
1827 }
1828
1829 static inline void xs_reclassify_socket6(struct socket *sock)
1830 {
1831         struct sock *sk = sock->sk;
1832
1833         sock_lock_init_class_and_name(sk, "slock-AF_INET6-RPC",
1834                 &xs_slock_key[1], "sk_lock-AF_INET6-RPC", &xs_key[1]);
1835 }
1836
1837 static inline void xs_reclassify_socket(int family, struct socket *sock)
1838 {
1839         WARN_ON_ONCE(sock_owned_by_user(sock->sk));
1840         if (sock_owned_by_user(sock->sk))
1841                 return;
1842
1843         switch (family) {
1844         case AF_LOCAL:
1845                 xs_reclassify_socketu(sock);
1846                 break;
1847         case AF_INET:
1848                 xs_reclassify_socket4(sock);
1849                 break;
1850         case AF_INET6:
1851                 xs_reclassify_socket6(sock);
1852                 break;
1853         }
1854 }
1855 #else
1856 static inline void xs_reclassify_socketu(struct socket *sock)
1857 {
1858 }
1859
1860 static inline void xs_reclassify_socket4(struct socket *sock)
1861 {
1862 }
1863
1864 static inline void xs_reclassify_socket6(struct socket *sock)
1865 {
1866 }
1867
1868 static inline void xs_reclassify_socket(int family, struct socket *sock)
1869 {
1870 }
1871 #endif
1872
1873 static void xs_dummy_setup_socket(struct work_struct *work)
1874 {
1875 }
1876
1877 static struct socket *xs_create_sock(struct rpc_xprt *xprt,
1878                 struct sock_xprt *transport, int family, int type, int protocol)
1879 {
1880         struct socket *sock;
1881         int err;
1882
1883         err = __sock_create(xprt->xprt_net, family, type, protocol, &sock, 1);
1884         if (err < 0) {
1885                 dprintk("RPC:       can't create %d transport socket (%d).\n",
1886                                 protocol, -err);
1887                 goto out;
1888         }
1889         xs_reclassify_socket(family, sock);
1890
1891         err = xs_bind(transport, sock);
1892         if (err) {
1893                 sock_release(sock);
1894                 goto out;
1895         }
1896
1897         return sock;
1898 out:
1899         return ERR_PTR(err);
1900 }
1901
1902 static int xs_local_finish_connecting(struct rpc_xprt *xprt,
1903                                       struct socket *sock)
1904 {
1905         struct sock_xprt *transport = container_of(xprt, struct sock_xprt,
1906                                                                         xprt);
1907
1908         if (!transport->inet) {
1909                 struct sock *sk = sock->sk;
1910
1911                 write_lock_bh(&sk->sk_callback_lock);
1912
1913                 xs_save_old_callbacks(transport, sk);
1914
1915                 sk->sk_user_data = xprt;
1916                 sk->sk_data_ready = xs_local_data_ready;
1917                 sk->sk_write_space = xs_udp_write_space;
1918                 sk->sk_error_report = xs_error_report;
1919                 sk->sk_allocation = GFP_ATOMIC;
1920
1921                 xprt_clear_connected(xprt);
1922
1923                 /* Reset to new socket */
1924                 transport->sock = sock;
1925                 transport->inet = sk;
1926
1927                 write_unlock_bh(&sk->sk_callback_lock);
1928         }
1929
1930         /* Tell the socket layer to start connecting... */
1931         xprt->stat.connect_count++;
1932         xprt->stat.connect_start = jiffies;
1933         return kernel_connect(sock, xs_addr(xprt), xprt->addrlen, 0);
1934 }
1935
1936 /**
1937  * xs_local_setup_socket - create AF_LOCAL socket, connect to a local endpoint
1938  * @xprt: RPC transport to connect
1939  * @transport: socket transport to connect
1940  * @create_sock: function to create a socket of the correct type
1941  */
1942 static int xs_local_setup_socket(struct sock_xprt *transport)
1943 {
1944         struct rpc_xprt *xprt = &transport->xprt;
1945         struct socket *sock;
1946         int status = -EIO;
1947
1948         current->flags |= PF_FSTRANS;
1949
1950         clear_bit(XPRT_CONNECTION_ABORT, &xprt->state);
1951         status = __sock_create(xprt->xprt_net, AF_LOCAL,
1952                                         SOCK_STREAM, 0, &sock, 1);
1953         if (status < 0) {
1954                 dprintk("RPC:       can't create AF_LOCAL "
1955                         "transport socket (%d).\n", -status);
1956                 goto out;
1957         }
1958         xs_reclassify_socketu(sock);
1959
1960         dprintk("RPC:       worker connecting xprt %p via AF_LOCAL to %s\n",
1961                         xprt, xprt->address_strings[RPC_DISPLAY_ADDR]);
1962
1963         status = xs_local_finish_connecting(xprt, sock);
1964         trace_rpc_socket_connect(xprt, sock, status);
1965         switch (status) {
1966         case 0:
1967                 dprintk("RPC:       xprt %p connected to %s\n",
1968                                 xprt, xprt->address_strings[RPC_DISPLAY_ADDR]);
1969                 xprt_set_connected(xprt);
1970         case -ENOBUFS:
1971                 break;
1972         case -ENOENT:
1973                 dprintk("RPC:       xprt %p: socket %s does not exist\n",
1974                                 xprt, xprt->address_strings[RPC_DISPLAY_ADDR]);
1975                 break;
1976         case -ECONNREFUSED:
1977                 dprintk("RPC:       xprt %p: connection refused for %s\n",
1978                                 xprt, xprt->address_strings[RPC_DISPLAY_ADDR]);
1979                 break;
1980         default:
1981                 printk(KERN_ERR "%s: unhandled error (%d) connecting to %s\n",
1982                                 __func__, -status,
1983                                 xprt->address_strings[RPC_DISPLAY_ADDR]);
1984         }
1985
1986 out:
1987         xprt_clear_connecting(xprt);
1988         xprt_wake_pending_tasks(xprt, status);
1989         current->flags &= ~PF_FSTRANS;
1990         return status;
1991 }
1992
1993 static void xs_local_connect(struct rpc_xprt *xprt, struct rpc_task *task)
1994 {
1995         struct sock_xprt *transport = container_of(xprt, struct sock_xprt, xprt);
1996         int ret;
1997
1998          if (RPC_IS_ASYNC(task)) {
1999                 /*
2000                  * We want the AF_LOCAL connect to be resolved in the
2001                  * filesystem namespace of the process making the rpc
2002                  * call.  Thus we connect synchronously.
2003                  *
2004                  * If we want to support asynchronous AF_LOCAL calls,
2005                  * we'll need to figure out how to pass a namespace to
2006                  * connect.
