net/sunrpc/xprt.c

   1 /*
   2  *  linux/net/sunrpc/xprt.c
   3  *
   4  *  This is a generic RPC call interface supporting congestion avoidance,
   5  *  and asynchronous calls.
   6  *
   7  *  The interface works like this:
   8  *
   9  *  -   When a process places a call, it allocates a request slot if
  10  *      one is available. Otherwise, it sleeps on the backlog queue
  11  *      (xprt_reserve).
  12  *  -   Next, the caller puts together the RPC message, stuffs it into
  13  *      the request struct, and calls xprt_call().
  14  *  -   xprt_call transmits the message and installs the caller on the
  15  *      socket's wait list. At the same time, it installs a timer that
  16  *      is run after the packet's timeout has expired.
  17  *  -   When a packet arrives, the data_ready handler walks the list of
  18  *      pending requests for that socket. If a matching XID is found, the
  19  *      caller is woken up, and the timer removed.
  20  *  -   When no reply arrives within the timeout interval, the timer is
  21  *      fired by the kernel and runs xprt_timer(). It either adjusts the
  22  *      timeout values (minor timeout) or wakes up the caller with a status
  23  *      of -ETIMEDOUT.
  24  *  -   When the caller receives a notification from RPC that a reply arrived,
  25  *      it should release the RPC slot, and process the reply.
  26  *      If the call timed out, it may choose to retry the operation by
  27  *      adjusting the initial timeout value, and simply calling rpc_call
  28  *      again.
  29  *
  30  *  Support for async RPC is done through a set of RPC-specific scheduling
  31  *  primitives that `transparently' work for processes as well as async
  32  *  tasks that rely on callbacks.
  33  *
  34  *  Copyright (C) 1995-1997, Olaf Kirch <okir@monad.swb.de>
  35  *
  36  *  TCP callback races fixes (C) 1998 Red Hat Software <alan@redhat.com>
  37  *  TCP send fixes (C) 1998 Red Hat Software <alan@redhat.com>
  38  *  TCP NFS related read + write fixes
  39  *   (C) 1999 Dave Airlie, University of Limerick, Ireland <airlied@linux.ie>
  40  *
  41  *  Rewrite of larges part of the code in order to stabilize TCP stuff.
  42  *  Fix behaviour when socket buffer is full.
  43  *   (C) 1999 Trond Myklebust <trond.myklebust@fys.uio.no>
  44  */
  45
  46 #include <linux/types.h>
  47 #include <linux/slab.h>
  48 #include <linux/capability.h>
  49 #include <linux/sched.h>
  50 #include <linux/errno.h>
  51 #include <linux/socket.h>
  52 #include <linux/in.h>
  53 #include <linux/net.h>
  54 #include <linux/mm.h>
  55 #include <linux/udp.h>
  56 #include <linux/tcp.h>
  57 #include <linux/sunrpc/clnt.h>
  58 #include <linux/file.h>
  59 #include <linux/workqueue.h>
  60 #include <linux/random.h>
  61
  62 #include <net/sock.h>
  63 #include <net/checksum.h>
  64 #include <net/udp.h>
  65 #include <net/tcp.h>
  66
  67 /*
  68  * Local variables
  69  */
  70
  71 #ifdef RPC_DEBUG
  72 # undef  RPC_DEBUG_DATA
  73 # define RPCDBG_FACILITY        RPCDBG_XPRT
  74 #endif
  75
  76 #define XPRT_MAX_BACKOFF        (8)
  77 #define XPRT_IDLE_TIMEOUT       (5*60*HZ)
  78 #define XPRT_MAX_RESVPORT       (800)
  79
  80 /*
  81  * Local functions
  82  */
  83 static void     xprt_request_init(struct rpc_task *, struct rpc_xprt *);
  84 static inline void      do_xprt_reserve(struct rpc_task *);
  85 static void     xprt_disconnect(struct rpc_xprt *);
  86 static void     xprt_connect_status(struct rpc_task *task);
  87 static struct rpc_xprt * xprt_setup(int proto, struct sockaddr_in *ap,
  88                                                 struct rpc_timeout *to);
  89 static struct socket *xprt_create_socket(struct rpc_xprt *, int, int);
  90 static void     xprt_bind_socket(struct rpc_xprt *, struct socket *);
  91 static int      __xprt_get_cong(struct rpc_xprt *, struct rpc_task *);
  92
  93 static int      xprt_clear_backlog(struct rpc_xprt *xprt);
  94
  95 #ifdef RPC_DEBUG_DATA
  96 /*
  97  * Print the buffer contents (first 128 bytes only--just enough for
  98  * diropres return).
  99  */
 100 static void
 101 xprt_pktdump(char *msg, u32 *packet, unsigned int count)
 102 {
 103         u8      *buf = (u8 *) packet;
 104         int     j;
 105
 106         dprintk("RPC:      %s\n", msg);
 107         for (j = 0; j < count && j < 128; j += 4) {
 108                 if (!(j & 31)) {
 109                         if (j)
 110                                 dprintk("\n");
 111                         dprintk("0x%04x ", j);
 112                 }
 113                 dprintk("%02x%02x%02x%02x ",
 114                         buf[j], buf[j+1], buf[j+2], buf[j+3]);
 115         }
 116         dprintk("\n");
 117 }
 118 #else
 119 static inline void
 120 xprt_pktdump(char *msg, u32 *packet, unsigned int count)
 121 {
 122         /* NOP */
 123 }
 124 #endif
 125
 126 /*
 127  * Look up RPC transport given an INET socket
 128  */
 129 static inline struct rpc_xprt *
 130 xprt_from_sock(struct sock *sk)
 131 {
 132         return (struct rpc_xprt *) sk->sk_user_data;
 133 }
 134
 135 /*
 136  * Serialize write access to sockets, in order to prevent different
 137  * requests from interfering with each other.
 138  * Also prevents TCP socket connects from colliding with writes.
 139  */
 140 static int
 141 __xprt_lock_write(struct rpc_xprt *xprt, struct rpc_task *task)
 142 {
 143         struct rpc_rqst *req = task->tk_rqstp;
 144
 145         if (test_and_set_bit(XPRT_LOCKED, &xprt->sockstate)) {
 146                 if (task == xprt->snd_task)
 147                         return 1;
 148                 goto out_sleep;
 149         }
 150         if (xprt->nocong || __xprt_get_cong(xprt, task)) {
 151                 xprt->snd_task = task;
 152                 if (req) {
 153                         req->rq_bytes_sent = 0;
 154                         req->rq_ntrans++;
 155                 }
 156                 return 1;
 157         }
 158         smp_mb__before_clear_bit();
 159         clear_bit(XPRT_LOCKED, &xprt->sockstate);
 160         smp_mb__after_clear_bit();
 161 out_sleep:
 162         dprintk("RPC: %4d failed to lock socket %p\n", task->tk_pid, xprt);
 163         task->tk_timeout = 0;
 164         task->tk_status = -EAGAIN;
 165         if (req && req->rq_ntrans)
 166                 rpc_sleep_on(&xprt->resend, task, NULL, NULL);
 167         else
 168                 rpc_sleep_on(&xprt->sending, task, NULL, NULL);
 169         return 0;
 170 }
 171
 172 static inline int
 173 xprt_lock_write(struct rpc_xprt *xprt, struct rpc_task *task)
 174 {
 175         int retval;
 176
 177         spin_lock_bh(&xprt->sock_lock);
 178         retval = __xprt_lock_write(xprt, task);
 179         spin_unlock_bh(&xprt->sock_lock);
 180         return retval;
 181 }
 182
 183
 184 static void
 185 __xprt_lock_write_next(struct rpc_xprt *xprt)
 186 {
 187         struct rpc_task *task;
 188
 189         if (test_and_set_bit(XPRT_LOCKED, &xprt->sockstate))
 190                 return;
 191         if (!xprt->nocong && RPCXPRT_CONGESTED(xprt))
 192                 goto out_unlock;
 193         task = rpc_wake_up_next(&xprt->resend);
 194         if (!task) {
 195                 task = rpc_wake_up_next(&xprt->sending);
 196                 if (!task)
 197                         goto out_unlock;
 198         }
 199         if (xprt->nocong || __xprt_get_cong(xprt, task)) {
 200                 struct rpc_rqst *req = task->tk_rqstp;
 201                 xprt->snd_task = task;
 202                 if (req) {
 203                         req->rq_bytes_sent = 0;
 204                         req->rq_ntrans++;
 205                 }
 206                 return;
 207         }
 208 out_unlock:
 209         smp_mb__before_clear_bit();
 210         clear_bit(XPRT_LOCKED, &xprt->sockstate);
 211         smp_mb__after_clear_bit();
 212 }
 213
 214 /*
 215  * Releases the socket for use by other requests.
