afs: Overhaul volume and server record caching and fileserver rotation

The current code assumes that volumes and servers are per-cell and are
never shared, but this is not enforced, and, indeed, public cells do exist
that are aliases of each other.  Further, an organisation can, say, set up
a public cell and a private cell with overlapping, but not identical, sets
of servers.  The difference is purely in the database attached to the VL
servers.

The current code will malfunction if it sees a server in two cells as it
assumes global address -> server record mappings and that each server is in
just one cell.

Further, each server may have multiple addresses - and may have addresses
of different families (IPv4 and IPv6, say).

To this end, the following structural changes are made:

 (1) Server record management is overhauled:

     (a) Server records are made independent of cell.  The namespace keeps
     	 track of them, volume records have lists of them and each vnode
     	 has a server on which its callback interest currently resides.

     (b) The cell record no longer keeps a list of servers known to be in
     	 that cell.

     (c) The server records are now kept in a flat list because there's no
     	 single address to sort on.

     (d) Server records are now keyed by their UUID within the namespace.

     (e) The addresses for a server are obtained with the VL.GetAddrsU
     	 rather than with VL.GetEntryByName, using the server's UUID as a
     	 parameter.

     (f) Cached server records are garbage collected after a period of
     	 non-use and are counted out of existence before purging is allowed
     	 to complete.  This protects the work functions against rmmod.

     (g) The servers list is now in /proc/fs/afs/servers.

 (2) Volume record management is overhauled:

     (a) An RCU-replaceable server list is introduced.  This tracks both
     	 servers and their coresponding callback interests.

     (b) The superblock is now keyed on cell record and numeric volume ID.

     (c) The volume record is now tied to the superblock which mounts it,
     	 and is activated when mounted and deactivated when unmounted.
     	 This makes it easier to handle the cache cookie without causing a
     	 double-use in fscache.

     (d) The volume record is loaded from the VLDB using VL.GetEntryByNameU
     	 to get the server UUID list.

     (e) The volume name is updated if it is seen to have changed when the
     	 volume is updated (the update is keyed on the volume ID).

 (3) The vlocation record is got rid of and VLDB records are no longer
     cached.  Sufficient information is stored in the volume record, though
     an update to a volume record is now no longer shared between related
     volumes (volumes come in bundles of three: R/W, R/O and backup).

and the following procedural changes are made:

 (1) The fileserver cursor introduced previously is now fleshed out and
     used to iterate over fileservers and their addresses.

 (2) Volume status is checked during iteration, and the server list is
     replaced if a change is detected.

 (3) Server status is checked during iteration, and the address list is
     replaced if a change is detected.

 (4) The abort code is saved into the address list cursor and -ECONNABORTED
     returned in afs_make_call() if a remote abort happened rather than
     translating the abort into an error message.  This allows actions to
     be taken depending on the abort code more easily.

     (a) If a VMOVED abort is seen then this is handled by rechecking the
     	 volume and restarting the iteration.

     (b) If a VBUSY, VRESTARTING or VSALVAGING abort is seen then this is
         handled by sleeping for a short period and retrying and/or trying
         other servers that might serve that volume.  A message is also
         displayed once until the condition has cleared.

     (c) If a VOFFLINE abort is seen, then this is handled as VBUSY for the
     	 moment.

     (d) If a VNOVOL abort is seen, the volume is rechecked in the VLDB to
     	 see if it has been deleted; if not, the fileserver is probably
     	 indicating that the volume couldn't be attached and needs
     	 salvaging.

     (e) If statfs() sees one of these aborts, it does not sleep, but
     	 rather returns an error, so as not to block the umount program.

 (5) The fileserver iteration functions in vnode.c are now merged into
     their callers and more heavily macroised around the cursor.  vnode.c
     is removed.

 (6) Operations on a particular vnode are serialised on that vnode because
     the server will lock that vnode whilst it operates on it, so a second
     op sent will just have to wait.

 (7) Fileservers are probed with FS.GetCapabilities before being used.
     This is where service upgrade will be done.

 (8) A callback interest on a fileserver is set up before an FS operation
     is performed and passed through to afs_make_call() so that it can be
     set on the vnode if the operation returns a callback.  The callback
     interest is passed through to afs_iget() also so that it can be set
     there too.

