afs: Add an address list concept

Add an RCU replaceable address list structure to hold a list of server
addresses.  The list also holds the

To this end:

 (1) A cell's VL server address list can be loaded directly via insmod or
     echo to /proc/fs/afs/cells or dynamically from a DNS query for AFSDB
     or SRV records.

 (2) Anyone wanting to use a cell's VL server address must wait until the
     cell record comes online and has tried to obtain some addresses.

 (3) An FS server's address list, for the moment, has a single entry that
     is the key to the server list.  This will change in the future when a
     server is instead keyed on its UUID and the VL.GetAddrsU operation is
     used.

 (4) An 'address cursor' concept is introduced to handle iteration through
     the address list.  This is passed to the afs_make_call() as, in the
     future, stuff (such as abort code) that doesn't outlast the call will
     be returned in it.

In the future, we might want to annotate the list with information about
how each address fares.  We might then want to propagate such annotations
over address list replacement.

Whilst we're at it, we allow IPv6 addresses to be specified in
colon-delimited lists by enclosing them in square brackets.

Signed-off-by: David Howells <dhowells@redhat.com>

12 files changed, 844 insertions, 561 deletions

diff --git a/fs/afs/Makefile b/fs/afs/Makefile
index 641148208e90..849383986d3b 100644
--- a/fs/afs/Makefile
+++ b/fs/afs/Makefile
@@ -7,6 +7,7 @@ afs-cache-$(CONFIG_AFS_FSCACHE) := cache.o
 
