[linux.git] / fs / afs / vl_rotate.c

/* Handle vlserver selection and rotation.
 *
 * Copyright (C) 2018 Red Hat, Inc. All Rights Reserved.
 * Written by David Howells ([email protected])
 *
 * This program is free software; you can redistribute it and/or
 * modify it under the terms of the GNU General Public Licence
 * as published by the Free Software Foundation; either version
 * 2 of the Licence, or (at your option) any later version.
 */

#include <linux/kernel.h>
#include <linux/sched.h>
#include <linux/sched/signal.h>
#include "internal.h"
#include "afs_vl.h"

/*
 * Begin an operation on a volume location server.
 */
bool afs_begin_vlserver_operation(struct afs_vl_cursor *vc, struct afs_cell *cell,
				  struct key *key)
{
	memset(vc, 0, sizeof(*vc));
	vc->cell = cell;
	vc->key = key;
	vc->error = -EDESTADDRREQ;
	vc->ac.error = SHRT_MAX;

	if (signal_pending(current)) {
		vc->error = -EINTR;
		vc->flags |= AFS_VL_CURSOR_STOP;
		return false;
	}

	return true;
}

/*
 * Begin iteration through a server list, starting with the last used server if
 * possible, or the last recorded good server if not.
 */
static bool afs_start_vl_iteration(struct afs_vl_cursor *vc)
{
	struct afs_cell *cell = vc->cell;

	if (wait_on_bit(&cell->flags, AFS_CELL_FL_NO_LOOKUP_YET,
			TASK_INTERRUPTIBLE)) {
		vc->error = -ERESTARTSYS;
		return false;
	}

	read_lock(&cell->vl_servers_lock);
	vc->server_list = afs_get_vlserverlist(
		rcu_dereference_protected(cell->vl_servers,
					  lockdep_is_held(&cell->vl_servers_lock)));
	read_unlock(&cell->vl_servers_lock);
	if (!vc->server_list || !vc->server_list->nr_servers)
		return false;

	vc->untried = (1UL << vc->server_list->nr_servers) - 1;
	vc->index = -1;
	return true;
}

/*
 * Select the vlserver to use.  May be called multiple times to rotate
 * through the vlservers.
 */
bool afs_select_vlserver(struct afs_vl_cursor *vc)
{
	struct afs_addr_list *alist;
	struct afs_vlserver *vlserver;
	struct afs_error e;
	u32 rtt;
	int error = vc->ac.error, i;

	_enter("%lx[%d],%lx[%d],%d,%d",
	       vc->untried, vc->index,
	       vc->ac.tried, vc->ac.index,
	       error, vc->ac.abort_code);

	if (vc->flags & AFS_VL_CURSOR_STOP) {
		_leave(" = f [stopped]");
		return false;
	}

	vc->nr_iterations++;

	/* Evaluate the result of the previous operation, if there was one. */
	switch (error) {
	case SHRT_MAX:
		goto start;

	default:
	case 0:
		/* Success or local failure.  Stop. */
		vc->error = error;
		vc->flags |= AFS_VL_CURSOR_STOP;
		_leave(" = f [okay/local %d]", vc->ac.error);
		return false;

	case -ECONNABORTED:
		/* The far side rejected the operation on some grounds.  This
		 * might involve the server being busy or the volume having been moved.
		 */
		switch (vc->ac.abort_code) {
		case AFSVL_IO:
		case AFSVL_BADVOLOPER:
		case AFSVL_NOMEM:
			/* The server went weird. */
			vc->error = -EREMOTEIO;
			//write_lock(&vc->cell->vl_servers_lock);
			//vc->server_list->weird_mask |= 1 << vc->index;
			//write_unlock(&vc->cell->vl_servers_lock);
			goto next_server;

		default:
			vc->error = afs_abort_to_error(vc->ac.abort_code);
			goto failed;
		}

	case -ERFKILL:
	case -EADDRNOTAVAIL:
	case -ENETUNREACH:
	case -EHOSTUNREACH:
	case -EHOSTDOWN:
	case -ECONNREFUSED:
	case -ETIMEDOUT:
	case -ETIME:
		_debug("no conn %d", error);
		vc->error = error;
		goto iterate_address;

