xprt.c 41.5 KB
Newer Older
L
Linus Torvalds 已提交
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727
/*
 *  linux/net/sunrpc/xprt.c
 *
 *  This is a generic RPC call interface supporting congestion avoidance,
 *  and asynchronous calls.
 *
 *  The interface works like this:
 *
 *  -	When a process places a call, it allocates a request slot if
 *	one is available. Otherwise, it sleeps on the backlog queue
 *	(xprt_reserve).
 *  -	Next, the caller puts together the RPC message, stuffs it into
 *	the request struct, and calls xprt_call().
 *  -	xprt_call transmits the message and installs the caller on the
 *	socket's wait list. At the same time, it installs a timer that
 *	is run after the packet's timeout has expired.
 *  -	When a packet arrives, the data_ready handler walks the list of
 *	pending requests for that socket. If a matching XID is found, the
 *	caller is woken up, and the timer removed.
 *  -	When no reply arrives within the timeout interval, the timer is
 *	fired by the kernel and runs xprt_timer(). It either adjusts the
 *	timeout values (minor timeout) or wakes up the caller with a status
 *	of -ETIMEDOUT.
 *  -	When the caller receives a notification from RPC that a reply arrived,
 *	it should release the RPC slot, and process the reply.
 *	If the call timed out, it may choose to retry the operation by
 *	adjusting the initial timeout value, and simply calling rpc_call
 *	again.
 *
 *  Support for async RPC is done through a set of RPC-specific scheduling
 *  primitives that `transparently' work for processes as well as async
 *  tasks that rely on callbacks.
 *
 *  Copyright (C) 1995-1997, Olaf Kirch <okir@monad.swb.de>
 *
 *  TCP callback races fixes (C) 1998 Red Hat Software <alan@redhat.com>
 *  TCP send fixes (C) 1998 Red Hat Software <alan@redhat.com>
 *  TCP NFS related read + write fixes
 *   (C) 1999 Dave Airlie, University of Limerick, Ireland <airlied@linux.ie>
 *
 *  Rewrite of larges part of the code in order to stabilize TCP stuff.
 *  Fix behaviour when socket buffer is full.
 *   (C) 1999 Trond Myklebust <trond.myklebust@fys.uio.no>
 */

#include <linux/types.h>
#include <linux/slab.h>
#include <linux/capability.h>
#include <linux/sched.h>
#include <linux/errno.h>
#include <linux/socket.h>
#include <linux/in.h>
#include <linux/net.h>
#include <linux/mm.h>
#include <linux/udp.h>
#include <linux/tcp.h>
#include <linux/sunrpc/clnt.h>
#include <linux/file.h>
#include <linux/workqueue.h>
#include <linux/random.h>

#include <net/sock.h>
#include <net/checksum.h>
#include <net/udp.h>
#include <net/tcp.h>

/*
 * Local variables
 */

#ifdef RPC_DEBUG
# undef  RPC_DEBUG_DATA
# define RPCDBG_FACILITY	RPCDBG_XPRT
#endif

#define XPRT_MAX_BACKOFF	(8)
#define XPRT_IDLE_TIMEOUT	(5*60*HZ)
#define XPRT_MAX_RESVPORT	(800)

/*
 * Local functions
 */
static void	xprt_request_init(struct rpc_task *, struct rpc_xprt *);
static inline void	do_xprt_reserve(struct rpc_task *);
static void	xprt_disconnect(struct rpc_xprt *);
static void	xprt_connect_status(struct rpc_task *task);
static struct rpc_xprt * xprt_setup(int proto, struct sockaddr_in *ap,
						struct rpc_timeout *to);
static struct socket *xprt_create_socket(struct rpc_xprt *, int, int);
static void	xprt_bind_socket(struct rpc_xprt *, struct socket *);
static int      __xprt_get_cong(struct rpc_xprt *, struct rpc_task *);

static int	xprt_clear_backlog(struct rpc_xprt *xprt);

#ifdef RPC_DEBUG_DATA
/*
 * Print the buffer contents (first 128 bytes only--just enough for
 * diropres return).
 */
static void
xprt_pktdump(char *msg, u32 *packet, unsigned int count)
{
	u8	*buf = (u8 *) packet;
	int	j;

	dprintk("RPC:      %s\n", msg);
	for (j = 0; j < count && j < 128; j += 4) {
		if (!(j & 31)) {
			if (j)
				dprintk("\n");
			dprintk("0x%04x ", j);
		}
		dprintk("%02x%02x%02x%02x ",
			buf[j], buf[j+1], buf[j+2], buf[j+3]);
	}
	dprintk("\n");
}
#else
static inline void
xprt_pktdump(char *msg, u32 *packet, unsigned int count)
{
	/* NOP */
}
#endif

/*
 * Look up RPC transport given an INET socket
 */
static inline struct rpc_xprt *
xprt_from_sock(struct sock *sk)
{
	return (struct rpc_xprt *) sk->sk_user_data;
}

/*
 * Serialize write access to sockets, in order to prevent different
 * requests from interfering with each other.
 * Also prevents TCP socket connects from colliding with writes.
 */
static int
__xprt_lock_write(struct rpc_xprt *xprt, struct rpc_task *task)
{
	struct rpc_rqst *req = task->tk_rqstp;

	if (test_and_set_bit(XPRT_LOCKED, &xprt->sockstate)) {
		if (task == xprt->snd_task)
			return 1;
		if (task == NULL)
			return 0;
		goto out_sleep;
	}
	if (xprt->nocong || __xprt_get_cong(xprt, task)) {
		xprt->snd_task = task;
		if (req) {
			req->rq_bytes_sent = 0;
			req->rq_ntrans++;
		}
		return 1;
	}
	smp_mb__before_clear_bit();
	clear_bit(XPRT_LOCKED, &xprt->sockstate);
	smp_mb__after_clear_bit();
out_sleep:
	dprintk("RPC: %4d failed to lock socket %p\n", task->tk_pid, xprt);
	task->tk_timeout = 0;
	task->tk_status = -EAGAIN;
	if (req && req->rq_ntrans)
		rpc_sleep_on(&xprt->resend, task, NULL, NULL);
	else
		rpc_sleep_on(&xprt->sending, task, NULL, NULL);
	return 0;
}

static inline int
xprt_lock_write(struct rpc_xprt *xprt, struct rpc_task *task)
{
	int retval;

	spin_lock_bh(&xprt->sock_lock);
	retval = __xprt_lock_write(xprt, task);
	spin_unlock_bh(&xprt->sock_lock);
	return retval;
}


static void
__xprt_lock_write_next(struct rpc_xprt *xprt)
{
	struct rpc_task *task;

	if (test_and_set_bit(XPRT_LOCKED, &xprt->sockstate))
		return;
	if (!xprt->nocong && RPCXPRT_CONGESTED(xprt))
		goto out_unlock;
	task = rpc_wake_up_next(&xprt->resend);
	if (!task) {
		task = rpc_wake_up_next(&xprt->sending);
		if (!task)
			goto out_unlock;
	}
	if (xprt->nocong || __xprt_get_cong(xprt, task)) {
		struct rpc_rqst *req = task->tk_rqstp;
		xprt->snd_task = task;
		if (req) {
			req->rq_bytes_sent = 0;
			req->rq_ntrans++;
		}
		return;
	}
out_unlock:
	smp_mb__before_clear_bit();
	clear_bit(XPRT_LOCKED, &xprt->sockstate);
	smp_mb__after_clear_bit();
}