 kafs-objs := \
 	$(afs-cache-y) \
+	addr_list.o \
 	callback.o \
 	cell.o \
 	cmservice.o \
diff --git a/fs/afs/addr_list.c b/fs/afs/addr_list.c
new file mode 100644
index 000000000000..ecb9c72aebd2
--- /dev/null
+++ b/fs/afs/addr_list.c
@@ -0,0 +1,308 @@
+/* Server address list management
+ *
+ * Copyright (C) 2017 Red Hat, Inc. All Rights Reserved.
+ * Written by David Howells (dhowells@redhat.com)
+ *
+ * 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/slab.h>
+#include <linux/ctype.h>
+#include <linux/dns_resolver.h>
+#include <linux/inet.h>
+#include <keys/rxrpc-type.h>
+#include "internal.h"
+#include "afs_fs.h"
+
+#define AFS_MAX_ADDRESSES						\
+	((unsigned int)((PAGE_SIZE - sizeof(struct afs_addr_list)) /	\
+			sizeof(struct sockaddr_rxrpc)))
+
+/*
+ * Release an address list.
+ */
+void afs_put_addrlist(struct afs_addr_list *alist)
+{
+	if (alist && refcount_dec_and_test(&alist->usage))
+		call_rcu(&alist->rcu, (rcu_callback_t)kfree);
+}
+
+/*
+ * Allocate an address list.
+ */
+struct afs_addr_list *afs_alloc_addrlist(unsigned int nr,
+					 unsigned short service,
+					 unsigned short port)
+{
+	struct afs_addr_list *alist;
+	unsigned int i;
+
+	_enter("%u,%u,%u", nr, service, port);
+
+	alist = kzalloc(sizeof(*alist) + sizeof(alist->addrs[0]) * nr,
+			GFP_KERNEL);
+	if (!alist)
+		return NULL;
+
+	refcount_set(&alist->usage, 1);
+
+	for (i = 0; i < nr; i++) {
+		struct sockaddr_rxrpc *srx = &alist->addrs[i];
+		srx->srx_family			= AF_RXRPC;
+		srx->srx_service		= service;
+		srx->transport_type		= SOCK_DGRAM;
+		srx->transport_len		= sizeof(srx->transport.sin6);
+		srx->transport.sin6.sin6_family	= AF_INET6;
+		srx->transport.sin6.sin6_port	= htons(port);
+	}
+
+	return alist;
+}
+
+/*
+ * Parse a text string consisting of delimited addresses.
+ */
+struct afs_addr_list *afs_parse_text_addrs(const char *text, size_t len,
+					   char delim,
+					   unsigned short service,
+					   unsigned short port)
+{
+	struct afs_addr_list *alist;
+	const char *p, *end = text + len;
+	unsigned int nr = 0;
+
+	_enter("%*.*s,%c", (int)len, (int)len, text, delim);
+
+	if (!len)
+		return ERR_PTR(-EDESTADDRREQ);
+
+	if (delim == ':' && (memchr(text, ',', len) || !memchr(text, '.', len)))
+		delim = ',';
+
+	/* Count the addresses */
+	p = text;
+	do {
+		if (!*p)
+			return ERR_PTR(-EINVAL);
+		if (*p == delim)
+			continue;
+		nr++;
+		if (*p == '[') {
+			p++;
+			if (p == end)
+				return ERR_PTR(-EINVAL);
+			p = memchr(p, ']', end - p);
+			if (!p)
+				return ERR_PTR(-EINVAL);
+			p++;
+			if (p >= end)
+				break;
+		}
+
+		p = memchr(p, delim, end - p);
+		if (!p)
+			break;
+		p++;
+	} while (p < end);
+
+	_debug("%u/%u addresses", nr, AFS_MAX_ADDRESSES);
+	if (nr > AFS_MAX_ADDRESSES)
+		nr = AFS_MAX_ADDRESSES;
+
+	alist = afs_alloc_addrlist(nr, service, port);
+	if (!alist)
+		return ERR_PTR(-ENOMEM);
+
+	/* Extract the addresses */
+	p = text;
+	do {
+		struct sockaddr_rxrpc *srx = &alist->addrs[alist->nr_addrs];
+		char tdelim = delim;
+
+		if (*p == delim) {
+			p++;
+			continue;
+		}
+
+		if (*p == '[') {
+			p++;
+			tdelim = ']';
+		}
+
+		if (in4_pton(p, end - p,
+			     (u8 *)&srx->transport.sin6.sin6_addr.s6_addr32[3],
+			     tdelim, &p)) {
+			srx->transport.sin6.sin6_addr.s6_addr32[0] = 0;
+			srx->transport.sin6.sin6_addr.s6_addr32[1] = 0;
+			srx->transport.sin6.sin6_addr.s6_addr32[2] = htonl(0xffff);
+		} else if (in6_pton(p, end - p,
+				    srx->transport.sin6.sin6_addr.s6_addr,
+				    tdelim, &p)) {
+			/* Nothing to do */
+		} else {
+			goto bad_address;
+		}
+
+		if (tdelim == ']') {
+			if (p == end || *p != ']')
+				goto bad_address;
+			p++;
+		}
+
+		if (p < end) {
+			if (*p == '+') {
+				/* Port number specification "+1234" */
+				unsigned int xport = 0;
+				p++;
+				if (p >= end || !isdigit(*p))
+					goto bad_address;
+				do {
+					xport *= 10;
+					xport += *p - '0';
+					if (xport > 65535)
+						goto bad_address;
+					p++;
+				} while (p < end && isdigit(*p));
+				srx->transport.sin6.sin6_port = htons(xport);
+			} else if (*p == delim) {
+				p++;
+			} else {
+				goto bad_address;
+			}
+		}
+
+		alist->nr_addrs++;
+	} while (p < end && alist->nr_addrs < AFS_MAX_ADDRESSES);
+
+	_leave(" = [nr %u]", alist->nr_addrs);
+	return alist;
+
+bad_address:
+	kfree(alist);
+	return ERR_PTR(-EINVAL);
+}
+
+/*
+ * Compare old and new address lists to see if there's been any change.
+ * - How to do this in better than O(Nlog(N)) time?
+ *   - We don't really want to sort the address list, but would rather take the
+ *     list as we got it so as not to undo record rotation by the DNS server.
+ */
+#if 0
+static int afs_cmp_addr_list(const struct afs_addr_list *a1,
+			     const struct afs_addr_list *a2)
+{
+}
+#endif
+
+/*
+ * Perform a DNS query for VL servers and build a up an address list.
+ */
+struct afs_addr_list *afs_dns_query(struct afs_cell *cell, time64_t *_expiry)
+{
+	struct afs_addr_list *alist;
+	char *vllist = NULL;
+	int ret;
+
+	_enter("%s", cell->name);
+
+	ret = dns_query("afsdb", cell->name, cell->name_len,
+			"ipv4", &vllist, _expiry);
+	if (ret < 0)
+		return ERR_PTR(ret);
+
+	alist = afs_parse_text_addrs(vllist, strlen(vllist), ',',
+				     VL_SERVICE, AFS_VL_PORT);
+	if (IS_ERR(alist)) {
+		kfree(vllist);
+		if (alist != ERR_PTR(-ENOMEM))
+			pr_err("Failed to parse DNS data\n");
+		return alist;
+	}
+
+	kfree(vllist);
+	return alist;
+}
+
+/*
+ * Get an address to try.
+ */
+bool afs_iterate_addresses(struct afs_addr_cursor *ac)
+{
+	_enter("%hu+%hd", ac->start, (short)ac->index);
+
+	if (!ac->alist)
+		return false;
+
+	if (ac->begun) {
+		ac->index++;
+		if (ac->index == ac->alist->nr_addrs)
+			ac->index = 0;
+
+		if (ac->index == ac->start) {
+			ac->error = -EDESTADDRREQ;
+			return false;
+		}
+	}
+
+	ac->begun = true;
+	ac->responded = false;
+	ac->addr = &ac->alist->addrs[ac->index];
+	return true;
+}
+
+/*
+ * Release an address list cursor.
+ */
+int afs_end_cursor(struct afs_addr_cursor *ac)
+{
+	if (ac->responded && ac->index != ac->start)
+		WRITE_ONCE(ac->alist->index, ac->index);
+
+	afs_put_addrlist(ac->alist);
+	ac->alist = NULL;
+	return ac->error;
+}
+
+/*
+ * Set the address cursor for iterating over VL servers.
+ */
+int afs_set_vl_cursor(struct afs_addr_cursor *ac, struct afs_cell *cell)
+{
+	struct afs_addr_list *alist;
+	int ret;
+
+	if (!rcu_access_pointer(cell->vl_addrs)) {
+		ret = wait_on_bit(&cell->flags, AFS_CELL_FL_NO_LOOKUP_YET,
+				  TASK_INTERRUPTIBLE);
+		if (ret < 0)
+			return ret;
+
+		if (!rcu_access_pointer(cell->vl_addrs) &&
+		    ktime_get_real_seconds() < cell->dns_expiry)
+			return cell->error;
+	}
+
+	read_lock(&cell->vl_addrs_lock);
+	alist = rcu_dereference_protected(cell->vl_addrs,
+					  lockdep_is_held(&cell->vl_addrs_lock));
+	if (alist->nr_addrs > 0)
+		afs_get_addrlist(alist);
+	else
+		alist = NULL;
+	read_unlock(&cell->vl_addrs_lock);
+
+	if (!alist)
+		return -EDESTADDRREQ;
+
+	ac->alist = alist;
+	ac->addr = NULL;
+	ac->start = READ_ONCE(alist->index);
+	ac->index = ac->start;
+	ac->error = 0;
+	ac->begun = false;
+	return 0;
+}
diff --git a/fs/afs/cell.c b/fs/afs/cell.c
index e83103e8a6fb..a0e08d3a108c 100644
--- a/fs/afs/cell.c
+++ b/fs/afs/cell.c
@@ -9,7 +9,6 @@
  * 2 of the License, or (at your option) any later version.
  */
 