	case -ECONNRESET:
		_debug("call reset");
		vc->error = error;
		vc->flags |= AFS_VL_CURSOR_RETRY;
		goto next_server;
	}

restart_from_beginning:
	_debug("restart");
	afs_end_cursor(&vc->ac);
	afs_put_vlserverlist(vc->cell->net, vc->server_list);
	vc->server_list = NULL;
	if (vc->flags & AFS_VL_CURSOR_RETRIED)
		goto failed;
	vc->flags |= AFS_VL_CURSOR_RETRIED;
start:
	_debug("start");

	if (!afs_start_vl_iteration(vc))
		goto failed;

	error = afs_send_vl_probes(vc->cell->net, vc->key, vc->server_list);
	if (error < 0)
		goto failed_set_error;

pick_server:
	_debug("pick [%lx]", vc->untried);

	error = afs_wait_for_vl_probes(vc->server_list, vc->untried);
	if (error < 0)
		goto failed_set_error;

	/* Pick the untried server with the lowest RTT. */
	vc->index = vc->server_list->preferred;
	if (test_bit(vc->index, &vc->untried))
		goto selected_server;

	vc->index = -1;
	rtt = U32_MAX;
	for (i = 0; i < vc->server_list->nr_servers; i++) {
		struct afs_vlserver *s = vc->server_list->servers[i].server;

		if (!test_bit(i, &vc->untried) || !s->probe.responded)
			continue;
		if (s->probe.rtt < rtt) {
			vc->index = i;
			rtt = s->probe.rtt;
		}
	}

	if (vc->index == -1)
		goto no_more_servers;

selected_server:
	_debug("use %d", vc->index);
	__clear_bit(vc->index, &vc->untried);

	/* We're starting on a different vlserver from the list.  We need to
	 * check it, find its address list and probe its capabilities before we
	 * use it.
	 */
	ASSERTCMP(vc->ac.alist, ==, NULL);
	vlserver = vc->server_list->servers[vc->index].server;
	vc->server = vlserver;

	_debug("USING VLSERVER: %s", vlserver->name);

	read_lock(&vlserver->lock);
	alist = rcu_dereference_protected(vlserver->addresses,
					  lockdep_is_held(&vlserver->lock));
	afs_get_addrlist(alist);
	read_unlock(&vlserver->lock);

	memset(&vc->ac, 0, sizeof(vc->ac));

	if (!vc->ac.alist)
		vc->ac.alist = alist;
	else
		afs_put_addrlist(alist);

	vc->ac.index = -1;

iterate_address:
	ASSERT(vc->ac.alist);
	/* Iterate over the current server's address list to try and find an
	 * address on which it will respond to us.
	 */
	if (!afs_iterate_addresses(&vc->ac))
		goto next_server;

	_debug("VL address %d/%d", vc->ac.index, vc->ac.alist->nr_addrs);

	_leave(" = t %pISpc", &vc->ac.alist->addrs[vc->ac.index].transport);
	return true;

next_server:
	_debug("next");
	afs_end_cursor(&vc->ac);
	goto pick_server;

no_more_servers:
	/* That's all the servers poked to no good effect.  Try again if some
	 * of them were busy.
	 */
	if (vc->flags & AFS_VL_CURSOR_RETRY)
		goto restart_from_beginning;

	e.error = -EDESTADDRREQ;
	e.responded = false;
	for (i = 0; i < vc->server_list->nr_servers; i++) {
		struct afs_vlserver *s = vc->server_list->servers[i].server;

		afs_prioritise_error(&e, READ_ONCE(s->probe.error),
				     s->probe.abort_code);
	}

failed_set_error:
	vc->error = error;
failed:
	vc->flags |= AFS_VL_CURSOR_STOP;
	afs_end_cursor(&vc->ac);
	_leave(" = f [failed %d]", vc->error);
	return false;
}