In general, record updating is done on an as-needed basis when we try to
access servers, volumes or vnodes rather than offloading it to work items
and special threads.

Notes:

 (1) Pre AFS-3.4 servers are no longer supported, though this can be added
     back if necessary (AFS-3.4 was released in 1998).

 (2) VBUSY is retried forever for the moment at intervals of 1s.

 (3) /proc/fs/afs/<cell>/servers no longer exists.

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

1 files changed, 213 insertions, 94 deletions

diff --git a/fs/afs/vlclient.c b/fs/afs/vlclient.c
index 1d1e7df77dd5..173c652fe875 100644
--- a/fs/afs/vlclient.c
+++ b/fs/afs/vlclient.c
@@ -16,14 +16,15 @@
 #include "internal.h"
 
 /*
- * deliver reply data to a VL.GetEntryByXXX call
+ * Deliver reply data to a VL.GetEntryByNameU call.
  */
-static int afs_deliver_vl_get_entry_by_xxx(struct afs_call *call)
+static int afs_deliver_vl_get_entry_by_name_u(struct afs_call *call)
 {
-	struct afs_cache_vlocation *entry;
-	__be32 *bp;
+	struct afs_uvldbentry__xdr *uvldb;
+	struct afs_vldb_entry *entry;
+	bool new_only = false;
 	u32 tmp;
-	int loop, ret;
+	int i, ret;
 
 	_enter("");
 
@@ -32,152 +33,270 @@ static int afs_deliver_vl_get_entry_by_xxx(struct afs_call *call)
 		return ret;
 
 	/* unmarshall the reply once we've received all of it */
+	uvldb = call->buffer;
 	entry = call->reply[0];
-	bp = call->buffer;
-
-	for (loop = 0; loop < 64; loop++)
-		entry->name[loop] = ntohl(*bp++);
-	entry->name[loop] = 0;
-	bp++; /* final NUL */
-
-	bp++; /* type */
-	entry->nservers = ntohl(*bp++);
-
-	for (loop = 0; loop < 8; loop++) {
-		entry->servers[loop].srx_family = AF_RXRPC;
-		entry->servers[loop].srx_service = FS_SERVICE;
-		entry->servers[loop].transport_type = SOCK_DGRAM;
-		entry->servers[loop].transport_len = sizeof(entry->servers[loop].transport.sin6);
-		entry->servers[loop].transport.sin6.sin6_family = AF_INET6;
-		entry->servers[loop].transport.sin6.sin6_port = htons(AFS_FS_PORT);
-		entry->servers[loop].transport.sin6.sin6_flowinfo = 0;
-		entry->servers[loop].transport.sin6.sin6_scope_id = 0;
-		entry->servers[loop].transport.sin6.sin6_addr.s6_addr32[0] = 0;
-		entry->servers[loop].transport.sin6.sin6_addr.s6_addr32[1] = 0;
-		entry->servers[loop].transport.sin6.sin6_addr.s6_addr32[2] = htonl(0xffff);
-		entry->servers[loop].transport.sin6.sin6_addr.s6_addr32[3] = *bp++;
+
+	for (i = 0; i < ARRAY_SIZE(uvldb->name) - 1; i++)
+		entry->name[i] = (u8)ntohl(uvldb->name[i]);
+	entry->name[i] = 0;
+	entry->name_len = strlen(entry->name);
+
+	/* If there is a new replication site that we can use, ignore all the
+	 * sites that aren't marked as new.
+	 */
+	for (i = 0; i < AFS_NMAXNSERVERS; i++) {
+		tmp = ntohl(uvldb->serverFlags[i]);
+		if (!(tmp & AFS_VLSF_DONTUSE) &&
+		    (tmp & AFS_VLSF_NEWREPSITE))
+			new_only = true;
 	}
 
-	bp += 8; /* partition IDs */
+	for (i = 0; i < AFS_NMAXNSERVERS; i++) {
+		struct afs_uuid__xdr *xdr;
+		struct afs_uuid *uuid;
+		int j;
 