/*
 * Releases the socket for use by other requests.
 */
static void
__xprt_release_write(struct rpc_xprt *xprt, struct rpc_task *task)
{
	if (xprt->snd_task == task) {
		xprt->snd_task = NULL;
		smp_mb__before_clear_bit();
		clear_bit(XPRT_LOCKED, &xprt->sockstate);
		smp_mb__after_clear_bit();
		__xprt_lock_write_next(xprt);
	}
}

static inline void
xprt_release_write(struct rpc_xprt *xprt, struct rpc_task *task)
{
	spin_lock_bh(&xprt->sock_lock);
	__xprt_release_write(xprt, task);
	spin_unlock_bh(&xprt->sock_lock);
}

/*
 * Write data to socket.
 */
static inline int
xprt_sendmsg(struct rpc_xprt *xprt, struct rpc_rqst *req)
{
	struct socket	*sock = xprt->sock;
	struct xdr_buf	*xdr = &req->rq_snd_buf;
	struct sockaddr *addr = NULL;
	int addrlen = 0;
	unsigned int	skip;
	int		result;

	if (!sock)
		return -ENOTCONN;

	xprt_pktdump("packet data:",
				req->rq_svec->iov_base,
				req->rq_svec->iov_len);

	/* For UDP, we need to provide an address */
	if (!xprt->stream) {
		addr = (struct sockaddr *) &xprt->addr;
		addrlen = sizeof(xprt->addr);
	}
	/* Dont repeat bytes */
	skip = req->rq_bytes_sent;

	clear_bit(SOCK_ASYNC_NOSPACE, &sock->flags);
	result = xdr_sendpages(sock, addr, addrlen, xdr, skip, MSG_DONTWAIT);

	dprintk("RPC:      xprt_sendmsg(%d) = %d\n", xdr->len - skip, result);

	if (result >= 0)
		return result;

	switch (result) {
	case -ECONNREFUSED:
		/* When the server has died, an ICMP port unreachable message
		 * prompts ECONNREFUSED.
		 */
	case -EAGAIN:
		break;
	case -ECONNRESET:
	case -ENOTCONN:
	case -EPIPE:
		/* connection broken */
		if (xprt->stream)
			result = -ENOTCONN;
		break;
	default:
		printk(KERN_NOTICE "RPC: sendmsg returned error %d\n", -result);
	}
	return result;
}

/*
 * Van Jacobson congestion avoidance. Check if the congestion window
 * overflowed. Put the task to sleep if this is the case.
 */
static int
__xprt_get_cong(struct rpc_xprt *xprt, struct rpc_task *task)
{
	struct rpc_rqst *req = task->tk_rqstp;

	if (req->rq_cong)
		return 1;
	dprintk("RPC: %4d xprt_cwnd_limited cong = %ld cwnd = %ld\n",
			task->tk_pid, xprt->cong, xprt->cwnd);
	if (RPCXPRT_CONGESTED(xprt))
		return 0;
	req->rq_cong = 1;
	xprt->cong += RPC_CWNDSCALE;
	return 1;
}

/*
 * Adjust the congestion window, and wake up the next task
 * that has been sleeping due to congestion
 */
static void
__xprt_put_cong(struct rpc_xprt *xprt, struct rpc_rqst *req)
{
	if (!req->rq_cong)
		return;
	req->rq_cong = 0;
	xprt->cong -= RPC_CWNDSCALE;
	__xprt_lock_write_next(xprt);
}

/*
 * Adjust RPC congestion window
 * We use a time-smoothed congestion estimator to avoid heavy oscillation.
 */
static void
xprt_adjust_cwnd(struct rpc_xprt *xprt, int result)
{
	unsigned long	cwnd;

	cwnd = xprt->cwnd;
	if (result >= 0 && cwnd <= xprt->cong) {
		/* The (cwnd >> 1) term makes sure
		 * the result gets rounded properly. */
		cwnd += (RPC_CWNDSCALE * RPC_CWNDSCALE + (cwnd >> 1)) / cwnd;
		if (cwnd > RPC_MAXCWND(xprt))
			cwnd = RPC_MAXCWND(xprt);
		__xprt_lock_write_next(xprt);
	} else if (result == -ETIMEDOUT) {
		cwnd >>= 1;
		if (cwnd < RPC_CWNDSCALE)
			cwnd = RPC_CWNDSCALE;
	}
	dprintk("RPC:      cong %ld, cwnd was %ld, now %ld\n",
			xprt->cong, xprt->cwnd, cwnd);
	xprt->cwnd = cwnd;
}

/*
 * Reset the major timeout value
 */
static void xprt_reset_majortimeo(struct rpc_rqst *req)
{
	struct rpc_timeout *to = &req->rq_xprt->timeout;

	req->rq_majortimeo = req->rq_timeout;
	if (to->to_exponential)
		req->rq_majortimeo <<= to->to_retries;
	else
		req->rq_majortimeo += to->to_increment * to->to_retries;
	if (req->rq_majortimeo > to->to_maxval || req->rq_majortimeo == 0)
		req->rq_majortimeo = to->to_maxval;
	req->rq_majortimeo += jiffies;
}

/*
 * Adjust timeout values etc for next retransmit
 */
int xprt_adjust_timeout(struct rpc_rqst *req)
{
	struct rpc_xprt *xprt = req->rq_xprt;
	struct rpc_timeout *to = &xprt->timeout;
	int status = 0;

	if (time_before(jiffies, req->rq_majortimeo)) {
		if (to->to_exponential)
			req->rq_timeout <<= 1;
		else
			req->rq_timeout += to->to_increment;
		if (to->to_maxval && req->rq_timeout >= to->to_maxval)
			req->rq_timeout = to->to_maxval;
		req->rq_retries++;
		pprintk("RPC: %lu retrans\n", jiffies);
	} else {
		req->rq_timeout = to->to_initval;
		req->rq_retries = 0;
		xprt_reset_majortimeo(req);
		/* Reset the RTT counters == "slow start" */
		spin_lock_bh(&xprt->sock_lock);
		rpc_init_rtt(req->rq_task->tk_client->cl_rtt, to->to_initval);
		spin_unlock_bh(&xprt->sock_lock);
		pprintk("RPC: %lu timeout\n", jiffies);
		status = -ETIMEDOUT;
	}

	if (req->rq_timeout == 0) {
		printk(KERN_WARNING "xprt_adjust_timeout: rq_timeout = 0!\n");
		req->rq_timeout = 5 * HZ;
	}
	return status;
}

/*
 * Close down a transport socket
 */
static void
xprt_close(struct rpc_xprt *xprt)
{
	struct socket	*sock = xprt->sock;
	struct sock	*sk = xprt->inet;

	if (!sk)
		return;

	write_lock_bh(&sk->sk_callback_lock);
	xprt->inet = NULL;
	xprt->sock = NULL;

	sk->sk_user_data    = NULL;
	sk->sk_data_ready   = xprt->old_data_ready;
	sk->sk_state_change = xprt->old_state_change;
	sk->sk_write_space  = xprt->old_write_space;
	write_unlock_bh(&sk->sk_callback_lock);

	sk->sk_no_check	 = 0;

	sock_release(sock);
}

static void
xprt_socket_autoclose(void *args)
{
	struct rpc_xprt *xprt = (struct rpc_xprt *)args;

	xprt_disconnect(xprt);
	xprt_close(xprt);
	xprt_release_write(xprt, NULL);
}

/*
 * Mark a transport as disconnected
 */
static void
xprt_disconnect(struct rpc_xprt *xprt)
{
	dprintk("RPC:      disconnected transport %p\n", xprt);
	spin_lock_bh(&xprt->sock_lock);
	xprt_clear_connected(xprt);
	rpc_wake_up_status(&xprt->pending, -ENOTCONN);
	spin_unlock_bh(&xprt->sock_lock);
}