2007                  */
2008                 rpc_exit(task, -ENOTCONN);
2009                 return;
2010         }
2011         ret = xs_local_setup_socket(transport);
2012         if (ret && !RPC_IS_SOFTCONN(task))
2013                 msleep_interruptible(15000);
2014 }
2015
2016 #ifdef CONFIG_SUNRPC_SWAP
2017 static void xs_set_memalloc(struct rpc_xprt *xprt)
2018 {
2019         struct sock_xprt *transport = container_of(xprt, struct sock_xprt,
2020                         xprt);
2021
2022         if (xprt->swapper)
2023                 sk_set_memalloc(transport->inet);
2024 }
2025
2026 /**
2027  * xs_swapper - Tag this transport as being used for swap.
2028  * @xprt: transport to tag
2029  * @enable: enable/disable
2030  *
2031  */
2032 int xs_swapper(struct rpc_xprt *xprt, int enable)
2033 {
2034         struct sock_xprt *transport = container_of(xprt, struct sock_xprt,
2035                         xprt);
2036         int err = 0;
2037
2038         if (enable) {
2039                 xprt->swapper++;
2040                 xs_set_memalloc(xprt);
2041         } else if (xprt->swapper) {
2042                 xprt->swapper--;
2043                 sk_clear_memalloc(transport->inet);
2044         }
2045
2046         return err;
2047 }
2048 EXPORT_SYMBOL_GPL(xs_swapper);
2049 #else
2050 static void xs_set_memalloc(struct rpc_xprt *xprt)
2051 {
2052 }
2053 #endif
2054
2055 static void xs_udp_finish_connecting(struct rpc_xprt *xprt, struct socket *sock)
2056 {
2057         struct sock_xprt *transport = container_of(xprt, struct sock_xprt, xprt);
2058
2059         if (!transport->inet) {
2060                 struct sock *sk = sock->sk;
2061
2062                 write_lock_bh(&sk->sk_callback_lock);
2063
2064                 xs_save_old_callbacks(transport, sk);
2065
2066                 sk->sk_user_data = xprt;
2067                 sk->sk_data_ready = xs_udp_data_ready;
2068                 sk->sk_write_space = xs_udp_write_space;
2069                 sk->sk_allocation = GFP_ATOMIC;
2070
2071                 xprt_set_connected(xprt);
2072
2073                 /* Reset to new socket */
2074                 transport->sock = sock;
2075                 transport->inet = sk;
2076
2077                 xs_set_memalloc(xprt);
2078
2079                 write_unlock_bh(&sk->sk_callback_lock);
2080         }
2081         xs_udp_do_set_buffer_size(xprt);
2082 }
2083
2084 static void xs_udp_setup_socket(struct work_struct *work)
2085 {
2086         struct sock_xprt *transport =
2087                 container_of(work, struct sock_xprt, connect_worker.work);
2088         struct rpc_xprt *xprt = &transport->xprt;
2089         struct socket *sock = transport->sock;
2090         int status = -EIO;
2091
2092         current->flags |= PF_FSTRANS;
2093
2094         /* Start by resetting any existing state */
2095         xs_reset_transport(transport);
2096         sock = xs_create_sock(xprt, transport,
2097                         xs_addr(xprt)->sa_family, SOCK_DGRAM, IPPROTO_UDP);
2098         if (IS_ERR(sock))
2099                 goto out;
2100
2101         dprintk("RPC:       worker connecting xprt %p via %s to "
2102                                 "%s (port %s)\n", xprt,
2103                         xprt->address_strings[RPC_DISPLAY_PROTO],
2104                         xprt->address_strings[RPC_DISPLAY_ADDR],
2105                         xprt->address_strings[RPC_DISPLAY_PORT]);
2106
2107         xs_udp_finish_connecting(xprt, sock);
2108         trace_rpc_socket_connect(xprt, sock, 0);
2109         status = 0;
2110 out:
2111         xprt_clear_connecting(xprt);
2112         xprt_wake_pending_tasks(xprt, status);
2113         current->flags &= ~PF_FSTRANS;
2114 }
2115
2116 /*
2117  * We need to preserve the port number so the reply cache on the server can
2118  * find our cached RPC replies when we get around to reconnecting.
2119  */
2120 static void xs_abort_connection(struct sock_xprt *transport)
2121 {
2122         int result;
2123         struct sockaddr any;
2124
2125         dprintk("RPC:       disconnecting xprt %p to reuse port\n", transport);
2126
2127         /*
2128          * Disconnect the transport socket by doing a connect operation
2129          * with AF_UNSPEC.  This should return immediately...
2130          */
2131         memset(&any, 0, sizeof(any));
2132         any.sa_family = AF_UNSPEC;
2133         result = kernel_connect(transport->sock, &any, sizeof(any), 0);
2134         trace_rpc_socket_reset_connection(&transport->xprt,
2135                         transport->sock, result);
2136         if (!result)
2137                 xs_sock_reset_connection_flags(&transport->xprt);
2138         dprintk("RPC:       AF_UNSPEC connect return code %d\n", result);
2139 }
2140
2141 static void xs_tcp_reuse_connection(struct sock_xprt *transport)
2142 {
2143         unsigned int state = transport->inet->sk_state;
2144
2145         if (state == TCP_CLOSE && transport->sock->state == SS_UNCONNECTED) {
2146                 /* we don't need to abort the connection if the socket
2147                  * hasn't undergone a shutdown
2148                  */
2149                 if (transport->inet->sk_shutdown == 0)
2150                         return;
2151                 dprintk("RPC:       %s: TCP_CLOSEd and sk_shutdown set to %d\n",
2152                                 __func__, transport->inet->sk_shutdown);
2153         }
2154         if ((1 << state) & (TCPF_ESTABLISHED|TCPF_SYN_SENT)) {
2155                 /* we don't need to abort the connection if the socket
2156                  * hasn't undergone a shutdown
2157                  */
2158                 if (transport->inet->sk_shutdown == 0)
2159                         return;
2160                 dprintk("RPC:       %s: ESTABLISHED/SYN_SENT "
2161                                 "sk_shutdown set to %d\n",
2162                                 __func__, transport->inet->sk_shutdown);
2163         }
2164         xs_abort_connection(transport);
2165 }
2166
2167 static int xs_tcp_finish_connecting(struct rpc_xprt *xprt, struct socket *sock)
2168 {
2169         struct sock_xprt *transport = container_of(xprt, struct sock_xprt, xprt);
2170         int ret = -ENOTCONN;
2171
2172         if (!transport->inet) {
2173                 struct sock *sk = sock->sk;
2174                 unsigned int keepidle = xprt->timeout->to_initval / HZ;
2175                 unsigned int keepcnt = xprt->timeout->to_retries + 1;
2176                 unsigned int opt_on = 1;
2177