-#include <linux/module.h>
 #include <linux/slab.h>
 #include <linux/key.h>
 #include <linux/ctype.h>
@@ -152,68 +151,33 @@ static struct afs_cell *afs_alloc_cell(struct afs_net *net,
 	init_rwsem(&cell->vl_sem);
 	INIT_LIST_HEAD(&cell->vl_list);
 	spin_lock_init(&cell->vl_lock);
-	seqlock_init(&cell->vl_addrs_lock);
-	cell->flags = (1 << AFS_CELL_FL_NOT_READY);
-
-	for (i = 0; i < AFS_CELL_MAX_ADDRS; i++) {
-		struct sockaddr_rxrpc *srx = &cell->vl_addrs[i];
-		srx->srx_family			= AF_RXRPC;
-		srx->srx_service		= VL_SERVICE;
-		srx->transport_type		= SOCK_DGRAM;
-		srx->transport.sin6.sin6_family	= AF_INET6;
-		srx->transport.sin6.sin6_port	= htons(AFS_VL_PORT);
-	}
+	cell->flags = ((1 << AFS_CELL_FL_NOT_READY) |
+		       (1 << AFS_CELL_FL_NO_LOOKUP_YET));
+	rwlock_init(&cell->vl_addrs_lock);
 
 	/* Fill in the VL server list if we were given a list of addresses to
 	 * use.
 	 */
 	if (vllist) {
-		char delim = ':';
-
-		if (strchr(vllist, ',') || !strchr(vllist, '.'))
-			delim = ',';
-
-		do {
-			struct sockaddr_rxrpc *srx = &cell->vl_addrs[cell->vl_naddrs];
-
-			if (in4_pton(vllist, -1,
-				     (u8 *)&srx->transport.sin6.sin6_addr.s6_addr32[3],
-				     delim, &vllist)) {
-				srx->transport_len = sizeof(struct sockaddr_in6);
-				srx->transport.sin6.sin6_addr.s6_addr32[0] = 0;
-				srx->transport.sin6.sin6_addr.s6_addr32[1] = 0;
-				srx->transport.sin6.sin6_addr.s6_addr32[2] = htonl(0xffff);
-			} else if (in6_pton(vllist, -1,
-					    srx->transport.sin6.sin6_addr.s6_addr,
-					    delim, &vllist)) {
-				srx->transport_len = sizeof(struct sockaddr_in6);
-				srx->transport.sin6.sin6_family	= AF_INET6;
-			} else {
-				goto bad_address;
-			}
+		struct afs_addr_list *alist;
 
-			cell->vl_naddrs++;
-			if (!*vllist)
-				break;
-			vllist++;
-
-		} while (cell->vl_naddrs < AFS_CELL_MAX_ADDRS && vllist);
+		alist = afs_parse_text_addrs(vllist, strlen(vllist), ':',
+					     VL_SERVICE, AFS_VL_PORT);
+		if (IS_ERR(alist)) {
+			ret = PTR_ERR(alist);
+			goto parse_failed;
+		}
 
-		/* Disable DNS refresh for manually-specified cells */
+		rcu_assign_pointer(cell->vl_addrs, alist);
 		cell->dns_expiry = TIME64_MAX;
-	} else {
-		/* We're going to need to 'refresh' this cell's VL server list
-		 * from the DNS before we can use it.
-		 */
-		cell->dns_expiry = S64_MIN;
 	}
 
 	_leave(" = %p", cell);
 	return cell;
 