/*
 * Dump cursor state in the case of the error being EDESTADDRREQ.
 */
static void afs_vl_dump_edestaddrreq(const struct afs_vl_cursor *vc)
{
	static int count;
	int i;

	if (!IS_ENABLED(CONFIG_AFS_DEBUG_CURSOR) || count > 3)
		return;
	count++;

	rcu_read_lock();
	pr_notice("EDESTADDR occurred\n");
	pr_notice("VC: ut=%lx ix=%u ni=%hu fl=%hx err=%hd\n",
		  vc->untried, vc->index, vc->nr_iterations, vc->flags, vc->error);

	if (vc->server_list) {
		const struct afs_vlserver_list *sl = vc->server_list;
		pr_notice("VC: SL nr=%u ix=%u\n",
			  sl->nr_servers, sl->index);
		for (i = 0; i < sl->nr_servers; i++) {
			const struct afs_vlserver *s = sl->servers[i].server;
			pr_notice("VC: server %s+%hu fl=%lx E=%hd\n",
				  s->name, s->port, s->flags, s->probe.error);
			if (s->addresses) {
				const struct afs_addr_list *a =
					rcu_dereference(s->addresses);
				pr_notice("VC:  - nr=%u/%u/%u pf=%u\n",
					  a->nr_ipv4, a->nr_addrs, a->max_addrs,
					  a->preferred);
				pr_notice("VC:  - pr=%lx R=%lx F=%lx\n",
					  a->probed, a->responded, a->failed);
				if (a == vc->ac.alist)
					pr_notice("VC:  - current\n");
			}
		}
	}

	pr_notice("AC: t=%lx ax=%u ac=%d er=%d r=%u ni=%u\n",
		  vc->ac.tried, vc->ac.index, vc->ac.abort_code, vc->ac.error,
		  vc->ac.responded, vc->ac.nr_iterations);
	rcu_read_unlock();
}

/*
 * Tidy up a volume location server cursor and unlock the vnode.
 */
int afs_end_vlserver_operation(struct afs_vl_cursor *vc)
{
	struct afs_net *net = vc->cell->net;

	if (vc->error == -EDESTADDRREQ ||
	    vc->error == -EADDRNOTAVAIL ||
	    vc->error == -ENETUNREACH ||
	    vc->error == -EHOSTUNREACH)
		afs_vl_dump_edestaddrreq(vc);

	afs_end_cursor(&vc->ac);
	afs_put_vlserverlist(net, vc->server_list);

	if (vc->error == -ECONNABORTED)
		vc->error = afs_abort_to_error(vc->ac.abort_code);