2178                 /* TCP Keepalive options */
2179                 kernel_setsockopt(sock, SOL_SOCKET, SO_KEEPALIVE,
2180                                 (char *)&opt_on, sizeof(opt_on));
2181                 kernel_setsockopt(sock, SOL_TCP, TCP_KEEPIDLE,
2182                                 (char *)&keepidle, sizeof(keepidle));
2183                 kernel_setsockopt(sock, SOL_TCP, TCP_KEEPINTVL,
2184                                 (char *)&keepidle, sizeof(keepidle));
2185                 kernel_setsockopt(sock, SOL_TCP, TCP_KEEPCNT,
2186                                 (char *)&keepcnt, sizeof(keepcnt));
2187
2188                 write_lock_bh(&sk->sk_callback_lock);
2189
2190                 xs_save_old_callbacks(transport, sk);
2191
2192                 sk->sk_user_data = xprt;
2193                 sk->sk_data_ready = xs_tcp_data_ready;
2194                 sk->sk_state_change = xs_tcp_state_change;
2195                 sk->sk_write_space = xs_tcp_write_space;
2196                 sk->sk_error_report = xs_error_report;
2197                 sk->sk_allocation = GFP_ATOMIC;
2198
2199                 /* socket options */
2200                 sk->sk_userlocks |= SOCK_BINDPORT_LOCK;
2201                 sock_reset_flag(sk, SOCK_LINGER);
2202                 tcp_sk(sk)->linger2 = 0;
2203                 tcp_sk(sk)->nonagle |= TCP_NAGLE_OFF;
2204
2205                 xprt_clear_connected(xprt);
2206
2207                 /* Reset to new socket */
2208                 transport->sock = sock;
2209                 transport->inet = sk;
2210
2211                 write_unlock_bh(&sk->sk_callback_lock);
2212         }
2213
2214         if (!xprt_bound(xprt))
2215                 goto out;
2216
2217         xs_set_memalloc(xprt);
2218
2219         /* Tell the socket layer to start connecting... */
2220         xprt->stat.connect_count++;
2221         xprt->stat.connect_start = jiffies;
2222         ret = kernel_connect(sock, xs_addr(xprt), xprt->addrlen, O_NONBLOCK);
2223         switch (ret) {
2224         case 0:
2225         case -EINPROGRESS:
2226                 /* SYN_SENT! */
2227                 if (xprt->reestablish_timeout < XS_TCP_INIT_REEST_TO)
2228                         xprt->reestablish_timeout = XS_TCP_INIT_REEST_TO;
2229         }
2230 out:
2231         return ret;
2232 }
2233
2234 /**
2235  * xs_tcp_setup_socket - create a TCP socket and connect to a remote endpoint
2236  * @xprt: RPC transport to connect
2237  * @transport: socket transport to connect
2238  * @create_sock: function to create a socket of the correct type
2239  *
2240  * Invoked by a work queue tasklet.
2241  */
2242 static void xs_tcp_setup_socket(struct work_struct *work)
2243 {
2244         struct sock_xprt *transport =
2245                 container_of(work, struct sock_xprt, connect_worker.work);
2246         struct socket *sock = transport->sock;
2247         struct rpc_xprt *xprt = &transport->xprt;
2248         int status = -EIO;
2249
2250         current->flags |= PF_FSTRANS;
2251
2252         if (!sock) {
2253                 clear_bit(XPRT_CONNECTION_ABORT, &xprt->state);
2254                 sock = xs_create_sock(xprt, transport,
2255                                 xs_addr(xprt)->sa_family, SOCK_STREAM, IPPROTO_TCP);
2256                 if (IS_ERR(sock)) {
2257                         status = PTR_ERR(sock);
2258                         goto out;
2259                 }
2260         } else {
2261                 int abort_and_exit;
2262
2263                 abort_and_exit = test_and_clear_bit(XPRT_CONNECTION_ABORT,
2264                                 &xprt->state);
2265                 /* "close" the socket, preserving the local port */
2266                 set_bit(XPRT_CONNECTION_REUSE, &xprt->state);
2267                 xs_tcp_reuse_connection(transport);
2268                 clear_bit(XPRT_CONNECTION_REUSE, &xprt->state);
2269
2270                 if (abort_and_exit)
2271                         goto out_eagain;
2272         }
2273
2274         dprintk("RPC:       worker connecting xprt %p via %s to "
2275                                 "%s (port %s)\n", xprt,
2276                         xprt->address_strings[RPC_DISPLAY_PROTO],
2277                         xprt->address_strings[RPC_DISPLAY_ADDR],
2278                         xprt->address_strings[RPC_DISPLAY_PORT]);
2279
2280         status = xs_tcp_finish_connecting(xprt, sock);
2281         trace_rpc_socket_connect(xprt, sock, status);
2282         dprintk("RPC:       %p connect status %d connected %d sock state %d\n",
2283                         xprt, -status, xprt_connected(xprt),
2284                         sock->sk->sk_state);
2285         switch (status) {
2286         default:
2287                 printk("%s: connect returned unhandled error %d\n",
2288                         __func__, status);
2289         case -EADDRNOTAVAIL:
2290                 /* We're probably in TIME_WAIT. Get rid of existing socket,
2291                  * and retry
2292                  */
2293                 xs_tcp_force_close(xprt);
2294                 break;
2295         case 0:
2296         case -EINPROGRESS:
2297         case -EALREADY:
2298                 xprt_clear_connecting(xprt);
2299                 current->flags &= ~PF_FSTRANS;
2300                 return;
2301         case -EINVAL:
2302                 /* Happens, for instance, if the user specified a link
2303                  * local IPv6 address without a scope-id.
2304                  */
2305         case -ECONNREFUSED:
2306         case -ECONNRESET:
2307         case -ENETUNREACH:
2308         case -ENOBUFS:
2309                 /* retry with existing socket, after a delay */
2310                 goto out;
2311         }
2312 out_eagain:
2313         status = -EAGAIN;
2314 out:
2315         xprt_clear_connecting(xprt);
2316         xprt_wake_pending_tasks(xprt, status);
2317         current->flags &= ~PF_FSTRANS;
2318 }
2319
2320 /**
2321  * xs_connect - connect a socket to a remote endpoint
2322  * @xprt: pointer to transport structure
2323  * @task: address of RPC task that manages state of connect request
2324  *
2325  * TCP: If the remote end dropped the connection, delay reconnecting.
2326  *
2327  * UDP socket connects are synchronous, but we use a work queue anyway
2328  * to guarantee that even unprivileged user processes can set up a
2329  * socket on a privileged port.
2330  *
2331  * If a UDP socket connect fails, the delay behavior here prevents
2332  * retry floods (hard mounts).