/*
 * Used to allow disconnection when we've been idle
 */
static void
xprt_init_autodisconnect(unsigned long data)
{
	struct rpc_xprt *xprt = (struct rpc_xprt *)data;

	spin_lock(&xprt->sock_lock);
	if (!list_empty(&xprt->recv) || xprt->shutdown)
		goto out_abort;
	if (test_and_set_bit(XPRT_LOCKED, &xprt->sockstate))
		goto out_abort;
	spin_unlock(&xprt->sock_lock);
	/* Let keventd close the socket */
	if (test_bit(XPRT_CONNECTING, &xprt->sockstate) != 0)
		xprt_release_write(xprt, NULL);
	else
		schedule_work(&xprt->task_cleanup);
	return;
out_abort:
	spin_unlock(&xprt->sock_lock);
}

static void xprt_socket_connect(void *args)
{
	struct rpc_xprt *xprt = (struct rpc_xprt *)args;
	struct socket *sock = xprt->sock;
	int status = -EIO;

	if (xprt->shutdown || xprt->addr.sin_port == 0)
		goto out;

	/*
	 * Start by resetting any existing state
	 */
	xprt_close(xprt);
	sock = xprt_create_socket(xprt, xprt->prot, xprt->resvport);
	if (sock == NULL) {
		/* couldn't create socket or bind to reserved port;
		 * this is likely a permanent error, so cause an abort */
		goto out;
	}
	xprt_bind_socket(xprt, sock);
	xprt_sock_setbufsize(xprt);

	status = 0;
	if (!xprt->stream)
		goto out;

	/*
	 * Tell the socket layer to start connecting...
	 */
	status = sock->ops->connect(sock, (struct sockaddr *) &xprt->addr,
			sizeof(xprt->addr), O_NONBLOCK);
	dprintk("RPC: %p  connect status %d connected %d sock state %d\n",
			xprt, -status, xprt_connected(xprt), sock->sk->sk_state);
	if (status < 0) {
		switch (status) {
			case -EINPROGRESS:
			case -EALREADY:
				goto out_clear;
		}
	}
out:
	if (status < 0)
		rpc_wake_up_status(&xprt->pending, status);
	else
		rpc_wake_up(&xprt->pending);
out_clear:
	smp_mb__before_clear_bit();
	clear_bit(XPRT_CONNECTING, &xprt->sockstate);
	smp_mb__after_clear_bit();
}

/*
 * Attempt to connect a TCP socket.
 *
 */
void xprt_connect(struct rpc_task *task)
{
	struct rpc_xprt	*xprt = task->tk_xprt;

	dprintk("RPC: %4d xprt_connect xprt %p %s connected\n", task->tk_pid,
			xprt, (xprt_connected(xprt) ? "is" : "is not"));

	if (xprt->shutdown) {
		task->tk_status = -EIO;
		return;
	}
	if (!xprt->addr.sin_port) {
		task->tk_status = -EIO;
		return;
	}
	if (!xprt_lock_write(xprt, task))
		return;
	if (xprt_connected(xprt))
		goto out_write;

	if (task->tk_rqstp)
		task->tk_rqstp->rq_bytes_sent = 0;

	task->tk_timeout = RPC_CONNECT_TIMEOUT;
	rpc_sleep_on(&xprt->pending, task, xprt_connect_status, NULL);
	if (!test_and_set_bit(XPRT_CONNECTING, &xprt->sockstate)) {
		/* Note: if we are here due to a dropped connection
		 * 	 we delay reconnecting by RPC_REESTABLISH_TIMEOUT/HZ
		 * 	 seconds
		 */
		if (xprt->sock != NULL)
			schedule_delayed_work(&xprt->sock_connect,
					RPC_REESTABLISH_TIMEOUT);
		else
			schedule_work(&xprt->sock_connect);
	}
	return;
 out_write:
	xprt_release_write(xprt, task);
}

/*
 * We arrive here when awoken from waiting on connection establishment.
 */
static void
xprt_connect_status(struct rpc_task *task)
{
	struct rpc_xprt	*xprt = task->tk_xprt;

	if (task->tk_status >= 0) {
		dprintk("RPC: %4d xprt_connect_status: connection established\n",
				task->tk_pid);
		return;
	}

	/* if soft mounted, just cause this RPC to fail */
	if (RPC_IS_SOFT(task))
		task->tk_status = -EIO;

	switch (task->tk_status) {
	case -ECONNREFUSED:
	case -ECONNRESET:
	case -ENOTCONN:
		return;
	case -ETIMEDOUT:
		dprintk("RPC: %4d xprt_connect_status: timed out\n",
				task->tk_pid);
		break;
	default:
		printk(KERN_ERR "RPC: error %d connecting to server %s\n",
				-task->tk_status, task->tk_client->cl_server);
	}
	xprt_release_write(xprt, task);
}

/*
 * Look up the RPC request corresponding to a reply, and then lock it.
 */
static inline struct rpc_rqst *
xprt_lookup_rqst(struct rpc_xprt *xprt, u32 xid)
{
	struct list_head *pos;
	struct rpc_rqst	*req = NULL;

	list_for_each(pos, &xprt->recv) {
		struct rpc_rqst *entry = list_entry(pos, struct rpc_rqst, rq_list);
		if (entry->rq_xid == xid) {
			req = entry;
			break;
		}
	}
	return req;
}

/*
 * Complete reply received.
 * The TCP code relies on us to remove the request from xprt->pending.
 */
static void
xprt_complete_rqst(struct rpc_xprt *xprt, struct rpc_rqst *req, int copied)
{
	struct rpc_task	*task = req->rq_task;
	struct rpc_clnt *clnt = task->tk_client;

	/* Adjust congestion window */
	if (!xprt->nocong) {
		unsigned timer = task->tk_msg.rpc_proc->p_timer;
		xprt_adjust_cwnd(xprt, copied);
		__xprt_put_cong(xprt, req);
		if (timer) {
			if (req->rq_ntrans == 1)
				rpc_update_rtt(clnt->cl_rtt, timer,
						(long)jiffies - req->rq_xtime);
			rpc_set_timeo(clnt->cl_rtt, timer, req->rq_ntrans - 1);
		}
	}

#ifdef RPC_PROFILE
	/* Profile only reads for now */
	if (copied > 1024) {
		static unsigned long	nextstat;
		static unsigned long	pkt_rtt, pkt_len, pkt_cnt;

		pkt_cnt++;
		pkt_len += req->rq_slen + copied;
		pkt_rtt += jiffies - req->rq_xtime;
		if (time_before(nextstat, jiffies)) {
			printk("RPC: %lu %ld cwnd\n", jiffies, xprt->cwnd);
			printk("RPC: %ld %ld %ld %ld stat\n",
					jiffies, pkt_cnt, pkt_len, pkt_rtt);
			pkt_rtt = pkt_len = pkt_cnt = 0;
			nextstat = jiffies + 5 * HZ;
		}
	}
#endif

	dprintk("RPC: %4d has input (%d bytes)\n", task->tk_pid, copied);
	list_del_init(&req->rq_list);
	req->rq_received = req->rq_private_buf.len = copied;