-bad_address:
-	printk(KERN_ERR "kAFS: bad VL server IP address\n");
-	ret = -EINVAL;
+parse_failed:
+	if (ret == -EINVAL)
+		printk(KERN_ERR "kAFS: bad VL server IP address\n");
 	kfree(cell);
 	_leave(" = %d", ret);
 	return ERR_PTR(ret);
@@ -325,7 +289,6 @@ cell_already_exists:
 	if (excl) {
 		ret = -EEXIST;
 	} else {
-		ASSERTCMP(atomic_read(&cursor->usage), >=, 1);
 		afs_get_cell(cursor);
 		ret = 0;
 	}
@@ -333,8 +296,10 @@ cell_already_exists:
 	kfree(candidate);
 	if (ret == 0)
 		goto wait_for_cell;
+	goto error_noput;
 error:
 	afs_put_cell(net, cell);
+error_noput:
 	_leave(" = %d [error]", ret);
 	return ERR_PTR(ret);
 }
@@ -396,78 +361,50 @@ int afs_cell_init(struct afs_net *net, const char *rootcell)
  */
 static void afs_update_cell(struct afs_cell *cell)
 {
+	struct afs_addr_list *alist, *old;
 	time64_t now, expiry;
-	char *vllist = NULL;
-	int ret;
 
 	_enter("%s", cell->name);
 
-	ret = dns_query("afsdb", cell->name, cell->name_len,
-			"ipv4", &vllist, &expiry);
-	_debug("query %d", ret);
-	switch (ret) {
-	case 0 ... INT_MAX:
-		clear_bit(AFS_CELL_FL_DNS_FAIL, &cell->flags);
-		clear_bit(AFS_CELL_FL_NOT_FOUND, &cell->flags);
-		goto parse_dns_data;
+	alist = afs_dns_query(cell, &expiry);
+	if (IS_ERR(alist)) {
+		switch (PTR_ERR(alist)) {
+		case -ENODATA:
+			/* The DNS said that the cell does not exist */
+			set_bit(AFS_CELL_FL_NOT_FOUND, &cell->flags);
+			clear_bit(AFS_CELL_FL_DNS_FAIL, &cell->flags);
+			cell->dns_expiry = ktime_get_real_seconds() + 61;
+			break;
 
-	case -ENODATA:
-		clear_bit(AFS_CELL_FL_DNS_FAIL, &cell->flags);
-		set_bit(AFS_CELL_FL_NOT_FOUND, &cell->flags);
-		cell->dns_expiry = ktime_get_real_seconds() + 61;
-		cell->error = -EDESTADDRREQ;
-		goto out;
+		case -EAGAIN:
+		case -ECONNREFUSED:
+		default:
+			set_bit(AFS_CELL_FL_DNS_FAIL, &cell->flags);
+			cell->dns_expiry = ktime_get_real_seconds() + 10;
+			break;
+		}
 
-	case -EAGAIN:
-	case -ECONNREFUSED:
-	default:
-		/* Unable to query DNS. */
-		set_bit(AFS_CELL_FL_DNS_FAIL, &cell->flags);
-		cell->dns_expiry = ktime_get_real_seconds() + 10;
 		cell->error = -EDESTADDRREQ;
-		goto out;
-	}
-
-parse_dns_data:
-	write_seqlock(&cell->vl_addrs_lock);
-
-	ret = -EINVAL;
-	do {
-		struct sockaddr_rxrpc *srx = &cell->vl_addrs[cell->vl_naddrs];
-
-		if (in4_pton(vllist, -1,
-			     (u8 *)&srx->transport.sin6.sin6_addr.s6_addr32[3],
-			     ',', (const char **)&vllist)) {
-			srx->transport_len = sizeof(struct sockaddr_in6);
-			srx->transport.sin6.sin6_addr.s6_addr32[0] = 0;
-			srx->transport.sin6.sin6_addr.s6_addr32[1] = 0;
-			srx->transport.sin6.sin6_addr.s6_addr32[2] = htonl(0xffff);
-		} else if (in6_pton(vllist, -1,
-				    srx->transport.sin6.sin6_addr.s6_addr,
-				    ',', (const char **)&vllist)) {
-			srx->transport_len = sizeof(struct sockaddr_in6);
-			srx->transport.sin6.sin6_family	= AF_INET6;
-		} else {
-			goto bad_address;
-		}
+	} else {
+		clear_bit(AFS_CELL_FL_DNS_FAIL, &cell->flags);
+		clear_bit(AFS_CELL_FL_NOT_FOUND, &cell->flags);
 
-		cell->vl_naddrs++;
-		if (!*vllist)
-			break;
-		vllist++;
+		/* Exclusion on changing vl_addrs is achieved by a
+		 * non-reentrant work item.
+		 */
+		old = rcu_dereference_protected(cell->vl_addrs, true);
+		rcu_assign_pointer(cell->vl_addrs, alist);
+		cell->dns_expiry = expiry;
 
-	} while (cell->vl_naddrs < AFS_CELL_MAX_ADDRS);
+		if (old)
+			afs_put_addrlist(old);
+	}
 
-	if (cell->vl_naddrs < AFS_CELL_MAX_ADDRS)
-		memset(cell->vl_addrs + cell->vl_naddrs, 0,
-		       (AFS_CELL_MAX_ADDRS - cell->vl_naddrs) * sizeof(cell->vl_addrs[0]));
+	if (test_and_clear_bit(AFS_CELL_FL_NO_LOOKUP_YET, &cell->flags))
+		wake_up_bit(&cell->flags, AFS_CELL_FL_NO_LOOKUP_YET);
 
 	now = ktime_get_real_seconds();
-	cell->dns_expiry = expiry;
-	afs_set_cell_timer(cell->net, expiry - now);
-bad_address:
-	write_sequnlock(&cell->vl_addrs_lock);
-out:
+	afs_set_cell_timer(cell->net, cell->dns_expiry - now);
 	_leave("");
 }
 