2333  */
2334 static void xs_connect(struct rpc_xprt *xprt, struct rpc_task *task)
2335 {
2336         struct sock_xprt *transport = container_of(xprt, struct sock_xprt, xprt);
2337
2338         if (transport->sock != NULL && !RPC_IS_SOFTCONN(task)) {
2339                 dprintk("RPC:       xs_connect delayed xprt %p for %lu "
2340                                 "seconds\n",
2341                                 xprt, xprt->reestablish_timeout / HZ);
2342                 queue_delayed_work(rpciod_workqueue,
2343                                    &transport->connect_worker,
2344                                    xprt->reestablish_timeout);
2345                 xprt->reestablish_timeout <<= 1;
2346                 if (xprt->reestablish_timeout < XS_TCP_INIT_REEST_TO)
2347                         xprt->reestablish_timeout = XS_TCP_INIT_REEST_TO;
2348                 if (xprt->reestablish_timeout > XS_TCP_MAX_REEST_TO)
2349                         xprt->reestablish_timeout = XS_TCP_MAX_REEST_TO;
2350         } else {
2351                 dprintk("RPC:       xs_connect scheduled xprt %p\n", xprt);
2352                 queue_delayed_work(rpciod_workqueue,
2353                                    &transport->connect_worker, 0);
2354         }
2355 }
2356
2357 /**
2358  * xs_local_print_stats - display AF_LOCAL socket-specifc stats
2359  * @xprt: rpc_xprt struct containing statistics
2360  * @seq: output file
2361  *
2362  */
2363 static void xs_local_print_stats(struct rpc_xprt *xprt, struct seq_file *seq)
2364 {
2365         long idle_time = 0;
2366
2367         if (xprt_connected(xprt))
2368                 idle_time = (long)(jiffies - xprt->last_used) / HZ;
2369
2370         seq_printf(seq, "\txprt:\tlocal %lu %lu %lu %ld %lu %lu %lu "
2371                         "%llu %llu %lu %llu %llu\n",
2372                         xprt->stat.bind_count,
2373                         xprt->stat.connect_count,
2374                         xprt->stat.connect_time,
2375                         idle_time,
2376                         xprt->stat.sends,
2377                         xprt->stat.recvs,
2378                         xprt->stat.bad_xids,
2379                         xprt->stat.req_u,
2380                         xprt->stat.bklog_u,
2381                         xprt->stat.max_slots,
2382                         xprt->stat.sending_u,
2383                         xprt->stat.pending_u);
2384 }
2385
2386 /**
2387  * xs_udp_print_stats - display UDP socket-specifc stats
2388  * @xprt: rpc_xprt struct containing statistics
2389  * @seq: output file
2390  *
2391  */
2392 static void xs_udp_print_stats(struct rpc_xprt *xprt, struct seq_file *seq)
2393 {
2394         struct sock_xprt *transport = container_of(xprt, struct sock_xprt, xprt);
2395
2396         seq_printf(seq, "\txprt:\tudp %u %lu %lu %lu %lu %llu %llu "
2397                         "%lu %llu %llu\n",
2398                         transport->srcport,
2399                         xprt->stat.bind_count,
2400                         xprt->stat.sends,
2401                         xprt->stat.recvs,
2402                         xprt->stat.bad_xids,
2403                         xprt->stat.req_u,
2404                         xprt->stat.bklog_u,
2405                         xprt->stat.max_slots,
2406                         xprt->stat.sending_u,
2407                         xprt->stat.pending_u);
2408 }
2409
2410 /**
2411  * xs_tcp_print_stats - display TCP socket-specifc stats
2412  * @xprt: rpc_xprt struct containing statistics
2413  * @seq: output file
2414  *
2415  */
2416 static void xs_tcp_print_stats(struct rpc_xprt *xprt, struct seq_file *seq)
2417 {
2418         struct sock_xprt *transport = container_of(xprt, struct sock_xprt, xprt);
2419         long idle_time = 0;
2420
2421         if (xprt_connected(xprt))
2422                 idle_time = (long)(jiffies - xprt->last_used) / HZ;
2423
2424         seq_printf(seq, "\txprt:\ttcp %u %lu %lu %lu %ld %lu %lu %lu "
2425                         "%llu %llu %lu %llu %llu\n",
2426                         transport->srcport,
2427                         xprt->stat.bind_count,
2428                         xprt->stat.connect_count,
2429                         xprt->stat.connect_time,
2430                         idle_time,
2431                         xprt->stat.sends,
2432                         xprt->stat.recvs,
2433                         xprt->stat.bad_xids,
2434                         xprt->stat.req_u,
2435                         xprt->stat.bklog_u,
2436                         xprt->stat.max_slots,
2437                         xprt->stat.sending_u,
2438                         xprt->stat.pending_u);
2439 }
2440
2441 /*
2442  * Allocate a bunch of pages for a scratch buffer for the rpc code. The reason
2443  * we allocate pages instead doing a kmalloc like rpc_malloc is because we want
2444  * to use the server side send routines.
2445  */
2446 static void *bc_malloc(struct rpc_task *task, size_t size)
2447 {
2448         struct page *page;
2449         struct rpc_buffer *buf;
2450
2451         WARN_ON_ONCE(size > PAGE_SIZE - sizeof(struct rpc_buffer));
2452         if (size > PAGE_SIZE - sizeof(struct rpc_buffer))
2453                 return NULL;
2454
2455         page = alloc_page(GFP_KERNEL);
2456         if (!page)
2457                 return NULL;
2458
2459         buf = page_address(page);
2460         buf->len = PAGE_SIZE;
2461
2462         return buf->data;
2463 }
2464
2465 /*
2466  * Free the space allocated in the bc_alloc routine
2467  */
2468 static void bc_free(void *buffer)
2469 {
2470         struct rpc_buffer *buf;
2471
2472         if (!buffer)
2473                 return;
2474
2475         buf = container_of(buffer, struct rpc_buffer, data);
2476         free_page((unsigned long)buf);
2477 }
2478
2479 /*
2480  * Use the svc_sock to send the callback. Must be called with svsk->sk_mutex
2481  * held. Borrows heavily from svc_tcp_sendto and xs_tcp_send_request.