	/* ... and wake up the process. */
	rpc_wake_up_task(task);
	return;
}

static size_t
skb_read_bits(skb_reader_t *desc, void *to, size_t len)
{
	if (len > desc->count)
		len = desc->count;
	if (skb_copy_bits(desc->skb, desc->offset, to, len))
		return 0;
	desc->count -= len;
	desc->offset += len;
	return len;
}

static size_t
skb_read_and_csum_bits(skb_reader_t *desc, void *to, size_t len)
{
	unsigned int csum2, pos;

	if (len > desc->count)
		len = desc->count;
	pos = desc->offset;
	csum2 = skb_copy_and_csum_bits(desc->skb, pos, to, len, 0);
	desc->csum = csum_block_add(desc->csum, csum2, pos);
	desc->count -= len;
	desc->offset += len;
	return len;
}

/*
 * We have set things up such that we perform the checksum of the UDP
 * packet in parallel with the copies into the RPC client iovec.  -DaveM
 */
int
csum_partial_copy_to_xdr(struct xdr_buf *xdr, struct sk_buff *skb)
{
	skb_reader_t desc;

	desc.skb = skb;
	desc.offset = sizeof(struct udphdr);
	desc.count = skb->len - desc.offset;

	if (skb->ip_summed == CHECKSUM_UNNECESSARY)
		goto no_checksum;

	desc.csum = csum_partial(skb->data, desc.offset, skb->csum);
728 729
	if (xdr_partial_copy_from_skb(xdr, 0, &desc, skb_read_and_csum_bits) < 0)
		return -1;
L
Linus Torvalds 已提交
730 731 732 733 734 735 736 737 738 739 740
	if (desc.offset != skb->len) {
		unsigned int csum2;
		csum2 = skb_checksum(skb, desc.offset, skb->len - desc.offset, 0);
		desc.csum = csum_block_add(desc.csum, csum2, desc.offset);
	}
	if (desc.count)
		return -1;
	if ((unsigned short)csum_fold(desc.csum))
		return -1;
	return 0;
no_checksum:
741 742
	if (xdr_partial_copy_from_skb(xdr, 0, &desc, skb_read_bits) < 0)
		return -1;
L
Linus Torvalds 已提交
743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825
	if (desc.count)
		return -1;
	return 0;
}

/*
 * Input handler for RPC replies. Called from a bottom half and hence
 * atomic.
 */
static void
udp_data_ready(struct sock *sk, int len)
{
	struct rpc_task	*task;
	struct rpc_xprt	*xprt;
	struct rpc_rqst *rovr;
	struct sk_buff	*skb;
	int err, repsize, copied;
	u32 _xid, *xp;

	read_lock(&sk->sk_callback_lock);
	dprintk("RPC:      udp_data_ready...\n");
	if (!(xprt = xprt_from_sock(sk))) {
		printk("RPC:      udp_data_ready request not found!\n");
		goto out;
	}

	dprintk("RPC:      udp_data_ready client %p\n", xprt);

	if ((skb = skb_recv_datagram(sk, 0, 1, &err)) == NULL)
		goto out;

	if (xprt->shutdown)
		goto dropit;

	repsize = skb->len - sizeof(struct udphdr);
	if (repsize < 4) {
		printk("RPC: impossible RPC reply size %d!\n", repsize);
		goto dropit;
	}

	/* Copy the XID from the skb... */
	xp = skb_header_pointer(skb, sizeof(struct udphdr),
				sizeof(_xid), &_xid);
	if (xp == NULL)
		goto dropit;

	/* Look up and lock the request corresponding to the given XID */
	spin_lock(&xprt->sock_lock);
	rovr = xprt_lookup_rqst(xprt, *xp);
	if (!rovr)
		goto out_unlock;
	task = rovr->rq_task;

	dprintk("RPC: %4d received reply\n", task->tk_pid);

	if ((copied = rovr->rq_private_buf.buflen) > repsize)
		copied = repsize;

	/* Suck it into the iovec, verify checksum if not done by hw. */
	if (csum_partial_copy_to_xdr(&rovr->rq_private_buf, skb))
		goto out_unlock;

	/* Something worked... */
	dst_confirm(skb->dst);

	xprt_complete_rqst(xprt, rovr, copied);

 out_unlock:
	spin_unlock(&xprt->sock_lock);
 dropit:
	skb_free_datagram(sk, skb);
 out:
	read_unlock(&sk->sk_callback_lock);
}

/*
 * Copy from an skb into memory and shrink the skb.
 */
static inline size_t
tcp_copy_data(skb_reader_t *desc, void *p, size_t len)
{
	if (len > desc->count)
		len = desc->count;
826 827 828
	if (skb_copy_bits(desc->skb, desc->offset, p, len)) {
		dprintk("RPC:      failed to copy %zu bytes from skb. %zu bytes remain\n",
				len, desc->count);
L
Linus Torvalds 已提交
829
		return 0;
830
	}
L
Linus Torvalds 已提交
831 832
	desc->offset += len;
	desc->count -= len;
833 834
	dprintk("RPC:      copied %zu bytes from skb. %zu bytes remain\n",
			len, desc->count);
L
Linus Torvalds 已提交
835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872
	return len;
}

/*
 * TCP read fragment marker
 */
static inline void
tcp_read_fraghdr(struct rpc_xprt *xprt, skb_reader_t *desc)
{
	size_t len, used;
	char *p;

	p = ((char *) &xprt->tcp_recm) + xprt->tcp_offset;
	len = sizeof(xprt->tcp_recm) - xprt->tcp_offset;
	used = tcp_copy_data(desc, p, len);
	xprt->tcp_offset += used;
	if (used != len)
		return;
	xprt->tcp_reclen = ntohl(xprt->tcp_recm);
	if (xprt->tcp_reclen & 0x80000000)
		xprt->tcp_flags |= XPRT_LAST_FRAG;
	else
		xprt->tcp_flags &= ~XPRT_LAST_FRAG;
	xprt->tcp_reclen &= 0x7fffffff;
	xprt->tcp_flags &= ~XPRT_COPY_RECM;
	xprt->tcp_offset = 0;
	/* Sanity check of the record length */
	if (xprt->tcp_reclen < 4) {
		printk(KERN_ERR "RPC: Invalid TCP record fragment length\n");
		xprt_disconnect(xprt);
	}
	dprintk("RPC:      reading TCP record fragment of length %d\n",
			xprt->tcp_reclen);
}

static void
tcp_check_recm(struct rpc_xprt *xprt)
{
873 874
	dprintk("RPC:      xprt = %p, tcp_copied = %lu, tcp_offset = %u, tcp_reclen = %u, tcp_flags = %lx\n",
			xprt, xprt->tcp_copied, xprt->tcp_offset, xprt->tcp_reclen, xprt->tcp_flags);
L
Linus Torvalds 已提交
875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918
	if (xprt->tcp_offset == xprt->tcp_reclen) {
		xprt->tcp_flags |= XPRT_COPY_RECM;
		xprt->tcp_offset = 0;
		if (xprt->tcp_flags & XPRT_LAST_FRAG) {
			xprt->tcp_flags &= ~XPRT_COPY_DATA;
			xprt->tcp_flags |= XPRT_COPY_XID;
			xprt->tcp_copied = 0;
		}
	}
}