 216  */
 217 static void
 218 __xprt_release_write(struct rpc_xprt *xprt, struct rpc_task *task)
 219 {
 220         if (xprt->snd_task == task) {
 221                 xprt->snd_task = NULL;
 222                 smp_mb__before_clear_bit();
 223                 clear_bit(XPRT_LOCKED, &xprt->sockstate);
 224                 smp_mb__after_clear_bit();
 225                 __xprt_lock_write_next(xprt);
 226         }
 227 }
 228
 229 static inline void
 230 xprt_release_write(struct rpc_xprt *xprt, struct rpc_task *task)
 231 {
 232         spin_lock_bh(&xprt->sock_lock);
 233         __xprt_release_write(xprt, task);
 234         spin_unlock_bh(&xprt->sock_lock);
 235 }
 236
 237 /*
 238  * Write data to socket.
 239  */
 240 static inline int
 241 xprt_sendmsg(struct rpc_xprt *xprt, struct rpc_rqst *req)
 242 {
 243         struct socket   *sock = xprt->sock;
 244         struct xdr_buf  *xdr = &req->rq_snd_buf;
 245         struct sockaddr *addr = NULL;
 246         int addrlen = 0;
 247         unsigned int    skip;
 248         int             result;
 249
 250         if (!sock)
 251                 return -ENOTCONN;
 252
 253         xprt_pktdump("packet data:",
 254                                 req->rq_svec->iov_base,
 255                                 req->rq_svec->iov_len);
 256
 257         /* For UDP, we need to provide an address */
 258         if (!xprt->stream) {
 259                 addr = (struct sockaddr *) &xprt->addr;
 260                 addrlen = sizeof(xprt->addr);
 261         }
 262         /* Dont repeat bytes */
 263         skip = req->rq_bytes_sent;
 264
 265         clear_bit(SOCK_ASYNC_NOSPACE, &sock->flags);
 266         result = xdr_sendpages(sock, addr, addrlen, xdr, skip, MSG_DONTWAIT);
 267
 268         dprintk("RPC:      xprt_sendmsg(%d) = %d\n", xdr->len - skip, result);
 269
 270         if (result >= 0)
 271                 return result;
 272
 273         switch (result) {
 274         case -ECONNREFUSED:
 275                 /* When the server has died, an ICMP port unreachable message
 276                  * prompts ECONNREFUSED.
 277                  */
 278         case -EAGAIN:
 279                 break;
 280         case -ECONNRESET:
 281         case -ENOTCONN:
 282         case -EPIPE:
 283                 /* connection broken */
 284                 if (xprt->stream)
 285                         result = -ENOTCONN;
 286                 break;
 287         default:
 288                 printk(KERN_NOTICE "RPC: sendmsg returned error %d\n", -result);
 289         }
 290         return result;
 291 }
 292
 293 /*
 294  * Van Jacobson congestion avoidance. Check if the congestion window
 295  * overflowed. Put the task to sleep if this is the case.
 296  */
 297 static int
 298 __xprt_get_cong(struct rpc_xprt *xprt, struct rpc_task *task)
 299 {
 300         struct rpc_rqst *req = task->tk_rqstp;
 301
 302         if (req->rq_cong)
 303                 return 1;
 304         dprintk("RPC: %4d xprt_cwnd_limited cong = %ld cwnd = %ld\n",
 305                         task->tk_pid, xprt->cong, xprt->cwnd);
 306         if (RPCXPRT_CONGESTED(xprt))
 307                 return 0;
 308         req->rq_cong = 1;
 309         xprt->cong += RPC_CWNDSCALE;
 310         return 1;
 311 }
 312
 313 /*
 314  * Adjust the congestion window, and wake up the next task
 315  * that has been sleeping due to congestion
 316  */
 317 static void
 318 __xprt_put_cong(struct rpc_xprt *xprt, struct rpc_rqst *req)
 319 {
 320         if (!req->rq_cong)
 321                 return;
 322         req->rq_cong = 0;
 323         xprt->cong -= RPC_CWNDSCALE;
 324         __xprt_lock_write_next(xprt);
 325 }
 326
 327 /*
 328  * Adjust RPC congestion window
 329  * We use a time-smoothed congestion estimator to avoid heavy oscillation.
 330  */
 331 static void
 332 xprt_adjust_cwnd(struct rpc_xprt *xprt, int result)
 333 {
 334         unsigned long   cwnd;
 335
 336         cwnd = xprt->cwnd;
 337         if (result >= 0 && cwnd <= xprt->cong) {
 338                 /* The (cwnd >> 1) term makes sure
 339                  * the result gets rounded properly. */
 340                 cwnd += (RPC_CWNDSCALE * RPC_CWNDSCALE + (cwnd >> 1)) / cwnd;
 341                 if (cwnd > RPC_MAXCWND(xprt))
 342                         cwnd = RPC_MAXCWND(xprt);
 343                 __xprt_lock_write_next(xprt);
 344         } else if (result == -ETIMEDOUT) {
 345                 cwnd >>= 1;
 346                 if (cwnd < RPC_CWNDSCALE)
 347                         cwnd = RPC_CWNDSCALE;
 348         }
 349         dprintk("RPC:      cong %ld, cwnd was %ld, now %ld\n",
 350                         xprt->cong, xprt->cwnd, cwnd);
 351         xprt->cwnd = cwnd;
 352 }
 353
 354 /*
 355  * Reset the major timeout value
 356  */
 357 static void xprt_reset_majortimeo(struct rpc_rqst *req)
 358 {
 359         struct rpc_timeout *to = &req->rq_xprt->timeout;
 360
 361         req->rq_majortimeo = req->rq_timeout;
 362         if (to->to_exponential)
 363                 req->rq_majortimeo <<= to->to_retries;
 364         else
 365                 req->rq_majortimeo += to->to_increment * to->to_retries;
 366         if (req->rq_majortimeo > to->to_maxval || req->rq_majortimeo == 0)
 367                 req->rq_majortimeo = to->to_maxval;
 368         req->rq_majortimeo += jiffies;
 369 }
 370
 371 /*
 372  * Adjust timeout values etc for next retransmit
 373  */
 374 int xprt_adjust_timeout(struct rpc_rqst *req)
 375 {
 376         struct rpc_xprt *xprt = req->rq_xprt;
 377         struct rpc_timeout *to = &xprt->timeout;
 378         int status = 0;
 379
 380         if (time_before(jiffies, req->rq_majortimeo)) {
 381                 if (to->to_exponential)
 382                         req->rq_timeout <<= 1;
 383                 else
 384                         req->rq_timeout += to->to_increment;
 385                 if (to->to_maxval && req->rq_timeout >= to->to_maxval)
 386                         req->rq_timeout = to->to_maxval;
 387                 req->rq_retries++;
 388                 pprintk("RPC: %lu retrans\n", jiffies);
 389         } else {
 390                 req->rq_timeout = to->to_initval;
 391                 req->rq_retries = 0;
 392                 xprt_reset_majortimeo(req);
 393                 /* Reset the RTT counters == "slow start" */
 394                 spin_lock_bh(&xprt->sock_lock);
 395                 rpc_init_rtt(req->rq_task->tk_client->cl_rtt, to->to_initval);
 396                 spin_unlock_bh(&xprt->sock_lock);
 397                 pprintk("RPC: %lu timeout\n", jiffies);
 398                 status = -ETIMEDOUT;
 399         }
 400
 401         if (req->rq_timeout == 0) {
 402                 printk(KERN_WARNING "xprt_adjust_timeout: rq_timeout = 0!\n");
 403                 req->rq_timeout = 5 * HZ;
 404         }
 405         return status;
 406 }
 407
 408 /*
 409  * Close down a transport socket
 410  */
 411 static void
 412 xprt_close(struct rpc_xprt *xprt)
 413 {
 414         struct socket   *sock = xprt->sock;
 415         struct sock     *sk = xprt->inet;
 416
 417         if (!sk)
 418                 return;
 419
 420         write_lock_bh(&sk->sk_callback_lock);
 421         xprt->inet = NULL;
 422         xprt->sock = NULL;
 423
 424         sk->sk_user_data    = NULL;
 425         sk->sk_data_ready   = xprt->old_data_ready;
 426         sk->sk_state_change = xprt->old_state_change;
 427         sk->sk_write_space  = xprt->old_write_space;
 428         write_unlock_bh(&sk->sk_callback_lock);
 429
 430         sk->sk_no_check  = 0;
 431