2482  */
2483 static int bc_sendto(struct rpc_rqst *req)
2484 {
2485         int len;
2486         struct xdr_buf *xbufp = &req->rq_snd_buf;
2487         struct rpc_xprt *xprt = req->rq_xprt;
2488         struct sock_xprt *transport =
2489                                 container_of(xprt, struct sock_xprt, xprt);
2490         struct socket *sock = transport->sock;
2491         unsigned long headoff;
2492         unsigned long tailoff;
2493
2494         xs_encode_stream_record_marker(xbufp);
2495
2496         tailoff = (unsigned long)xbufp->tail[0].iov_base & ~PAGE_MASK;
2497         headoff = (unsigned long)xbufp->head[0].iov_base & ~PAGE_MASK;
2498         len = svc_send_common(sock, xbufp,
2499                               virt_to_page(xbufp->head[0].iov_base), headoff,
2500                               xbufp->tail[0].iov_base, tailoff);
2501
2502         if (len != xbufp->len) {
2503                 printk(KERN_NOTICE "Error sending entire callback!\n");
2504                 len = -EAGAIN;
2505         }
2506
2507         return len;
2508 }
2509
2510 /*
2511  * The send routine. Borrows from svc_send
2512  */
2513 static int bc_send_request(struct rpc_task *task)
2514 {
2515         struct rpc_rqst *req = task->tk_rqstp;
2516         struct svc_xprt *xprt;
2517         u32                     len;
2518
2519         dprintk("sending request with xid: %08x\n", ntohl(req->rq_xid));
2520         /*
2521          * Get the server socket associated with this callback xprt
2522          */
2523         xprt = req->rq_xprt->bc_xprt;
2524
2525         /*
2526          * Grab the mutex to serialize data as the connection is shared
2527          * with the fore channel
2528          */
2529         if (!mutex_trylock(&xprt->xpt_mutex)) {
2530                 rpc_sleep_on(&xprt->xpt_bc_pending, task, NULL);
2531                 if (!mutex_trylock(&xprt->xpt_mutex))
2532                         return -EAGAIN;
2533                 rpc_wake_up_queued_task(&xprt->xpt_bc_pending, task);
2534         }
2535         if (test_bit(XPT_DEAD, &xprt->xpt_flags))
2536                 len = -ENOTCONN;
2537         else
2538                 len = bc_sendto(req);
2539         mutex_unlock(&xprt->xpt_mutex);
2540
2541         if (len > 0)
2542                 len = 0;
2543
2544         return len;
2545 }
2546
2547 /*
2548  * The close routine. Since this is client initiated, we do nothing
2549  */
2550
2551 static void bc_close(struct rpc_xprt *xprt)
2552 {
2553 }
2554
2555 /*
2556  * The xprt destroy routine. Again, because this connection is client
2557  * initiated, we do nothing
2558  */
2559
2560 static void bc_destroy(struct rpc_xprt *xprt)
2561 {
2562         dprintk("RPC:       bc_destroy xprt %p\n", xprt);
2563
2564         xs_xprt_free(xprt);
2565         module_put(THIS_MODULE);
2566 }
2567
2568 static struct rpc_xprt_ops xs_local_ops = {
2569         .reserve_xprt           = xprt_reserve_xprt,
2570         .release_xprt           = xs_tcp_release_xprt,
2571         .alloc_slot             = xprt_alloc_slot,
2572         .rpcbind                = xs_local_rpcbind,
2573         .set_port               = xs_local_set_port,
2574         .connect                = xs_local_connect,
2575         .buf_alloc              = rpc_malloc,
2576         .buf_free               = rpc_free,
2577         .send_request           = xs_local_send_request,
2578         .set_retrans_timeout    = xprt_set_retrans_timeout_def,
2579         .close                  = xs_close,
2580         .destroy                = xs_destroy,
2581         .print_stats            = xs_local_print_stats,
2582 };
2583
2584 static struct rpc_xprt_ops xs_udp_ops = {
2585         .set_buffer_size        = xs_udp_set_buffer_size,
2586         .reserve_xprt           = xprt_reserve_xprt_cong,
2587         .release_xprt           = xprt_release_xprt_cong,
2588         .alloc_slot             = xprt_alloc_slot,
2589         .rpcbind                = rpcb_getport_async,
2590         .set_port               = xs_set_port,
2591         .connect                = xs_connect,
2592         .buf_alloc              = rpc_malloc,
2593         .buf_free               = rpc_free,
2594         .send_request           = xs_udp_send_request,
2595         .set_retrans_timeout    = xprt_set_retrans_timeout_rtt,
2596         .timer                  = xs_udp_timer,
2597         .release_request        = xprt_release_rqst_cong,
2598         .close                  = xs_close,
2599         .destroy                = xs_destroy,
2600         .print_stats            = xs_udp_print_stats,
2601 };
2602
2603 static struct rpc_xprt_ops xs_tcp_ops = {
2604         .reserve_xprt           = xprt_reserve_xprt,
2605         .release_xprt           = xs_tcp_release_xprt,
2606         .alloc_slot             = xprt_lock_and_alloc_slot,
2607         .rpcbind                = rpcb_getport_async,
2608         .set_port               = xs_set_port,
2609         .connect                = xs_connect,
2610         .buf_alloc              = rpc_malloc,
2611         .buf_free               = rpc_free,
2612         .send_request           = xs_tcp_send_request,
2613         .set_retrans_timeout    = xprt_set_retrans_timeout_def,
2614         .close                  = xs_tcp_close,
2615         .destroy                = xs_destroy,
2616         .print_stats            = xs_tcp_print_stats,
2617 };
2618
2619 /*
2620  * The rpc_xprt_ops for the server backchannel
2621  */
2622
2623 static struct rpc_xprt_ops bc_tcp_ops = {
2624         .reserve_xprt           = xprt_reserve_xprt,
2625         .release_xprt           = xprt_release_xprt,
2626         .alloc_slot             = xprt_alloc_slot,
2627         .buf_alloc              = bc_malloc,
2628         .buf_free               = bc_free,
2629         .send_request           = bc_send_request,
2630         .set_retrans_timeout    = xprt_set_retrans_timeout_def,
2631         .close                  = bc_close,
2632         .destroy                = bc_destroy,
2633         .print_stats            = xs_tcp_print_stats,
2634 };
2635
2636 static int xs_init_anyaddr(const int family, struct sockaddr *sap)
2637 {
2638         static const struct sockaddr_in sin = {
2639                 .sin_family             = AF_INET,
2640                 .sin_addr.s_addr        = htonl(INADDR_ANY),
2641         };
2642         static const struct sockaddr_in6 sin6 = {
2643                 .sin6_family            = AF_INET6,
2644                 .sin6_addr              = IN6ADDR_ANY_INIT,
2645         };
2646
2647         switch (family) {
2648         case AF_LOCAL:
2649                 break;
2650         case AF_INET:
2651                 memcpy(sap, &sin, sizeof(sin));
2652                 break;
2653         case AF_INET6:
2654                 memcpy(sap, &sin6, sizeof(sin6));
2655                 break;
2656         default:
2657                 dprintk("RPC:       %s: Bad address family\n", __func__);
2658                 return -EAFNOSUPPORT;
2659         }
2660         return 0;
2661 }
2662
2663 static struct rpc_xprt *xs_setup_xprt(struct xprt_create *args,
2664                                       unsigned int slot_table_size,
2665                                       unsigned int max_slot_table_size)
2666 {
2667         struct rpc_xprt *xprt;
2668         struct sock_xprt *new;
2669
2670         if (args->addrlen > sizeof(xprt->addr)) {
2671                 dprintk("RPC:       xs_setup_xprt: address too large\n");
2672                 return ERR_PTR(-EBADF);
2673         }
2674
2675         xprt = xprt_alloc(args->net, sizeof(*new), slot_table_size,
2676                         max_slot_table_size);
2677         if (xprt == NULL) {
2678                 dprintk("RPC:       xs_setup_xprt: couldn't allocate "
2679                                 "rpc_xprt\n");
2680                 return ERR_PTR(-ENOMEM);
2681         }
2682
2683         new = container_of(xprt, struct sock_xprt, xprt);
2684         memcpy(&xprt->addr, args->dstaddr, args->addrlen);
2685         xprt->addrlen = args->addrlen;
2686         if (args->srcaddr)
2687                 memcpy(&new->srcaddr, args->srcaddr, args->addrlen);
2688         else {
2689                 int err;
2690                 err = xs_init_anyaddr(args->dstaddr->sa_family,
2691                                         (struct sockaddr *)&new->srcaddr);
2692                 if (err != 0) {
2693                         xprt_free(xprt);
2694                         return ERR_PTR(err);
2695                 }
2696         }
2697
2698         return xprt;
2699 }
2700
2701 static const struct rpc_timeout xs_local_default_timeout = {
2702         .to_initval = 10 * HZ,
2703         .to_maxval = 10 * HZ,
2704         .to_retries = 2,
2705 };
2706
2707 /**
2708  * xs_setup_local - Set up transport to use an AF_LOCAL socket
2709  * @args: rpc transport creation arguments
2710  *
2711  * AF_LOCAL is a "tpi_cots_ord" transport, just like TCP
2712  */
2713 static struct rpc_xprt *xs_setup_local(struct xprt_create *args)
2714 {
2715         struct sockaddr_un *sun = (struct sockaddr_un *)args->dstaddr;
2716         struct sock_xprt *transport;
2717         struct rpc_xprt *xprt;
2718         struct rpc_xprt *ret;
2719
2720         xprt = xs_setup_xprt(args, xprt_tcp_slot_table_entries,
2721                         xprt_max_tcp_slot_table_entries);
2722         if (IS_ERR(xprt))
2723                 return xprt;
2724         transport = container_of(xprt, struct sock_xprt, xprt);
2725
2726         xprt->prot = 0;
2727         xprt->tsh_size = sizeof(rpc_fraghdr) / sizeof(u32);
2728         xprt->max_payload = RPC_MAX_FRAGMENT_SIZE;
2729
2730         xprt->bind_timeout = XS_BIND_TO;
2731         xprt->reestablish_timeout = XS_TCP_INIT_REEST_TO;
2732         xprt->idle_timeout = XS_IDLE_DISC_TO;
2733
2734         xprt->ops = &xs_local_ops;
2735         xprt->timeout = &xs_local_default_timeout;
2736
2737         INIT_DELAYED_WORK(&transport->connect_worker,
2738                         xs_dummy_setup_socket);
2739
2740         switch (sun->sun_family) {
2741         case AF_LOCAL:
2742                 if (sun->sun_path[0] != '/') {
2743                         dprintk("RPC:       bad AF_LOCAL address: %s\n",
2744                                         sun->sun_path);
2745                         ret = ERR_PTR(-EINVAL);
2746                         goto out_err;
2747                 }
2748                 xprt_set_bound(xprt);
2749                 xs_format_peer_addresses(xprt, "local", RPCBIND_NETID_LOCAL);
2750                 ret = ERR_PTR(xs_local_setup_socket(transport));
2751                 if (ret)
2752                         goto out_err;
2753                 break;
2754         default:
2755                 ret = ERR_PTR(-EAFNOSUPPORT);
2756                 goto out_err;
2757         }
2758
2759         dprintk("RPC:       set up xprt to %s via AF_LOCAL\n",
2760                         xprt->address_strings[RPC_DISPLAY_ADDR]);
2761
2762         if (try_module_get(THIS_MODULE))
2763                 return xprt;
2764         ret = ERR_PTR(-EINVAL);
2765 out_err:
2766         xs_xprt_free(xprt);
2767         return ret;
2768 }
2769
2770 static const struct rpc_timeout xs_udp_default_timeout = {
2771         .to_initval = 5 * HZ,
2772         .to_maxval = 30 * HZ,
2773         .to_increment = 5 * HZ,
2774         .to_retries = 5,
2775 };
2776
2777 /**
2778  * xs_setup_udp - Set up transport to use a UDP socket
2779  * @args: rpc transport creation arguments
2780  *
2781  */
2782 static struct rpc_xprt *xs_setup_udp(struct xprt_create *args)
2783 {
2784         struct sockaddr *addr = args->dstaddr;
2785         struct rpc_xprt *xprt;
2786         struct sock_xprt *transport;
2787         struct rpc_xprt *ret;
2788
2789         xprt = xs_setup_xprt(args, xprt_udp_slot_table_entries,
2790                         xprt_udp_slot_table_entries);
2791         if (IS_ERR(xprt))
2792                 return xprt;
2793         transport = container_of(xprt, struct sock_xprt, xprt);
2794
2795         xprt->prot = IPPROTO_UDP;
2796         xprt->tsh_size = 0;
2797         /* XXX: header size can vary due to auth type, IPv6, etc. */
2798         xprt->max_payload = (1U << 16) - (MAX_HEADER << 3);
2799
2800         xprt->bind_timeout = XS_BIND_TO;
2801         xprt->reestablish_timeout = XS_UDP_REEST_TO;
2802         xprt->idle_timeout = XS_IDLE_DISC_TO;
2803
2804         xprt->ops = &xs_udp_ops;
2805
2806         xprt->timeout = &xs_udp_default_timeout;
2807
2808         switch (addr->sa_family) {
2809         case AF_INET:
2810                 if (((struct sockaddr_in *)addr)->sin_port != htons(0))
2811                         xprt_set_bound(xprt);
2812
2813                 INIT_DELAYED_WORK(&transport->connect_worker,
2814                                         xs_udp_setup_socket);
2815                 xs_format_peer_addresses(xprt, "udp", RPCBIND_NETID_UDP);
2816                 break;
2817         case AF_INET6:
2818                 if (((struct sockaddr_in6 *)addr)->sin6_port != htons(0))
2819                         xprt_set_bound(xprt);
2820
2821                 INIT_DELAYED_WORK(&transport->connect_worker,