-	for (loop = 0; loop < 8; loop++) {
-		tmp = ntohl(*bp++);
-		entry->srvtmask[loop] = 0;
+		tmp = ntohl(uvldb->serverFlags[i]);
+		if (tmp & AFS_VLSF_DONTUSE ||
+		    (new_only && !(tmp & AFS_VLSF_NEWREPSITE)))
+			continue;
 		if (tmp & AFS_VLSF_RWVOL)
-			entry->srvtmask[loop] |= AFS_VOL_VTM_RW;
+			entry->fs_mask[i] |= AFS_VOL_VTM_RW;
 		if (tmp & AFS_VLSF_ROVOL)
-			entry->srvtmask[loop] |= AFS_VOL_VTM_RO;
+			entry->fs_mask[i] |= AFS_VOL_VTM_RO;
 		if (tmp & AFS_VLSF_BACKVOL)
-			entry->srvtmask[loop] |= AFS_VOL_VTM_BAK;
-	}
+			entry->fs_mask[i] |= AFS_VOL_VTM_BAK;
+		if (!entry->fs_mask[i])
+			continue;
+
+		xdr = &uvldb->serverNumber[i];
+		uuid = (struct afs_uuid *)&entry->fs_server[i];
+		uuid->time_low			= xdr->time_low;
+		uuid->time_mid			= htons(ntohl(xdr->time_mid));
+		uuid->time_hi_and_version	= htons(ntohl(xdr->time_hi_and_version));
+		uuid->clock_seq_hi_and_reserved	= (u8)ntohl(xdr->clock_seq_hi_and_reserved);
+		uuid->clock_seq_low		= (u8)ntohl(xdr->clock_seq_low);
+		for (j = 0; j < 6; j++)
+			uuid->node[j] = (u8)ntohl(xdr->node[j]);
 
-	entry->vid[0] = ntohl(*bp++);
-	entry->vid[1] = ntohl(*bp++);
-	entry->vid[2] = ntohl(*bp++);
+		entry->nr_servers++;
+	}
 
-	bp++; /* clone ID */
+	for (i = 0; i < AFS_MAXTYPES; i++)
+		entry->vid[i] = ntohl(uvldb->volumeId[i]);
 
-	tmp = ntohl(*bp++); /* flags */
-	entry->vidmask = 0;
+	tmp = ntohl(uvldb->flags);
 	if (tmp & AFS_VLF_RWEXISTS)
-		entry->vidmask |= AFS_VOL_VTM_RW;
+		__set_bit(AFS_VLDB_HAS_RW, &entry->flags);
 	if (tmp & AFS_VLF_ROEXISTS)
-		entry->vidmask |= AFS_VOL_VTM_RO;
+		__set_bit(AFS_VLDB_HAS_RO, &entry->flags);
 	if (tmp & AFS_VLF_BACKEXISTS)
-		entry->vidmask |= AFS_VOL_VTM_BAK;
-	if (!entry->vidmask)
-		return -EBADMSG;
+		__set_bit(AFS_VLDB_HAS_BAK, &entry->flags);
 
+	if (!(tmp & (AFS_VLF_RWEXISTS | AFS_VLF_ROEXISTS | AFS_VLF_BACKEXISTS))) {
+		entry->error = -ENOMEDIUM;
+		__set_bit(AFS_VLDB_QUERY_ERROR, &entry->flags);
+	}
+
+	__set_bit(AFS_VLDB_QUERY_VALID, &entry->flags);
 	_leave(" = 0 [done]");
 	return 0;
 }
 
-/*
- * VL.GetEntryByName operation type
- */
-static const struct afs_call_type afs_RXVLGetEntryByName = {
-	.name		= "VL.GetEntryByName",
-	.deliver	= afs_deliver_vl_get_entry_by_xxx,
-	.destructor	= afs_flat_call_destructor,
-};
+static void afs_destroy_vl_get_entry_by_name_u(struct afs_call *call)
+{
+	kfree(call->reply[0]);
+	afs_flat_call_destructor(call);
+}
 
 /*
- * VL.GetEntryById operation type
+ * VL.GetEntryByNameU operation type.
  */
-static const struct afs_call_type afs_RXVLGetEntryById = {
-	.name		= "VL.GetEntryById",
-	.deliver	= afs_deliver_vl_get_entry_by_xxx,
-	.destructor	= afs_flat_call_destructor,
+static const struct afs_call_type afs_RXVLGetEntryByNameU = {
+	.name		= "VL.GetEntryByNameU",
+	.deliver	= afs_deliver_vl_get_entry_by_name_u,
+	.destructor	= afs_destroy_vl_get_entry_by_name_u,
 };
 