@@ -482,6 +419,7 @@ static void afs_cell_destroy(struct rcu_head *rcu)
 
 	ASSERTCMP(atomic_read(&cell->usage), ==, 0);
 
+	afs_put_addrlist(cell->vl_addrs);
 	key_put(cell->anonymous_key);
 	kfree(cell);
 
@@ -515,6 +453,15 @@ void afs_cells_timer(struct timer_list *timer)
 }
 
 /*
+ * Get a reference on a cell record.
+ */
+struct afs_cell *afs_get_cell(struct afs_cell *cell)
+{
+	atomic_inc(&cell->usage);
+	return cell;
+}
+
+/*
  * Drop a reference on a cell record.
  */
 void afs_put_cell(struct afs_net *net, struct afs_cell *cell)
diff --git a/fs/afs/fsclient.c b/fs/afs/fsclient.c
index 680c02d510f7..6614d0a78daa 100644
--- a/fs/afs/fsclient.c
+++ b/fs/afs/fsclient.c
@@ -297,7 +297,7 @@ static const struct afs_call_type afs_RXFSFetchStatus = {
 /*
  * fetch the status information for a file
  */
-int afs_fs_fetch_file_status(struct afs_server *server,
+int afs_fs_fetch_file_status(struct afs_fs_cursor *fc,
 			     struct key *key,
 			     struct afs_vnode *vnode,
 			     struct afs_volsync *volsync,
@@ -325,9 +325,9 @@ int afs_fs_fetch_file_status(struct afs_server *server,
 	bp[2] = htonl(vnode->fid.vnode);
 	bp[3] = htonl(vnode->fid.unique);
 
-	call->cb_break = vnode->cb_break + server->cb_s_break;
-	afs_use_fs_server(call, server);
-	return afs_make_call(&server->addr, call, GFP_NOFS, async);
+	call->cb_break = vnode->cb_break + fc->server->cb_s_break;
+	afs_use_fs_server(call, fc->server);
+	return afs_make_call(&fc->ac, call, GFP_NOFS, async);
 }
 
 /*
@@ -502,7 +502,7 @@ static const struct afs_call_type afs_RXFSFetchData64 = {
 /*
  * fetch data from a very large file
  */
-static int afs_fs_fetch_data64(struct afs_server *server,
+static int afs_fs_fetch_data64(struct afs_fs_cursor *fc,
 			       struct key *key,
 			       struct afs_vnode *vnode,
 			       struct afs_read *req,
@@ -536,15 +536,15 @@ static int afs_fs_fetch_data64(struct afs_server *server,
 	bp[7] = htonl(lower_32_bits(req->len));
 
 	atomic_inc(&req->usage);
-	call->cb_break = vnode->cb_break + server->cb_s_break;
-	afs_use_fs_server(call, server);
-	return afs_make_call(&server->addr, call, GFP_NOFS, async);
+	call->cb_break = vnode->cb_break + fc->server->cb_s_break;
+	afs_use_fs_server(call, fc->server);
+	return afs_make_call(&fc->ac, call, GFP_NOFS, async);
 }
 
 /*
  * fetch data from a file
  */
-int afs_fs_fetch_data(struct afs_server *server,
+int afs_fs_fetch_data(struct afs_fs_cursor *fc,
 		      struct key *key,
 		      struct afs_vnode *vnode,
 		      struct afs_read *req,
@@ -557,7 +557,7 @@ int afs_fs_fetch_data(struct afs_server *server,
 	if (upper_32_bits(req->pos) ||
 	    upper_32_bits(req->len) ||
 	    upper_32_bits(req->pos + req->len))
-		return afs_fs_fetch_data64(server, key, vnode, req, async);
+		return afs_fs_fetch_data64(fc, key, vnode, req, async);
 
 	_enter("");
 
@@ -581,9 +581,9 @@ int afs_fs_fetch_data(struct afs_server *server,
 	bp[5] = htonl(lower_32_bits(req->len));
 
 	atomic_inc(&req->usage);
-	call->cb_break = vnode->cb_break + server->cb_s_break;
-	afs_use_fs_server(call, server);
-	return afs_make_call(&server->addr, call, GFP_NOFS, async);
+	call->cb_break = vnode->cb_break + fc->server->cb_s_break;
+	afs_use_fs_server(call, fc->server);
+	return afs_make_call(&fc->ac, call, GFP_NOFS, async);
 }
 
 /*
@@ -625,7 +625,7 @@ static const struct afs_call_type afs_RXFSCreateXXXX = {
 /*
  * create a file or make a directory
  */
-int afs_fs_create(struct afs_server *server,
+int afs_fs_create(struct afs_fs_cursor *fc,
 		  struct key *key,
 		  struct afs_vnode *vnode,
 		  const char *name,
@@ -677,8 +677,8 @@ int afs_fs_create(struct afs_server *server,
 	*bp++ = htonl(mode & S_IALLUGO); /* unix mode */
 	*bp++ = 0; /* segment size */
 