2822                                         xs_udp_setup_socket);
2823                 xs_format_peer_addresses(xprt, "udp", RPCBIND_NETID_UDP6);
2824                 break;
2825         default:
2826                 ret = ERR_PTR(-EAFNOSUPPORT);
2827                 goto out_err;
2828         }
2829
2830         if (xprt_bound(xprt))
2831                 dprintk("RPC:       set up xprt to %s (port %s) via %s\n",
2832                                 xprt->address_strings[RPC_DISPLAY_ADDR],
2833                                 xprt->address_strings[RPC_DISPLAY_PORT],
2834                                 xprt->address_strings[RPC_DISPLAY_PROTO]);
2835         else
2836                 dprintk("RPC:       set up xprt to %s (autobind) via %s\n",
2837                                 xprt->address_strings[RPC_DISPLAY_ADDR],
2838                                 xprt->address_strings[RPC_DISPLAY_PROTO]);
2839
2840         if (try_module_get(THIS_MODULE))
2841                 return xprt;
2842         ret = ERR_PTR(-EINVAL);
2843 out_err:
2844         xs_xprt_free(xprt);
2845         return ret;
2846 }
2847
2848 static const struct rpc_timeout xs_tcp_default_timeout = {
2849         .to_initval = 60 * HZ,
2850         .to_maxval = 60 * HZ,
2851         .to_retries = 2,
2852 };
2853
2854 /**
2855  * xs_setup_tcp - Set up transport to use a TCP socket
2856  * @args: rpc transport creation arguments
2857  *
2858  */
2859 static struct rpc_xprt *xs_setup_tcp(struct xprt_create *args)
2860 {
2861         struct sockaddr *addr = args->dstaddr;
2862         struct rpc_xprt *xprt;
2863         struct sock_xprt *transport;
2864         struct rpc_xprt *ret;
2865         unsigned int max_slot_table_size = xprt_max_tcp_slot_table_entries;
2866
2867         if (args->flags & XPRT_CREATE_INFINITE_SLOTS)
2868                 max_slot_table_size = RPC_MAX_SLOT_TABLE_LIMIT;
2869
2870         xprt = xs_setup_xprt(args, xprt_tcp_slot_table_entries,
2871                         max_slot_table_size);
2872         if (IS_ERR(xprt))
2873                 return xprt;
2874         transport = container_of(xprt, struct sock_xprt, xprt);
2875
2876         xprt->prot = IPPROTO_TCP;
2877         xprt->tsh_size = sizeof(rpc_fraghdr) / sizeof(u32);
2878         xprt->max_payload = RPC_MAX_FRAGMENT_SIZE;
2879
2880         xprt->bind_timeout = XS_BIND_TO;
2881         xprt->reestablish_timeout = XS_TCP_INIT_REEST_TO;
2882         xprt->idle_timeout = XS_IDLE_DISC_TO;
2883
2884         xprt->ops = &xs_tcp_ops;
2885         xprt->timeout = &xs_tcp_default_timeout;
2886
2887         switch (addr->sa_family) {
2888         case AF_INET:
2889                 if (((struct sockaddr_in *)addr)->sin_port != htons(0))
2890                         xprt_set_bound(xprt);
2891
2892                 INIT_DELAYED_WORK(&transport->connect_worker,
2893                                         xs_tcp_setup_socket);
2894                 xs_format_peer_addresses(xprt, "tcp", RPCBIND_NETID_TCP);
2895                 break;
2896         case AF_INET6:
2897                 if (((struct sockaddr_in6 *)addr)->sin6_port != htons(0))
2898                         xprt_set_bound(xprt);
2899
2900                 INIT_DELAYED_WORK(&transport->connect_worker,
2901                                         xs_tcp_setup_socket);
2902                 xs_format_peer_addresses(xprt, "tcp", RPCBIND_NETID_TCP6);
2903                 break;
2904         default:
2905                 ret = ERR_PTR(-EAFNOSUPPORT);
2906                 goto out_err;
2907         }
2908
2909         if (xprt_bound(xprt))
2910                 dprintk("RPC:       set up xprt to %s (port %s) via %s\n",
2911                                 xprt->address_strings[RPC_DISPLAY_ADDR],
2912                                 xprt->address_strings[RPC_DISPLAY_PORT],
2913                                 xprt->address_strings[RPC_DISPLAY_PROTO]);
2914         else
2915                 dprintk("RPC:       set up xprt to %s (autobind) via %s\n",
2916                                 xprt->address_strings[RPC_DISPLAY_ADDR],
2917                                 xprt->address_strings[RPC_DISPLAY_PROTO]);
2918
2919         if (try_module_get(THIS_MODULE))
2920                 return xprt;
2921         ret = ERR_PTR(-EINVAL);
2922 out_err:
2923         xs_xprt_free(xprt);
2924         return ret;
2925 }
2926
2927 /**
2928  * xs_setup_bc_tcp - Set up transport to use a TCP backchannel socket
2929  * @args: rpc transport creation arguments
2930  *
2931  */
2932 static struct rpc_xprt *xs_setup_bc_tcp(struct xprt_create *args)
2933 {
2934         struct sockaddr *addr = args->dstaddr;
2935         struct rpc_xprt *xprt;
2936         struct sock_xprt *transport;
2937         struct svc_sock *bc_sock;
2938         struct rpc_xprt *ret;
2939
2940         xprt = xs_setup_xprt(args, xprt_tcp_slot_table_entries,
2941                         xprt_tcp_slot_table_entries);
2942         if (IS_ERR(xprt))
2943                 return xprt;
2944         transport = container_of(xprt, struct sock_xprt, xprt);
2945
2946         xprt->prot = IPPROTO_TCP;
2947         xprt->tsh_size = sizeof(rpc_fraghdr) / sizeof(u32);
2948         xprt->max_payload = RPC_MAX_FRAGMENT_SIZE;
2949         xprt->timeout = &xs_tcp_default_timeout;
2950
2951         /* backchannel */
2952         xprt_set_bound(xprt);
2953         xprt->bind_timeout = 0;
2954         xprt->reestablish_timeout = 0;
2955         xprt->idle_timeout = 0;
2956
2957         xprt->ops = &bc_tcp_ops;
2958
2959         switch (addr->sa_family) {
2960         case AF_INET:
2961                 xs_format_peer_addresses(xprt, "tcp",
2962                                          RPCBIND_NETID_TCP);
2963                 break;
2964         case AF_INET6:
2965                 xs_format_peer_addresses(xprt, "tcp",
2966                                    RPCBIND_NETID_TCP6);
2967                 break;
2968         default:
2969                 ret = ERR_PTR(-EAFNOSUPPORT);
2970                 goto out_err;
2971         }
2972
2973         dprintk("RPC:       set up xprt to %s (port %s) via %s\n",
2974                         xprt->address_strings[RPC_DISPLAY_ADDR],
2975                         xprt->address_strings[RPC_DISPLAY_PORT],
2976                         xprt->address_strings[RPC_DISPLAY_PROTO]);
2977
2978         /*
2979          * Once we've associated a backchannel xprt with a connection,
2980          * we want to keep it around as long as the connection lasts,
2981          * in case we need to start using it for a backchannel again;
2982          * this reference won't be dropped until bc_xprt is destroyed.