 /*
- * dispatch a get volume entry by name operation
+ * Dispatch a get volume entry by name or ID operation (uuid variant).  If the
+ * volname is a decimal number then it's a volume ID not a volume name.
  */
-int afs_vl_get_entry_by_name(struct afs_net *net,
-			     struct afs_addr_cursor *ac,
-			     struct key *key,
-			     const char *volname,
-			     struct afs_cache_vlocation *entry,
-			     bool async)
+struct afs_vldb_entry *afs_vl_get_entry_by_name_u(struct afs_net *net,
+						  struct afs_addr_cursor *ac,
+						  struct key *key,
+						  const char *volname,
+						  int volnamesz)
 {
+	struct afs_vldb_entry *entry;
 	struct afs_call *call;
-	size_t volnamesz, reqsz, padsz;
+	size_t reqsz, padsz;
 	__be32 *bp;
 
 	_enter("");
 
-	volnamesz = strlen(volname);
 	padsz = (4 - (volnamesz & 3)) & 3;
 	reqsz = 8 + volnamesz + padsz;
 
-	call = afs_alloc_flat_call(net, &afs_RXVLGetEntryByName, reqsz, 384);
-	if (!call)
-		return -ENOMEM;
+	entry = kzalloc(sizeof(struct afs_vldb_entry), GFP_KERNEL);
+	if (!entry)
+		return ERR_PTR(-ENOMEM);
+
+	call = afs_alloc_flat_call(net, &afs_RXVLGetEntryByNameU, reqsz,
+				   sizeof(struct afs_uvldbentry__xdr));
+	if (!call) {
+		kfree(entry);
+		return ERR_PTR(-ENOMEM);
+	}
 
 	call->key = key;
 	call->reply[0] = entry;
+	call->ret_reply0 = true;
 
-	/* marshall the parameters */
+	/* Marshall the parameters */
 	bp = call->request;
-	*bp++ = htonl(VLGETENTRYBYNAME);
+	*bp++ = htonl(VLGETENTRYBYNAMEU);
 	*bp++ = htonl(volnamesz);
 	memcpy(bp, volname, volnamesz);
 	if (padsz > 0)
-		memset((void *) bp + volnamesz, 0, padsz);
+		memset((void *)bp + volnamesz, 0, padsz);
 
-	/* initiate the call */
-	return afs_make_call(ac, call, GFP_KERNEL, async);
+	return (struct afs_vldb_entry *)afs_make_call(ac, call, GFP_KERNEL, false);
 }
 