-	afs_use_fs_server(call, server);
-	return afs_make_call(&server->addr, call, GFP_NOFS, async);
+	afs_use_fs_server(call, fc->server);
+	return afs_make_call(&fc->ac, call, GFP_NOFS, async);
 }
 
 /*
@@ -717,7 +717,7 @@ static const struct afs_call_type afs_RXFSRemoveXXXX = {
 /*
  * remove a file or directory
  */
-int afs_fs_remove(struct afs_server *server,
+int afs_fs_remove(struct afs_fs_cursor *fc,
 		  struct key *key,
 		  struct afs_vnode *vnode,
 		  const char *name,
@@ -756,8 +756,8 @@ int afs_fs_remove(struct afs_server *server,
 		bp = (void *) bp + padsz;
 	}
 
-	afs_use_fs_server(call, server);
-	return afs_make_call(&server->addr, call, GFP_NOFS, async);
+	afs_use_fs_server(call, fc->server);
+	return afs_make_call(&fc->ac, call, GFP_NOFS, async);
 }
 
 /*
@@ -797,7 +797,7 @@ static const struct afs_call_type afs_RXFSLink = {
 /*
  * make a hard link
  */
-int afs_fs_link(struct afs_server *server,
+int afs_fs_link(struct afs_fs_cursor *fc,
 		struct key *key,
 		struct afs_vnode *dvnode,
 		struct afs_vnode *vnode,
@@ -840,8 +840,8 @@ int afs_fs_link(struct afs_server *server,
 	*bp++ = htonl(vnode->fid.vnode);
 	*bp++ = htonl(vnode->fid.unique);
 
-	afs_use_fs_server(call, server);
-	return afs_make_call(&server->addr, call, GFP_NOFS, async);
+	afs_use_fs_server(call, fc->server);
+	return afs_make_call(&fc->ac, call, GFP_NOFS, async);
 }
 
 /*
@@ -882,7 +882,7 @@ static const struct afs_call_type afs_RXFSSymlink = {
 /*
  * create a symbolic link
  */
-int afs_fs_symlink(struct afs_server *server,
+int afs_fs_symlink(struct afs_fs_cursor *fc,
 		   struct key *key,
 		   struct afs_vnode *vnode,
 		   const char *name,
@@ -943,8 +943,8 @@ int afs_fs_symlink(struct afs_server *server,
 	*bp++ = htonl(S_IRWXUGO); /* unix mode */
 	*bp++ = 0; /* segment size */
 
-	afs_use_fs_server(call, server);
-	return afs_make_call(&server->addr, call, GFP_NOFS, async);
+	afs_use_fs_server(call, fc->server);
+	return afs_make_call(&fc->ac, call, GFP_NOFS, async);
 }
 
 /*
@@ -986,7 +986,7 @@ static const struct afs_call_type afs_RXFSRename = {
 /*
  * create a symbolic link
  */
-int afs_fs_rename(struct afs_server *server,
+int afs_fs_rename(struct afs_fs_cursor *fc,
 		  struct key *key,
 		  struct afs_vnode *orig_dvnode,
 		  const char *orig_name,
@@ -1045,8 +1045,8 @@ int afs_fs_rename(struct afs_server *server,
 		bp = (void *) bp + n_padsz;
 	}
 
-	afs_use_fs_server(call, server);
-	return afs_make_call(&server->addr, call, GFP_NOFS, async);
+	afs_use_fs_server(call, fc->server);
+	return afs_make_call(&fc->ac, call, GFP_NOFS, async);
 }
 
 /*
@@ -1094,7 +1094,7 @@ static const struct afs_call_type afs_RXFSStoreData64 = {
 /*
  * store a set of pages to a very large file
  */
-static int afs_fs_store_data64(struct afs_server *server,
+static int afs_fs_store_data64(struct afs_fs_cursor *fc,
 			       struct afs_writeback *wb,
 			       pgoff_t first, pgoff_t last,
 			       unsigned offset, unsigned to,
@@ -1147,14 +1147,14 @@ static int afs_fs_store_data64(struct afs_server *server,
 	*bp++ = htonl(i_size >> 32);
 	*bp++ = htonl((u32) i_size);
 
-	afs_use_fs_server(call, server);
-	return afs_make_call(&server->addr, call, GFP_NOFS, async);
+	afs_use_fs_server(call, fc->server);
+	return afs_make_call(&fc->ac, call, GFP_NOFS, async);
 }
 
 /*
  * store a set of pages
  */
-int afs_fs_store_data(struct afs_server *server, struct afs_writeback *wb,
+int afs_fs_store_data(struct afs_fs_cursor *fc, struct afs_writeback *wb,
 		      pgoff_t first, pgoff_t last,
 		      unsigned offset, unsigned to,
 		      bool async)
@@ -1183,7 +1183,7 @@ int afs_fs_store_data(struct afs_server *server, struct afs_writeback *wb,
 	       (unsigned long long) i_size);
 
 	if (pos >> 32 || i_size >> 32 || size >> 32 || (pos + size) >> 32)
-		return afs_fs_store_data64(server, wb, first, last, offset, to,
+		return afs_fs_store_data64(fc, wb, first, last, offset, to,
 					   size, pos, i_size, async);
 
 	call = afs_alloc_flat_call(net, &afs_RXFSStoreData,
@@ -1221,8 +1221,8 @@ int afs_fs_store_data(struct afs_server *server, struct afs_writeback *wb,
 	*bp++ = htonl(size);
 	*bp++ = htonl(i_size);
 