/*
 * TCP read xid
 */
static inline void
tcp_read_xid(struct rpc_xprt *xprt, skb_reader_t *desc)
{
	size_t len, used;
	char *p;

	len = sizeof(xprt->tcp_xid) - xprt->tcp_offset;
	dprintk("RPC:      reading XID (%Zu bytes)\n", len);
	p = ((char *) &xprt->tcp_xid) + xprt->tcp_offset;
	used = tcp_copy_data(desc, p, len);
	xprt->tcp_offset += used;
	if (used != len)
		return;
	xprt->tcp_flags &= ~XPRT_COPY_XID;
	xprt->tcp_flags |= XPRT_COPY_DATA;
	xprt->tcp_copied = 4;
	dprintk("RPC:      reading reply for XID %08x\n",
						ntohl(xprt->tcp_xid));
	tcp_check_recm(xprt);
}

/*
 * TCP read and complete request
 */
static inline void
tcp_read_request(struct rpc_xprt *xprt, skb_reader_t *desc)
{
	struct rpc_rqst *req;
	struct xdr_buf *rcvbuf;
	size_t len;
919
	ssize_t r;
L
Linus Torvalds 已提交
920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939

	/* Find and lock the request corresponding to this xid */
	spin_lock(&xprt->sock_lock);
	req = xprt_lookup_rqst(xprt, xprt->tcp_xid);
	if (!req) {
		xprt->tcp_flags &= ~XPRT_COPY_DATA;
		dprintk("RPC:      XID %08x request not found!\n",
				ntohl(xprt->tcp_xid));
		spin_unlock(&xprt->sock_lock);
		return;
	}

	rcvbuf = &req->rq_private_buf;
	len = desc->count;
	if (len > xprt->tcp_reclen - xprt->tcp_offset) {
		skb_reader_t my_desc;

		len = xprt->tcp_reclen - xprt->tcp_offset;
		memcpy(&my_desc, desc, sizeof(my_desc));
		my_desc.count = len;
940
		r = xdr_partial_copy_from_skb(rcvbuf, xprt->tcp_copied,
L
Linus Torvalds 已提交
941
					  &my_desc, tcp_copy_data);
942 943
		desc->count -= r;
		desc->offset += r;
L
Linus Torvalds 已提交
944
	} else
945
		r = xdr_partial_copy_from_skb(rcvbuf, xprt->tcp_copied,
L
Linus Torvalds 已提交
946 947
					  desc, tcp_copy_data);

948 949 950 951 952
	if (r > 0) {
		xprt->tcp_copied += r;
		xprt->tcp_offset += r;
	}
	if (r != len) {
953 954 955 956 957 958 959 960 961 962
		/* Error when copying to the receive buffer,
		 * usually because we weren't able to allocate
		 * additional buffer pages. All we can do now
		 * is turn off XPRT_COPY_DATA, so the request
		 * will not receive any additional updates,
		 * and time out.
		 * Any remaining data from this record will
		 * be discarded.
		 */
		xprt->tcp_flags &= ~XPRT_COPY_DATA;
963 964 965 966
		dprintk("RPC:      XID %08x truncated request\n",
				ntohl(xprt->tcp_xid));
		dprintk("RPC:      xprt = %p, tcp_copied = %lu, tcp_offset = %u, tcp_reclen = %u\n",
				xprt, xprt->tcp_copied, xprt->tcp_offset, xprt->tcp_reclen);
967 968 969
		goto out;
	}

970 971 972 973 974
	dprintk("RPC:      XID %08x read %u bytes\n",
			ntohl(xprt->tcp_xid), r);
	dprintk("RPC:      xprt = %p, tcp_copied = %lu, tcp_offset = %u, tcp_reclen = %u\n",
			xprt, xprt->tcp_copied, xprt->tcp_offset, xprt->tcp_reclen);

L
Linus Torvalds 已提交
975 976 977 978 979 980 981
	if (xprt->tcp_copied == req->rq_private_buf.buflen)
		xprt->tcp_flags &= ~XPRT_COPY_DATA;
	else if (xprt->tcp_offset == xprt->tcp_reclen) {
		if (xprt->tcp_flags & XPRT_LAST_FRAG)
			xprt->tcp_flags &= ~XPRT_COPY_DATA;
	}

982
out:
L
Linus Torvalds 已提交
983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005
	if (!(xprt->tcp_flags & XPRT_COPY_DATA)) {
		dprintk("RPC: %4d received reply complete\n",
				req->rq_task->tk_pid);
		xprt_complete_rqst(xprt, req, xprt->tcp_copied);
	}
	spin_unlock(&xprt->sock_lock);
	tcp_check_recm(xprt);
}

/*
 * TCP discard extra bytes from a short read
 */
static inline void
tcp_read_discard(struct rpc_xprt *xprt, skb_reader_t *desc)
{
	size_t len;

	len = xprt->tcp_reclen - xprt->tcp_offset;
	if (len > desc->count)
		len = desc->count;
	desc->count -= len;
	desc->offset += len;
	xprt->tcp_offset += len;
1006
	dprintk("RPC:      discarded %u bytes\n", len);
L
Linus Torvalds 已提交
1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400 1401 1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 1545 1546 1547 1548 1549 1550 1551 1552 1553 1554 1555 1556 1557 1558 1559 1560 1561 1562 1563 1564 1565 1566 1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 1636 1637 1638 1639 1640 1641 1642 1643 1644 1645 1646 1647 1648 1649 1650 1651 1652 1653 1654 1655 1656 1657 1658 1659 1660 1661 1662 1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680 1681 1682 1683 1684 1685 1686 1687 1688 1689 1690 1691 1692 1693 1694 1695 1696 1697 1698 1699 1700 1701 1702 1703 1704 1705 1706 1707 1708 1709 1710 1711 1712 1713 1714 1715
	tcp_check_recm(xprt);
}

/*
 * TCP record receive routine
 * We first have to grab the record marker, then the XID, then the data.
 */
static int
tcp_data_recv(read_descriptor_t *rd_desc, struct sk_buff *skb,
		unsigned int offset, size_t len)
{
	struct rpc_xprt *xprt = rd_desc->arg.data;
	skb_reader_t desc = {
		.skb	= skb,
		.offset	= offset,
		.count	= len,
		.csum	= 0
       	};

	dprintk("RPC:      tcp_data_recv\n");
	do {
		/* Read in a new fragment marker if necessary */
		/* Can we ever really expect to get completely empty fragments? */
		if (xprt->tcp_flags & XPRT_COPY_RECM) {
			tcp_read_fraghdr(xprt, &desc);
			continue;
		}
		/* Read in the xid if necessary */
		if (xprt->tcp_flags & XPRT_COPY_XID) {
			tcp_read_xid(xprt, &desc);
			continue;
		}
		/* Read in the request data */
		if (xprt->tcp_flags & XPRT_COPY_DATA) {
			tcp_read_request(xprt, &desc);
			continue;
		}
		/* Skip over any trailing bytes on short reads */
		tcp_read_discard(xprt, &desc);
	} while (desc.count);
	dprintk("RPC:      tcp_data_recv done\n");
	return len - desc.count;
}

static void tcp_data_ready(struct sock *sk, int bytes)
{
	struct rpc_xprt *xprt;
	read_descriptor_t rd_desc;

	read_lock(&sk->sk_callback_lock);
	dprintk("RPC:      tcp_data_ready...\n");
	if (!(xprt = xprt_from_sock(sk))) {
		printk("RPC:      tcp_data_ready socket info not found!\n");
		goto out;
	}
	if (xprt->shutdown)
		goto out;