 432         sock_release(sock);
 433 }
 434
 435 static void
 436 xprt_socket_autoclose(void *args)
 437 {
 438         struct rpc_xprt *xprt = (struct rpc_xprt *)args;
 439
 440         xprt_disconnect(xprt);
 441         xprt_close(xprt);
 442         xprt_release_write(xprt, NULL);
 443 }
 444
 445 /*
 446  * Mark a transport as disconnected
 447  */
 448 static void
 449 xprt_disconnect(struct rpc_xprt *xprt)
 450 {
 451         dprintk("RPC:      disconnected transport %p\n", xprt);
 452         spin_lock_bh(&xprt->sock_lock);
 453         xprt_clear_connected(xprt);
 454         rpc_wake_up_status(&xprt->pending, -ENOTCONN);
 455         spin_unlock_bh(&xprt->sock_lock);
 456 }
 457
 458 /*
 459  * Used to allow disconnection when we've been idle
 460  */
 461 static void
 462 xprt_init_autodisconnect(unsigned long data)
 463 {
 464         struct rpc_xprt *xprt = (struct rpc_xprt *)data;
 465
 466         spin_lock(&xprt->sock_lock);
 467         if (!list_empty(&xprt->recv) || xprt->shutdown)
 468                 goto out_abort;
 469         if (test_and_set_bit(XPRT_LOCKED, &xprt->sockstate))
 470                 goto out_abort;
 471         spin_unlock(&xprt->sock_lock);
 472         /* Let keventd close the socket */
 473         if (test_bit(XPRT_CONNECTING, &xprt->sockstate) != 0)
 474                 xprt_release_write(xprt, NULL);
 475         else
 476                 schedule_work(&xprt->task_cleanup);
 477         return;
 478 out_abort:
 479         spin_unlock(&xprt->sock_lock);
 480 }
 481
 482 static void xprt_socket_connect(void *args)
 483 {
 484         struct rpc_xprt *xprt = (struct rpc_xprt *)args;
 485         struct socket *sock = xprt->sock;
 486         int status = -EIO;
 487
 488         if (xprt->shutdown || xprt->addr.sin_port == 0)
 489                 goto out;
 490
 491         /*
 492          * Start by resetting any existing state
 493          */
 494         xprt_close(xprt);
 495         sock = xprt_create_socket(xprt, xprt->prot, xprt->resvport);
 496         if (sock == NULL) {
 497                 /* couldn't create socket or bind to reserved port;
 498                  * this is likely a permanent error, so cause an abort */
 499                 goto out;
 500         }
 501         xprt_bind_socket(xprt, sock);
 502         xprt_sock_setbufsize(xprt);
 503
 504         status = 0;
 505         if (!xprt->stream)
 506                 goto out;
 507
 508         /*
 509          * Tell the socket layer to start connecting...
 510          */
 511         status = sock->ops->connect(sock, (struct sockaddr *) &xprt->addr,
 512                         sizeof(xprt->addr), O_NONBLOCK);
 513         dprintk("RPC: %p  connect status %d connected %d sock state %d\n",
 514                         xprt, -status, xprt_connected(xprt), sock->sk->sk_state);
 515         if (status < 0) {
 516                 switch (status) {
 517                         case -EINPROGRESS:
 518                         case -EALREADY:
 519                                 goto out_clear;
 520                 }
 521         }
 522 out:
 523         if (status < 0)
 524                 rpc_wake_up_status(&xprt->pending, status);
 525         else
 526                 rpc_wake_up(&xprt->pending);
 527 out_clear:
 528         smp_mb__before_clear_bit();
 529         clear_bit(XPRT_CONNECTING, &xprt->sockstate);
 530         smp_mb__after_clear_bit();
 531 }
 532
 533 /*
 534  * Attempt to connect a TCP socket.
 535  *
 536  */
 537 void xprt_connect(struct rpc_task *task)
 538 {
 539         struct rpc_xprt *xprt = task->tk_xprt;
 540
 541         dprintk("RPC: %4d xprt_connect xprt %p %s connected\n", task->tk_pid,
 542                         xprt, (xprt_connected(xprt) ? "is" : "is not"));
 543
 544         if (xprt->shutdown) {
 545                 task->tk_status = -EIO;
 546                 return;
 547         }
 548         if (!xprt->addr.sin_port) {
 549                 task->tk_status = -EIO;
 550                 return;
 551         }
 552         if (!xprt_lock_write(xprt, task))
 553                 return;
 554         if (xprt_connected(xprt))
 555                 goto out_write;
 556
 557         if (task->tk_rqstp)
 558                 task->tk_rqstp->rq_bytes_sent = 0;
 559
 560         task->tk_timeout = RPC_CONNECT_TIMEOUT;
 561         rpc_sleep_on(&xprt->pending, task, xprt_connect_status, NULL);
 562         if (!test_and_set_bit(XPRT_CONNECTING, &xprt->sockstate)) {
 563                 /* Note: if we are here due to a dropped connection
 564                  *       we delay reconnecting by RPC_REESTABLISH_TIMEOUT/HZ
 565                  *       seconds
 566                  */
 567                 if (xprt->sock != NULL)
 568                         schedule_delayed_work(&xprt->sock_connect,
 569                                         RPC_REESTABLISH_TIMEOUT);
 570                 else {
 571                         schedule_work(&xprt->sock_connect);
 572                         if (!RPC_IS_ASYNC(task))
 573                                 flush_scheduled_work();
 574                 }
 575         }
 576         return;
 577  out_write:
 578         xprt_release_write(xprt, task);
 579 }
 580
 581 /*
 582  * We arrive here when awoken from waiting on connection establishment.
 583  */
 584 static void
 585 xprt_connect_status(struct rpc_task *task)
 586 {
 587         struct rpc_xprt *xprt = task->tk_xprt;
 588
 589         if (task->tk_status >= 0) {
 590                 dprintk("RPC: %4d xprt_connect_status: connection established\n",
 591                                 task->tk_pid);
 592                 return;
 593         }
 594
 595         switch (task->tk_status) {
 596         case -ECONNREFUSED:
 597         case -ECONNRESET:
 598                 dprintk("RPC: %4d xprt_connect_status: server %s refused connection\n",
 599                                 task->tk_pid, task->tk_client->cl_server);
 600                 break;
 601         case -ENOTCONN:
 602                 dprintk("RPC: %4d xprt_connect_status: connection broken\n",
 603                                 task->tk_pid);
 604                 break;
 605         case -ETIMEDOUT:
 606                 dprintk("RPC: %4d xprt_connect_status: connect attempt timed out\n",
 607                                 task->tk_pid);
 608                 break;
 609         default:
 610                 dprintk("RPC: %4d xprt_connect_status: error %d connecting to server %s\n",
 611                                 task->tk_pid, -task->tk_status, task->tk_client->cl_server);
 612                 xprt_release_write(xprt, task);
 613                 task->tk_status = -EIO;
 614                 return;
 615         }
 616
 617         /* if soft mounted, just cause this RPC to fail */
 618         if (RPC_IS_SOFT(task)) {
 619                 xprt_release_write(xprt, task);
 620                 task->tk_status = -EIO;
 621         }
 622 }
 623
 624 /*
 625  * Look up the RPC request corresponding to a reply, and then lock it.
 626  */
 627 static inline struct rpc_rqst *
 628 xprt_lookup_rqst(struct rpc_xprt *xprt, u32 xid)
 629 {
 630         struct list_head *pos;
 631         struct rpc_rqst *req = NULL;
 632
 633         list_for_each(pos, &xprt->recv) {
 634                 struct rpc_rqst *entry = list_entry(pos, struct rpc_rqst, rq_list);
 635                 if (entry->rq_xid == xid) {
 636                         req = entry;
 637                         break;
 638                 }
 639         }
 640         return req;
 641 }
 642
 643 /*
 644  * Complete reply received.