2983          */
2984         xprt_get(xprt);
2985         args->bc_xprt->xpt_bc_xprt = xprt;
2986         xprt->bc_xprt = args->bc_xprt;
2987         bc_sock = container_of(args->bc_xprt, struct svc_sock, sk_xprt);
2988         transport->sock = bc_sock->sk_sock;
2989         transport->inet = bc_sock->sk_sk;
2990
2991         /*
2992          * Since we don't want connections for the backchannel, we set
2993          * the xprt status to connected
2994          */
2995         xprt_set_connected(xprt);
2996
2997         if (try_module_get(THIS_MODULE))
2998                 return xprt;
2999
3000         args->bc_xprt->xpt_bc_xprt = NULL;
3001         xprt_put(xprt);
3002         ret = ERR_PTR(-EINVAL);
3003 out_err:
3004         xs_xprt_free(xprt);
3005         return ret;
3006 }
3007
3008 static struct xprt_class        xs_local_transport = {
3009         .list           = LIST_HEAD_INIT(xs_local_transport.list),
3010         .name           = "named UNIX socket",
3011         .owner          = THIS_MODULE,
3012         .ident          = XPRT_TRANSPORT_LOCAL,
3013         .setup          = xs_setup_local,
3014 };
3015
3016 static struct xprt_class        xs_udp_transport = {
3017         .list           = LIST_HEAD_INIT(xs_udp_transport.list),
3018         .name           = "udp",
3019         .owner          = THIS_MODULE,
3020         .ident          = XPRT_TRANSPORT_UDP,
3021         .setup          = xs_setup_udp,
3022 };
3023
3024 static struct xprt_class        xs_tcp_transport = {
3025         .list           = LIST_HEAD_INIT(xs_tcp_transport.list),
3026         .name           = "tcp",
3027         .owner          = THIS_MODULE,
3028         .ident          = XPRT_TRANSPORT_TCP,
3029         .setup          = xs_setup_tcp,
3030 };
3031
3032 static struct xprt_class        xs_bc_tcp_transport = {
3033         .list           = LIST_HEAD_INIT(xs_bc_tcp_transport.list),
3034         .name           = "tcp NFSv4.1 backchannel",
3035         .owner          = THIS_MODULE,
3036         .ident          = XPRT_TRANSPORT_BC_TCP,
3037         .setup          = xs_setup_bc_tcp,
3038 };
3039
3040 /**
3041  * init_socket_xprt - set up xprtsock's sysctls, register with RPC client
3042  *
3043  */
3044 int init_socket_xprt(void)
3045 {
3046 #ifdef RPC_DEBUG
3047         if (!sunrpc_table_header)
3048                 sunrpc_table_header = register_sysctl_table(sunrpc_table);
3049 #endif
3050
3051         xprt_register_transport(&xs_local_transport);
3052         xprt_register_transport(&xs_udp_transport);
3053         xprt_register_transport(&xs_tcp_transport);
3054         xprt_register_transport(&xs_bc_tcp_transport);
3055
3056         return 0;
3057 }
3058
3059 /**
3060  * cleanup_socket_xprt - remove xprtsock's sysctls, unregister
3061  *
3062  */
3063 void cleanup_socket_xprt(void)
3064 {
3065 #ifdef RPC_DEBUG
3066         if (sunrpc_table_header) {
3067                 unregister_sysctl_table(sunrpc_table_header);
3068                 sunrpc_table_header = NULL;
3069         }
3070 #endif
3071
3072         xprt_unregister_transport(&xs_local_transport);
3073         xprt_unregister_transport(&xs_udp_transport);
3074         xprt_unregister_transport(&xs_tcp_transport);
3075         xprt_unregister_transport(&xs_bc_tcp_transport);
3076 }
3077
3078 static int param_set_uint_minmax(const char *val,
3079                 const struct kernel_param *kp,
3080                 unsigned int min, unsigned int max)
3081 {
3082         unsigned int num;
3083         int ret;
3084
3085         if (!val)
3086                 return -EINVAL;
3087         ret = kstrtouint(val, 0, &num);
3088         if (ret == -EINVAL || num < min || num > max)
3089                 return -EINVAL;
3090         *((unsigned int *)kp->arg) = num;
3091         return 0;
3092 }
3093
3094 static int param_set_portnr(const char *val, const struct kernel_param *kp)
3095 {
3096         return param_set_uint_minmax(val, kp,
3097                         RPC_MIN_RESVPORT,
3098                         RPC_MAX_RESVPORT);
3099 }
3100
3101 static struct kernel_param_ops param_ops_portnr = {
3102         .set = param_set_portnr,
3103         .get = param_get_uint,
3104 };
3105
3106 #define param_check_portnr(name, p) \
3107         __param_check(name, p, unsigned int);
3108
3109 module_param_named(min_resvport, xprt_min_resvport, portnr, 0644);
3110 module_param_named(max_resvport, xprt_max_resvport, portnr, 0644);
3111
3112 static int param_set_slot_table_size(const char *val,
3113                                      const struct kernel_param *kp)
3114 {
3115         return param_set_uint_minmax(val, kp,
3116                         RPC_MIN_SLOT_TABLE,
3117                         RPC_MAX_SLOT_TABLE);
3118 }
3119
3120 static struct kernel_param_ops param_ops_slot_table_size = {
3121         .set = param_set_slot_table_size,
3122         .get = param_get_uint,
3123 };
3124
3125 #define param_check_slot_table_size(name, p) \
3126         __param_check(name, p, unsigned int);
3127
3128 static int param_set_max_slot_table_size(const char *val,
3129                                      const struct kernel_param *kp)
3130 {
3131         return param_set_uint_minmax(val, kp,
3132                         RPC_MIN_SLOT_TABLE,
3133                         RPC_MAX_SLOT_TABLE_LIMIT);
3134 }
3135
3136 static struct kernel_param_ops param_ops_max_slot_table_size = {
3137         .set = param_set_max_slot_table_size,
3138         .get = param_get_uint,
3139 };
3140
3141 #define param_check_max_slot_table_size(name, p) \
3142         __param_check(name, p, unsigned int);
3143
3144 module_param_named(tcp_slot_table_entries, xprt_tcp_slot_table_entries,
3145                    slot_table_size, 0644);
3146 module_param_named(tcp_max_slot_table_entries, xprt_max_tcp_slot_table_entries,
3147                    max_slot_table_size, 0644);
3148 module_param_named(udp_slot_table_entries, xprt_udp_slot_table_entries,
3149                    slot_table_size, 0644);
3150