	/* We use rd_desc to pass struct xprt to tcp_data_recv */
	rd_desc.arg.data = xprt;
	rd_desc.count = 65536;
	tcp_read_sock(sk, &rd_desc, tcp_data_recv);
out:
	read_unlock(&sk->sk_callback_lock);
}

static void
tcp_state_change(struct sock *sk)
{
	struct rpc_xprt	*xprt;

	read_lock(&sk->sk_callback_lock);
	if (!(xprt = xprt_from_sock(sk)))
		goto out;
	dprintk("RPC:      tcp_state_change client %p...\n", xprt);
	dprintk("RPC:      state %x conn %d dead %d zapped %d\n",
				sk->sk_state, xprt_connected(xprt),
				sock_flag(sk, SOCK_DEAD),
				sock_flag(sk, SOCK_ZAPPED));

	switch (sk->sk_state) {
	case TCP_ESTABLISHED:
		spin_lock_bh(&xprt->sock_lock);
		if (!xprt_test_and_set_connected(xprt)) {
			/* Reset TCP record info */
			xprt->tcp_offset = 0;
			xprt->tcp_reclen = 0;
			xprt->tcp_copied = 0;
			xprt->tcp_flags = XPRT_COPY_RECM | XPRT_COPY_XID;
			rpc_wake_up(&xprt->pending);
		}
		spin_unlock_bh(&xprt->sock_lock);
		break;
	case TCP_SYN_SENT:
	case TCP_SYN_RECV:
		break;
	default:
		if (xprt_test_and_clear_connected(xprt))
			rpc_wake_up_status(&xprt->pending, -ENOTCONN);
		break;
	}
 out:
	read_unlock(&sk->sk_callback_lock);
}

/*
 * Called when more output buffer space is available for this socket.
 * We try not to wake our writers until they can make "significant"
 * progress, otherwise we'll waste resources thrashing sock_sendmsg
 * with a bunch of small requests.
 */
static void
xprt_write_space(struct sock *sk)
{
	struct rpc_xprt	*xprt;
	struct socket	*sock;

	read_lock(&sk->sk_callback_lock);
	if (!(xprt = xprt_from_sock(sk)) || !(sock = sk->sk_socket))
		goto out;
	if (xprt->shutdown)
		goto out;

	/* Wait until we have enough socket memory */
	if (xprt->stream) {
		/* from net/core/stream.c:sk_stream_write_space */
		if (sk_stream_wspace(sk) < sk_stream_min_wspace(sk))
			goto out;
	} else {
		/* from net/core/sock.c:sock_def_write_space */
		if (!sock_writeable(sk))
			goto out;
	}

	if (!test_and_clear_bit(SOCK_NOSPACE, &sock->flags))
		goto out;

	spin_lock_bh(&xprt->sock_lock);
	if (xprt->snd_task)
		rpc_wake_up_task(xprt->snd_task);
	spin_unlock_bh(&xprt->sock_lock);
out:
	read_unlock(&sk->sk_callback_lock);
}

/*
 * RPC receive timeout handler.
 */
static void
xprt_timer(struct rpc_task *task)
{
	struct rpc_rqst	*req = task->tk_rqstp;
	struct rpc_xprt *xprt = req->rq_xprt;

	spin_lock(&xprt->sock_lock);
	if (req->rq_received)
		goto out;

	xprt_adjust_cwnd(req->rq_xprt, -ETIMEDOUT);
	__xprt_put_cong(xprt, req);

	dprintk("RPC: %4d xprt_timer (%s request)\n",
		task->tk_pid, req ? "pending" : "backlogged");

	task->tk_status  = -ETIMEDOUT;
out:
	task->tk_timeout = 0;
	rpc_wake_up_task(task);
	spin_unlock(&xprt->sock_lock);
}

/*
 * Place the actual RPC call.
 * We have to copy the iovec because sendmsg fiddles with its contents.
 */
int
xprt_prepare_transmit(struct rpc_task *task)
{
	struct rpc_rqst	*req = task->tk_rqstp;
	struct rpc_xprt	*xprt = req->rq_xprt;
	int err = 0;

	dprintk("RPC: %4d xprt_prepare_transmit\n", task->tk_pid);

	if (xprt->shutdown)
		return -EIO;

	spin_lock_bh(&xprt->sock_lock);
	if (req->rq_received && !req->rq_bytes_sent) {
		err = req->rq_received;
		goto out_unlock;
	}
	if (!__xprt_lock_write(xprt, task)) {
		err = -EAGAIN;
		goto out_unlock;
	}

	if (!xprt_connected(xprt)) {
		err = -ENOTCONN;
		goto out_unlock;
	}
out_unlock:
	spin_unlock_bh(&xprt->sock_lock);
	return err;
}

void
xprt_transmit(struct rpc_task *task)
{
	struct rpc_clnt *clnt = task->tk_client;
	struct rpc_rqst	*req = task->tk_rqstp;
	struct rpc_xprt	*xprt = req->rq_xprt;
	int status, retry = 0;


	dprintk("RPC: %4d xprt_transmit(%u)\n", task->tk_pid, req->rq_slen);

	/* set up everything as needed. */
	/* Write the record marker */
	if (xprt->stream) {
		u32	*marker = req->rq_svec[0].iov_base;

		*marker = htonl(0x80000000|(req->rq_slen-sizeof(*marker)));
	}

	smp_rmb();
	if (!req->rq_received) {
		if (list_empty(&req->rq_list)) {
			spin_lock_bh(&xprt->sock_lock);
			/* Update the softirq receive buffer */
			memcpy(&req->rq_private_buf, &req->rq_rcv_buf,
					sizeof(req->rq_private_buf));
			/* Add request to the receive list */
			list_add_tail(&req->rq_list, &xprt->recv);
			spin_unlock_bh(&xprt->sock_lock);
			xprt_reset_majortimeo(req);
		}
	} else if (!req->rq_bytes_sent)
		return;

	/* Continue transmitting the packet/record. We must be careful
	 * to cope with writespace callbacks arriving _after_ we have
	 * called xprt_sendmsg().
	 */
	while (1) {
		req->rq_xtime = jiffies;
		status = xprt_sendmsg(xprt, req);

		if (status < 0)
			break;

		if (xprt->stream) {
			req->rq_bytes_sent += status;