 645  * The TCP code relies on us to remove the request from xprt->pending.
 646  */
 647 static void
 648 xprt_complete_rqst(struct rpc_xprt *xprt, struct rpc_rqst *req, int copied)
 649 {
 650         struct rpc_task *task = req->rq_task;
 651         struct rpc_clnt *clnt = task->tk_client;
 652
 653         /* Adjust congestion window */
 654         if (!xprt->nocong) {
 655                 unsigned timer = task->tk_msg.rpc_proc->p_timer;
 656                 xprt_adjust_cwnd(xprt, copied);
 657                 __xprt_put_cong(xprt, req);
 658                 if (timer) {
 659                         if (req->rq_ntrans == 1)
 660                                 rpc_update_rtt(clnt->cl_rtt, timer,
 661                                                 (long)jiffies - req->rq_xtime);
 662                         rpc_set_timeo(clnt->cl_rtt, timer, req->rq_ntrans - 1);
 663                 }
 664         }
 665
 666 #ifdef RPC_PROFILE
 667         /* Profile only reads for now */
 668         if (copied > 1024) {
 669                 static unsigned long    nextstat;
 670                 static unsigned long    pkt_rtt, pkt_len, pkt_cnt;
 671
 672                 pkt_cnt++;
 673                 pkt_len += req->rq_slen + copied;
 674                 pkt_rtt += jiffies - req->rq_xtime;
 675                 if (time_before(nextstat, jiffies)) {
 676                         printk("RPC: %lu %ld cwnd\n", jiffies, xprt->cwnd);
 677                         printk("RPC: %ld %ld %ld %ld stat\n",
 678                                         jiffies, pkt_cnt, pkt_len, pkt_rtt);
 679                         pkt_rtt = pkt_len = pkt_cnt = 0;
 680                         nextstat = jiffies + 5 * HZ;
 681                 }
 682         }
 683 #endif
 684
 685         dprintk("RPC: %4d has input (%d bytes)\n", task->tk_pid, copied);
 686         list_del_init(&req->rq_list);
 687         req->rq_received = req->rq_private_buf.len = copied;
 688
 689         /* ... and wake up the process. */
 690         rpc_wake_up_task(task);
 691         return;
 692 }
 693
 694 /*
 695  * Input handler for RPC replies. Called from a bottom half and hence
 696  * atomic.
 697  */
 698 static void
 699 udp_data_ready(struct sock *sk, int len)
 700 {
 701         struct rpc_task *task;
 702         struct rpc_xprt *xprt;
 703         struct rpc_rqst *rovr;
 704         struct sk_buff  *skb;
 705         int err, repsize, copied;
 706         u32 _xid, *xp;
 707
 708         read_lock(&sk->sk_callback_lock);
 709         dprintk("RPC:      udp_data_ready...\n");
 710         if (!(xprt = xprt_from_sock(sk))) {
 711                 printk("RPC:      udp_data_ready request not found!\n");
 712                 goto out;
 713         }
 714
 715         dprintk("RPC:      udp_data_ready client %p\n", xprt);
 716
 717         if ((skb = skb_recv_datagram(sk, 0, 1, &err)) == NULL)
 718                 goto out;
 719
 720         if (xprt->shutdown)
 721                 goto dropit;
 722
 723         repsize = skb->len - sizeof(struct udphdr);
 724         if (repsize < 4) {
 725                 printk("RPC: impossible RPC reply size %d!\n", repsize);
 726                 goto dropit;
 727         }
 728
 729         /* Copy the XID from the skb... */
 730         xp = skb_header_pointer(skb, sizeof(struct udphdr),
 731                                 sizeof(_xid), &_xid);
 732         if (xp == NULL)
 733                 goto dropit;
 734
 735         /* Look up and lock the request corresponding to the given XID */
 736         spin_lock(&xprt->sock_lock);
 737         rovr = xprt_lookup_rqst(xprt, *xp);
 738         if (!rovr)
 739                 goto out_unlock;
 740         task = rovr->rq_task;
 741
 742         dprintk("RPC: %4d received reply\n", task->tk_pid);
 743
 744         if ((copied = rovr->rq_private_buf.buflen) > repsize)
 745                 copied = repsize;
 746
 747         /* Suck it into the iovec, verify checksum if not done by hw. */
 748         if (csum_partial_copy_to_xdr(&rovr->rq_private_buf, skb))
 749                 goto out_unlock;
 750
 751         /* Something worked... */
 752         dst_confirm(skb->dst);
 753
 754         xprt_complete_rqst(xprt, rovr, copied);
 755
 756  out_unlock:
 757         spin_unlock(&xprt->sock_lock);
 758  dropit:
 759         skb_free_datagram(sk, skb);
 760  out:
 761         read_unlock(&sk->sk_callback_lock);
 762 }
 763
 764 /*
 765  * Copy from an skb into memory and shrink the skb.
 766  */
 767 static inline size_t
 768 tcp_copy_data(skb_reader_t *desc, void *p, size_t len)
 769 {
 770         if (len > desc->count)
 771                 len = desc->count;
 772         if (skb_copy_bits(desc->skb, desc->offset, p, len)) {
 773                 dprintk("RPC:      failed to copy %zu bytes from skb. %zu bytes remain\n",
 774                                 len, desc->count);
 775                 return 0;
 776         }
 777         desc->offset += len;
 778         desc->count -= len;
 779         dprintk("RPC:      copied %zu bytes from skb. %zu bytes remain\n",
 780                         len, desc->count);
 781         return len;
 782 }
 783
 784 /*
 785  * TCP read fragment marker
 786  */
 787 static inline void
 788 tcp_read_fraghdr(struct rpc_xprt *xprt, skb_reader_t *desc)
 789 {
 790         size_t len, used;
 791         char *p;
 792
 793         p = ((char *) &xprt->tcp_recm) + xprt->tcp_offset;
 794         len = sizeof(xprt->tcp_recm) - xprt->tcp_offset;
 795         used = tcp_copy_data(desc, p, len);
 796         xprt->tcp_offset += used;
 797         if (used != len)
 798                 return;
 799         xprt->tcp_reclen = ntohl(xprt->tcp_recm);
 800         if (xprt->tcp_reclen & 0x80000000)
 801                 xprt->tcp_flags |= XPRT_LAST_FRAG;
 802         else
 803                 xprt->tcp_flags &= ~XPRT_LAST_FRAG;
 804         xprt->tcp_reclen &= 0x7fffffff;
 805         xprt->tcp_flags &= ~XPRT_COPY_RECM;
 806         xprt->tcp_offset = 0;
 807         /* Sanity check of the record length */
 808         if (xprt->tcp_reclen < 4) {
 809                 printk(KERN_ERR "RPC: Invalid TCP record fragment length\n");
 810                 xprt_disconnect(xprt);
 811         }
 812         dprintk("RPC:      reading TCP record fragment of length %d\n",
 813                         xprt->tcp_reclen);
 814 }
 815
 816 static void
 817 tcp_check_recm(struct rpc_xprt *xprt)
 818 {
 819         dprintk("RPC:      xprt = %p, tcp_copied = %lu, tcp_offset = %u, tcp_reclen = %u, tcp_flags = %lx\n",
 820                         xprt, xprt->tcp_copied, xprt->tcp_offset, xprt->tcp_reclen, xprt->tcp_flags);
 821         if (xprt->tcp_offset == xprt->tcp_reclen) {
 822                 xprt->tcp_flags |= XPRT_COPY_RECM;
 823                 xprt->tcp_offset = 0;
 824                 if (xprt->tcp_flags & XPRT_LAST_FRAG) {
 825                         xprt->tcp_flags &= ~XPRT_COPY_DATA;
 826                         xprt->tcp_flags |= XPRT_COPY_XID;
 827                         xprt->tcp_copied = 0;
 828                 }
 829         }
 830 }
 831
 832 /*
 833  * TCP read xid
 834  */
 835 static inline void
 836 tcp_read_xid(struct rpc_xprt *xprt, skb_reader_t *desc)
 837 {
 838         size_t len, used;
 839         char *p;
 840
 841         len = sizeof(xprt->tcp_xid) - xprt->tcp_offset;
 842         dprintk("RPC:      reading XID (%Zu bytes)\n", len);
 843         p = ((char *) &xprt->tcp_xid) + xprt->tcp_offset;
 844         used = tcp_copy_data(desc, p, len);
 845         xprt->tcp_offset += used;
 846         if (used != len)
 847                 return;
 848         xprt->tcp_flags &= ~XPRT_COPY_XID;
 849         xprt->tcp_flags |= XPRT_COPY_DATA;
 850         xprt->tcp_copied = 4;
 851         dprintk("RPC:      reading reply for XID %08x\n",
 852                                                 ntohl(xprt->tcp_xid));
 853         tcp_check_recm(xprt);
 854 }
 855
 856 /*
 857  * TCP read and complete request
 858  */
 859 static inline void