 /*
- * dispatch a get volume entry by ID operation
+ * Deliver reply data to a VL.GetAddrsU call.
+ *
+ *	GetAddrsU(IN ListAddrByAttributes *inaddr,
+ *		  OUT afsUUID *uuidp1,
+ *		  OUT uint32_t *uniquifier,
+ *		  OUT uint32_t *nentries,
+ *		  OUT bulkaddrs *blkaddrs);
  */
-int afs_vl_get_entry_by_id(struct afs_net *net,
-			   struct afs_addr_cursor *ac,
-			   struct key *key,
-			   afs_volid_t volid,
-			   afs_voltype_t voltype,
-			   struct afs_cache_vlocation *entry,
-			   bool async)
+static int afs_deliver_vl_get_addrs_u(struct afs_call *call)
 {
+	struct afs_addr_list *alist;
+	__be32 *bp;
+	u32 uniquifier, nentries, count;
+	int i, ret;
+
+	_enter("{%u,%zu/%u}", call->unmarshall, call->offset, call->count);
+
+again:
+	switch (call->unmarshall) {
+	case 0:
+		call->offset = 0;
+		call->unmarshall++;
+
+		/* Extract the returned uuid, uniquifier, nentries and blkaddrs size */
+	case 1:
+		ret = afs_extract_data(call, call->buffer,
+				       sizeof(struct afs_uuid__xdr) + 3 * sizeof(__be32),
+				       true);
+		if (ret < 0)
+			return ret;
+
+		bp = call->buffer + sizeof(struct afs_uuid__xdr);
+		uniquifier	= ntohl(*bp++);
+		nentries	= ntohl(*bp++);
+		count		= ntohl(*bp);
+
+		nentries = min(nentries, count);
+		alist = afs_alloc_addrlist(nentries, FS_SERVICE, AFS_FS_PORT);
+		if (!alist)
+			return -ENOMEM;
+		alist->version = uniquifier;
+		call->reply[0] = alist;
+		call->count = count;
+		call->count2 = nentries;
+		call->offset = 0;
+		call->unmarshall++;
+
+		/* Extract entries */
+	case 2:
+		count = min(call->count, 4U);
+		ret = afs_extract_data(call, call->buffer,
+				       count * sizeof(__be32),
+				       call->count > 4);
+		if (ret < 0)
+			return ret;
+
+		alist = call->reply[0];
+		bp = call->buffer;
+		for (i = 0; i < count; i++)
+			if (alist->nr_addrs < call->count2)
+				afs_merge_fs_addr4(alist, *bp++);
+
+		call->count -= count;
+		if (call->count > 0)
+			goto again;
+		call->offset = 0;
+		call->unmarshall++;
+		break;
+	}
+
+	_leave(" = 0 [done]");
+	return 0;
+}
+
+static void afs_vl_get_addrs_u_destructor(struct afs_call *call)
+{
+	afs_put_server(call->net, (struct afs_server *)call->reply[0]);
+	kfree(call->reply[1]);
+	return afs_flat_call_destructor(call);
+}
+
+/*
+ * VL.GetAddrsU operation type.
+ */
+static const struct afs_call_type afs_RXVLGetAddrsU = {
+	.name		= "VL.GetAddrsU",
+	.deliver	= afs_deliver_vl_get_addrs_u,
+	.destructor	= afs_vl_get_addrs_u_destructor,
+};
+
+/*
+ * Dispatch an operation to get the addresses for a server, where the server is
+ * nominated by UUID.
+ */
+struct afs_addr_list *afs_vl_get_addrs_u(struct afs_net *net,
+					 struct afs_addr_cursor *ac,
+					 struct key *key,
+					 const uuid_t *uuid)
+{
+	struct afs_ListAddrByAttributes__xdr *r;
+	const struct afs_uuid *u = (const struct afs_uuid *)uuid;
 	struct afs_call *call;
 	__be32 *bp;
+	int i;
 
 	_enter("");
 
-	call = afs_alloc_flat_call(net, &afs_RXVLGetEntryById, 12, 384);
+	call = afs_alloc_flat_call(net, &afs_RXVLGetAddrsU,
+				   sizeof(__be32) + sizeof(struct afs_ListAddrByAttributes__xdr),
+				   sizeof(struct afs_uuid__xdr) + 3 * sizeof(__be32));
 	if (!call)
-		return -ENOMEM;
+		return ERR_PTR(-ENOMEM);
 
 	call->key = key;
-	call->reply[0] = entry;
+	call->reply[0] = NULL;
+	call->ret_reply0 = true;
 
-	/* marshall the parameters */
+	/* Marshall the parameters */
 	bp = call->request;
-	*bp++ = htonl(VLGETENTRYBYID);
-	*bp++ = htonl(volid);
-	*bp   = htonl(voltype);
+	*bp++ = htonl(VLGETADDRSU);
+	r = (struct afs_ListAddrByAttributes__xdr *)bp;
+	r->Mask		= htonl(AFS_VLADDR_UUID);
+	r->ipaddr	= 0;
+	r->index	= 0;
+	r->spare	= 0;
+	r->uuid.time_low			= u->time_low;
+	r->uuid.time_mid			= htonl(ntohs(u->time_mid));
+	r->uuid.time_hi_and_version		= htonl(ntohs(u->time_hi_and_version));
+	r->uuid.clock_seq_hi_and_reserved 	= htonl(u->clock_seq_hi_and_reserved);
+	r->uuid.clock_seq_low			= htonl(u->clock_seq_low);
+	for (i = 0; i < 6; i++)
+		r->uuid.node[i] = ntohl(u->node[i]);
 
-	/* initiate the call */
-	return afs_make_call(ac, call, GFP_KERNEL, async);
+	return (struct afs_addr_list *)afs_make_call(ac, call, GFP_KERNEL, false);
 }