	return vc->error;
}
Commit	Line	Data
0a5143f2 DH	1	/* Handle vlserver selection and rotation.
	2	*
	3	* Copyright (C) 2018 Red Hat, Inc. All Rights Reserved.
	4	* Written by David Howells ([email protected])
	5	*
	6	* This program is free software; you can redistribute it and/or
	7	* modify it under the terms of the GNU General Public Licence
	8	* as published by the Free Software Foundation; either version
	9	* 2 of the Licence, or (at your option) any later version.
	10	*/
	11
	12	#include <linux/kernel.h>
	13	#include <linux/sched.h>
	14	#include <linux/sched/signal.h>
	15	#include "internal.h"
	16	#include "afs_vl.h"
	17
	18	/*
	19	* Begin an operation on a volume location server.
	20	*/
	21	bool afs_begin_vlserver_operation(struct afs_vl_cursor vc, struct afs_cell cell,
	22	struct key *key)
	23	{
	24	memset(vc, 0, sizeof(*vc));
	25	vc->cell = cell;
	26	vc->key = key;
	27	vc->error = -EDESTADDRREQ;
	28	vc->ac.error = SHRT_MAX;
	29
	30	if (signal_pending(current)) {
	31	vc->error = -EINTR;
	32	vc->flags \|= AFS_VL_CURSOR_STOP;
	33	return false;
	34	}
	35
	36	return true;
	37	}
	38
	39	/*
	40	* Begin iteration through a server list, starting with the last used server if
	41	* possible, or the last recorded good server if not.
	42	*/
	43	static bool afs_start_vl_iteration(struct afs_vl_cursor *vc)
	44	{
	45	struct afs_cell *cell = vc->cell;
	46
	47	if (wait_on_bit(&cell->flags, AFS_CELL_FL_NO_LOOKUP_YET,
	48	TASK_INTERRUPTIBLE)) {
	49	vc->error = -ERESTARTSYS;
	50	return false;
	51	}
	52
	53	read_lock(&cell->vl_servers_lock);
	54	vc->server_list = afs_get_vlserverlist(
	55	rcu_dereference_protected(cell->vl_servers,
	56	lockdep_is_held(&cell->vl_servers_lock)));
	57	read_unlock(&cell->vl_servers_lock);
	58	if (!vc->server_list \|\| !vc->server_list->nr_servers)
	59	return false;
	60
3bf0fb6f DH	61	vc->untried = (1UL << vc->server_list->nr_servers) - 1;
3bf0fb6f DH	62	vc->index = -1;
0a5143f2 DH	63	return true;
	64	}
	65
	66	/*
	67	* Select the vlserver to use. May be called multiple times to rotate
	68	* through the vlservers.
	69	*/
	70	bool afs_select_vlserver(struct afs_vl_cursor *vc)
	71	{
	72	struct afs_addr_list *alist;
	73	struct afs_vlserver *vlserver;
4584ae96	74	struct afs_error e;
3bf0fb6f	75	u32 rtt;
4584ae96	76	int error = vc->ac.error, i;
0a5143f2	77
3bf0fb6f DH	78	_enter("%lx[%d],%lx[%d],%d,%d",
	79	vc->untried, vc->index,
	80	vc->ac.tried, vc->ac.index,
0a5143f2 DH	81	error, vc->ac.abort_code);
	82
	83	if (vc->flags & AFS_VL_CURSOR_STOP) {
	84	_leave(" = f [stopped]");
	85	return false;
	86	}
	87
744bcd71 DH	88	vc->nr_iterations++;
744bcd71 DH	89
0a5143f2 DH	90	/* Evaluate the result of the previous operation, if there was one. */
	91	switch (error) {
	92	case SHRT_MAX:
	93	goto start;
	94
	95	default:
	96	case 0:
	97	/* Success or local failure. Stop. */
	98	vc->error = error;
	99	vc->flags \|= AFS_VL_CURSOR_STOP;
	100	_leave(" = f [okay/local %d]", vc->ac.error);
	101	return false;
	102
	103	case -ECONNABORTED:
	104	/* The far side rejected the operation on some grounds. This
	105	* might involve the server being busy or the volume having been moved.
	106	*/
	107	switch (vc->ac.abort_code) {
	108	case AFSVL_IO:
	109	case AFSVL_BADVOLOPER:
	110	case AFSVL_NOMEM:
	111	/* The server went weird. */
	112	vc->error = -EREMOTEIO;