-	afs_use_fs_server(call, server);
-	return afs_make_call(&server->addr, call, GFP_NOFS, async);
+	afs_use_fs_server(call, fc->server);
+	return afs_make_call(&fc->ac, call, GFP_NOFS, async);
 }
 
 /*
@@ -1279,7 +1279,7 @@ static const struct afs_call_type afs_RXFSStoreData64_as_Status = {
  * set the attributes on a very large file, using FS.StoreData rather than
  * FS.StoreStatus so as to alter the file size also
  */
-static int afs_fs_setattr_size64(struct afs_server *server, struct key *key,
+static int afs_fs_setattr_size64(struct afs_fs_cursor *fc, struct key *key,
 				 struct afs_vnode *vnode, struct iattr *attr,
 				 bool async)
 {
@@ -1319,15 +1319,15 @@ static int afs_fs_setattr_size64(struct afs_server *server, struct key *key,
 	*bp++ = htonl(attr->ia_size >> 32);	/* new file length */
 	*bp++ = htonl((u32) attr->ia_size);
 
-	afs_use_fs_server(call, server);
-	return afs_make_call(&server->addr, call, GFP_NOFS, async);
+	afs_use_fs_server(call, fc->server);
+	return afs_make_call(&fc->ac, call, GFP_NOFS, async);
 }
 
 /*
  * set the attributes on a file, using FS.StoreData rather than FS.StoreStatus
  * so as to alter the file size also
  */
-static int afs_fs_setattr_size(struct afs_server *server, struct key *key,
+static int afs_fs_setattr_size(struct afs_fs_cursor *fc, struct key *key,
 			       struct afs_vnode *vnode, struct iattr *attr,
 			       bool async)
 {
@@ -1340,8 +1340,7 @@ static int afs_fs_setattr_size(struct afs_server *server, struct key *key,
 
 	ASSERT(attr->ia_valid & ATTR_SIZE);
 	if (attr->ia_size >> 32)
-		return afs_fs_setattr_size64(server, key, vnode, attr,
-					     async);
+		return afs_fs_setattr_size64(fc, key, vnode, attr, async);
 
 	call = afs_alloc_flat_call(net, &afs_RXFSStoreData_as_Status,
 				   (4 + 6 + 3) * 4,
@@ -1367,15 +1366,15 @@ static int afs_fs_setattr_size(struct afs_server *server, struct key *key,
 	*bp++ = 0;				/* size of write */
 	*bp++ = htonl(attr->ia_size);		/* new file length */
 
-	afs_use_fs_server(call, server);
-	return afs_make_call(&server->addr, call, GFP_NOFS, async);
+	afs_use_fs_server(call, fc->server);
+	return afs_make_call(&fc->ac, call, GFP_NOFS, async);
 }
 
 /*
  * set the attributes on a file, using FS.StoreData if there's a change in file
  * size, and FS.StoreStatus otherwise
  */
-int afs_fs_setattr(struct afs_server *server, struct key *key,
+int afs_fs_setattr(struct afs_fs_cursor *fc, struct key *key,
 		   struct afs_vnode *vnode, struct iattr *attr,
 		   bool async)
 {
@@ -1384,8 +1383,7 @@ int afs_fs_setattr(struct afs_server *server, struct key *key,
 	__be32 *bp;
 
 	if (attr->ia_valid & ATTR_SIZE)
-		return afs_fs_setattr_size(server, key, vnode, attr,
-					   async);
+		return afs_fs_setattr_size(fc, key, vnode, attr, async);
 
 	_enter(",%x,{%x:%u},,",
 	       key_serial(key), vnode->fid.vid, vnode->fid.vnode);
@@ -1409,8 +1407,8 @@ int afs_fs_setattr(struct afs_server *server, struct key *key,
 
 	xdr_encode_AFS_StoreStatus(&bp, attr);
 
-	afs_use_fs_server(call, server);
-	return afs_make_call(&server->addr, call, GFP_NOFS, async);
+	afs_use_fs_server(call, fc->server);
+	return afs_make_call(&fc->ac, call, GFP_NOFS, async);
 }
 
 /*
@@ -1607,7 +1605,7 @@ static const struct afs_call_type afs_RXFSGetVolumeStatus = {
 /*
  * fetch the status of a volume
  */
-int afs_fs_get_volume_status(struct afs_server *server,
+int afs_fs_get_volume_status(struct afs_fs_cursor *fc,
 			     struct key *key,
 			     struct afs_vnode *vnode,
 			     struct afs_volume_status *vs,
@@ -1640,8 +1638,8 @@ int afs_fs_get_volume_status(struct afs_server *server,
 	bp[0] = htonl(FSGETVOLUMESTATUS);
 	bp[1] = htonl(vnode->fid.vid);
 
-	afs_use_fs_server(call, server);
-	return afs_make_call(&server->addr, call, GFP_NOFS, async);
+	afs_use_fs_server(call, fc->server);
+	return afs_make_call(&fc->ac, call, GFP_NOFS, async);
 }
 