			/* If we've sent the entire packet, immediately
			 * reset the count of bytes sent. */
			if (req->rq_bytes_sent >= req->rq_slen) {
				req->rq_bytes_sent = 0;
				goto out_receive;
			}
		} else {
			if (status >= req->rq_slen)
				goto out_receive;
			status = -EAGAIN;
			break;
		}

		dprintk("RPC: %4d xmit incomplete (%d left of %d)\n",
				task->tk_pid, req->rq_slen - req->rq_bytes_sent,
				req->rq_slen);

		status = -EAGAIN;
		if (retry++ > 50)
			break;
	}

	/* Note: at this point, task->tk_sleeping has not yet been set,
	 *	 hence there is no danger of the waking up task being put on
	 *	 schedq, and being picked up by a parallel run of rpciod().
	 */
	task->tk_status = status;

	switch (status) {
	case -EAGAIN:
		if (test_bit(SOCK_ASYNC_NOSPACE, &xprt->sock->flags)) {
			/* Protect against races with xprt_write_space */
			spin_lock_bh(&xprt->sock_lock);
			/* Don't race with disconnect */
			if (!xprt_connected(xprt))
				task->tk_status = -ENOTCONN;
			else if (test_bit(SOCK_NOSPACE, &xprt->sock->flags)) {
				task->tk_timeout = req->rq_timeout;
				rpc_sleep_on(&xprt->pending, task, NULL, NULL);
			}
			spin_unlock_bh(&xprt->sock_lock);
			return;
		}
		/* Keep holding the socket if it is blocked */
		rpc_delay(task, HZ>>4);
		return;
	case -ECONNREFUSED:
		task->tk_timeout = RPC_REESTABLISH_TIMEOUT;
		rpc_sleep_on(&xprt->sending, task, NULL, NULL);
	case -ENOTCONN:
		return;
	default:
		if (xprt->stream)
			xprt_disconnect(xprt);
	}
	xprt_release_write(xprt, task);
	return;
 out_receive:
	dprintk("RPC: %4d xmit complete\n", task->tk_pid);
	/* Set the task's receive timeout value */
	spin_lock_bh(&xprt->sock_lock);
	if (!xprt->nocong) {
		int timer = task->tk_msg.rpc_proc->p_timer;
		task->tk_timeout = rpc_calc_rto(clnt->cl_rtt, timer);
		task->tk_timeout <<= rpc_ntimeo(clnt->cl_rtt, timer) + req->rq_retries;
		if (task->tk_timeout > xprt->timeout.to_maxval || task->tk_timeout == 0)
			task->tk_timeout = xprt->timeout.to_maxval;
	} else
		task->tk_timeout = req->rq_timeout;
	/* Don't race with disconnect */
	if (!xprt_connected(xprt))
		task->tk_status = -ENOTCONN;
	else if (!req->rq_received)
		rpc_sleep_on(&xprt->pending, task, NULL, xprt_timer);
	__xprt_release_write(xprt, task);
	spin_unlock_bh(&xprt->sock_lock);
}

/*
 * Reserve an RPC call slot.
 */
static inline void
do_xprt_reserve(struct rpc_task *task)
{
	struct rpc_xprt	*xprt = task->tk_xprt;

	task->tk_status = 0;
	if (task->tk_rqstp)
		return;
	if (!list_empty(&xprt->free)) {
		struct rpc_rqst	*req = list_entry(xprt->free.next, struct rpc_rqst, rq_list);
		list_del_init(&req->rq_list);
		task->tk_rqstp = req;
		xprt_request_init(task, xprt);
		return;
	}
	dprintk("RPC:      waiting for request slot\n");
	task->tk_status = -EAGAIN;
	task->tk_timeout = 0;
	rpc_sleep_on(&xprt->backlog, task, NULL, NULL);
}

void
xprt_reserve(struct rpc_task *task)
{
	struct rpc_xprt	*xprt = task->tk_xprt;

	task->tk_status = -EIO;
	if (!xprt->shutdown) {
		spin_lock(&xprt->xprt_lock);
		do_xprt_reserve(task);
		spin_unlock(&xprt->xprt_lock);
		if (task->tk_rqstp)
			del_timer_sync(&xprt->timer);
	}
}

/*
 * Allocate a 'unique' XID
 */
static inline u32 xprt_alloc_xid(struct rpc_xprt *xprt)
{
	return xprt->xid++;
}

static inline void xprt_init_xid(struct rpc_xprt *xprt)
{
	get_random_bytes(&xprt->xid, sizeof(xprt->xid));
}

/*
 * Initialize RPC request
 */
static void
xprt_request_init(struct rpc_task *task, struct rpc_xprt *xprt)
{
	struct rpc_rqst	*req = task->tk_rqstp;

	req->rq_timeout = xprt->timeout.to_initval;
	req->rq_task	= task;
	req->rq_xprt    = xprt;
	req->rq_xid     = xprt_alloc_xid(xprt);
	dprintk("RPC: %4d reserved req %p xid %08x\n", task->tk_pid,
			req, ntohl(req->rq_xid));
}

/*
 * Release an RPC call slot
 */
void
xprt_release(struct rpc_task *task)
{
	struct rpc_xprt	*xprt = task->tk_xprt;
	struct rpc_rqst	*req;

	if (!(req = task->tk_rqstp))
		return;
	spin_lock_bh(&xprt->sock_lock);
	__xprt_release_write(xprt, task);
	__xprt_put_cong(xprt, req);
	if (!list_empty(&req->rq_list))
		list_del(&req->rq_list);
	xprt->last_used = jiffies;
	if (list_empty(&xprt->recv) && !xprt->shutdown)
		mod_timer(&xprt->timer, xprt->last_used + XPRT_IDLE_TIMEOUT);
	spin_unlock_bh(&xprt->sock_lock);
	task->tk_rqstp = NULL;
	memset(req, 0, sizeof(*req));	/* mark unused */

	dprintk("RPC: %4d release request %p\n", task->tk_pid, req);

	spin_lock(&xprt->xprt_lock);
	list_add(&req->rq_list, &xprt->free);
	xprt_clear_backlog(xprt);
	spin_unlock(&xprt->xprt_lock);
}

/*
 * Set default timeout parameters
 */
static void
xprt_default_timeout(struct rpc_timeout *to, int proto)
{
	if (proto == IPPROTO_UDP)
		xprt_set_timeout(to, 5,  5 * HZ);
	else
		xprt_set_timeout(to, 5, 60 * HZ);
}

/*
 * Set constant timeout
 */
void
xprt_set_timeout(struct rpc_timeout *to, unsigned int retr, unsigned long incr)
{
	to->to_initval   = 
	to->to_increment = incr;
	to->to_maxval    = incr * retr;
	to->to_retries   = retr;
	to->to_exponential = 0;
}

unsigned int xprt_udp_slot_table_entries = RPC_DEF_SLOT_TABLE;
unsigned int xprt_tcp_slot_table_entries = RPC_DEF_SLOT_TABLE;

/*
 * Initialize an RPC client
 */
static struct rpc_xprt *
xprt_setup(int proto, struct sockaddr_in *ap, struct rpc_timeout *to)
{
	struct rpc_xprt	*xprt;
	unsigned int entries;
	size_t slot_table_size;
	struct rpc_rqst	*req;

	dprintk("RPC:      setting up %s transport...\n",
				proto == IPPROTO_UDP? "UDP" : "TCP");

	entries = (proto == IPPROTO_TCP)?
		xprt_tcp_slot_table_entries : xprt_udp_slot_table_entries;

	if ((xprt = kmalloc(sizeof(struct rpc_xprt), GFP_KERNEL)) == NULL)
		return ERR_PTR(-ENOMEM);
	memset(xprt, 0, sizeof(*xprt)); /* Nnnngh! */
	xprt->max_reqs = entries;
	slot_table_size = entries * sizeof(xprt->slot[0]);
	xprt->slot = kmalloc(slot_table_size, GFP_KERNEL);
	if (xprt->slot == NULL) {
		kfree(xprt);
		return ERR_PTR(-ENOMEM);
	}
	memset(xprt->slot, 0, slot_table_size);

	xprt->addr = *ap;
	xprt->prot = proto;
	xprt->stream = (proto == IPPROTO_TCP)? 1 : 0;
	if (xprt->stream) {
		xprt->cwnd = RPC_MAXCWND(xprt);
		xprt->nocong = 1;
		xprt->max_payload = (1U << 31) - 1;
	} else {
		xprt->cwnd = RPC_INITCWND;
		xprt->max_payload = (1U << 16) - (MAX_HEADER << 3);
	}
	spin_lock_init(&xprt->sock_lock);
	spin_lock_init(&xprt->xprt_lock);
	init_waitqueue_head(&xprt->cong_wait);