	113	//write_lock(&vc->cell->vl_servers_lock);
	114	//vc->server_list->weird_mask \|= 1 << vc->index;
	115	//write_unlock(&vc->cell->vl_servers_lock);
	116	goto next_server;
	117
	118	default:
	119	vc->error = afs_abort_to_error(vc->ac.abort_code);
	120	goto failed;
	121	}
	122
4584ae96 DH	123	case -ERFKILL:
4584ae96 DH	124	case -EADDRNOTAVAIL:
0a5143f2 DH	125	case -ENETUNREACH:
0a5143f2 DH	126	case -EHOSTUNREACH:
4584ae96	127	case -EHOSTDOWN:
0a5143f2 DH	128	case -ECONNREFUSED:
	129	case -ETIMEDOUT:
	130	case -ETIME:
	131	_debug("no conn %d", error);
	132	vc->error = error;
	133	goto iterate_address;
	134
	135	case -ECONNRESET:
	136	_debug("call reset");
	137	vc->error = error;
	138	vc->flags \|= AFS_VL_CURSOR_RETRY;
	139	goto next_server;
	140	}
	141
	142	restart_from_beginning:
	143	_debug("restart");
	144	afs_end_cursor(&vc->ac);
	145	afs_put_vlserverlist(vc->cell->net, vc->server_list);
	146	vc->server_list = NULL;
	147	if (vc->flags & AFS_VL_CURSOR_RETRIED)
	148	goto failed;
	149	vc->flags \|= AFS_VL_CURSOR_RETRIED;
	150	start:
	151	_debug("start");
	152
0a5143f2 DH	153	if (!afs_start_vl_iteration(vc))
	154	goto failed;
	155
3bf0fb6f DH	156	error = afs_send_vl_probes(vc->cell->net, vc->key, vc->server_list);
	157	if (error < 0)
	158	goto failed_set_error;
	159
	160	pick_server:
	161	_debug("pick [%lx]", vc->untried);
	162
	163	error = afs_wait_for_vl_probes(vc->server_list, vc->untried);
	164	if (error < 0)
	165	goto failed_set_error;
	166
	167	/* Pick the untried server with the lowest RTT. */
	168	vc->index = vc->server_list->preferred;
	169	if (test_bit(vc->index, &vc->untried))
	170	goto selected_server;
	171
	172	vc->index = -1;
	173	rtt = U32_MAX;
	174	for (i = 0; i < vc->server_list->nr_servers; i++) {
	175	struct afs_vlserver *s = vc->server_list->servers[i].server;
	176
	177	if (!test_bit(i, &vc->untried) \|\| !s->probe.responded)
	178	continue;
	179	if (s->probe.rtt < rtt) {
	180	vc->index = i;
	181	rtt = s->probe.rtt;
	182	}
	183	}
	184
	185	if (vc->index == -1)
	186	goto no_more_servers;
	187
	188	selected_server:
	189	_debug("use %d", vc->index);
	190	__clear_bit(vc->index, &vc->untried);
	191
0a5143f2 DH	192	/* We're starting on a different vlserver from the list. We need to
	193	* check it, find its address list and probe its capabilities before we
	194	* use it.
	195	*/
	196	ASSERTCMP(vc->ac.alist, ==, NULL);
	197	vlserver = vc->server_list->servers[vc->index].server;
3bf0fb6f	198	vc->server = vlserver;
0a5143f2 DH	199
	200	_debug("USING VLSERVER: %s", vlserver->name);
	201
	202	read_lock(&vlserver->lock);
	203	alist = rcu_dereference_protected(vlserver->addresses,
	204	lockdep_is_held(&vlserver->lock));
	205	afs_get_addrlist(alist);
	206	read_unlock(&vlserver->lock);
	207
	208	memset(&vc->ac, 0, sizeof(vc->ac));
	209
0a5143f2 DH	210	if (!vc->ac.alist)
	211	vc->ac.alist = alist;
	212	else
	213	afs_put_addrlist(alist);
	214
3bf0fb6f	215	vc->ac.index = -1;
0a5143f2 DH	216
	217	iterate_address:
	218	ASSERT(vc->ac.alist);
0a5143f2 DH	219	/* Iterate over the current server's address list to try and find an
	220	* address on which it will respond to us.
	221	*/
	222	if (!afs_iterate_addresses(&vc->ac))
	223	goto next_server;
	224
3bf0fb6f DH	225	_debug("VL address %d/%d", vc->ac.index, vc->ac.alist->nr_addrs);
3bf0fb6f DH	226