 860 tcp_read_request(struct rpc_xprt *xprt, skb_reader_t *desc)
 861 {
 862         struct rpc_rqst *req;
 863         struct xdr_buf *rcvbuf;
 864         size_t len;
 865         ssize_t r;
 866
 867         /* Find and lock the request corresponding to this xid */
 868         spin_lock(&xprt->sock_lock);
 869         req = xprt_lookup_rqst(xprt, xprt->tcp_xid);
 870         if (!req) {
 871                 xprt->tcp_flags &= ~XPRT_COPY_DATA;
 872                 dprintk("RPC:      XID %08x request not found!\n",
 873                                 ntohl(xprt->tcp_xid));
 874                 spin_unlock(&xprt->sock_lock);
 875                 return;
 876         }
 877
 878         rcvbuf = &req->rq_private_buf;
 879         len = desc->count;
 880         if (len > xprt->tcp_reclen - xprt->tcp_offset) {
 881                 skb_reader_t my_desc;
 882
 883                 len = xprt->tcp_reclen - xprt->tcp_offset;
 884                 memcpy(&my_desc, desc, sizeof(my_desc));
 885                 my_desc.count = len;
 886                 r = xdr_partial_copy_from_skb(rcvbuf, xprt->tcp_copied,
 887                                           &my_desc, tcp_copy_data);
 888                 desc->count -= r;
 889                 desc->offset += r;
 890         } else
 891                 r = xdr_partial_copy_from_skb(rcvbuf, xprt->tcp_copied,
 892                                           desc, tcp_copy_data);
 893
 894         if (r > 0) {
 895                 xprt->tcp_copied += r;
 896                 xprt->tcp_offset += r;
 897         }
 898         if (r != len) {
 899                 /* Error when copying to the receive buffer,
 900                  * usually because we weren't able to allocate
 901                  * additional buffer pages. All we can do now
 902                  * is turn off XPRT_COPY_DATA, so the request
 903                  * will not receive any additional updates,
 904                  * and time out.
 905                  * Any remaining data from this record will
 906                  * be discarded.
 907                  */
 908                 xprt->tcp_flags &= ~XPRT_COPY_DATA;
 909                 dprintk("RPC:      XID %08x truncated request\n",
 910                                 ntohl(xprt->tcp_xid));
 911                 dprintk("RPC:      xprt = %p, tcp_copied = %lu, tcp_offset = %u, tcp_reclen = %u\n",
 912                                 xprt, xprt->tcp_copied, xprt->tcp_offset, xprt->tcp_reclen);
 913                 goto out;
 914         }
 915
 916         dprintk("RPC:      XID %08x read %Zd bytes\n",
 917                         ntohl(xprt->tcp_xid), r);
 918         dprintk("RPC:      xprt = %p, tcp_copied = %lu, tcp_offset = %u, tcp_reclen = %u\n",
 919                         xprt, xprt->tcp_copied, xprt->tcp_offset, xprt->tcp_reclen);
 920
 921         if (xprt->tcp_copied == req->rq_private_buf.buflen)
 922                 xprt->tcp_flags &= ~XPRT_COPY_DATA;
 923         else if (xprt->tcp_offset == xprt->tcp_reclen) {
 924                 if (xprt->tcp_flags & XPRT_LAST_FRAG)
 925                         xprt->tcp_flags &= ~XPRT_COPY_DATA;
 926         }
 927
 928 out:
 929         if (!(xprt->tcp_flags & XPRT_COPY_DATA)) {
 930                 dprintk("RPC: %4d received reply complete\n",
 931                                 req->rq_task->tk_pid);
 932                 xprt_complete_rqst(xprt, req, xprt->tcp_copied);
 933         }
 934         spin_unlock(&xprt->sock_lock);
 935         tcp_check_recm(xprt);
 936 }
 937
 938 /*
 939  * TCP discard extra bytes from a short read
 940  */
 941 static inline void
 942 tcp_read_discard(struct rpc_xprt *xprt, skb_reader_t *desc)
 943 {
 944         size_t len;
 945
 946         len = xprt->tcp_reclen - xprt->tcp_offset;
 947         if (len > desc->count)
 948                 len = desc->count;
 949         desc->count -= len;
 950         desc->offset += len;
 951         xprt->tcp_offset += len;
 952         dprintk("RPC:      discarded %Zu bytes\n", len);
 953         tcp_check_recm(xprt);
 954 }
 955
 956 /*
 957  * TCP record receive routine
 958  * We first have to grab the record marker, then the XID, then the data.
 959  */
 960 static int
 961 tcp_data_recv(read_descriptor_t *rd_desc, struct sk_buff *skb,
 962                 unsigned int offset, size_t len)
 963 {
 964         struct rpc_xprt *xprt = rd_desc->arg.data;
 965         skb_reader_t desc = {
 966                 .skb    = skb,
 967                 .offset = offset,
 968                 .count  = len,
 969                 .csum   = 0
 970         };
 971
 972         dprintk("RPC:      tcp_data_recv\n");
 973         do {
 974                 /* Read in a new fragment marker if necessary */
 975                 /* Can we ever really expect to get completely empty fragments? */
 976                 if (xprt->tcp_flags & XPRT_COPY_RECM) {
 977                         tcp_read_fraghdr(xprt, &desc);
 978                         continue;
 979                 }
 980                 /* Read in the xid if necessary */
 981                 if (xprt->tcp_flags & XPRT_COPY_XID) {
 982                         tcp_read_xid(xprt, &desc);
 983                         continue;
 984                 }
 985                 /* Read in the request data */
 986                 if (xprt->tcp_flags & XPRT_COPY_DATA) {
 987                         tcp_read_request(xprt, &desc);
 988                         continue;
 989                 }
 990                 /* Skip over any trailing bytes on short reads */
 991                 tcp_read_discard(xprt, &desc);
 992         } while (desc.count);
 993         dprintk("RPC:      tcp_data_recv done\n");
 994         return len - desc.count;
 995 }
 996
 997 static void tcp_data_ready(struct sock *sk, int bytes)
 998 {
 999         struct rpc_xprt *xprt;
1000         read_descriptor_t rd_desc;
1001
1002         read_lock(&sk->sk_callback_lock);
1003         dprintk("RPC:      tcp_data_ready...\n");
1004         if (!(xprt = xprt_from_sock(sk))) {
1005                 printk("RPC:      tcp_data_ready socket info not found!\n");
1006                 goto out;
1007         }
1008         if (xprt->shutdown)
1009                 goto out;
1010
1011         /* We use rd_desc to pass struct xprt to tcp_data_recv */
1012         rd_desc.arg.data = xprt;
1013         rd_desc.count = 65536;
1014         tcp_read_sock(sk, &rd_desc, tcp_data_recv);
1015 out:
1016         read_unlock(&sk->sk_callback_lock);
1017 }
1018
1019 static void
1020 tcp_state_change(struct sock *sk)
1021 {
1022         struct rpc_xprt *xprt;
1023
1024         read_lock(&sk->sk_callback_lock);
1025         if (!(xprt = xprt_from_sock(sk)))
1026                 goto out;
1027         dprintk("RPC:      tcp_state_change client %p...\n", xprt);
1028         dprintk("RPC:      state %x conn %d dead %d zapped %d\n",
1029                                 sk->sk_state, xprt_connected(xprt),
1030                                 sock_flag(sk, SOCK_DEAD),
1031                                 sock_flag(sk, SOCK_ZAPPED));
1032
1033         switch (sk->sk_state) {
1034         case TCP_ESTABLISHED:
1035                 spin_lock_bh(&xprt->sock_lock);
1036                 if (!xprt_test_and_set_connected(xprt)) {
1037                         /* Reset TCP record info */
1038                         xprt->tcp_offset = 0;
1039                         xprt->tcp_reclen = 0;
1040                         xprt->tcp_copied = 0;
1041                         xprt->tcp_flags = XPRT_COPY_RECM | XPRT_COPY_XID;
1042                         rpc_wake_up(&xprt->pending);
1043                 }
1044                 spin_unlock_bh(&xprt->sock_lock);
1045                 break;
1046         case TCP_SYN_SENT:
1047         case TCP_SYN_RECV:
1048                 break;
1049         default:
1050                 xprt_disconnect(xprt);
1051                 break;
1052         }
1053  out:
1054         read_unlock(&sk->sk_callback_lock);
1055 }
1056
1057 /*
1058  * Called when more output buffer space is available for this socket.