 /*
@@ -1696,7 +1694,7 @@ static const struct afs_call_type afs_RXFSReleaseLock = {
 /*
  * get a lock on a file
  */
-int afs_fs_set_lock(struct afs_server *server,
+int afs_fs_set_lock(struct afs_fs_cursor *fc,
 		    struct key *key,
 		    struct afs_vnode *vnode,
 		    afs_lock_type_t type,
@@ -1723,14 +1721,14 @@ int afs_fs_set_lock(struct afs_server *server,
 	*bp++ = htonl(vnode->fid.unique);
 	*bp++ = htonl(type);
 
-	afs_use_fs_server(call, server);
-	return afs_make_call(&server->addr, call, GFP_NOFS, async);
+	afs_use_fs_server(call, fc->server);
+	return afs_make_call(&fc->ac, call, GFP_NOFS, async);
 }
 
 /*
  * extend a lock on a file
  */
-int afs_fs_extend_lock(struct afs_server *server,
+int afs_fs_extend_lock(struct afs_fs_cursor *fc,
 		       struct key *key,
 		       struct afs_vnode *vnode,
 		       bool async)
@@ -1755,14 +1753,14 @@ int afs_fs_extend_lock(struct afs_server *server,
 	*bp++ = htonl(vnode->fid.vnode);
 	*bp++ = htonl(vnode->fid.unique);
 
-	afs_use_fs_server(call, server);
-	return afs_make_call(&server->addr, call, GFP_NOFS, async);
+	afs_use_fs_server(call, fc->server);
+	return afs_make_call(&fc->ac, call, GFP_NOFS, async);
 }
 
 /*
  * release a lock on a file
  */
-int afs_fs_release_lock(struct afs_server *server,
+int afs_fs_release_lock(struct afs_fs_cursor *fc,
 			struct key *key,
 			struct afs_vnode *vnode,
 			bool async)
@@ -1787,8 +1785,8 @@ int afs_fs_release_lock(struct afs_server *server,
 	*bp++ = htonl(vnode->fid.vnode);
 	*bp++ = htonl(vnode->fid.unique);
 
-	afs_use_fs_server(call, server);
-	return afs_make_call(&server->addr, call, GFP_NOFS, async);
+	afs_use_fs_server(call, fc->server);
+	return afs_make_call(&fc->ac, call, GFP_NOFS, async);
 }
 
 /*
@@ -1812,6 +1810,7 @@ static const struct afs_call_type afs_RXFSGiveUpAllCallBacks = {
  * Flush all the callbacks we have on a server.
  */
 int afs_fs_give_up_all_callbacks(struct afs_server *server,
+				 struct afs_addr_cursor *ac,
 				 struct key *key,
 				 bool async)
 {
@@ -1831,5 +1830,5 @@ int afs_fs_give_up_all_callbacks(struct afs_server *server,
 	*bp++ = htonl(FSGIVEUPALLCALLBACKS);
 
 	/* Can't take a ref on server */
-	return afs_make_call(&server->addr, call, GFP_NOFS, async);
+	return afs_make_call(ac, call, GFP_NOFS, async);
 }
diff --git a/fs/afs/internal.h b/fs/afs/internal.h
index 51e3825b5ffb..df52bf18a263 100644
--- a/fs/afs/internal.h
+++ b/fs/afs/internal.h
@@ -71,6 +71,17 @@ enum afs_call_state {
 };
 
 /*
+ * List of server addresses.
+ */
+struct afs_addr_list {
+	struct rcu_head		rcu;		/* Must be first */
+	refcount_t		usage;
+	unsigned short		nr_addrs;
+	unsigned short		index;		/* Address currently in use */
+	struct sockaddr_rxrpc	addrs[];
+};
+
+/*
  * a record of an in-progress RxRPC call
  */
 struct afs_call {
@@ -283,16 +294,15 @@ struct afs_cell {
 #define AFS_CELL_FL_NO_GC	1		/* The cell was added manually, don't auto-gc */
 #define AFS_CELL_FL_NOT_FOUND	2		/* Permanent DNS error */
 #define AFS_CELL_FL_DNS_FAIL	3		/* Failed to access DNS */
+#define AFS_CELL_FL_NO_LOOKUP_YET 4		/* Not completed first DNS lookup yet */
 	enum afs_cell_state	state;
 	short			error;
 
 	spinlock_t		vl_lock;	/* vl_list lock */
 
 	/* VLDB server list. */
-	seqlock_t		vl_addrs_lock;
-	unsigned short		vl_naddrs;	/* number of VL servers in addr list */
-	unsigned short		vl_curr_svix;	/* current server index */
-	struct sockaddr_rxrpc	vl_addrs[AFS_CELL_MAX_ADDRS];	/* cell VL server addresses */
+	rwlock_t		vl_addrs_lock;	/* Lock on vl_addrs */
+	struct afs_addr_list	__rcu *vl_addrs; /* List of VL servers */
 	u8			name_len;	/* Length of name */
 	char			name[64 + 1];	/* Cell name, case-flattened and NUL-padded */
 };
@@ -343,7 +353,7 @@ struct afs_vlocation {
 struct afs_server {