2feeaf84	227	_leave(" = t %pISpc", &vc->ac.alist->addrs[vc->ac.index].transport);
0a5143f2 DH	228	return true;
	229
	230	next_server:
	231	_debug("next");
	232	afs_end_cursor(&vc->ac);
3bf0fb6f	233	goto pick_server;
0a5143f2	234
3bf0fb6f	235	no_more_servers:
0a5143f2 DH	236	/* That's all the servers poked to no good effect. Try again if some
	237	* of them were busy.
	238	*/
	239	if (vc->flags & AFS_VL_CURSOR_RETRY)
	240	goto restart_from_beginning;
	241
4584ae96 DH	242	e.error = -EDESTADDRREQ;
4584ae96 DH	243	e.responded = false;
3bf0fb6f DH	244	for (i = 0; i < vc->server_list->nr_servers; i++) {
3bf0fb6f DH	245	struct afs_vlserver *s = vc->server_list->servers[i].server;
3bf0fb6f	246
4584ae96 DH	247	afs_prioritise_error(&e, READ_ONCE(s->probe.error),
4584ae96 DH	248	s->probe.abort_code);
3bf0fb6f DH	249	}
3bf0fb6f DH	250
3bf0fb6f DH	251	failed_set_error:
3bf0fb6f DH	252	vc->error = error;
0a5143f2 DH	253	failed:
	254	vc->flags \|= AFS_VL_CURSOR_STOP;
	255	afs_end_cursor(&vc->ac);
	256	_leave(" = f [failed %d]", vc->error);
	257	return false;
	258	}
	259
744bcd71 DH	260	/*
	261	* Dump cursor state in the case of the error being EDESTADDRREQ.
	262	*/
	263	static void afs_vl_dump_edestaddrreq(const struct afs_vl_cursor *vc)
	264	{
	265	static int count;
	266	int i;
	267
	268	if (!IS_ENABLED(CONFIG_AFS_DEBUG_CURSOR) \|\| count > 3)
	269	return;
	270	count++;
	271
	272	rcu_read_lock();
	273	pr_notice("EDESTADDR occurred\n");
3bf0fb6f DH	274	pr_notice("VC: ut=%lx ix=%u ni=%hu fl=%hx err=%hd\n",
3bf0fb6f DH	275	vc->untried, vc->index, vc->nr_iterations, vc->flags, vc->error);
744bcd71 DH	276
	277	if (vc->server_list) {
	278	const struct afs_vlserver_list *sl = vc->server_list;
	279	pr_notice("VC: SL nr=%u ix=%u\n",
	280	sl->nr_servers, sl->index);
	281	for (i = 0; i < sl->nr_servers; i++) {
	282	const struct afs_vlserver *s = sl->servers[i].server;
3bf0fb6f DH	283	pr_notice("VC: server %s+%hu fl=%lx E=%hd\n",
3bf0fb6f DH	284	s->name, s->port, s->flags, s->probe.error);
744bcd71 DH	285	if (s->addresses) {
	286	const struct afs_addr_list *a =
	287	rcu_dereference(s->addresses);
3bf0fb6f	288	pr_notice("VC: - nr=%u/%u/%u pf=%u\n",
744bcd71	289	a->nr_ipv4, a->nr_addrs, a->max_addrs,
3bf0fb6f DH	290	a->preferred);
	291	pr_notice("VC: - pr=%lx R=%lx F=%lx\n",
	292	a->probed, a->responded, a->failed);
744bcd71 DH	293	if (a == vc->ac.alist)
	294	pr_notice("VC: - current\n");
	295	}
	296	}
	297	}
	298
3bf0fb6f DH	299	pr_notice("AC: t=%lx ax=%u ac=%d er=%d r=%u ni=%u\n",
	300	vc->ac.tried, vc->ac.index, vc->ac.abort_code, vc->ac.error,
	301	vc->ac.responded, vc->ac.nr_iterations);
744bcd71 DH	302	rcu_read_unlock();
	303	}
	304
0a5143f2 DH	305	/*
	306	* Tidy up a volume location server cursor and unlock the vnode.
	307	*/
	308	int afs_end_vlserver_operation(struct afs_vl_cursor *vc)
	309	{
	310	struct afs_net *net = vc->cell->net;
	311
744bcd71	312	if (vc->error == -EDESTADDRREQ \|\|
4584ae96	313	vc->error == -EADDRNOTAVAIL \|\|
744bcd71 DH	314	vc->error == -ENETUNREACH \|\|
	315	vc->error == -EHOSTUNREACH)
	316	afs_vl_dump_edestaddrreq(vc);
	317
0a5143f2 DH	318	afs_end_cursor(&vc->ac);
	319	afs_put_vlserverlist(net, vc->server_list);
	320
	321	if (vc->error == -ECONNABORTED)
	322	vc->error = afs_abort_to_error(vc->ac.abort_code);
	323
	324	return vc->error;
	325	}