1059  * We try not to wake our writers until they can make "significant"
1060  * progress, otherwise we'll waste resources thrashing sock_sendmsg
1061  * with a bunch of small requests.
1062  */
1063 static void
1064 xprt_write_space(struct sock *sk)
1065 {
1066         struct rpc_xprt *xprt;
1067         struct socket   *sock;
1068
1069         read_lock(&sk->sk_callback_lock);
1070         if (!(xprt = xprt_from_sock(sk)) || !(sock = sk->sk_socket))
1071                 goto out;
1072         if (xprt->shutdown)
1073                 goto out;
1074
1075         /* Wait until we have enough socket memory */
1076         if (xprt->stream) {
1077                 /* from net/core/stream.c:sk_stream_write_space */
1078                 if (sk_stream_wspace(sk) < sk_stream_min_wspace(sk))
1079                         goto out;
1080         } else {
1081                 /* from net/core/sock.c:sock_def_write_space */
1082                 if (!sock_writeable(sk))
1083                         goto out;
1084         }
1085
1086         if (!test_and_clear_bit(SOCK_NOSPACE, &sock->flags))
1087                 goto out;
1088
1089         spin_lock_bh(&xprt->sock_lock);
1090         if (xprt->snd_task)
1091                 rpc_wake_up_task(xprt->snd_task);
1092         spin_unlock_bh(&xprt->sock_lock);
1093 out:
1094         read_unlock(&sk->sk_callback_lock);
1095 }
1096
1097 /*
1098  * RPC receive timeout handler.
1099  */
1100 static void
1101 xprt_timer(struct rpc_task *task)
1102 {
1103         struct rpc_rqst *req = task->tk_rqstp;
1104         struct rpc_xprt *xprt = req->rq_xprt;
1105
1106         spin_lock(&xprt->sock_lock);
1107         if (req->rq_received)
1108                 goto out;
1109
1110         xprt_adjust_cwnd(req->rq_xprt, -ETIMEDOUT);
1111         __xprt_put_cong(xprt, req);
1112
1113         dprintk("RPC: %4d xprt_timer (%s request)\n",
1114                 task->tk_pid, req ? "pending" : "backlogged");
1115
1116         task->tk_status  = -ETIMEDOUT;
1117 out:
1118         task->tk_timeout = 0;
1119         rpc_wake_up_task(task);
1120         spin_unlock(&xprt->sock_lock);
1121 }
1122
1123 /*
1124  * Place the actual RPC call.
1125  * We have to copy the iovec because sendmsg fiddles with its contents.
1126  */
1127 int
1128 xprt_prepare_transmit(struct rpc_task *task)
1129 {
1130         struct rpc_rqst *req = task->tk_rqstp;
1131         struct rpc_xprt *xprt = req->rq_xprt;
1132         int err = 0;
1133
1134         dprintk("RPC: %4d xprt_prepare_transmit\n", task->tk_pid);
1135
1136         if (xprt->shutdown)
1137                 return -EIO;
1138
1139         spin_lock_bh(&xprt->sock_lock);
1140         if (req->rq_received && !req->rq_bytes_sent) {
1141                 err = req->rq_received;
1142                 goto out_unlock;
1143         }
1144         if (!__xprt_lock_write(xprt, task)) {
1145                 err = -EAGAIN;
1146                 goto out_unlock;
1147         }
1148
1149         if (!xprt_connected(xprt)) {
1150                 err = -ENOTCONN;
1151                 goto out_unlock;
1152         }
1153 out_unlock:
1154         spin_unlock_bh(&xprt->sock_lock);
1155         return err;
1156 }
1157
1158 void
1159 xprt_transmit(struct rpc_task *task)
1160 {
1161         struct rpc_clnt *clnt = task->tk_client;
1162         struct rpc_rqst *req = task->tk_rqstp;
1163         struct rpc_xprt *xprt = req->rq_xprt;
1164         int status, retry = 0;
1165
1166
1167         dprintk("RPC: %4d xprt_transmit(%u)\n", task->tk_pid, req->rq_slen);
1168
1169         /* set up everything as needed. */
1170         /* Write the record marker */
1171         if (xprt->stream) {
1172                 u32     *marker = req->rq_svec[0].iov_base;
1173
1174                 *marker = htonl(0x80000000|(req->rq_slen-sizeof(*marker)));
1175         }
1176
1177         smp_rmb();
1178         if (!req->rq_received) {
1179                 if (list_empty(&req->rq_list)) {
1180                         spin_lock_bh(&xprt->sock_lock);
1181                         /* Update the softirq receive buffer */
1182                         memcpy(&req->rq_private_buf, &req->rq_rcv_buf,
1183                                         sizeof(req->rq_private_buf));
1184                         /* Add request to the receive list */
1185                         list_add_tail(&req->rq_list, &xprt->recv);
1186                         spin_unlock_bh(&xprt->sock_lock);
1187                         xprt_reset_majortimeo(req);
1188                         /* Turn off autodisconnect */
1189                         del_singleshot_timer_sync(&xprt->timer);
1190                 }
1191         } else if (!req->rq_bytes_sent)
1192                 return;
1193
1194         /* Continue transmitting the packet/record. We must be careful
1195          * to cope with writespace callbacks arriving _after_ we have
1196          * called xprt_sendmsg().
1197          */
1198         while (1) {
1199                 req->rq_xtime = jiffies;
1200                 status = xprt_sendmsg(xprt, req);
1201
1202                 if (status < 0)
1203                         break;
1204
1205                 if (xprt->stream) {
1206                         req->rq_bytes_sent += status;
1207
1208                         /* If we've sent the entire packet, immediately
1209                          * reset the count of bytes sent. */
1210                         if (req->rq_bytes_sent >= req->rq_slen) {
1211                                 req->rq_bytes_sent = 0;
1212                                 goto out_receive;
1213                         }
1214                 } else {
1215                         if (status >= req->rq_slen)
1216                                 goto out_receive;
1217                         status = -EAGAIN;
1218                         break;
1219                 }
1220
1221                 dprintk("RPC: %4d xmit incomplete (%d left of %d)\n",
1222                                 task->tk_pid, req->rq_slen - req->rq_bytes_sent,
1223                                 req->rq_slen);
1224
1225                 status = -EAGAIN;
1226                 if (retry++ > 50)
1227                         break;
1228         }
1229
1230         /* Note: at this point, task->tk_sleeping has not yet been set,
1231          *       hence there is no danger of the waking up task being put on
1232          *       schedq, and being picked up by a parallel run of rpciod().