	INIT_LIST_HEAD(&xprt->free);
	INIT_LIST_HEAD(&xprt->recv);
	INIT_WORK(&xprt->sock_connect, xprt_socket_connect, xprt);
	INIT_WORK(&xprt->task_cleanup, xprt_socket_autoclose, xprt);
	init_timer(&xprt->timer);
	xprt->timer.function = xprt_init_autodisconnect;
	xprt->timer.data = (unsigned long) xprt;
	xprt->last_used = jiffies;
	xprt->port = XPRT_MAX_RESVPORT;

	/* Set timeout parameters */
	if (to) {
		xprt->timeout = *to;
	} else
		xprt_default_timeout(&xprt->timeout, xprt->prot);

	rpc_init_wait_queue(&xprt->pending, "xprt_pending");
	rpc_init_wait_queue(&xprt->sending, "xprt_sending");
	rpc_init_wait_queue(&xprt->resend, "xprt_resend");
	rpc_init_priority_wait_queue(&xprt->backlog, "xprt_backlog");

	/* initialize free list */
	for (req = &xprt->slot[entries-1]; req >= &xprt->slot[0]; req--)
		list_add(&req->rq_list, &xprt->free);

	xprt_init_xid(xprt);

	/* Check whether we want to use a reserved port */
	xprt->resvport = capable(CAP_NET_BIND_SERVICE) ? 1 : 0;

	dprintk("RPC:      created transport %p with %u slots\n", xprt,
			xprt->max_reqs);
	
	return xprt;
}

/*
 * Bind to a reserved port
 */
static inline int xprt_bindresvport(struct rpc_xprt *xprt, struct socket *sock)
{
	struct sockaddr_in myaddr = {
		.sin_family = AF_INET,
	};
	int		err, port;

	/* Were we already bound to a given port? Try to reuse it */
	port = xprt->port;
	do {
		myaddr.sin_port = htons(port);
		err = sock->ops->bind(sock, (struct sockaddr *) &myaddr,
						sizeof(myaddr));
		if (err == 0) {
			xprt->port = port;
			return 0;
		}
		if (--port == 0)
			port = XPRT_MAX_RESVPORT;
	} while (err == -EADDRINUSE && port != xprt->port);

	printk("RPC: Can't bind to reserved port (%d).\n", -err);
	return err;
}

static void
xprt_bind_socket(struct rpc_xprt *xprt, struct socket *sock)
{
	struct sock	*sk = sock->sk;

	if (xprt->inet)
		return;

	write_lock_bh(&sk->sk_callback_lock);
	sk->sk_user_data = xprt;
	xprt->old_data_ready = sk->sk_data_ready;
	xprt->old_state_change = sk->sk_state_change;
	xprt->old_write_space = sk->sk_write_space;
	if (xprt->prot == IPPROTO_UDP) {
		sk->sk_data_ready = udp_data_ready;
		sk->sk_no_check = UDP_CSUM_NORCV;
		xprt_set_connected(xprt);
	} else {
		tcp_sk(sk)->nonagle = 1;	/* disable Nagle's algorithm */
		sk->sk_data_ready = tcp_data_ready;
		sk->sk_state_change = tcp_state_change;
		xprt_clear_connected(xprt);
	}
	sk->sk_write_space = xprt_write_space;

	/* Reset to new socket */
	xprt->sock = sock;
	xprt->inet = sk;
	write_unlock_bh(&sk->sk_callback_lock);

	return;
}

/*
 * Set socket buffer length
 */
void
xprt_sock_setbufsize(struct rpc_xprt *xprt)
{
	struct sock *sk = xprt->inet;

	if (xprt->stream)
		return;
	if (xprt->rcvsize) {
		sk->sk_userlocks |= SOCK_RCVBUF_LOCK;
		sk->sk_rcvbuf = xprt->rcvsize * xprt->max_reqs *  2;
	}
	if (xprt->sndsize) {
		sk->sk_userlocks |= SOCK_SNDBUF_LOCK;
		sk->sk_sndbuf = xprt->sndsize * xprt->max_reqs * 2;
		sk->sk_write_space(sk);
	}
}

/*
 * Datastream sockets are created here, but xprt_connect will create
 * and connect stream sockets.
 */
static struct socket * xprt_create_socket(struct rpc_xprt *xprt, int proto, int resvport)
{
	struct socket	*sock;
	int		type, err;

	dprintk("RPC:      xprt_create_socket(%s %d)\n",
			   (proto == IPPROTO_UDP)? "udp" : "tcp", proto);

	type = (proto == IPPROTO_UDP)? SOCK_DGRAM : SOCK_STREAM;

	if ((err = sock_create_kern(PF_INET, type, proto, &sock)) < 0) {
		printk("RPC: can't create socket (%d).\n", -err);
		return NULL;
	}

	/* If the caller has the capability, bind to a reserved port */
	if (resvport && xprt_bindresvport(xprt, sock) < 0) {
		printk("RPC: can't bind to reserved port.\n");
		goto failed;
	}

	return sock;

failed:
	sock_release(sock);
	return NULL;
}

/*
 * Create an RPC client transport given the protocol and peer address.
 */
struct rpc_xprt *
xprt_create_proto(int proto, struct sockaddr_in *sap, struct rpc_timeout *to)
{
	struct rpc_xprt	*xprt;

	xprt = xprt_setup(proto, sap, to);
	if (IS_ERR(xprt))
		dprintk("RPC:      xprt_create_proto failed\n");
	else
		dprintk("RPC:      xprt_create_proto created xprt %p\n", xprt);
	return xprt;
}

/*
 * Prepare for transport shutdown.
 */
static void
xprt_shutdown(struct rpc_xprt *xprt)
{
	xprt->shutdown = 1;
	rpc_wake_up(&xprt->sending);
	rpc_wake_up(&xprt->resend);
	rpc_wake_up(&xprt->pending);
	rpc_wake_up(&xprt->backlog);
	wake_up(&xprt->cong_wait);
	del_timer_sync(&xprt->timer);
}

/*
 * Clear the xprt backlog queue
 */
static int
xprt_clear_backlog(struct rpc_xprt *xprt) {
	rpc_wake_up_next(&xprt->backlog);
	wake_up(&xprt->cong_wait);
	return 1;
}

/*
 * Destroy an RPC transport, killing off all requests.
 */
int
xprt_destroy(struct rpc_xprt *xprt)
{
	dprintk("RPC:      destroying transport %p\n", xprt);
	xprt_shutdown(xprt);
	xprt_disconnect(xprt);
	xprt_close(xprt);
	kfree(xprt->slot);
	kfree(xprt);

	return 0;
}