1233          */
1234         task->tk_status = status;
1235
1236         switch (status) {
1237         case -EAGAIN:
1238                 if (test_bit(SOCK_ASYNC_NOSPACE, &xprt->sock->flags)) {
1239                         /* Protect against races with xprt_write_space */
1240                         spin_lock_bh(&xprt->sock_lock);
1241                         /* Don't race with disconnect */
1242                         if (!xprt_connected(xprt))
1243                                 task->tk_status = -ENOTCONN;
1244                         else if (test_bit(SOCK_NOSPACE, &xprt->sock->flags)) {
1245                                 task->tk_timeout = req->rq_timeout;
1246                                 rpc_sleep_on(&xprt->pending, task, NULL, NULL);
1247                         }
1248                         spin_unlock_bh(&xprt->sock_lock);
1249                         return;
1250                 }
1251                 /* Keep holding the socket if it is blocked */
1252                 rpc_delay(task, HZ>>4);
1253                 return;
1254         case -ECONNREFUSED:
1255                 task->tk_timeout = RPC_REESTABLISH_TIMEOUT;
1256                 rpc_sleep_on(&xprt->sending, task, NULL, NULL);
1257         case -ENOTCONN:
1258                 return;
1259         default:
1260                 if (xprt->stream)
1261                         xprt_disconnect(xprt);
1262         }
1263         xprt_release_write(xprt, task);
1264         return;
1265  out_receive:
1266         dprintk("RPC: %4d xmit complete\n", task->tk_pid);
1267         /* Set the task's receive timeout value */
1268         spin_lock_bh(&xprt->sock_lock);
1269         if (!xprt->nocong) {
1270                 int timer = task->tk_msg.rpc_proc->p_timer;
1271                 task->tk_timeout = rpc_calc_rto(clnt->cl_rtt, timer);
1272                 task->tk_timeout <<= rpc_ntimeo(clnt->cl_rtt, timer) + req->rq_retries;
1273                 if (task->tk_timeout > xprt->timeout.to_maxval || task->tk_timeout == 0)
1274                         task->tk_timeout = xprt->timeout.to_maxval;
1275         } else
1276                 task->tk_timeout = req->rq_timeout;
1277         /* Don't race with disconnect */
1278         if (!xprt_connected(xprt))
1279                 task->tk_status = -ENOTCONN;
1280         else if (!req->rq_received)
1281                 rpc_sleep_on(&xprt->pending, task, NULL, xprt_timer);
1282         __xprt_release_write(xprt, task);
1283         spin_unlock_bh(&xprt->sock_lock);
1284 }
1285
1286 /*
1287  * Reserve an RPC call slot.
1288  */
1289 static inline void
1290 do_xprt_reserve(struct rpc_task *task)
1291 {
1292         struct rpc_xprt *xprt = task->tk_xprt;
1293
1294         task->tk_status = 0;
1295         if (task->tk_rqstp)
1296                 return;
1297         if (!list_empty(&xprt->free)) {
1298                 struct rpc_rqst *req = list_entry(xprt->free.next, struct rpc_rqst, rq_list);
1299                 list_del_init(&req->rq_list);
1300                 task->tk_rqstp = req;
1301                 xprt_request_init(task, xprt);
1302                 return;
1303         }
1304         dprintk("RPC:      waiting for request slot\n");
1305         task->tk_status = -EAGAIN;
1306         task->tk_timeout = 0;
1307         rpc_sleep_on(&xprt->backlog, task, NULL, NULL);
1308 }
1309
1310 void
1311 xprt_reserve(struct rpc_task *task)
1312 {
1313         struct rpc_xprt *xprt = task->tk_xprt;
1314
1315         task->tk_status = -EIO;
1316         if (!xprt->shutdown) {
1317                 spin_lock(&xprt->xprt_lock);
1318                 do_xprt_reserve(task);
1319                 spin_unlock(&xprt->xprt_lock);
1320         }
1321 }
1322
1323 /*
1324  * Allocate a 'unique' XID
1325  */
1326 static inline u32 xprt_alloc_xid(struct rpc_xprt *xprt)
1327 {
1328         return xprt->xid++;
1329 }
1330
1331 static inline void xprt_init_xid(struct rpc_xprt *xprt)
1332 {
1333         get_random_bytes(&xprt->xid, sizeof(xprt->xid));
1334 }
1335
1336 /*
1337  * Initialize RPC request
1338  */
1339 static void
1340 xprt_request_init(struct rpc_task *task, struct rpc_xprt *xprt)
1341 {
1342         struct rpc_rqst *req = task->tk_rqstp;
1343
1344         req->rq_timeout = xprt->timeout.to_initval;
1345         req->rq_task    = task;
1346         req->rq_xprt    = xprt;
1347         req->rq_xid     = xprt_alloc_xid(xprt);
1348         dprintk("RPC: %4d reserved req %p xid %08x\n", task->tk_pid,
1349                         req, ntohl(req->rq_xid));
1350 }
1351
1352 /*
1353  * Release an RPC call slot
1354  */
1355 void
1356 xprt_release(struct rpc_task *task)
1357 {
1358         struct rpc_xprt *xprt = task->tk_xprt;
1359         struct rpc_rqst *req;
1360
1361         if (!(req = task->tk_rqstp))
1362                 return;
1363         spin_lock_bh(&xprt->sock_lock);
1364         __xprt_release_write(xprt, task);
1365         __xprt_put_cong(xprt, req);
1366         if (!list_empty(&req->rq_list))
1367                 list_del(&req->rq_list);
1368         xprt->last_used = jiffies;
1369         if (list_empty(&xprt->recv) && !xprt->shutdown)
1370                 mod_timer(&xprt->timer, xprt->last_used + XPRT_IDLE_TIMEOUT);
1371         spin_unlock_bh(&xprt->sock_lock);
1372         task->tk_rqstp = NULL;
1373         memset(req, 0, sizeof(*req));   /* mark unused */
1374
1375         dprintk("RPC: %4d release request %p\n", task->tk_pid, req);
1376
1377         spin_lock(&xprt->xprt_lock);
1378         list_add(&req->rq_list, &xprt->free);
1379         xprt_clear_backlog(xprt);
1380         spin_unlock(&xprt->xprt_lock);
1381 }
1382
1383 /*
1384  * Set default timeout parameters
1385  */
1386 static void
1387 xprt_default_timeout(struct rpc_timeout *to, int proto)
1388 {
1389         if (proto == IPPROTO_UDP)
1390                 xprt_set_timeout(to, 5,  5 * HZ);
1391         else
1392                 xprt_set_timeout(to, 2, 60 * HZ);
1393 }
1394
1395 /*
1396  * Set constant timeout
1397  */
1398 void
1399 xprt_set_timeout(struct rpc_timeout *to, unsigned int retr, unsigned long incr)
1400 {
1401         to->to_initval   =
1402         to->to_increment = incr;
1403         to->to_maxval    = to->to_initval + (incr * retr);
1404         to->to_retries   = retr;
1405         to->to_exponential = 0;
1406 }
1407
1408 unsigned int xprt_udp_slot_table_entries = RPC_DEF_SLOT_TABLE;
1409 unsigned int xprt_tcp_slot_table_entries = RPC_DEF_SLOT_TABLE;
1410
1411 /*
1412  * Initialize an RPC client
1413  */
1414 static struct rpc_xprt *
1415 xprt_setup(int proto, struct sockaddr_in *ap, struct rpc_timeout *to)
1416 {
1417         struct rpc_xprt *xprt;
1418         unsigned int entries;
1419         size_t slot_table_size;
1420         struct rpc_rqst *req;
1421
1422         dprintk("RPC:      setting up %s transport...\n",
1423                                 proto == IPPROTO_UDP? "UDP" : "TCP");
1424
1425         entries = (proto == IPPROTO_TCP)?
1426                 xprt_tcp_slot_table_entries : xprt_udp_slot_table_entries;
1427
1428         if ((xprt = kmalloc(sizeof(struct rpc_xprt), GFP_KERNEL)) == NULL)
1429                 return ERR_PTR(-ENOMEM);
1430         memset(xprt, 0, sizeof(*xprt)); /* Nnnngh! */
1431         xprt->max_reqs = entries;
1432         slot_table_size = entries * sizeof(xprt->slot[0]);
1433         xprt->slot = kmalloc(slot_table_size, GFP_KERNEL);
1434         if (xprt->slot == NULL) {
1435                 kfree(xprt);
1436                 return ERR_PTR(-ENOMEM);
1437         }
1438         memset(xprt->slot, 0, slot_table_size);
1439
1440         xprt->addr = *ap;
1441         xprt->prot = proto;
1442         xprt->stream = (proto == IPPROTO_TCP)? 1 : 0;
1443         if (xprt->stream) {
1444                 xprt->cwnd = RPC_MAXCWND(xprt);
1445                 xprt->nocong = 1;
1446                 xprt->max_payload = (1U << 31) - 1;
1447         } else {
1448                 xprt->cwnd = RPC_INITCWND;
1449                 xprt->max_payload = (1U << 16) - (MAX_HEADER << 3);
1450         }
1451         spin_lock_init(&xprt->sock_lock);
1452         spin_lock_init(&xprt->xprt_lock);
1453         init_waitqueue_head(&xprt->cong_wait);
1454
1455         INIT_LIST_HEAD(&xprt->free);
1456         INIT_LIST_HEAD(&xprt->recv);
1457         INIT_WORK(&xprt->sock_connect, xprt_socket_connect, xprt);
1458         INIT_WORK(&xprt->task_cleanup, xprt_socket_autoclose, xprt);
1459         init_timer(&xprt->timer);
1460         xprt->timer.function = xprt_init_autodisconnect;
1461         xprt->timer.data = (unsigned long) xprt;
1462         xprt->last_used = jiffies;
1463         xprt->port = XPRT_MAX_RESVPORT;
1464
1465         /* Set timeout parameters */
1466         if (to) {
1467                 xprt->timeout = *to;
1468         } else
1469                 xprt_default_timeout(&xprt->timeout, xprt->prot);
1470
1471         rpc_init_wait_queue(&xprt->pending, "xprt_pending");
1472         rpc_init_wait_queue(&xprt->sending, "xprt_sending");
1473         rpc_init_wait_queue(&xprt->resend, "xprt_resend");
1474         rpc_init_priority_wait_queue(&xprt->backlog, "xprt_backlog");
1475
1476         /* initialize free list */
1477         for (req = &xprt->slot[entries-1]; req >= &xprt->slot[0]; req--)
1478                 list_add(&req->rq_list, &xprt->free);
1479
1480         xprt_init_xid(xprt);
1481
1482         /* Check whether we want to use a reserved port */
1483         xprt->resvport = capable(CAP_NET_BIND_SERVICE) ? 1 : 0;
1484
1485         dprintk("RPC:      created transport %p with %u slots\n", xprt,
1486                         xprt->max_reqs);
1487
1488         return xprt;
1489 }
1490
1491 /*
1492  * Bind to a reserved port
1493  */
1494 static inline int xprt_bindresvport(struct rpc_xprt *xprt, struct socket *sock)
1495 {
1496         struct sockaddr_in myaddr = {
1497                 .sin_family = AF_INET,
1498         };
1499         int             err, port;
1500
1501         /* Were we already bound to a given port? Try to reuse it */
1502         port = xprt->port;
1503         do {
1504                 myaddr.sin_port = htons(port);
1505                 err = sock->ops->bind(sock, (struct sockaddr *) &myaddr,
1506                                                 sizeof(myaddr));
1507                 if (err == 0) {
1508                         xprt->port = port;
1509                         return 0;
1510                 }
1511                 if (--port == 0)
1512                         port = XPRT_MAX_RESVPORT;
1513         } while (err == -EADDRINUSE && port != xprt->port);
1514
1515         printk("RPC: Can't bind to reserved port (%d).\n", -err);
1516         return err;
1517 }
1518
1519 static void
1520 xprt_bind_socket(struct rpc_xprt *xprt, struct socket *sock)
1521 {
1522         struct sock     *sk = sock->sk;
1523
1524         if (xprt->inet)
1525                 return;
1526
1527         write_lock_bh(&sk->sk_callback_lock);
1528         sk->sk_user_data = xprt;
1529         xprt->old_data_ready = sk->sk_data_ready;
1530         xprt->old_state_change = sk->sk_state_change;
1531         xprt->old_write_space = sk->sk_write_space;
1532         if (xprt->prot == IPPROTO_UDP) {
1533                 sk->sk_data_ready = udp_data_ready;
1534                 sk->sk_no_check = UDP_CSUM_NORCV;
1535                 xprt_set_connected(xprt);
1536         } else {
1537                 tcp_sk(sk)->nonagle = 1;        /* disable Nagle's algorithm */
1538                 sk->sk_data_ready = tcp_data_ready;
1539                 sk->sk_state_change = tcp_state_change;
1540                 xprt_clear_connected(xprt);
1541         }
1542         sk->sk_write_space = xprt_write_space;
1543
1544         /* Reset to new socket */
1545         xprt->sock = sock;
1546         xprt->inet = sk;
1547         write_unlock_bh(&sk->sk_callback_lock);
1548
1549         return;
1550 }
1551
1552 /*
1553  * Set socket buffer length
1554  */
1555 void
1556 xprt_sock_setbufsize(struct rpc_xprt *xprt)
1557 {
1558         struct sock *sk = xprt->inet;
1559
1560         if (xprt->stream)
1561                 return;
1562         if (xprt->rcvsize) {
1563                 sk->sk_userlocks |= SOCK_RCVBUF_LOCK;
1564                 sk->sk_rcvbuf = xprt->rcvsize * xprt->max_reqs *  2;
1565         }
1566         if (xprt->sndsize) {
1567                 sk->sk_userlocks |= SOCK_SNDBUF_LOCK;
1568                 sk->sk_sndbuf = xprt->sndsize * xprt->max_reqs * 2;
1569                 sk->sk_write_space(sk);
1570         }
1571 }
1572
1573 /*
1574  * Datastream sockets are created here, but xprt_connect will create
1575  * and connect stream sockets.
1576  */
1577 static struct socket * xprt_create_socket(struct rpc_xprt *xprt, int proto, int resvport)
1578 {
1579         struct socket   *sock;
1580         int             type, err;
1581
1582         dprintk("RPC:      xprt_create_socket(%s %d)\n",
1583                            (proto == IPPROTO_UDP)? "udp" : "tcp", proto);
1584
1585         type = (proto == IPPROTO_UDP)? SOCK_DGRAM : SOCK_STREAM;
1586
1587         if ((err = sock_create_kern(PF_INET, type, proto, &sock)) < 0) {
1588                 printk("RPC: can't create socket (%d).\n", -err);
1589                 return NULL;
1590         }
1591
1592         /* If the caller has the capability, bind to a reserved port */
1593         if (resvport && xprt_bindresvport(xprt, sock) < 0) {
1594                 printk("RPC: can't bind to reserved port.\n");
1595                 goto failed;
1596         }
1597
1598         return sock;
1599
1600 failed:
1601         sock_release(sock);
1602         return NULL;
1603 }
1604
1605 /*
1606  * Create an RPC client transport given the protocol and peer address.
1607  */
1608 struct rpc_xprt *
1609 xprt_create_proto(int proto, struct sockaddr_in *sap, struct rpc_timeout *to)
1610 {
1611         struct rpc_xprt *xprt;
1612
1613         xprt = xprt_setup(proto, sap, to);
1614         if (IS_ERR(xprt))
1615                 dprintk("RPC:      xprt_create_proto failed\n");
1616         else
1617                 dprintk("RPC:      xprt_create_proto created xprt %p\n", xprt);
1618         return xprt;
1619 }
1620
1621 /*
1622  * Prepare for transport shutdown.
1623  */
1624 static void
1625 xprt_shutdown(struct rpc_xprt *xprt)
1626 {
1627         xprt->shutdown = 1;
1628         rpc_wake_up(&xprt->sending);
1629         rpc_wake_up(&xprt->resend);
1630         rpc_wake_up(&xprt->pending);
1631         rpc_wake_up(&xprt->backlog);
1632         wake_up(&xprt->cong_wait);
1633         del_timer_sync(&xprt->timer);
1634
1635         /* synchronously wait for connect worker to finish */
1636         cancel_delayed_work(&xprt->sock_connect);
1637         flush_scheduled_work();
1638 }
1639
1640 /*
1641  * Clear the xprt backlog queue
1642  */
1643 static int
1644 xprt_clear_backlog(struct rpc_xprt *xprt) {
1645         rpc_wake_up_next(&xprt->backlog);
1646         wake_up(&xprt->cong_wait);
1647         return 1;
1648 }
1649
1650 /*
1651  * Destroy an RPC transport, killing off all requests.
1652  */
1653 int
1654 xprt_destroy(struct rpc_xprt *xprt)
1655 {
1656         dprintk("RPC:      destroying transport %p\n", xprt);
1657         xprt_shutdown(xprt);
1658         xprt_disconnect(xprt);
1659         xprt_close(xprt);
1660         kfree(xprt->slot);
1661         kfree(xprt);
1662
1663         return 0;
1664 }